SP2 Study Mat

Life Insurance
Specialist Principles
Combined Materials Pack
for exams in 2019
The Actuarial Education Company

on behalf of the Institute and Faculty of Actuaries
All study material produced by ActEd is copyright and is sold for
the exclusive use of the purchaser. The copyright is owned by
Institute and Faculty Education Limited, a subsidiary of the
Institute and Faculty of Actuaries.
Unless prior authority is granted by ActEd, you may not hire out,
lend, give out, sell, store or transmit electronically or photocopy
any part of the study material.
You must take care of your study material to ensure that it is

not used or copied by anybody else.
Legal action will be taken if these terms are infringed. In

addition, we may seek to take disciplinary action through the
profession or through your employer.
These conditions remain in force after you have finished using

the course.
© IFE: 2019 Examinations The Actuarial Education Company

SP2: Study Guide Page 1
Subject SP2
2019 Study Guide
Introduction
This Study Guide has been created to help guide you through Subject SP2. It contains all the
information that you will need before starting to study Subject SP2 for the 2019 exams and you
may also find it useful to refer to throughout your Subject SP2 journey.
The guide is split into two parts:

 Part 1 contains general information about the Specialist Principles subjects
 Part 2 contains specific information about Subject SP2.
Please read this Study Guide carefully before reading the Course Notes, even if you have studied
for some actuarial exams before.
Contents
Part 1 Section 1 Before you start Page 2

Section 2 Core study material Page 3
Section 3 ActEd study support Page 5
Section 4 Study skills Page 11
Section 5 The examination Page 16
Section 6 Queries and feedback Page 17
Part 2 Section 1 Subject SP2 – background Page 18
Section 2 Subject SP2 – Syllabus and Core Reading Page 19
Section 3 Subject SP2 – the course structure Page 27
Section 4 Subject SP2 – summary of ActEd products Page 29
Section 5 Subject SP2 – skills and assessment Page 30
Section 6 Subject SP2 – frequently asked questions Page 31
The Actuarial Education Company © IFE: 2019 Examinations

Page 2 SP2: Study Guide
1.1 Before you start

When studying for the UK actuarial exams, you will need:
 a copy of the Formulae and Tables for Examinations of the Faculty of Actuaries and the
Institute of Actuaries, 2nd Edition (2002) – these are often referred to as simply the
Yellow Tables or the Tables
 a ‘permitted’ scientific calculator – you will find the list of permitted calculators on the
profession’s website. Please check the list carefully, since it is reviewed each year.
These are both available from the Institute and Faculty of Actuaries’ eShop. Please visit
www.actuaries.org.uk.

1.2 Core study material

This section explains the role of the Syllabus, Core Reading and supplementary ActEd text. It also
gives guidance on how to use these materials most effectively in order to pass the exam.
Some of the information below is also contained in the introduction to the Core Reading
produced by the Institute and Faculty of Actuaries.
Syllabus
The Syllabus for Subject SP2 has been produced by the Institute and Faculty of Actuaries. The
relevant individual Syllabus Objectives are included at the start of each course chapter and a
complete copy of the Syllabus is included in Section 2.2 of this Study Guide. We recommend that
you use the Syllabus as an important part of your study.
Core Reading
The Core Reading has been produced by the Institute and Faculty of Actuaries. The purpose of
the Core Reading is to ensure that tutors, students and examiners understand the requirements
of the syllabus for the qualification examinations for Fellowship of the Institute and Faculty of
Actuaries.
The Core Reading supports coverage of the Syllabus in helping to ensure that both depth and
breadth are re-enforced. It is therefore important that students have a good understanding of
the concepts covered by the Core Reading.
The examinations require students to demonstrate their understanding of the concepts given in
the syllabus and described in the Core Reading; this will be based on the legislation, professional
guidance etc that are in force when the Core Reading is published, ie on 31 May in the year
preceding the examinations.
Therefore the exams in April and September 2019 will be based on the Syllabus and Core Reading
as at 31 May 2018. We recommend that you always use the up-to-date Core Reading to prepare
for the exams.
Examiners will have this Core Reading when setting the papers. In preparing for examinations,
students are advised to work through past examination questions and may find additional tuition
helpful. The Core Reading will be updated each year to reflect changes in the syllabus and current
practice, and in the interest of clarity.
Accreditation
The Institute and Faculty of Actuaries would like to thank the numerous people who have helped
in the development of the material contained in this Core Reading.

ActEd text
Core Reading deals with each syllabus objective and covers what is needed to pass the exam.
However, the tuition material that has been written by ActEd enhances it by giving examples and
further explanation of key points. Here is an excerpt from some ActEd Course Notes to show you
how to identify Core Reading and the ActEd material. Core Reading is shown in this bold font.
Note that in the example given above, the index will fall if the actual share price goes below the
theoretical ex-rights share price. Again, this is consistent with what would happen to an
underlying portfolio.
This is
After allowing for chain-linking, the formula for the investment index then becomes: ActEd
text
 Ni ,t Pi ,t
I (t )  i This is Core
B(t ) Reading
where Ni ,t is the number of shares issued for the ith constituent at time t;
B (t ) is the base value, or divisor, at time t.
Copyright
All study material produced by ActEd is copyright and is sold for the exclusive use of the
purchaser. The copyright is owned by Institute and Faculty Education Limited, a subsidiary of the
Institute and Faculty of Actuaries. Unless prior authority is granted by ActEd, you may not hire
out, lend, give out, sell, store or transmit electronically or photocopy any part of the study
material. You must take care of your study material to ensure that it is not used or copied by
anybody else.
Legal action will be taken if these terms are infringed. In addition, we may seek to take
disciplinary action through the Institute and Faculty of Actuaries or through your employer.
These conditions remain in force after you have finished using the course.

1.3 ActEd study support

This section gives a description of the products offered by ActEd.
Successful students tend to undertake three main study activities:

1. Learning – initial study and understanding of subject material
2. Revision – learning subject material and preparing to tackle exam-style questions
3. Rehearsal – answering exam-style questions, culminating in answering questions at exam
speed without notes.
Different approaches suit different people. For example, you may like to learn material gradually
over the months running up to the exams or you may do your revision in a shorter period just
before the exams. Also, these three activities will almost certainly overlap.
We offer a flexible range of products to suit you and let you control your own learning and exam
preparation. The following table shows the products that we produce. Note that not all products
are available for all subjects.
LEARNING LEARNING & REVISION REVISION & REHEARSAL

REVISION REHEARSAL
Course Notes X Assignments Flashcards Revision Notes Mock Exam
Combined ASET Mock Exam

Materials Pack Marking
(CMP)
X Assignment
Marking
Tutorials
Online
Classroom
The products and services are described in more detail below.

‘Learning’ products
Course Notes
The Course Notes will help you develop the basic knowledge and understanding of principles
needed to pass the exam. They incorporate the complete Core Reading and include full
explanation of all the syllabus objectives, with worked examples and questions (including some
past exam questions) to test your understanding.
Each chapter includes:

 the relevant syllabus objectives
 a chapter summary
 practice questions with full solutions.
‘Learning & revision’ products

X Assignments
The Series X Assignments are written assessments that cover the material in each part of the
course in turn. They can be used to both develop and test your understanding of the material.
Combined Materials Pack (CMP)
The Combined Materials Pack (CMP) comprises the Course Notes and the Series X Assignments.
The CMP is available in eBook format for viewing on a range of electronic devices. eBooks can be
ordered separately or as an addition to paper products. Visit www.ActEd.co.uk for full details
about the eBooks that are available, compatibility with different devices, software requirements
and printing restrictions.
X Assignment Marking
We are happy to mark your attempts at the X assignments. Marking is not included with the
Assignments or the CMP and you need to order it separately. You should submit your script as a
PDF attached to an email. Your script will be marked electronically and you will be able to download
your marked script via a secure link on the internet.
Don’t underestimate the benefits of doing and submitting assignments:

 Question practice during this phase of your study gives an early focus on the end goal of
answering exam-style questions.
 You’re incentivised to keep up with your study plan and get a regular, realistic assessment
of your progress.
 Objective, personalised feedback from a high quality marker will highlight areas on which
to work and help with exam technique.
In a recent study, we found that students who attempt more than half the assignments have
significantly higher pass rates.

There are two different types of marking product: Series Marking and Marking Vouchers.
Series Marking
Series Marking applies to a specified subject, session and student. If you purchase Series Marking,
you will not be able to defer the marking to a future exam sitting or transfer it to a different subject
or student.
We typically send out full solutions with the Series X Assignments. However, if you order Series
Marking at the same time as you order the Series X Assignments, you can choose whether or not
to receive a copy of the solutions in advance. If you choose not to receive them with the study
material, you will be able to download the solutions via a secure link on the internet when your
marked script is returned (or following the final deadline date if you do not submit a script).
If you are having your attempts at the assignments marked by ActEd, you should submit your scripts
regularly throughout the session, in accordance with the schedule of recommended dates set out in
information provided with the assignments. This will help you to pace your study throughout the
session and leave an adequate amount of time for revision and question practice.
The recommended submission dates are realistic targets for the majority of students. Your scripts
will be returned more quickly if you submit them well before the final deadline dates.
Any script submitted after the relevant final deadline date will not be marked. It is your
responsibility to ensure that we receive scripts in good time.
Marking Vouchers
Marking Vouchers give the holder the right to submit a script for marking at any time, irrespective of
the individual assignment deadlines, study session, subject or person.
Marking Vouchers can be used for any assignment. They are valid for four years from the date of
purchase and can be refunded at any time up to the expiry date.
Although you may submit your script with a Marking Voucher at any time, you will need to adhere
to the explicit Marking Voucher deadline dates to ensure that your script is returned before the date
of the exam. The deadline dates are provided with the assignments.
Tutorials
Our tutorials are specifically designed to develop the knowledge that you will acquire from the
course material into the higher-level understanding that is needed to pass the exam.
We run a range of different tutorials including face-to-face tutorials at various locations, and Live
Online tutorials. Full details are set out in our Tuition Bulletin, which is available on our website at
www.ActEd.co.uk.
Regular and Block Tutorials
In preparation for these tutorials, we expect you to have read the relevant part(s) of the Course
Notes before attending the tutorial so that the group can spend time on exam questions and
discussion to develop understanding rather than basic bookwork.

You can choose one of the following types of tutorial:

 Regular Tutorials spread over the session.
 A Block Tutorial held two to eight weeks before the exam.
Online Classroom
The Online Classroom acts as either a valuable add-on or a great alternative to a face-to-face or
Live Online tutorial.
At the heart of the Online Classroom in each subject is a comprehensive, easily-searched collection
of tutorial units. These are a mix of:
 teaching units, helping you to really get to grips with the course material, and
 guided questions, enabling you to learn the most efficient ways to answer questions and
avoid common exam pitfalls.
The best way to discover the Online Classroom is to see it in action. You can watch a sample of
the Online Classroom tutorial units on our website at www.ActEd.co.uk.
‘Revision’ products
For most subjects, there is a lot of material to revise. Finding a way to fit revision into your
routine as painlessly as possible has got to be a good strategy.
Flashcards
Flashcards are an inexpensive option that can provide a massive boost. They can also provide a
variation in activities during a study day, and so help you to maintain concentration and
effectiveness.
Flashcards are a set of A6-sized cards that cover the key points of the subject that most students
want to commit to memory. Each flashcard has questions on one side and the answers on the
reverse. We recommend that you use the cards actively and test yourself as you go.
Flashcards are available in eBook format for viewing on a range of electronic devices. eBooks can
be ordered separately or as an addition to paper products. Visit www.ActEd.co.uk for full details
about the eBooks that are available, compatibility with different devices, software requirements
and printing restrictions.
The following questions and comments might help you to decide if flashcards are suitable for you:
Flashcards
 Do you have a regular train or bus journey?

Flashcards are ideal for regular bursts of revision on the move.
 Do you want to fit more study into your routine?

Flashcards are a good option for ‘dead time’, eg using flashcards on your phone or sticking
them on the wall in your study.

 Do you find yourself cramming for exams (even if that’s not your original plan)?
Flashcards are an extremely efficient way to do your pre-exam memorising.
If you are retaking a subject, then you might consider using flashcards if you didn’t use them on a
previous attempt.
‘Revision & rehearsal’ products

Revision Notes
Our Revision Notes have been designed with input from students to help you revise efficiently.
They are suitable for first-time sitters who have worked through the ActEd Course Notes or for
retakers (who should find them much more useful and challenging than simply reading through
the course again).
The Revision Notes are a set of A5 booklets – perfect for revising on the train or tube to work.
Each booklet covers one main theme or a set of related topics from the course and includes:
 Core Reading with a set of integrated short questions to develop your bookwork
knowledge
 relevant past exam questions with concise solutions from the last ten years
 other useful revision aids.
ActEd Solutions with Exam Technique (ASET)
The ActEd Solutions with Exam Technique (ASET) contains our solutions to eight past exam
papers, plus comment and explanation. In particular. it highlights how questions might have been
analysed and interpreted so as to produce a good solution with a wide range of relevant points.
This will be valuable in approaching questions in subsequent examinations.
‘Rehearsal’ products
Mock Exam
The Mock Exam is a 100-mark mock exam paper that provides a realistic test of your exam
preparation.
Mock Marking
We are happy to mark your attempts at the mock exams. The same general principles apply as for
the X Assignment Marking. In particular:
 Mock Exam Marking is available for the Mock Exam and it applies to a specified subject,
session and student
 Marking Vouchers can be used for the Mock Exam.

Recall that:
 marking is not included with the products themselves and you need to order it separately
 you should submit your script as a PDF attached to an email
 your script will be marked electronically and you will be able to download your marked
script via a secure link on the internet.

1.4 Skills
The Subject SP exams

It is important to recognise that the SP subject exams are very different from the Core Principles
subject exams in both the nature of the material covered and the skills being examined.
Both the Core Reading and the exam papers themselves are generally much less numerical and
more ‘wordy’ than the Core Principles subjects. The exam will primarily require you to explain a
particular point in words and sentences, rather than to manipulate formulae or do calculations.
Numerical questions typically account for only a small part of each exam paper. If you haven’t sat
this type of exam for some time, you need to start practising again now. Many students find that
it takes time to adjust to the different style of the SP subject exam questions. As ever, practice is
the key to success.
The aim of the exams is to test your ability to apply your knowledge and understanding of the key
principles described in the Core Reading to specific situations presented to you in the form of
exam questions. Therefore your aim should be to identify and understand the key principles, and
then to practise applying them. You will also need a good knowledge of the Core Reading to score
well and quickly on any bookwork questions.
Study skills
Overall study plan
We suggest that you develop a realistic study plan, building in time for relaxation and allowing
some time for contingencies. Be aware of busy times at work, when you may not be able to take
as much study leave as you would like. Once you have set your plan, be determined to stick to it.
You don’t have to be too prescriptive at this stage about what precisely you do on each study day.
The main thing is to be clear that you will cover all the important activities in an appropriate
manner and leave plenty of time for revision and question practice.
Aim to manage your study so as to allow plenty of time for the concepts you meet in this course
to ‘bed down’ in your mind. Most successful students will probably aim to complete the course at
least six weeks before the exam, thereby leaving a sufficient amount of time for revision. By
finishing the course as quickly as possible, you will have a much clearer view of the big picture. It
will also allow you to structure your revision so that you can concentrate on the important and
difficult areas of the course.
You can also try looking at our discussion forum on the internet, which can be accessed at
www.ActEd.co.uk/forums (or use the link from our home page at www.ActEd.co.uk). There are
some good suggestions from students on how to study.
Study sessions
Only do activities that will increase your chance of passing. Try to avoid including activities for the
sake of it and don’t spend time reviewing material that you already understand. You will only
improve your chances of passing the exam by getting on top of the material that you currently
find difficult.

In particular, you may already be familiar with the content of some of the chapters (from the Core
Principles (CS, CM or CB subjects), Subject CP1 or other SP subjects). Try to cover these chapters
quickly to give yourself more time on the material with which you are less comfortable. Where
chapters refer back to material from the Core Principles subjects, you don’t have to follow these
links up unless you are feeling curious or clueless.
Ideally, each study session should have a specific purpose and be based on a specific task,
eg ’Finish reading Chapter 3 and attempt Practice Questions 1.4, 1.7 and 1.12 ’, as opposed to a
specific amount of time, eg ‘Three hours studying the material in Chapter 3’.
Try to study somewhere quiet and free from distractions (eg a library or a desk at home dedicated
to study). Find out when you operate at your peak, and endeavour to study at those times of the
day. This might be between 8am and 10am or could be in the evening. Take short breaks during
your study to remain focused – it’s definitely time for a short break if you find that your brain is
tired and that your concentration has started to drift from the information in front of you.
Order of study
We suggest that you work through each of the chapters in turn. To get the maximum benefit from
each chapter you should proceed in the following order:
1. Read the Syllabus Objectives. These are set out in the box at the start of each chapter.
2. Read the Chapter Summary at the end of each chapter. This will give you a useful overview
of the material that you are about to study and help you to appreciate the context of the
ideas that you meet.
3. Study the Course Notes in detail, annotating them and possibly making your own notes. Try
the self-assessment questions as you come to them. As you study, pay particular attention
to the listing of the Syllabus Objectives and to the Core Reading.
4. Read the Chapter Summary again carefully. If there are any ideas that you can’t
remember covering in the Course Notes, read the relevant section of the notes again to
refresh your memory.
5. Attempt (at least some of) the Practice Questions that appear at the end of the chapter.
It’s a fact that people are more likely to remember something if they review it several times. So,
do look over the chapters you have studied so far from time to time. It is useful to re-read the
Chapter Summaries or to try the Practice Questions again a few days after reading the chapter
itself. It’s a good idea to annotate the questions with details of when you attempted each one. This
makes it easier to ensure that you try all of the questions as part of your revision without repeating
any that you got right first time.
Once you’ve read the relevant part of the notes and tried a selection of questions from the
Practice Questions (and attended a tutorial, if appropriate) you should attempt the corresponding
assignment. If you submit your assignment for marking, spend some time looking through it
carefully when it is returned. It can seem a bit depressing to analyse the errors you made, but
you will increase your chances of passing the exam by learning from your mistakes. The markers
will try their best to provide practical comments to help you to improve.

To be really prepared for the exam, you should not only know and understand the Core Reading but
also be aware of what the examiners will expect. Your revision programme should include plenty of
question practice so that you are aware of the typical style, content and marking structure of exam
questions. You should attempt as many past exam questions as you can.
Active study
Here are some techniques that may help you to study actively.
1. Don’t believe everything you read. Good students tend to question everything that they
read. They will ask ‘why, how, what for, when?’ when confronted with a new concept,
and they will apply their own judgement. This contrasts with those who unquestioningly
believe what they are told, learn it thoroughly, and reproduce it (unquestioningly?) in
response to exam questions.
2. Another useful technique as you read the Course Notes is to think of possible questions
that the examiners could ask. This will help you to understand the examiners’ point of
view and should mean that there are fewer nasty surprises in the exam room. Use the
Syllabus to help you make up questions.
3. Annotate your notes with your own ideas and questions. This will make you study more
actively and will help when you come to review and revise the material. Do not simply
copy out the notes without thinking about the issues.
4. As you study each chapter, condense the key points (not whole chunks of text) on to a
double side of A4 or less. This is essential as otherwise, when you come to revision, you
will end up having to re-read the whole course again, and there won’t be time.
5. Try to use memory aids, such as mind maps and acronyms, to help remember the material
when you come back to it later.
6. Attempt the questions in the notes as you work through the course. Write down your
answer before you refer to the solution.

7. Attempt other questions and assignments on a similar basis, ie write down your answer
before looking at the solution provided. Attempting the assignments under exam
conditions has some particular benefits:
 It forces you to think and act in a way that is similar to how you will behave in the
exam.
 When you have your assignments marked it is much more useful if the marker’s
comments can show you how to improve your performance under exam conditions
than your performance when you have access to the notes and are under no time
pressure.
 The knowledge that you are going to do an assignment under exam conditions and
then submit it (however good or bad) for marking can act as a powerful incentive to
make you study each part as well as possible.
 It is also quicker than trying to write perfect answers.
8. Sit a mock exam four to six weeks before the real exam to identify your weaknesses and
work to improve them. You could use a mock exam written by ActEd or a past exam
paper.
You can find further information on how to study in the profession’s Student Handbook, which
you can download from their website at:
www.actuaries.org.uk/studying
Revision and exam skills

Revision skills
You will have sat many exams before and will have mastered the exam and revision techniques
that suit you. However it is important to note that due to the high volume of work involved in
Subject SP2, it is not possible to leave all your revision to the last minute. Students who prepare
well in advance have a better chance of passing the exam on the first sitting.
We recommend that you prepare for the exam by practising a large number of exam-style
questions under exam conditions. This will:
 help you to develop the necessary knowledge and understanding of the key principles
described in the Core Reading
 highlight exactly which are the key principles that crop up time and time again in many
different contexts and questions
 help you to practise the specific skills that you will need to pass the exam.
There are many sources of exam-style questions. You can use past exam papers, the Practice
Questions at the end of each chapter (which include many past exam questions), assignments,
mock exams, the Revision Notes and ASET.

Exam question skill levels
Exam questions are not designed to be of similar difficulty. The Institute and Faculty of Actuaries
specifies different skill levels that questions may be set with reference to.
Questions may be set at any skill level:

 Knowledge – demonstration of a detailed knowledge and understanding of the topic
 Application – demonstration of an ability to apply the principles underlying the topic
within a given context
 Higher Order – demonstration of an ability to perform deeper analysis and assessment of
situations, including forming judgements, taking into account different points of view,
comparing and contrasting situations, suggesting possible solutions and actions, and
making recommendations.
Command verbs
The Institute and Faculty of Actuaries use command verbs (such as ‘Define’, ‘Discuss’ and
‘Explain’) to help students to identify what the question requires. The profession has produced a
document, ‘Command verbs used in the Associate and Fellowship written examinations’, to help
students to understand what each command verb is asking them to do.
It also gives the following advice:

 The use of a specific command verb within a syllabus objective does not indicate that this
is the only form of question which can be asked on the topic covered by that objective.
 The Examiners may ask a question on any syllabus topic using any of the agreed command
verbs, as are defined in the document.
You can find the relevant document on the profession’s website at:
https://www.actuaries.org.uk/studying/prepare-your-exams

1.5 The examination
What to take to the exam

IMPORTANT NOTE: The following information was correct at the time of printing, however it is
important to keep up-to-date with any changes. See the profession’s website for the latest
guidance.
For the written exam, the examination room will be equipped with:
 the question paper
 an answer booklet
 rough paper
 a copy of the Yellow Tables.
Remember to take with you:

 black pens
 a permitted scientific calculator – please refer to www.actuaries.org.uk for the latest
advice.
Past exam papers

You can download some past exam papers and Examiners’ Reports from the profession’s website
at www.actuaries.org.uk. However, please be aware that these exam papers are for the
pre-2019 syllabus and not all questions will be relevant.

1.6 Queries and feedback
Questions and queries

From time to time you may come across something in the study material that is unclear to you.
The easiest way to solve such problems is often through discussion with friends, colleagues and
peers – they will probably have had similar experiences whilst studying. If there’s no-one at work
to talk to then use our discussion forum at www.ActEd.co.uk/forums (or use the link from our
home page at www.ActEd.co.uk).
Our online forum is dedicated to actuarial students so that you can get help from fellow students
on any aspect of your studies from technical issues to study advice. You could also use it to get
ideas for revision or for further reading around the subject that you are studying. ActEd tutors
will visit the site from time to time to ensure that you are not being led astray and we also post
other frequently asked questions from students on the forum as they arise.
If you are still stuck, then you can send queries by email to the relevant subject email address (see
Section 2.6), but we recommend that you try the forum first. We will endeavour to contact you as
soon as possible after receiving your query but you should be aware that it may take some time to
reply to queries, particularly when tutors are away from the office running tutorials. At the
busiest teaching times of year, it may take us more than a week to get back to you.
If you have many queries on the course material, you should raise them at a tutorial or book a
personal tuition session with an ActEd tutor. Information about personal tuition is set out in our
current brochure. Please email ActEd@bpp.com for more details.
Feedback
If you find an error in the course, please check the corrections page of our website
(www.ActEd.co.uk/paper_corrections.html) to see if the correction has already been dealt with.
Otherwise please send details via email to the relevant subject email address (see Section 2.6).
Each year our tutors work hard to improve the quality of the study material and to ensure that
the courses are as clear as possible and free from errors. We are always happy to receive
feedback from students, particularly details concerning any errors, contradictions or unclear
statements in the courses. If you have any comments on this course please email them to the
relevant subject email address (see Section 2.6).
Our tutors also work with the profession to suggest developments and improvements to the
Syllabus and Core Reading. If you have any comments or concerns about the Syllabus or Core
Reading, these can be passed on via ActEd. Alternatively, you can send them directly to the
Institute and Faculty of Actuaries’ Examination Team by email to
education.services@actuaries.org.uk.

2.1 Subject SP2 – background
History
The Specialist Principles subjects are new subjects in the Institute and Faculty of Actuaries 2019
Curriculum.
Subject SP2 is Life Insurance.
Predecessors
The Specialist Principles subjects cover content that was previously in the equivalent Specialist
Technical subjects. So:
 Subject SP2 replaces Subject ST2.
Exemptions
You will need to have passed or been granted an exemption from Subject ST2 to be eligible for a
pass in Subject SP2 during the transfer process.
Links to other subjects

 Subject CM1 – Actuarial Mathematics 1 provides principles and tools that are built upon in
Subject SP2.
 Subject CP1 – Subject SP2 develops a number of the general principles that were
introduced in the Actuarial Practice subject.
 Subject SA2 – The Life Insurance Advanced subject uses the concepts developed in Subject
SP2 to solve more complex problems, to produce coherent advice, and to make
recommendations in specific practice areas.

2.2 Subject SP2 – Syllabus and Core Reading
Syllabus
The Syllabus for Subject SP2 is given here. To the right of each objective are the chapter numbers
in which the objective is covered in the ActEd course.
Aim
The aim of the Life Insurance Principles subject is to instil in successful candidates the main
principles of actuarial planning and control, and mathematical and economic techniques, relevant
to life insurance companies. The student should gain the ability to apply the knowledge and
understanding, in simple situations, to the operation, on sound financial lines, of life insurance
companies. The life insurance products covered by this subject exclude health and care insurance
products covered by the Health and Care Principles subject.
Competences
On successful completion of this subject, a student will be able to:

1. understand the main principles and techniques of actuarial management and control that
are relevant to life insurance
2. apply these principles to simple situations within the context of life insurance
3. analyse hypothetical scenarios, including using judgement to assess the implications of
possible actions and to develop appropriate proposals or recommendations relating to
the management of life insurance business.
Syllabus topics
1. Life insurance products and general business environment (15%)

2. Product design and specific features (25%)
3. Risks and risk management (30%)
4. Models and valuation (15%)
5. Monitoring experience and setting assumptions (15%)
The weightings are indicative of the approximate balance of the assessment of this subject
between the main syllabus topics, averaged over a number of examination sessions.

The weightings also have a correspondence with the amount of learning material underlying each
syllabus topic. However, this will also reflect aspects such as:
 the relative complexity of each topic, and hence the amount of explanation and support
required for it
 the need to provide thorough foundation understanding on which to build the other
objectives
 the extent of prior knowledge which is expected
 the degree to which each topic area is more knowledge or application based.
Detailed syllabus objectives
0. Introduction
0.1 Define the principal terms used in life insurance. (Chapter 32)
1. Life insurance products and general business environment (15%)
1.1 Describe the main types of life insurance products. (Chapters 1 to 4 and 23)
1.1.1 Describe the main types of life insurance products that provide benefits
on death, survival to a specified point in time or continued survival.
1.1.2 Describe the following life insurance product bases:
 conventional without-profits
 with-profits
 unit-linked
 index-linked.
1.1.3 Outline typical guarantees and options that may be offered on life
insurance products.
1.2 Assess the main types of life insurance products in terms of: (Chapters 1 to 4)
 the needs of consumers and key risks for the insured

 the purpose and key risks for the insurer.

1.3 Assess the effect of the general business environment on the management of life
insurance business, in terms of: (Chapters 8 and 9)
 propensity of consumers to purchase products

 local culture
 methods of sale
 remuneration of sales channels
 types of expenses and commissions, including influence of inflation
 economic environment
 legal environment
 regulatory environment
 taxation regime
 professional guidance.
2. Product design and specific features (25%)
2.1 Demonstrate an understanding of life insurance product design, including:

(Chapter 16)
2.1.1 Describe the factors to consider in determining a suitable design, in terms

of premiums, benefits and charges, for a life insurance product.
2.1.2 Determine a suitable design for a product in a given situation.
2.1.3 Discuss the relative merits of different product designs.
2.2 Demonstrate an understanding of with-profits business management.

(Chapters 5 to 7)
2.2.1 Describe methods of distributing profits to with-profits policyholders.
2.2.2 Explain the main uses of asset shares and how they may be built up using
a recursive formula.
2.3 Describe the principles of unit pricing for internal unit-linked funds. (Chapter 13)
2.4 Determine discontinuance and alteration terms for without-profits contracts.

(Chapters 21 and 22)
2.4.1 State the principles of setting discontinuance and alteration terms.
2.4.2 Describe different methods of determination of discontinuance and

alteration terms.
2.4.3 Assess the extent to which these methods meet the principles in 2.4.1.
2.4.4 Calculate surrender values and alteration terms for conventional

without-profits contracts using reserves or by equating policy values.

3. Risks and risk management (30%)
3.1 Assess how the following can be a source of risk to a life insurance company:
(Chapters 10 to 12)
 policy and other data

 mortality rates
 investment performance
 expenses, including the effect of inflation
 persistency
 mix of new business
 volume of new business
 guarantees and options
 competition
 actions of the board of directors
 actions of distributors
 failure of appropriate management systems and controls
 counterparties
 legal, regulatory and tax developments
 fraud
 aggregation and concentration of risk.
3.2 Demonstrate the application of reinsurance as a risk management technique.

(Chapters 24 and 25)
3.2.1 Describe the purposes of reinsurance.
3.2.2 Describe the different types and structures of reinsurance.
3.2.3 Discuss the factors that should be considered before taking out
reinsurance.
3.3 Demonstrate the application of underwriting as a risk management technique.

(Chapter 26)
3.3.1 Outline the purposes of underwriting.
3.3.2 Describe the different approaches by which underwriting is applied.
3.3.3 Discuss the factors that should be considered when determining the level
of underwriting to use.

3.4 Demonstrate the application of asset-liability matching as a risk management

technique. (Chapter 28)
3.4.1 State the principles of investment for a life insurance company.
3.4.2 Analyse life insurance liabilities into different types for asset-liability
matching purposes.
3.4.3 Propose an appropriate asset-liability matching strategy for different

types of liability.
3.5 Propose further ways of managing the risks in 3.1, including:

(Chapters 7, 15, 27 and 29)
 policy data checks

 choice of with-profits bonus method
 capital management
 expense control
 policy retention activity
 management of new business mix and volumes
 management of options
 systematic risk assessment and management strategies.
4. Models and valuation (15%)
4.1 Describe the main features of a life insurance model. (Chapters 14 and 15)
4.1.1 Outline the objectives and basic features of a life insurance model.
4.1.2 Compare stochastic and deterministic approaches.
4.1.3 Explain the use of sensitivity analysis.
4.2 Demonstrate the different uses of actuarial models for decision-making purposes
in life insurance, including: (Chapters 15, 18 and 28)
 pricing products
 developing investment strategy
 projecting solvency
 calculating embedded value.
4.3 Demonstrate methods of determining the cost of options and guarantees.

(Chapter 23)
4.3.1 Describe the use of stochastic simulation and the use of option prices to
determine the cost of an investment guarantee.
4.3.2 Describe the assessment of the cost of simple mortality options.

4.4 Discuss the determination of supervisory reserves and solvency capital

requirements for a life insurance company. (Chapters 19 and 20)
4.4.1 Describe how supervisory reserves and solvency capital requirements may
be determined, including:
 market-consistent valuation
 non-unit reserves
 Value at Risk (VaR) capital assessment.
4.4.2 Discuss the interplay between the strength of the supervisory reserves
and the level of solvency capital required.
4.4.3 Compare passive and active valuation approaches, including the valuation
of assets.
5. Monitoring experience and setting assumptions (15%)
5.1 Describe the principles of setting assumptions for life insurance business.
(Chapters 15, 17, 18 and 20)
5.1.1 Describe the principles of setting assumptions for pricing life insurance
contracts, including profit requirements.
5.1.2 Describe the principles of setting assumptions for determining liabilities.
5.1.3 Explain why the assumptions used for supervisory reserves may be
different from those used in pricing.
5.1.4 Outline the principles of setting assumptions for determining embedded

value.
5.2 Demonstrate the relevance of experience monitoring to a life insurance company.

(Chapter 30)
5.2.1 Explain why it is important for a life insurance company to monitor its
experience.
5.2.2 Describe how the actual mortality, persistency, expense and investment
experience of a life insurance company should be monitored, including
the data required.
5.3 Demonstrate the relevance of analysis of surplus or profit. (Chapter 30)
5.3.1 Give reasons for undertaking an analysis of surplus and an analysis of

embedded value profit.
5.3.2 Suggest ways in which the results of such analyses can be used.

6. Solving problems
6.1 Analyse hypothetical examples and scenarios in relation to the financial

management of life insurance companies. (Chapter 31)
6.1.1 Propose solutions and actions that are appropriate to the given context,
with justification where required.
6.1.2 Suggest possible reasons why certain actions have been chosen.
6.1.3 Assess the implications of actions within a given scenario.
6.1.4 Discuss the advantages and disadvantages of suggested actions, taking

into account different perspectives.

Core Reading
The Subject SP2 Course Notes include the Core Reading in full, integrated throughout the course.
Further reading
The exam will be based on the relevant Syllabus and Core Reading and the ActEd course material
will be the main source of tuition for students.

2.3 Subject SP2 – the course structure

There are six parts to the Subject SP2 course. The parts cover related topics and have broadly
equal marks in the exam. The parts are broken down into chapters.
The following table shows how the parts, the chapters and the syllabus items relate to each other.
The end columns show how the chapters relate to the days of the regular tutorials. We have also
given you a broad indication of the length of each chapter. This table should help you plan your
progress across the study session.

No of Syllabus 3 full
Part Chapter Title
pages objectives days
1 Life insurance products 1 28 1.1, 1.2
1 4 Life insurance products 4 46 1.1, 1.2
5 Asset shares 22 2.2
6 With-profits surplus distribution 1 32 2.2
1
7 With-profits surplus distribution 2 33 2.2, 3.5
8 The general business environment 1 25 1.3
9 The general business environment 2 29 1.3
2 10 Risk 1 24 3.1
11 Risk 2 34 3.1
12 Risk 3 14 3.1
13 Unit pricing 20 2.3
14 Models 1 15 4.1
3 15 Models 2 39 3.5, 4.1, 4.2,
5.1
16 Product design 36 2.1
17 Setting assumptions 1 47 5.1 2
18 Setting assumptions 2 24 4.2, 5.1
19 Supervisory reserves & capital requirements 1 20 4.4
4
20 Supervisory reserves & capital requirements 2 34 4.4, 5.1
21 Surrender values 27 2.4
22 Alterations 31 2.4
23 Cost of guarantees and options 27 1.1, 4.3
24 Reinsurance 1 29 3.2
5 25 Reinsurance 2 33 3.2
26 Underwriting 33 3.3
27 Policy data checks 14 3.5 3
28 Investment 42 3.4, 4.2
29 Further risk management 26 3.5
6
30 Monitoring experience 54 5.2, 5.3
31 Problem solving 43 6.1
32 Glossary 13 0.1

2.4 Subject SP2 – summary of ActEd products

The following products are available for Subject SP2:
 Course Notes
 X Assignments – six assignments:
– X1-X3: 80-mark tests (you are allowed 2 hours to complete these)
– X4-X6: 100-mark tests (you are allowed 3¼ hours to complete these)
and X Assignment marking (Series Marking and Marking Vouchers)
 Online Classroom – over 50 units
 Flashcards
 Revision Notes – five A5 booklets
 ASET – four years’ exam papers, ie eight papers, covering the period April 2014 to
September 2017
 Mock Exam and marking (Series Marking and Marking Vouchers).
We will endeavour to release as much material as possible but unfortunately some revision
products may not be available until the September 2019 or even April 2020 exam sessions.
Please check the ActEd website or email ActEd@bpp.com for more information.
The following tutorials are typically available for Subject SP2:

 regular tutorials (three days)
 block tutorials (three days).
Full details are set out in our Tuition Bulletin, which is available on our website at
www.ActEd.co.uk.

2.5 Subject SP2 – skills and assessment
Exam skills
Exam question skill levels
In the SP subjects, the approximate split of assessment across the three skill types is:
 Knowledge – 25%
 Application – 50%
 Higher Order skills – 25%.
Assessment
The Specialist Principles examinations are in the form of 3¼-hour paper-based examinations.

2.6 Subject SP2 – frequently asked questions

Q: What knowledge of earlier subjects should I have?
A: The Course Notes have been written assuming that you have already studied, or been
exempted from, the Core Principles subjects (the CS, CM and CB subjects) or the equivalent
CT subjects. The key area that you will need in studying Subject SP2 is the material in
Subject CM1, or equivalently Subject CT5.
Q: I have studied Subject CP1 and/or SP1. Will this help?
A: Yes, it will help, but we don’t recommend that you alter your study methods. Read the
course as you would any other. You may be familiar with some parts of the notes, and
this will enable you to work through these more quickly so that you can spend more time
looking at the less familiar material.
Q: What is your advice if I am simultaneously studying Subject SA2?
A: Subject SA2 builds on the common principles developed in Subject SP2, but requires a
much greater depth of knowledge and understanding. Consequently, there is a degree of
overlap between the two subjects – both in the Core Reading and also possibly in the
types of questions that are likely to appear on the exam papers. It is therefore important
to assimilate the key ideas presented in Subject SP2 before tackling the same ground in
Subject SA2.
We suggest that you aim to cover the Subject SP2 course as quickly as possible, so as to
get a general feel for the principles underlying life insurance, together with an overview of
the course content. It also makes sense to quickly review the relevant Subject SP2
material prior to working through each chapter in Subject SA2.
From time to time over the study session, and particularly at the revision stage, it might
also be a good idea to review the Subjects SP2 & SA2 Course Notes at the same time,
along with the Practice Questions. In particular, it is always worth thinking about how
each idea or principle is presented in each of SP2 and SA2 and hence how it might
consequently be examined in either exam.
Q: What should I do if I discover an error in the course?
A: If you find an error in the course, please check our website at:
www.ActEd.co.uk/paper_corrections.html
to see if the correction has already been dealt with. Otherwise please send details via
email to SP2@bpp.com.
Q: Who should I send feedback to?
A: We are always happy to receive feedback from students, particularly details concerning
any errors, contradictions or unclear statements in the courses.
If you have any comments on this course in general, please email to SP2@bpp.com.

If you have any comments or concerns about the Syllabus or Core Reading, these can be
passed on to the profession via ActEd. Alternatively, you can send them directly to the
Institute and Faculty of Actuaries’ Examination Team by email to
education.services@actuaries.org.uk.

SP2-01: Life insurance products (1) Page 1
Life insurance products (1)

Syllabus objectives
1.1 Describe the main types of life insurance products.
1.1.1 Describe the main types of life insurance products that provide benefits on
death, survival to a specified point in time or continued survival.
 with-profits
 unit-linked
 index-linked.
1.2 Assess the main types of life insurance products in terms of:
(Covered in part in this chapter.)

Page 2 SP2-01: Life insurance products (1)
0 Introduction
We begin our study of life insurance by taking a look at the types of product that life insurance
companies around the world typically offer. Each market will have its own varieties, but we shall
be concentrating on the generic forms.
The products typically offered in a life insurance market have evolved over time so as to
represent reasonably profitable risks to the insurer whilst providing useful benefits to the
consumer.
The contracts discussed in these four chapters are variations on the basic theme of life
insurance.
Life insurance can be defined as follows:
In return for one or more premiums paid by a consumer, the insurer contracts to pay
benefits which are in some way contingent upon human life – payment occurs on death or
survival.
Insurance contracts that pay benefits depending on the health of the policyholder are described
in Subject SP1 (Health and Care Insurance), and will not be discussed (in any depth) here.
The simplest contracts provide specified guaranteed benefits in return for the payment of
specified premiums. Variations on this simple basis are discussed later in this chapter (and
also later in the course), of which the most notable are with-profits and unit-linked
contracts.
The simplest contracts referred to by the above Core Reading are known as non-profit non-linked
(or without-profits non-linked) contracts – which are shortened (usually) to non-profit (or without-
profits) respectively. On first reading, this terminology seems to suggest that these contracts are
not profitable – but this is not what it means. The name is simply telling us that the policy is
neither with-profits nor unit-linked.
The terms and conditions of the contract will specify both the amount of each premium and for
how long the premiums have to be paid. When a contract is paid for by regular premiums
(typically of level monthly or annual amounts), the premiums would usually be payable for a fixed
number of years or until earlier death.
In practice, contracts may also provide benefits on surrender, but this is usually a non-
contractual payment of non-guaranteed amount.
A surrender is said to occur when a policyholder fails to pay all of the premiums required under
the contract, and receives a lump sum (surrender value) in compensation for the premiums paid
to date. The amount paid (at the date of surrender) would normally not be specified in the
contract, but would be at the discretion of the company at that time. However, contracts that
give guaranteed surrender values can exist.
The payment of a surrender value is just one example of what can happen when a policyholder
stops paying his or her contractual premiums in this way. As an alternative, it is usually possible
for the policy to continue, without paying any more premiums, but for a reduced benefit amount.
This is called making the policy paid up.

Some policies do not pay any benefits if premiums cease before the contractual time: in these
cases the contracts are said to lapse.
Surrender, paid up and lapse are all examples of withdrawal or policy discontinuance, although
there is no universally accepted definition of these terms. We will describe how insurance
companies work out their discontinuance terms (eg how much surrender value to pay) later in the
course.
The type of contractual benefit offered provides a useful way of listing the main products.
For each contract type, we first give a description of how the benefits provided may be
useful to a consumer.
The basic customer needs met by life insurance contracts are protection and savings. Many
contracts protect people (or their dependants) from the financial consequences of unwelcome
events, such as death. Other contracts are essentially investments, allowing the policyholder to
build up funds for specific things such as an income in retirement, the repayment of a loan, or just
a lump sum to spend as the policyholder wishes. Some contracts provide a mixture of savings and
protection.
Secondly an outline is given of the risks associated specifically with each product. These
‘micro’ risks are one part of the overall risk picture. The ‘macro’ aspects are covered in
later chapters.
The risks that the Core Reading is referring to here are mainly risks to the insurance company,
rather than the policyholder, although the risks that policyholders take on when taking out life
insurance are also important. An example of a ‘micro’ risk (to the insurance company) would be if
more deaths than expected occurred on an assurance contract, leading to the insurance company
making less profit than expected on the contract. An example of a ‘macro’ risk would be the
company becoming insolvent.
The risks borne by the insurance company in writing insurance business vary considerably
depending on the type of contract involved. In Chapters 1 to 4 we shall study how and why this
variation occurs for the different products that life insurers sell. As suggested by the Core
Reading, a ‘big picture’ view of the risks involved in being a life insurance company will be
discussed later in the course.
A familiarity from earlier subjects of the concept of new business strain, and the
consequent requirement for capital to write a new contract, is assumed.
However, just in case you are feeling a little rusty……
Consider first the cashflows that occur when a contract is written. A large proportion of the costs
associated with the contract may occur at the start of the contract, because of items such as
marketing, underwriting (the process of assessing risks) and initially setting the policy up on the
company’s computer systems. In many markets it is also common practice to pay a commission
to the provider of the business, for example the company’s own salesperson or a financial adviser.
This means that the cashflow (ie income less outgo) in the first year will be low or, often,
negative. This is the amount of money the company has on day 1 of the policy. (This is also
known as the policy’s asset share as at day 1 – more about asset shares later.)

A further issue is that the insurance supervisory authorities may require that money (reserves) be
specifically set aside to ensure that the company is able to meet its obligations to policyholders.
In some countries the supervisory authorities require that these reserves are calculated on a
prudent basis. Capital is then tied up because of this need to set aside reserves larger than those
the company really believes are needed to meet payouts to policyholders.
The company may also have to maintain a minimum amount of assets in addition to those backing
its assessment of liabilities. Such a ‘solvency capital requirement’ ties up further capital.
The combination of the initial cash outflow, any prudence in the reserves and the need to
establish the required solvency capital means that money has to be found initially in order to
write the business. Hence there is an ‘initial capital strain’.
Because we all like formulae, this initial capital strain can be written:
C0  V0  E0  P0
where C0  capital strain at time 0 
V0  supervisory reserves and required solvency capital at time 0+
E0  expenses and commission incurred by time 0+
P0  premium paid by time 0+
and time 0+ is the point immediately after the policy has been issued, after the first premium has
been paid, and all the initial expenses have been incurred. P0 would be the single premium, or
the first regular (annual or monthly) premium.
(The asset share at time 0+ can be written:
A0  P0  E0 .
Therefore the initial capital strain is also the excess of the supervisory reserve and required
solvency capital over the asset share on day 0+.)
Depending on the contract design and the particular supervisory regulations, this capital may be
recouped quickly or only very slowly. All other things being equal, a company is likely to prefer
contract types and designs that recoup the invested capital quickly, so that this capital can be
used to fund further profitable business.
Question
Taken to its logical conclusion, this would imply that companies would be selling contracts with
very high premiums or charges in the first year of any contract. In practice, though, premiums (or
charges) are often level or even gradually rising during the policy term.
Explain why this is the case.

Solution
Products with very high first-year premiums or initial charges may not be marketable. Another
way of thinking about this is that such a product design does not meet the needs of the public; for
example policyholders may be less able to afford high costs early on in the policy term than later
on in the term. Another factor is the expectations of the public in the light of companies’ past
practice: if life insurance has traditionally been paid for by issuing policies with level premiums,
then it is hard to do things in a very different way subsequently.
Examples like this often happen in real life: the interests of policyholder and insurance company
often differ and cannot both be satisfied fully by the decision reached. In this case the
‘compromise’ reflects the policyholders’ requirements, as the alternative could well result in no
policies being sold at all.
Because of its importance, therefore, we shall be considering the capital requirements of the
different types of contract throughout these first four chapters.

1 The personal financial life cycle

The personal financial life cycle can give an indication of the needs of customers for various life
and health and care products at different times in an individual’s life, eg young adult, mid-life
adult etc.
The history of the life insurance industry has been a gradual evolution in which companies have
been designing, and redesigning, their products to meet the needs of the consumer. All the
products that are currently sold in any insurance market are there because they meet an
important consumer need. It is therefore important to understand the life insurance needs of the
consumer, and how these may change both over personal lifetimes and over calendar time.
This is, of course, an important part of the general commercial and economic environment for the
insurance company. A key point is that consumer needs do not remain static, but change as
lifestyles, standard of living, education and technology each change over time. For example, if
more people become financially better off over time, then the demand for savings products
should rise and that for protection may fall. It may also allow more people to accept more risk in
the policies they buy (with the greater potential rewards they provide). This will tend to increase
demand for the more risky products and reduce demand for the (more expensive) fully
guaranteed products. Changes in technology and a more educated public also allow more flexible
and complicated products to be sold. These are all changes that affect demand for products over
calendar time.
There are also important changes in personal financial needs that affect demand for different
types of products over individual lifetimes. This helps to explain the variety of products that are
sold simultaneously in any life insurance marketplace. A typical personal financial life cycle (in a
country with a developed economy) may proceed as follows (although the actual age-bands may
differ quite widely between individuals):
 Ages 16 – 25
May still be in higher education, first job, may or may not have partner, probably does not
own home or have dependants.
Financial needs: may still have some support from parents; may be saving towards future
family needs, such as buying a home; paying off student debts; likes to spend money if
possible.
(This period can be very short for those who do not continue into higher education.)

 Ages 25 – 40
May have partners and sometimes children, large debts (eg if they have borrowed money
to buy a home), moderate income and often high expenditure (cost of raising children)
but often not much wealth.
Financial needs: loans to meet cost of buying home and/or other high expenditure,
worried about what will happen to dependants should earner(s) become sick or die. May
start saving, possibly for the benefit of children as they grow older. Far-sighted
individuals may perceive the need to start saving for retirement.
 Ages 40 – 65
Children become older and ultimately independent. Debts reduce, loans are paid off.
Income may increase and could well outstrip expenditure. Periods of redundancy may
also occur.
Biggest financial need is to save for retirement. May also wish to provide for the cost of
future long-term care (to protect against the consequences of long-term sickness or
disability), and how best to manage the transfer of wealth to the next generation (on or
before death). May have material other additional disposable income and freedom of
choice in how this is used.
 Ages over 65
Move from employment into retirement. Few debts, but much lower income. Should
have more accumulated wealth. Main risks are running out of money if become very old,
and an increasingly imminent need for long-term care. Still wish to save disposable
income for leisure activities and for wealth transfer.
As you study the remainder of this and the next three chapters, think about where each of the
products you meet fits into the personal financial life cycle. Also think about whether you would
expect the products described to be popular in your own country. (We will return to this again at
the end of Chapter 4.)

2 Contracts that insurers sell

The contracts we shall be looking at in Chapters 1 to 4 are:
Chapter 1 Endowment assurances
Chapter 2 Whole life assurances
Term assurances
Convertible or renewable term assurances
Chapter 3 Immediate annuity contracts
Deferred annuity contracts
Chapter 4 Unit-linked contracts
Index-linked contracts
Although unit-linked and index-linked contracts have separate sections to themselves (in
Chapter 4), linking is essentially an approach to product design that can be used with most of the
contracts covered in Chapters 1 to 3. Hence they are not strictly products in their own right.
Most of the contracts in Chapters 1 to 3 could come in linked or non-linked form. We shall
therefore use our first contract, the endowment assurance, to illustrate both linked and
non-linked designs. For the other contracts we study in Chapters 1 to 3 we shall concentrate
largely on the non-linked versions. When we come to study unit-linked contracts specifically, in
Chapter 4, we shall use one or two of the contract types to illustrate the key features of
unit-linked designs and how they differ from non-linked designs.

3 Endowment assurances
3.1 The needs of consumers

An endowment is a contract to pay a benefit on survival to a known date and hence
operates as a savings vehicle, for example to provide a lump sum on retirement, or a means
of repaying the capital on an interest-only loan.
The contract may also provide a significant benefit on the death of the life insured before
that date and, in this case, operates also as a vehicle for providing protection for
dependants and is known as an endowment assurance.
Endowment assurances are often used as a means of transferring wealth from, say, parents to
children. The basic idea is that a parent would pay the premiums but the benefit would be
payable to the child.
Question
Explain why an endowment assurance could be a more attractive means of transferring wealth
than a non-insured savings method.
Solution
If the sum assured on death is chosen to be the same as at maturity, then the endowment
assurance provides the guarantee that a substantial wealth transfer will be made whether or not
the policyholder survives the intended savings period. A non-insured savings method would
simply have accumulated the contributions paid in by the date of death, and this could be very
small (eg if the person died after only a short time).
An endowment assurance is sometimes used as a means of repaying the capital on an

interest-only loan. Where this is done, the borrower pays interest (only) to the lender while the
loan is outstanding. The borrower also takes out an endowment policy, the proceeds of which are
designed to repay the debt.
Another use of endowments is to provide a vehicle for saving money for retirement. (We
describe this use under the heading of deferred annuities in Chapter 3.)
Typically, a surrender value would also be available.
An endowment assurance could come in without-profits, with-profits or unit-linked form. Each

has significantly different characteristics and will meet different consumer needs. Here is a brief
description of each.

Without-profits endowment assurance

A without-profits endowment assurance offers a guaranteed amount of money (the ‘sum
assured’) at the end of the contract in exchange for a single premium at the start of the contract
or a series of regular premiums throughout the contract. If the policyholder dies before the
contract term ends then usually the same sum assured is paid on death. However, the contract
could be structured with a sum assured paid on death different from that paid at maturity.
With-profits endowment assurance

(With-profits contracts are also known as participating insurance contracts.)
The structure of what we know as conventional with-profits contracts is similar to that of the
without-profits contract, except that the initial sum assured is expected to be enhanced by the
declaration of bonuses to the policyholder. In the UK, bonuses usually take the form of additions
to the amount of benefit, and this is also true in many other countries. In some other countries
bonuses are given as regular cash payments, or as a reduction in future premiums.
As well as the conventional type of with-profits policy, there is also an accumulating type of with-
profits structure. We will cover this when we look at with-profits business in detail later in the
course.
Question
A conventional with-profits endowment and a without-profits endowment both have an initial

sum assured of £10,000.
State, with justification, which would have the higher initial premium, all else being equal.
Solution
The premium for the with-profits contract should be larger, probably significantly so.
This is because the with-profits premium is set to provide an expected benefit that is much larger
than the initial guaranteed amount, whether as additions to the eventual benefit, as cash
bonuses, or as reductions in future premiums.
Unit-linked
Unit-linked contracts operate by paying policyholder premiums into pooled investment funds.
The benefit payable at maturity depends on the performance of the underlying assets and the
level of charges levied by the insurance company.

The benefit on death might be, for example:

 a fixed sum (eg £10,000), or
 the value of units, or
 some percentage (eg 120%) of the value of units.
This makes unit-linked policies very versatile: if the first option is chosen, with a very high sum
assured (relative to the premium), then the policy can be almost entirely protection. Choose a
lower level of sum assured (or, equally, a higher premium), then the policyholder could end up
with unit funds accumulating up to the sum assured level after a period of, say, 20 years. This
would be similar to the non-linked endowment described above, with death benefit equal to, or
at least the same order of magnitude as, the survival benefit. (This version would be chosen if the
unit-linked endowment were to be used for paying off a loan.)
With either of the second two versions the emphasis would be on savings, so these might be
appropriate if the policy was being used as a retirement savings policy, for example.
All three versions are commonly found in practice, because they are used to meet different
needs.
The company levies its charges in order to cover expenses and the cost of life cover, and
(probably) to make a profit. The charges may be taken from premiums before they are invested
in units and/or as deductions from the unit fund.
A range of investment funds may be offered, for example International Equities, European Bonds,
Venezuelan Tin Futures etc.
We cover unit-linked contracts in more detail a little later in the course. You might also like to
revisit your Subject CM1 or CT5 course to remind you of how these contracts work.
Group endowment assurances

‘Group’ contracts usually arise as a way for an employer to provide some form of insurance cover
or savings benefit to employees as part of their overall remuneration package. Group contracts
might also be sold to ‘affinity groups’ (clubs, societies etc).
A group endowment assurance would enable, for example, an employer to provide benefits
for employees at retirement, and maybe also on death in service.
In this case, the contract would effectively consist of a collection of individual endowment
assurances, one for each employee.
Question
Suggest, with reasons, a practical problem with such an arrangement that would not arise on an
individual contract.

Solution
The main problem would be mobility of the workforce (by ‘mobility’ we mean ‘the ability to
change jobs frequently’). With a highly mobile workforce, the administration of such an
arrangement would become costly. It might also result in poor surrender values for employees
who leave after a short time, if the individual contracts have to be surrendered on leaving.
(Alternatively, it might mean losses for the insurance company if for any reason, such as
government legislation, it has to pay over-generous surrender values to employees who leave.)
As with individual contracts, group endowment assurances could be without-profits, with-profits

or unit-linked.
Example
A large unit-linked group endowment assurance might have the following features:
 All contributions, ie premiums, are paid by the sponsoring employer.
 The only charge is a monthly percentage of each member’s fund.
 The scheme provides a cash fund for each member, equal to the value of units, which is
used to purchase an annuity at retirement.
3.2 The risks
Investment risk
The savings nature of the contract introduces a greater investment risk than for a term
assurance, but the extent depends on whether the contract is unit-linked (and whether any
guaranteed benefits are given), index-linked, without-profits or with-profits.
We are concentrating here on the amount of investment risk borne by the life insurance
company, rather than the policyholder. In other words, if the assets held to meet the liabilities
perform badly, to what extent is the company adversely affected?
The way in which investment risk varies under without-profits, with-profits and unit-linked
endowment assurances is explored in one of the Practice Questions at the end of this chapter.
Index-linked products are covered in Chapter 4.
Mortality risk
The nature of the death benefit provided will determine whether there is a mortality risk and
the nature of that risk. Possible approaches include:
 A significant benefit, for example a without-profits contract where the benefit on

death equals that payable on survival. There will be a significant mortality risk at the
start of the contract, but this will reduce as duration in force increases.

Question
Explain why there is this pattern of risk over time.
Solution
The sum at risk (ie the sum assured minus the reserve) reduces as the reserve underlying the
contract builds up. So, for every death that occurs near the start of the contract, the company
loses more money than for every death nearer the end of the contract.
Associated with this mortality risk will be an anti-selection risk. The extent of this
risk will depend on the extent of the actual, or perceived, choice the policyholder
had in effecting the contract.
Anti-selection can involve, for example, an applicant for insurance using knowledge he or
she has about their own state of health to gain favourable terms from an insurer. For
example, a person who knows that they are seriously ill may apply for insurance. Life
insurance companies protect themselves against this risk in various ways, the chief of
which is underwriting. We shall consider anti-selection and underwriting in more detail
later in the course.
 Return of premiums or ‘fund’. The mortality risk is likely to be insignificant except

near the start of the contract.
The exact interpretation of ‘fund’ will vary from contract to contract, but broadly it is an
accumulation of premiums and similar in concept to the ‘asset share’. In the case of a
unit-linked contract the most likely interpretation of ‘fund’ would be the value of the
policyholder’s unit account at the time of death.
Question
Explain why this mortality risk might be more significant near the start of a contract.
Solution
This will mostly occur where the death benefit is a return of premiums. Near the start of the
contract the reserve might be less than the total premiums paid, leading to a positive sum at risk
and hence a mortality risk. Later in the policy term the reserve would almost certainly exceed the
death benefit and hence there would be no mortality risk (in fact, there would be a slight
longevity risk).
 No death benefit (ie a ‘pure endowment’ contract). Here there will be a longevity
risk, the significance of which will increase with duration in force.

Clearly, if the contract pays a benefit only on survival to maturity then the company risks losing
money if fewer policyholders than expected die. In effect the death strain at risk is negative, and
becomes increasingly so as the policy approaches maturity. This is why the significance of the risk
increases with duration in force.
Expense risk
There is an expense risk in that the actual marginal costs of administering the contract need
to be met.
The marginal costs of a contract are those costs which are incurred because the policy exists. The
most obvious example of a marginal cost is commission, which is incurred only when the policy is
in force (both at the start of the policy, and also often when further premiums are paid). But
there are others: eg costs of postage and telephone calls to policyholders, printing of policy
documents or statements – these are all marginal costs provided they would not have been
incurred if the policy was not in existence.
The non-marginal costs are referred to as fixed costs or overheads. The vast majority of
staff-related costs are fixed, at least in the short term, as staff are not hired or fired continuously
in proportion to the volume of business administered by the company. For example, the number
of actuaries the company employs will remain relatively stable over a period of time, whether the
company is in a sales boom or a sales depression.
Meeting the marginal costs is the very least that the company should require from a contract. In
most cases, a contribution to the company’s fixed costs and profit will also be required. Strictly,
the risk is that the total actual expenses (over a period) are higher than the total contributions
received towards those expenses from the charges and/or premiums received by the company.
As these charges or premiums reflect the assumptions made for future expenses when the
products were priced, then this can be regarded as the risk that the company’s expenses are
greater than the levels assumed when pricing the business.
Expenses might prove higher than expected for many reasons, for example because of higher
than expected price inflation, or because of lower than expected sales of business.
Question
Explain why lower than expected sales of business can lead to expenses being higher than
expected.
Solution
When a contract is priced, an assumption will be made about the expected average amount of
expenses incurred per policy in force. Because some of the expenses are fixed (ie independent of
the number of policies in force) then the actual average expenses per policy will rise if fewer
policies are in force – which will generally be the case if fewer policies are sold. However, each
policy’s contribution to the expected expenses will remain the same, so that, on a per-policy basis,
the actual expenses will have increased relative to the expected expenses.

Withdrawal (persistency) risk

Persistency risk is the risk that the number of withdrawals is different to that expected.
Withdrawals could include not only surrenders and lapses, but also partial withdrawals and paid-
up policies.
At times when the asset share is negative, there is a financial risk from withdrawal. At other
times, whether there is such a risk depends on how any withdrawal benefit paid compares
with the asset share.
Asset shares are covered fully a little later in the course. However, broadly speaking, the asset
share of a policy is the accumulation of the premiums paid under the contract (ie including any
investment return earned), minus the expenses incurred and the cost of the benefits provided. It
is sometimes referred to as the ‘earned asset share’. An important point about an asset share is
that it is built up from actual cashflows attributed to individual policies, accumulated at actual
earned rates of investment return.
Question
Calculate the asset share of a ten-year endowment assurance policy issued at age 60, for a
policyholder surviving exactly three years (after payment of all surrender and death claims due at
the end of the year), given the following information:
Sum assured, payable at end of year of death: £10,000

Annual premium: £1,000
Policy’s share of expenses, incurred at start of year: £25
Asset share per policy at the end of the second policy year: £2,000
Rate of return on assets during the year (net of tax): 11.27%
Expected mortality: AM92 Ultimate
A recent investigation of the company’s relevant mortality experience has shown that mortality
has been running at an average of 65% of the expected rate. An investigation of the company’s
withdrawals has shown that 7% of all policyholders with policies in force surrender their policy on
the third policy anniversary, without paying the premium then due. An amount equal to 90% of
the current asset share is paid as a surrender value in each case.
Solution
The asset share at the start of the year increases by the premium paid, less the actual expenses
incurred, plus the investment return during the year:
  2,000  1,000  25   1.1127  3,310.28
The actual mortality rate experienced is:
AM92
0.65  q62  0.65  0.010112  0.0065728

The policy’s share of the death cost is:
0.0065728  10,000  65.73
The asset share per survivor is:
3,310.28  65.73
 3,266.02
1  0.0065728
Allowing for surrenders of 7% we end up with a final asset share, per policy surviving to the end of
the year, of:
3,266.02  0.07  0.9  3,266.02

 £3,290.60
1  0.07
As stated in the Core Reading, one example of a situation where withdrawal risk arises is when
the asset share is negative.
Question
State when the asset share of an endowment assurance policy is most likely to be negative, giving
a reason for your answer.
Solution
Asset shares are most likely to be negative in the first few months or years of a regular premium
contract, because of high initial expenses.
When the asset share is negative on withdrawal, the company will lose money even if it pays the
policyholder nothing. When the asset share is positive, the company will lose money if the
withdrawal benefit is larger than the asset share.
Question
When we were talking about mortality risk, we looked at the death strain – the difference
between the death benefit and the reserve of the policy at the time of death. But when we talk
about withdrawal risk, we are looking at the difference between the withdrawal benefit and the
asset share at the time of claim. This suggests that the loss made by a death and a withdrawal
would be different even if they paid out the same amounts and related to identical policies,
because the reserve and the asset share would not be the same.
Explain the difference between these two ways of measuring loss.

Solution
The reserve (by which we mean the supervisory reserve) is how much physical money the
company is actually holding towards the liability under the policy. The asset share represents
how much money the company has accumulated to date from the historical cashflows from the
policy.
Suppose a policy has current asset share £900 and a reserve of £1,200. A claim arises which is
settled by a lump sum payment of £1,300.
First look at the asset share. The company has accumulated £900 from the policy, but has to pay
out £1,300: so it has made a total accumulated loss of £400.
However, because the company was holding reserves of £1,200 but will need no reserve for the
policy going forwards, the company only needs to find a further £100 to meet the claim cost.
That is, the company has a capital shortfall of £100, and must fund this sum from its available
capital resources (or failing that, become insolvent).
So we can see that:

 [claim cost] – [reserve] is the capital loss incurred by the policy at the time of claim
 [claim cost] – [asset share] is the overall accumulated loss from the policy at the point
immediately after the claim has been paid.
If the above amounts are negative, then:

 [reserve] – [claim cost] is the amount of capital (or supervisory profit) released by the
policy at the date of claim
 [asset share] – [claim cost] is the total amount of profit that the policy has accumulated
for the company at the point immediately after the claim has been paid.
It is quite possible for one of the above to give a loss and the other to give a profit.
The determination of [claim cost] – [asset share] might ignore the cost of capital. The actual loss
to the shareholders should include the opportunity cost of holding reserves.
Don’t worry if you didn’t get this solution on your first reading through – the concepts involved will
be developed during the course and should become clearer later.
Risks under group contracts

The group version of the contract adds no additional risks and any anti-selection risk is
likely to be much reduced, particularly if it is compulsory for all eligible members to join the
group contract.
However, the grouped nature of the contract does mean that a concentration of risk may arise.
For example, the contract might cover employees in the same workplace, where an industrial
accident could result in a number of claims.

The reduced anti-selection risk follows from the restricted choice that individuals may have where
group schemes are concerned. For example, depending on the product and the market in which
it is sold:
 Membership of the group may be compulsory (eg for all employees of a company).
 There may be restrictions on the level of cover each member can have (perhaps related to
salary).
3.3 Capital requirements

How much initial capital is needed to write a contract depends on a number of factors. The key
factors given in the Core Reading below apply not only to endowment assurances but also much
more generally.
The capital requirement will depend on the:
 design of the contract
 frequency of payment of the premium
 relationship between the pricing and supervisory reserving bases
 additional solvency capital requirements
 level of the initial expenses (including any initial commission).
We shall consider each of the Core Reading points in turn.
Contract design
The key issue in contract design, as far as capital requirements are concerned, is whether the
design enables reserves and solvency capital requirements to be kept low. Broadly speaking, the
lower the initial reserves, the lower the initial capital requirement. The slower the increase in
reserves over the contract’s term, the faster any invested capital is released. This issue therefore
interacts with the issue of how supervisory reserves are calculated.
Example
In many countries, unit-linked contracts can be made very capital efficient. Under one possible
design a very low (even zero) percentage of the early premiums paid is allocated to unit funds.
Because the unit-related liability is low (or zero), the supervisory reserve needed to cover it is very
low (or zero). Most or all of the early premiums can be used to recoup the insurance company’s
initial expenses and provide a profit.
In addition, because unit-linked business usually gives low guarantees, the solvency capital
requirement is usually low.

Question
One of the advantages of a unit-linked contract over a non-linked contract is said to be better
value for money to the policyholder.
Explain whether this would be expected to be the case for the ultra capital efficient policy
described in the example above.
Solution
A product that is very capital efficient will not need to generate as much profit (per £ of premium)
to service the capital. So it should be cheaper for the policyholder and, in this respect, the more
capital efficient the better.
In practice, products might not be designed in this most capital efficient way, even if legislation
permits it. This is because policyholders might take a dim view if they were to see almost all of
their first year’s premium disappearing in company expenses and profits, and so the policy might
not be marketable.
This of course seems illogical following the arguments of the previous question. Proper education
and advice on the merits of the case would be needed in order to make the product marketable in
practice. It also requires trust in the insurance company that the company will pass back a fair
share of the saving in capital cost that it makes, and not just pay higher returns to its capital
providers. Whether these products can be sold successfully does therefore depend on the
prevailing conditions of the market.
Frequency of premium payment
Question
Set out the following in decreasing order of capital required, for an endowment assurance:
 regular annual payment in advance
 regular monthly payment in advance
 single premium.
Give a brief explanation for your answer.

Solution
Regular monthly premium (most capital needed).
Regular annual premium.
Single premium (least capital needed).
Essentially the ratio of the premiums received by time 0+, to the amount of initial expenses
incurred at time 0+, increases as we go down the list. Typically the asset share at time 0+ would
be severely negative for the monthly premium case (remember only one monthly premium will
have been received), a bit less negative for the annual premium case, and heavily positive for the
single premium contract.
There is some compensation for this, as credit for the expected present value of future premiums
would be taken in the supervisory reserves. However the insurance company may not be able to
take full credit for the future premiums under the first two contracts, as there is a risk that the
policyholder withdraws or dies before the premiums are paid, and so additional solvency capital
or higher reserves may need to be held to cover this risk. But with the single premium case, full
credit for all possible premiums occurs – because they have, of course, all been received.
This assumes that there are no unlikely distortions such as particularly heavy single premium
commission.
Relationship between pricing and supervisory reserving bases

We shall illustrate the key idea under this heading with a simple example. Imagine that a
company is writing single premium without-profits endowment assurances and is aiming to break
even. Let us also imagine that the company’s experience (in particular, the amount of its initial
expenses) is exactly as expected according to its pricing basis for the contract. If the company
calculates its supervisory reserves using its premium basis assumptions, and there is no solvency
capital requirement, then there will be no initial capital strain. This follows from the fact that the
premium has been calculated to pay the initial expenses and then be exactly sufficient, on the
premium basis (= reserving basis), to meet ongoing expenses and the final benefit.
However, if the reserving basis is stronger than the pricing basis then the premium charged will
seem insufficient, on the reserving basis, to meet the expenses and the benefit. Capital will
therefore be needed to set up the required reserves at outset. The stronger the reserving basis
compared with the pricing basis, the more capital is needed.
Question
Explain what will happen if the reserving basis is weaker than the pricing basis.

Solution
The premium would exceed the amount needed to be set aside as reserves and so some profit
would immediately be released to the insurance company’s free assets, rather than capital having
to be found from the insurance company’s free assets.
However the requirement to hold some solvency capital, so that in total the sum of the supervisory
reserves plus solvency capital is prudent, would be expected to prevent this happening.
Additional solvency capital requirements

The supervisory authority will require that each insurance company holds enough assets to cover
the sum of the supervisory reserves and the required solvency capital.
In some countries the supervisory reserves may be large (based on very prudent assumptions)
and only a relatively low level of additional solvency capital may be required. In other countries
reserves may be lower but solvency capital requirements may be much higher. Although these
two approaches may appear to be different, the impact on the total amount of capital the
insurance company needs to hold may be very similar.
We realise that some of the ideas in this and the previous subsection are quite difficult if you are
studying the course for the first time, so we recommend that you come back to this chapter after
you have read Part 4 of the course, which covers supervisory reserving and solvency capital in
detail.
Level of initial expenses

The higher the level of initial expenses, the less the initial asset share will be. If nothing else
changes, then clearly this will increase the capital requirement. If the increase in initial expenses
has been anticipated in the pricing basis – ie the premiums charged include the relevant loading
for these initial expenses – then in theory we can allow for this in the reserves by taking credit for
the expected present value of these loadings that we expect to receive in the future. Hence, if
this credit were to be taken in full, then the reduction in the asset share would be balanced by the
reduction in the (net) reserves required, resulting in no change to the capital requirement. (The
use of negative non-unit reserves is an example of this process: but more about these later.)
However, this rarely, if ever, works out entirely in practice. Firstly, it may not be possible to
reduce the supervisory reserve. This would ultimately depend on the supervisory regulations of
the country in question, but could easily arise if, for example, the asset share is negative and the
supervisory reserve is at its minimum allowable level. Secondly, initial expenses might be higher
than expected in the pricing basis, in which case the future expense loadings are not increased (in
other words, a real loss has been made).

The chapter summary starts on the next page so that you can
keep all the chapter summaries together for revision purposes.

Chapter 1 Summary
Life insurance products are designed to meet personal financial needs, and these needs vary
over personal lifetimes (the ‘personal financial life cycle’) and over calendar time.
The need may be for savings, for protection or a mixture of the two. The contract we
studied in this chapter is primarily a savings contract, but also offers some element of life
cover.
Endowment assurances allow policyholders to save a lump sum by a known maturity date,
with the possibility of a lump sum being paid on earlier death. The life company may also
offer surrender values or ‘paid-up’ values, either on guaranteed or non-guaranteed terms.
Endowment assurances may come in without-profits, with-profits or unit-linked form. They

might be sold to individuals or to groups.
The main risks relating to these contracts involve:

 investment returns
 expenses
 withdrawals (especially when the asset share is negative)
 mortality (including anti-selection risk).
The extent of the risks for the life company depends on the contract design. The risk is
arguably greatest for without-profits contracts (where everything is guaranteed), lower for
with-profits contracts (where bonuses may be reduced) and lowest for unit-linked contracts,
where much of the risk may be passed back to the policyholder.
The capital requirements of the business depend on the:

 contract design
 premium payment frequency
 relationship between the pricing and supervisory reserving bases
 additional solvency capital requirements
 level of initial expenses and initial commission.

The practice questions start on the next page so that you can
keep the chapter summaries together for revision purposes.

Chapter 1 Practice Questions

1.1 A life insurance company writes all types of without-profits non-linked business.
Exam style
State the reasons why new business strain is likely to arise. [4]
1.2 When using an endowment assurance to pay off a loan, a with-profits policy is much more often
used than a without-profits policy.
1.3 Comment on how much risk a life insurance company faces from adverse investment
performance in the cases of without-profits, with-profits and unit-linked endowment assurances.

The solutions start on the next page so that you can

separate the questions and solutions.

Chapter 1 Solutions
1.1 New business strain occurs when the difference between the day-one asset share, and the sum of
the supervisory reserves and the required solvency capital, is negative. Any such strain has to be
made good from the company’s free assets. [1]
For regular premium contracts the asset share is normally negative in the early years because
there is a high level of expenses incurred at the start of the contract, in particular from
commission, sales & marketing, underwriting and policy documentation. [1]
The sum of the reserves and solvency capital is likely to be positive, because these will be
calculated to give a low probability that the insurer is unable to meet its liabilities. Thus liabilities
exceed the assets in value and this is known as new business strain. [1]
The prudence implicit in the solvency calculation (reserves plus solvency capital) is usually greater
than the margins for risk and profit in the premium basis. For this reason it is common to have
new business strain even for single premium contracts, despite the positive net cash inflow. [1]
[Total 4]
1.2 Although the premium for the same sum assured is higher under with-profits, the expectation is
that the final payout on maturity (including the bonuses to be added in future) will represent a
substantially better return on the policyholder’s premiums than under the without-profits
contract. When the loan is paid off, the extra money would be payable to the policyholder.
Alternatively, a smaller policy could be taken out to start with (with smaller premiums and smaller
initial sum assured), on the basis that this sum assured plus expected future bonuses will together
be enough to pay off the loan at maturity, and give the possibility of additional benefit to the
policyholder. This leads to a policyholder risk, however, because the future bonuses would not be
guaranteed to be enough to pay off the loan at maturity.
This risk has, unfortunately, become reality for many endowment mortgage policyholders over
recent years in the UK. The repercussions for the UK life insurance industry have been severe.
Much bad press has arisen from allegations of mis-selling, with insurers and their intermediaries
accused of not properly communicating the risks of these arrangements to policyholders. The end
result has been a virtual cessation of all such business (for this purpose) in the UK market.
In summary, it is the possibility of additional reward that (in theory) makes the with-profits
version attractive. The non-profit version offers nothing to the policyholder that a standard
repayment mortgage plus decreasing term assurance does not. In fact, the endowment route is
probably worse, as it introduces (to the policyholder) the third party risk that the insurer will
default before the mortgage is repaid, and probably incurs higher costs via the expense and
guarantee loadings in the premium rate. (We describe decreasing term assurances in Chapter 2.)
Don't worry if all this wasn't obvious straight away. It should become clear when you study
with-profits policies in detail later in the course.

1.3 Don’t be discouraged if you only got a few of the points made below. Some of the issues are
picked up in more detail later in the course. However, it is important that you think about these
issues first, rather than simply have the notes tell you everything. Treat the question as a learning
experience.
On a without-profits contract the benefit is guaranteed, and so if the achieved investment returns
are lower than allowed for in the premiums then the insurance company will make a loss, all
other things being equal.
The investment risk is rather lower on with-profits contracts. If investment returns are lower than
expected then the insurance company may be able to reduce bonuses.
However:
 There is still a guaranteed element that the insurance company must meet.
 Policyholder expectations may limit the scope and speed of bonus reductions.
 It is possible that shareholders (where there are any) share in the investment profits of
the company. If so, they would lose from reduced investment return.
 Poor investment returns may discourage future sales of with-profits policies. The extent
to which this is true depends on the market and on the way in which the policies are sold.
(Because this affects sales, we can actually think of this as a marketing risk.)
On unit-linked business the investment risk is essentially borne by the policyholder, not the
insurance company. However:
 The insurance company will have to meet the excess of any guaranteed benefit over the
value of the units (although many contracts do not contain guarantees of this type).
 If there are charges linked to fund size then these charges are reduced by poor
investment performance.
 As for with-profits contracts, poor performance relative to the competition may damage
future sales.
Even so, the direct exposure of the insurance company to investment risk on unit-linked contracts
is usually lower than for without-profits or with-profits contracts.


Syllabus objectives
 with-profits
 unit-linked
 index-linked.
insurance products.

0 Introduction
This chapter covers products paying out on death of the life insured.
The products covered are:

 whole life assurances
 term assurances
 convertible or renewable term assurances.

1 Whole life assurances

A whole life assurance is a contract to pay a benefit on the death of the life insured
whenever that might occur. Typically, a surrender value would also be available.
As with endowment assurances, there may also be a ‘paid-up policy’ option.
For consumers, it is useful as a means of providing for funeral expenses or for meeting any
liability to tax, for example inheritance tax or death duties, arising on the death of the life
insured. It is a general purpose contract for providing long-term protection for dependants.
In the last respect a whole life assurance is particularly useful as a means of protecting some of
the expected transfer of wealth that a parent would be aiming to make to his or her children
when he or she died. Without the contract, the wealth transfer would be likely to be very small if
the parent died young. Such contracts can also be a tax efficient way of transferring wealth, at
any age, depending on legislation (often reducing the liability to death duties or inheritance tax).
Question
Discuss whether a whole life contract is primarily a savings contract or a protection contract.
Solution
On one hand, quite significant reserves can build up on policies of long duration in force, and the
contracts could be seen as ways of saving for a known outgo such as funeral expenses.
However, given that the contracts pay out on death and not at a fixed time it is probably better to
think of the contract as providing protection.
We shall be concentrating in this section on without-profits and with-profits versions of the

contract. The unit-linked version can be sufficiently different to need separate consideration. We
shall consider it further in Chapter 4.
Without-profits contracts offer a guaranteed sum assured on death. On with-profits whole life
contracts the initial benefits may be increased by bonuses, or cash refunds may be given.
There would not seem to be a consumer need for a group version of this contract.
Group contracts are usually funded by employers for their employees. It is therefore natural to
restrict life cover to the period of employment, rather than offer it on a whole of life basis. We
shall therefore consider only individual whole life contracts.

1.2 The risks

The relative and absolute significance of the investment and mortality risks depend on the
age at entry and the duration in force of the contract.
Examples of these relationships can be found in the Practice Questions at the end of this chapter.
A mortality risk may also arise from selective withdrawals.
The policyholders most likely to withdraw from the contract are those in good health, leaving the
insurance company with a sub-standard group of lives. It can try to allow for this in the premiums
it charges and in the terms it offers to withdrawing policyholders, but it cannot completely
eliminate the risk.
(Selective withdrawal is actually a form of anti-selection; that is, where the actions of individuals
(policyholders) adversely affect the financial condition of the company.)
The risk of selective withdrawals is not unique to whole life contracts. The risk can be important
for any contract offering high levels of protection.
There will also be an anti-selection risk.
Question
Explain how the anti-selection risk arises.
Solution
The risk is that those who take out the policy are the ones who ‘expect’ to have heavier than
average mortality. Without underwriting, the insurance company would suffer earlier claims, on
average, resulting in lower profits.
The effects of inflation with regard to the expense loadings in the premium payable and the
size of the premium itself can lead to a significant expense risk for contracts of long
duration in force.
A contract taken out when a policyholder is aged thirty may still be in force when he or she is
ninety. For a without-profits contract in particular, if paid for by level, guaranteed premiums, this
produces considerable exposure to inflation far into an unpredictable future.
It is even possible that at long durations the costs of administering the policy may be greater than
the amount of premium collected. This is one reason why whole life contracts are often written
with premium payments ceasing at some age, for example at eighty. (Another reason is that at
very long durations, policyholders may find that they have paid more in premiums than the
amount of sum assured, which is not good for customer relations.)
At times when the asset share is negative, there is a financial risk from withdrawals. At
other times, whether there is such a risk depends on how any withdrawal benefit paid
compares with the asset share.
So the issues here are much as for an endowment assurance - and just about any other contract.


Capital requirements might be similar to those of an endowment assurance contract.
Question
State the five main factors affecting capital requirements.
Solution
The capital requirements of the business depend on:

 the contract design (in particular the size of reserves that the design requires)
 premium payment frequency (in particular whether single or regular premium)
 the relationship between the pricing and supervisory reserving bases
 the additional solvency capital requirements
 the level of initial expenses (including any initial commission).

2 Term assurance

A term assurance is a contract to pay a benefit on the death of the life insured within the
term of the contract (chosen at outset). Typically, no benefit is available on surrender.
The benefit on death is often, but not necessarily, a lump sum. The term of the contract can be
anything from one year to twenty-five years or more. The lump sum usually remains at a fixed
(ie level) monetary amount throughout the policy term (in which case the policy is often referred
to as level term assurance). Premiums may be paid regularly over the term (or some shorter
period) or as a single premium at outset. For regular premiums, the premiums may be level or
increase at regular intervals with age.
For regular premium term assurances, premiums are usually guaranteed. Alternatively, it is
possible for companies to issue contracts with reviewable premiums, where premiums can be
changed at intervals during the policy term, in response to changing claim experience.
Generally, there is no payment if the policyholder survives to the end of the contract, although
variations have been offered in some markets where some proportion of premiums paid (usually
a token sum) is payable on survival.
Question
Term assurances don’t usually pay any benefits on withdrawal.
Explain why most policyholders will accept this as perfectly reasonable.
Solution
Policyholders naturally tend to visualise their level (or increasing) premiums as paying for their
level (or increasing) death benefit each year. It will therefore come as no surprise to them that no
compensation is available from past premiums on withdrawal.
In most markets term assurance contracts are without-profits, although with-profits, unit-linked
and index-linked versions could be offered.
Where the contract is taken out by an individual, it provides protection against financial loss
for dependants, at a lower cost than under an endowment or whole life assurance contract
for the same level of benefit.
Question
Explain why the cost is lower for a term assurance.

Solution
The cost is lower because on endowments and whole life contracts a claim is expected on every
contract written (with the exception, perhaps, of early withdrawals). On a term assurance, claims
are expected only on a relatively small proportion of policies written, because there will be no (or
only a token) maturity payment.
The most common alternative to a level term assurance is a decreasing term assurance.
A decreasing term assurance (ie the benefit amount reduces periodically during the policy
term) can be used to meet two specific such needs. First, it can be used to repay the
balance outstanding under a repayment loan and, secondly, it can be used to provide an
income for a family with children, until they become independent adults.
Example
A decreasing term assurance could be structured as follows to protect a young family against
financial loss on the death of the main breadwinner.
Policyholder: Aged thirty years
Term of contract: Twenty years
Benefit: $5,000 per month from death until the end of the contract term; no
benefit on survival of life assured to the end of the term
Premium: $200 per month until the earlier of death or fifteen years
The income payable from the date of death until the end of the term is equivalent to a decreasing
lump sum payment, hence the contract is a decreasing term assurance. In many countries, such
contracts are known as family income benefit policies.
Where the policyholder is a corporate body or partnership, the contract can be used to
provide protection against the financial loss that might arise on the death of a key person
within the organisation.
The company or partnership takes out a contract (which could be a level or decreasing term
assurance) where the life insured is a specified individual who is thought to be key to the success
of the organisation. The contract might give the company protection against such things as the
loss of important contracts, delay to a major project, and the cost of recruiting a suitable
replacement.

The group equivalent of the term assurance contract can be used by an employer to provide
a benefit to dependants on the death, whilst in employment, of an employee.
Question
Two possibilities for meeting an employer’s needs here would be:

(a) one-year grouped term assurances, renewable on non-guaranteed terms
(b) a collection of longer term assurances, covering each employee up to his or her expected
retirement date.
Suggest one major advantage for the employer of each approach.
Solution
(a) The major advantage of the one-year assurances would be their greater suitability for a
mobile workforce. This approach would also handle salary increases better, if the benefits
were linked to salary. (In fact, variations on the yearly term assurance approach are used
in a number of markets for group life cover.)
(b) The main advantage for the employer of the longer term assurances would be the known
cost of cover, at least for current employees. It could also be simpler to administer.
(These would only really be advantages if the death benefit was fixed in monetary terms,
at least over significant periods of time, and was not directly linked to salary.)
A group term assurance can also be used by a credit card company, for example, to provide
a benefit on death equal to the balance outstanding on a credit card.
Thus the insurance company in effect indemnifies the credit card company against the loss when
a credit card balance is not repaid because of the death of the credit card holder. The benefit
could be defined, for example, as the full amount outstanding on death, or as the average
monthly balance over a period prior to death.
2.2 The risks

The main risk for the company with both the individual and group versions of the contract
arises from mortality. There will be a significant anti-selection risk with the individual
version of the contract and a much reduced such risk in the case of the group version.
Question
Explain why anti-selection risk:

 is particularly important for term assurance contracts
 is much reduced for group contracts.

Solution
Anti-selection is a particular worry because there is so much to gain for the policyholder (or rather
the policyholder’s dependants). Large sums assured can be obtained for relatively small
premiums.
Anti-selection is less of a concern on group business because of the reduced element of choice for
the individual. For example:
 Membership of the group may be compulsory, for example for all employees of a
company.
 There may be restrictions on the level of cover each member can have (perhaps related to
salary).
In addition, there will be a mortality risk from selective withdrawals.
This is one reason why companies are unlikely to offer any surrender or paid-up benefits on
withdrawal.
With both types of contract there is a financial risk from withdrawals at times where the
asset share is negative. The risk is exacerbated in the case of decreasing term assurances
if the cost of benefit exceeds the premium being charged early in the term.
Withdrawal risk under decreasing term assurance contracts

If level premiums are paid over the whole term of the contract, then it is quite likely that the cost
of the (very high) benefits payable on death in the early policy years would exceed the level
premium payable. Later on in the term the premiums would exceed the cost of the (much lower)
death benefits, and these would (in theory) make good the earlier shortfalls.
This has the effect of producing a negative asset share for a significant part of the term of the
policy, which means that there will be many more opportunities for a loss to occur on withdrawal
(except near the very end of the policy term).
But there is a further factor which makes the financial withdrawal risk much greater: as the policy
progresses, and the sum assured reduces, the policyholder has less and less of an incentive to
continue paying premiums.
There would probably even be a point at which the policyholder would be better off lapsing the
policy and taking out a new level term assurance for the same premium for the rest of the term –
he or she could end up with a higher level of cover by doing so. There is therefore a financial
incentive to withdraw (often referred to as a lapse and re-entry option), and furthermore,
exercise of that option would almost certainly lead to losses for the insurer.

The lapse and re-entry problem can be avoided by limiting the term over which premiums are
paid (which of course increases the regular premium payable). This also reduces the period over
which the asset share is negative, and would be designed to produce positive asset shares (and
hence profits) at a much earlier point in the policy term. Hence this one product design feature
reduces two aspects of the withdrawal risk: it reduces the amount of loss on withdrawal (by
increasing the asset share), and it reduces the incidence of withdrawal (by removing the financial
incentive).
Expense risk for term assurance contracts

With both types of contract there will be an expense risk in that the actual marginal costs of
administering the contract need to be met.
There is expense risk for both group and individual business. It can be substantial, particularly if
the policy term is long, as the premiums are often relatively small and may well be of fixed
amount throughout the term of the policy.
Question
Explain why there is no mention of investment risk for term assurances.
Solution
The funds that build up under a term assurance are small compared with the sum assured. The
funds are generally too small for the investment return to have a significant effect, particularly if
non-volatile assets are held.

The basic reserves for term assurance contracts are relatively small, but solvency capital
requirements can be more significant in some jurisdictions.
As mentioned earlier, most supervisory authorities require that additional assets in excess of
reserves be held as an extra ‘cushion’ against adverse experience. Thus a minimum amount of
solvency capital is required. There is a great variety of approaches by the supervisory authorities
of different countries. Generally, relatively low levels of prudence in the reserving bases are (or at
least should be) balanced by relatively higher solvency capital requirements.
The capital needed to meet the solvency capital requirements is in effect every bit as tied up as
capital directly backing the liabilities, and so solvency capital requirements increase the initial
capital strain.
For example, in the EU the required solvency capital is calculated as the amount of capital the
insurance company needs so that its assets will exceed its liabilities in one year’s time with
probability of 0.995, ie the amount of capital to withstand an event that happens once in every
two hundred years. So a company that specialised in term assurance might need to hold
sufficient capital to withstand much higher mortality rates than it would normally expect, eg to
withstand the kind of epidemic that occurs only once in every two hundred years.

We discuss solvency capital requirements in much more detail in Chapter 20. In particular, we will
describe the Value at Risk approach that was introduced in the example above.
Initial capital strain can also arise for regular premium policies due to high initial expenses,
such as underwriting costs and initial commission.
Question
Discuss whether capital requirements are high or low for term assurances.
Solution
Under some supervisory regimes, the capital required for single premium term assurance will be
low.
However, where solvency capital is required and up-front commission is paid, capital
requirements are large for regular premium policies when compared with the level of premium.

3 Convertible and renewable term assurance

A renewable term assurance is a term assurance with the option to renew (ie take out a further
term assurance) at the end of the original contract. The appeal of this option lies in the fact that
the renewal can be made without further medical underwriting. (Although in South Africa, for
example, renewable contracts sometimes allow for an AIDS test at renewal.)
One would also expect some sort of guarantee concerning the premium rates for the renewal
contract, for example that they will be the same as those for new business at the time at which
the option is taken up.
A convertible term assurance allows the policyholder to convert the term assurance into another
type of contract, such as a whole life or endowment assurance. At what point(s) conversion is
allowed will vary depending on the particular policy conditions. Conversion may be allowed on
only one date, on any of several dates, or at any time during the original term assurance contract.
For individuals, these contracts combine the attractions of a term assurance, in terms of
obtaining low-cost death cover, with the certainty of being able either to convert to a
permanent form of contract, ie an endowment or whole life assurance, when it can be
afforded or to renew the original contract for a further period of years, all without health
evidence being provided (unless the benefit level is increased).
The Core Reading comment ‘when it can be afforded’ reflects the fact that for a given sum
assured the premiums for an endowment or whole life contract will be considerably more than
for a term assurance.
A particular contract may offer only the renewal option, only the conversion option, or both.
There would not seem to be a need for a group form of these contracts, other than a
continuation option following cessation of employment.
A continuation option allows employees leaving an employer (and hence leaving a group contract)
to take out an individual policy without medical underwriting. The policy is usually a whole life or
endowment assurance.
Surrender values would not normally be paid pre-conversion.
This is for the same reasons as for normal term assurances.
3.2 The risks

These contracts have the risks associated with the underlying term assurance contract.
The existence of the option to convert or renew gives rise to a significant additional
anti-selection risk at the date of exercise.

Question
A life insurance company allows policyholders to renew their term assurance policies, at the
normal premium rates it would charge for new policies at the date of renewal. No further
underwriting is required, even though the company expects those taking up the option to be less
healthy on average than its normal insured population.
Explain why this might be less dangerous than it sounds.
Solution
The company can charge for any expected loss following renewal by charging an extra premium
over the course of the original contract. (Deciding what amount of extra premium to charge is
covered later in the course.)
Also, the renewal may involve much lower expenses than completely new business, and so the
excess expense loading in the normal premium rates can provide some cushion against the
mortality loss.

Capital requirements may be higher than for a basic term assurance, depending on the
extent of any additional reserve required.
Question
Explain why an additional reserve might be required:

 at the start of the contract
 towards the end of the contract.
Solution
Higher reserves at the start of the contract
At outset, the reason is not because the expected outgo after renewal or conversion is expected
to be high due to poor mortality experience. This higher outgo should be covered by a loading in
the future premiums. (The exact effect of this will depend on the particular rules for calculating
supervisory reserves for contracts with options.)
However, the existence of options (which are, in effect, additional guarantees) creates extra risks
and uncertainty with regard to future costs. So the insurance company may wish to hold higher
reserves out of prudence (or alternatively, may wish to hold more solvency capital), or may be
required to do so by the supervisory authorities.

Higher reserves towards the end of the contract
By the time the contract comes to the renewal date, the company must have sufficient reserves
available to pay for the additional cost of the higher future mortality experience of those
policyholders who exercise the renewal option. The company should be able to pay for these
reserves from the premium loadings that have been received in the premiums of the original
policy.

Chapter 2 Summary
Whole life assurance
A whole life assurance has no fixed term and pays out when the policyholder dies. It is most
useful for meeting payments that may arise on death, such as funeral expenses or
inheritance tax, or for transferring wealth between generations.
Mortality and investment risks are both significant for the non-linked versions of the
contract, although how important they are varies by age at entry and duration in force.
Withdrawals and expense inflation are also risks.
Capital requirements are similar to endowments and may be significant.
Term assurance
This provides a benefit on death within a specified term. The payment might be a lump sum
or an income for a specified period.
If the income is for the outstanding term of the contract, the contract is effectively a form of
decreasing term assurance. A decreasing term assurance could also take the form of a lump
sum, for example to pay off the balance of a loan.
Term assurances could be sold in group form, for example to enable an employer to offer life
cover within an employee benefits package.
The main risk on term assurance is of worse-than-expected mortality experience.

Withdrawals and expenses also present risks of loss.
There is a significant risk from selective withdrawal.
Contracts may contain options to renew at the end of the term, or convert to a whole life or
endowment assurance contract, on guaranteed terms. The existence of the options,
particularly renewal, leads to increased anti-selection risks.
Capital requirements may be high depending on the supervisory reserving, and solvency
capital, requirements.



2.1 Comment on the extent of the mortality and investment risks that a company faces on:
(a) A portfolio of regular premium without-profits whole life contracts taken out thirty years
ago by men then aged thirty.
(b) A portfolio of regular premium without-profits whole life contracts taken out in the last
year by men aged sixty.
2.2 Asset shares under (level) term assurances tend to be positive for roughly the final two-thirds or
three-quarters of the policy term.
Outline possible justifications for the company to not pay out at least some of this asset share
when a policy withdraws at later durations.
2.3 State, with justification, which option under a term assurance product is likely to pose the greater
risk to the life insurance company: that to convert or that to renew.
2.4 A young couple are looking for appropriate life insurance. They have just bought their first house
Exam style
and have two young children. One of them is in paid work, the other looks after the children.
(i) Explain why a convertible term assurance policy may be the most appropriate product for
this couple. [5]
(ii) Discuss whether you would expect the product to be written on a single life, joint life
payable on the first death, or joint life payable on the second death basis. [4]
[Total 9]



Chapter 2 Solutions
2.1 (a) Now aged sixty, policies written thirty years ago
The mortality risk is likely to be low, despite the fact that the policy was underwritten many years
ago and the portfolio will include a mix of healthy and unhealthy lives. The reason is because the
policy will have accumulated a large reserve by this point.
There is the risk to the company’s capital from the strain (loss) caused by a death claim, ie from
the excess of the death benefit paid over the supervisory reserves (and any solvency capital
requirement). The reserve will be large enough to cover the expected strain according to the
mortality rates assumed in the reserving basis; only where the actual death strain is greater than
this (and the associated solvency capital) will the company have to provide additional finance.
This risk will not be very large if the reserving mortality basis is prudent (or if substantial solvency
capital can be released on death).
In the above we have considered the risk of a capital strain caused by earlier death. There is also
the risk that early deaths lead to a loss of future profits – we will return to this risk in Chapter 18
when we consider embedded values.
On the other hand, the large fund means that there is an investment risk. (The techniques of
matching and immunisation, which you have studied in earlier subjects, may enable this risk to be
limited.)
(b) Aged sixty, policies just taken out
As the policyholders have been recently underwritten, the probability of claiming is lower than in
(a). However, the risk of making losses from higher than expected mortality remains high,
because the sum assured is very large relative to the size of the reserve.
The risk of needing additional capital on death is again mitigated by any prudence in the reserving
basis and the ability to release the solvency capital associated with the contract on death.
There is little risk related to the investment of the accumulated funds (there aren’t any). There is
a risk relating to the unknown rates of return that may be earned on future premiums.
2.2 The following are possible reasons:

 The company will have made a loss on early withdrawals where the asset share would
have been negative, often substantially so. So the company feels justified in recouping
some of these withdrawal losses by making profits from later withdrawals.
 There will be a significant selective withdrawal effect. The retained asset share can then
go towards meeting the increased costs due to the higher average mortality of the
remaining insured lives.
 Asset shares are never very large on a term assurance contract, and payment of very
small surrender values would not justify the cost of administering them.

 Term assurance asset shares will tend to be quite volatile (fluctuating with claims
experience each year), making any fair scale of surrender values equally volatile. This
could seriously upset policyholders.
 Even at later durations asset shares are not always positive – eg where the policy is loss-
making.
We would be surprised if you got all of the above points. Don’t worry: we suggest you look at the
question again once you have studied the chapter on asset shares (Chapter 5).
2.3 The option to renew poses the greater risk, because the term assurance offers the greater
incentive for anti-selection.
For example, a policyholder who converts to an endowment or a whole of life contract will
immediately start to pay a much higher premium for the same level of cover, which will quickly
serve to reduce the death strain at risk and hence the mortality risk in total. Such policyholders
are unlikely to be in poor health anyway, because if they were they would choose the cheapest
way of maximising their death benefit in the short term. This would be achieved either by not
converting the original term assurance policy in the first place or, if the end of the term has been
reached, by renewing into another term assurance. Hence, the renewal option imposes (by far)
the higher anti-selection risk.
2.4 (i) Why convertible term assurance?
To explain why a convertible term assurance is the most appropriate product, we also need to
consider why other products are not as suitable.
Why term assurance?
A term assurance policy provides a lump sum benefit on death. Such a lump sum would be able
to pay off any mortgage on the house and pay for bringing up the children until such time as they
can be independent financially. [1]
Since a term assurance pays out only on death, it is much cheaper than alternative savings
products, such as an endowment or whole of life policy. [½]
The couple are unlikely to have much extra disposable income after financing the house purchase
and paying family bills each month. The couple are young and so the wage earner is likely to still
be in the early stage of their career, being paid accordingly. [½]
Therefore, savings products probably aren’t appropriate for the couple at the moment. [½]
It is vital to protect the income of the lone wage earner. [½]
If the mortgage is a ‘repayment’ mortgage (monthly payments include interest and capital
elements), then a decreasing term assurance might be appropriate, … [½]
… but if it is ‘interest only’ then a level term assurance would be better. [½]
The total remaining cost of looking after the family until the children are self-sufficient will be
decreasing with time, so this need is best met by a decreasing term assurance. [½]

A without-profits version is the most common type of term assurance policy, but the death
benefit will be eroded by inflation over time. [1]
It is possible to get unit-linked or with-profits term assurances, but these are much more
complicated to understand and carry more investment risk. [1]
Why convertible?
The couple might be looking to effect a savings policy at some time in the future, when their
disposable income has increased (due to pay rises in excess of inflation or the homemaker going
back to work), or their expenditure decreased (due to children moving out). [1]
However, they might be worried about the possibility of not meeting the conditions for standard
terms. This might be because they’re worried about becoming ill or about underwriting standards
becoming more stringent in the future. [1]
Opting for a convertible policy gives them the ability to take out a savings policy at some future
date without having any further medical underwriting. [½]
It might also be possible to take out further term assurances at set dates in the future, giving
further flexibility to meet changing needs and the option to mitigate the effect of inflation. [½]
[Maximum 5]
(ii) Basis of life cover
Cost
A joint life, first death policy will pay out as soon as either of the couple dies and so will be the
most expensive option. [½]
A joint life, second death policy won’t pay out until the couple have both died, and so will be the
cheapest option. [½]
A single life policy will pay out on the death of the life assured, which could be the first or second
death. The cost of this option will therefore be in between the other two. [½]
Benefits
If the first death is the wage earner, then all the household income will be lost. [½]
If the first death is the other life, there will be no-one at home to look after the children, so child
care will be needed. [½]
A joint life, first death policy will provide capital to generate income for living off or for child care.
[½]
If the couple has family who would be willing to provide child care, then a single life policy might
meet the couple’s needs. [½]
This would probably be on the life of the wage earner, but if the other parent could get cheaper
cover and would be willing and able to generate a suitable income, then their life might be
insured instead. [1]

If the wage earner’s employer provides a sufficient death-in-service benefit, then the couple
might decide to insure just the other parent’s life. [½]
A joint life, second death policy wouldn’t pay anything on first death, so would only be suitable if
there’s a plan in place to cope with this situation, eg grandparents willing to look after children,
an employer’s death-in-service benefit and/or the other parent able to go back to work. [1]
[Maximum 4]


Syllabus objectives
 with-profits
 unit-linked
 index-linked.
insurance products.

0 Introduction
This chapter covers products paying out on continued survival of the life insured.
The products covered are:

 immediate annuity contracts
 deferred annuity contracts.

1 Immediate annuity contracts

An annuity is a contract that pays out regular amounts of benefit.
An immediate annuity is a contract to pay out regular amounts of benefit provided the life
insured is alive at the time of payment. The word ‘immediate’ indicates that the contract
starts payments immediately, without a deferred period. Payments may be made in advance
or in arrears, so that the maximum period until the first payment is the frequency period of
payment. For example, annually in arrears would mean payment occurs at the end of each
year.
So, someone who buys an immediate annuity payable monthly in arrears will have to wait one
month before receiving the first benefit payment.
As the purpose of these contracts is to provide regular benefit payments, the contracts will
be purchased in advance by a single premium. This premium may be the proceeds of
another regular (or single) premium contract.
For example, it may come from the maturity proceeds of an endowment: see Section 2.

The main purpose of the contract for the consumer is to convert capital into lifetime income,
and remove the uncertainty of how quickly capital should be spent to provide income over
the consumer’s remaining lifetime. The pensions legislation of a country may specify that
an annuity must be purchased from part or all of a pension fund accumulated to the
retirement date. An immediate annuity ensures that the individual provides for a regular
pension in retirement rather than allowing the capital to be otherwise spent in the short
term.
The pension fund referred to by the Core Reading would again most likely be the proceeds of an
endowment.
When not compulsory, the main reason someone would wish to buy an immediate annuity is
therefore to protect the level of their future income (and hence standard of living) from the risk of
living too long.

Features of immediate annuities

The main features of immediate annuity contracts are given in the table below:
 single life
Basis of purchase  joint life first death
 joint life last survivor (can be used to provide income for
dependants)
Several options:
Duration of cover
 temporary for specific purpose (eg for school fees of
insured’s children)
 fixed for payments to be made for a set number of years,
irrespective of insured’s survival in that period
 guaranteed for shortfall between premium paid and annuity
received to become payable on death of insured within
specified period.
Level or variable It is possible for payments to be fixed or inflation-linked to protect

payments? against the erosion of benefits by inflation.
Not normally available, although it may be possible to ‘sell’ an

Surrender values?
annuity in payment within a secondary market and convert
payments to a cash lump sum.
So immediate annuities may be purchased on a single life, joint life first death or joint life last
survivor basis. For example, a last survivor annuity can be used to provide for dependants’
income following the death of the main life. When an annuity has a last survivor benefit, the level
of payment may reduce after the first death.
Most annuities pay the benefits for as long as the insured life or lives survive. Payments then stop
on death. However, there are some other possible durations of cover, which we discuss next.
A temporary annuity would usually be paid until death or the end of the policy term, whichever
came first. However, it is possible for the term of the annuity to be fixed, so that payments would
continue to be made until maturity regardless of whether the insured had died.
Even for annuities that do not have a fixed term, there may be a guarantee that, on death within a
specified period, any shortfall between the initial premium paid and the amounts of annuity
received to the date of death will become payable.
Another alternative is for the payments to be made for an initial number of years irrespective of
whether the life insured survives the initial period. For example, the annuity could be paid for five
years or until death, whichever is longer.

Question
(i) Explain why a minimum payment period, and a return of the balance of premium on
death, are common features of annuities.
(ii) State the main disadvantage of having such a feature.
Solution
(i) If the annuity had neither of these features and the annuitant died a short time after
taking it out, then the annuitant would have received very little money in return for the
probably substantial single premium that would have been paid. This can lead to bad
feeling between the insurance company and any surviving relatives of the annuitant.
(ii) It makes the annuity more expensive (the annuitant gets a smaller annuity for the same
single premium), assuming the company reflects the extra benefit in its premium rates.
The regular benefit payments may be level or variable. The most common way to make the
benefits variable is to link them to an inflation index.
Surrender values are not normally available.
Question
Explain why surrender values are not normally available.
Solution
It is because of the anti-selection risk. The policyholder who stands to gain most from
surrendering an annuity is the one who is in severe ill-health and may not expect to receive many
more annuity payments anyway.
This is a very strong argument for not offering a withdrawal benefit on individual annuities.
However, it may be possible in some jurisdictions to ‘sell’ an annuity in payment within a

secondary market and thus convert the outstanding regular benefit payments into a cash lump
sum.
The purchaser will need to assess the health of the policyholder and will pay substantially less to
those in ill-health.
A group version of the contract can be used by an employer to fund for pensions for
employees at or after retirement.
Essentially, such contracts are just collections of individual annuities.

In many markets annuity contracts are predominantly conventional without-profits or index-

linked, though with-profits and unitised annuities are also possible. For a with-profits annuity the
income paid to the policyholder will be a guaranteed amount plus a bonus added by the insurer.
For a unit-linked annuity, the insurer guarantees paying the value of a number of units. However,
the policyholder’s income is not guaranteed (in monetary terms), because the income is equal to
the number of units multiplied by the unit price, which will vary on a daily basis.
In some countries, so-called impaired life annuities are offered. Applicants who can demonstrate
that their health is significantly impaired can be given higher levels of annuity benefit, on the
expectation of shorter lifetime.
1.2 The risks

The main risk, usually, with these contracts is a longevity risk, particularly in understating
the rate of improvement of life expectancy.
If annuitants live longer on average than the life insurance company has allowed for in its pricing,
then the insurance company will lose money.
Associated with this is an anti-selection risk, the extent of which will depend on the extent
of free choice available to the policyholder regarding the purchase of the contract.
Question
(i) Explain the nature of the anti-selection risk here, assuming that individuals are free to
choose whether or not to buy an annuity.
(ii) Compare the nature of the anti-selection risk on an annuity to that on a term assurance.
Solution
(i) The risk is that when there is the choice, those most likely to buy an annuity are
individuals who are in reasonably good health, or who have a family history of longevity.
This might be termed ‘self-selection’. (Of course, the insurance company will try to allow
for this effect in its mortality assumptions.)
(ii) The principle as applied to annuities is not really very different from that applying to term
assurances. The individual is using knowledge about his or her own state of health to
decide whether or not to take out a contract.
However, the selection is rather less active than it can be on term assurance, for which
insurers have to be on the lookout for downright dishonesty on the part of applicants.
Applicants for immediate annuities are unlikely to be rejected for being in exceptionally
good health and, indeed, it can be difficult to identify such individuals.
The less free choice there is, the greater will be the extent of compulsory purchase and the less
the company will suffer from anti-selection, as the policyholders involved will have to buy an
annuity whether or not they are in good or bad health.

In addition, there will be investment and expense risks.
Investment risk
The nature of the investment risk will depend on the extent to which the annuity payments under
the contracts have been ‘matched’ by suitable assets.
We will cover the subject of matching in detail much later in the course. However, it is useful just
to give an outline of what happens at this stage.
Under an immediate annuity, a lump sum is available for the company to invest at the start of the
policy. By the time the ‘average’ policyholder has died, this money should have been used up in
providing the regular annuity payments and paying the expenses incurred over the lifetime of the
policy, (assuming that the company has priced its annuities correctly). The income earned on the
lump sum will usually be insufficient on its own to cover the annuity and expense outgo, so each
year part of the lump sum will be disinvested in order to fund the balance. Assuming that
annuitants die exactly as expected, then the fund will decline to zero just as the last annuitant
dies.
In theory, the insurance company could invest the lump sum in fixed interest securities at the
start of the contract, choosing a spread of redemption dates such that the redemption amounts
being paid each year are exactly equal to the amount by which the fund needs to be disinvested
in that year. Hence, again assuming that the annuitants die as expected, the company will earn
exactly the redemption yields currently obtainable at the start of the contract and there will be no
investment risk.
Question
Explain why there will not be any investment risk (ie explain the last sentence above).
Solution
There is no investment risk (in theory) because the company should not need to buy, or sell, any of
its assets during the term. With fixed interest securities (ignoring risk of default) investment risk
arises due to changes to market prices (and hence to the current redemption yields). Hence if the
company has to sell its assets after yields have risen then it has to sell for lower value, thereby
making a loss; if the company has to buy assets when yields have fallen, then the yield on this
investment will be lower than was initially expected in the pricing basis, hence a loss also.
This is an example of perfect matching in theory. Of course in practice the annuitants will not die
exactly according to plan, so that some assets will necessarily have to be bought and sold during
the policy term. Nevertheless, the financial impact of this mismatching risk should be slight. The
major practical difficulty may be a shortage of appropriate securities – as will be considered when
we revisit this topic later in the course.

Expense risk
To understand the expense risk we need to understand how the expenses will be paid for under
the contract. As there are no regular premiums, all the future expenses have to be met from the
single premium. Hence part of the single premium will be invested (to provide an expense
reserve) and this will be drawn upon over the lifetime of the policy to pay the expenses as they
arise.
There will therefore be a risk that this reserve will prove inadequate, which can be caused by
future inflation of the company’s expenses (at higher than the expected rate), or by the company
operating less efficiently than it expected (eg causing fixed costs to rise as a proportion of the in
force business).
As with any contract, there is also the risk that the initial expenses will be higher than the
allowance made for them in the pricing basis.
Withdrawal risk
A withdrawal risk would arise if withdrawal is permitted.
Question
(i) Explain how an insurance company could permit benefits to be paid on the withdrawal of
an annuity, without risking very heavy losses.
(ii) Explain the nature of the withdrawal risk in this case.
Solution
(i) The payment of a withdrawal benefit would be possible only if the company could
ascertain evidence of good health at the time of withdrawal, ie withdrawal would need to
be subject to medical underwriting.
(ii) There will still be an anti-selection risk, but the extent of this has been much reduced by
underwriting the withdrawal benefit. Other than this, the company risks making a loss to
the extent that the lump sum surrender value exceeds the asset share (if at all).

The contracts can give rise to significant capital requirements, depending on the
relationship between the pricing and supervisory reserving bases and on the additional
solvency capital requirements.
Capital strain can arise on single premium contracts even though the full premium for the
contract is received at outset. If the supervisory authorities require the insurance company to set
up reserves and solvency capital larger than the single premium (less any expenses), then capital
strain does arise.

Question
A life insurance company sells an immediate annuity assuming initial commission and expenses of
5% of single premium. (Renewal expenses are ignored.) The interest rate used for pricing is
5.5% pa. At this rate, a life annuity of 1 pa using the premium basis mortality is 11.39 for the
policyholder concerned.
The insurance company is required to reserve at a valuation interest rate of 4% pa. The life
annuity for the policyholder at this rate of interest and using the valuation mortality basis is
12.85. The insurance company is also required to hold solvency capital of 4% of reserves.
On sale, the commission and expenses were equal to those assumed in the pricing basis.
Calculate the capital strain on the sale as a percentage of the single premium.
Solution
Answer: 16.5%
This is derived by considering all of the cashflow components and reserving / solvency capital
requirements, as shown below.
For a premium of 100:
Positives: Premium 100.0
Negatives: Commission and expenses 5.0

Setting up reserve 107.2
Required solvency capital 4.3
Total: –16.5
where the reserve is derived as (100 – 5)  12.85 / 11.39 = 107.2
and the solvency capital is derived as 107.2  4% = 4.3.
So what does this mean? If the life insurance company wants to sell £100 million of such policies
this year, it will require £16.5 million to finance those sales. Therefore the difference between
the pricing and the reserving bases costs 12.2% (from 107.2 minus 95) – although we should bear
in mind that these numbers are entirely artificial.

2 Deferred annuity contracts

A deferred annuity is a contract to pay out regular amounts of benefit provided the life
insured is alive at the end of the deferred period (the ‘vesting’ date) when payments
commence, and subsequently alive at the future times of payment.
As there is a deferred period, regular or single premiums may be payable up to the vesting
date.
Under individual contracts, a single premium is naturally only payable once, at the beginning of
the contract. A person may, however, prefer the flexibility of taking out a new single premium
policy each year until the chosen retirement age, rather than be committed to paying a given level
of premium each year as would be required under a regular premium contract.
Surrender values are not normally available post vesting but may be payable during the
deferred period. Where the deferred annuity is an investment to provide a pension,
legislation is likely to require that the surrender value be reinvested in a pension
arrangement rather than being paid directly to the insured.
Some deferred annuity contracts are sometimes eligible for certain concessions (usually tax
concessions) in the country concerned. To qualify for these concessions, the policy terms have to
meet certain criteria – such as the surrender requirement stated in the above Core Reading. This
requirement ensures that the policyholder’s (generously treated) contributions can only be used
for buying post-retirement income. (The details of the possible tax treatment of insurance and
annuity contracts are not required for this subject.)
Deferred annuities may be without-profits or with-profits. A with-profits deferred annuity will

provide a guaranteed level of regular income, and bonus additions to this income level may be
made while the policy is in deferment. (In other words, bonuses are added to the benefit level
from profits earned during the period leading up to the vesting date.)
For individuals, the contract enables them to build up a pension that becomes payable on
their retirement from gainful employment. At the vesting date of the annuity, the alternative
of a lump sum may be offered in lieu of part or all of the pension, thereby meeting any need
for a cash sum at that point, for example to pay off a house purchase loan.
In practice the same aims can be achieved more flexibly by combining an endowment
assurance with an immediate annuity starting at the maturity date.
The endowment assurance (whether with-profits, without-profits or unit-linked) would be used as

a savings vehicle, and the proceeds used to buy an annuity at maturity (or possibly earlier, if
reasonable surrender values are given on the endowment). Alternatively, (but only if legislation
allows), the policyholder could spend it all on a huge party and a world cruise.
Question
Suggest two disadvantages from the policyholder’s perspective of the endowment + immediate
annuity structure, despite the added flexibility.

Solution
If the proceeds from the endowment must be converted to income at rates unknown until the
annuity is purchased, this may introduce considerable uncertainty for the policyholder.
It may also mean higher expenses, for example two lots of commission.
The group equivalent of a deferred annuity can be used by an employer to fund for
employee pensions.
2.2 The risks

To the extent that the contract can be decomposed into a combination of an endowment
assurance and an immediate annuity, the risks are essentially the combination of those that
apply to those two contracts.
Question
Outline the main risks associated with an endowment assurance, as were described in an earlier
chapter.
Solution
The main risks are:

 investment returns – returns being lower than expected, because the contract has a large
savings element (although the risk for the insurance company is reduced on with-profits
endowments if bonuses can be cut)
 mortality – if a significant benefit is payable on death then the risk is of higher death rates
than expected; if no benefit is given on death then the risk is of longevity (for return of
‘fund’ on death there is little mortality risk)
 withdrawals – particularly the risk of higher than expected withdrawals when the asset
share is negative, or where generous guaranteed surrender terms have been given
 expenses – being higher than allowed for in pricing.
There will be additional investment and mortality risks at the vesting date if the terms for
converting between a pension and a lump sum are subject to a guarantee.
It is a general principle that giving a guarantee of any sort will increase the risks faced by a life
insurance company.

For example, in the case of the endowment plus annuity combination, there may be a minimum
guaranteed rate at which the lump sum will be converted into an annuity at the maturity date.
Should this guaranteed rate exceed the rate the company is normally offering at the time of
maturity, then the company will have to provide a higher benefit than it can comfortably afford,
with a resulting loss of profit. To put this another way, it will have converted these lump sums
into income on the basis of assumptions which, collectively, are too optimistic in relation to
expected future mortality and investment returns (and, also, expenses). Hence the investment
and mortality (and expense) risks have, collectively, been increased.
In the case of the deferred annuity (that is, the contract whose benefits are expressed in terms of
income per annum), there may be a guaranteed rate at which the income would be converted to
an alternative lump sum at the retirement date.

Capital requirements are as for an endowment assurance contract in so far as covering the
lump sum benefit.
These will vary slightly according to the pattern of death benefits provided during the period
before the vesting date (eg return of premiums with or without interest, etc).
Additional capital may be required to cover any guaranteed terms for converting between a
lump sum and a pension.
This is identical in principle to the extra capital required to cover the (guaranteed) renewal option
under the renewable term assurance that we described earlier: whilst the expected cost of the
guarantee should be met from loadings in the future premiums, a higher reserve (or solvency
capital) may be required at outset to cover the greater uncertainty and (hence) risk involved.
The same applies where a deferred annuity (funding for income) provides a guaranteed lump sum
conversion rate.

Chapter 3 Summary
Immediate annuity
An annuity provides an income for life or for a set period. The income is purchased by a
single premium. The possible needs met include an income in retirement and payment of
school fees.
An annuity protects the policyholder from overspending his or her personal capital.
The main risk is that mortality is lighter than expected. This may be affected by self-selection
on the part of those who buy annuities.
Other risks are investment and expenses. Investment risk can be reduced if appropriate
matching assets are held by the insurance company. If a withdrawal benefit is allowed then
this presents a significant anti-selection risk.
Although immediate annuities are single premium, capital strain may arise through
regulatory requirements for supervisory reserves or solvency capital requirements.
Annuities may be written on an individual or group basis.
Deferred annuity
An income starting at a future date (the vesting date) is purchased by a single premium or,
more likely, regular premiums payable up to the vesting date.
After the vesting date the risks are much the same as those for an immediate annuity.
Before the vesting date the risks may be very similar to those of an endowment assurance.
However, the nature and size of the mortality risk will depend on the death benefit given (if
any) and how this compares with the accumulated fund.
Contracts may either provide for income from the vesting date (in which case conversion
from income to a lump sum at that date may be allowed), or they essentially operate as
endowment assurances where the maturity payouts are used to purchase an immediate
annuity at the vesting date.
Additional risks, in particular in respect of mortality, investment and expenses, arise if

guaranteed terms are given for converting between income and lump sum.
Capital requirements are broadly similar to those of endowment assurances.



3.1 Standard mortality tables for annuitant lives often incorporate projected future improvements to
the mortality rates determined from the observed experience. On the other hand, the standard
tables that are based on assured lives data do not usually incorporate future improvements.
3.2 Comment on the following quote: ‘Buying a without-profits annuity to provide income in
Exam style
retirement is foolish. These contracts will be backed with fixed-interest securities and so the
policyholder would do better to buy these directly and cut out the insurer’s expenses and profit.’
[4]
3.3 A regular premium, individual, without-profits deferred annuity contract is being sold to a woman
aged forty. The contract will vest on the woman’s sixtieth birthday. The contract has a guarantee
that, at the vesting date, some or all of the regular income can be exchanged for a lump sum. For
a woman aged sixty, the contract guarantees that each 1 pa of regular income given up will
produce a lump sum of 9.
Explain in detail the risks that the insurance company takes in giving such a guarantee.
3.4 Explain what is meant by ‘selective withdrawal’ and the extent to which it may pose a risk to a life
Exam style
insurer on the following individual contracts:
(a) endowment assurances
(b) term assurances
(c) immediate annuities. [5]



Chapter 3 Solutions
3.1 The single-word answer to this is – prudence. Standard mortality tables are frequently used by
life insurance companies for pricing and reserving purposes. Simply using the graduated past-
experience data, as a basis for the pricing or reserving of annuity contracts, will be imprudent for
most of the developed world, where mortality has persistently improved over time.
On the other hand, prudence will be built in to the premium or reserving basis of assurance
contracts if the assumption is for higher than expected mortality to occur. Provided that the
expectation is for mortality to improve, then using the graduated past-experience data as the
basis will automatically build in a margin for prudence.
This is not the same as saying that standard tables build in sufficient margins for prudence just
because of any adjustment (or lack of adjustment) that has been made within the tables
themselves. Each case has to be treated on its own merits, taking account of any adjustments that
may have already been made within the tables. We will return to the subject of setting
assumptions in much more detail later in the course.
3.2 The principal objection to this comment concerns the absence of insurance if someone were to
take this advice. An annuity may be thought of as insurance against the financial consequences of
living too long (ie beyond the point at which the individual’s money runs out). Without an
annuity, the individual is taking the ‘longevity risk’ himself or herself. [2]
Other points that might be mentioned:

 The expenses of buying fixed interest directly may be higher for the individual. [½]
 But it is true that the insurer’s administrative expenses and profit would be cut out
(assuming that the insurer is making a profit). [½]
 The individual might not know which are the best fixed-interest securities to buy. [½]
 There may be tax advantages for the individual in buying an annuity. [½]
 If the individual were in poor health then the advice might be reasonable, although there
might be impaired-life annuities available that might give better terms. [½]
 The purchase price of the annuity may have come from the maturity proceeds of some
savings policy, a condition of which might have been to buy the annuity with the money
(compulsory purchase). [½]
[Maximum 4]
3.3 The risk to the company is that the cost of buying an annuity of 1 pa, at the vesting date, is lower
than 9. This would make the terms on offer over-generous, and costly to the company. This
might arise if:
 future annuitant mortality were heavier than assumed in the conversion terms, or
 current interest rates were higher than assumed in the conversion terms, or
 future expenses were lower than assumed in the conversion terms.

Detailed explanation
The actual cost to the company if the annuity is taken, is:
( X  E ) 
a60
where X is the annual amount of annuity that would be paid from age 60, E is the actual annual
amount of expenses that would be incurred, and a  is based on the actual mortality and interest
60
rates that would be achieved.
The actual cost to the company if the lump sum is taken is simply the amount of the lump sum.
This is calculated as (say):
LS  X (1  e) a
60
where e is the (proportionate) annual expense loading assumed when fixing the conversion terms
and a60 is based on the mortality and interest assumptions assumed in the conversion terms. (In
this example (1  e) a
60  9 .)
Taking the lump sum will be the more costly option to the company if the lump sum is higher than
the realistic cost of paying the annuity
ie if X (1  e) a
60  ( X  eX ) a
60  ( X  E) a  .
60
This could happen if E  e X and/or a  a

60 60 , ie if the actual future experience would have been
more favourable than the basis assumed in the conversion terms, as stated in the bullet points
above.
3.4 ‘Selective withdrawal’ refers to the fact that those who withdraw from a contract might be
expected to exhibit different mortality experience from those who remain, … [½]
… to the detriment of the insurer. [½]
Whilst the insurer can allow for this expected effect in pricing the contract … [½]
… or in the terms it offers on withdrawal, … [½]
… it may not make enough allowance. [½]
The risk for an endowment assurance is that those in better health withdraw. [½]
Given that the contract is primarily for savings, the risk may not be a serious worry in practice. [½]
The risk for a term assurance is that those in better health withdraw. [½]
Given that term assurance is pure protection this is a more serious concern. [½]
The risk on withdrawal from an immediate annuity is that those withdrawing are in worse health
than those who remain. [½]

The company is at risk of making a lump sum payment to a policyholder who was about to stop
receiving payments anyway, … [½]
… so the average mortality of the remaining annuitants could be significantly lighter as a

result. [½]
[Maximum 5]

All study material produced by ActEd is copyright and is sold
for the exclusive use of the purchaser. The copyright is
owned by Institute and Faculty Education Limited, a
subsidiary of the Institute and Faculty of Actuaries.
Unless prior authority is granted by ActEd, you may not hire

out, lend, give out, sell, store or transmit electronically or
photocopy any part of the study material.
You must take care of your study material to ensure that it

is not used or copied by anybody else.

addition, we may seek to take disciplinary action through
the profession or through your employer.
These conditions remain in force after you have finished

using the course.


Syllabus objectives
 with-profits
 unit-linked
 index-linked.
insurance products.

0 Introduction
This chapter covers the basis of the products and risks to the insurer and insured.
By ‘basis of the products’ we mean the underlying structures by which the benefits under a
contract can be determined. These have already been introduced to you in the preceding
chapters. In this chapter we start to fill in some of the detail.
It is important not to confuse this use of the word ‘basis’ with its other, more familiar use, in
which it refers to the set of assumptions used to determine the price or the reserve (say) for a
product. In these cases we would probably specifically talk about the ‘pricing basis’, or ‘reserving
basis’, whereas here we are talking about the basis of the product – ie its structural basis.
For the purpose of the syllabus for this subject, it may be assumed that the products
described so far may be written on the following main bases:
 conventional without-profits – this is the oldest form of writing insurance contracts,

characterised by fully guaranteed benefits and, usually, level regular premiums
(except for immediate annuities)
Immediate annuities would generally be paid for by a single premium.
 with-profits – a policy where the policyholder has an entitlement to part or all of any
future surplus which arises under the contract (or more widely to share in surplus
arising within the with-profits fund). The policy may be ‘conventional with-profits’ or
‘accumulating with-profits’. The difference between conventional and accumulating
with-profits policies is characterised by how bonuses are added to the policies  this
is covered in more detail in Chapter 6
 unit-linked – a policy where the benefits are linked directly to the investment
performance of a specified fund, and characterised by a lower level of guarantees on
benefits and premiums
 index-linked – a policy where the benefits are linked directly to a specified

investment index or economic index, and are guaranteed to move in line with the
performance of that index.
Question
Explain what we mean when we describe a product as being ‘conventional’.
Solution
A conventional product has benefits that are expressed in terms of values as at the expected
claim date (ie maturity or death), rather than in current terms.
The other types – accumulating with-profits and unit-linked – have their benefits defined as some
accumulated value of premiums paid to date of claim, and so are not known in advance.
We discussed in the preceding chapters the kinds of basis that are found for each of the main
contract types. The next paragraph of Core Reading provides a summary.

Typically, term assurances would be written only on a conventional without-profits basis.

Whole life and endowment assurances are typically written on a unit-linked or with-profits
basis, with new conventional without-profits business being rare. It should be noted that
single premium unitised savings products are often called ‘bonds’. Annuities tend mainly
to be written on a conventional without-profits basis, although are often also available on an
index-linked basis (and in some countries also on a with-profits basis).
The type of basis used for a product has important implications for:
 the cost of the product, and the risks of the product, to the policyholder
 the amount of flexibility that is possible in the product design
 the risks of the product to the insurance company, including capital requirements, and the
potential for profit.
We’ve talked about some of these things already in earlier chapters. So here’s a bit of revision …
Question
State, with a reason in each case, whether a unit-linked or a without-profits benefit structure is
most likely to have:
(a) the higher expected cost for the policyholder
(b) the higher possible cost for the policyholder
(c) the higher risk for the insurance company
(d) the higher potential profitability for the insurance company
(e) the higher capital requirements for the insurance company.
Solution
(a) A without-profits contract – because the insurance company has to include significant
margins in its prices in order to cover the risk of making losses from future adverse
experience.
(b) A unit-linked contract – because it tends to have least guarantees, the company should be
able to pass on much of the costs of future adverse experience to the policyholder: so if
the experience is very bad, the cost to the policyholder can be very high.
(c) A without-profits contract – because all future losses are borne by the company.
(d) This is more difficult. The obvious answer might seem to be the without-profits contract,
because the margins in the premium would be expected to accrue as profits for the
company. On the other hand, should the policy become very profitable then
policyholders may withdraw their policies and take out cheaper ones with competitors:
it’s no good having a very expensive product if no-one will buy it. Unit-linked policies may
actually present the company with greater total profit potential, if it can sell much larger
volumes despite smaller per-policy (or marginal) profits being made.

(e) A without-profits contract – because the guarantees are greater, the company will need
to use more conservative assumptions in its reserving basis or will be required to hold
more solvency capital.
Conventional without-profits products can contain significant risks for the insurer due to
the high level of guarantee that they provide. Mortality risk is particularly high for term
assurances and longevity risk is high for annuities.
Broadly speaking, the risks which the insurer carries on with-profits contracts lie between
those of the conventional without-profits and unit-linked bases.
Question
Explain why this is usually the case.
Solution
Part of the future benefits payable under a with-profits contract is guaranteed, the rest (the
future bonuses, or other profit distributions) is not. Should experience be worse than assumed in
the pricing basis, then the company can pay out less future bonus than originally planned. Under
a without-profits contract, the company has no such ability to reduce future claim costs, and so is
likely to make larger losses than from the with-profits design.
Future payouts under with-profits policies can only be reduced to a certain extent: part of the
future benefit is (fully) guaranteed. So if future experience is worse than required to pay for the
currently guaranteed benefits, then a with-profits policy can produce losses for the company. For
unit-linked policies there is no such (theoretical) limit: any amount of poor investment returns will
simply be passed on to the policyholder through parallel reductions to their unit fund value, and
the company can increase its charges without limit (in theory) to cover deteriorating expenses
and claim costs.
Question
Give an example of a risk that may be higher for a with-profits policy than for an equivalent
without-profits one.

Solution
With-profits endowments (and to an extent whole life policies) are essentially savings vehicles.
Quite a large proportion of the benefits that policyholders ultimately receive will depend on what
the company pays out in future bonuses. It is quite possible that these payouts will be lower than
hoped for, and will disappoint the policyholders concerned. This makes bad publicity for the
company, especially if other companies have done better for their policyholders, and this bad
feeling can result in poor sales of new policies and higher withdrawals from existing ones. This is
an example of a marketing risk.
There is no such problem with a without-profits policy, as policyholders are fully aware when they
bought the policy what they would get from the policy, and no future disappointment is possible
(unless, tragically, the insurance company were to go insolvent before the claim fell due).
Well done if you got most of the points for this and the last question. If you didn’t, don’t worry: we
are going to cover unit-linked and with-profits structures in much more detail shortly. Make a
note to come back to these questions again when you have got further through the course.
In this chapter we concentrate on some important concepts relating to unit-linking and give some
examples of unit-linked contracts. (Some of this will be revision of material you have already
studied in earlier subjects.)
We then cover index-linked contracts in a little more detail.
We then discuss in more depth the different risks taken on by the policyholder for each kind of
product basis, and also look at the way products are used to meet the purposes of the insurance
company.
We cover the with-profits benefit structure, and some of its main variants, in Chapters 6 and 7.

1 Unit-linked contracts
1.1 Introduction to unit-linked contracts

The key features of a unit-linked contract are:
 The policyholder’s premium is paid into an investment fund and at that time buys a
number of ‘units’, which represents a share of that fund. The value of the fund depends
directly on the value of the assets underlying the investment fund. The value of one unit
(the unit price) will be calculated daily as the value of the assets changes. However, the
policyholder’s share (ie the number of units) is unchanged until some policy cashflow
occurs. This could be the payment of a premium, the deduction of a charge, a switch of
units from one fund to another or through payment of a claim (including partial
withdrawals). The total value of an individual policyholder’s fund at any time is the
number of units multiplied by the unit price.
 The insurance company will deduct charges from the policyholder’s fund. These may be
deducted from the premiums before they are invested. For example:
– Less than 100% of premiums may be allocated to units (ie the allocation rate is less
than 100%)
– There may be a ‘bid-offer spread’, ie a difference between the price at which the
insurance company sells the units (the offer price) and that at which it will buy them
back (the bid price). (This is equivalent to a reduced percentage of premiums
allocated to units.)
– A fixed amount may be deducted from each premium paid.
 Charges may also be deducted from the unit funds. For example:
– a percentage of the fund value, taken on a regular basis – this is the regular fund
charge (often referred to as a fund management charge)
– a regular fixed charge (probably as an alternative to a fixed deduction from
premiums)
– regular charges taken to cover risk benefits (eg life cover) in excess of the value of the
unit funds.
These charges will be met by the insurer cancelling units of an equal value to the charge
levied.
 The contracts can be used to provide protection benefits. Where these cost more than
the value of the unit fund at the claim date (as they usually would) then the insurer will
meet the balance of cost. The insurer will need to recoup the cost of these claims (plus a
margin for profit) from the policy charges.

 There are various ways in which initial expenses can be paid for under unit-linked
contracts. The main ways are:
– have a very low, or zero, allocation rate for a short period at the start of the contract
– have a (moderately) reduced allocation rate (eg 93%) for a significant part of the term
of the policy
– have a high regular fund charge.
 The value of the policy at maturity is usually the (bid) value of units.
 A surrender penalty may be deducted from the value of the units if the policyholder
withdraws from the contract.
 The charges are kept by the insurer to cover expenses and risk benefits. The insurer will
aim for charges to exceed expenses and insurance benefits, to give profit to the insurer.
Unit and non-unit funds

A key thing to get clear regarding unit-linked policies is the distinction between the unit and
non-unit funds.
The (bid) value of the policyholder’s unit fund at any point in time is the amount of money that
the company would pay to the policyholder on a claim under the policy. This might be on death,
on the maturity of an endowment, or on surrender (ignoring any guaranteed claim amounts or
the effect of any penalties that might be imposed). Hence the unit fund defines the policyholder’s
(basic) benefit, and the company has a liability to pay this amount (whatever it should be) at the
time of claim.
The non-unit fund simply refers to the company’s ‘other’ money. We can think of it in total as
being the accumulated value of all the charges the company has taken out of its unit-linked
policies, less all the actual costs it has incurred on behalf of those contracts, less any distributions
of profit it has made to its providers of capital, plus any capital injections paid in (for example, in
order to pay for setting up reserves). The actual costs will include actual expenses, plus any
additional claim costs over and above the amounts of unit fund paid out (eg the extra amounts
required to make up any total guaranteed sum assured for those policyholders who die). At the
policy level, we could think of a policy’s contribution to (or share of) the non-unit fund at any time
as being the difference between the total policy asset share and the unit fund of the policy at that
point.
It is important to be completely clear about the distinction between charges and actual costs: the
charges are what the company asks the policyholder to pay, while the actual costs are what the
company has to pay for.
Examples given in Section 1.2 will further illustrate these ideas.


From the consumer’s point of view, the guarantees of conventional without-profits products
come with a higher cost. This type of product also tends to be the least flexible in terms of
how premiums or benefits can be altered. At the other extreme, a unit-linked contract
enables consumers either to obtain a higher expected level of benefit for a given premium
or to pay a lower expected level of premium for a given level of benefit, than under a
comparable non-linked version of the contract.
Question
Explain the last sentence of Core Reading, with reference to a without-profits endowment policy
(having the same death and maturity benefits) and a unit-linked endowment policy (which has the
same guaranteed death benefit as the without-profits policy).
Solution
The life insurance company is likely to be giving lower guarantees on the unit-linked contracts
than on the non-linked contract (eg investment risk is mostly passed to the policyholder and, if
the charges are reviewable, the charges might be increased at a later date if expenses or mortality
costs turn out to be higher than expected). Because risk is being passed to the policyholder this
means that, for example:
The insurance company may take smaller margins in its assumptions, allowing lower charges for a
given level of profit.
100% investment in equities is possible (if that is what the policyholder wants), and these have a
higher expected return than, say, the fixed interest securities that the company is likely to invest
in for the non-linked product.
Hence we have a combination of lower charges, and possibly higher expected returns from the
assets, both of which could lead either to lower premiums and/or higher expected maturity
values.
Avoiding the need for large margins is particularly useful in pricing contracts where there is
considerable uncertainty, for example on contracts offering innovative types of benefit. If the
insurance company is bearing the risk in a highly uncertain situation (ie it is a non-linked
without-profits policy) then very large margins may be necessary. So this also explains the first
sentence of the Core Reading paragraph.
The word ‘expected’ is very important in the last sentence of the Core Reading above. While the
expected maturity value might be higher under a unit-linked contract, the actual maturity value
could be lower. This is what is meant by the policyholder taking more risk.

An important point to note from the above question is the effect of the charges being guaranteed
or reviewable. If charges are guaranteed then the company cannot change them during the
lifetime of any one contract. This is not the same as saying that charges are level over the policy
term. For example, we could have a contract that had guaranteed expense charges rising in line
with a published price inflation index, or which had a guaranteed rate of mortality charge for each
possible age of the policyholder (which would vary each year both because of age and also
because the sum at risk – the difference between the guaranteed death benefit and the unit fund
– would change).
In addition, a unit-linked contract can offer flexibility in the types and levels of cover
included and the ability to vary premiums according to need.
A flexible design for a unit-linked whole life contract is illustrated in the following example.
Example
Unit-linked regular premium whole life contract.
Premiums: payable monthly in advance.
Allocation rate of premiums to units: 25% for first 18 months, 97% thereafter.
Range of twelve unit-linked investment funds to choose from, such as International Equities,
Japanese Equities, German Fixed Interest etc.
£2.00 per month deducted from the premium before allocation to units, inflating annually in line
with a suitable index of prices.
Annual fund management charge taken from the fund: 1% of value of unit fund.
Male aged twenty-five, non-smoker.
Premium: £20 per month. Reviewable by the insurance company every ten years.
Policyholder may choose sum assured up to a maximum of £200,000.
The cost of life cover to be met by cancelling units monthly.
The insurance company guarantees to pay the benefits, provided that the required premiums are
paid.
Withdrawal is allowed at any time. Surrender value will be the value of the unit fund.
It is possible to cover the expected cost of risk benefits from the charging structure without
having explicit unit cancellations for this. Where explicit cancellation is used (as here) the broad
principle is to cancel units monthly to a value of
qx
 (sum assured – unit value)
12
where (sum assured – unit value) is the sum at risk, or death strain at risk.

A key element of the flexibility in the contract illustrated is the choice of level of life cover for a
given premium. At the maximum level of cover there may be little unit fund left by the time of
the first premium review after ten years; at a low level of cover the fund would be larger.
Question
Explain what would be expected to happen to the premium at the first review for a case with a
very high sum assured.
Solution
If the same sum assured is to be maintained, the premium is likely to rise, perhaps significantly.
Little reserve has been built up and the policyholder is now ten years older. The cost of life cover
will therefore have increased significantly.
Alternatively, the policyholder might be allowed to continue paying the same premium but with a
reduced sum assured.
If a very small sum assured is chosen, unit cancellations would be much lower, and so a larger
savings element would build up. The contract is almost like several products in one, with the
policyholder able to select the desired mix of savings and protection, and possibly able to change
this mix as his or her needs evolve. This sort of unit-linked structure may be used for anything
from pure savings to (virtually) pure protection.
Question
The amount of unit fund might become negative during the period between premium reviews.
(i) Explain how this might happen.
(ii) Describe the implications for the insurance company and policyholder.
Solution
(i) The charges may exceed, in aggregate, the value of the accumulating premiums. This
could happen with high inflation (the premium deduction is linked to inflation while
premiums remain constant), and low investment return (so the unit fund grows less
quickly). A lower unit fund will also increase the mortality charge (as the sum at risk will
be larger), further reducing the unit fund value.
If the mortality experience gets worse, the company might increase its mortality charges.
This will again produce further reductions in the unit fund.
(ii) There are a number of implications. If the policy were still in force by the review date,
then a substantial rise in premiums would be required to repay the accumulated deficit.
Policyholders may not accept the increase and might lapse their policies, with a loss
(equal to the negative value of the unit fund) to the insurer.

Were the policy to lapse at any time while the fund was in deficit then the insurer would
make a loss.
The insurer would be obliged to continue the policy until the next review date regardless
of the state of the unit fund, because it had guaranteed that premiums would not be
increased between review dates.
Not all unit-linked products are as flexible as the one illustrated above, while others may be more
flexible. This is just one example. The following is an example of a rather less flexible unit-linked
contract. It is less flexible from both the policyholder’s and insurance company’s points of view.
Example
Unit-linked endowment assurance contract.
Term: ten years.
Premiums: payable monthly in advance.
Allocation rate of premiums to units: between 93% and 96% throughout, dependent on the
policyholder’s age and sex.
Fund management charge: 1.5% pa.
Sum assured on death before the end of the policy term: ten times the annual premium, or the
value of the units if higher.
The cost of life cover to be met from the overall product charges given above.
The surrender value will be the value of the unit fund minus a surrender penalty, dependent on
the premium level and the duration in force at the time of surrender.
Question
Discuss the advantages and disadvantages to the company of not explicitly charging for mortality
in the above contract design.
Solution
Advantages
Greater simplicity may help the marketing of the contract, particularly where the method of sale
does not allow much explanation, or the target market is not financially sophisticated.
The greater simplicity should make it less expensive to administer and so better value for the
policyholder, which should further increase marketability. Profits could also be increased.

Disadvantages
The charges will not be sensitive to changes in the sum at risk. For example, when unit fund
values fall (eg due to poor investment returns), then the company’s mortality costs will rise (the
sum at risk has increased). However, under these charging arrangements, the company’s charges
will actually fall (because of the large fund management charge component).
The charges each year may not match very closely the cost of death claims each year. If the
expected claim costs exceed the charges expected to be received in some future years, then the
company would have to hold some non-unit reserves (ie reserve some money in addition to the
unit-fund values) to cover these future losses. As these reserves would be subject to supervisory
regulation, then this is likely to increase the contract’s capital requirement.
Don’t worry if you did not get this last point. This is the kind of thing you will get more familiar
with as your knowledge of the subject increases.
1.3 Risks of the product to the insurer

From the insurer’s point of view, the risks under unit-linked contracts may be lower than
under non-linked contracts due to the reduced level of guarantees offered. The nature and
extent of any financial risks from investment, expenses and demographic assumptions will
depend on the nature and level of any guarantees given and any constraints imposed on the
level of charges by marketing or legislative requirements.
Most of the investment risk on unit-linked contracts is borne by the policyholders through the
value of their units. The life insurance company will have some exposure if there are charges
based on fund value, but even this would be a second-order effect. (Nevertheless this last point
can be important, especially if the fund management charge is a large part of all the charges paid
under a particular contract design.)
However, the investment risk for the insurance company may be significantly increased if it gives
some sort of non-unit-related guarantee, such as a minimum maturity value equal to premiums
paid, or equal to premiums paid accumulated at 3% pa, for example.
The insurance company may give itself some protection against future increases in expenses by
having some charges linked to an index of inflation, or by giving itself the right to increase certain
charges at its discretion (that is, by having reviewable charges, as we mentioned in the previous
section). However, it may be necessary to keep charges low in order to be competitive or to be
fair to policyholders, who may not have expected charges to be increased. Therefore poor
expense performance on unit-linked products may in practice be as damaging as on non-linked
products.
Any specific legislative restrictions on increases to charges would clearly also increase the risk.

The guarantees relating to demographic assumptions such as mortality may be lower on

unit-linked contracts. For example, in the whole life contract illustrated in the example in
Section 1.2, the insurance company might be able to vary its unit cancellations for life cover at its
discretion (the information did not state that the rates of calculating the mortality charge were
guaranteed in any way). Again, however, it may be necessary to keep charges low in order to be
competitive.
It should also not be forgotten that the company is still guaranteeing a certain level of life cover
for ten years for a given premium, and a serious deterioration in experience is likely to cost it
money. This could happen if the policy lapses when the unit fund value is negative.
The anti-selection risk for a unit-linked policy is the same as for the comparable non-linked
contract. The selective withdrawal risk is likely to be higher, due to the more open charging
structures of the contracts, than under the comparable non-linked contract.
Question
‘But surely if the insurance company has the right to review the charges regularly, the
anti-selection risk is greatly reduced? If the mortality experience is bad the company can just
increase the mortality charges.’
Discuss the validity of this comment.
Solution
This comment is wrong. Reviewable charges do provide some protection against a general
deterioration in mortality, but not against anti-selection. An insurance company that had weak
underwriting would soon find its experience declining relative to its competitors. Attempts to
increase charges would probably result in selective lapses, as the healthy would seek a better deal
elsewhere.
The last point is also referred to in the Core Reading: the fact that under many types of unit-linked
contract the amount of mortality charge is explicitly identified means that policyholders are fully
aware of how much they are spending on their insurance protection. This is likely to lead to
increased selective withdrawal, especially as policyholders get older and their mortality charges
are increased every year.
At times when the asset share is negative, there is a financial risk from withdrawal. At other
times, whether there is such a risk depends on how any withdrawal benefit paid compares
with the asset share. Although the withdrawal benefit may not be guaranteed in amount, the
method of its calculation will usually be guaranteed.

At this point, it is important to make sure that you appreciate the difference between the asset
share and the unit fund of a unit-linked policy. You should refer back to our description of the
unit fund in Section 1.1 if in any doubt: the unit fund is a definition of the policy benefits at a point
in time, while the asset share is the accumulated value of total premiums less actual costs to the
same point in time. While the unit fund is calculated in a similar way, it involves the accumulation
of allocated (not total) premiums, and charges (not actual costs). The unit fund and asset share
can then be quite different, depending on the charging structure, and there is no reason why the
unit fund should not considerably exceed the asset share, particularly near the start of a contract.
A surrender penalty may then be imposed in order to reduce the withdrawal risk.
For unit-linked policies, the extent of the withdrawal risk will depend on the incidence of
expenses relative to charges, and on surrender value penalties.
The company will wish to define the surrender penalty so that wherever possible the payment to
the policyholder is no more than the asset share, and preferably below it so as to produce a
profit. In fact, at some durations, the company may aim to make the same amount of profit on
the withdrawing policy as it would have done had the contract stayed in force.
The contract design and the pattern of expenses incurred on the contracts will be the main
factors determining whether or not a surrender penalty is needed, to what extent it depends on
term and duration in force, and the form it takes (cash amount or related to units).
The following three simplified example contracts illustrate this. For all three contracts we assume
that the bulk of expenses are incurred at the outset of the contract and that the contract term is
ten years.
Example contract 1
The allocation rate is 95% throughout. The fund management charge is ½% pa.
With this design the initial expenses are largely recouped by the 5% from each premium that is
not allocated to units. Therefore a suitable surrender penalty might be the present value of the
‘lost’ future monetary amounts, for example 20% of one year’s premium after six years, ignoring
discounting.
Example contract 2
The allocation rate is 100% throughout. The fund management charge is 2% pa.
On this contract the initial expenses are recouped through a high charge related to the value of
the units. If the company wishes to recoup the lost future expected charges (so that surrendering
and continuing policies have equal expected profitability), it will need a surrender penalty equal
to some percentage of the unit fund, with the percentage reducing as the contract nears
maturity.
Example contract 3
The allocation rate is 0% for 2 years, 100% thereafter. The fund management charge is ½% pa.

Question
State, with justification, the surrender penalty that would be necessary with this contract design.
Solution
Probably none. There is no fund from which to take the penalty in the first two years anyway.
After two years the company should have recouped its initial expenses and made its profit (and
the asset share should then exceed the unit fund value).
Whatever design is used, the company cannot always avoid a loss on surrender. The fund may be
too small in the first few months, say, for an adequate surrender penalty to be applied. This is
really just another way of saying that the asset share is likely to be negative early on, and so even
giving no surrender value cannot prevent a loss.
One way to reduce the surrender penalties required is to reduce initial expenses. For example, it
may be possible to reduce commission paid at the start of the contract in favour of more
commission over the term of the contract.
There is also a significant marketing risk under unit-linked products due to the higher
variance of the level of benefit for a given premium, as compared with the equivalent
non-linked contract.
In other words, the reduced guarantees which may produce a better expected deal also produce
greater uncertainty of return for the policyholder. If things turn out badly then the company or
the industry as a whole may suffer criticism and ill feeling. This is particularly the case if
policyholders have not understood the risks they were taking.

The capital requirement depends critically on the design structure of the contract.
As far as regular premium contracts are concerned, a design that involves a low allocation of
premiums to units at the start of the contract is particularly capital efficient (see ‘Example
contract 3’ in the previous section). This is because the early premiums can be used to provide a
positive cashflow to the life insurance company, rather than building up the unit fund. This
means that initial capital strain is quickly repaid.
Other designs may recoup initial expenses and produce profit more slowly, for example through
regular deductions from premiums (Example contract 1) or from the unit fund (Example
contract 2) over the course of a contract.
It may be possible to take credit for these future charges in advance, although this will depend on
the particular supervisory regulations. The basic principle is to take credit, in the reserves, for the
expected present value of those charges that will be received in the future which are designed to
recoup the initial expenses.

As a formula we have:
[Adjusted reserve]  [Normal reserve]  EPV [Initial expense charges]
As the reserve appears as a liability in the company’s balance sheet, the expected present value of
the initial expense charges will appear as an asset, reducing the company’s capital requirement.
In the case of Example contract 1, we would calculate this asset in the form of a negative non-unit
reserve. We return to the calculation of such a reserve in Chapter 19.
In the case of Example contract 2, we would calculate the adjusted reserve using a technique
called actuarial funding. The use of actuarial funding was much more prevalent in the past, and
you may come across references to it in past exam questions. However, actuarial funding is no
longer part of the Subject SP2 syllabus.

2 Index-linked contracts

As mentioned earlier, index-linked contracts provide benefits that are guaranteed to move in line
with the performance of an index specified in the contract.
Suitable investment indices might be the major domestic equity market indices of any country, or
more broadly based international equity indices. Links might also be made to other asset classes
such as fixed interest.
Typical economic indices that might be used include retail or other appropriate price indices.
Linking benefits to a retail price index, for example, promises the consumer a benefit that moves
exactly in line with retail price inflation.
Examples
Index-linked annuity
An annuity paid for the lifetime of the policyholder whose regular payment moves in proportion
to a published index of prices or earnings. It is paid for by a single premium, or it may be in the
form of a deferred annuity paid for by regular premiums during deferment. In this case the
premiums may or may not be linked to the same index.
The company will take its margins for expenses, mortality and profit in the rates it offers for
conversion from the premium to the initial level of annuity.
Index-linked investment bond
Usually a single premium is paid. The contract guarantees to pay the value of the single premium,
increased in line with a published investment index, at a given future maturity date or on earlier
death. Early surrender may be possible.
The company takes its margins for expenses, mortality and profit from the difference between
the investment return obtained by the company and the return determined by the progress of the
index over the period.
2.2 The risks

The main risk to the life insurance company with respect to index-linked contracts relates to
investment. The company may not be able to invest in a way that precisely matches the
benefit guarantee it is giving.
The lack of precise matching may result simply from practical considerations. If the index
concerned is an investment index covering several hundred stocks, including some that are not
actively traded, it may be impractical and very expensive always to keep exactly the same
weightings in the actual portfolio as in the index. Periodic changes to the constituents of the
index would increase this problem.

Even where the index is based on fewer shares, the life insurance company may choose to try to
match (broadly) the index movements with a carefully chosen selection of shares from the index.
This would reduce dealing expenses.
Problems in matching the movements in the index will also arise if the index is not based on
weighted market capitalisation.
In some markets it will be possible to use derivatives to get a very close match to movements in
the underlying indices, although this will come with a cost.
Where the benefit is linked to the change in an economic index, the company will need to invest
in assets whose value will follow as closely as possible the movements of that index.
The risk that the assets held do not move in line with the investment or economic index is
borne by the life company. This can be contrasted with the unit-linked design, where the
asset risk is borne by the policyholder.
The risk of adverse movements in the index value itself is a risk to the policyholder, ie if the index
value increases slowly (or even falls) then the policyholder receives lower benefits than expected.
Question
A single premium contract guarantees to pay, after five years or on earlier death, the value of the
single premium adjusted by the change in the national Retail Prices Index over the period up to
the date of payment.
Explain briefly the likely nature of the investment risk arising from this contract from the point of
view of:
(a) the policyholder
(b) the insurance company.

Solution
(a) Policyholder’s investment risk
A return equal to the growth in retail prices over time obtains for policyholders a guaranteed zero
real rate of return on their investment. The greatest risk of disappointment arises from inflation
unexpectedly and permanently taking a fall at some stage during the policy term. On the other
hand, the policyholder will gain most from unexpected periods of high future inflation, compared
with alternative investments.
(b) Insurance company’s investment risk
The insurer’s risk depends mainly on the degree of matching achieved. 100% investment in
index-linked government securities of the appropriate term would be the best match. However,
the real return offered by indexed-linked government securities will probably be quite low. So
there will be little money left over, after paying for expenses, to make a profit.
The company might therefore adopt an investment strategy that, whilst ensuring a broad match
with the index, would lead to an expectation of making a much higher real return. A mix of
domestic equities and index-linked government securities might achieve the required balance,
noting that equities over a long period may be expected to provide high real returns. However,
this leaves the company with a risk of making investment losses.
The life insurance company could, even with a unit-linked or (investment) index-linked design,
choose to invest in assets other than those that determine the company’s liability to the
policyholder. However, this is normally unwise and legislation may even require that companies
match as closely as possible.

3 Risks of the products to the insured
3.1 Conventional without-profits contracts

In the following paragraph of Core Reading, we consider the risk taken on by the policyholder,
should he or she take out a without-profits policy that provides level benefits (eg a fixed sum
assured on death) in return for level regular premiums.
In the case of conventional without-profits contracts, the main risk borne by the
policyholder is that the amount of benefit provided – which is fixed and guaranteed –
eventually turns out to be insufficient. Given the long-term nature of the contracts, this risk
is exacerbated by the effects of inflation over time. A subsidiary risk is that the insurer
becomes insolvent and unable to meet the guaranteed benefits in full. Further risks arise
from the inflexibility of the product to keep pace with the changing disposable income of the
policyholder, and the changing amounts of benefit needed throughout the financial life
cycle.
Question
Explain what the Core Reading means by ‘insufficient’ in relation to a benefit payable under a
term assurance contract, taken out by someone with a young family.
Solution
It means that the benefit paid out does not fully meet the needs for which it was intended.
A term assurance benefit would be expected to cover items of expenditure, such as repayments
of a mortgage, education fees for children, and general living costs. If these expenditure items
turn out to have much greater cost than anticipated when the contract was taken out, then the
benefit would be insufficient. Just some of the many possible reasons for this might be:
 has more children than expected
 university tuition fees have been unexpectedly introduced (which have to be paid for by
the parent)
 has a bigger mortgage than originally expected
 inflation has been higher than expected, so all costs have increased
 spouse has adopted a celebrity lifestyle.
Question
It would seem from the Core Reading above that insurer insolvency is a policyholder risk.
Explain whether this means that insolvency is not a risk for the company.

Solution
Insolvency is most definitely a risk for the insurance company. If the company is insolvent, then
all other parties with financial interests in the company will be adversely affected too. These will
include shareholders, employees, and other policyholders (not just the without-profits holders
being referred to here).
The government, and hence the public at large, will also be very concerned to keep insurance
companies solvent. Insurance is a vital financial service and society would be significantly worse
off without it. A failed insurer may reduce market capacity, reduce market competition, or both,
all with possible detrimental effects on the public.
Finally, the policyholder is exposed to the risk of being unable to maintain premiums due to
accident, sickness, redundancy, or other loss of income (although other forms of insurance
may be available to provide cover against these eventualities).
Various health insurance contracts are available to cover loss of income, or increased costs, due
to sickness or disability of a policyholder. Part of the benefits from these policies could be used to
pay for the continuing premiums on the person’s life insurance policies. Alternatively, life
insurance policies can include a waiver of premium benefit, where the insurance company will
effectively pay the policyholder’s premium for them during periods of sickness. Such a life
insurance policy is said to have a waiver of premium rider benefit. Further discussion of these
benefits is beyond the scope of this course, but is covered in Subject SP1.
Bear in mind that not all without-profits contracts have level benefits or premiums. For example,
they might be linked to an inflation index.
Index-linked benefits, if linked to an inflation index, can reduce the risk of erosion of the
benefit value, although there remains a risk that the selected index does not replicate the
rate of spending inflation to which the individual policyholder is exposed.
There are also without-profits contracts that provide fixed increases (eg 3% compound per
annum), though with these there is clearly no guarantee that the increases will match the
policyholder’s rate of spending inflation.

3.2 With-profits contracts

With-profits contracts can provide some protection against the ultimate benefits being
eroded by inflation, to the extent that the policyholder does not also choose to reduce the
guaranteed level of benefit in anticipation of the future value of surpluses which he or she
might enjoy.
The nature of this protection does depend on the way that profits are distributed under the
policies concerned. Methods that involve increasing the benefit levels over time will certainly
have the effect described in the Core Reading. Methods that pay out profits as annual cash
dividends, however, will not appear to provide the same protection against inflation at the time
of contractual claim, as this benefit may remain at the contractual level or only be slightly
increased. On the other hand, some distribution methods involve increasing the premium as well
as the benefit payable each year, which will increase the inflation-protection achieved. (This is
what happens with the method known as revalorisation: this and other methods will be described
in Chapters 6 and 7.)
However, further inflation-protection can be obtained through the indirect investment achieved
for a policyholder through his or her with-profits policy. The extent varies significantly between
the distribution methods involved, but two of the methods (the additions to benefits and the
contribution methods, to be precise) involve at least some backing investment in real assets, such
as equities and property. To the extent that equity and property investment provide real returns,
these will be passed on through profit distributions to the policyholders, who thereby achieve an
element of real returns on the premiums they have invested through the insurance company.
This is very different from the without-profits case, where the performance of the actual assets
backing the contract have no effect on the benefits that are paid out.
Question
‘The assets the insurer holds in reserve for its without-profits policies have no bearing on the
benefits paid out under those policies.’
Comment on this statement.
Solution
Once a without-profits policy has been issued, its benefit amount is fixed by the policy conditions.
So, whether the company invests in government fixed-interest securities or in back-copies of
Batman comics, this will have no consequential effect on the company’s contractual obligation
under the contract. This therefore supports the statement.
On the other hand, the company will take a great deal of account of the assets expected to be
invested in when determining the price to charge for the contract – ie when determining the
amount of benefit to provide in return for a given premium. (This comes under the heading of
setting assumptions, which we discuss in detail later in the course.)

Under a with-profits policy, the policyholder still carries some risk of insurer insolvency,
although this should be less than under a conventional without-profits contract to the
extent that future surpluses can now be used to maintain solvency, before being distributed
to policyholders.
Don’t worry if this paragraph of Core Reading doesn’t make much sense on first reading – we will
discuss this in detail later.
The other risks outlined above remain largely unchanged.
3.3 Unit-linked contracts

Unit-linked contracts may provide some protection against the ultimate benefits being
eroded by inflation, but equally the policyholder would accept the possibility of poor
investment performance – either long term, or because of reduced investment values at the
point of benefit payment. Any minimum guaranteed death benefit will tend to be expressed
in monetary terms and vulnerable to erosion by inflation.
Question
Describe the circumstances that would mean that a unit-linked contract provided the most
protection of the policyholder’s benefit level against erosion by inflation.
Solution
Most protection against inflation is provided where:

 the unit fund is invested in secure index-linked securities
 the premiums paid in to the fund increase in line with the appropriate external inflation
index.
Less secure protection is provided if the unit funds are invested in real assets, such as equities and
property. If only level premiums are paid then, even if the unit funds are invested in index-linked
stocks, only partial protection is produced. This is because, whilst the returns obtained on the
money invested would be inflation-proofed, the money being invested (the premium) would be
reducing in real value.
The Core Reading is making a distinction here between long-term poor investment performance,
and the effect of the policy maturing on a ‘bad day’.
Long-term investment performance can be thought of as some sort of smoothed annual return
obtained over each year of the policy term, while the ‘bad-day’ effect is a function of the short-
term volatility of the stock market concerned. Of course, some policyholders (presumably around
half) will be lucky and the maturity of their policy will occur on a relatively ‘good’ day.
Question
Explain what the Core Reading actually means by poor investment performance.

Solution
It means achieving lower returns than reasonably expected from intelligent investment of the
policyholders’ money over the period, allowing for the company’s (reasonably expected) charges.
This usually boils down to considering returns relative to some standard, such as the performance
of a relevant investment index, or the average performance of competitors’ products.
The policyholder still carries some risk of insurer insolvency.
Unit-linked contracts tend to be designed in a flexible manner to allow the policyholder to

vary premiums and benefits over his or her lifetime. However, increases in minimum
guaranteed death benefits are likely to be made subject to satisfactory evidence of health,
so there is a risk to the policyholder that such increases are declined or made subject to
special terms.
In fact, this last point would be the norm for any kind of contract that provides a benefit on death,
and not just for unit-linked policies. The difference is that, in the case of conventional contracts,
the increase would normally involve issuing a new policy for the additional amount of cover
required.
Despite the risks of benefit erosion described previously, death benefit sums assured under unit-
linked policies tend to be reduced over time, rather than increased, so that the need for additional
health evidence does not arise.
Question
Explain why policyholders tend to reduce, rather than increase, their levels of death benefit over
time.
Solution
This reflects policyholders’ reduced desire for lump sum protection on death as they become
older. Three main factors are at work:
 Policyholders tend to become more affluent over time, reducing the need for insurance
(they can, in effect, self-insure from their own wealth).
 The need for death benefit reduces: children become older and have fewer future years
of dependence on parental income, personal debt reduces as mortgages are paid off, and
so on.
 The death benefit becomes visibly more costly with increasing age, as the charging rates
can be seen to rise annually under unit-linked policies. So policyholders prefer to redirect
their money into their own savings. This can easily (and temptingly) be achieved by
reducing the death benefit sum assured on their unit-linked policies, and simply allowing
more of their premium to be invested in their unit funds.

Unit-linked contracts often allow premiums to be missed for ad hoc periods of time.
This is what really makes a unit-linked product stand out from a conventional product design from
the point of view of flexibility.
Question
Explain why this premium flexibility can be an important feature of a unit-linked policy, for both
the policyholder and the insurance company.
Solution
It is important to have flexibility of payments where those payments are being used for saving
(including saving for retirement). If premiums could not be reduced without the policy being
made paid up permanently or surrendered, people would be confined to paying only the level of
premium they could anticipate being able to afford in every future year. Flexibility, on the other
hand, means that the policyholders can contribute more to saving in ‘good’ years (ie when they
have extra disposable income), knowing that in ‘bad’ years they could reduce or even stop their
premium without jeopardising the returns they will get from the premiums they have already
paid.
This should also maximise the amounts of premiums the insurer will receive under any individual
policy, and by making the policy more attractive encourage more sales. Both of these things
should help increase profits.
In addition, or instead, waiver of premium benefit can provide for premiums to continue in
the event of sickness of the policyholder.
(Waiver of premium benefits can be found in some non unit-linked contracts as well.)
Unit-linked contract charges may be variable at the discretion of the insurance company, so
there is a risk to the policyholder that these are increased.
As we saw in Section 1.3, the insurer is exposed to the risk of high inflation on its expenses and
possibly mortality risk too. It may be able to pass these risks on to the policyholders through
higher charges.
3.4 Index-linked contracts

We have already discussed the advantage to the policyholder of having benefits linked to an
inflation index, at the end of Section 3.1.
For policies with benefits linked to an investment index, the investment risks to the policyholder
are very similar to those from unit-linked products, except that they depend on the performance
of the index used rather than on that of a specifically designated asset fund.

Differences in policyholder risk for products that are index-linked as compared to unit-linked
include:
 the insurer tends to have greater investment risk with index-linked products (see Sections
1.3 and 2.2), which will increase the risk from insurer insolvency
 there is much less scope for reviewing charges for existing index-linked products, which
also increases the insolvency risk
 index-linked products generally have much less flexibility.

4 Purpose of the products to the insurer

The insurer is likely, in the normal course of events of a free economy, to seek to offer an
attractive range of products which maximises its profits. In the long run, profits are likely to
be maximised when the utility of the products to the consumer is also maximised. Within
this, the insurer will seek to take on further profitable risks within its available capital to
maximise economies of scale. The insurer will seek to diversify and control the overall
risks wherever possible, using an appropriate volume and mix of suitably designed
contracts.
Here we are talking about the single most important aspect of the insurance company’s
management. Selling life insurance products is the business of life insurance companies. The
types (designs) of products that are sold, in what volumes and at what prices, are the key factors
in determining the company’s profitability and risk profile. They are not the only factors, though…
Question
List eight other aspects of insurance company management that affect its profitability and risk
profile (ie other than those to do with product design, pricing, selling and marketing).
Solution
Some possible aspects, in no particular order, are:

 investment policy
 underwriting practice
 reinsurance strategy
 approach to reserving and profit distribution
 discontinuance terms offered
 effectiveness of monitoring and feedback systems in reacting to events
 human resources management and remuneration
 administration procedures, data handling and maintenance
 capital management.
Well done if you got most of these: it means you are already thinking like a life insurance actuary,
because you have a good idea of the important management issues.
But don’t worry if you didn’t get so many: these are the things (along with design and pricing) that
we are going to learn about as we go through the course.
When we go on to talk about macro risks later in the course, we consider the volume and mix of
new business as a source of risk. This is because it is largely through the new policies being taken
on that the liability profile of the company gradually changes over time, and this in turn will
change the risk (and profitability) characteristics of the company.

Question
Explain the point in the above paragraph of Core Reading regarding maximising the utility of the
products to the consumer.
Solution
The Core Reading is best explained by first looking at an example.
When a person buys a term assurance (for example), he or she agrees to pay a series of small
premiums (contingent on survival over a period), in exchange for a very large lump sum
(contingent on dying during the period). Because insurers incur expenses and need to service
their capital (make profits), the expected cost of the premiums paid by the policyholder will
invariably exceed the expected value of the benefit received. Nevertheless, he or she will still
take out the policy.
The reasoning goes like this. Should the policyholder actually die during the term, the loss to the
dependants of the person’s income is perceived as catastrophic. The claim amount paid by the
policy is therefore vital: the dependants cannot manage without it.
On the other hand, the premiums are so small as to be almost unnoticeable to the policyholder,
even if he or she survives for the whole term and no benefit is paid. Paying these premiums will
not incur the policyholder or his or her family with any noticeable inconvenience, whereas to die
without insurance would be a disaster.
Utility theory explains the above by describing the benefits as having higher expected utility to the
policyholder than do the premiums. So what matters most in terms of product design is the way
that the costs (premiums) are converted into rewards (benefits). Finding a product design that
provides the greatest utility difference between the two will enable the company to sell more
business to more people and with bigger margins for profit.

5 Package products
These first four chapters of the course have covered the basic contract types on which the
examiners will base their questions. However, questions may be asked where two or more of
these basic contract types are combined together as a package.
When answering questions featuring such ‘package’ products you will find it helpful to think
about the separate underlying component contracts.
(You will find an example of such a question in the Practice Questions at the end of this chapter.)

6 Products for the personal financial life cycle

We now return to the subject of the personal financial life cycle that we first introduced at the
start of Chapter 1.
Question
For each stage of the life cycle described in Section 1 of Chapter 1, suggest the type(s) of product
that would be most suitable (and therefore be in most demand) to meet the needs described.
Solution
Ages 16 - 25
Life insurance needs are probably zero, as there are likely to be no dependants. There is possibly
a need for savings (eg towards house purchase) but this is likely to be too short term and too
small to be well served by life insurance products.
Ages 25 - 40
This is a key period for life insurance protection, because of the high debt, high expenditure and
financial dependence of the family on continued employment. Low wealth means that the
consumer is unlikely to want to take on very much financial risk.
There is therefore a big need for sickness and death benefit protection, so there should be a
demand for term assurance at this stage. Policy durations should be at least as long as the main
expected period of dependency (eg until children become independent). Policyholders’ general
risk-aversion will tend to favour sales of products with guaranteed terms (which may still be
cheap enough to afford because of the temporary cover being sought).
Policyholders with uncertain future needs for death benefit protection may well choose
guaranteed convertible, increasable or renewable term assurance contracts. These will provide
the facility to increase benefits in the future even if the policyholder’s health were to deteriorate.
Much life assurance may be sold to this age group in relation to loans, especially those taken out
for the purpose of house purchase. In particular, term assurance may be required to repay the
loan on death within the repayment period, both as a protection for dependants (to prevent
repossession of the property should they be unable to maintain repayments on the death of the
policyholder), and also as additional security to the lenders. This would often be a form of
decreasing term assurance.
Some consumers (possibly those with higher incomes) may be sufficiently risk tolerant to take on
some risks in their products. For example, the protection products described above might be
taken out on a unit-linked basis, especially as this provides more flexibility to change levels of
cover in future as needs change over time.

Such individuals may also decide to fund the full repayment of their loan (ie on surviving the
repayment term or on earlier death) using an endowment assurance. To do this using a
without-profits policy would usually be more expensive than a simple loan repayment with the
lender, so the products would normally be with-profits or unit-linked. With these policies there is
usually a risk of the policy providing inadequate payment at maturity to repay the loan in full, but
this is usually so far distant as to be an acceptable risk for people to take on. However, companies
should ensure that their policyholders fully understand this risk.
Other savings vehicles may be bought, often associated with children: for example to help provide
for children’s education costs (eg school and/or higher education fees) or simply as a ‘present’.
An endowment assurance, probably on a with-profits or unit-linked basis, would usually be used
for this. A start to saving for retirement is also possible, particularly for those with no other
source of pension.
This is also an ideal time to sell group life insurance benefits to employers. This will include death
benefits, and endowment assurances or deferred annuities, each of which would be designed to
terminate at retirement. (The last two contracts would be designed to provide a post-retirement
income for the employees.) Provided the employer is large enough much more risky contracts
could be taken on, especially for the endowment contract, because the employer should be more
tolerant of risk than the individual and should be keen to provide best value for money. A unit-
linked policy (without guaranteed charges) is likely to provide this.
Ages 40 - 65
This is the biggest period of demand for retirement, wealth transfer and other savings vehicles. In
all cases the emphasis will be on the levels of return provided on the investment, so with-profits
or unit-linked products would be normal.
For retirement provision, either endowments or deferred annuities would be used, with the more
risk-averse investors generally choosing the latter. There will also be more emphasis on
guarantees nearer to retirement, so people with unit-linked policies may well switch their funds
into more secure investment types at that time (eg from equities into fixed interest or cash).
Whole life assurance would be most useful for accumulating wealth to be transferred on death,
again on a unit-linked or with-profits basis.
Opportunity for other investment is possible, because of the higher levels of disposable income
for some people in this age bracket. Here single premium investments could be attractive,
providing a lump sum benefit in the future through any of the possible product types
(eg unit-linked, with-profits, or index-linked).
Ages over 65
Immediate annuities will be sold to convert assets into income, in order to protect individuals
from the risks of running out of money into old age. The extent of this will depend on the amount
of deferred annuity bought in advance of retirement, any other assets the individual has, and any
lump sum available from other pre-retirement savings. Legislation may also dictate how much of
these lump sums has to be converted into annuities.

Fixed income, index-linked, unit-linked and with-profits annuities are all possibilities, although
often the desire to have maximum income in the short term leads to the fixed-income versions
often being most popular in practice. Additional features, such as a five-year minimum payment
period and/or last survivor benefits, would often be chosen.
Needs for wealth transfer vehicles and other savings will continue as for the previous age group,
but probably at lower levels of demand as the amount of income usually falls after retirement.

Chapter 4 Summary
Unit-linked contracts
Unit-linked contracts generally offer lower guarantees than their non-linked equivalents.
Most of the investment risk and possibly some of the other risks are passed back to the
policyholder. This allows the insurance company to take smaller margins in its pricing
assumptions, and this results in a better expected deal for the policyholder.
However, the passing of risk back to the policyholder increases uncertainty for the
policyholder, who may suffer more from adverse experience than would be the case under
non-linked designs. This may cause ill-feeling towards the company, particularly if the
policyholder did not understand the risks involved, which is then a marketing risk for the
company.
The financial risks for the life company from investment, expenses, and demographic
assumptions will depend on the nature and level of any guarantees given. The company’s
ability to respond to adverse experience by increasing charges may be constrained by
marketing considerations or by legislation.
As with non-linked designs, a loss will result from a withdrawal if the payout on withdrawal
exceeds the asset share, which is unavoidable when the asset share is negative.
An important feature of unit-linked contracts is the flexibility they can offer, for example in
terms of design and in the level and range of benefits offered.
Unit-linked contracts can be very capital efficient, depending on the charging structure and
the ability (if necessary) to take credit in the supervisory reserves for future initial expense
charges, using actuarial funding or negative non-unit reserves, as appropriate. The ability to
do this may be constrained by legislation.
Index-linked contracts
These contracts link the policyholder’s returns to an investment or economic index.
The main risk to the life insurance company is that the assets held, if they are not an exact
match, may not move in line with the index.
Risks of the products to the insured

The main risk is that the benefits ultimately paid out do not meet the need for which they
were originally intended.
This is made worse by the effects of inflation eroding the real value of the monetary benefits
provided.

Unit-linked policies generally provide greatest potential for inflation protection, conventional
(fixed-benefit) without-profits contracts provide the least, while the with-profits design
provides some protection.
Index-linking of benefits, including those of without-profits policies, can provide guaranteed

inflation protection.
Premium and benefit flexibility is also important in ensuring needs are met: unit-linked
policies are generally greatly superior from this point of view.
Purpose of the products to the insurer

The insurer aims to sell products that will maximise its profitability for an acceptable degree
of risk. Product designs that have greatest utility to policyholders should achieve this aim,
because they should sell in higher volumes and/or permit the company to build in higher
profit margins, than other less attractive designs.


4.1 A life insurance company writes regular premium unit-linked contracts.
Exam style
(i) List the different types of charge that could be levied to meet the renewal expenses and
other regular costs incurred under the contract. [3]
(ii) Describe briefly the factors that would be taken into account in deciding which of the
types of charge mentioned in part (i) should be used to cover the renewal expenses. [4]
[Total 7]
4.2 At a time of historically high yields on government fixed-interest securities, a life insurance
company has launched a packaged product that consists of two contracts:
Exam style
(a) A regular premium ten-year with-profits endowment assurance. The benefit at maturity or
death is the initial sum assured plus any attaching reversionary bonuses. A terminal bonus
may also be given at maturity or on death.
(b) A ten-year single premium level temporary annuity.
A single premium at outset will purchase the annuity and the fixed monthly payments from this
annuity will be used to pay the premiums on the endowment assurance.
The company usually invests annuity premiums in fixed-interest securities, while a large part of its
investments for with-profits endowments is in equities.
(i) Discuss the extent to which this package would meet the consumer needs of a potential
policyholder whose prime aim is to produce a capital sum for the holiday of a lifetime ten
years from now. In your solution comment briefly on the two alternatives of investing in
contract (a) only or investing in a single premium with-profits endowment assurance
only. [7]
(ii) Discuss the risks for the company in offering this package product, including the new
business strain that might arise. [11]
[Total 18]

4.3 A life insurance company writes unit-linked immediate annuity contracts. Consumers purchase
the contracts to provide themselves with an income in retirement.
Exam style
The contracts operate as follows. The policyholder pays a single premium. In exchange the
company guarantees to pay the policyholder the value of x units in its International Equities unit
fund every month until the policyholder’s death. The company determines the number x, which
depends on the premium paid and the policyholder’s age and sex, when the contract is taken out.
The International Equities fund is invested in a range of domestic and overseas equities.
The company invests the single premium into the International Equities fund at outset and so
buys units to back its expected liabilities. The values of the policyholder’s monthly payments
reflect two factors:
 the underlying changes in the value of the equities in the fund
 the effect of the insurance company’s fund management charge, which reduces the value
of each unit on a daily basis, by an amount equivalent to 1% pa.
The company expects to cover its expenses and make a profit through this fund management
charge, and by calculating the value of x using slightly prudent mortality assumptions.
No payment is made on death and no withdrawal benefit is offered.
(i) Comment on the extent to which this contract is suitable for a policyholder seeking
income in retirement. [4]
(ii) Discuss the risks to the life insurance company in writing this contract. [6]
[Total 10]

4.4 A house-owner is considering taking out an endowment assurance to repay the capital sum under
a mortgage of £60,000 due in 25 years’ time. During the term of the loan, interest payments are
Exam style
made to the lender, but no capital is repaid.
The house-owner first considers:
(a) A regular premium without-profits 25-year pure endowment where the only benefit
payable is £60,000 on survival to the maturity date.
The house-owner then considers two other products:
(b) A regular premium 25-year with-profits endowment assurance where the basic sum
assured is £40,000 payable on death or survival to the maturity date. Regular and
terminal bonuses are payable. In recent years the regular bonus has been a simple bonus
of 3% of the basic sum assured. On death before the maturity date the minimum benefit
payable is guaranteed to be £60,000.
(c) A regular premium 25-year unit-linked endowment assurance where the sum assured
payable on death before the maturity date is £60,000, or the value of units if greater. On
the maturity date the value of units is payable.
(i) Describe the risks avoided and accepted by the investor in opting for product (a) and
outline any insurance products available to overcome the risks accepted. [7]
(ii) Describe the additional risks avoided and accepted by the investor in opting for each of
products (b) and (c) compared with (a). [6]
(iii) Compare the premiums payable under each product (b) and (c) with (a), giving reasons for
the possible differences. [5]
[Total 18]



Chapter 4 Solutions
4.1 (i) Types of charge
Often on a list question, there would be no more than ½ a mark available per point, especially on a
list like this which is relatively straightforward.
 reduced allocation rate

 bid-offer spread
 administration fee / policy charge / policy fee (could be fixed or could escalate each year;
could be taken from the premium or from the unit fund)
 fund management charge
 switch charge
 mortality charge
 charges for other benefits
[½ mark per point, maximum 3]
(ii) Factors to consider when deciding on the renewal expense charges
In this part we need to think about more than just actuarial issues. The key idea is the conflict
between profitability and marketability.
Don’t worry if you don’t get all of these ideas first time through, as many are covered in more
detail in Part 3 of the course, when we deal with pricing and product design.
The main principle that should be followed is that the charge matches the expenses that are being
incurred. [½]
For example:
 Renewal commission is a percentage of premium and so the best match is a bid-offer
spread or reduced allocation. [1]
 Administration costs will be a fixed amount per policy and will rise each year. Hence the
best match is an increasing administration fee. [1]
 However the income from the fund management charge will also rise as the fund grows,
so this may also be a reasonable match. [½]
 A switch charge is the best match for the cost of administering a policyholder’s request to
switch between funds. [½]
Other considerations:
Marketing: What is the trade off between price and sales? Is there any benefit in having an
administration fee which does not inflate? or not to have one? or to offer one free
switch per year? [1]
Simplicity: For administration and marketing. [½]

Cross subsidy: How much does the company want to subsidise small policies from big ones [½]
[Maximum 4]
The following points are a bit more obscure. You probably wouldn’t be expected to get them with
just four marks on offer:
Market norm: Consider charging structures offered by competitors.
Collection cost: If the IT cost is more than the income, don’t have an admin fee.
4.2 (i) Meeting needs
The contract will provide a capital sum in ten years, because the benefit is the maturity benefit
under the ten-year endowment assurance. Therefore it meets the basic need. [1]
The indirect investment in equities should produce returns in excess of inflation, which is
appropriate to the purpose. The risk of doing so for the policyholder is reduced to some extent by
the guarantees under the contract and any smoothing of benefits by the life insurance company.
[1½]
Depending on the policyholder’s circumstances, the life cover under the endowment may not be
required. It certainly means that not all of the policyholder’s money would be directed towards
the prime objective. [1]
If the individual has a lump sum available for the single premium then the contract may be
suitable. If not, then saving through the regular premium endowment alone may be more
suitable. [1]
If the individual has a lump sum available then an alternative would be a single premium
endowment (if available) rather than the packaged product. The differences in doing so would
be:
 Different investment risk for the policyholder. For the single premium endowment, the
return from the policy will mostly reflect yields obtained on investments made at the start
of the contract. For the packaged contract, some of the return will depend on yields
obtained from future investment (of the regular annuity payments), which are subject to
more uncertainty. [1]
 On the other hand, with the package more of the return is determined from fixed-interest
assets. Provided that the products are consistently priced, this should tend to produce
higher guarantees and less potential return from the package. [1]
 The overall effect of these two opposing factors may well make the resulting package very
similar to the single premium endowment. [½]
 The policyholder may be subject to two sets of expense charges under the package, but
only one if he or she chooses the straight endowment. So the package might be worse
value for money (unless the company prices products differently when issued in a
package). [1]
[Maximum 7]

(ii) Risks
The risks for the combined contract are essentially those of the two constituent contracts, (a) and
(b). However, in some cases the risks may offset each other.
Investment return
(a) The company has some investment risk because of the guaranteed sum assured and any
reversionary bonuses granted over the term. To the extent that it can reduce bonuses it
passes some of the investment risk back to the policyholder. However, the scope for this
may be limited by smoothing of payouts and policyholder expectations. [1½]
(b) Investment return is an important assumption because the single premium produces a
significant fund. It is therefore a potential source of risk. [½]
In practice the risk may be low if the company can match the payments with suitable
fixed-interest investments. If it chooses to use more volatile investments such as equities
then the risk of not achieving the necessary return is greatly increased. [1]
We shall study life insurance investment later in the course, but the ideas underlying this should be
familiar from earlier subjects.
Expenses
(a) The company is vulnerable to higher expenses (for example through high inflation), but
the ability to reduce bonuses provides some protection. The protection may be limited by
competition or policyholder expectations. [1]
Also, high inflation may be associated with high nominal investment returns. This might
offset the need to reduce bonuses. [½]
(b) Because the contract guarantees payments, high expenses are a potential risk. Inflation
ten years away may not be easy to predict. [1]
Mortality
(a) The death benefit will be higher than the asset share at early durations and so for this
part of the contract looked at in isolation, higher-than-expected deaths would lead to a
loss. [1]
(b) Mortality is a significant assumption for annuities and so lighter than expected mortality
would lead to a loss. This might arise from general improvements in mortality (greater
than those allowed for) or from random fluctuations. [1]
Viewed overall, these risks will to some extent offset each other, but some mortality risk is likely
to remain. [½]

Withdrawals (persistency)
(a) Higher-than-expected withdrawals early on, when the asset share is negative, are likely to
lead to a loss. This might be recouped by reducing bonuses to continuing policyholders.
At later durations, whether or not a loss is made will depend on how any surrender
payment made compares with the asset share. [1]
(b) Selective withdrawals would be a major risk if withdrawal from this part of the contract
alone is allowed. Those in poor health would be most likely to withdraw. [1]
Therefore the company will probably only allow withdrawal from the whole package, not from
contract (b) alone. It could then control the overall surrender value to eliminate the loss from
early withdrawals and to try to reflect the risk of selective withdrawals. [1]
New business strain
(a) Although the guarantees are reasonably low, the monthly premium structure and
probably high initial expenses will lead to significant initial strain, after taking account of
the supervisory solvency requirements. [1]
(b) While the single premium will easily cover the initial expenses for this contract, the fully
guaranteed nature of the product will lead to a requirement for large reserves and/or
solvency capital and hence almost certainly an initial strain. [1]
It might be possible to reduce the total strain on the combined package, provided the supervisory
regulations will allow this, and if appropriate. For example, reductions in mortality risk overall
may reduce the reserves or solvency capital required.
[Maximum 11]
You may have come up with additional points, for example on business volumes and mix.
However, it would not be fair on first-timers to include such points in the marking schedule at this
stage in the course. Similarly the last point under new business strain above, while valid, has not
been included in the schedule. These points will be picked up later in the course.
4.3 This is quite a tricky question. It is important, though, that you practise applying your
understanding to unfamiliar situations.
The particular details here are just one possibility for structuring a unit-linked annuity. It is always
important to read carefully the particular details in any question, and answer accordingly.
(i) Suitability to meet the needs
The contract does provide an income for life, and so in this sense is suitable. [1]
The lack of an investment guarantee, and investment in equities, should mean a higher expected
return. [1]
However, the annuity payments may be volatile, as they are invested in equities. If the payments
are the policyholder’s main source of income then a unit-linked design may not be the most
suitable one. [1½]

The volatility of the payments may be increased by the currency risk of the overseas equities. [½]
Over the long term the value of the payments would be expected to increase at least as fast as
inflation. This is appropriate, especially if the country involved is prone to inflation. [1]
For policyholders who die soon after the annuity is started, very little money might be paid out in
benefits. The policyholders’ impression may be that, in such a case, some or even a lot of the
initial investment had not been used, and they would expect the balance to be paid back on
death. The product does not meet this possible need. [1]
In practice, products like this might be offered with a guaranteed minimum payment period
(eg five years), or with a return of balance of premium on death.
[Maximum 4]
(ii) Company risks
Mortality
This is probably the key risk. The risk is that policyholder mortality is lighter than expected, and
so the company has to pay out the value of more units than it was able to buy with the single
premium. This could reduce profits or even produce a loss. [2]
Also, if the unit fund becomes depleted more quickly than expected, the fund management
charge income is reduced. [½]
Expenses
The company is at risk of higher than expected expenses. [½]
If this is due to high inflation then it might be offset by higher nominal investment returns and
hence larger fund management charges. [½]
The company may also have the right to increase the fund charge. [½]
Investment return
This risk has largely been passed to the policyholder. However, lower fund management charges
would reduce the contract’s profitability. [1]
Other
The absence of a withdrawal payment eliminates the risk of selective withdrawals. [½]
There might be a marketing risk if the fund performs badly, even over a fairly short period. It
could lead to hardship for policyholders, ill feeling and bad publicity, especially if policyholders
had not fully appreciated the risks. [1]
The lack of any compensation on early death might lead to ill-feeling on the part of any surviving
relatives, which can again lead to a bad press with adverse effects on the company’s future sales
prospects. [1]

There might be an initial strain as there are no initial charges to offset the inevitable initial
expenses. [½]
[Maximum 6]
As in the solution to the previous question, further points on business volume and mix would be
reasonable here, but these issues are covered later in the course.
4.4 (i) Risks of (a) to the investor
Risks avoided
Investment risk is avoided as the amount payable on maturity is guaranteed to repay the total
mortgage of £60,000. [1]
There is no risk of premiums changing, so the investor has a known outgo for the next 25 years
and is therefore protected against any inflation risk. [1]
Any changes in health (which would have been unknown at the time of underwriting) will have no
effect on either the premiums payable or the benefits received. [½]
Risks accepted and products available
The main risk is of death before maturity, where the policyholder’s dependants would receive
nothing. They would be left with a property mortgaged with a loan on which interest is still
payable, and no money to repay the capital unless they sell the property. [1]
Two main products exist to cover this problem. The first is a level term assurance, which would
pay a fixed lump sum (in this case £60,000) in the event of death before maturity. [1]
An alternative, possibly cheaper option might be to combine the existing contract with the term
assurance to get a without-profits endowment, where the £60,000 is paid on maturity or earlier
death. [1]
A further risk accepted is that on surrender. There may not be a surrender benefit payable (if, for
instance, the policyholder sold the house and repaid the mortgage). [½]
One product to overcome this problem is that outlined in part (b). [½]
An alternative to accepting a poor surrender value, if it were available, would be to employ the
second-hand endowment market.
There is the risk that the company will not exist in 25 years to pay the maturity claim, although
the probability of this happening with no compensation scheme available will probably be very
small. [½]
There is a significant redundancy and sickness risk, covering the general inability to pay premiums
as a result of being unable to work. [½]

This risk can be mitigated by:

 waiver of premium benefit, which will waive the payment of premiums during such
periods [½]
 other kinds of sickness or disability insurance. [½]
Although sickness insurance is not covered explicitly by the Subject SP2 syllabus, waiver of
premium is mentioned in the Subject SP2 Core Reading, and you are expected to be generally
aware that insurance contracts exist to meet various long-term sickness and disability needs.
[Maximum 7]
(ii) Extra risks for options (b) and (c)
Option (b)
The main risk avoided is that of death before maturity. With this product, the mortgage could be
repaid on death at any time. [½]
Indeed, if bonuses have been particularly buoyant, there may even be more than enough to repay
the mortgage. This is particularly likely later on in the duration of the policy. [1]
The risk accepted is the investment risk that the amount payable on maturity would not be
sufficient to repay the £60,000. [1]
In recent years the bonus has been more than enough, since 3% simple for 25 years would accrue
to far more than £60,000, particularly once terminal bonus has been added. [½]
However, in the long term there is no guarantee, and the return will depend on investment
performance generally over the term of the policy. [½]
Option (c)
The risk avoided in the unit-linked option is again the death benefit before maturity. There is a
possibility that the death benefit would more than cover the mortgage amount, depending on the
performance of the underlying investments. [1]
The risk accepted is again the investment risk. The maturity amount will depend on the value of
the unit fund, which in turn depends on the investment performance over the term of the policy.
[1]
This means that whilst there is a chance of benefits being greater than £60,000, there is also a
chance of benefits being lower than this amount, in which case the investor would have to
supplement the benefits with money from other sources such as savings. [1]
The extent of this risk depends on the premium size and if the policy is reviewable. [1]
[Maximum 6]

(iii) Comparison of premiums
(b) compared with (a)
The premiums are likely to be higher for (b) because:

 (b) is with-profits, and although the initial guaranteed maturity benefit is somewhat
smaller at £40,000, it has the added benefit of bonuses, giving the chance of a maturity
benefit in excess of £60,000 [½]
 (b) has a death benefit [½]
 working in the opposite direction, the guaranteed maturity benefit in (a) poses a risk to
the insurer, which will be reflected in a prudent premium basis and thus higher premiums.
However, the first two factors are likely to outweigh this. [1]
(c) compared with (a)
The premiums for (c) might be bigger due to the additional death benefit. However, this benefit
costs little at young ages, and the sum at risk reduces over the policy term. [1]
The level of premium would be set at the outset by assuming a certain level of future investment
growth consistent with a £60,000 maturity amount. The investment return assumption will
heavily influence the premium. [½]
As the risk of not meeting the maturity amount is considerable, the investor may err on the side
of caution and choose to pay higher premiums (if the choice is available). [1]
Alternatively, a low initial premium could be taken, and the policy reviewed at various intervals so
that the risk of not meeting the maturity amount could be assessed and the premium increased if
necessary. [½]
The overall premium size thus depends on these factors and may be greater or smaller than that
in (a). [½]
[Maximum 5]

SP2-05: Asset shares Page 1
Asset shares
Syllabus objective
2.2.2 Explain the main uses of asset shares and how they may be built up using a
recursive formula.

Page 2 SP2-05: Asset shares
0 Introduction
In this chapter we introduce the concept of the asset share. The asset share is a widely used tool
in the management of with-profits business. In a nutshell, the asset share of any policy is the
accumulation of premiums paid rolled up with interest, less expenses and the cost of any cover. It
is therefore very similar to a gross premium retrospective reserve studied in Subject CM1 or CT5.
However, when we calculate asset shares we normally use the experience of the policy to date,
rather than any assumed basis.
The concept of asset share can be applied to individual policies, or to a group of policies. We can
calculate the asset share for an individual policy just as we can calculate the retrospective reserve
for an individual policy. Or, we could consider a group of similar policies all issued at roughly the
same time and monitor their progress.
To avoid tedious repetition, we shall normally consider the asset shares of individual policies, but
most things in this chapter could equally be done in respect of a group of policies.

1 Determining asset shares

Asset shares are most commonly used for with-profits contracts, but can also be
determined for without-profits contracts. Asset shares (sometimes called earned asset
shares) can be evaluated for individual policies or for a block of policies issued with similar
terms and conditions.
At the most generic level, the asset share for a life insurance policy is the accumulation of monies
in less monies out in respect of that policy. In other words, the accumulated cashflow in respect
of a policy.
1.1 Components of the asset share

The ‘monies in’ for a policy will obviously include the premiums paid to date, and the investment
income. Investment income would include any unrealised capital appreciation on the assets
concerned.
A less obvious source of ‘monies in’ occurs when we consider with-profits business. In this
context, the life insurance company may be organised so that profits from without-profits
business flow into the with-profits fund and become the property of with-profits policyholders. In
this case, the profits from such business would be added to the asset shares of the with-profits
policies.
The ‘monies out’ will include the expenses involved with the contract, and the cost of any cover
provided. However, there are several other ways in which the contract can be deemed to cost
the life insurance company some money, as introduced in the following Core Reading. We shall
then consider what these deductions really mean.
The asset share is the accumulation of premiums less the deductions associated with the
contract, all accumulated at the actual rate of return earned on investments.
Deductions include all expenditure associated with the contract(s), in particular:
 expenses incurred and any commissions paid
 the cost of providing all benefits in excess of asset share (eg life cover or any other
guarantees or options granted) possibly on a smoothed, rather than current cost,
basis
 tax (if appropriate) including any reserves made for future tax liabilities
 transfers of profit to shareholders
 the costs of any capital necessary to support contracts in the early years
 any contribution to the free assets which, in turn, support the smoothing of bonuses
and the ability to exercise greater investment flexibility.
If other than asset share is paid out on average on surrender, then the calculation could
also include addition of surrender profits (or deduction of losses) from with-profits
business.
A further potential addition to the asset share is an allocation of profits / losses on without-
profits business, depending on the participation structure.

We now consider some of these in more detail.
Expenses incurred and any commissions paid

We need to deduct all expenses associated with the policy. This will include acquisition expenses
such as any commission payable to financial advisers, and similarly any renewal commission. It
would be both normal and fair to use all of the expenses incurred in respect of the policy,
including its share of the life insurance company’s overheads.
The cost of providing all benefits in excess of asset share

The cost of providing death benefits for an individual policy is not an immediately obvious thing to
grasp.
Question
‘The cost of having provided cover for a life policy is obviously zero if there hasn’t been a claim on
the policy, and is equal to the sum insured if there has been a claim.’
State whether this is true, providing justification of your answer.
Solution
The suggested approach is not valid, because it is measuring the outcome of the policy’s cover in
respect of some risk, rather than the risk itself. It would lead to meaningless results: there is not
much use in knowing that the asset share of a policy which has just claimed is –£87,934; while
asset shares for policies surviving to maturity would be overstated because no allowance would
have been made for the risk that the company took in insuring their lives over the term to
maturity.
It follows from the solution to the above question, that the cost of providing death benefits must
lie somewhere between zero and the sum assured. The form of this deduction will depend on
whether we are calculating asset shares on an individual policy basis or in respect of a group of
policies (ie on an aggregate basis).
Let’s start by considering an aggregate basis. Imagine a group of n identical policies, each with an
asset share of ASt at time t.

If we consider the cashflows for this group of policies over the next year say, we can develop an
expression for the aggregate asset share at time t+1:
n  ASt  P  Et 1  it   Snqx t
where P = annual premium

Et = per-policy expenses incurred at time t
it = investment return achieved in year t to t+1
S = death benefit sum assured

qx t = actual mortality rate in the year t to t+1.
If we want to know the asset share of an individual policy at time t+1, then we need to divide this
aggregate asset share by the number of policies that are still in force at time t+1. Doing this gives:
n  ASt  P  Et 1  it   Snqx t

ASt 1 
n(1  qx t )
Rearranging this gives:
ASt 1  (1  qx t )   ASt  P  Et  1  i   qx t S
or:
ASt 1   ASt  P  Et  1  i   qx t  S  ASt 1 
As this formula shows, when doing individual asset share calculations, we deduct the death
benefit in excess of the asset share. Notice that when doing the aggregate asset share
calculations, we deducted the full death benefit sum assured.
Question
Describe how the cost of cover might be calculated in order to determine the asset share for a
very large group of policies.

Solution
If we had a very large group of policies (say at least several thousand) then we could get
meaningful results for the asset share of that group by taking ‘total claim payments in the year’ as
the cost of cover in any particular year, ie in the aggregate asset share calculation deducting
‘number of actual deaths  S ‘ rather than S nqx t . However, the group would have to be very
large for there not to be any noticeable random fluctuation from year to year.
In practice, even large life companies would not follow this approach but would first use their
claims data to determine crude mortality rates, graduate these rates and then calculate the cost
of cover using these smoothed rates, ie the qx t rates. (So this is what the Core Reading is
meaning when it says that the cost of benefits is deducted on a ‘smoothed, rather than current
cost, basis’.)
In addition to death benefits, we should also deduct something in respect of any guarantees or
options.
For example, suppose we are calculating the asset share of an endowment insurance policy and
the contract gives the policyholder a scale of guaranteed surrender values. We should allow for
the fact that some policyholders might have exercised the option to surrender when economic
conditions were bad and the guaranteed surrender value exceeded the policy’s asset share. This
would cause a net loss, which would have to be shared between (ie deducted from) the asset
shares of the continuing policyholders.
How we cost such guarantees and options is considered in Chapter 23.
Question
Explain what happens to the asset share when surrender profits are made.

Solution
The asset shares of continuing policies would be increased. In theory the increase per policy will
be equal to the total surrender profit made, shared between the surviving policyholders. For
example, suppose the surrender of one policy at a particular time produces a profit of £70, and 25
out of 1,000 similar policyholders surrender. The increase in the asset share of each continuing
policy would be:
70  25
 1.79
975
70  0.025

1  0.025
qw
 SP 
1  qw
where qw is the rate of withdrawal and SP is the surrender profit from one policy.
Tax (if appropriate)

The adjustment made in respect of tax will vary from country to country, and may differ for
different types of contract. The important idea here is that the calculation of the asset share
reflects how the policies are taxed (where we mean tax that the company pays, not that which
the policyholder pays).
For example, any tax already paid on the investment income that has gone into the positive side
of the asset share ‘equation’ should be deducted for consistency.
Less obviously, if we have included unrealised capital appreciation as a plus in the asset share
then we should also deduct any reserves required for the future tax liability that we would incur
on realising that capital appreciation.
Transfers of profit to shareholders

Clearly if part of the surplus has been transferred to the shareholders then that money is no
longer around, and should be deducted from the asset share.
Cost of capital
We have already encountered the phenomenon of new business strain, whereby the company
makes a loss on writing a new policy. While there are a number of things a company can do to
reduce new business strain, such as negative non-unit reserves (described later in the course), it is
very unlikely to be able to eliminate it altogether. Most life insurance companies therefore find
that writing business causes some new business strain.

Where does the money come from to fund this strain? In effect, it comes from the surplus built
up over previous years from other policies but not yet distributed, and/or from the shareholders’
free assets. These other policyholders, and shareholders, have in effect ‘invested in’ the new
policy by putting up the money to fund the new business strain. They therefore require some
money back from this ‘investment’.
We therefore need to deduct from the asset share of a policy the cost of this ‘loan’ provided at
inception by other policyholders and/or shareholders. There is then the issue of who should
benefit from the returns thus earned on the capital. The shareholders will naturally get money
back from with-profits policyholders via the shareholders’ transfers, and this would normally be
designed to provide the necessary return on their capital. The capital provided by policyholders
cannot be easily assigned to individual existing policies, particularly as most of the capital will
have come from policies that are no longer in force. However, at times the insurer may wish to
reduce (or increase) its existing capital supply, and so may make further (gradual) adjustments to
existing policies’ asset shares to this effect.
For any single policy, therefore, we would expect deductions from asset shares at early durations
(to pay for the capital being required at those times). There may be further additions or
deductions at later durations, reflecting the company’s overall policy to expand or contract the
company’s capital base in addition to any retention of shareholders’ profits for this purpose.
This is not an exact science, and not all companies do this. Details are not required for the Subject
SP2 exam.
Whatever approach is used, the main reason behind the adjustments made to asset shares is to
ensure that the different groups of with-profits policyholders are treated fairly.
Contribution to the free assets

As we will see in later chapters, it is common to smooth payouts on with-profits policies to
protect the policyholders from fluctuations in the value of the underlying assets. The payout to a
with-profits policyholder shortly after a downturn in the investment markets would normally be
greater than the asset share of that policy; while the payout shortly after a market boom would
be less than asset share.
This system is going to work only if there is a large amount of free assets to use to effect this
smoothing – if we want to pay more than asset share at times, then we need spare assets
(ie capital) to fund the difference.
Additionally, the presence of free assets also increases the company’s ability to take on
investment risk, and so allows greater investment flexibility. This is covered more fully in the
investment chapter (Chapter 28).

As with the new business strain described above, the providers of capital should be suitably
reimbursed for the use of this capital. Unlike that case, however, this capital is required
throughout the policy term, and in fact more towards the end rather than at the beginning. For a
stable portfolio of policies, it is likely that deductions from and additions to asset shares for the
cost of this capital might be expected to cancel out and may quite probably be ignored. However,
as before there may well be times when the pool of free assets needs enlarging or contracting, in
order to ensure the most effective use of that capital, in which case net deductions (or net
additions) to asset shares would be made.
1.2 Asset share calculation

The asset share can be calculated recursively on a year to year basis. Initially the earned
asset share is zero. Each year the cashflows including premiums received, any
miscellaneous profits to be allocated, and deductions made (as listed in Section 1.1 above)
are recorded. A suitable rate of return on investments is used to accumulate the asset
shares plus premiums less deductions plus miscellaneous profit allocations to the year
end, to determine the asset share. This process is repeated for subsequent years.
Question
A number of with-profits regular premium endowment assurance policies were issued on

1 January 2015 with premiums of £1,000 pa paid annually and sum insured of £15,000 under each
policy. The following average experience occurred:
Year 2015 2016 2017

Investment return 12.6% 8.5% 7.9%
Mortality 0.0014 0.0015 0.0017
Expenses £1,320 £84 £89
The net effect of other sources of profit and charges for use of capital etc was a reduction in the
yield of 0.3% pa.
Assume that expenses occur on premium payment (at the start of the year) and that claims are
paid at the end of the year.
Calculate the asset share for each policy at the end of 2017.

Solution
The asset share at the end of 2015, per policy then in force, was (using the formula from page 5):
AS1 
 0  1,000  1,320   1.123  15,000  0.0014  380.89
.
1  0.0014
AS2 
 380.89  1,000  84   1.082  15,000  0.0015  557.32
Similarly
1  0.0015
AS3 
 557.32  1,000  89  1.076  15,000  0.0017  1,557.06
and
1  0.0017
so the asset share at the end of 2017 is £1,557.
As before, each year we divide by the proportion of policyholders who survive the year. This is
because asset shares are, by definition, expressed per policy in force at any given time.
The result for year 1 in the solution above might be surprising. We shall discuss the idea of
negative asset shares in the next section.
1.3 Typical asset share developments

What do asset shares normally look like as they develop over time?
For a regular premium policy, one of the most important influences is the incidence of expenses.
We normally find very heavy initial expenses, caused by the cost of acquiring the business, and
then much lighter expenses subsequent to that. As the policyholder ages, we expect the cost of
life cover to increase year by year. Investment income will have a compounding effect, unless
deductions due to shareholder transfers are particularly big.
At maturity, we would expect the asset share to exceed the guaranteed sum insured. The excess
is the profit on the policy, which would then be distributed to shareholders and/or the
policyholder depending on whether the policy is with or without-profits.

For example, for a regular annual premium ten-year endowment assurance policy, with sum
insured £10,000, we would expect the asset share to look like:
12,000
10,000
8,000
6,000
4,000
2,000
-2,000
0 1 2 3 4 5 6 7 8 9 10
Year
We can see how the impact of high initial expenses gives a negative asset share immediately after
inception. Thereafter this asset share grows in a compound fashion.
How can an asset share be negative? This might seem a counter-intuitive idea, but given our
definition of asset share, if a policy has incurred expenses to date in excess of the premiums then
it seems reasonable to say that the policy represents a negative for the company. The
phenomenon of negative asset shares at early duration for regular premium contracts is a big
problem for life companies, because of the risk of policies lapsing before the asset share grows
above zero. If this happens, the company will have made a loss on that policy.
How the asset share moves after inception will depend primarily on how the three major
components of renewal expenses, investment income and cost of life cover all balance out. In the
example above, investment income was greater than the other items and so the asset share
compounded up. If investment yields had been much lower, or expenses or cost of cover much
higher, we might have seen a convex shape where the asset share grows but at a gradually
reducing rate.
Question
Sketch the asset share development for a regular annual premium term assurance contract.

Solution
We should see something like the following:
1,600
800
-800
0 1 2 3 4 5 6 7 8 9 10
Year
The steep falls every year represent the cost of death claims; the rises at each policy anniversary
represent the annual premium.
If we just plotted the end-of-year values it would look like the following graph, which shows the
overall trend:
1,600
800
-800
0 1 2 3 4 5 6 7 8 9 10
Year
The asset share is initially negative because of very high initial expenses.
It then grows because each year’s premium exceeds the cost of life cover in the early years. Then,
later on, as the policyholder ages, the cost of cover rises above the level of the premiums and the
asset share decreases.
At the end of the policy’s term the asset share is small and positive. There is no maturity benefit
to provide, but the company should be making some profit on the contract.

To give an idea of the relative values involved, the annual premium used for the above example
was £900, the sum insured was £100,000.
In real life, most term assurance policies have monthly premiums. This will make the steps much
closer together and smaller, and producing a graph looking more like the overall trend in
appearance.
Question
Comment on how the following might affect the individual (per-policy) asset shares for a group of
regular premium whole life with-profits policies at the same duration:
(i) several of the policyholders surrender
(ii) the market value of assets decreases by 10%
(iii) the shareholders’ required rate of return on capital increases
(iv) the supervisory authority strengthens the reserving basis for the policies, so reserves
increase
(v) one of the policyholders dies.
Solution
(i) Surrender of policyholders
The impact on the asset shares for remaining policyholders depends entirely on how the
surrendering policyholders were treated. If they were paid surrender values equal to asset share
less the administration costs of the surrender then the asset shares of the remaining
policyholders will be unaffected. If, on the other hand, they were paid surrender values lower
than that, the company will have made some profit from their surrender and this should be
reflected in higher asset shares for the remaining with-profits policyholders.
(ii) Market value of assets down 10%
The immediate answer would be ‘asset shares down by 10%’. However, if the fall in market
values also reduces the tax liability on total unrealised capital gains, then asset shares will fall by
less than 10%.
(iii) Shareholders rate of return increase
This will increase the cost of the shareholders’ capital. As a result, the shareholders’ share of the
distributed profit may be increased, increasing the amount of shareholders’ transfer each year
(although the extent to which this can be done may be heavily restricted by regulation). This will
reduce asset shares.

(iv) Reserves increased
At an immediate level, there will be no change because asset shares are affected only by
cashflows, and the level of the reserves is therefore irrelevant.
However, a number of second-order effects may occur:

 A strengthening of the reserving basis will mean that the capital required to sell a policy
will now be higher. Larger deductions should therefore be made from asset shares at
early policy durations, to pay for the increased cost of capital, for all policies that require
more capital than before. This will reduce the future asset shares for such policies.
 An increase in reserves will reduce the insurer’s investment freedom. This will reduce the
future investment return potential, which will in turn result in lower asset shares
ultimately.
(v) Policyholder dies
If the group asset share is being calculated by totalling actual cashflows for the group, then the
asset share will clearly go down a little. If the approach taken is that of determining the mortality
rates experienced by the group, and then using those rates to calculate asset share, then asset
share will go down if our estimate of mortality rates increases as a result of the death. In practice,
one death would have no influence on our rates and so the asset share would not be affected.

2 Asset shares and surrender values

The asset share will, over a period of time and allowing for smoothing, be the upper limit on
a policy’s surrender value.
We have mentioned in earlier chapters (and also in the previous section) that when a policy is
surrendered, the company makes a profit equal to the difference between the then asset share of
the policy and its surrender value. For many conventional life insurance products, the amount to
pay as a surrender value is at the discretion of the company. So one thing the company will
always do when determining the surrender value is to try to ensure that it is not bigger than the
asset share.
Question
Explain why the surrender value should not be bigger than the asset share.
Solution
If we pay a benefit on surrender that is bigger than the asset share, then the overall financial
effect on the company of that policy’s existence is to cause it a loss, equal to the excess of the
surrender value over the asset share as at the date of surrender.
The details of calculating surrender values are discussed later in the course.

3 Asset share and bonus distribution

Asset shares are a key element in the process of determining profit distribution: this is described
in Chapter 6.

Chapter 5 Summary
The asset share of a policy is the accumulation of
 premiums
 investment income
 miscellaneous profits
less
 expenses and any commissions
 cost of all benefits in excess of asset share
 tax
 profit transfers to shareholders
 cost of capital required to support new business strain
 any contribution to the free assets required to support smoothing of bonuses and
increase investment freedom.
The calculation can be summarised by the following recursive formula:
ASt 1 
 ASt  P  Et 1  it   S qx t
1  qx t
where:
ASt  asset share at t years
qx t  mortality rate for year t to t  1
Et  per-policy expenses incurred at time t
P  premium
S  death benefit sum assured

it  investment return achieved in year t to t  1.
The asset share is the maximum surrender value payable under any policy in order to
prevent the policy from causing a loss to the company.
The asset share is a key tool in the determination of with-profits bonus rates.



5.1 (i) Define the term ‘earned asset share’. [2]
Exam style
(ii) List the key components involved in the calculation of an earned asset share. [4]
[Total 6]
5.2 In the context of a life insurance contract, show how an asset share may be built up using a
recursive formula.
5.3 On 1 January a life insurance company sells 1,000 identical term assurance policies, all to lives
Exam style
aged 30. The policies are for a term of 30 years, and have a sum assured of £50,000. There is one
death in the first year and two deaths in the second year. The office experience is as follows:
 interest: 5% per annum
 premiums: £120 per annum, payable at the start of each year
 expenses: £100 initially, £10 at the start of each subsequent year.
(i) Calculate the aggregate asset share for the group of policies at the end of each of the first
two years. [2]
(ii) Calculate the asset share per policy at the end of each of the first two years. [1]
[Total 3]



Chapter 5 Solutions
5.1 (i) Earned asset share
The earned asset share is the accumulation of the premiums paid, less deductions, plus
allocations of miscellaneous profits, at a suitable rate of investment return. [1½]
The level of allocations and deductions and the rates of investment return will usually reflect the
company’s past experience. [½]
[Total 2]
(ii) Key components of the earned asset share
These are:
 premiums
 investment return earned, including capital gains
 expenses paid
 commission paid
 accrued tax
 allowance for death and surrender claims paid
 charges for guarantees
 charge to cover capital used to smooth bonuses (etc)
 transfers to shareholders
 miscellaneous profits from other sources (eg the returns from capital provided, and from
without-profits business).
[½ per point, maximum 4]
5.2 We use, for example, a whole of life assurance contract.
Defining:
Pt  premiums received during year t (assumed received at start of year)
Et  the policy’s share of the actual expenses during year t (assumed incurred at the start of
the year)
S sum assured payable on death (claims assumed to be payable at the end of the year of
death)
At  policy asset share at the end of year t
qt  actual proportion of policyholders dying over year t
it  actual rate of investment return over year t.

The value of the assets that build up by the end of year 1, from one policy issued at the start of
the year, is:
 P1  E1 1  i1   q1 S
The policy’s share of these assets at time 1 (the asset share) is:
 P  E 1  i1   q1 S
A1  1 1
1  q1
because no assets are required, at the end of the year, for those policyholders who died during
the year (the cost of all amounts due to these policyholders have already been deducted through
q1 S).
The next year’s cashflow will accumulate with the existing asset share to produce next year’s
value, ie:
 A1  P2  E2 1  i2   q2 S
A2 
1  q2
and so on for successive years.
5.3 (i) Taking the actual cashflows and accumulating them gives the following:
Year 1: 1,000(120  100)  1.05  50,000  29,000 [1]
Year 2: (29,000  999(120  10))  1.05  100,000  15,066 [1]

[Total 2]
(We have assumed here that the benefit is payable at the end of the year of death.)
(ii) So dividing by the number of policies in force at the end of each year:
Year 1: 29,000 / 999  £29.03 [½]
Year 2: 15,066 / 997  £15.11 [½]

[Total 1]

SP2-06: With-profits surplus distribution (1) Page 1
With-profits surplus
distribution (1)
Syllabus objective

Page 2 SP2-06: With-profits surplus distribution (1)
0 Introduction
In this chapter and the next we shall be considering three different approaches to distributing
profits on with-profits contracts. In this context, profits are also often referred to as ‘surplus’, or
more accurately as ‘surplus arising’.
In Section 2 of this chapter we give a brief description of what is meant by surplus. In Section 3
we consider the ‘additions to benefits’ approach to distributing surplus, as used in the UK (and
some other countries). This approach applies to two distinct types of benefit structure:
conventional with-profits and accumulating with-profits.
We shall consider two further methods of distributing profit, used in other parts of the world, in
Chapter 7.

1 Possible ways of distributing profits to with-profits policyholders

Notwithstanding any prior constraints placed upon the insurer as to how profits must be
distributed, the main logical possibilities are one (or a combination) of:
 cash bonus
 premium reduction (which can be viewed as a form of cash bonus)
 benefit increase.
Question
A life insurance company has distributed its with-profits surplus as cash bonuses for many years.
For all new policies to be issued after a certain future date, it now intends to distribute surplus by
benefit increases only.
State four possible constraints that may prevent the company from bringing in this approach, or
at least may make it difficult for it to do so.
Solution
Possible reasons include:

 Legislation – there may be a legal requirement that prevents profits from being
distributed in this way.
 Policyholders’ reasonable expectations – even though we are considering new policies
only, the market (of existing and potential future policyholders and their advisers) will not
be expecting this new approach. So these products may not appeal to the company’s
normal clientele, and any inability to win them over (or nurture new clients to replace
them) will put a marketing pressure on the company against introducing the new
structure.
 Competition – if none of the company’s competitors have a similar system then the
product may appear uncompetitive, and suffer marketing failure.
 Systems limitations – current systems may not be able to cope with handling the different
data required for the new system (or the expense of conversion may be prohibitive).
 Management limitations – management of the new structure may require different
techniques and expertise: lack of this (or the expense of introducing it) may be
prohibitive.
 Articles of association – there may be conditions set out in the company’s articles of
association, or constitution, that dictate how policy profits must be distributed.

1.1 Deferring the distribution of profits

From the point of view of the insurer, the probability of remaining solvent is increased by
reducing and delaying the distribution of any available profits. To the extent that greater
investment freedom results from delaying the distribution of profits, the long-term trend of
available bonuses may also be higher as a result.
This is an important point.
Question
Explain how delaying profit distribution can reduce the probability of insolvency.
Solution
At a superficial level, this seems easy to explain: if the company doesn’t pay out profits now, then
it must retain them; if it retains profits then it is basically increasing its current level of free assets,
so that it is more likely to withstand a financial calamity.
At a deeper level, it may not be so clear. It depends on what we mean by ‘delaying’. Delaying a
particular amount of profit may still imply that it is going to be paid out at some future date. In
such a case the amount becomes a fixed liability for the company and is then unlikely to improve
the company’s solvency position to any significant extent.
On the other hand, if delaying the distribution also implies that the company is deferring its
decision on how much it is going to pay out when the time eventually comes to do so, then the
profit amount is not yet promised even if the company has an obligation to pay out what it will
eventually earn. This, logically, must reduce the risk: it is saying that the company should
eventually have to pay out something in respect of those retained profits but the monetary
amount can be, for example, reduced if those profits have fallen in value by the time they have to
paid.
It is important to recognise that giving a bonus as a benefit increase (to be paid out on a future
claim), rather than paying it as immediate cash, is not a delay in the profit distribution.
Question
‘Surely the policyholder will receive the bonus later, so of course the distribution is delayed.’
Explain the fallacy of this argument.

Solution
The delay here is in the payment of the profit amount, whereas the profit is distributed as soon as
the company announces the amount of profit that will be paid. (This announcement is referred to
as a profit declaration.) This is because, by making the declaration, the company has made a
contractually binding promise to pay the stated amount of profit on the policy claim date. So,
whatever happens, say, to policy asset shares between the date of declaration and the date of
claim payment, the insurer will be obliged to pay out the monetary amount of bonus declared.
If delaying profit distribution reduces the probability of insolvency, then this will give the
company the freedom to invest in more risky assets. So, if we were to invest our retained profit in
equities and property, for example, and those assets had reduced in value by the time the policies
came to claim, then the company would simply pay lower bonuses as a result. Of course, the only
point of investing in higher risk assets is to improve returns, so the hope is that future bonus rates
can be higher to reflect the higher returns that we would have hoped to have achieved. This is
what the last sentence of the above Core Reading is referring to.
So on the one hand we are saying:

 pay out less profit now than we’ve earned, in order to defer our profit distribution
while on the other hand we are saying:
 total payouts will be increased because of the higher bonuses declared.
How can both be achieved at the same time?
The answer is (mainly) by using a terminal profit distribution. So, all the profits we retain from
underpaying bonuses throughout the policy term, we pay back in a big lump (hopefully) as an
extra distribution on the day the policy terminates. The higher returns that we hope to have
made in the meantime should result in higher terminal distributions being paid, at least in the
long run.
However, such an approach is likely to be relatively unpopular with policyholders and will
make the marketing of the insurer’s contracts more difficult.
This is because people dislike the uncertainty of not knowing how much the terminal distribution
is going to be. In theory this uncertainty lasts until the day the claim benefits are paid out. In
practice, it is not as bad as this. A company that reduces terminal distributions heavily over short
time periods will find itself very unpopular with policyholders. This may give the company a bad
name, which could have a serious adverse impact on future policy sales. This has given rise to the
practice of smoothing the terminal profit distribution over time relative to actual profits.
1.2 Meeting policyholders’ needs

Assuming that the policyholder has an eventual use for the ultimate policy benefits, it may
be considered unfortunate to put policyholders in a position of receiving cash, or premium
reductions, only later to find that future bonuses are reduced and the ultimate policy
benefits are insufficient for the policyholder’s needs. Avoiding this situation would suggest
benefit increases in preference to the alternatives for distributing profits.

The assumption at the start of this paragraph follows logically if we are considering a policy with
annual premiums and a lump sum benefit. Paying back some of the policyholder’s money as a
reduction in premiums would seem contrary to his or her apparent desire to save, and this would
be particularly sad if the lump sum benefit turns out to be inadequate to meet the required needs
at the time of claim.
Question
(i) Suggest two situations for which a with-profits policy, paying annual cash dividends,
would be well suited.
(ii) Outline, for each suggestion, the most appropriate version of this policy to meet the
customer’s needs.
Solution
(i) Cash dividend with-profits policy
This would be good for a policyholder who has a large amount of assets to invest now, but who
has little future income supply.
Another example would be a person with a fixed monetary liability at a future date (eg a loan
repayment).
(ii) Suitable versions
In the first case, a single premium with-profits endowment with annual cash bonus distributions
would meet the need of supplying income, while the lump sum survival or death benefit would
essentially protect the policyholder’s capital.
In the second case, a regular premium with-profits endowment policy, probably with premium
reductions rather than cash bonuses, would be suitable, with the sum assured payable on death
or maturity equal to the amount of the loan. It would not be necessary to increase the benefits,
but the reductions in premiums would keep the cost of the loan repayment to a minimum.
In practice, alternative methods of distributing bonuses have built up over time in different
countries and markets. This chapter and Chapter 7 discuss three methods used for
distributing profits:
 additions to benefits method
 revalorisation method
 contribution method.

2 Distribution of profits
Before we can talk more about distributing profits on with-profits business, we need to know
exactly what profits we are talking about. For the most part we are talking about the excess
investment return on the premiums paid, over and above the guaranteed return.
For example, the initial level of guaranteed benefit on a conventional with-profits contract would
normally be set well below the total expected affordable benefit. The initial guarantee may well
be affordable given a very low achieved investment return. Any excess actually earned can be
used to pay more than the initial guarantee, in some form or other. The scope for increasing
benefits is not a surprise, but an expectation.
Profits may also arise from other sources, for example from mortality or expense experience
being better than allowed for, although the extent to which these will be distributed to
policyholders will depend on the profit distribution method, country and even insurance company
concerned. In any case, these sources will usually be much less significant than the investment
profit. The main reason for this is that long-term mortality and expense experience can be more
accurately predicted than investment performance. Therefore, the premium basis can be based
on a more realistic forecast for these, compared with the deliberately low investment return that
companies guarantee on with-profits policies.
Of course ‘profits’ are not inevitably positive. Losses may also arise.
Although we are discussing the distribution of surplus to with-profits policyholders, some of this
surplus may have arisen from without-profits or unit-linked policies. Whether or not with-profits
policyholders are entitled to profits made on this business depends on the particular
circumstances of each company.
There is one further aspect we shall consider in the distribution of profits on with-profits policies.
When a company has shareholders (ie is not a mutual), consideration must be given as to how
profits are to be divided between shareholders and policyholders. This issue is resolved in various
ways in different countries, and in many cases is restricted by legislation of some form.

3 Additions to benefits approach
3.1 Introduction
The name of this method reflects the fact that profits are distributed in direct relation to the
current benefits under each contract. It is the method used for UK with-profits business
and has also been adopted by other countries, eg parts of Asia.
As we mentioned at the start of the chapter, this profit distribution method takes two distinct
forms. In this section and in Sections 3.2, 3.3 and 3.4, we describe the conventional with-profits
design. We describe accumulating with-profits in Section 3.5.
Conventional with-profits contracts were commonly sold in the UK until the 1990s, and even
following their subsequent decline in popularity many insurers have a lot of policies still in force
(mostly pension plans and whole life assurances as these have the longest terms).
The conventional additions to benefits approach is the system of distributing profits to

policyholders, using reversionary and terminal bonuses, in which the (usually annual) bonus
added to each policy is defined as some proportion of the benefits currently payable on a claim
under that policy. The bonuses so calculated are then added to the contractual benefits payable
under each contract. Hence, the distributed profits are paid out along with the original sum
assured when there is a contractual claim under the contract, for example at death or maturity.
This is not, in fact, the only method which can involve adding to benefits: both the revalorisation
and contribution methods (covered in Chapter 7) can or always involve this also, but have other
particular features that distinguish them both from this method and from each other.
Under the ‘additions to benefits’ method (as defined above), the initial guaranteed sum assured
(‘basic benefit’) may be increased by bonuses of three kinds:
 regular reversionary bonuses, added throughout the contract term
 a special reversionary bonus, added as a ‘one-off’ from time to time
 a terminal bonus, paid when the contract reaches maturity and possibly also on death or
surrender.
3.2 Regular reversionary bonuses

A regular reversionary bonus is a reversionary bonus that is declared on a regular basis,
usually each year, throughout the lifetime of a contract. Once declared it becomes attached
to the basic benefits and is guaranteed, ie it cannot be taken away.
The granting of a regular reversionary bonus is regarded as a distribution of surplus to the

policyholder, as we explained earlier, even though the policyholder will not actually receive the
benefit until the policy becomes a claim.

The amount of the bonus can be calculated in one of three ways:
 simple – the bonus is expressed as a percentage of the basic benefit under the
contract
 compound – the bonus is expressed as a percentage of the basic benefit plus any
already attaching bonuses
 super-compound – the bonus is expressed in terms of two percentages: one applied

to the basic benefit and a second applied to any already attaching bonuses. Where
this method is used, the second percentage is typically higher than the first.
The super-compound bonus structure is also known as a two-tier bonus system.
The following table shows examples of simple, compound and super-compound bonus structures
that might be suitable today given recent investment performance in the UK.
Type of bonus Bonus on initial guaranteed Bonus on past bonus

sum assured
Simple 2.0% Not applicable
Compound 1.0%
1.0% 2.0%
Super-compound 0.5% 1.0%
0.0% 0.0%
However, much higher rates of bonus were seen in the past, when investment conditions were
more favourable.
For a given total amount of reversionary bonus at the maturity date of an endowment
assurance, the super-compound approach defers the distribution of surplus – as bonus –
more than does the compound approach. The compound approach in turn defers it more
than does the simple approach.
In the first few years of the contract, the initial guaranteed sum assured is much larger than the
attaching bonuses. Therefore super-compound bonus methods, which give a lower percentage
bonus to the initial guaranteed sum assured, produce lower amounts of bonus in the early years.
Once significant bonuses have built up under the contract (towards maturity), the higher figure on
bonuses becomes more significant. The bonuses given late in the contract will be higher than
under a simple or compound system, all other things being equal.

Question
Sketch, on the graph below, the approximate build up of the guaranteed sum assured plus
reversionary bonus for a twenty-five year with-profits endowment assurance, under each of the
following bonus approaches:
 simple
 compound
 super-compound.
Assume that:
 the rates for the different systems are such as to produce the same eventual total of
reversionary bonus, namely £15,000 bonus on an initial guarantee of £10,000
 the bonus rates, under each system, are level over the policy term.
[Once you have sketched the build up of benefits, you may like to check your sketch using a
spreadsheet package to produce a graph.]
Sum assured
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Duration

Solution
30,000
Sum assured
25,000
20,000
15,000
10,000
5,000
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Duration
Simple Compound Super-compound
Thus the bonus rates used with the simple, compound and super-compound methods can be set
to give the same total bonus over the policy term. However, deferring the build up of bonuses
under the contract keeps the guarantees under the contract lower. This in turn keeps the
supervisory reserves lower. Therefore super-compound systems are more capital efficient than
compound systems, which in turn are more efficient than simple bonus systems.
It is important to note two other things about this example. Firstly, the assumption of level bonus
rates throughout the term of a contract would be very unusual in practice, unless investment and
economic conditions were very stable over time. Secondly we have assumed that each system
involves the same basic sum assured, whereas in practice companies operating different bonus
systems are likely also to offer different guaranteed sum assured levels for otherwise equivalent
contracts.
Question
State the system that would have the highest basic sum assured, and the system that would have
the lowest.
Justify your answer.

Solution
The highest basic sum assured would be expected with the simple bonus system, the lowest with
the super-compound system.
For example, if we start with the same sum assured and apply the same rates of simple or
compound bonus, then the compound bonus will produce a higher maturity value. Hence, to
produce the same maturity value (using similar rates of bonus) the compound system requires a
lower initial sum assured.
This is assuming that policyholders are expecting their policies to provide similar levels of financial
guarantee towards the maturity date of their policies, rather than at the outset. If the converse
were the case, then the basic sum assured would be similar under all three systems.
By offering lower basic sums assured, the reserves at the start of the contract are reduced still
further, which will further improve the capital efficiency of the contract.
Determining regular reversionary bonus rates

The Core Reading and Syllabus for Subject SP2 do not explicitly require you to demonstrate a
knowledge of how bonus rates are determined, although the Core Reading does expect you to
understand what the bonus earning capacity of a block of with-profits business is. However, an
understanding of the general way in which bonus rates are determined (and their equivalent for
the other methods) will help your understanding of this subject. Hence an outline (only) is given
below.
One of the features of this system of bonus distribution is that the regular reversionary bonus
rates tend not to vary greatly from one year to another, and when they do change, they tend to
change gradually over a period of years, rather than suddenly over just one year. This practice
has come to be expected by people who take out such policies, (and form part of what are known
as policyholders’ reasonable expectations, or PRE).
A company will therefore aim to declare, if at all possible, regular bonus rates which are the same
each year. This it could do easily if long-term investment conditions remained constant over time.
In reality, this is not the case. Hence the company needs to look forward into the future and
estimate the level of regular bonus that it could reasonably afford to continue paying into the
future according to its assessment of its future experience. This is the bonus earning capacity,
which is defined by the Core Reading in the glossary as follows:
The bonus earning capacity of a block of contracts is the rate – or rates – of bonus that
those contracts can sustain over their future lifetime, on the basis of a set of assumptions
with regard to future experience.
The target long-term regular bonus rate would almost certainly be less than the total bonus
earning capacity of the business, in order to allow for a significant proportion of the total profit to
be ultimately paid out as terminal bonus (see below).
The bonus declared this year must also be affordable in the short term (in other words, sufficient
profits must have been earned during previous years to pay for the bonus).

The company will also look at its current bonus rate, and if this differs from the long-term
sustainable rate, may consider changing it to be nearer to the long-term rate. The progression of
the company’s regular bonus rates over the last few years would also be taken into account when
considering the rate to declare for the current year.
3.3 Special reversionary bonuses

A company may declare part or all of a reversionary bonus as a one-off ‘special’, ie in
addition to any regular reversionary bonus that it is giving. This may, for example, occur as
the result of restructuring a with-profits fund.
The intention of a special reversionary bonus is to grant a one-off increase in benefits without
creating an expectation that a similar increase will be granted in the future.
Although future rates of regular reversionary bonuses are not usually guaranteed, there is an
expectation that the insurance company will continue to pay them and that rates will not change
too much, too quickly, as we described above. A special reversionary bonus might therefore be
suitable for distributing any one-off profits, or in other special circumstances that are unlikely to be
repeated, eg as may arise from a demutualisation or takeover.
3.4 Terminal bonuses

The great advantage of terminal bonus is that it defers the distribution of surplus until the end of
the contract, and so slows down the build-up of guarantees under the contract. It is particularly
useful for distributing profits that come from volatile sources, such as capital gains on equity
shares.
The amount of a terminal bonus is determined when the insured event occurs. In theory
this implies a potentially constantly changing bonus. In practice this does not happen, but
even so a company will not guarantee to maintain the bonus at any particular level.
Companies might normally set terminal bonus rates once a year but decrease them if, for
example, there were a market crash during the year.
The aim of with-profits business is to distribute profits to the policyholders, allowing for any
shareholders’ interests and other aspects such as cost of capital. So we want to give a payout
equal to the asset share. To do this we set the terminal bonus to be equal to the difference
between the asset share and benefits guaranteed to date at the policy maturity date (initial
guaranteed plus reversionary bonuses). Insurance companies track the asset shares of sample
policies (to represent, for example, different types and terms of policy) in order to determine this
affordable terminal bonus.
To a greater or lesser extent companies will then ‘smooth’ the payouts, to reduce the impact on
the policyholder of fluctuations in investment markets. This means that the terminal bonus for a
particular policy may be more or less than that needed to bring the payout exactly up to the asset
share.
The company may also grant some terminal bonus on surrender or death.

The terminal bonus to give to a particular contract may be specified in a number of different
ways, for example:
 A percentage of total attaching reversionary bonuses, including any special

reversionary bonuses. The percentage may vary depending on duration in force and
original term of contract.
 A percentage of the total claim amount before addition of terminal bonus. The
percentage will vary according to duration in force.
Question
Explain why the Core Reading says that the percentage would possibly vary by duration in force
for the first option, but implies that this would definitely be so for the second option.
Solution
One might think of the reversionary bonus as distributing part of the investment surplus, with a
part held back to give scope for terminal bonus. Therefore, given a reasonably smooth,
non-volatile investment return, the undistributed surplus should be approximately the same
proportion of the distributed surplus at all durations. Therefore a terminal bonus that is
expressed as a fixed percentage of the reversionary bonus might bring the payout up reasonably
close to the asset shares of policies of varying terms.
However, there is no similar argument to justify the undistributed surplus being a fixed proportion
of {distributed surplus + initial guaranteed sum assured}.
If investment returns are volatile then it may be necessary to make even a terminal bonus
expressed as a percentage of reversionary bonus variable by term, if the company wishes to avoid
considerable cross-subsidy between policies of different terms.
The form of expression of terminal bonus is not especially important. The same actual amount of
terminal bonus for a given term can be achieved with either of the above approaches. (Although,
as mentioned in the solution to the above question, a percentage of reversionary bonus not
varying by duration might be rather restrictive.)
Example
Twenty-five year with-profits endowment assurance, maturing now
Guaranteed sum assured (basic benefit): £5,000
Attaching reversionary bonus: £5,950
Terminal bonus as a percentage of reversionary bonus only: 58%
Terminal bonus as a percentage of sum assured plus reversionary bonus: 31.5%
Terminal bonus in either case: £3,450
Total payout in either case: £14,400

Question
Explain why the system is not simplified so that bonus is allocated only as terminal bonus, with no
reversionary bonus.
Solution
Policyholders like reversionary bonus. With no reversionary bonus, policyholders would not feel
that they were really accruing any bonus, and their financial planning (how much will I have when
my policy matures?) will be more difficult.
Question
Explain why the system is not simplified so that bonus is allocated only as reversionary bonus,
with no terminal bonus.
Solution
By having a system which involves non-guaranteed terminal bonus, the company’s investment
freedom is greater – the company can invest the assets in respect of the non-guaranteed terminal
bonus as it wants, allowing investment in high risk / high return assets. So the eventual payout to
policyholders should be greater. In addition, terminal bonus allows smoothing which would be
impossible to do at a significant level with only reversionary bonuses.
3.5 Accumulating with-profits

An accumulating with-profits contract is a with-profits policy to which bonuses are added
annually in relation to the premiums payable to date plus previously declared bonuses. A
terminal bonus may be added when the policy becomes a claim on maturity, death or
surrender.
This differs from conventional with-profits contracts, which we have been describing up to now,
in the following way.
When a policyholder pays premiums into a conventional with-profits contract there is no

immediately obvious relationship between the stated benefit (a distant sum assured plus
attaching bonuses) and any present value of the policy. This is particularly true of regular
premium policies, where any present value of the policy would have to deduct the value of future
premiums from the value of benefits.
An accumulating with-profits contract, on the other hand, looks more like a bank deposit account.
The individual’s with-profits account starts at zero and is increased (broadly) by the amount of the
premiums paid and by bonuses (which are expressed as percentages of the value of the account).
We can crudely think of these bonuses as ‘interest’ payments on our account, although in reality
the sources of profit giving rise to them may include such things as mortality and expense profits,
as well as genuine investment returns.

This means that the values actually seen by the policyholder are in present value terms, and the
policy can be described as having a readily identifiable current benefit. However, as we shall see
below, the policyholder may not be able to obtain the full value of the account if the policy is
surrendered.
As we shall also describe below, this structure may be presented in various ways.
The most common form of such contracts is what is described, for marketing purposes, as
a ‘unitised with-profits’ contract.
However, it is not necessary for the contract to be unitised.
A non-unitised accumulating with-profits contract can look and operate very much like a
conventional with-profits contract with recurring single premiums. An essential difference
is that (for accumulating with-profits contracts) there is an explicit relationship between
each single premium paid and the addition to the benefit to which it gives rise.
In the case of conventional single premium contracts, the premium rate provides the implicit
relationship between each single premium paid and the benefit granted.
The explicit relationship between premium paid and increase in benefits described in the Core
Reading above is essentially an allocation rate. Although this is a term normally associated with
unitised business, the question of unitised versus non-unitised accumulating with-profits is mostly
just one of presentation.
The following shows the recursive relationship between policy benefit values:
Ft 1   Ft  Pt  ct  (1  bt )
where:
Ft  policy benefit (excluding terminal bonus) at policy time t; F0  0
Pt  premium paid at policy time t
ct  charge deducted at time t
bt  regular bonus added between times t and t  1 .
If the charge is defined as an allocation rate, then ct would just be a constant proportion of each
year’s premium. On the other hand, the charge could be in monetary form, such as a fixed policy
fee of, say, £3 per month on a monthly premium contract. Either charging structure could be
found on both unitised or non-unitised contracts.
Another difference is that the guarantees under conventional with-profits are likely to be
greater than many, but not all, guarantees provided by accumulating with-profits products.
This reflects the fact that bonuses are declared on the ‘current benefit’ rather than on a sum
assured. However, there may be other factors that can influence the relative extent of guarantee
under the two types of product. For example in the UK, companies have typically held back less
surplus for terminal bonus under accumulating with-profits policies than under conventional with-
profits policies, thereby increasing the guarantees.

The contract may specify a guaranteed minimum rate of accumulation, although this is now
rare for new contracts.
This guarantee may be a positive number (as assumed in the above Core Reading), or may just be
a guarantee that the value of the account will not fall. Even such a 0% guarantee is worthwhile,
compared for example with the situation on a unit-linked policy. The guaranteed rate can be
thought of as producing a maturity payment analogous to the initial guaranteed sum assured on a
conventional with-profits policy.
Where the contract is unitised, it looks and operates very much like a unit-linked contract.
The most important difference relates to the way in which the company determines the price
of the units and hence the benefit payable on the happening of the insured event.
There are two basic ways in which the unit part of the contract could operate:
 The price of a unit remains constant.
The company allocates additional units to each contract, usually annually at the
bonus declaration. These are made up of a guaranteed addition, which could be
zero, and a ‘bonus’ addition which could also be zero, especially if there are
guaranteed additions. The number of bonus units is determined at the discretion of
the company.
With this approach the price of units remains fixed, for example, at £1.00.
 The price of a unit changes.
The company, instead of allocating additional units, changes the price of a unit,
usually on a daily basis. The increase is made up of a guaranteed part (possibly
zero) and a bonus part.
The difference between these two approaches is one of presentation (the first option ‘looking’
like a with-profits contract and the second approach ‘looking’ like a unit-linked contract). The key
thing that distinguishes both of these approaches from unit-linked business is the discretion that
the company has over the bonuses granted, in whatever form. The direct link to a specified pool
of assets that we see in unit-linked business does not exist in unitised with-profits business.
Instead, the company will take a longer-term view of investment returns and smooth the bonuses
it allocates. The bonuses will of course have an indirect link to a pool of assets: the insurance
company’s with-profits fund or some subset of it.
Under both methods, these bonus additions are akin to the regular bonus given to a
conventional with-profits contract. When the insured event happens the company may add
a terminal bonus to the bid price of the units.
Thus the benefit paid to the policyholder does not depend solely on the bid value of units.
Terminal bonus might be added at maturity, death or surrender. The contracts may also offer life
cover payable by explicit risk benefit charging (ie unit cancellation) in the same way that
unit-linked contracts do. In this case the benefit on death could far exceed the bid value of units.
There is a second important difference between the unitised version of the contracts and
unit-linked contracts, which relates to the benefit payable on surrender. With unit-linked
contracts, the company has no discretion as to the amount payable on surrender. It will be
the bid price of the allocated units less any surrender penalty specified in the contract.

For unitised accumulating with-profits contracts, the surrender value will usually be defined
in a similar way, but the company may also retain the right to apply a market value
reduction (also known as ‘MVR’). The size of the adjustment is determined at the discretion
of the company.
This adjustment is also sometimes referred to as a ‘market value adjustment’ (MVA).
Companies might also retain the right to apply an MVR to switches out of an accumulating with-
profits fund into a unit-linked fund.
Question
Explain why an MVR may be needed on a unitised with-profits contract but not on a unit-linked
one.
Solution
Essentially an MVR may be needed because the investment return (and therefore the value of
units) is smoothed on a unitised with-profits contract rather than linked directly to the value of
assets. This means that the unit value of a unitised with-profits policy, even with a surrender
penalty deducted, may at times be above the value of the underlying asset share.
In these circumstances the insurance company would be vulnerable to surrenders and switches,
and would be at risk of investment selection by policyholders.
Such an adjustment may also apply in the case of a non-unitised accumulating with-profits
contract, depending upon what the policy conditions say regarding any benefit that might
be available on early termination of a contract.
Given that the difference between unitised and non-unitised accumulating with-profits contracts
is essentially just presentation, an MVR is equally likely to be needed, all other things being equal,
on non-unitised accumulating with-profits contracts.
On death, the benefit will depend on the type of contract and could, for example, be a
guaranteed sum assured, a return of premiums, or a return of the fund value.
A guaranteed sum assured would often operate as a minimum monetary amount, such that the
higher of this amount and the unit fund (including terminal bonus if applicable) would be payable
on death.
Premiums may be payable as a single lump sum, recurring lump sums, or as regular
monthly or annual amounts.

The charging structures may be as for unit-linked business, ie any combination of:
 policy charge
 percentage allocation during an initial period
 a different percentage allocation after the initial period
 bid-offer spread
 charge for risk benefits
 annual management charge.
The policy charge may instead be called a ‘policy fee’. It may be taken from either the premium
or the fund.
The recursive relation we showed earlier would be made more complicated by the introduction of
different types of charges, but its essential structure would not alter.
Alternatively, the charges could be taken implicitly through the bonus rate, with no explicit
charging structure.
This alternative is less common, at least in the UK where there has been an increasing demand for
transparency within with-profits business.
Question
A unitised with-profits contract has a regular annual premium of £8,000, and a guaranteed death
benefit of £200,000. 102% of the premium is allocated to units.
A particular policy is subject to the following additional charges in a particular year:

 a bid-offer spread of 4%
 a fixed policy fee of £60, deducted from the fund at the start of the policy year
 a fund management charge of 0.5%, deducted at the end of the policy year
 a charge for risk benefits, deducted at the start of the policy year, equal to the sum at risk
at the start of the year (after the payment of any premium and deduction of all other
relevant charges) multiplied by 0.004.
The unit fund value is £34,500 at the start of the year, immediately before the payment of the
next annual premium.
Calculate the unit value of the policy at the end of the year, assuming that a regular bonus of 5%
is added over the year and ignoring terminal bonus throughout.

Solution
So that we can work out the risk charge, we first need to work out the unit fund value at the start
of the year, after the premium has been allocated and other charges have been deducted. This is:
34,500  8,000  1.02  0.96  60  42,273.60
The risk charge is then:
200,000  42,273.60   0.004  £630.91

Using the recursive relation, we then have the fund value at the end of the year to be:
F   42,273.60  630.91  1.05  0.995  £43,506.20
Question
(i) Calculate the amount of benefit that would be paid if the policyholder in the previous
question were to die at the end of the policy year, assuming that terminal bonuses of 15%
of fund values are currently payable on death.
(ii) Calculate the surrender value that would be payable if the policyholder in the previous
question were to surrender the policy at the end of the year, assuming that:
 the policy value is subject to an MVR of 3%
 there is a separate monetary surrender penalty of £750.
Solution
(i) Death benefit
Ignoring the guaranteed minimum value, the death benefit would be:
43,506.20  1.15  £50,032.13
This is lower than the guaranteed minimum value of £200,000 and so £200,000 would be paid.
(ii) Surrender benefit
The surrender value would be:
43,506.20  0.97  750  £41,451.01
Special bonuses under accumulating with-profits policies

There is no reason, in theory, why the company could not distribute any one-off source of profit
by means of a special one-off immediate bonus. The rationale behind such a bonus would be
identical to that which we described for conventional contracts in Section 3.3 above.

4 Bonus distribution considerations

The Core Reading in this section is generally applicable to any of the distribution methods covered
in the course (ie also including those described in Chapter 7). However, the notes below refer
specifically to the additions to benefits method of distribution (as covered in this chapter).
Any proposed bonus distribution should be consistent with policyholders’ reasonable

expectations, be equitable between different categories and generations of policyholders
and not threaten the future business plans, investment strategy or solvency of the
company.
‘Policyholders’ reasonable expectations’ are described in the next chapter.
Broadly, a bonus distribution is ‘equitable’ if it is fair. In this regard we need to consider to what
extent the profit distributed to different categories and generations of policyholders actually
reflects the fair shares of the profits their policies have earned for the company.
Possible categories include policies with the same policy type, term, policy size, sex and territory.
However, it is not uncommon for there to be some deliberate inequity in some respects: for
example, separate asset shares might not be calculated for males and females, despite their
different mortality experience, because declaring different bonuses for the two sexes could be
‘politically’ unacceptable.
Equity over different generations relates to the extent to which payouts might be smoothed over
time, relative to asset shares, as described in Chapter 5.
How can bonus distribution affect the future business plans, investment strategy and solvency of
the company?
When we declare a reversionary bonus, we increase the guarantees under all the policies
involved, and in turn this increases the insurer’s risk of insolvency. So, for a life insurance
company writing a significant amount of with-profits business, the bonuses declared will have a
large impact on the solvency of the company.
Should the guarantees become particularly onerous (eg reversionary bonus rates too high for too
long), the company’s free assets could become restricted and so reduce the capital available for
other ventures – such as for providing the finance needed for selling high volumes of profitable
without-profits business.
The amount of profit deferral achieved will affect the company’s investment freedom, as we
mentioned at the start of this chapter.
So in deciding on a bonus declaration, we must be sure that the cost of the profit distribution will
not interfere with either the investment strategy, by reducing investment freedom, or the future
business plans, by reducing our ability to write new business.
Question
Explain how the bonus declaration can be modified to assist future investment strategy and
business plans.

Solution
By declaring less surplus now as reversionary bonus, and so keeping more back for eventual
terminal bonus.

5 Asset shares and bonus distribution

Over time, the asset share will represent the amount that the insurance company has
available in respect of with-profits policies. Also, to the extent that the calculation of asset
share can reflect different blocks of policies issued with similar terms and conditions, it can
be used as a tool for determining what might be reasonable and equitable to pay out to the
different generations and classes of with-profits policyholders over time. In practice, asset
shares are thus used as a guide in determining rates of terminal bonus.
We’ve already seen in Section 3.4 how the asset share is used for determining terminal bonus
rates.
However, asset shares are just one tool in the list of considerations just described.
Question
Explain why asset shares are used for determining payouts only for with-profits products.
Solution
We have seen that asset share is of great use in helping decide on bonus payments for
with-profits contracts, where the bonus system is of the additions to benefits type.
We could calculate asset shares for unit-linked and conventional without-profits contracts.
However, the benefits for these are determined by other things: for unit-linked products, by the
size of the unit account in conjunction with any guarantees, and for conventional without-profits
contracts by the guaranteed sum assured (or however else the benefit is defined). So, for these
contracts we do not need to calculate asset shares.
The only use of the asset share for such contracts would be to retrospectively monitor
profitability once a contract has reached maturity – comparing asset share against the benefit
paid out at maturity will tell us if the contract has given us a profit or a loss – and to use as a limit
on the calculation of surrender values, as we mentioned earlier.


Chapter 6 Summary
Surplus (profits)
Surpluses are expected to arise under with-profits business. This is because the premiums
are based on conservative assumptions, so the company only has to achieve very modest
experience (investment, mortality, expenses, etc) in order to meet the guaranteed benefits.
In return for these (high) premiums, policyholders expect a share of the profits earned.
The surplus can be distributed by adding to the benefits available on claim, by an immediate
cash refund, or by reducing future premiums payable.
The key decisions for the company are:

 how much surplus it can afford to distribute
 if it has shareholders, how it will split surplus between shareholders and
policyholders (although this might be controlled by legislation, the company’s own
rules or by established practice)
 how it will divide surplus between different groups of policyholders.
Profit deferral
For the insurer it is desirable to defer the distribution of surplus. This reduces insolvency
risk, and also increases investment freedom and potential returns. Most deferral is possible
when a terminal distribution is paid.
Policyholders may not like the uncertainty involved with terminal distributions, so deferral
can be unpopular and lead to marketing risk.
Meeting customer needs

Additions to benefits are appropriate for a regular premium savings vehicle with a lump sum
benefit. Cash bonuses may be better for a policyholder who is asset rich and income
deficient, and for whom a single premium structure would be suitable. Regular premium
policies with cash bonuses may be a better structure for meeting guaranteed monetary
liabilities (eg loan repayments) at minimum cost.

Additions to benefits – conventional with-profits

Bonuses can be either reversionary or terminal. Reversionary bonuses are declared (added)
to the benefit during the term of the policy, and once added become guaranteed (cannot be
removed).
A terminal bonus may be paid at the claim date but its amount is not guaranteed.
Regular reversionary bonuses are declared throughout the term in the following ways:
 Simple: as a percentage of basic benefit only
 Compound: as a percentage of basic benefit plus attaching bonus
 Super-compound: different percentages of basic benefit and attaching bonus.
Super-compound bonus defers the distribution of surplus more than compound bonus,
which in turn achieves greater deferral than simple bonus.
Special reversionary bonuses can be presented as ‘one-off’ so as to limit expectations that

they will be repeated.
Terminal bonus may be expressed as:

 a percentage of attaching bonus only, in which case the percentage may vary by
duration in force and original term of contract, or
 a percentage of basic benefit and attaching bonus, with the percentage varying by
duration in force.
Terminal bonus defers the distribution of surplus, which helps to make this kind of with-
profits business less risky for the insurer.
Both reversionary and terminal bonuses are smoothed over time, the reversionary bonuses
especially so. This provides with-profits policyholders with considerable protection from
short-term fluctuations in returns, which they would otherwise experience if they invested
directly in the investment markets.
An alternative to conventional with-profits is accumulating (eg unitised) with-profits.
Additions to benefits – accumulating with-profits

Accumulating with-profits contracts may come in unitised or non-unitised form, and in either
case may have an explicit charging structure like that of a unit-linked contract.
They operate, superficially, rather like a deposit account, with premiums being allocated to a
policyholder’s account, which is then increased at a guaranteed rate of accumulation (which
may be zero) and by regular discretionary bonuses. A terminal bonus may be added at
maturity, death or surrender. A market value reduction (MVR) may be applied on surrender.
On unitised with-profits contracts the regular bonus may be given as extra units, with the
price of a unit remaining fixed, or as an increase in the price of the units.

Bonus distribution considerations

Bonus distributions should:
 be in accord with policyholders’ reasonable expectations
 satisfy the requirements for equity between different groups of policyholders,
including the different generations
 not interfere with the company’s new business plans, investment strategy or
solvency.
Use of the asset share

The asset share is particularly important for helping to determine terminal bonus rates and in
assessing the equity of any proposed bonus distribution.



6.1 A life insurance company has issued conventional with-profits endowment and whole life
assurance policies for many years, and uses the additions to benefits method of surplus
Exam style
distribution. The company has recently had difficulties with this product, following several years
of poor investment performance in the equity and property markets.
The insurer therefore proposes to close its existing with-profits product to new business and to
sell accumulating with-profits products in the future, which it has not sold before.
(i) Explain the difficulties the insurance company might have been having with the current
type of with-profits business. [5]
(ii) Discuss whether the proposed course of action will help the insurer resolve these
problems. [8]
[Total 13]



Chapter 6 Solutions
6.1 This is a tough question at this stage of the course as it refers to some ideas that will be covered in
greater depth later. Have another go at this question as part of your revision.
(i) Current difficulties
Conventional with-profits will have its terminal bonus liabilities, in particular, backed by equity
and properties. [½]
Falling equity and property values would then have led to falling asset share values. [½]
The company would then have to reduce terminal bonuses in order to reduce losses. [½]
Policyholders’ reasonable expectations (PRE) of benefit smoothing may have delayed the changes
in bonus rates so incurring losses on claims in the meantime. [1]
Competitive and marketing pressures may have prevented the insurer from reducing bonuses by
as much or as soon as it should have done, so increasing losses. [½]
The company may be finding it very difficult to maintain its policyholders’ confidence in their
policies with falling bonus rates, which may be increasing withdrawal rates. [½]
Sales of new business may be falling significantly for a contract where returns appear to be falling
over the foreseeable future. Brokers, in particular, may stop recommending this product for their
clients. [1]
Alternatively, the company may still be paying excessively high bonus rates in order to meet these
marketing pressures, and the losses incurred may ultimately threaten solvency and so be
untenable in the long term. [½]
These problems may have extended to reversionary bonuses, if the current rates are no longer
sustainable in current investment conditions. [½]
The insurer may find it even harder to reduce reversionary bonuses in the light of PRE, especially
if these bonuses have never been reduced before. [½]
A lack of transparency in the way in which the bonuses are declared may also have led to a
general dissatisfaction with the product amongst policyholders. [½]
[Maximum 5]
(ii) Proposed course of action
Advantages
Selling a new tranche of with-profits business can enable the company to ring-fence the existing
losses, so that new business will not have to adhere to the same bonus rates. [½]
Having completely different benefit and bonus structures should make it easier for the insurer to
pay more realistic bonuses on the new contracts. [½]

If part of the problem is to do with mis-managed PRE, then the new structure gives the company a
fresh opportunity to get PRE properly set from the outset, for example during the sales process,
and so reduce the risk of similar problems happening in the future. [1]
The accumulating with-profits (AWP) structure is more transparent to policyholders. This will
make it easier for policyholders to appreciate when bonus rates are high relative to current
market interest rates (for example), so possibly making a future reduction in bonus rates (if it
follows market movements) more acceptable. [1]
Policyholders may therefore come to accept more variability in bonus rates under the new
contract, in the same way as they can generally accept variable interest rates in such things as
deposit savings accounts. [½]
By severing the link with the existing business, the company may be able to reduce bonus rates
(and hence losses) under its existing business without prejudicing its new business prospects by as
much. [½]
Disadvantages
The company will not have experience in managing the new type of structure and may make
mistakes. [½]
For example, it may start declaring rates of bonus that are too high and create similar difficulties
as before by causing unsustainable expectations of future rates. [½]
If less smoothing is anticipated, then the company will need to make sure it passes more of its
good returns to policyholders in good years, as well as poor returns, when applicable. [½]
It will probably find it harder to defer surplus in the same way as under conventional policies,
which will increase its investment risk or require a different investment policy, for example one
involving less investment in real assets. [½]
The reduced terminal bonuses that result may then disappoint policyholders. [½]
There are likely to be admin and IT issues (eg if a unitised structure is being used) that could cause
errors and poor service, again causing customer dissatisfaction. [½]
Policyholders may find the new design hard to understand. [½]
The new design may turn out to be less attractive to policyholders than the old product. [½]
The new product may introduce new marketing and PRE issues, eg on surrender when a ‘natural’
expectation might be that the full fund value would be payable. In fact, the company may have to
impose a market value reduction on surrender, so as to prevent surrenders from making losses,
which will need careful explaining to avoid disappointment. [1]
The company may find it hard to create different PRE for this product despite the new structure,
and the legacy of the old product may persist to prevent the change from achieving the possible
benefits for the company described above. [1]
[Maximum 8]

With-profits surplus
distribution (2)
Syllabus objectives
3.5 Propose further ways of managing the risks in Syllabus Objective 3.1, including:
 choice of with-profits bonus method.
(Only this part of Syllabus Objective 3.5 is covered in this chapter.)

0 Introduction
In this chapter we shall study two further approaches to the distribution of surplus, which are
used in countries other than the UK.
When studying these methods, it is worth bearing in mind that, compared with the UK, life
insurers in most countries have held relatively low proportions of their investments in equities.
Most policyholder liabilities have been backed with fixed-interest stocks, giving a relatively stable
investment surplus.
We also consider some of the various ways in which a company’s bonus distribution can be used
as part of its overall financial management.

1 Revalorisation method
This method is used in parts of continental Europe, eg Italy.
The profit, or surplus, to be given to a particular contract is expressed as a percentage, r%

say, of that contract’s supervisory reserve. The benefit under the contract and the premium
payable by the policyholder are then increased by the same amount.
So, bonus is distributed to policyholders by means of increasing their reserves by r%. In fact this
increase in reserves is the aim of the method. The increase is effected by increasing the sum
assured by r% and the future premiums payable by r%.
Question
Explain why premiums must increase by r%.
Solution
Reserve V  S A  P 
a
Hence
1  r V  1  r  S A  1  r  P a
In other words it is necessary for premiums to increase by the same proportion, otherwise the
company would not be able to afford the increase to the sum assured, ie the right hand side of
the above equation (which is the expected present value of the policy liabilities) would be greater
than the left hand side (which is the amount of money available to cover them).
Question
A policyholder has made the following comment:
‘My premium has gone up by 2%, so of course my sum assured should go up by 2% too. I don’t
see how this is a bonus.’
Outline how the company should respond.
Solution
If the policyholder had paid 2% more from outset then the sum assured would have been 2%
higher. So one way of thinking about the bonus is that the policyholder gets 2% more in benefits
without having paid 2% more on his or her past premiums.

In fact, most countries operating this bonus system offer the option of constant premium policies.
Many policyholders will feel uncomfortable with the idea of a long-term commitment to unknown
future premium increases. Obviously if less money is put in then less can be expected in benefits,
all other things being equal. Sums assured will therefore increase more slowly than on policies
with ‘revaluable’ premiums.
Question
Explain how the revalorisation method would work for a single premium product.
Solution
Nothing subtle: the sum assured increases by r%.
Where this method is used, it is not unusual for the profit of the life insurance company to
be divided into a ‘savings’ profit and an ‘insurance’ profit.
The ‘savings’ profit represents the profit from the assets and can be distributed, in whole or
in part, by the ‘revalorisation’ method.
The profit from the assets is essentially the excess of the actual investment return over the
expected investment return multiplied by the assets held. The method uses the end-year reserve
as a close approximation for the amount of assets held during the year, and r reflects the excess
of actual investment return over expected, ie:
r  i  i
where:
i   actual interest rate on the assets
i  expected interest rate on the assets
The distributed savings profit would then be typically of the form:
distributed profit  r .t 1 V
As the Core Reading says, it may be only part of the savings profit that is distributed to each
policyholder. So, companies typically use formulae such as r  k  i   i  or r   k.i   i  , where k
might be in the range 0.75 to 0.95. The proportion k is normally specified in the policy conditions
as a minimum that will be declared to policyholders every year. The value of r would be subject
to a minimum of zero (as negative bonuses cannot be declared when the actual return is less than
expected).
The calculation of i´ will vary from country to country. Typically, it will be derived from
investment income, realised gains (or losses), and some allowance in respect of unrealised gains
(or losses).
The ‘expected’ interest rate is the valuation interest rate. In countries employing this method this
is usually the same as the premium basis interest rate.

The bonus percentage k offered under a product will be a significant marketing aspect of a
product. Clearly this will only be true in a country where companies have some freedom to set
their own bonus allocation rates, rather than being constrained by regulation.
The ‘insurance’ profit is that arising from actual experience being better than expected for
all sources of profit other than the return on the assets. This profit might then typically be
retained by the company for distribution to shareholders, as reward for the pure insurance
risks that they have borne.
Although the insurance profit is typically retained for shareholders, it is sometimes the case that,
like the savings profit, it is divided between policyholders and shareholders.
To make the above ideas a bit more concrete we examine two case studies, Italy and France.
Example 1
In Italy the system normally works as follows.
For every calendar year, a yield i' is calculated on the relevant policyholders’ fund. This yield is
determined by reference to investment income and realised capital gains, as a percentage of the
fund’s average (adjusted) book value over the year.
Policies will have their reserves increased by  k.i   i  , where i is the valuation investment return
assumption. The bonus allocation k is normally 80, 85 or 90% depending on the company. The
increase in reserves is subject to a minimum of zero, as bonuses cannot be negative.
The bonus (ie the increase in reserves) is distributed at the policy anniversary, at the end of the
policy year to which that bonus relates. If a policy surrenders or claims before then, some pro
rata adjustment is made.
The increase in reserves is made by increasing the guaranteed sum assured. ‘Revaluable’
premium policies will see the full  k.i   i  increase to their sum assured, and their premiums will
increase by the same percentage amount. Constant premium policies will see a smaller increase
to their sum assured, but their premiums will be unaltered.
Thus the bonus is calculated purely from ‘savings’ profit, and no adjustment is made in respect of
‘insurance’ profit. Shareholders therefore automatically take all insurance profit, and a share of
the savings profit.

Example 2
In France the system is slightly different. The amount to be distributed is calculated as the sum of
part of the savings profit and part of the insurance profit.
Savings profit is based on the investment income and the realised capital gains accrued to the
policyholders’ funds.
Insurance profit is the sum of mortality, expense and surrender profits.
The life insurance company must distribute 85% of the savings profit and 90% of the insurance
profit. However if the insurance result is a loss then it can be offset against the savings profit.
The bonus is then distributed in the normal revalorisation way via an appropriate increase to
guaranteed sums assured (and possibly premiums).
We conclude this section by considering some of the advantages and disadvantages of the
revalorisation method.
The major advantages are:

 It is simple to apply.
 The method codifies exactly how a company should declare part of its profit as bonus to
with-profits policyholders. Very little judgement is normally required, and should
therefore be relatively cheap to administer.
 Having a codified method generally protects policyholders from ungenerous life insurance
companies.
 By taking assets at book values, including appropriately smoothed writing-up (or writing-
down) adjustments, a smooth emergence of investment profit is usually achieved.
The principal disadvantages are:

 The company has no discretion in its profit distribution (except to the extent of spreading
of one-off costs, if insurance profit is distributed to policyholders).
 The method tends to discourage equity investment. This is because of the fact that there
is no deferral of profit distribution. This means that all investment losses would be borne
by the company and would constitute an unacceptable insolvency risk. There would also
be a problem regarding the treatment of unrealised gains, which are not easy to distribute
directly under current revalorisation systems.
 Versions that do not share insurance profit with policyholders go against the principle of
mutuality. Example 2 (France) is obviously an exception to this.
 It is not very easy to explain to policyholders with ‘constant premium’ policies who see
very small additions to their guaranteed benefits early in the policy term.

Question
Explain the differences, if any, in the way the revalorisation method works for a mutual life
insurance company.
Solution
Insurance profit would need to be added to the savings profit and expressed as a percentage of
reserves, and then passed on to policyholders by increasing premiums and benefits in the same
proportion, as usual. (This is very like the French model described in Example 2 above.)
The proportion of total profit distributed to policyholders would probably be much closer to
100%, because there would be no shareholders with whom to share the profit.
Distributing the insurance profit in proportion to policy reserves is unlikely to be a very fair method
of allocating this source of surplus between different groups of policyholders.

2 Contribution method
This method evolved in the United States.
The ‘contribution principle’, which underlies this approach, is that distributable surplus should be
distributed among policies in the same proportion as those policies are judged to have
contributed to surplus.
The dividend given to a particular contract is calculated using the following formula:
dividend  V0  P   i   i    q  q    S  V1    E  1  i   E   1  i   
where V0  value of contract at beginning of year on valuation basis
V1  value of contract at end of year on valuation basis
P  gross premium
i   actual rate of interest earned
i  valuation basis rate of interest

q   actual rate of mortality experienced
q  valuation basis rate of mortality
S  sum assured
E   actual expenses experienced under the contract
E  valuation basis expenses
The gross premium in the above formula is the actual premium paid by the policyholder.
Question
Explain in words the three components of the formula.
Solution
Excess interest on the {reserve plus premium}, ie interest surplus. (The premium is assumed
payable annually in advance in this formula.)
The expected death strain less the actual death strain, ie mortality surplus.
Excess of expected expenses over actual expenses, accumulated to the end of the year,
ie expense surplus.
(For simplicity the formula has assumed that everything happens either at the start or the end of
the year. Now that the concept is clear, it should be possible to adjust the formula for different
assumptions about incidence of premiums, expenses and claims.)

The amount of distributable surplus will be determined by the company, and the dividend
formula will be applied with respect to this surplus. The proportion of the total surplus which is
distributed will vary from year to year, in order to obtain a more stable (ie smoothed) progression
of dividends over time. Some profit may also be retained for shareholders, held against future
adverse contingencies and/or held for future distribution as a terminal dividend.
A key point is that the method results in an extremely equitable distribution of profit, by relating
the distribution to each policy’s contribution to the total profit.
There are several different possible but mathematically equivalent ways of expressing the
dividend formula, so do not be put off if the formula is expressed slightly differently elsewhere
(eg in the exam). The important thing is to remember that the purpose of the formula is to divide
the total amount of distributable surplus between the policyholders in as fair a way as possible, in
accordance with the policy’s contribution to the company’s investment, mortality and loading
surplus. Dividends would probably need to be adjusted in practice (eg to allow for miscellaneous
sources of surplus) in order for the total dividends paid to equal the distributable surplus
required.
Unlike in the UK, it is quite common for the dividend to be paid out as a cash sum, or used to
reduce future premiums, but this is not always the case.
The dividend is often converted into a ‘paid-up’ addition to the benefit, instead of being paid
out in cash.
When the dividend is converted into a paid-up addition to the benefit, it makes the distribution
(although not its determination) rather like the UK reversionary bonus system. The conversion to
paid-up benefit would be calculated on the valuation assumptions (otherwise an additional profit
or loss would arise through the difference between the conversion basis and the valuation basis).
In practice, the actual mortality, expense and interest rates used in the dividend formula will be
based on the experience of (relatively homogeneous) groups of policyholders.
Question
Suggest one theoretical and one practical reason why policies will be grouped.
Solution
If we treated every policy individually we should lose the concept of pooling of risk. (For example,
policyholders who survived the year would show a mortality profit, while those who died would
show a huge loss. Hence we use instead the idea of the actual experience being the average
experience of all policies in a homogeneous group.)
Practical limitations on the amount of calculation we wish to do would restrict the number of
groups we should wish to have.
We also need to ensure that there are sufficient data in each group to make the experience
credible (ie that the average experience is not too distorted by random sampling error).

The issue of how contracts should be grouped for this purpose is one on which actuaries in the US
have had considerable debate. No definitive answer can be given.
On the whole, the approach leads to much more of an attempt to give surplus back to specific
groups of policyholders than is the case in the UK.
For example, in both the US and the UK, allowance for differences in mortality at different ages is
made in the with-profits premium rates. In the UK that is pretty much the end of the story. It
would be unlikely that different bonuses would be paid to different age groups, even if, say,
experience for older lives turned out significantly better than expected while that for younger
lives was worse than expected. It would be treated as part of the pooling of risk.
However, in this instance it would be quite likely in the US that a positive mortality contribution
would be attributed to older lives and a negative one to younger lives.
Other approaches exist for determining the amount of the dividend, but these are beyond
the scope of Subject SP2.
A terminal dividend may be given, which effectively returns to the policyholder at maturity
part of the difference between the true earned asset share at that point and the sum
assured.
In principle, this terminal dividend may be substantial.
When the contract goes off the books, the contingency provisions that have been held back on
the contract are no longer required. It is widely accepted that at least a part of this money should
be released to the policyholder. The company may itself retain part, on the grounds that the
policyholder has benefited from the protection and support of capital built up from earlier
generations of policyholder. He or she should therefore contribute to the protection of
continuing and future policyholders.
Similar arguments are sometimes used in the UK, where companies may aim to pay slightly less
than asset share on average in order to build up their free capital, or may make an annual charge
on asset shares as compensation for the use of capital.
There does tend to be less subjective actuarial judgement in applying the contribution approach
compared with that of the UK ‘additions to benefits’ method. With contribution dividends there
may be a more objective application of the ‘rules’ concerning surplus distribution. Some features
of this (for example the proportion of mortality surplus given to policyholders) may be guaranteed
in policy literature.
One should not be too dogmatic about this though, as far as general principles are concerned.
Hence, while mechanical rules may be applied in dividing the distributable surplus between
policyholders under the contribution method, the total amount of surplus to distribute each year
may be subject to some discretion and hence will require judgement. Likewise UK-style bonuses
could be determined by mechanical rules.

3 Choice of with-profits bonus method

In this and the previous chapter we have studied three different bonus methods: additions to
benefits, revalorisation and contribution method dividends.
Question
Explain how the three systems compare, in particular with regard to:
 equity and smoothing
 flexibility (ie discretion)
 simplicity
 investment freedom.

Solution
Additions to benefits Revalorisation Contribution

Equity and Deliberate smoothing of Equitable distribution of Very equitable with
smoothing asset share volatility investment profit. regard to investment,
over time. Sharing of Mortality and expenses mortality and expense
mortality and expense not usually included. surplus. Some
experience is only very Some smoothing over smoothing.
broadly equitable. time achieved by
valuation method.
Flexibility Maximum flexibility to No discretion (*) to do Considerable flexibility

(discretion) the life insurance anything other than for total distributed; total
company – how much apply the set formula. is split by fairly fixed
bonus to give and when rules, although some
(now or at end). judgement as to
groupings for application
of the formula.
Simplicity Simple to tell policy- Simple both to apply Complex to apply, and
holders what the and to present to likewise to explain (but
company is doing, more policyholders. perhaps easier than UK
difficult to explain / as not reversionary).
justify; they need to
trust the industry.
Investment Can be lots, especially if Very little, because Intermediate: some
freedom conventional, use high surplus is distributed as surplus may be held back
TB and super- it arises. for terminal dividend, but
compound RB. limited by transparency.
(*) This assumes an accounting environment that does not allow any flexibility to write up the
book value of the assets held, other than by an amount which ensures that the values of
the assets gradually blend into their nominal values by their redemption dates.
The different approaches have arisen largely by historical accident. In practice, the approach a life
insurance company uses will be forced on it by prevailing custom (and possibly legislation or
regulation) in the country concerned.
However, within a particular with-profits custom, there may remain considerable scope for an
insurer to use its bonus distribution as a means to furthering the general management aims of the
company. In this section, we consider the extent to which different bonus methods enable an
insurer to manage various risks.

3.1 Margins for future adverse experience

The existence of profit distribution to policyholders may mean that the premium rates
charged contain implicit or explicit margins to generate that profit (particularly for
conventional with-profits business). These margins for profit could equally be looked at as
margins against adverse future experience. How far they may, in practice, be viewed this
way may depend on the expectations of policyholders in this respect.
For instance, suppose mortality experience on one of a company’s with-profits contracts is much
worse than anticipated. To what extent can the company reduce bonuses accordingly without
going against policyholders’ reasonable expectations?
Question
State the risk being addressed in the Core Reading.
Solution
The risk is that of experience being worse than assumed in the pricing basis. In general the
experience could be mortality, expenses, withdrawals or investment return.
It makes a big difference to the risks of running a with-profits portfolio if the insurance company
can reduce bonuses to reflect poor experience. While it could be thought that the whole nature
of with-profits policies should allow this to happen, the extent to which it is possible can vary
considerably. For example:
 the revalorisation method usually only distributes investment profits so any expense or
mortality losses must be borne by the company. (However, as these products are usually
prudently priced, the likelihood of the company suffering large losses from this source is
quite small.)
 policyholders’ reasonable expectations of payouts may make it very difficult for
companies to reflect poor experience fully in the bonuses they declare.
 guarantees under all with-profits contracts mean that there is some level of adverse
experience beyond which any further losses cannot be recouped (ie the company cannot
declare negative bonuses).
3.2 Business objectives of the company

A life insurance company is likely to have as one of its business objectives the
maximisation of the profit distribution to policyholders to improve its competitive position.
The actuary advising the company on its distribution of profit needs to balance this with
other objectives.
From a competitive perspective, the aim of the company will be to maximise the eventual
benefits payable to policyholders. Once a policy has been written, the only way to improve the
eventual benefits to the policyholder is by maximising the bonus distribution.

Conversely, any move of the bonus rates downwards could be expected to injure the company’s
competitive position and so lead to a reduction in new business. However, there may be times
when the company should reduce bonus rates to reflect the current financial reality; and there
may be times when the company must reduce bonus rates to ensure that it remains solvent.
Question
(i) Outline the risk being controlled here.
(ii) Discuss the extent that the risk can be controlled under
(a) the conventional additions to benefits method.
(b) the accumulating with-profits method.
(c) revalorisation method.
(d) the contribution method.
Solution
(i) Risk
The risk is that a low bonus decision, while actuarially sound, has unforeseen repercussions on
new business; as a result per-policy expenses increase for current policyholders. This is a
marketing risk.
(ii)(a) Conventional additions to benefits
The actuary can smooth bonus rates from year to year. Thus, in a year in which the surplus was
very low and this was seen as an anomalous result rather than the start of a new economic trend,
the bonus rates might be kept unchanged – ie declare more than indicated by asset share
projections.
(ii)(b) Accumulating with-profits
There will be a marketing risk if the regular bonus is reduced by too much in any one year.
However, some smoothing of bonuses is very likely to be practised (and expected) under this
method, but to a lesser extent than under conventional additions to benefits. If the company is
also declaring a significant terminal bonus, then it will have more scope to smooth and hence to
manage this risk.
(ii)(c) Revalorisation method
The question does not arise, since the company must declare bonuses in accordance with the
surplus accrued.
(ii)(d) Contribution method
The situation regarding smoothing is very similar to the accumulating with-profits method.

3.3 Policyholder expectations

This section considers policyholder expectations, which we’ve referred to at various points
throughout these two chapters on surplus distribution.
Policyholders may have expectations about the form of the profit distribution and the level
of the bonuses or dividends given. Such expectations may be built up from:
 documentation issued by the life insurance company
 the company’s actual past practice
 general practice in the life insurance market.
The documentation referred to here that will influence policyholders’ attitudes is any marketing
literature shown at the time of sale, and in particular any projections implying a specific yield on
the policyholders’ fund, or eventual benefit at maturity.
The company’s past practice relates to its bonus distributions in the past few years. One
consequence of this is that, on finding itself with a larger than expected surplus, a life company
might not want to declare all of that surplus in one year because to do so could raise policyholder
expectations that that bonus rate is typical and repeatable.
A lower than expected surplus in any year will tend to be met by the opposite response: over-
payment of profit in that year to reflect expectations of smoothing. But it would be against
policyholder expectations for the over-payments to persist for any length of time, as this might
result in much larger bonus reductions later, or the company having to compromise its financial
security. (A compromise in financial security would (almost certainly) be contrary to policyholder
expectations.)
‘General practice in the life insurance market’ refers to the bonuses that other life companies are
declaring. The company should bear in mind that policyholders expect any company’s investment
performance to be roughly in line with that of the market, and so expect the company’s resultant
bonus declarations to be similarly in line with other bonus declarations seen in the market.
Failure to meet these expectations will lead to policyholder dissatisfaction and the risk of
losing existing and/or new business. It may also be deemed to be failure to treat customers
fairly, which in some countries is grounds for intervention by the insurance supervisory
authority.
Question
(i) Outline the risk being controlled here.
(ii) Discuss the extent that the risk can be controlled under
(a) the conventional additions to benefits method.
(b) the accumulating with-profits method.
(c) revalorisation method.
(d) the contribution method.

Solution
(i) Risk
The risk is that the bonus declaration will not live up to policyholder expectations, and that those
expectations are reasonable. The risk from that is that new business levels may drop, the
company may suffer adverse publicity, and the supervisory authority might intervene in the
interests of disgruntled policyholders. (Here we are talking about all kinds of distributions,
including both regular and terminal bonuses, where applicable.)
(ii)(a) Conventional additions to benefits
Under this system the company has the greatest discretion to control this risk once policyholders’
expectations have been formed, through the management of its bonus distributions.
(ii)(b) Accumulating with-profits
All future non-guaranteed profit distributions are discretionary, so in theory the position is similar
to the conventional system. However, there might be greater expectations of bonuses being
linked to underlying (and hence market) investment returns. While this would not prevent
insurers from over-declaring, it might make it difficult for the insurer to balance the equation by
under-declaring during the good times.
(ii)(c) Revalorisation method
Here the actuary cannot really influence bonus rates, but that makes it all the more important for
policyholders’ expectations not to be raised misleadingly due to rash marketing literature etc.
(iii)(d) Contribution method
Here there is certainly a marketing risk if regular dividend distributions are lower than
policyholders’ expectations. It is this that actually prevents a company from deferring profit
distributions by more than they actually do under this method. In cases where there is a
significant terminal dividend, then the company can exercise some control over this risk with
regard to the total payouts, in much the same way as under additions to benefits, though the
scope for this will almost certainly be less under the contribution method. Again, in many
respects, the issues for this method are similar to those for accumulating with-profits, although
the link to market interest rates is not so obvious (the cash dividend is just an amount of money,
whereas for accumulating with-profits it is declared as an interest rate).
3.4 Provision of capital

This section considers how the surplus distribution system can assist the company in providing
working capital. Under certain circumstances, it may be possible to defer the distribution of
surplus to policyholders, and in the meantime to use that non-distributed surplus as working
capital.

The pace at which profit arises and the pace at which it is distributed may or may not be the
same. If part of the profit is deferred to some future date before being distributed, it may
augment the company’s free assets in the meantime and increase its ability to take on risk.
Depending on the constitution of the company and the regulatory regime, surplus in respect
of with-profits business may be available to support new business. Even in jurisdictions
where reserves must be held for undistributed profit which is earmarked for future bonuses,
it is generally the case that the overall reserves would be lower than had the profit been
distributed.
We described how deferral of profit distribution occurs at the start of the previous chapter.
The above Core Reading is saying that such deferral may lead to an increase in the company’s free
assets, and so enable it to write more new business.
To see how the distribution of profit might affect the build up of reserves, let’s consider two
possible reserving regimes. Under Regime A, the required reserves increase in line with
guaranteed benefits, so declaring a lower regular bonus now (with the aim of having a higher
terminal bonus later) will lead to lower reserves and so more working capital.
In contrast under Regime B, insurance companies have to hold a reserve equal to asset share. A
reduction in the current bonus declaration won’t have a direct impact on the size of the asset
share. Having to hold the asset share effectively means that with-profits policies won’t be a
source of working capital in such companies.
Where profit is not being distributed as and when it arises, there will be years when the
amount distributed exceeds the amount generated and vice versa. However, over time it
would be expected that there should be a balance between the two. Sustained
over-distribution could lead to an excessive drain on the free assets.
There are two concepts to consider here. First, if the combination of product design and bonus
distribution system is such that profit is not distributed as it arises, how can the company manage
this mismatch over time?
Secondly, on average are bonuses paid to policyholders in line with what they have a right to
(which we can quantify with asset share methodology) or are the bonuses too great or too little?
If bonuses are too great – perhaps because bonus declaration levels have been massaged
upwards for commercial reasons – then the capital situation of the company might be badly
affected. If bonuses are too small, then policyholders are getting a bad deal. In either case, there
may be a problem with equity: one group of policyholders is getting too much, another group is
getting too little.
The extent to which it is possible to defer the distribution of profit depends on the form of
the distribution.
As mentioned in Section 1.1 of Chapter 6, deferral of profit distribution also reduces the
probability of insolvency.
Additions to benefits
For a given total amount of bonus allocated to a contract, the more that is allocated in
terminal form the greater the deferral of profit distribution.

Within the reversionary bonus element of a conventional with-profits policy,

super-compound will defer the distribution of profit more than a compound bonus, which in
turn defers more than a simple bonus.
For a conventional policy, the amount of deferral of profit distribution can be quite considerable,
if the conditions are right. For instance, in the past a twenty-year contract in the UK might have
around half of the final benefit derived from terminal bonus. The proportion of the final benefit
represented by the initially guaranteed benefit could be very low – perhaps just 15-20% of the
final benefit.
Greater deferral of profit distribution from the super-compound bonus method gives the life
insurance company greater investment freedom, which should lead to higher final benefits.
Question
Explain why the life company has more investment freedom in this instance.
Solution
A policy where more of the benefits are payable as terminal bonus has less liability that is
guaranteed in monetary terms and more liability that is currently discretionary in nature. The
greater amount of discretionary liability permits a greater degree of investment in higher-risk /
higher-expected-return assets such as equities and property. (More explanation of this is given
later in the course, in the investment chapter.)
Question
‘Any free assets created from non-declared policyholder profits really belong to those
policyholders, and so should be used to improve investment freedom to give higher eventual
returns for those policyholders, not used to write more new business. In fact policyholder equity
implies that life companies should close to new business.’
Solution
This statement is not valid.
Although at first glance any surplus created on a policyholder’s account could be thought of as
‘their’ surplus, and so to be used only for their immediate benefit (by increasing ultimate real
investment returns), we should not forget that the policyholder ‘owes’ something to the fund in
having had their new business strain financed. Just as any policy will have been initially financed
by the fund, so it should expect to be used equally to finance other policies. So that over the
policy’s lifetime, from inception to maturity, the policy has paid a cost of capital equal to that
which it would have incurred anywhere else.

Comparing the regular annual bonuses for conventional and accumulating with-profits, the
distribution of profits will tend to be deferred more for the former than the latter, but in
practice this will depend on the bonuses actually declared. The present values of the
bonuses could be very similar.
Question
Explain why the amount of profit deferral typically achieved under accumulating with-profits
contracts might be expected to be less than that achieved under conventional with-profits
policies.
Solution
This is because the regular profit distribution is expressed in the form of a bonus interest rate,
that is, in current money terms. It will be hard for the insurer to hold back too much of the
current year’s surplus, because this will make the bonus interest rate look small and lead to
marketing risk.
For conventional business, the regular bonuses are reversionary, so they are expressed in terms
of their monetary value at the time of claim – ie often many years into the future. They then give
the impression of high monetary value, while only incurring a modest current cost. There is
therefore more scope for reducing reversionary bonus rates, without upsetting policyholders, and
ultimately deferring much more of the profit.
Revalorisation method
There is less scope for deferral of profit with this method.
Contribution method dividends

Depending on the balance between regular and final dividends, this method may also lead
to significant deferral of profit distribution, although typically this will be to a lesser extent
than under the additions to benefits method for conventional with-profits contracts.
The company has some scope with this method to defer profit distributions. When the company
determines the total amount of profit it wishes to distribute in any year, it can deliberately
withhold some of its profit for later distribution by a terminal dividend. The extent to which this
can be achieved, however, will be limited by the policyholders’ expectations of the level of cash
dividend to be received in each year, and so the company cannot defer too much profit as this will
lead to policyholder dissatisfaction and a resulting marketing risk.
We can recognise a similarity here with the accumulating with-profits design, because again the
regular profit distribution is made in terms of the current cash value (see the solution to the
previous question). The extent of deferral might therefore be expected to be quite similar for the
two methods.


Chapter 7 Summary
Revalorisation
Bonuses under the revalorisation approach are granted by increasing the reserves, benefits
and premiums of with-profits contracts by a percentage, r% say.
In determining this percentage, it is common to divide surplus into a ‘savings’ profit

(ie investment surplus) and an ‘insurance’ profit (ie surplus from other sources).
A (high) proportion of the savings profit is usually given to policyholders, with the rest
retained for shareholders. Depending on the market, all the insurance profit may go to
shareholders, or it may be divided between shareholders and policyholders.
Contribution method dividends

The contribution principle is that each policy receives a share of distributable surplus in
proportion to its contribution to surplus. However, for this purpose policies are classified
into reasonably homogeneous groups.
Various methods exist for determining each policy’s contribution to surplus. One approach
uses a formula to identify the investment, mortality and expense profit from sample policies.
Dividends may be paid as a cash sum, converted to an addition to the benefits, or used to
reduce future premiums.
A terminal dividend may be given when the policy becomes a claim.
The surplus distribution system may allow the actuary some scope for tackling some of the
problems that can face a life company. The extent to which this is so will differ greatly
according to the actual surplus distribution method.
Use of bonus distribution in the company’s financial management

With-profits products will normally have been designed with significant margins to allow
surplus to accumulate under normal conditions. If experience is much worse than
anticipated, these margins may be used to compensate the company – depending on
policyholders’ expectations.
Decisions on surplus distribution should bear in mind the possible impact on new business
levels and have regard to policyholder expectations, which may be built up from:
 documentation issued by the life insurance company,
 the company’s actual past practice, and
 the current practice in the life insurance market.
If the surplus distribution method permits deferral of bonus, such deferral may increase the
company’s free assets. Any ‘extra’ free assets can, until needed for policyholder benefits, be
used to increase investment freedom and to finance new business.



7.1 A company uses the contribution method of distributing surplus, with a cash dividend.
Exam style
Outline the advantages and disadvantages for the policyholder and the company compared with
the use of reversionary and terminal bonuses. Taxation should be ignored. [5]
7.2 Describe how the dividends might be determined under the contribution method of distributing
Exam style
surplus, defining any formula that might be used. [4]
7.3 Outline the various methods available for distribution of surplus to with-profits policyholders, and
Exam style
their main relative advantages and disadvantages. [15]
7.4 (i) Describe how savings profits are distributed using the revalorisation method. [2]
Exam style
An individual aged 35, who expects to retire at 60 and who is currently in stable employment,
buys an endowment assurance under which profits are distributed using the revalorisation
method.
(ii) Explain the extent to which the product might be suitable for this individual’s needs,
assuming they are buying the contract with the specific aim of providing a replacement
income following their retirement. [8]
[Total 10]
7.5 Recently, an insurer (Company A) paid a maturity value of £20,375 to one of its with-profits
Exam style
policyholders. The policy was taken out ten years previously for a level annual premium of
£1,000. The insurance company distributes profits according to the additions to benefits method.
A sales representative of a different insurer (Company B), which distributes profits using the
revalorisation method, has claimed that the policyholder would have received a much higher
maturity value had the policy been taken out with Company B instead.
Assume that the two companies operate in the same market, sell only conventional business, and
are broadly similar in all material respects other than those which would be appropriate to the
different methods of profit distribution adopted.
(i) Describe the profit distribution strategy that each company would be likely to adopt,
including in each case an explanation of:
 the impact of this strategy on policy payouts
 the circumstances under which each method will produce better value for money
than the other, in terms of the yield that will be obtained by policyholders who
keep their policies until the maturity date. [12]
(ii) Discuss whether the policyholder might have grounds for complaint against Company A,
with regard to the size of the maturity value that has been paid. [3]
[Total 15]



Chapter 7 Solutions
7.1 Advantages to the policyholder
The bonus is received straightaway instead of when the insured event arises, which may or may
not be an advantage depending on when the policyholder wants the money. [½]
The bonus may appear fairer to the policyholder. [½]
Disadvantages to the policyholder
Reduced return likely due to limited deferral of distribution of surplus. [½]
The policyholder may not want a reduction in premium, the policyholder may want to save a fixed
amount (or fixed percentage of salary) each month. [½]
The final benefit is lower in real terms. [½]
Advantages to the company
This method might be attractive to the policyholder and so would boost sales. [½]
It also could be marketed as new and innovative (if applicable in the market concerned). [½]
It could be used by a mutual company to combat other companies which have demutualised,
since the policyholder is receiving a ‘dividend’. [½]
Disadvantages to the company
A system of cash bonuses distributed early will reduce the level of funds under management and
so may place constraints on investment policy (eg by reducing the extent of diversification
possible). [½]
Affordable cash bonuses may seem low in relation to reversionary benefits that can be granted
for the same cost. [½]
This tends to make the amount of profit deferral less for the contribution method than for
additions to benefits, reducing the company’s potential investment freedom and adversely
affecting overall returns. [1]
Administration of cash bonuses would be more difficult and more expensive than an addition to
benefits. [½]
[Maximum 5]

7.2 The dividend might be determined using a formula, as follows:
dividend = (V0 + P)(i" – i) + (q – q")(S – V1) + [E(1 + i) – E"(1 + i")]
where V0 = value of contract at beginning of year on valuation basis

V1 = value of contract at end of year on valuation basis
P = gross premium
i" = actual rate of interest earned
i = valuation basis rate of interest
q" = actual rate of mortality experienced
q = valuation basis rate of mortality
S = sum assured
E" = actual expenses experienced under the contract
E = expected expenses under the contract according to the valuation basis.
[2]
By calculating the actual experience rates for suitably homogeneous groups of policies we can
directly attribute investment, mortality and expense profits to the policies on which they have
arisen. [½]
Some surplus may be held back as a contingency. [½]
The proportion held back may vary from year to year, to produce a smoother progression of
dividend payments. [½]
Some or all of this deferred surplus may be released to the policyholder at maturity in the form of
a terminal dividend. [½]
[Total 4]
7.3 There are three methods: additions to benefits, contribution and revalorisation. [½]
Additions to benefits method
The method differs depending on whether contracts are conventional or accumulating with-
profits. [½]
Conventional with-profits
Surplus arising on the policyholders’ fund from interest, mortality and expenses (and possibly
lapses) is distributed. [½]
This is distributed by means of:

 regular reversionary bonus, which is a percentage addition to the sum insured and is
guaranteed once added, ... [½]
... and the rates of these bonuses generally vary only gradually over time [½]

 special reversionary bonus, a percentage addition to the sum insured in respect of any
one-off unusual accumulation of surplus, and becomes guaranteed once added [½]
 terminal bonus, an addition to the sum insured added at the time of a claim (eg on death,
maturity, or surrender). [½]
Accumulating with-profits (AWP)
Here the policy benefit takes the form of an accumulating fund of premiums. [½]
Profits are distributed by means of:

 regular bonus, applied as a proportion of the fund (and often on a daily basis) [½]
 terminal bonus, added to the fund at the time of claim. [½]
Under AWP it is more common for investment surplus only to be distributed, although other
sources could be included. [½]
The accumulating benefit fund may have a unitised structure. [½]
Both types of additions to benefit
The surplus distribution is normally determined for fairly homogenous groups of policies, eg by
product and (for terminal bonus) by duration. [½]
Comparison is made with smoothed asset shares at maturity to determine appropriate payouts
for each group. [½]
The method allows:

 smoothing of bonuses over time [½]
 a high proportion of equity / property investment (usually higher for conventional then
accumulating) [½]
 actuary to apply judgmental considerations [½]
... but any perceived inequity by policyholders may be difficult to explain / justify. [½]
AWP can be made more transparent than the conventional method, and may therefore be easier
for policyholders to understand. [½]
The contribution method
This method identifies the investment, mortality and expense surplus arising on individual
policies. [½]
The actual experience rates used to calculate these individual surplus contributions would be
based on the experience of broadly homogeneous groups of policyholders. [½]

The surplus could be distributed by means of:

 an addition to the guaranteed benefit [½]
 reduction in premium [½]
 cash back to policyholders [½]
 retention of surplus for eventual ‘terminal dividend’ distribution on claim / maturity. [½]
The total distributable surplus each year is determined by the company, so there is some
discretion for smoothing of the profit distribution over time. [½]
The subdivision of the surplus between the homogeneous groups is mostly driven by a pre-set
formula, which results in profits being distributed in proportion to the groups’ contributions to
that profit. [½]
Some judgement will be necessary in allocating any unusual sources of surplus to policies (eg from
a large capital gain from property). [½]
The ability to invest in volatile assets such as equities and property would normally depend on the
relative amount of terminal dividend being used. [½]
The revalorisation method
Under this method, the investment surplus on the policyholders’ fund is distributed to all with-
profits policies by means of an increase in reserves. [½]
This increase in reserves is effected by means of an increase in benefit, and there may also be an
increase in premiums. [½]
The surplus may also be calculated to include a mortality and expense contribution, … [½]
... although such items may be considered as a reward for the insurance risks undertaken and
hence be reasonably distributed entirely to the shareholders. [½]
The method treats all policies in the same way, and is formula driven (objective). [½]
No profit deferral occurs, hence discouraging investment in more volatile asset classes. [½]
[Maximum 15]
7.4 (i) Distributing savings profit
The company will calculate its actual investment return i  for the year. [½]
The savings profit to be distributed for the year (as a percentage of reserves) would then be
calculated as the excess of this return over that expected according to the valuation basis. [½]
Examples of such calculations are:
r  k (i   i ) and: r  k i  i
where k is some contractual constant (eg 85%, 90%), and i is the valuation rate of interest.
r has a minimum value of zero. [½]

The amount of profit distributed to a policyholder whose policy has an end-year reserve of t 1V
is:
r t 1V [½]
This profit is distributed by increasing both the sum assured, St , and future annual premiums, Pt ,
by proportion r, ie:
St 1  St (1  r)
and:
Pt 1  Pt (1  r) [½]
Alternatively, increase the sum assured, only, by proportion b, ie:
St 1  St (1  b)
Pt 1  Pt
where b is such that:
(1  r ) t 1V  St (1  b) Ax t 1  Pt a
x t 1
assuming a whole life policy for age x at entry, premiums payable for life, and benefits payable at
the end of the year of death. [½]
[Maximum 2]
(ii) Suitability to meet needs
First it depends on the nature of the death benefit. If the death benefit sum assured is the same
as the maturity benefit sum assured, then this is an unnecessary diversion of cost away from the
main aim. If the contract could be arranged with, say, just a return of premiums with interest on
death, then this would be better. [1]
The individual would probably wish to pay a similar proportion of their income into the policy
each year (or possibly a rising proportion). The structure of increasing premiums each year may
approximately meet this need. [1]
On the other hand, lack of flexibility of premium amount may mean that periods of overpayment
or underpayment relative to desired levels are very likely to happen. [1]
The product is likely to be backed 100% with fixed-interest assets, which are likely to provide
lower returns on premiums over such a long period than would real assets, such as equities. The
product is therefore unlikely to be appropriate from this point of view. [1]
The need for guarantees nearer to retirement is important, and the product matches this need
well, at least in terms of guaranteeing a lump-sum payment. [½]

The form of the benefit as a lump sum is not itself appropriate: it will need to be converted into
an income at retirement. If an annuity is purchased with the lump sum for this purpose, there is a
risk that the company will be offering poor annuity rates at that time and that this will result in a
poor income for the policyholder. [1]
This risk could be mitigated to an extent if guaranteed annuity conversion terms were included in
the endowment contract. [½]
On the other hand, a lump sum provides the policyholder with maximum flexibility: eg they could
spend some, give some away (eg to children), invest some in non-insured savings and buy an
annuity with the rest. [½]
They could also wait until annuity rates improve before converting the cash into an annuity. [½]
As it stands, the product would not be very flexible if the policyholder decides that they want to
retire earlier or later than their 60th birthday, as it has a fixed maturity date. [½]
Early surrender may be possible to provide a lump sum if they want to retire early, but is unlikely
to be as good value for money. [½]
The policy does not appear to be very flexible to deal with any change in employment
circumstances before retirement age, such as becoming unemployed, or changing jobs, both of
which are likely to require changes in premium levels, and quite possibly more than one change
could occur over the period. [1]
The policy could probably be made paid-up, but premium variation beyond this is likely to be
problematic. Also unlikely to be cheap to do. [½]
[Maximum 8]
7.5 (i) Profit distribution and investment strategies
Additions to benefits method
The company will probably aim to keep the guarantees (the sum assured plus reversionary
bonuses) low, producing a deferral of the distribution of profit, and hence giving the company the
freedom to invest in real assets such as equities and property. [1]
This should increase the expected investment returns thereby increasing the potential payout
under the contract. [½]
A strategy of higher guarantees, with correspondingly less real investment content, is also
possible. [½]
Other sources of profit, such as from expenses and mortality, would be included in the surplus
distributed to policyholders, thereby increasing expected returns. [½]
The policyholders might also share in the profits from the company’s without-profits business,
again increasing the expected return. [½]
The company would aim to achieve broad equity of profit distribution compared with other
cohorts / tranches of with-profits policyholders. [½]

Payouts could therefore be affected by cross-subsidy either way. [½]
Payouts could be considerably smoothed over time (although the extent would depend upon the
company concerned), hence individual years’ payouts could see considerable under- or over-
payment on maturity compared with asset shares. [1]
Some contribution might be made from the policy’s profits to the company’s estate (or vice versa)
over the duration of the policy, decreasing (or increasing) the likely level of policy payouts
accordingly. [½]
Revalorisation method
A guaranteed sum assured defines the minimum level of payout under the contract. [½]
Some (high) fixed percentage of each year’s investment profit will be added to the reserves. [½]
Often, the proportionate increase in reserves will cause both benefits and premiums to be
increased to the same extent. [½]
Alternatively the benefit only will be increased, but by a lower proportion. [½]
No other sources of profit would usually be distributed. [½]
Because there is very little profit deferral, the company would probably need to invest in
matching assets to protect solvency, which would mean investing mostly in fixed-interest stocks
of appropriate term. [1]
However the distribution of the investment profit is likely to be very equitable between
policyholders (as it is being distributed proportionately to reserves) [½]
Comparison of yield on premiums between the two types
It is likely that the revalorisation method will produce lower expected returns than the additions
to benefits approach, in relation to premiums paid, because: [½]
 investments are likely to be weighted more towards fixed interest rather than real
assets [½]
 expense and mortality profits are not distributed [½]
 profits from other business are not distributed. [½]
But on some occasions the method can lead to better actual returns in relation to premiums paid.
[½]
This can arise if, under the additions to benefits method:

 the returns from equities and properties have been worse than fixed interest [½]
 net losses have been incurred from expenses and mortality [½]
 policyholders have shared in losses that have been incurred from the non-profit
business [½]

 a particular policy class or tranche was receiving lower than its fair share of the actual
contribution to profit (under the ‘‘broad equity’’ approach) [½]
 payouts were lower than asset shares due to smoothing [½]
 the company adopted a strategy of making a net contribution to the estate over the
duration of the policy. [½]
[Maximum 12]
(ii) Grounds for complaint
The revalorisation method usually involves increasing the premium as well as benefits each year,
hence if the sales representative is quoting for the same starting level of premium (ie £1,000 in
the first year) then the claim of a higher maturity payout is quite feasible. This, however, gives no
grounds for complaint. [1]
Even if the policyholder of Company A did obtain worse value for money than they would have
done under Company B (for reasons such as set out at the end of part (i)), there may still be no
grounds for complaint provided the causes were clearly part of the policyholder’s reasonable
expectations when the policy was taken out. [1]
Hence genuine reasons for complaint might be where:

 any or all of these losses were to occur when the policyholders had not been reasonably
informed about their possibility in advance [½]
 the insurer invested inappropriately and produced inferior long-term returns contrary to
reasonable expectations [½]
 the insurer operated inefficiently (high expenses) and/or engaged in inappropriate use of
capital (eg failed product launches, etc), if contrary to reasonable expectations [1]
 the company adopted a different bonus distribution strategy from that expected, perhaps
with a different approach to smoothing, or to contributing to the estate, or to equity,
etc. [½]
[Maximum 3]

End of Part 1
What next?
1. Briefly review the key areas of Part 1 and/or re-read the summaries at the end of
Chapters 1 to 7.
2. Ensure you have attempted some of the Practice Questions at the end of each chapter in
Part 1. If you don’t have time to do them all, you could save the remainder for use as part
of your revision.
3. Attempt Assignment X1.
Time to consider …
… ‘learning and revision’ products
Marking – Recall that you can buy Series Marking or more flexible Marking Vouchers to have
your assignments marked by ActEd. Results of surveys suggest that attempting the
assignments and having them marked improves your chances of passing the exam. One
student said:
‘The insight into my interpretation of the questions compared with that of

the model solutions was helpful. Also, the pointers as to how to shorten
the amount of work required to reach an answer were appreciated.’
You can find lots more information on our website at www.ActEd.co.uk.
Buy online at www.ActEd.co.uk/estore





using the course.

SP2-08: The general business environment (1) Page 1
The general business

environment (1)
Syllabus objective
insurance business, in terms of:
 local culture
 methods of sale
 taxation regime
(The first four points are covered in this chapter.)

Page 2 SP2-08: The general business environment (1)
0 Introduction
It is important for actuaries to be aware of the general business environment whenever they
attempt to solve an actuarial problem. In this chapter and the next we shall be studying some
aspects of that environment.
We start by considering the degree to which people seek to buy life insurance and the desire of
companies to sell.
We then look at how life insurance companies go about selling the products they have to offer
through the various distribution channels that are used.
The distribution channels we shall be studying are:

 insurance intermediaries, who select products for their clients from all or most of those
available on the market
 tied agents, who offer the products of one life insurance company or a small number of
life insurance companies
 own sales force, usually employed by a particular company to sell its products direct to
the public
 direct marketing via press advertising, over the telephone, internet or mailshots.

1 Propensity of consumers to purchase products

In general it is not compulsory for consumers to purchase life insurance products, and so
the likelihood of a consumer to purchase depends upon the combined forces of their own
natural inclination to buy, and the power of the insurer to sell.
In the life insurance marketplace, there is a large imbalance of power between buyer and seller.
Power in this sense means knowledge and information. For example, an insurance company will
be fully aware of the extent of the charges for expenses and profit in its prices. Furthermore it
will have good knowledge about how competitive its products are. The customer, on the other
hand, will have little knowledge of these things. So an unscrupulous insurance company (or
intermediary) could misuse its position of power to persuade an unwary customer to buy a
particular product without sharing this information. Regulations have been imposed in many
territories to prevent just this kind of malpractice from occurring (see Chapter 9 for more
information).
On the other hand, many people can benefit from having insurance products, and when properly
informed will choose to buy.
Question
Life insurance is often described as being sold, not bought.
Explain what is meant by this and why it is often the case.
Solution
People usually decide to buy something through a feeling of need for, or desire for, the product
(eg food, clothing). With life insurance, many people may not be aware that they need the
product (at least, not until it’s too late), so they will not always deliberately go and buy it. So,
much life insurance is sold following approaches made by insurers and intermediaries to the
public, rather than the other way round.
In summary, the reasons include:

 public unaware of need, either through ignorance, or from a belief that the event ‘won’t
happen to them’
 death may be a taboo subject
 products are difficult to understand: this puts people off
 difficult marketplace – the presence of many providers makes it hard for individuals to
know which product(s) and company(ies) would be best to meet their specific need
 general need for advice before making long-term financial commitments, which makes
the whole process very time consuming.

In the long run, a life insurer will run a risk to profits if its sales effort is not aligned to
selling products with useful benefits to consumers.
That is, with the proper sharing of information, customers will only be expected to buy life
insurance products that provide them with useful benefits. So companies have to supply
products that do just that, if they are going to do successful business.
Nevertheless, consumers are likely to need to be educated and informed as to the benefits
of life insurance and the particular aspects of the insurance company’s products. Often this
will involve a salesperson who receives remuneration from the sale of an insurance
company’s product.
Thus the role of the salesperson is to bridge the gap between the lack of knowledge and
awareness of the consumer and the commercial power of the product providers.
Question
Outline any problems with using a salesperson for this purpose.
Solution
There might be a problem because of the way in which the salesperson is remunerated.
The reason is that the salesperson is rewarded for making a sale, not for just giving advice and
information (at least in the situation described in this Core Reading). So a salesperson has a
financial incentive to sell a product, whereas it may be that the customer would be best advised
not to buy anything.
The situation is more complex in practice where salespeople receive different remuneration
depending on the type – and size – of product sold (which is normally the case) – this would
‘encourage’ the sale of the more expensive and/or larger products.
The insurer runs the risk that a policy is sold which does not meet the agreed needs of a
policyholder. In such an event, the key risks to the insurer are:
 persistency risk, and consequent financial losses – including the possibility of

compensation
Persistency is defined in the Glossary:
In a life insurance company, ‘persistency’ is used to refer to the rate of retention of

policies that is experienced by the company. If a company has ‘poor persistency’,
this indicates a high level of lapses, surrenders, partial withdrawals and/or
conversions to paid-up status.
So an example of the persistency risk mentioned in the Core Reading above is the risk that
the policyholder lapses or surrenders early. Persistency risk is also referred to as
withdrawal risk.
 reputational risk.

Question
Explain what is meant by ‘reputational risk’ and what the financial impact of such a risk might be.
Solution
It means the risk that the insurance company will earn itself a bad name.
‘Bad news’ travels fast around the insurance marketplace, and could significantly damage a
company’s ability to sell business. Poor sales leads to inadequate premiums to cover the
company’s fixed expenses. (We will meet this idea again in later chapters on ‘risk’.)
Persistency and reputational risks can be minimised provided that the main elements of the
sales process are clearly documented, and the policy literature is clearly explained whilst
remaining legally sound.
The risks are also minimised by insurers and salespeople adopting a fully professional approach to
their roles in the sales process.
Changes in the types of products consumers purchase (for a variety of reasons) lead to
changes in the assumed business volume or mix, and therefore presents a new business
risk for the insurer.
These risks are fully discussed in Chapter 11.
Local culture
Local culture and religious beliefs can impact the propensity of consumers to purchase insurance
products. For example, the extent to which products are purchased can depend on:
 the degree of trust that individuals have in financial services companies, particularly large
multinational corporations
 the level of education and financial awareness, which itself depends on the extent to
which education is seen as being important
 the extent to which families provide support to each other, such as during times of
unemployment, ill health and old age
 the age at which children become self-supporting
 cultural attitudes towards death.

2 The main distribution channels
2.1 Insurance intermediaries

Insurance intermediaries are salespeople who must act independently of any life insurance
company (although they can be owned by one). Their aim is to find the contract that best
meets, in terms of benefits and premiums, the needs and situation of their clients. They
may be remunerated, via commission payments, by the companies whose products they
sell, or they may receive a fee from their clients.
It will certainly be the case that the most professional intermediaries will seek the best contracts
for their clients. But in some markets, it has been known for less scrupulous intermediaries to be
influenced more by the levels of commission payable than by the appropriateness of a particular
product.
Very often, when faced with several similar products which are all appropriate for the client and
which will differ primarily in eventual investment performance (which is unknown), the
intermediary will be strongly influenced by commission. For this reason some intermediaries
offer their services for a fee (in some countries this may be a legal requirement), to demonstrate
that they will not be influenced by commission levels.
A good intermediary can provide a valuable service for the client. Most individuals will lack the
time and expertise to find the products which are the best value and most appropriate for their
needs, but the intermediary can perform the necessary research.
It will often be the client who initiates the sale. However, intermediaries are also likely to
promote themselves actively to existing clients by, for example, instigating a periodic
review of finances.
Promotion may also involve mailshots to the existing client base, for example when an innovative
product is launched or tax rules change. Intermediaries may also promote themselves to new
customers, for example by press advertising or through trade associations.
Alternative names for insurance intermediaries are insurance brokers and (independent)
financial advisers.
2.2 Tied agents

Tied agents are salespeople who work solely on behalf of one, or sometimes several, life
insurance companies; that is, they offer to their clients only the products of those
companies. Typically they may be the employees of a bank or other similar financial
institution.
Question
Explain the key advantage for a life insurer of linking up with a bank in this way.

Solution
This gives access to the ‘warm’ customer base that can be accessed through the bank’s branches
or through mailshots, and hence increased sales. (‘Warm’ customers are those with whom the
company already has an established relationship, and who are likely to be favourably disposed to
the company and to buying their products.)
A bank might even set up its own insurance company (if local legislation permits), to which it then
becomes a ‘tied agent’, marketing insurance products through its bank branches and through
mailshots to customers.
A bank is not the only possible type of tied agent. In Continental Europe the traditional
distribution method is through individuals who act as tied agents. Legally they work on behalf of
the insurance company, without being employees. This contrasts with insurance intermediaries,
who in a legal sense act on behalf of their client.
Where the tie is to more than one company, it will sometimes be the case that the product
ranges of the companies are mutually exclusive, but more often there will be an overlap.
Tied agents are remunerated by the companies to which they are tied. The remuneration
could be in the form of commission payments or by salary plus bonuses.
It will often be the client who will initiate the sale, but tied agents may actively engage in
selling.
The insurance company will certainly be hoping that their tied agents are actively engaged in
selling.
2.3 Own salesforce

Members of an own salesforce will usually be employees of a life insurance company and
will only sell the products of that company. They may be remunerated by commission or
salary or a mixture of both.
It will usually be the salesperson who initiates a sale, making use of client lists or
purchased leads. However, once a rapport has been established with a client, it will then
often be the client themselves who initiates further sales.
Many salespeople will also be active in trying to make further sales. It is often seen as helpful for
a company to have a full range of products if it is selling through an own salesforce. This way,
once it has a ‘warm’ client (one with whom it has an established relationship), it can try to sell
that client life protection, savings contracts and other insurance products (such as healthcare or
general insurance).
Of course, lists of clients have to be acquired in the first place. This might be achieved, for
example, through press advertising, personal contacts, or by purchasing a list from a marketing
company (who might, for example, provide a list of, say, 5,000 people who have moved house in
the last 12 months).

2.4 Direct marketing

This takes four main forms:
 mailshots
 telephone selling
 press advertising
 internet selling.
To send out a mailshot, the life insurer will get a list of names and addresses, for example from its
database of current policyholders or other affinity groups such as holders of a specific credit card.
The life insurer will write to all the names on the list inviting them to take out a policy. The letter
will include details of the contract being offered and enclosing an application form.
In the case of mailshots, arguably the life insurance company initiates the sale.
In the case of telephone selling, it would be either the prospective policyholder or the life
insurance company that initiates the sale.
Thus the sales process could involve the ‘cold-calling’ of prospective customers, where this is
allowed. Alternatively the customer might call in response to a newspaper or television
advertisement, with the sale then being completed over the telephone. In this latter case,
telephone selling and press advertising are really parts of the same selling process.
In the case of press advertising, it is debatable who initiates the sale.
If a company advertises and someone responds, who is ‘initiating the sale’?
Press advertising may take various forms, for example:

 It may include a short application form for the customer to complete and send in.
 It may give a telephone number or address from which further information and an
application form may be obtained.
 It may give a telephone number to call for the sale to be completed on the telephone, as
described above.
There is now considerable use of the internet, both for marketing and for selling products.
Internet selling is mainly limited to simple products to date, but it can also be used to
provide information leading to sales by telephone or other means.
It is most useful for without-profits contracts, such as term assurances and annuities, where there
is a fixed benefit for a fixed premium. It is usually possible to get a quote on line, and some sites
offer a choice of companies (for example, giving quotes for the ten cheapest premiums available
for the policy requested). The facility to apply on line for a chosen quote is then usually provided.
Internet quotes and applications can take into account basic underwriting information, such as
age, gender, height, weight, smoking habits, and a few brief questions about health and family
history.

Question
Discuss who initiates the sale when the internet is used.
Solution
Normally we would say this has to be the customer – he or she has to type in the words ‘life
insurance’ (or something similar), which is difficult to do by accident.
On the other hand, pop-up screens can be used to advertise products in the user’s face, as it
were, though care has to be taken not to irritate potential customers by the intrusion caused.
A quick search under ‘life insurance’ or ‘annuity rates’ will generate examples of what’s available.
Some direct marketing is now being performed using other forms of social media, such as
Twitter and Facebook.

3 The effect of different distribution channels
3.1 Demographic profile

Different channels are likely to appeal to different people, according to their level of
financial sophistication and level of income. These differences will then be reflected in the
resulting demographic experience of the lives taking out contracts through each channel,
ie there will be socio-economic selection.
Socio-economic selection is sometimes called class selection.
Question
Comment on likely differences in financial sophistication and income levels between the
following:
(a) customers of an insurance intermediary
(b) customers of a major bank acting as tied agent for an insurance company
(c) customers reached by advertising in a downmarket tabloid newspaper.
Solution
(a) Those seeking the advice of an insurance intermediary may tend to be more financially
sophisticated and wealthier on average than the public as a whole. As such they will tend
to seek out advice in managing their more complex financial affairs.
(b) A major bank would reach a broad cross-section of people with different levels of financial
sophistication and income. A certain minimum level might be implied by their having a
bank account. It may still be those at the upper end of this range who seek further advice,
or who are specifically targeted by the bank for insurance sales.
(c) Readers of the tabloid would probably be the least financially sophisticated of the three
groups, and have relatively low levels of income. (Although perhaps in some less
developed countries the ability to read at all will indicate some level of education, and
perhaps therefore relative wealth.)
One cannot be too dogmatic about levels of income and financial sophistication associated with a
particular distribution channel. For example, an own salesforce might be targeted specifically at
professional people or at people in the lower socio-economic groupings.
Effectively it is the target market rather than the distribution channel itself that directly affects
the levels of income and financial sophistication of the customers reached. However, it is via the
distribution channels that the target markets are reached and certain channels will tend to be
associated with particular target markets.
The levels of income and financial sophistication of the customers will tend to be correlated with
their mortality, sickness and withdrawal experience.

Question
Consider mortality experience on term assurance contracts.
Explain how this is likely to differ between customers of an insurance intermediary and those of a
direct salesforce targeted at a broad range of society.
Solution
If the customers of intermediaries are generally wealthier than average then this might lead to
lighter mortality experience for a range of reasons:
 better living conditions and lifestyle (although too much claret can be bad for you)
 better access to health care
 more likely to be professional / managerial than manual work, and so less risk from
work-related hazards
 a correlation with education, in that wealthier people tend to have had a better education
and so be more aware of health issues (this is linked to the lifestyle aspect of the first
point).
(As we shall see later in this chapter, there may also be more stringent underwriting on business
placed by insurance intermediaries.)
The answer to the previous question illustrates the ‘socio-economic selection’ referred to in the
Core Reading. ‘Socio-economic selection’ simply refers to the fact that people can be usefully
classified by certain socio-economic attributes that affect their mortality or sickness experience.
For example smokers exhibit higher mortality than non-smokers do. In the situation we are
considering here, the attribute is the distribution channel by which the business has been
acquired.
The sales channel can have a direct effect on the withdrawal experience. Different channels
employ different sales processes, which can include differences in terms of aggressiveness of the
selling approach, the extent to which customer needs and ability to pay are considered, and who
initiates the sale. These factors can have an effect on persistency, particularly near the start of
the policy, when people who have been ‘over-sold’ their policies are most likely to cancel. Lowest
withdrawal rates should then be associated with insurance intermediaries, especially through the
effect of client-initiated selling and the sophistication of the client base, although there can be
many exceptions (eg where a salesforce targets a sophisticated market and adopts a carefully
focused customer-needs based approach).

3.2 Contract design

In broad terms, the higher the level of financial sophistication of the client base the greater
can be the complexity of the products being sold.
Question
The following table cross-tabulates some life insurance products with different distribution
channels.
Identify the products that are suitable for sale through each distribution channel, using a tick
where the product would be suitable and a cross where it would not be.
In some cases, things may not be so clear-cut. Explain why, where this is the case.
Distribution Channel
Insurance Own salesforce Telephone selling
intermediary
Without-profits
term assurance
Product
Without-profits
annuities
Unit-linked
flexible whole life

Solution
Don’t treat the following table as definitive, but a sensible answer would be:
Distribution Channel
Insurance Own salesforce Telephone selling
intermediary
Without-
profits term
  
assurance (Note 1)
Without-
 
Product
profits ?
annuities (Note 2)
Unit-linked
flexible
 ? 
whole life (Note 3) (Note 4)
Notes
(1) Although products can be complex when sold to a sophisticated target market, they don’t
have to be.
(2) Annuity contracts are rather more complex than term assurance. There may be more
options, and policy conditions will be more complex. There may also be tax and other
legal requirements that have to be followed, for example when dealing with the (often
common) situation where the applicant is using the maturity proceeds of a savings
contract to fund the annuity. We may therefore have something that is too complex to
explain and deal with over the telephone.
(3) It is tempting to put a tick here, and it would perhaps be equally reasonable to do so.
Once a salesperson and potential customer are face to face almost anything could be sold.
(4) This is almost certainly too complex for telephone selling. (Although a South African
reviewer pointed out that these contracts are a market standard in South Africa and for
this reason might be sold through all channels.)
Different products will suit different methods of sale. For example, the dynamics of selling
via the telephone, press advertisements or the internet mean that the products sold in this
way may be less complicated than products sold face to face.
This confirms the answer to the previous question, but we wanted you to think about it for
yourself first.
Perhaps the most important thing is that the product should appear simple to the prospective
policyholder, even if, in actuarial terms, it is quite complex.

An insurance company using more than one distribution channel may therefore sell
different versions of the same product, varying by channel.
3.3 Contract pricing

We shall consider two aspects of the effect of distribution channels on contract pricing:
 the effect on demographic assumptions (including the effect of underwriting)
 the effect on the need for competitive terms.
The effect on demographic assumptions

The level of underwriting exercised will be linked to the marketing strategy used. In general
terms, the stringency of the underwriting will follow the order of the four methods in
Section 0 above, with sales through insurance intermediaries being subject to the most
stringent.
Question
Suggest why business from intermediaries is likely to be subject to most underwriting.
Solution
Insurance intermediaries are representing the interests of their clients, not a particular insurance
company. The insurance company needs to be aware of the possibility that an intermediary
might encourage anti-selection.
It should also be borne in mind that it is the intermediary or his / her customer who is ‘initiating
the sale’, which may also increase the risk of anti-selection, compared with, say, when the
customer is targeted by a company salesperson.
Also, insurance intermediaries are more likely to have customers with high net worth (rich
people), with consequent need for higher insurance cover.
Prices will need to be competitive, (as intermediaries have access to, and in theory knowledge of,
the whole market), which may only be achievable by careful selection of good risks through
stringent underwriting.
(One factor working against this is that there will be competitive pressures not to make
underwriting too stringent for business sold through intermediaries, for fear of discouraging
business. In other words, if too great a proportion of applicants are refused acceptance at
standard rates by a particular company, the intermediary will stop recommending that company
for most of his or her clients.)
Some products will need little or no underwriting whatever the distribution channel used.
Question
Suggest products that will need relatively little underwriting.

Solution
Really, any contract which is basically savings or has a small sum at risk, so:
 annuities (except where favourable terms are offered to lives with impaired health)
 unit-linked endowments with small sums assured on death
 depending on how one interprets ‘relatively little’, perhaps also other (non-linked)
endowments, unless the death benefit is substantially higher than the maturity benefit.
Without-profits endowments are likely to need more underwriting than with-profits, as
the former are likely to be purchased by relatively risk averse policyholders who are more
concerned about insurance protection than investment potential.
(Even whole life contracts may have relatively little underwriting. This can be acceptable if the
sum assured is small and some other device is used to protect the insurance company, such as
giving a death benefit of return of premiums in the first two years of the contract.)
One reason why sales through direct marketing may have relatively low underwriting is that in
many markets the sums assured on offer have been relatively low. Where this is not the case,
quite stringent underwriting would be necessary. Another reason for low underwriting is that
complex application forms with lots of underwriting questions are a major barrier to sales.
The level of underwriting will then be reflected in the demographic assumptions used for
pricing. The socio-economic selection mentioned above will also affect these assumptions.
All other things being equal, the more stringently a life insurance company underwrites its
potential customers the better its mortality and sickness experience is likely to be.
However, not all things are equal. The underlying health of those applying for insurance is also
important, and as we have seen this may vary by distribution channel. This therefore interacts
with the insurance company’s efforts to improve its experience through underwriting.
Question
A life insurance company markets a regular premium without-profits whole life policy designed to
cover funeral expenses. It places advertisements in a newspaper read mainly by manual workers.
The contract offers guaranteed acceptance of all applications, but the benefit in the first twelve
months is only a return of premiums paid.
Comment on the likely mortality experience under this contract.

Solution
Experience would probably be poor. The target population may experience relatively poor health
and there is no underwriting. Experience might be worse than general population mortality in the
first few years until the effect of anti-selection has diminished.
The low death benefit in the first twelve months would at least discourage some attempts at
anti-selection by the seriously ill, although twelve months may not be a sufficiently long period.
Given that the life insurance company expects experience to differ depending on the types of
customer it reaches and the extent of underwriting used, it is clear that it will wish to reflect this
in its assumptions when it prices contracts.
Withdrawal rates are likely to be affected by the level of financial sophistication and whether
it was the policyholder who initiated the sale.
A major reason for customers withdrawing from contracts, particularly at early durations, is the
realisation that what they have bought was not suitable for their needs in the first place. This is
the key to both of the factors mentioned in the Core Reading above.
An unsophisticated customer is more likely to buy an unsuitable product, either because he or she
has not sought appropriate advice or through deliberate mis-selling. Also, if a customer has
actively sought out a product to meet a perceived need it is more likely that the product is
suitable, compared with, say, an aggressive sales push for a particular product.
The effect on the need for competitive terms

The competitiveness of premium rates is, in broad terms, likely to follow the order of the
methods in Section 0 above.
That is:
 insurance intermediaries
 tied agents
 own sales force
 direct marketing via press advertising or over the telephone.
However, as we shall see below, direct marketing will not always be subject to the least
competitive pressures.
Insurance intermediaries will recommend to their clients the companies with the most
competitive rates, other things being equal.
The most important thing that would need to be equal would be commission. In some markets,
particularly those that are not well regulated, it may be more important to offer good commission
terms than competitive premium rates in order to secure business.

A bank will want the products sold by its employees, as tied agents, to be reasonably
competitive or they could damage its good name.
However, there is not the same direct comparison with other companies that exists with
intermediaries.
Members of an own sales force will not usually be in a competitive situation.
For example, a salesperson may correctly recommend that a client needs a 15-year term
assurance contract. It may well be that most or all of the clients who decide to take out such a
contract do so without looking round the market for a better deal. That may not be very bright,
but that’s the way it is.
It is not clear where direct marketing should fit into this picture of competitiveness. It really
depends on the target market. For example, press advertising in a financial magazine or
heavyweight newspaper may reach financially sophisticated customers who are keen to compare
rates. Advertising in a more downmarket newspaper and to the financially unsophisticated may
involve much less competitive pressure.
Having competitive rates or charges is not the be-all and end-all for every product, even when
selling through intermediaries. For example:
 Products may be differentiated from the competition through innovative features or
attractive options.
 More complex products may be difficult to compare across companies.
 Some savings contracts may compete as much on past investment performance as on
premium rates or charges.
 Some products may compete on the level of customer service or admin support – eg for
group life or pension products.
Question
A life insurance company markets the following contracts through insurance intermediaries only:
(a) without-profits term assurance
(b) unit-linked endowment assurance.
Comment on the need for competitive rates / charges for these two contracts.

Solution
Selling through intermediaries generally requires competitive terms. In particular:
(a) Term assurance rates are easily compared and everybody’s product is pretty much like
everybody else’s. It is very important for an insurance company to have competitive rates
if it wishes to sell large volumes. Companies may increase competitiveness in other ways,
such as by including renewal or conversion options or other kinds of benefits, but this can
only be of limited effectiveness (eg not all policyholders require these add-ons).
(b) Competitive expense and other charges are important. However, competition will not
only be on charges but also on past investment performance.

4 Distribution channels and the actuarial control cycle
Question
Describe how the control cycle can influence the management of a life insurer’s sales distribution.
Solution
At a ‘macro’ level a company needs to decide which distribution channel or channels it should use
(specify the problem). The company needs to choose a distribution strategy which should help it
to meet its targets for future profitability whilst maintaining an acceptable level of risk. An
analysis of the possible channels available in the market (and those being used by the company’s
competitors) should indicate likely successful methods (general commercial environment).
The choice of distribution channels will affect a number of important assumptions in the
company’s projection models (developing the solution). These will include:
 the volume, types and mix of business likely to be sold in the future, including contract
design, pricing (ie profitability) of those contracts, and policy size
 the mortality, morbidity and withdrawal rates likely to be experienced by the new
business.
The strategy that best achieves the company’s aims, according to these modelling exercises,
might be adopted. However, in these models, the costs and risks involved in making major
changes to the distribution system at the macro level should be fully taken into account. It would
be vital to test fully the impact of variations from our central assumptions before adopting a new
strategy, particularly in terms of the first of the above bullet points (which is extremely uncertain
and subjective).
The impact of the sales distribution strategy being used at the moment must be continually
analysed (monitoring and feedback). Have the levels, type and quality of new business (including
the resulting demographic experience) materialised as expected in our models? If not, then our
assumptions may need to be changed, and this in turn could affect the ongoing decisions being
made.
At a micro level, the company should always be striving to make improvements within its overall
chosen sales distribution strategy. Analysing the experience from individual sales outlets, such as
from individual intermediaries, tied agents, sales persons or direct marketing routes, can point to
areas of poor and good performance. Poor performance might then be improved by adopting
better training or administration procedures, or by providing better sales information, for
example.
A professional approach to sales distribution is necessary (professionalism). The consumers’

interests should be borne fully in mind. Selling methods that are considered to be unethical
should be avoided, and individual examples of poor practice need to be identified and rectified.
Good practice should be fostered.

One thing that should be observed from this example is that, whilst the actuary has some interest
in the company’s sales distribution strategy, he or she may have little personal involvement in the
decision making process. This shows how the ethos of the control cycle should be embraced by all
decision makers in a company, and is not something that just the actuary has to worry about.

Chapter 8 Summary
Consumer inclination to buy life insurance
Consumers have a natural tendency to avoid the issue of buying most life insurance products
because it may seem a distant need, it may be a taboo subject, or people are simply unaware
that the products exist.
There is an imbalance of power between buyers and sellers of life insurance in terms of
knowledge and information about products, prices and the market. The market is very
complex with a confusing array of products and providers.
It is often necessary for intermediaries to help advise customers about their life insurance
needs, and to sort out for them which products, and possibly companies, would best meet
their requirements.
Persons given proper advice and information will usually recognise any real need for life
insurance and would then seek to buy.
Failure by insurers to sell contracts that meet real consumer needs will ultimately result in
high policy withdrawals, a bad name, and poor new business sales.
Distribution channels
Distribution channels are the routes through which life insurers sell their products. The main
distribution channels are:
 insurance intermediaries, who select products for their clients from all those
available on the market
 tied agents, who offer the products of only one life insurance company or of a small
number of life insurance companies
 own sales force, usually employed by a particular company to sell its products direct
to the public
 direct marketing via mailshots, press advertising, over the telephone or internet.
Insurance intermediaries may be rewarded by commission from the companies whose

products they sell or by fees from their clients. Tied agents are usually rewarded through
commission. The company’s own salesforce may be rewarded by salary or commission or a
mixture of the two.

Effects of distribution channels

Different distribution channels tend to reach different types of people. The customers may
differ in levels of financial sophistication, keenness to take out a contract and levels of
income. Consequently the mortality and withdrawal experience of business secured through
different channels may not be the same.
The level of underwriting, which may also vary by distribution channel, will also affect this
experience.
These differences in expected experience should be reflected in assumptions used in contract

pricing.
The need for competitive terms varies by distribution channel, with the greatest degree of
competition being in business sold through insurance intermediaries. The importance of this
varies by type of contract.
Where a channel tends to reach a more financially sophisticated target market then more
complex product designs are possible. However, contracts sold through press advertising,
telephone selling or the internet have to be fairly simple, even when selling to a sophisticated
market.


8.1 Describe how the four main distribution channels used by life insurance companies operate. [8]
Exam style



Chapter 8 Solutions
8.1 The four distribution channels are:
Insurance intermediaries
These are independent of any particular life insurance company. They can advise their clients on
the best contracts for their needs from among all the contracts available. [1]
They may be remunerated by commission from the insurance companies whose products they
sell or by fees from their clients. [1]
Tied agents
These are salespeople who offer the products of only one or a limited number of insurance
companies. If the tie is to more than one company then usually the product ranges of the
companies are mutually exclusive. [1]
Typically, tied agents are financial institutions such as banks. Tied agents are remunerated by the
companies to which they are tied. The remuneration could be in the form of commission
payments or by salary plus bonuses. [1]
Own salesforce
These are usually employees of a life insurance company and so only sell the products of that
company. They may be paid by salary or commission or a mixture of the two. [1]
Direct marketing
The main forms of direct marketing are telephone selling, press advertising, mailshots, and the
Internet. [1]
Telephone selling might involve ‘cold-calling’ by the company or might be in response to an

advertisement (in which case press advertising and telephone selling are part of the same
process). [1]
Press advertising might include an application form or invite requests for further information.
Mailshots will definitely include application forms. [1]
Internet selling may be linked to advertising, and is essentially web-based application

processing. [½]
[Maximum 8]





using the course.

The general business

environment (2)
Syllabus objective
insurance business, in terms of:
 local culture
 methods of sale
 taxation regime
(The last six points are covered in this chapter.)

0 Introduction
In this chapter we describe the main kinds of expenses, and we study the economic, legal,
regulatory, taxation and professional influences that may affect the way that life insurance
companies operate.

1 Types of expenses and commissions, including influence of

inflation
A life insurance company will incur costs in running its business: commission payable to
salespeople (unless the business is written on a fee basis) and management expenses.
Initial commission may be payable on the acquisition of a new policy, with renewal
commission payable each time a renewal premium is paid. If a policy lapses, part of the
initial commission may be recoverable and there will be a risk of non-recovery.
Management expenses consist of expenses that are incurred directly when new policies are
written (new business) or maintained (in-force business), and overhead expenses that a
company will incur regardless of the amount of new business it writes and the business it
has in force. These overheads would include, for example, the costs of general
management, at least part of the costs of a company’s service departments (such as IT and
HR), and the cost of accommodation.
There may also be expenses associated with terminating a policy, for example the cost involved in
making a death claim payment on a whole life insurance.
These expenses would be split in various ways to determine, for example, the expense
loadings to be built into premium rates. This is considered in more detail in Section 3.4 of
Chapter 30.
The term ‘overheads’ can be used to mean a number of different things. Here it is firstly being
used to refer to any type of expense that is not specifically associated with policy activity (ie with
its set up, maintenance or termination). An example might be the cost of electricity used to run
the company’s offices and computer systems.
On the other hand, we can argue that, without policy activity we would not need offices or
computers – in other words the company’s electricity is also a cost associated with policy activity.
Moreover, we might even be able to establish how much of the electricity bill might be
apportionable to initial, renewal or termination activities.
Question
Suggest how the company’s lighting and heating costs might be apportioned between initial,
renewal and termination expenses.

Solution
Heating and lighting costs are essentially proportional to the amount of (office) space that is
heated and lit. So we would need to decide what proportion of the company’s office space is
devoted to initial, renewal, and termination activities.
For some office areas this might be straightforward. For example, underwriting and marketing
departments may be solely occupied with new business activity (initial expenses), while there may
be claim departments solely involved in policy terminations. Other areas would be more difficult
to apportion, such as the actuarial department or general management offices. For these it might
be sensible to assess how much of the occupants’ (staff) time is spent on the different activities so
as to get a realistic split.
(In the above question we have started to think about expense analysis, which is a very important
aspect of experience monitoring. We will return to this in detail towards the end of the course.)
As indicated by the above, it is possible to associate any kind of expense with one or several of the
three main areas of policy activity.
The Core Reading then goes on to say that overheads represent costs that are (broadly) fixed, and
this is, perhaps, the more normal way of considering them. For example, the same fixed initial
costs will be incurred during the next month whether we sell 100 new policies or 150 new policies
during the month. Those initial costs that change according to the number of policies sold would
be the so-called variable initial costs. (We have introduced this idea of overheads already in
Chapter 1.)
Question
Give an example of each of the following:

(a) a fixed cost
(b) a variable cost.
Solution
(a) Fixed cost
An example would be the salaries of the actuarial department staff (indeed, the salaries of nearly
all staff would be deemed as fixed costs).
(b) Variable cost
Writing, printing and postage of policy documents for a new policy (or for any other documents
sent to the policyholder once the policy is in force, or at time of claim).

There is, of course, a limit on what can be considered as fixed costs. For example, a company that
has the capacity to sell 100 new policies a month would probably be able to cope with occasional
months in which sales may be pushing 200. But it may not be able to cope with, say, 500 new
policies a month on a regular basis without recruiting additional staff. So the ‘fixed’ costs remain
strictly fixed only within a range of business activity.
There is a profitability risk that these loadings will prove insufficient to meet the actual
expenses incurred. See also Section 4 of Chapter 10.
At a simpler level, risks can arise if a company finds that it cannot contain costs, including
the rate at which they increase due to inflation.
Inflation affects underlying costs, which in turn influence the level of expenses allocated to
policies.
Question
Explain why we are interested in allocating expenses to policies.
Solution
This is so that we can work out, for example, how much each policy costs the company to
administer, which should enable us to calculate an appropriate premium (or reserve value) for the
policy.
Many underlying costs are directly or indirectly linked to wage and salary levels. Others are
influenced by the general level of prices or by the prices of particular commodities. Publicly
available data, eg retail price index, national average earnings index, and similar data
internal to the life insurance company, can be used to decide what assumptions to make
about general price, future wage or specific price inflation.
A particularly difficult aspect of managing a life insurer’s costs arises from the fact that new
business may be cyclical compared with the more predictable base of ongoing renewal and
in-force business. A significant part of operating expenses arises from the need to employ
experienced staff; if a downturn in new business forces the need for redundancies,
re-employing a similar quality of staff may prove difficult when the new business improves
again.
This contrast between new business and renewal costs is a function of the average duration of the
policies that the company sells. If all policies have 20-year terms, then (all else being equal) a
fluctuation in new business volume will be diluted approximately 20 times in terms of its effect on
the in-force volume (the policies issued over the past 19 years will be unaffected by the volume of
business issued over the past year). Renewal expenses will therefore be much less volatile over
time than expenses that are purely associated with new business sales.
The company would therefore like to have much more flexibility with regard to the number of
new business staff that it employs from time to time, whereas this is much less of an issue with
regard to staff employed on renewals.

Question
Explain how the demand for claim staff would vary over time, compared with the demand for
staff to process new business and renewals.
Solution
Some claim costs, in particular policy maturities, may be unevenly distributed over time. An
example would be where a company has had sales campaigns in the past for policies of specific
terms – eg of ten-year endowment assurance policies sold in 2012. So in this case there would be
an unusually large number of maturity claims to deal with in 2022.
On the other hand, death claims and surrender claims will be spread over time regardless of
whether a vast number of additional policies were sold in 2012 or not.
Altogether this will make demands on claim staff time less volatile than for new business, but
more volatile than for renewals.

2 Economic environment
The state of the economy in which an insurer operates will carry risks for the insurer. The
availability of asset types, and their short- and long-term expected yields, will determine
how well an insurer can choose investments and the probability of securing the return
assumed when setting premium rates.
Question
Explain how the whole of the previous sentence might affect the return that should be assumed
when setting premium rates in the first place.
Solution
The expected yields on the assets that the company intends to invest in will form the best
estimate of the future investment return, and this is a key factor in setting the interest rate
assumption.
The uncertainty associated with the future returns, either inherent in the asset classes themselves
or as a result of any lack of availability, will lead to the company needing to allow some margin in
the basis, to cover the event that yields turn out lower than best estimate.
There are other factors to take into account when setting the interest rate assumption – more
about this later in the course.
From the consumers’ point of view, insurance products may be seen as more or less
attractive compared with other available investments (or not investing at all).
Economies in which the investment markets are more volatile will tend to have more
expensive insurance products and possibly less take-up of them. The insurers will tend to
have relatively higher capital requirements because of increased uncertainty of investment
return.
This will increase the cost of capital and so increase product charges and premiums.
An insurer investing in more risky / speculative markets is likely to seek a greater expected
rate of return on capital, leading to a relatively greater risk of the required return not being
achieved.
This is an additional reason for increasing the cost of capital, as it will mean a higher risk discount
rate will be used for discounting profits.

Question
In the country of Niceandeasyland, insurance companies sell regular premium without-profits

endowment assurances and without-profits term assurances. Niceandeasyland has a stable
economy with historically little volatility in its investment markets.
In neighbouring Boomandbustland, where mad government succeeds mad government and

investment markets are unpredictable, insurance companies offer the same two types of
contract. The expected rate of return on like investments is the same in both countries.
Explain how the prices of the two contracts are likely to differ between the two countries.
Solution
The endowment assurances will be vastly more expensive in Boomandbustland. This is because
companies will need to hold much larger reserves / solvency capital in this country to cover the
significant risk that future investment returns will prove insufficient to cover maturity values. The
high reserves delay the emergence of profits, reducing their value to shareholders, and hence
requiring higher premiums to compensate. Furthermore, the equity markets will demand higher
risk premiums for shareholders, further increasing the cost of capital.
The prices of the term assurances will be much more similar to each other, as investment returns
have a relatively minor impact on profitability and hence little effect on overall risk. This should
lead to much more similar capital requirements (though there will still be some difference) and
the difference in premiums will be much less as a result.
Even the risk discount rate may not need to be so high for the term assurance in Boomandbustland
compared with that for the endowment, because of the less significant increase in risk.
You may not have got all of this solution on your first attempt, particularly the parts in italics. We
cover the subject of determining risk discount rates in later chapters, and suggest you attempt this
question again once you have studied the relevant sections.

3 Legal environment
Assuming a politically stable operating environment in which the legal processes of
contract law operate efficiently, the written contract between insurer and policyholder
should normally avoid any significant legal risk.
However, care must be taken with regard to those areas of the contract where the insurer
has discretion. In this event, the insurer is at risk of:
 Some principle related to policyholders’ reasonable expectations acting

unfavourably to the insurer – for example, the flexibility it would like to adopt in
declaring differing levels of bonus across with-profits policies may be constrained.
A company may well be legally required to distribute profits in a way that is ‘consistent with PRE’.
However, PRE is very difficult to pin down in a legal sense, and interpretation might only be
settled by court rulings in particular cases. (Nevertheless, such things can and do happen, so they
can have as much effect in constraining insurance company behaviour as a more precise
requirement.) In any case, companies would not wish to be taken to court in the first place, and
they will always have a desire to keep their policyholders happy so as to improve their company
image and increase sales. In this way, a rather loosely defined PRE requirement might even have
greater constraining influence than might a more rigidly defined rule.
 Unfair contract terms voiding clauses of the contract – for example, if a contract
confers a right for the charges on unit-linked contracts to be increased by a fixed
annual rate which is considered unfair in the country concerned, the legal process
may act to void the clause. A particular risk is that the ability to review charges at all
is removed completely, rather than the clause being replaced with a percentage
increase which is acceptably lower than before.
In most cases it would be hoped that these kinds of issues would be sorted out during the
development of any product. So if legal requirements prevented certain kinds of charge increases
from being made, then the product design would not incorporate such features. Naturally this
puts a constraint on product design, which could ultimately mean that whole classes of business
might not be offered if insurers felt they were too risky.
A compounding difficulty here is that insurance contracts span many years and are open to
developing legal cultures, interpretations, and court judgments.
So one of the biggest risks here is that new legislation could be introduced that could change the
legal contract between the insurer and its existing policyholders. For example, it might prevent a
company from making further increases in charges on existing policies under which increases
could hitherto have been made.
Question
Explain why the above scenario is much worse than the case where charging rate reviews were
known to be not allowed before the policy was issued.

Solution
In the second case the company would issue the policy with much larger margins in its charging
rates from the start – ie the charges would be higher. In the first case the existing charges would
have much less margin – ie they would be lower – based on the (now false) assumption that
future charges could be increased were experience to deteriorate.
Beyond the written contract, legal risks may arise from inconsistencies between the policy
document and any other relevant representations made by the company or its agents.
For example, an intermediary may have inadvertently told a client that a policy had a certain
feature that, in reality, it did not possess. The client, having taken out (and maybe even claimed
under) the policy, later finds out the discrepancy and makes a complaint to the regulator. The
insurer may well lose money (and face) if the complaint were to be upheld.

4 Regulatory environment
4.1 Regulatory restrictions

Governments may impose restrictions on the way in which life insurance companies
operate. The aim of such restrictions is usually stated to be the protection of the
policyholder.
Historical experience has shown that insurance companies could not always be trusted to manage
their affairs appropriately without legislative control. People who take out long-term insurance
contracts are giving large sums of money to insurance companies, often over very long periods of
time, in return for a ‘promise’ by the company to pay the contractual sum assured at some (often
distant) future date. The public needs to have well-founded confidence that insurance companies
will still be in business and able to honour their obligations when policyholders claim. Without
this confidence people will be reluctant to buy insurance, and will therefore be denied this
essential financial service as a result (in other words, taken to the extreme, the industry will fail).
Countries have therefore found it necessary to regulate their insurance industries in order to
ensure the security of the policyholders’ interests (ie that the companies should not become
insolvent), thereby achieving the necessary public confidence in the industry to the benefit of
both parties (the public receive a reliable service and the insurance companies do business).
Although the restrictions will usually meet this aim, they may also have the effect of either
restricting innovation or reducing the benefits that could otherwise be given to
policyholders.
The following are the more common of such regulatory restrictions.
1. the types of contract that a life insurance company can offer
For example, in Italy there are six classes of life insurance product, based on contract type
– unit-linked business is one of these classes. Companies have to be authorised
separately for each class. Some companies may be authorised to write only one or two of
the six classes.
Another example is that life insurance contracts in South Africa have to be of term at least
five years. This was introduced to prevent competition with banks for short-term
contracts.
2. the premium rates, or charges, that can be used for some types of contract
Such restrictions have been common in many European countries and in some States in
the United States. The rates themselves might be restricted, or certain elements of the
premium rate basis, such as mortality and interest, might be controlled.
Question
Discuss whether restrictions on premium rates for term assurances would be expected to specify
upper or lower limits for the rates, if policyholder protection were the aim.

Solution
Theoretically, either would be possible. The authorities may feel that:

 rates that are too low might threaten the financial security of the company and hence the
security of its policyholders
 rates that are too high might give poor value for money.
(Depending on the product, the market and the distribution channel, the authorities might feel
that competition could take care of the second point.)
3. rating factors that can be used to calculate premiums, for example gender or age
The March 2011 European Court of Justice (ECJ) decision on an insurance opt-out
provision from the EU Gender Directive is a good example of restrictions on rating factors.
The insurance opt-out provision from the directive meant that, provided they met certain
conditions, EU insurance companies were allowed to use gender as a premium rating
factor. However, the ECJ ruled that this opt-out was not valid and should be removed,
with the result that from December 2012 EU insurance companies were no longer able to
use gender as a rating factor.
4. the terms and conditions of the contracts offered, for example with regard to how
paid-up policy and surrender values are to be calculated
Customers and regulators in a number of countries have been concerned about poor
value for money on early withdrawal from life insurance contracts. High initial expenses,
including up-front commission, are responsible for this problem. One possible solution is
for regulators to impose minimum values (for example specified percentages of premiums
paid). Alternatively, where policy values are calculated on a prospective basis, the basis
for the calculation could be specified or restricted.
Question
Discuss how much of a risk to the life insurer such minimum guaranteed surrender values would
represent.

Solution
It depends. The basic risk is that the asset share of the policy is below the guaranteed value,
which depends on such things as:
 how large the guaranteed values are
 how large initial expenses are
 the volatility of the assets underlying the policy
 how likely withdrawal is (which could depend on a host of other factors, such as how the
policy was sold, tax treatment on withdrawal and economic conditions).
5. the channels through which life insurance may be sold or requirements as to the
procedures to be followed or the information required to be given as part of the
selling process
For example this might involve:

● minimum training requirements for insurance salespeople
● the right of cancellation of contract by the policyholder
● the illustration of possible maturity values and surrender values, perhaps on
specified or restricted bases.
The UK, for example, has chosen to deal with the question of poor early surrender values
through a process of disclosure and competitive pressure, rather than direct control of
the values. However, as discussed under point 4 above, this is not the only possible
approach.
Where bases for illustrative values are not restricted, experience shows that (wildly)
over-optimistic projections can result.
6. the ability to underwrite, for example a prohibition on the use of the results of
genetic testing, or prohibition of use of past claims history or medical history
For example, in New Zealand anti-discrimination law prevents life insurers refusing to
insure anyone, whatever their state of health. However, insurers are allowed to charge
an appropriate premium for the risk.
7. the amount of business that may be written
In most countries there are regulations regarding the minimum level of mathematical
reserves that must be held, often combined with minimum solvency capital
requirements.
We have already considered this issue of capital strain in the first four chapters of the
course.
These regulations have the effect of limiting the capital available within a company
to write new business and effectively placing a minimum requirement on the finance
required to write a contract.

The capital that a company has available to write new business is, broadly speaking, the
supervisory value of its assets minus the supervisory value of its liabilities (including any
required solvency capital). This difference is sometimes referred to as the ‘free assets’,
‘free capital’ or ‘free reserves’ of the company.
If the company uses up more than the amount of its free assets in writing new business
(or any other way) it will be insolvent in the supervisors’ eyes. Quite simply, the larger the
value of supervisory reserves plus solvency capital that the insurance company has to
hold to satisfy the regulators on the existing in-force business, the smaller the free assets
and hence the capital available for writing new business.
What is more, the bigger the reserves plus solvency capital that have to be set up for any
new contract, the greater the amount of the limited capital available that will be used up
when the new contracts are written.
8. Investment restrictions are also to be expected, such as:
 the types of assets in which a life insurance company can invest
 the amount of any type of asset that can be taken into account for the
purpose of demonstrating solvency
 the extent to which mismatching is allowed at all.
For example, in some countries there is a requirement to hold minimum proportions of

total assets in certain types of investment, such as government stock or bank deposits.
Investment in assets perceived to be risky, such as equities, property and derivatives,
might be restricted (eg to x% of total funds or liabilities) or even forbidden.
An alternative is to allow largely unrestricted investment, but to permit only a certain

percentage of some asset classes to count when demonstrating statutory solvency. For
example, equities might only be allowed to count for 25% of assets in demonstrating
solvency. So a company with, say, 40m in equities and 60m in gilts would only be able to
show assets of 80m in its statutory balance sheet.
Another approach adopted in some countries focuses on preventing over-concentrations

in individual assets (eg the shares of one particular company), rather than in particular
asset classes. For example, a restriction that shares in any one company may amount to
no more than 1% of the assets used to demonstrate solvency.
Question
Outline how such restrictions might protect policyholders.

Solution
The restrictions could prevent or limit investment in very high-risk assets and so reduce the risk of
insolvency.
They could prevent over-concentration in a particular:

 country
 asset class
 stock-market sector
 individual asset
which again will help to reduce the risk to the policyholder.
The regulatory environment can also affect the choice of assets through their
relationship with the investment assumptions used to value the liabilities. A
particular asset selection may allow a company to use a higher investment yield
assumption and thereby reduce the value of the liabilities and increase the free
assets. Typically, however, such asset selections will not enable the company to
maximise the expected investment return.
For example, in some jurisdictions, the rate that has been used to value the liabilities has
been higher if they were backed with government stock than if they were backed with
equities.
Finally there may be a regulatory requirement to allow for mismatching. This could
involve the possible setting up of an investment mismatching reserve. The more a
company decides to invest in riskier assets with a higher expected return, the higher
could be any such resulting reserve. This would increase the value of the liabilities
and reduce the available free assets.
There may be regulations specifying what changes in conditions should be considered for
the purpose of determining a mismatching reserve, or such guidance might come from a
professional actuarial association or the regulatory authorities. Alternatively the decision
might be left to the individual actuary’s professional judgement.
Another regulatory constraint which might be important is the method required to value
the assets. For instance, in countries where assets are valued at book value, there is a
disincentive to invest in property because any increase in value in the properties could
only be seen after sale, and having to sell frequently would be expensive and
inconvenient. Even simple equity investment is less attractive in such a country, since
gains can only be seen on sale and there will be times when the company might not want
to sell.
(The subject of investment, including the principle of matching, is covered in detail later in
the course.)

4.2 Wider regulatory environment

The wider regulatory environment in which institutions are allowed to transact life insurance
business is also important. In practice, life insurance companies are likely to have the
monopoly of providing pure protection benefits, but not of providing savings benefits.
Other pooled savings vehicles are likely to exist, for example unit trusts and investment trusts in
the UK, or their approximate equivalents in the US, mutual funds and investment companies.
Direct investment in shares or other assets will also be an option for some individuals.
The other institutions offering savings contracts, for example banks, will usually be subject
to different regulatory controls from life insurance companies, leading to differences in the
terms upon which such contracts can be offered.
The differences may be to the benefit or detriment of insurance companies. Possible examples
include:
 different rules for demonstrating capital adequacy
 different rules on advertising of products
 different rules for benefit illustrations.
The regulatory environment is likely to have a significant effect on the design of the
contracts sold by life insurance companies, as the companies will want to make the best
use of any regulatory opportunities available to them. Conversely, contract design will have
to allow for any constraints imposed.
Question
The regulatory authorities in a particular country have just introduced a requirement for life
insurers to illustrate, to prospective policyholders, withdrawal values for each of the first ten
years of contracts.
Comment on how this change may affect contract design and benefits paid to policyholders on
unit-linked endowment assurances.
Solution
The most likely effect is that the life insurance company will have to improve early surrender
values, if these have been poor to date. Not to do so might discourage potential policyholders.
This might mean:

 reducing charges early in the contract, or
 reducing any penalties applied to unit funds on surrender.
All other things being equal, higher benefits paid to surrendering policies will mean lower benefits
for maturing policies. This would come about through a different structure of charges.
Policyholders who surrender will be charged less than before, and those who don’t, more.

5 Taxation regime
5.1 Approaches to taxation

Life insurance business may be taxed in a number of different ways. The most common
methods are:
 A tax on the annual profits of the business, where profits means the excess of the
change in the value of the assets over the change in the value of the liabilities.
 Tax payable on investment income / gains less some or all of the operating
expenses of the company.
In addition, there may be a tax on premium income.
One way of looking at the two above approaches is that they recognise different aspects of the
nature of a life insurer. The ‘profits’ approach recognises that a life insurer, at least if it has
shareholders, is a company trying to make a profit like any other.
The ‘investment return less expenses’ approach could be thought of as treating the insurer as a
group of individuals pooling their resources for investment. If the investment return of
individuals is taxed then it is logical that this return should be taxed if it is earned within a life
insurance company.
We shall now consider these two approaches further.
5.2 The profits approach

The profits calculation described above essentially measures taxable profit as the increase in the
free assets of the company over the year, where free assets is defined as the value of assets
minus the value of liabilities. (Any required solvency capital may or may not be added to the
value of liabilities in this calculation; this makes little difference to our discussion here.) Let us
define:
A0 = Assets at start of year
A1 = Assets at end of year
V0 = Liabilities at start of year
V1 = Liabilities at end of year
Then profit could be stated as:
1. (A1 – A0) – (V1 – V0) In symbol form, the definition of profit given in the Core
Reading above
2. (A1 – V1) – (A0 – V0) ie increase in free assets over the year

With an approach based on profit the focus is usually on shareholders’ profit. Profit distributed to
policyholders on with-profits policies may well be automatically excluded from the profit
calculation. For example, where policyholder bonuses increase the liabilities at the end of the
year (V1) the value of those bonuses would be excluded from the profit figure. Similarly, if cash
bonuses are paid out to policyholders these reduce the end of year asset figure (A1).
5.3 The investment return less expenses approach

The taxation rules will specify which types of investment income, investment gains and expenses
are included in the taxation calculation and any adjustments that need to be made.
For example, in the UK the investment return on equities and property, for life business tax
purposes, includes realised gains but not unrealised gains. In Ireland, on the other hand,
unrealised gains are notionally realised each year and are also included, albeit with the amount of
tax paid spread over seven years.
Allowing a deduction for expenses is reasonable because such expenses must reduce any returns
actually available to policyholders or shareholders.
Question
Comment on how sensible the ‘investment return less expenses’ approach would be as a basis for
taxing regular premium term assurance business.
Solution
It would be fine from the insurance company’s perspective - but the tax authorities might not be
so happy, as there is unlikely to be any tax generated. The funds building up under term
assurance are small, as we explained earlier in the course. This means that very little investment
return is earned and may well be smaller than expenses (depending on how much of an offset for
expenses is allowed).
But the business may still be profitable, and so a basis that produces no tax might not be
considered suitable by the tax authorities, particularly for companies with shareholders.
To get a full picture of the impact of taxation on investment return one must consider the total
effect of taxation within the life insurance company and on the benefits received by the
policyholder.
For example, in the UK it is broadly the case that policyholder and shareholder life insurance
funds (other than pension funds) suffer tax on investment return within the company, but
policyholders are not taxed on the benefits they receive. In Continental Europe it is much more
common for funds attributed to policyholders (as opposed to shareholders) to receive gross
returns. However, the benefits paid to policyholders may then be taxable.

5.4 The effect of the taxation regime

Within a country, different types of life insurance business may be taxed by different
methods.
‘Different types’ here could refer to various possible classifications of business. For example,
business classed as ‘pensions’ might be taxed differently from other business. Other possibilities
could include different treatment of with-profits and unit-linked business, or of ‘savings’ and
‘protection’ policies.
This can mean that it is at a lower cost for the consumer if certain forms of benefit can be
offered as one type of business rather than another.
For example, in the UK it was possible to offer term assurance as pensions business (subject to
certain limits) as well as life insurance business. Because of the tax concessions on contributions
to pension funds the term assurance cover could be offered more cheaply.
The taxation treatment of life insurance business may make life insurance more or less
attractive as a savings medium than contracts offered by other savings institutions, subject
to a different fiscal regime.
Tax concessions available to individuals may make the sale of certain types of contract
easier.
The tax treatment of policy proceeds when they are received by policyholders can distort
buying habits.
The last three paragraphs of Core Reading illustrate the point that the overall attractiveness of a
life insurance product compared with other products (be they from life insurers or other savings
institutions) depends on a combination of:
1. the taxation treatment of premiums paid, in particular whether the premiums are
deductible from the individual’s taxable income in full, in part or not at all
2. the taxation of the life insurer’s funds during the life of the contract
3. the taxation treatment of the eventual policy benefits.
Example
In a certain country, taxation of life insurance policies was known to be as follows:
 premiums are tax deductible up to a certain amount
 premiums incur a 2½% premium tax
 investment return is gross
 the benefit is taxed at 12½% of excess of benefit over premiums paid.
Overall, this represented a subsidy on savings effected in the form of a life insurance policy
compared with other savings channels such as banks or direct investments.
(As a result, many banks established their own life insurance subsidiaries in order to offer their
customers these advantages over normal bank deposit account savings.)

As with the regulatory environment, product design will want to make the best use of any
opportunities provided by the fiscal environment. On the other hand, the ability to
maximise favourable taxation treatment may force constraints on product design.
For example, tax authorities may be keen that pure savings business should not be ‘dressed up’ as
life insurance in order to secure favourable tax treatment where this exists. If so, they might
therefore specify minimum levels of life cover necessary to secure tax concessions.
This would then represent both an opportunity and a constraint. The tax concessions might allow
a competitive product to be produced, but the design would have to include at least the minimum
level of life cover.
It should be noted that both the taxation and regulatory environments are very significant drivers
of product design in the insurance industry.
Example
In South Africa, products called ‘Deferred Compensation’ are offered. Employers buy these to
provide benefits to key employees at retirement or at the end of a specified term. The contracts
give the employees an incentive to stay with the company over the long term. The premiums are
a tax-deductible expense for the employer, but only if the contracts satisfy the following
conditions:
 only one life insured
 maximum 15% pa premium escalation
 minimum life cover equal to 80% of annual premium multiplied by term.
Taxation risk
Taxation changes over time, so it is important to bear this in mind when benefits are
guaranteed over the long term.
Existing policies are usually not immune from the effects of changes in taxation (methods or
rates), so the company has a risk of making less (net) profit than anticipated from any policy as a
result of future taxation changes.

6 Professional guidance
Actuarial associations will often issue professional guidance for actuaries advising life
insurance companies. These will typically give such actuaries a framework of points that
they need to consider in carrying out their responsibilities so as to maintain professional
standards.
The extent to which this is considered necessary will depend on the extent to which actuaries’
scope for judgement is limited by legislation, which will vary from country to country.
As such, professional guidance should not unduly restrict the actions of actuaries and may
even provide protection against pressure from proprietors to agree to courses of action that
may not be in the best interests of policyholders.
For example, planned new business levels might require more capital than the company has or
can raise, threatening the long-term survival of the company. The actuary would presumably be
arguing for a change in the plans in any case, but the existence of professional guidance on the
matter may strengthen his or her hand in discussions with the proprietors.
Actuarial associations may also issue professional guidance on the interpretation of

government regulations. This typically arises where the government does not want to be
overly prescriptive in its requirements so as not to restrict unduly the actions of life
insurance companies. The government will then look to the actuarial profession to set
limits, as it were, on what would be acceptable behaviour.
To the extent that professional guidance adds safeguards as to the proper running of life
insurers, it effectively provides an opportunity for insurers to encourage consumers to
invest in life insurance products. Such products should then be lower cost and more
flexible as a result of an appropriate, but not burdensome and costly, regulatory regime.
Professional guidance and regulations are both ultimately designed for the same end, as we
described in Section 4: to generate consumer confidence and hence create a thriving life
insurance industry for the benefit both of consumers and providers.


Chapter 9 Summary
Expenses and commission
Commissions may be initial and/or renewal. Other expenses can be categorised into initial,
renewal and termination expenses.
Initial expenses and commission are usually the largest expense categories.
Overheads can be any expenses that are not specifically related to a particular policy activity.
More often they are the ‘fixed’ expenses, that do not vary with volume of business within a
given scale of company operation.
Inflation is a key element of the uncertainty and risk associated with future expenses.
The main different expenses are salaries, property, computer and investment expenses.
Many are linked in some way to earnings inflation. Price and (particularly) earnings indices
can be used to help predict the relevant inflation rates.
New business volume is more volatile than in-force volume or claim numbers. New business
may fluctuate in line with economic cycles. Initial, renewal and claim expenses reflect this
pattern of volatility.
Economic environment
The available assets will influence investment policy and interest rate assumptions.
Volatility of investment markets will increase cost of capital for insurers, so increasing policy
prices. This can also affect the demand for life insurance products.
Legal environment
Life insurance companies’ operations will be influenced by PRE. PRE can have legal force
through court rulings, so it can place considerable restrictions on companies.
General contract legislation may outlaw certain policy conditions, so restricting a company’s
ability to manage risk and/or issue certain product designs.
New legislation that has retrospective force could be a serious problem.
Any inaccuracies in the information given to the policyholder regarding contractual terms
can lead to a risk of consequent legal action against the company.

The regulatory regime

Governments may impose restrictions on life insurance companies, usually with the stated
aim of protecting policyholders. The more common restrictions are:
 the types of contract that a life insurance company can offer
 the premium rates, or charges, for some types of contract
 rating factors that can be used to calculate premiums
 the terms and conditions of the contracts
 the channels through which life insurance can be sold, sales procedures or
information given at the point of sale
 the ability to underwrite (eg prohibition of the use of genetic test results)
 the amount of business that may be written, via minimum reserving or solvency
capital requirements
 the types of asset or the amount of any particular asset in which the life company
may invest for the purpose of demonstrating solvency.
The wider regulation of different sales media also has an impact on life insurance companies
and the products that they sell.
The taxation regime

The most common approaches to life insurance company taxation are:
 a tax on the annual profits of the business
 tax payable on investment income / gains less some or all of the operating expenses
of the company.
Different tax treatment of different types of life insurance contract can make it cheaper to
provide some benefits as one form of business than another.
Different tax treatment of life insurance business and other forms of saving may create
opportunities for a life insurer, but may also impose restrictions on contract design.
In comparing the tax advantages of different products one must consider:

 the taxation treatment of premiums paid, in particular whether the premiums are
deductible from the individual’s taxable income and whether there is a premium tax
 the taxation of the life insurer’s funds during the life of the contract
 the taxation treatment of the eventual policy benefits.
Professional guidance
Actuarial associations will often issue professional guidance for actuaries advising life
insurance companies. This will typically give actuaries a framework of points to be
considered in carrying out their responsibilities in order to maintain professional standards.
Actuarial associations may also issue professional guidance on the interpretation of

government regulations.


9.1 (i) State the most common types of regulatory restriction that are imposed on life insurance
Exam style
companies. [6]
(ii) Comment on how the possibility of changes to taxation and the taxation system might
represent a risk to a life insurer. [5]
[Total 11]
9.2 (i) List four different types of expenses incurred by life insurance companies.
(ii) Explain the difference between management expenses and overhead expenses, giving
one example of each.
9.3 An actuary has been asked by a non-actuarial colleague about why an actuarial association might
Exam style
issue professional guidance to actuaries advising life insurance companies.
Set out the points that should be mentioned in response. [5]
9.4 A life company has recently changed from paying high initial commission on its regular premium
Exam style
contracts to paying a lower rate of level commission throughout the term without altering the
charging structure of the products.
Describe the financial risks to the company of this change in policy. [4]



Chapter 9 Solutions
9.1 (i) Types of regulatory restriction
The following are common regulatory restrictions:

 Restrictions on the type of contract that a life insurance company can offer. [½]
 Restrictions on the premium rates or charges that can be used for contracts. [½]
 Requirements relating to the terms and conditions of the contracts offered, for example
specifying how paid-up policy and surrender values are to be calculated. [½]
 Restrictions on the channels through which life insurance may be sold, … [½]
… or specified procedures which must be followed, … [½]
… or information which must be given as part of the selling process. [½]
 Restrictions on the ability to underwrite, for example a prohibition on the use of gender
as a rating factor or on the use of the results of genetic testing. [½]
 An indirect constraint on the amount of business that may be written. [½]
This arises from minimum requirements for reserves and solvency capital, which decrease
available capital and increase the capital needed to fund new business. [1]
 Restrictions on the types of asset … [½]
… or the amount of any particular asset in which the life company may invest, … [½]
… or which it can include for the purpose of demonstrating solvency. [½]
[Maximum 6]
(ii) Taxation risks
A life insurance company is at risk from unexpected changes in taxation methods ... [½]
… or in rates, … [½]
… which may increase its tax bill above what it had been assuming. [½]
For example there might be a change from taxing on profits to investment return (or vice-versa)
or the introduction of a tax on premiums. [1]
Favourable taxation of certain life assurance contracts compared with others may be reduced or
removed. [½]
Any of the above points could render profitable / attractive contracts unprofitable / unattractive.
[½]
Products may then have to be redesigned or even scrapped, causing wasted or additional
expense. [½]
Similarly, changes to taxation treatment of other savings institutions or their products might
suddenly put a life insurer’s products at a relative disadvantage. [½]

Changes to personal taxation of benefits or tax relief on premiums could have a similar effect. [½]
The extent of the risk also depends on whether or not changes may be made retrospective. [½]
If the effect is an increase in withdrawals and/or a fall in sales then this might increase per-policy
expenses. [½]
The withdrawals might also be selective, further reducing profits. [½]
Changes to methods of accounting for / collecting tax might increase the company’s expenses and
reduce profits. [½]
[Maximum 5]
9.2 (i) Types of expenses
The four different types covered in Chapter 9 are initial, renewal, termination, and overheads.
However, it would be legitimate to have included investment expenses instead of overheads, as

this combination of expense types is mentioned later in the course under the heading of
experience monitoring.
(ii) Management expenses vs overhead expenses
Management expenses consist of expenses that are incurred as a direct consequence of the
existence of a policy. They occur when policies are written (new business), maintained (in-force
business) or terminate (eg death, surrender, maturity).
Examples include: printing and postage costs associated with policy documentation, claims, etc;
cost of obtaining an independent medical examination; commission.
An overhead expense is one that will occur regardless of whether particular individual policies
exist or not. So these costs will be invariant of the numbers of policies issued / in force /
terminating.
Examples include: most salaries, except possibly for temporary staff or others on very short-term
contracts; rental of the company’s office premises; accounting costs.
9.3 The main aim is to give actuaries a framework of issues that they need to consider in carrying out
their professional responsibilities. This helps to ensure consistency and maintain professional
standards. [1]
This is essential in order to maintain a high degree of trust between the public and the insurance
companies, which in turn should encourage individuals to take out the insurance protection that
they need. [1]
The guidance can also add weight to the actuary’s arguments if he or she needs to resist pressure
from proprietors in order to protect the best interests of policyholders. [1]
Actuarial associations may also issue guidance on the interpretation of government regulations.
In fact, the legislation may be framed with the expectation that the actuarial association will set
out an appropriate interpretation. [1]

Guidance should be less onerous and expensive than regulation. This should make life insurance
products lower cost and more flexible on account of enjoying an appropriate, but not undue and
over-costly, regulatory regime. [1]
[Total 5]
9.4 The basic financial effect of the change is to increase first year profits and decrease profit in all
subsequent years.
There is a risk of withdrawals and paid up rates being lighter than expected (so that the cost of
the renewal commission is higher than expected). However this may benefit the company since
overall profitability may improve if withdrawals are low. [1]
There is a similar risk of lighter than expected mortality, but again this could benefit the
company. [½]
There is a risk that intermediaries will prefer initial commission, so own salesforce staff and tied
agents might leave the company. Brokers may recommend products of other companies, so
overall volumes (and hence profits) will fall. [1]
The company will be making an assumption regarding the rate at which the renewal commission
will be discounted. The company might not achieve this return. [1]
If the tax rates change in the future after the policies have been issued (so that tax on profits
falls), then the profit to the company will less under the new structure than under the old. [½]
Similarly, if tax relief on expenses is removed (if applicable), then the profit will be less under the
new structure than under the old. [½]
[Maximum 4]





using the course.

SP2-10: Risk (1) Page 1
Risk (1)
Syllabus objective
 mortality rates
 persistency
 competition
 counterparties
 fraud
(The items in italics are covered in this chapter.)

Page 2 SP2-10: Risk (1)
0 Introduction
This chapter and the following two chapters look at the macro risks with which a life
insurance company might be faced.
In this chapter and Chapter 11 we shall look at the possible sources of those risks. In Chapter 12
we shall consider how the individual risks combine and their overall effect on the way the
company is managed.
We have already spent some time discussing risks earlier in the course, in the context of specific
contracts. We identified those risks as ‘micro’ risks, but they might equally well be thought of as
specific examples of some of the ‘macro’ risks covered in this chapter.
For example, uncertainty over future investment returns is a general risk for a life insurer and so
we shall be discussing it here. However, we have already seen that its importance as a risk varies
depending on the type of contract involved.
What we shall be doing in this chapter is looking in more detail at how and why our assumptions
about the future may turn out to be incorrect. Indeed, the general theme of much of this chapter
and the next is:
1. Actuaries are constantly involved in investigations that require them to make

assumptions about a wide range of future events. These investigations form the basis of
the company’s strategic decision-making (for example for pricing, product design,
investment strategy, reinsurance strategy, etc).
2. For various reasons reality will turn out differently from assumed (ie from expected). This
will mean that the decisions the company made on the basis of its assumptions may no
longer be appropriate (for example, with hindsight, it may have adopted a different
strategy from the one actually adopted on the basis of foresight only). The use of the
‘wrong’ decision will lead to less desirable outcomes (eg of profit and solvency) and the
extent of this deviation from the desired outcomes is what is ultimately meant by ‘risk’.
Hence each assumption represents a source of risk.
We shall also consider some risks that do not fit into this framework, for example the effects of
competition and the management of the company.
Although some of the risks are specific to life companies, other risks are more general to
financial services companies or simply to companies operating in free, well-regulated,
economies.
The identification of risks is an essential part of a life insurance company’s control cycle. Any
business will wish to identify threats to its financial well-being and profit stream.
Having identified the risks, the company can put in place mechanisms to monitor them. The
company can also do a number of things to limit and control the risks that it faces. In this part of
the course we are still concentrating primarily on the sources of risk rather than the control and
monitoring. These other aspects of the control cycle will be covered later.

1 Policy and other data
1.1 Policy data

Having high quality policy data is vital for the good management of an insurance company.
A life insurance company will maintain a database of the policies it has on its books. For each
policy this will include items such as age at entry, policy term, duration in force, sum assured and
many others, dependent on policy type.
One important use of policy data is to carry out a regular valuation of the liabilities, usually,
but not exclusively, for insurance supervisory purposes. These data will be maintained by
the life insurance company and the actuary carrying out a valuation will not, usually, have
direct control over them. There is, therefore, a risk that the company will not maintain the
adequate, accurate and complete records required by the actuary and hence the result of
the valuation will not itself be accurate.
Question
Discuss how serious these data problems could be.
Solution
Potentially, this could be extremely serious. Clearly this depends on the size and nature of the
error. In the extreme, an insurance company might declare itself to be solvent when it was
insolvent, or vice versa.
Less serious errors could still lead to incorrect decisions, for example through underestimating the
capital available to fund new business.
A similar point relates to policy data required for any internal investigations carried out by
the actuary in order to give appropriate advice to the company.
Clearly, if the data are so inaccurate as to lead the actuary to give incorrect advice then this is
potentially very serious, both for the company and the actuary. Therefore the actuary would
normally carry out some checks on the quality of the data. It is imperative that the actuary is at
least aware of any data problems.
For some investigations, model points representing all or part of the business may be used.
There will be a risk that these model points do not adequately represent the underlying
policy data.
It may not always be practical or cost effective to carry out an investigation on all the policies on a
company’s books, or even for a particular class of policy. A large insurer in a major market might
have millions of in-force policies. So the insurer may instead choose to use model points.

For example, all without-profits endowment assurances sold to males, of original term
twenty-five years and duration in force of eleven years, might be represented by a single
specimen policy (this is the model point mentioned above) with the average sum assured and
policyholder age. Results from these model points can then be scaled up to give a result for the
full portfolio.
The insurer’s model might use only a few hundred model points or tens of thousands, depending
on the purpose and the available computer power. The fewer policies used, the less accurate the
representation of the in-force business. This introduces scope for errors in the results and for
incorrect business decisions.
Even if the model points are an accurate representation of the policy data, the insurer is still
exposed to model or modelling risk, ie the risk that the model is an over-simplification of the real
company and, as a consequence, behaves unrealistically. We return to the subject of model risk
in Section 2 of this chapter, and to modelling in general in Chapter 14.
1.2 Other data

Other data useful to actuarial investigations might include information on asset holdings
and external data.
The latter may be needed to determine actuarial assumptions such as those for mortality.
External sources include insurance industry data, reinsurer’s data, national statistics or an
overseas market.
It may be that the data internal to the company will not be adequate to provide reliable
indicators of future experience, because of inaccuracy and/or insufficient volume. Even
where adequate, data may prove to be inappropriate because of differences between the
population on which it was based and the population for which it will be used.
Internal data might be inadequate for various reasons, for example:

 there may be missing data
 the data may be inaccurate
 there may not be enough data for credibility.
Even if the data are of good quality, they may not be suitable as a basis for future assumptions
(unless they are significantly modified), because they may be based on an inappropriate class of
lives.
Question
A life insurer has written large numbers of with-profits endowment assurances over many years.
It has performed a careful analysis of the average expense per policy of this business. It is about
to start writing without-profits term assurances for the first time, and it plans to use the
per-policy expenses from its endowments as its expense assumption for the term assurances.
Suggest reasons why this might not be appropriate.

Solution
The expenses of the term assurances are likely to be different from the endowments, for a
number of reasons:
 More extensive underwriting on term assurances will mean higher initial expenses.
 However, the simpler nature of the contract may reduce ongoing expenses.
 The mix of business by distribution channel may differ.
 Commission payable for the contracts might be different.
We must allow for inflation that has occurred since the time period that has been analysed …
… and for some future inflation, if the premium rates are to be used unchanged for some time.
Per-policy expenses will also be affected by relative business volumes.
The insurance company’s expenses may have changed over time and may be expected to change
in the future for any number of reasons, for example greater efficiency.
Alternatively, any external data used may be inadequate or unreliable, and even where it is
reliable may prove to be inappropriate for the same reason as above.
That is, due to differences between the population on which it was based and the population for
which it will be used.

2 Mortality
In most of the investigations carried out by an actuary, assumptions will need to be made
regarding future mortality. There are three risks associated with these assumptions:
 a ‘model’ risk that the model, typically a probability distribution, chosen to represent
future mortality, etc, may not be appropriate or may contain errors
 a ‘parameter’ risk that the parameters used with the model may not adequately
reflect the future experience of the class of lives insured or to be insured, even
though the underlying model may be appropriate
 a ‘random fluctuations’ risk that the actual future experience may not correspond
with the model and parameters adopted, even though these adequately reflect the
class of lives insured or to be insured.
The first two risks always exist, as actuaries cannot predict the future with complete
certainty. However, the extent of the risk will differ according to the reliability and
applicability of any existing data.
The risk here is that mortality turns out to be more adverse than assumed in the models we have
used for decision-making. The three types of risk listed above help us to understand the various
ways in which the risk can arise. This in turn can help us to devise suitable means of controlling
the risk. In other words, understanding where the risk has come from is half of the battle itself.
There are no dogmatic rights or wrongs about how much of the risk associated with a particular
variable is due to model, parameter or random fluctuations risks. It all depends on how the
variable is expected to behave and how it is modelled.
Example
As an example, we might assume that a group of 1,000 policyholders all aged 40 will suffer
mortality of q40  0.003 over the next year (in other words, we are assuming that 3 of the 1,000
people die during the year). How would we assess the risk of the actual number of people dying
being different from this?
To do this, we need to specify the model we have made, more precisely. We could suppose that
the policyholders are identical and independent (a typical statistical assumption), and define the
probability of each dying as 0.003. If these assumptions are correct, then the number of deaths
will follow a binomial probability distribution, with parameters of 1,000 and 0.003. This will
enable us to quantify the random fluctuations risk of this portfolio, because we can predict such
things as the probability of more than 10 of the 1,000 lives dying during the year solely as a result
of the natural, or random, fluctuations inherent in the system.
But there are other, possibly more significant reasons, why we might get more deaths than
expected. Probably the most important is the parameter risk: for example, we might have lives
who are identical and independent but whose actual probability of dying is 0.006. This will
significantly increase the probability that more than 10 of the people die, for example.

The uncertainty surrounding the parameter value of this model (ie the probability of dying, q40 )
can itself be thought of in terms of a probability distribution or, more prosaically, as an expanding
funnel of doubt. This can be thought of as an expanding funnel because, if we were trying to
predict the value of q40 in ten years’ time, rather than next year, say, then we would consider a
broader range of values of q40 to be possible. (In other words, the variance of the parameter
q40 would increase the further into the future we are trying to predict it.)
The other main source of risk is the model risk. For example, if our lives were not independent,
then the variance of the number of deaths would increase. This is because the binomial model is
only valid if the lives are independent with respect to their mortality, so the random fluctuations
risk should actually be higher than we would estimate using the binomial model. Further error
would arise if our population were heterogeneous – in other words, the lives were not identical
with respect to their mortality.
As the Core Reading says, the likelihood of our assumptions for the future being right depends on
the quality and relevance of the data available. In other words, poor or irrelevant data will
increase the parameter risk. At one extreme we could be dealing with a well-established type of
contract, with reasonably stable and predictable experience, and lots of relevant data on which to
base assumptions. This would be low risk. At the other extreme we might be dealing with a
contract type that is completely new to our market, and have to rely on a mixture of intelligent
guesswork and heavy adaptation of what little relevant data exist. This would be high risk.
However good the past data, the possibility of new diseases or sudden advances in medical
treatment can never be ruled out. In other words, there will always be an expanding funnel of
doubt, however good the experience data may be.
The ‘random fluctuations’ risk is most likely to arise if the numbers exposed to risk are not
large enough for the ‘law of large numbers’ to apply.
The idea here, that unpredictable fluctuations are larger in small samples, should need little
further explanation. But even in large samples there will be some random fluctuation, and so
there will always be random fluctuations risk.
Question
Discuss the extent to which mortality represents a risk on without-profits term assurance, for an
individual insurance company operating in an established insurance market in a country with a
developed economy.

Solution
Mortality is the most important element in the basis and so is a significant risk.
On the other hand, in an established market there is likely to be good data on which to base
assumptions. Also, population mortality has been reasonably stable, or at least subject to steady
trends, in developed countries. So there should be low model and parameter risk.
Even so, the possibility of a new disease represents a risk that is difficult to quantify.
The risk to a particular insurance company will also depend on the portfolio size. The bigger the
portfolio, the bigger the potential loss. On the other hand, a larger portfolio should reduce
random fluctuations risk.
The risk will also depend on any steps the insurer has taken to reduce the risk, but we shall be
looking at this aspect later in the course.

3 Investment performance
In most of the investigations carried out by actuaries, assumptions, either on a
deterministic or stochastic basis, will need to be made as to the rate of return, allowing for
both income and capital appreciation, on amounts to be invested in the future. Both ‘model’
and ‘parameter’ risks can, therefore, arise.
Question
List possible sources of ‘amounts to be invested in the future’, as referred to in the Core Reading.
Solution
Possibilities are:
 reinvestment of proceeds from the assets
 future premiums on existing business
 future premiums on new business.
Traditionally, actuaries have made deterministic assumptions about future investment returns.
For example, a fixed rate of 6% pa may be assumed. The effect of possible variations has been
explored by sensitivity testing.
Stochastic modelling is now an alternative, under which the investment returns on asset classes in
any given period are modelled as the outcomes of random variables with specified probability
distributions. Such projections can easily be performed on a computer by allowing random
drawings to be made, from the assumed probability distributions, using the computer’s random
number generator. The method is that of Monte Carlo simulation, as described in earlier subjects.
We give an example of the use of such a model in Chapter 28.
A key advantage of a stochastic approach is that it enables us to attempt to model the inherent
volatility (ie random fluctuations risk) in some asset classes, notably equities. Hence, with
stochastic modelling we have to choose not just the mean returns but the complete probability
density function. Furthermore, we may wish to specify relationships between the returns on
different asset classes, or limit the extent to which returns may vary from one year to the next.
In other words, when we make projections using a stochastic model, more aspects of the
probability distribution of the variable being modelled are included. Whatever parameters are
required for the projection (which might include one or more parameters for each of the mean,
variance, skewness, correlation with other variables, etc) then there will be a parameter risk (or
uncertainty) associated with each one.
As we have seen, ‘model’ risk, in the most general sense, is the risk that the model is
inappropriate. No model can reflect perfectly the complexity of the real world. So perhaps it is
better to say that the model is not a sufficiently accurate representation of reality to give reliable
results.

Examples of an error in a stochastic investment model might be:

 The probability density function chosen is inappropriate, for example normal when it
should have been log-normal.
 The time-series relationships between outcomes at different times are not specified
appropriately.
When we use a deterministic approach we are essentially using just one parameter, the expected
value. The range of values we use in sensitivity testing says something about our view of
underlying variability, but we do not really model this explicitly.
The separate projection of income and capital gains is common. One reason for this is that the
tax treatment of the two may differ.
Some investigations will require assumptions, again either on a deterministic or stochastic

basis, as to the future income and capital appreciation in respect of all, or parts of, the
existing assets of the company. Again, there can be ‘model’ and ‘parameter’ risks.
Investigations involving a comparison of the value of the liabilities with the value of the
assets will be subject also to a capital values risk in respect of the existing assets being
valued.
Investigations of this sort tend to be of a ‘snapshot’ nature, at one point in time. The risk is that
an insurance company that appears solvent on one day could become insolvent the following day
after a change in asset values.
Question
Explain whether ‘change’ should be ‘fall’ in the last sentence.
Solution
No. A fall or a rise could lead to insolvency if liabilities are valued at market rates. If assets were
invested with a shorter discounted mean term than the liabilities, then on a fall in interest rates
the value of the liabilities would rise by more than the value of the assets.
Another way of looking at this is that if interest rates fall then the proceeds from the assets
(which on average are available earlier than required), can no longer be reinvested at the rates
needed to meet the liabilities.

4 Expenses, including the effect of inflation

In most investigations carried out by actuaries, assumptions will need to be made as to the
future expenses to be incurred by the company. These will allow, implicitly or explicitly,
either on a deterministic or stochastic basis, for inflation. There is therefore a ‘parameter’
risk. In addition, there will be a ‘model’ risk if stochastic assumptions are being used.
When a stochastic approach to investment returns is used it is normal to model inflation

stochastically. In fact, the model may be ‘driven’ by a stochastically generated inflation rate, with
investment returns linked in some way to the projected inflation. The inflation of life insurance
company expenses may not be the same as general price inflation in the economy, but can be
linked to it in the model.
The company will include amounts designed to cover its expenses, either explicitly or
implicitly, in the premiums that it charges to policyholders. These charges may be
augmented by others made, again either explicitly or implicitly, against the funds built up by
contracts.
On non-linked contracts the charges for expenses may be invisible to the policyholder, as they are
loaded within the quoted premium rate. Sometimes, a fixed policy fee to cover some part of the
expenses will be quoted explicitly.
On unit-linked contracts, as we saw earlier in the course, charges are explicit. They may be
deducted from premiums (eg by allocating less than 100% of the premiums to units) or from the
fund (eg a regular policy fee paid for by unit cancellation).
It is not likely to be possible to design the charges (implicit or explicit) so that they correspond
exactly to the expenses of administering the product, and change in line with any changes in
expenses.
There is, therefore, a risk that the charges accruing to the company in a year will not cover
the actual expenses of the company in that year.
A particularly important example of this is when heavy up-front expenses are recouped gradually
by charges made over the term of the policy. The insurance company may then be vulnerable, for
example, to the policy withdrawing. Another problem exists on policies of long duration, where
higher-than-expected inflation may mean that the original expense loadings prove to be
inadequate.
Another very important aspect of expense risk involves the influence of business volumes on the
ability of the company to cover its fixed expenses (which we defined earlier in Chapter 1).
So we can see that expense risk can arise in many different ways and this can affect the way that
the risk is classified. There is no single unique way to classify risk, but a common approach is
described below.
The risk that actual expenses are higher than expected or allowed for is classified as
expense risk.
This can be due to the effects of inflation, but it is not the only factor at play.
For example, actual expenses could be higher than expected due to inefficient new processes.

The risk that the charges received are lower than expected or allowed for is more typically
classified as other types of risk, depending on the underlying driver of the shortfall. For
example, it may be
 investment performance risk (eg if the charges are fund-based and the shortfall is
due to low fund values because of poor investment returns)
 persistency risk (eg if charges required to recoup initial expenses are not received
due to high withdrawal rates)
 new business mix or volume risk (eg to the extent that charges are linked to average
size or volume of new business and this is lower than expected).
New business mix and volume are considered further in the next chapter.

5 Persistency
In most investigations carried out by actuaries, assumptions will need to be made as to the
future persistency rates under the contracts issued by the company. This could include full
withdrawal (surrender or lapse) rates, partial (or income) withdrawal rates and paid-up rates,
as applicable.
So persistency risk is the risk that the number of withdrawals (including partial withdrawals and
paid-up policies) is different to that expected. For this reason, persistency risk may also be
referred to as withdrawal risk.
There is, therefore, a ‘parameter’ risk and there may also be a ‘model’ risk.
For example, we might have a model for withdrawals that links withdrawals to stochastically
generated investment returns (perhaps as a proxy for general economic conditions). We might be
mistaken in our view that withdrawals would be linked to investment returns in this way. This
would be an example of model risk.
Even if the structure of our model is correct we might get the parameters wrong, for example by
generally underestimating the level of withdrawals.
One can also imagine random fluctuations risk affecting withdrawals, especially within small
portfolios.
The mortality parameters mentioned in Section 2 will be based on a model of the selective
effect of withdrawals. Departures from the latter may invalidate the former.
Question
Explain the last paragraph of Core Reading.
Solution
Selective withdrawals on assurance contracts are a risk because they leave behind a
‘sub-standard’ group of insured people. Insurers know this and so can allow for the expected
effect in pricing.
However, the risk of more selective withdrawals than expected is still there. If this happened then
it would lead to worse experience for mortality than had been assumed.
The parameters for future expenses, mentioned in Section 4, will be based on a model of
future withdrawals. Departures from the latter may invalidate the former.
Higher than expected withdrawals mean that fixed costs have to be spread over fewer policies.
This may produce higher per-policy expenses than the company had allowed for in its pricing.
Also such departures may be a cause of the mismatch, as mentioned in Section 4, between
the charges accruing to the company and its actual expenses.

Question
Explain the previous sentence.
Solution
If a company is recouping initial expenses gradually over the term of a contract then there is a
mismatch in the timing of income and outgo. The amount of charges to recoup the initial
expenses will have been set, when the contract was priced, on the basis of assumed rates of
future withdrawal. Higher than expected withdrawals would then make the future income from
these charges inadequate to repay the initial expenses.
It may be possible to counter this at some durations by giving the policyholder a surrender value
low enough for the insurance company to recoup its expenses, and perhaps even make its required
profit.
In summary, the key things about persistency risk are:

 the financial risk that the surrender value is higher than the asset share at the time of
withdrawal (this risk will often be made worse by the mismatching of the initial expenses
and the charges made to recoup those expenses)
 the risk to the mortality experience due to the selective effect of withdrawals
 the risk of increasing the per-policy fixed expenses due to the loss of business volume
from withdrawals.

Chapter 10 Summary
In most investigations, actuaries have to make assumptions about the future. Because we
cannot predict the future with certainty, there are risks attached to making these
assumptions.
These risks may often be usefully classified as follows:

 model risk – the risk that the underlying model is not appropriate or may contain
errors
 parameter risk – the risk that the parameters assumed for the underlying model are
incorrect
 random fluctuations risk – the risk of unpredictable fluctuations arising from sample
error; this is greater the smaller the sample.
The following are some of the risks faced by a life insurer:
Policy data
Inadequate, inaccurate or incomplete policy data could lead to incorrect results and
recommendations in actuarial investigations, including those performed for the supervisory
authorities.
Where model points are used, they may not represent the in-force policy data accurately
enough.
Other data
Data used in the formulation of actuarial assumptions may be inadequate. Even where they
are adequate in themselves, they may not be applicable to the purpose for which the
actuary requires them.
Mortality rates
The future parameters can never be predicted with certainty, because of, for example, new
diseases and advances in medical treatment. However, where there are good quality and
relevant data this reduces the risk of errors in parameter estimation.
A random fluctuations risk also exists.

Expenses
There is a risk of higher than expected expenses.
Higher than expected inflation of expenses leads to an expense risk. This can have model,
parameter, and random components.
Contributions to expenses from premiums and charges may be significantly mismatched with
the actual expenses incurred, over time.
Expense risk can be related to investment risk, persistency risk, or new business mix and
volume risk depending on the cause.
Investment return
Investment return assumptions may be needed for the existing assets or future investment
(and probably both).
Investment return might be modelled deterministically or stochastically. In either case there

are model, parameter and random fluctuations risks.
In a stochastic approach the random fluctuations are modelled explicitly. In a deterministic

approach this risk can be explored through sensitivity testing.
There is also a ‘capital values risk’ in any investigation comparing existing assets with
liabilities.
Persistency
The unpredictability of withdrawals makes it a risk. Early withdrawals usually represent a
particular risk of loss, because of the failure to recoup initial expenses and asset shares can
be very low or negative as a result.
If withdrawal experience is different from expected it may invalidate assumptions about the
effects of selective withdrawals on mortality experience. Expense assumptions may also be
invalidated due to the effect of withdrawals on the volume of business in force.


10.1 Explain how variations in levels of withdrawals on a portfolio of term assurance contracts might
Exam style
adversely affect the experience for mortality and expenses on that portfolio. [5]
10.2 A life insurance company sells only term assurances. It sells through both the insurance
Exam style
intermediary and own salesforce distribution channels.
(i) Explain how persistency can be a source of risk to this life insurance company. [4]
(ii) Discuss the differences in persistency experience for the two distribution channels. [2]
[Total 6]
10.3 The insurance supervisory authorities in Mesopotamia have decided to introduce minimum
surrender values on regular premium with-profits endowments (the main form of life insurance
Exam style
savings contract in Mesopotamia).
The minimum values will be:

 50% of all premiums paid, on withdrawal in the first twelve months, and
 100% of all premiums paid, on withdrawal thereafter.
(i) Explain briefly how this change will affect the persistency risk faced by life insurers. [5]
(ii) Comment on how the change will affect the benefits paid to policyholders. [4]
[Total 9]
10.4 A life insurance company writes without-profits immediate annuity contracts in a market where
consumers have complete freedom over whether or not to buy an annuity. In order to give the
Exam style
company a distinctive marketing advantage, the marketing manager has suggested that the
company should provide lump-sum benefits on withdrawal from these contracts.
Before allowing a withdrawal, the company would require a signed statement of good health
from the policyholder. Where the withdrawal payment would be higher than a certain sum of
money, the company would require a report on the policyholder’s health from his or her doctor.
Above a certain higher figure, the company would arrange a full medical examination of the
policyholder before allowing withdrawal.
The company would not publish externally the limits above which it will require a report or
medical examination.
(i) Discuss the principal additional risks for the company that would be involved. [6]
(ii) Explain the possible effects on the main items of experience if this change were
introduced. [7]
[Total 13]

10.5 A life insurance company has written individual term assurance contracts in its domestic market
for many years, and has a significant share of that market. It now intends to start writing this
Exam style
business in an overseas market. In that market, individual term assurance contracts are rare,
although there is a substantial amount of group term assurance business written.
Discuss the problems related to the various sources of mortality data that the company will face
in entering this market. [9]

Chapter 10 Solutions
10.1 Effect on mortality
The key effect here is that of selective withdrawals, ie the withdrawal of lives that are healthier on
average than those that keep their contracts in force, leaving behind a ‘sub-standard’ group of
insured people. [1]
There is a risk that rates of selective withdrawals are greater than has been allowed for in the
contract pricing, causing mortality to be worse than assumed. [1]
Effect on expenses
More withdrawals than expected means that fixed costs are spread over fewer policies than
expected. This may increase per-policy expenses above the allowance made in pricing. [1]
If the average size of withdrawing policies is larger than those that remain, then there will be less
contribution per remaining policy, on average, to cover per-policy expenses. So profits may fall if
withdrawal rates increase, even if the amount of per-policy expenses were to be unaffected. [1]
There may also be a mismatch between the incidence of expenses and the recouping of those
expenses through the product’s charges. [½]
In particular, high initial expenses may be recouped over the term and therefore may not be
recouped on early withdrawal. (This depends on when in the term the withdrawal occurs and the
basis for calculating any surrender value.) [1]
Also, there will be a small cost associated with the withdrawal itself. [½]
[Maximum 5]
10.2 (i) Persistency risk
In most investigations carried out by actuaries, assumptions will need to be made as to the future
persistency / withdrawal rates under the contracts issued by the company. [½]
This could include full withdrawal (surrender or lapse) rates, partial (or income) withdrawal rates
and paid-up rates, as applicable. [½]
Lapses are the most likely form of withdrawal for term assurances … [½]
… but other forms of withdrawal might be possible, eg surrenders for single premium policies or
reductions in sum assured. [½]
There is therefore a ‘parameter’ risk and there may also be a ‘model’ risk. [½]
The parameters for other decrements (eg mortality) will be based on a model of the selective
effect of withdrawals. Departures from the latter may invalidate the former. [½]

For products that pay a withdrawal benefit, whether there is a persistency risk depends on how
the benefit compares with the policy asset share. For term assurance, there is normally no
withdrawal benefit, so providing the asset share of the policy is positive (ie not too early on in the
policy), a withdrawal would be a source of surplus. [1]
However, the company can be open to selection if healthy lives leave (for example if they find
cheaper rates elsewhere), since the remaining lives may be of poorer health. [½]
The parameters for future expenses will also be based on a model of future withdrawals.
Departures from the latter may invalidate the former. [½]
Such departures may also be a cause of the mismatch between the charges accruing to the
company and its actual expenses. [½]
Random fluctuation risk may also arise, whereby the actual future withdrawal experience may not
correspond with the model and parameters adopted, even though these adequately reflect the
class of lives insured or to be insured. [½]
[Maximum 4]
(ii) Persistency experience
Persistency experience might be best (ie withdrawals lowest) for the insurance intermediaries
channel … [½]
… as the policyholder will have initiated the sale and so is more likely to take out a contract that
meets his or her needs. [½]
Persistency experience might be low (ie withdrawals high) for the own salesforce channel if the
salesperson pressurised the policyholder into buying a contract that he or she didn’t really want.
[½]
Persistency experience might also be better for the insurance intermediaries channel as their
clients are generally financially sophisticated and will have received a high level of advice. [½]
However, term assurances are relatively simple contracts and so the level of financial
sophistication and advice might have little impact on persistency … [½]
… unless the term assurances had a renewal or conversion option. [½]
Clients of insurance intermediaries also tend to have higher net worth and so will be more able to
continue to pay the premiums. [½]
However, term assurances have relatively small premiums, so affordability may be less of an
issue. [½]
[Maximum 2]

10.3 (i) Effect on persistency risk
Assuming that the guaranteed values are greater than asset share at some times, and are more
than the company would otherwise pay, persistency risk would be increased. [1]
It is highly likely that at the start of the contract this is the case, because the asset share will be
negative as a result of initial expenses. This means that on withdrawal initial expenses are not
recouped. [1]
There is also a risk of a surge in withdrawals just after one year, when policyholders can get all
their money back. This may be more than the company can afford to give, and will increase
per-policy expenses for the remaining policies. [1½]
The new rules might encourage more people who are likely to lapse to take out policies.
Therefore general withdrawal experience may deteriorate. [1]
There is also a risk (although it may be small) that a large fall in asset values at other durations
would make the guarantee valuable. Policyholders may feel that they could surrender the policy
and invest the money elsewhere in depressed markets. [1]
[Maximum 5]
(ii) Effect on policy benefits
For those policyholders withdrawing early, it means improved benefits (assuming guarantees are
greater than the company would otherwise have paid). [½]
However, unless the company is willing to accept lower overall profitability, then other
policyholders will have to pay for this. [1]
This might mean lower withdrawal benefits for those withdrawing later in the term, when asset
share may comfortably exceed the minimum guarantee. [½]
Alternatively there may be reduced benefits for policyholders who stay until the contract reaches
maturity or die during the term. [½]
One counter-argument is that if the guarantee encourages higher sales, per-policy expenses may
be reduced, to everyone’s benefit. However, this is only true if this is not associated with a
significant increase in withdrawals. [1]
Benefits may also be affected by changes to investment policy. For example the company may
invest in cash or other liquid assets to cover an expected level of early withdrawals. This could
mean lower investment returns and so reduce the expected benefit. [1]
[Maximum 4]

10.4 (i) Additional risks
This question perhaps strays a little into territory we shall cover in more detail when we study
underwriting later in the course. However, almost all the points included below could be obtained
from what we have covered to date on anti-selection and underwriting, plus some applied
common sense.
The main risk is of anti-selection on withdrawal. Policyholders who know that they are in poor
health are most likely to try to take up this option. This would leave a group of policyholders with
relatively long life expectancy. [1½]
The company would try to counter this with its underwriting proposals. However, these have a
number of potential problems:
 If the policyholder turns out to have been lying when signing the declaration of health it
may be impossible for the company to recover the money. (The company might not even
know the policyholder has died.) [1]
 Even if it could do so then ill-informed publicity may damage the company. [½]
 There would be much more protection for the cases where a doctor’s report is obtained.
Even so, a particular doctor’s sympathies may lie more with his or her patient than the
insurance company. Also, he or she may regard certain information as confidential. The
company may therefore not get the full picture. [1½]
 The medical examination offers very good protection against the anti-selection risk. It is
unlikely to completely eliminate the risk, and in these (large) cases the potential loss for
the company is greatest. [1]
The company could also protect itself by offering withdrawal values significantly lower than its
best estimate values, although this would reduce the perceived value of the feature. [½]
Also, if the new feature does increase sales volume, then the company will be subject to increased
new business strain. This will increase the risk of supervisory insolvency, all else being equal. [1]
[Maximum 6]
(ii) Effects on experience
Withdrawals would, of course, go up from the current level of zero. [½]
By how much will depend on how generous the withdrawal payments are and how strictly the
company applies its underwriting criteria. [1]
The effect may be least at shorter durations. Despite the withdrawal option, those who know
themselves to be in poor health are still unlikely to buy an annuity (unless favourable terms are
offered). [1]
Alternatively, the contracts may show the same tendency as other types of contract, with
withdrawals being highest early on. [½]

Expenses
These would increase because of the additional underwriting costs. [½]
The additional costs may be substantial, especially if the company sets its medical limits low, as it
will wish to do to limit anti-selection risk. [½]
Any withdrawals that are allowed will increase per-policy expenses for the remaining policies,
because some costs are fixed. [½]
However, if the contract proves popular and sales volumes are good then this may reduce
per-policy expenses. [½]
Mortality
Mortality experience is likely to deteriorate from the company’s perspective, ie get lighter. This is
because of the selective effect discussed in part (i). [1]
The effect will be greater the weaker the underwriting. [½]
The effect could be considerable, especially some years into the contract. [½]
Investment return
This may be little affected. One issue is that the greater uncertainty over the timing of payment
outgo will make matching impossible. The company may have to invest shorter on average and
this might change the overall return. [1]
[Maximum 7]
10.5 Mortality data
The main risk relates to the setting of mortality rates for pricing in the overseas market. [½]
There are various sources of data that the company might draw on, but all of these have
weaknesses.
For example, the company’s data from its own in-force policies could be of some use, … [½]
… but the overseas market may differ because of different:

 general health of population [½]
 underwriting levels and approach [½]
 target market reached, eg because of different distribution methods [½]
 provision by the State [½]
 lifestyles and attitudes to risky activities. [½]

The company might be able to use overseas industry data (if available) … [½]
… or reinsurers’ data in the overseas country. [½]
These would also be subject to problems, and hence error, because:

 the data might be very heterogeneous, reflecting a range of different policy conditions,
underwriting practices, free cover levels and target markets [1]
 the different contract design (individual versus group) may attract a different
cross-section of customers, group relating to (some) employed workers. [1]
Use might be made of any national mortality statistics in the overseas country. [½]
These would also have to be treated with caution as a guide to the mortality of an insured
population, because of the different class of lives involved. [½]
Any of these data might also be of poor quality in themselves, for example there may be missing
or inaccurate data, … [½]
… or there may not be sufficient data for reliable results. [½]
Although the company is a significant player in its own market and so its data ought to be good
quality. [½]
Some data may also be too far out-of-date. [½]
Overall
The main overall problem resulting from the above is that it is very difficult to price the contract
with confidence. [½]
This means that the company is at risk of making a loss on the contract, … [½]
… and so may need to price with significant margins, or incorporate a reviewable premium facility
to correct things if they start to go wrong. [½]
[Maximum 9]

Risk (2)
Syllabus objective
 mortality rates
 persistency
 competition
 counterparties
 fraud
 aggregation and concentration.
(The items in italics are covered in this chapter.)

0 Introduction
In this chapter we continue our study of the sources of risk to a life insurer.

1 New business mix
1.1 By nature and size of contract

A significant unintended change in the mix of new business by nature or by size of contract
could lead to a significant change in the risk profile or capital needs of the company that
were not within the resources available to it.
A change in the mix by nature might involve, for example, a change in the mix by:
 class of business (eg the proportion of pensions business written)
 type of contract (eg term assurances vs endowments)
 contract design (eg with-profits vs unit-linked)
 premium frequency (single vs regular premium).
If the risks involved in the business or the capital required to write it vary by any of these aspects,
then the overall risks and capital needs of the company will change if the mix changes.
If contracts are smaller on average than had been assumed, the income from charges and
premium loadings will be reduced, and this is unlikely to be matched entirely by a reduction in
expenses. This will occur especially on non-linked products, where most of the charges are likely
to be related to the premium size, while a larger proportion of the expenses may be per policy
and hence independent of size.
Question
Discuss whether companies could charge a fixed fee for contracts, irrespective of their size.
Solution
Well, they might do. It would get round the problem of mix by size of contract.
But, a significant problem is that a fixed fee can make smaller policies seem too expensive and
uncompetitive. Even though small contracts will probably be less profitable than large ones,
insurance companies are still interested in writing smaller contracts in order to get some
contribution to overheads.
A possible solution is to charge a fixed fee that covers part of the expenses, while still accepting
some cross-subsidy from large contracts to small ones.
A significant unintended change in the mix may be another cause of the mismatch, as
mentioned in Section 4 of Chapter 10, between the charges accruing to the company and its
actual expenses.
A mismatch between charges and expenses might result from a change in the mix of contracts by
nature and size. For example, an insurance company may have to accept a relatively low
contribution to total overhead (ie fixed) expenses from a particular type of contract. This could
arise if that contract were subject to greater competition, keeping premium rates or charges low.

If the mix of business were to change so that more of this contract were sold and less of other
contracts carrying higher overhead loadings, the insurance company might not receive a sufficient
overall contribution to cover overheads.
1.2 By source
The parameters adopted for the mortality and expense assumptions, mentioned in
Sections 2 and 4 of Chapter 10, will be based, where appropriate, on an expected mix of new
business by source. A change in the mix by source may invalidate those parameters.
By ‘source’ here we mean distribution channel. As we saw when we studied distribution channels
earlier in the course, the different populations reached by different distribution channels can
exhibit different experience for mortality and other parameters.
It is also common for withdrawal experience to vary by distribution channel.
Question
Suggest two key factors related to distribution channel that may cause withdrawal experience to
vary between channels.
Solution
 The level of financial sophistication of the customer.

 The degree of pressure of the sale, which may be affected by who initiated the sale.
These were the two key factors given by the Core Reading in the chapter on distribution channels
(Chapter 8).
Expenses of distribution also vary by channel. For example, the amounts of the salespeople’s
remuneration may differ.
A life insurance company may wish to charge the same price for a product through all its
distribution channels, despite the different expected experience. It would therefore have to use
some sort of weighted average of the experience, dependent on the expected mix by source. If
the mix turned out differently then the company might make or lose money, and this uncertainty
creates a risk.
Conversely, if a life insurance company prices differently for the different distribution channels,
and can achieve the same level of profitability in the different channels, then it is not exposed to
this risk of changes in the mix by source.

2 The volume of new business

We shall see in this section that a life insurance company may experience problems as a result of
writing too much or too little business.
A life insurance company will need to ensure that the capital and administrative
requirements of writing new business are within the resources available to it. An
inappropriately high volume of new business could, therefore, lead to the available
resources becoming inadequate.
Question
(i) Define the capital that a life insurer has available for writing new business.
(ii) Explain why writing unexpectedly large volumes of business might cause a problem.
Solution
(i) Broadly speaking, its available capital is the excess of its assets over its supervisory
liabilities including any required solvency capital.
(ii) Writing new business uses up this capital, largely because of prudent supervisory
reserving, solvency capital requirements and high initial expenses. If the company writes
too much business then it might have to impose limits on further new business.
In the extreme, it could make itself insolvent.
Question
Describe the risks, associated with administration, caused by:
(a) too much new business
(b) too little new business.
Solution
(a) Too much new business suggests over-stretched resources. Buying in extra capacity may
be difficult, expensive and inefficient (eg untrained staff, or existing staff working
overtime). This can lead to delays and/or mistakes (eg underwriting), which can generate
bad public relations and marketing risk, and/or other risks.
(b) Too little new business suggests excess capacity: costs may not be covered by
premiums / charges and profit per policy will fall. This is generally the principle that some
costs are ‘fixed’ (or ‘overhead’), at least in the short term, to the extent that they will be
incurred regardless of business volume.

On the other hand, inappropriately low volumes of a newly introduced product line may
pose a risk of not recovering the fixed development costs that will already have been
incurred.
This is related to the solution to part (b) of the previous question, but is particularly concerned
with the additional risk posed by new products being sold in inadequate volumes, as opposed to
just new business in general.
Question
Explain the difference between ‘new products’ and ‘new business’.
Solution
A new product is a type or class of policy that the company is about to sell and that it has never
sold before. New business just refers to the company’s policy sales over a period of time (as
opposed to policies that are now in force but were taken out before the period in question).
The parameters for future expenses, mentioned in Section 4 of Chapter 10, will be based on
a model of future new business. Departures from the latter may invalidate the former. Also,
such departures may be another cause of the mismatch, as mentioned in that section,
between the charges accruing to the company and its actual expenses.
The parameters for future expenses referred to in the Core Reading relate not only to future new
business but also to in-force business.
Question
Explain why the expense assumptions for in-force business may depend on a model of future new
business volumes.
Solution
There are some expenses involved in running the business that are not directly related to volumes
nor obviously associated with either new business or in-force business. (An example would be the
Chief Executive’s salary.) The more business that comes onto the books, the more policies there
are over which we can spread such expenses, so the smaller the per-policy expense assumption
will need to be.
This is another example of the effect of fixed or overhead costs that we introduced in the earlier
question.
Don’t worry if you did not get these points about fixed costs. We shall cover expense analysis in
detail later in the course.
As in any other industry, a significant fall in volumes of new business will damage profitability,
competitiveness and perhaps even the long-term viability of the insurer.

3 Guarantees and options

In the case of both guarantees and options, the life insurance company is offering terms in
advance of the happening of the event to which those terms relate. The actuary will need to
determine the cost to the company of doing this, using parameters for the future experience
in respect of the assumptions relevant to the situation, combined with a model of how to
calculate that cost. There will then be risks associated with each of these.
(That is, there will be risks associated with the choice of parameters and the choice of model.)
We will describe how guarantees and options are valued in Chapter 23.
An example will help to illustrate these ideas.
Example
A life insurer offers a ten-year, regular premium, unit-linked endowment assurance. The insurer
guarantees that the maturity proceeds from the contract will not be less than premiums paid
accumulated at 3% pa. The ‘happening of the event’, in the Core Reading terminology, is
therefore maturity, in this case.
Question
State precisely the additional risk that the insurance company faces as a result of giving this
guarantee.
Solution
The risk is that the unit fund at maturity, net of any charges, is less than the value of premiums
accumulated at 3% pa. The insurance company would then have to make up any shortfall.
This effectively means that the unit fund must grow on average at a rate rather larger than 3% pa
in order to offset the charges on the funds. The company may not achieve this rate. (One reason
might be a temporary fall in the value of the assets backing the unit fund just before maturity.)
‘The risk is that investment returns are less than 3% pa’ would not be a strictly correct answer,
because of the policy charges.
Even if the guarantee appears fairly low, there must be some risk that the insurance company will
have to find some money to cover a shortfall in the value of the unit fund. The expected cost (in a
statistical sense) must be determined.
Could we use a deterministic model to do this? A deterministic model is a projection of one

possible outcome. If we project on the basis of best estimate unit growth rates, then provided
these comfortably exceed 3% pa, the guarantee will appear to have no cost. But a best estimate
(ie an expected value) of the cost of the guarantee must be positive (unless the probability of a
shortfall occurring at maturity is zero.)

We could use a range of possible deterministic scenarios, ranging from pessimistic to optimistic,
assign some probabilities to each outcome and take an expectation based on this. This would be
better, but would probably be highly subjective in terms of the scenarios chosen and the
probabilities assigned to them.
A better approach might be to use a stochastic model. The company would have to choose a
suitable stochastic model to model unit growth. Suitable parameters would also have to be
chosen, such as the mean and variance of the random element.
The model would be run using simulations. This could produce several thousand possible values
for the unit fund at maturity, and hence of the cost of the guarantee, derived from the investment
returns generated by the model. An estimate of the expected value of the cost of the guarantee
would then be an average of all the simulated outcomes.
The results will only be valid if the model and the chosen parameters are appropriate. Either
might be inappropriate, in which case the company might charge too little or too much.
It is also worth mentioning that the insurance company could greatly reduce the risk it faces from
the guarantee if it could protect itself with a suitable derivative.

4 Competition
The need to compete, in a free market, may lead the management of a life insurance
company to take decisions that increase its risk profile beyond that which can be supported
by the available resources. Some of the decisions that it might take are:
 reduce premium rates or charges under new business contracts
 offer additional guarantees and options under new business contracts
 increase bonuses under existing contracts
 increase salaries or commissions in the respective distribution channels
 on existing business with reviewable charges, either do not increase the charges or
reduce their rate of growth relative to what may have been originally intended.
This last bullet means that, for marketing reasons, the company may not review charges upwards
by as much as it should in order to cover worsening future experience.
The impact of the above decisions can then be compounded if greater volumes of new
business result.
Question
A life insurance company has decided to add a renewal option to its ten-year without-profits term
assurance contracts. It will charge for this option by increasing the premium for the original
contract to cover the expected cost of the option.
Explain whether the risk for the company has increased by giving this option.
Solution
Yes. Even though the company has charged for the expected cost of the option, there are still
extra dimensions of uncertainty. The costing of the option will involve a model and parameters
that may turn out to be incorrect.
There may be unforeseen changes over the course of the ten years of the original contract.
There may have been weaknesses in the original underwriting that are compounded by allowing
customers to take out further contracts.
The increased premium might be viewed negatively in the market, leading to lower sales volumes.
It is of course not always the case that the example decisions given in the Core Reading above will
be inappropriate or will stretch a life insurance company beyond its means. Where this is the
case, the actuary’s course of action will depend on local regulation and professional guidance, and
on the actuary’s professional conscience.

Equally, the actions of competitors can have adverse implications for new business
volumes.
So a company’s new business volumes are likely to fall if it does not follow any of the decisions in
the list above but its competitors do.

5 Actions of the board of directors

The term ‘directors’ refers to those individuals who have the legal responsibility to:
 make decisions affecting the running of a company; and
 impose proper systems of management and control on the financial operations of

the company.
The actuary of a life insurance company will make recommendations as to how it should
operate so that its risk profile stays within the resources available to it. It is usually the
case, however, that the directors do not have to follow these recommendations. This might
arise:
 for competitive reasons, as mentioned in Section 4
 due to strategic company goals such as maximising new business volumes or

amount of funds under management
 to maximise shareholder earnings.
There is nothing in itself wrong with seeking to maximise shareholders’ (expected, risk-adjusted)
returns. However, the actuary may be in a particularly good position to judge the financial risks
involved in a particular course of action.
The actuary may also have a regulatory or professional responsibility to ensure that inappropriate
risks are not taken, particularly where the security and fair treatment of policyholders are
concerned.
There may in fact be a fundamental conflict between shareholders’ interests and those of
policyholders, because the two groups may have very different attitudes to the trade-off between
expected profit and risk of insolvency. For example, a without-profits policyholder has nothing to
gain from the insurance company pursuing high profits at high risk, but a great deal to lose if the
result is insolvency.
There may be a closer alignment of interests in other cases, for example a with-profits fund in
which all profits are shared between policyholders and shareholders.

6 Actions of distributors
The different types of distributor of a company’s products were considered in Chapter 8.
These distributors may take actions that are in their own interests and/or the interests of
their clients, but which may give rise to financial risk to a life insurance company. This
might occur particularly where the distributor is independent of the company. For example:
 encouraging business to lapse and re-enter where there are no exit penalties or
there is no clawback of commission payments – the earlier this occurs in the life of a
policy, generally the greater the loss for the life insurance company
Lapse and re-entry causes a loss if the profit made on surrender or lapse (which may be a
loss), plus the profit made from the new policy, are together less than the profit that
would have been made from the old policy had it not been discontinued.
 taking advantage of loopholes in product design
A particular example of this relates to risk classification, where intermediaries will

naturally try to find the best premium rates available in the market for their clients. For
example, insurers’ non-smoker / smoker definitions may allow some variation in
interpretation, and if an intermediary can find a way of getting (say) a formerly heavy
smoker accepted at non-smoker rates, then clearly they will take advantage of this.
 taking advantage of opportunities that arise due to timing effects in unit pricing
practices.
For example, suppose unit prices are fixed daily at the start of each day, and market
values fall heavily during the day. Should units be cashed at the end of the day and then
repurchased at the start of the following day (when they are much cheaper), a significant
profit could be made at the expense of the insurance company.
The company needs to design its products and processes to minimise these risks, as well
as engaging with distributors to discourage bad practice.

7 Failure of appropriate management systems and controls

Beyond the assumption that enough systems of control have been implemented (with
proper documentation and training to support the intended interpretation), risks arise from
the possibility that one or more of the controls is contravened, perhaps due to insufficient
monitoring.
In other words the company has done all it can to put control systems in place and to act by them,
yet there is always a finite risk that a mistake is made.
The failure of controls may result in:
 financial losses for the insurer
 regulatory intervention
 reputational damage.
It should be noted that control failures may lead to financial gains for the insurer but
nevertheless should be regarded as posing an ongoing risk.
There are two different levels of risk here, which can be illustrated by considering the example of
product pricing.
When a price is fixed for a particular contract, there will be a risk that it is inadequate to generate
the intended profit for the company.
But in coming to decide on this price, the company will have gone through a decision-making
process – ie a control system. A company that fails to implement its control system properly may
end up with a different price from that which should have been delivered by the system had it
been correctly implemented. In most cases we would expect this error to leave the company in a
financially worse situation (hence the three bullets of Core Reading listed above) but occasionally
a management error can be fortuitously beneficial (hence the final part of the above Core
Reading). Overall, though, the risk arising from the pricing process is increased by failures of
control.
This also leads into the idea of optimal decision-making strategies (or control systems). In theory,
ever more elaborate (and detailed) management control systems could be devised that lead to
ever reduced risk for the company.
Question
‘The optimal management control system is the one that leads to the smallest resulting risk.’

Solution
This statement ignores the aspect of expense: ever more detailed control systems involve ever
increasing expense, so that there comes a point at which the cost of reducing risk by a further x%
in this way becomes prohibitive. The point at which the loss of profit (due to increased expense)
first outweighs the benefit from the reduction in risk, identifies the optimal control system.
This point will itself depend on the company’s relative preference for reduced risk over increased
profitability – and this is also the point beyond which SP2 exam candidates do not need to be
further concerned.

8 Counterparties
When a life insurance company enters into an agreement with another entity, it is relying on
that entity to meet its obligations under the agreement. There is a risk that the entity will
not be able to do this, and that it may either fully or partially default on its obligations or
perform them to an unacceptable standard. This is counterparty risk.
Examples could include:
 reinsurance agreements (eg reinsurer default)
 outsourcing arrangements (eg poor quality service from a company providing

outsourced administration services)
 corporate bonds held as investments (eg non-payment of coupons and/or capital by

the issuing entity)
 distribution arrangements (eg non-recovery of broker balances).
We look at each of these examples in turn.
Reinsurance is a process by which part of the liability under a contract is passed on to another
insurer. Usually the full liability to the policyholder rests with the company that sold the business
in the first place (the ‘direct writer’). If the reinsurer is unable to pay its share of a reinsured
claim, then the direct writer faces a greater cost than expected.
We discuss reinsurance in more detail in Chapters 24 and 25.
Insurance companies often outsource functions such as IT, investment management and policy
administration. Poor service from the company providing outsourced administration services
could, for example, result in complaints from policyholders and costs for the insurance company
to put things right.
Insurance companies often invest in bonds or loans. The insurance company is then exposed to
the possibility that the issuing entity (ie the entity that has issued the bond or borrowed through
the loan) defaults on its repayments. This form of counterparty risk is often also referred to as
credit risk.
Often brokers collect the premiums and then pass them on to the insurance company, perhaps at
the end of each month. The insurance company is then exposed to the risk that there is a delay in
receiving these premiums, or that it never receives them at all.

9 Legal, regulatory and tax developments

As discussed in Chapter 9, the legal, regulatory and taxation regimes can pose risks to the
extent that such environments and rules are changed adversely, whether they apply to
policyholders, or insurers, or both.

10 Fraud
A general type of control failure, caused by deliberate intent of one or more parties, is fraud.
The main parties who might perpetrate fraud are as follows:
 directors or staff
These parties will have special access to the financial (including banking) systems
of the insurer and to the computer programs and data which support the business.
 policyholders
The main risk here relates to fraudulent claims, the risk increasing in proportion to
the difficulty of detection. Countries with secure processes for certifying deaths
represent reduced risks. There is also a risk to the insurer arising from criminal
activities such as money laundering.
 other outside parties.

Such parties may effectively obtain some access to the computer systems of the
insurer, particularly where there are external components such as website access.
Money laundering is the process that criminals use to conceal the source of their assets. It is not
unknown for criminals to attempt to buy insurance contracts with stolen money. They can then
surrender the insurance contract to give the impression that their assets were from a legitimate
source. Most countries require insurance companies to have strict controls in place to identify
suspicious transactions which may be connected to criminal activity.


Chapter 11 Summary
The following are some of the risks faced by a life insurer:
The mix of business by nature and size of contract

The unpredictability of the mix of new business by nature and size of contract is a risk for the
insurer.
Examples of a contract’s nature are: class of business, type of contract, contract design, and
premium frequency.
Unexpected variation in nature or size may change the risk profile of the company, which
may lead to an overstretching of the company’s capital and other resources available to
cover the risk.
Coverage of per-policy expenses is particularly vulnerable to a reduction in the average size

of policies issued.
The mix of business by source

Different distribution channels involve different sales methods and reach different
populations. As a result, the demographic and expense experience of the various channels is
likely to differ. Variation from the insurer’s assumed mix by source could therefore
invalidate the insurer’s demographic and expense assumptions.
The volume of business

Insurers may experience difficulties as a result of:
 too much business, so that its capital resources or administrative capacity are
exceeded
 too little business, so that it fails to cover its overhead expenses.
Guarantees and options

To calculate the cost of guarantees and options, a life insurance company will use a model.
Model, parameter and random fluctuations risks therefore occur.
Competition
The need to compete may lead management to take unacceptable risks. This might involve
decisions to:
 reduce premium rates or charges under new business contracts
 offer additional guarantees and options under new business contracts
 increase bonuses under existing contracts
 increase salaries or commissions in the respective distribution channels
 keep charges too low under existing, reviewable, contracts.

The management of the company

Management risks can arise because:
 the directors have made a conscious decision to ignore sound risk-management
advice in pursuit of other competing aims
 the control systems in place are inadequate or are not properly followed.
Mismanagement can lead to financial losses, regulatory intervention, and damage to the
company’s reputation.
Distributors
Distributors may:
 encourage lapse and re-entry where this favours the policyholder
 take advantage of loopholes in product design
 take advantage of timing loopholes in unit pricing practices.
Fraud
Directors, staff, policyholders and even external parties can all perpetrate fraud and so cause
loss to the insurer.
Counterparties
If an insurer has an agreement with another entity then it faces the risk that the entity either
fully or partially defaults on its obligations or performs them to an unacceptable standard.
This is counterparty risk.
Legal, regulatory and tax risks

Future changes to any of these can adversely affect the insurer and/or its policyholders (see
also Chapter 9.)


11.1 Describe the risks to a life insurance company arising from the general business environment.
Exam style
[14]
11.2 A small life insurance company has recently repriced its product range. As a result, it is selling
Exam style
significantly higher volumes of new business.
Describe the sources of risk to the company that may result from this. [4]
11.3 A life insurance company has successfully been selling a unit-linked single premium savings
contract through insurance intermediaries. The company is now considering introducing a new
Exam style
unit-linked regular premium savings contract in order to appeal to an even wider range of
investors.
The following design is being considered:

 fund management charge: 1% per annum
 minimum annual premium: £2,000
 minimum term: ten years
 surrender penalty: 0.5   x  10  y  % of the face value of the units at the

date of surrender, in the first ten years only (x = entry age,
y = age at surrender)
 all charges are guaranteed for the duration of the contract
 the face value of the units is payable on survival to the end of the term or on earlier
death.
Describe the risks to the insurance company from launching this product. [10]

11.4 A life insurance company sells without-profits level immediate annuity contracts through its own
salesforce and through insurance intermediaries. It is now planning to introduce with-profits
Exam style
annuities, using an ‘additions to benefits’ approach, as an additional product. The with-profits
contracts will operate as follows.
In the absence of bonuses, the income from the with-profits annuity will fall each year by 5% of
the annuity income payable in the first year. However, the company expects to grant regular
bonuses each year. These will be expressed as a percentage of the first year’s income, and once
granted cannot be taken away.
Therefore if bonuses were declared at a rate of 5% in each year, a level annuity would be
produced. However, bonuses are discretionary and might be more or less than 5%. The company
currently expects bonuses to average around 7% pa, given its expectations for future investment
returns. No terminal bonus will be given.
The company invests all the assets backing its without-profits annuities in fixed-interest securities.
It intends to invest the assets backing its with-profits annuities in a mixture of fixed interest and
equities.
Withdrawals are not allowed under either contract.
(i) Compare the risks for the company under the two contracts. [9]
(ii) Discuss how well the two contracts would meet the needs of consumers seeking a
principal source of income for life. [3]
The company is planning to use the same assumptions for mortality and expenses as it uses for its
without-profits annuities.
(iii) Discuss briefly how appropriate this is. [5]

[Total 17]

11.5 A life insurance company proposes to launch a new single premium unit-linked whole of life
product with the following features:
Exam style
 99% of the single premium is allocated to units at the offer price.

 The sum assured is payable on death.
 Before age 60, the sum assured is equal to the greater of:
– the value of units; and
– x% of the single premium paid, where x% is chosen by the policyholder at outset, with
a minimum of 100% and a maximum of 300%.
 From age 60 onwards, the sum assured is equal to the value of units.
 There is no medical underwriting if x% is 100%.
 There is a monthly management charge of 1% of the value of units and monthly
deductions are made from the value of units for the cost of life cover. The charges and
deductions are variable in certain circumstances.
 A guaranteed minimum surrender value is payable on surrender after age 60.
(i) State the principal elements of the experience which would be analysed for this
product. [3]
(ii) Describe the main risks that the company will bear on this product. [11]
[Total 14]



11.1 Risks arising through the distribution channel(s)
Possibility of selling inappropriate products to customers. This can lead to poor persistency and
damage to reputation, producing marketing failure. [½]
Risk that the channels choose to stop selling that company’s business, and it has no alternative
routes available through which to make sales. [½]
Some intermediaries might channel the products to high risk sectors of the customer market:
eg to take advantage of an underwriting loophole, resulting in anti-selection. [½]
Competition risk
Sales can be significantly reduced by competitive pressures. [½]
Commission costs to intermediaries may need to be increased to keep up with market rates. [½]
Keeping up with the competition can lead to increased risk in other areas, for example:
 by reducing contingency and/or profit margins in the pricing basis [½]
 providing additional guarantees or options [½]
 relaxing underwriting requirements [½]
 preventing the company from increasing reviewable charges / premiums when otherwise
indicated [½]
 paying out too much in discretionary payments. [½]
Risks from expenses
Labour and other costs are essentially part of the business environment. General inflation,
shortage of appropriate skilled labour, and increased costs of property, IT and investment, can all
lead to loss of profit for the company. [1½]
Risks from investment
Suitable assets for the insurer’s purposes (ie to back its liabilities) may not always be available. [½]
This will increase the cost of investment, reduce returns, and increase investment risk, all else
being equal. [1½]
Poor returns can feed back into higher premiums or charges, and/or poorer payouts to
policyholders, which leads to a marketing risk. [½]
Increased volatility of returns leads to higher capital costs, which again can increase premiums
and/or reduce policyholders’ returns. [½]
It also increases the investment risk, and will require more capital to be held. This in turn can
increase the risk of supervisory insolvency. [½]

One of the biggest investment risks comes from a large change in investment conditions, leading
to financial options coming into the money, where these are offered under life insurance
contracts (eg guaranteed maturity values under unit-linked contracts, guaranteed annuity
conversion options). [1]
Legal and regulation risks
Companies may run the risk of facing legal action, and having to pay large compensation bills, if
discretionary benefits fail to live up to PRE. [½]
It is possible that some of the contractual terms may be judged to be legally unfair at the time of
implementation, despite being written in the contract, which could cause the company significant
additional costs (eg may nullify a company’s expected ability to increase charges during the
lifetime of a policy). [½]
Any regulatory change that could affect a policy since its inception can be a risk. For example, a
change in the supervisory regulations could cause large increases in reserves (valuation strain)
and a risk to supervisory solvency. [½]
Taxation risks
A change in the way that insurance companies are taxed, or to the tax rates themselves, can
cause losses. [½]
Failure by insurance companies to take advantage of any opportunities made available by

regulation or taxation rules, can cause marketing risk. [½]
Data
The business environment may provide a source of data, eg for estimating future mortality and
lapse experience for pricing. There is a risk to the insurance company to the extent that these
data may not be fully relevant to its future experience (or have errors or omissions). [1]
Business risks
The company is at risk from downswings in sales due to adverse changes in economic
conditions. [½]
This risk will be made worse if the company is unable to make rapid adjustments to new-business
related costs, for example, by transferring staff to other functions and/or laying them off (or
recruiting them) as required by the economic cycle. [½]
The risks in this regard are much less for total in-force volume and the associated renewal and
termination expenses, as the in-force volumes will be much more stable (reflecting new business
volumes over the past 10-25 years, say). [½]

Counterparty risks
If the company enters into an agreement with another entity, then there is a risk that the entity
fails to fulfil its obligations. Examples include: reinsurance agreements, outsourcing
arrangements, corporate bonds and distribution arrangements. [1]
Downgrading of credit rating
Should the credit rating of the company fall, due to an increase in aggregate risk, this may
adversely affect a company’s ability to sell business and/or to raise capital. The latter will make
capital more expensive to the company, reducing the company’s competitive position. [1]
[Maximum 14]
11.2 The two key themes of this question are two of the key themes behind the whole course: profit and
capital. Four marks are on offer and so make sure you think of a few other ideas too.
If the increase in business is due to lower prices then there is a large risk that the company is
writing lots of loss-making business. [1]
This is especially true for without-profits policies and linked business with guaranteed charges. [½]
The company may also be exposed to an increased risk of anti-selection if the repricing exercise
has resulted in the company offering prices that are too low for certain types of policyholder. [1]
There is also a risk that a significant increase in new business puts pressure on the solvency of the
company, especially as a small company. [1]
There is a risk that administration staff cannot cope with the increased volume of business. [½]
In particular this may put pressure on the underwriters, which could lead to substandard lives
being accepted on standard terms. This could also increase the risk of anti-selection. [1]
The mix of business by contract type or by distribution channel may change significantly
invalidating the pricing assumptions used. [½]
There is a risk that reinsurance arrangements are not adequate. [½]

[Maximum 4]
11.3 Investment risk
All charges are linked to fund value: so there is risk of lower than expected income and profit if
investment returns are poor. [½]
If the policyholder chooses more risky assets the profit volatility will increase. [½]
If the policyholder chooses a poorer performing asset link then overall profitability will be lower.
[½]
Marketing risk
Persistently low investment returns will lead to a marketing risk. [½]

Poor returns produce policyholder dissatisfaction, which can ultimately lead to increased
withdrawals and lower future sales. [1]
Effect of guaranteed charges
Future losses are potentially unlimited because the company cannot increase charges in future, if
experience is worse than expected, so all risks are made worse. [½]
On the other hand, the product may be uncompetitive on price because the charging rate
probably includes a safety margin for adverse experience, so it may not sell well. [½]
Expense risk
There is a risk from selling fewer policies than expected, because this will produce higher than
expected per-policy expenses due to the company’s fixed expenses. [½]
It could also mean that the product development, marketing and launch costs are not recouped.
[½]
Whilst all of the charges are proportionate to fund size, at least some of the expenses will be per
policy (ie will be at a particular level regardless of the size of the policy). [½]
This therefore produces a risk from the new policies being smaller, on average, than assumed in
the pricing basis. [½]
This risk is, however, limited by the minimum premium condition imposed. [½]
The total income received from charges will be a function of how long a policy stays in force: the
longer in force, the more charges are received. Some of the expenses, especially initial, will not
vary by duration, so profitability will be adversely affected if only short-term policies are sold. [1]
This risk is limited by the minimum required term of ten years. [½]
Inflation
Expenses will increase with inflation. There is some protection from the effect of this on profit
because the charges are linked to fund growth, which is likely to be correlated with inflation if the
policyholder chooses real assets to link to, such as equities. [1]
But there remains the risk that income growth will not match inflation, especially in the short
term. [½]
If policyholder chooses particularly volatile or risky assets, then the risk is made worse. [½]
Persistency risk
If the asset share is less than the surrender value at withdrawal, then the company makes an
overall loss on the contract. [½]
The high initial expenses will mean that the asset share will be less than the unit fund for a
number of years. [½]

The surrender penalty should mean that, after a few years, the position will reverse and a profit
will be made on surrender. [½]
But the first year’s penalty (5% of one premium) is unlikely to be sufficient to cover the initial
expenses (including possibly significant levels of initial commission for the intermediaries), so the
company will definitely lose money on early surrender. [1]
At later durations (ie after ten years) the lack of any surrender penalty will mean that the
company will lose future profits on any surrender. So more surrenders than expected at later
durations will result in lower profits than expected. [1]
Valuation strain (capital strain) risk
There is a significant initial strain, as there are no charges at the start of the policy to offset the
initial expenses. All of the first premium is required to set up the unit reserve. [½]
The guaranteed charges may mean that non-unit reserves have to be held at some stage during
the policy, further reducing capital efficiency. [½]
New business strain means that the company is at risk from too high a rate of new business
growth. [½]
Mortality
There is a small mortality risk, because the death benefit will exceed the asset share for a number
of years. [½]
However, this is likely to be very minor, because the sum at risk is small, and mortality rates are
likely to be low (savings contract, so little risk of anti-selection). [½]
There is a possible risk from an adverse mix of business if mainly elderly persons choose to take
out the policy, as there is no price differential due to age and nor is there any maximum entry age
specified. [½]
[Maximum 10]
11.4 (i) Comparing the risks
Policy and other data
The new product will carry more risk because it is unfamiliar to the company. The company will
have no past data relating to the contract or have any experience in pricing it. [1]
Investment return
The without-profits annuity guarantees all benefits and so carries more risk, all other things being
equal. [½]
On the with-profits contracts the bonuses could be reduced, although guaranteed benefits would
still have to be covered and policyholder expectations might limit changes in bonus rates. [1]

However, it may be possible to match the outgo on the without-profits annuities, eliminating the
investment risk. This could also be done for the guaranteed element of the with-profits annuities,
although a high proportion of fixed interest would then have to be held. [1½]
If the company opts for a high proportion of equities to back its with-profits annuities then the
greater volatility of equities will give increased risk. This is especially so given that profit
distribution is not to be deferred using terminal bonus (unless the company plans to retain some
surplus). [1]
There is also a marketing risk linked to investment return on the with-profits contract. If bonuses
were particularly poor then annuity income could fall. Even if this possibility has been made clear
to policyholders, it might lead to bad publicity for the company. [1]
Mortality
Data considerations aside, the mortality risk should be lower on with-profits contracts because of
the ability to cut bonuses. [½]
One counter-argument here is that the with-profits annuities have been structured with an
expectation that annual amounts of income will increase. This weights the payments towards
long durations and so increases longevity risk. [½]
Expenses / business volumes
The company will have to spend money developing and marketing this new contract and so is at
risk that business volumes will be disappointing. [1]
However, the general point that bonuses provide a cushion against poor experience, as for
mortality, applies here. [½]
Competition
The nature of the risk is different here. There is immediate pressure to compete on premium
rates for the without-profits annuities, which are easily compared across companies (particularly
in the intermediary sector). However, if other companies are also writing with-profits annuities
then comparisons of amounts paid out will inevitably arise over time. [1½]
Capital strain risk
All else being equal, there should be less strain from the sales of the with-profits annuities
because of the lower levels of guarantee provided (due to the smaller supervisory reserves and/or
solvency capital required). [1]
The new product could still cause capital problems, however, if it turns out to be very popular in
the market. [½]
[Maximum 9]

(ii) Meeting the needs of consumers
The contracts are expected to provide an income for life and so in this sense both meet the
policyholder need. [½]
Given that the annuity is to provide the main source of income, a high degree of security and
predictability is required. In nominal terms, the without-profits annuity provides the greater
certainty. [1]
There is a risk that the income will fall on the with-profits annuity, which might concern potential
policyholders. However, there are still significant guarantees under the contract, and these
provide some security. [1]
The level without-profits annuity does not meet the policyholder’s need for protection against
inflation. The with-profits annuity would be expected to do this better because it is partly backed
with equities. [1]
[Maximum 3]
(iii) Using same assumptions as for without-profits
Mortality
This may not be an unreasonable assumption. However, the main risk would be that the different
types of annuity might appeal to different types of people. In particular, the with-profits annuity
might appeal to the more financially sophisticated, who might be better off and therefore have
higher life expectancy. [1½]
Expenses
A consequence of the above point is that the mix of sales by distribution channel may change.
This could lead to different average expenses. [½]
Commission rates could be higher for the new contract, to provide suitable remuneration to
intermediaries to reflect the greater complexity of the product. [1]
The expenses of the with-profits annuity will be higher because of the need to determine the
amounts of bonus, change payment details etc. [½]
The company will also have to decide how it will load for the initial cost of developing the
product, training sales staff, explaining the product to intermediaries etc. [1]
Both
For both expenses and mortality there is an argument for larger margins in assumptions to reflect
the greater uncertainty of the new product. [½]
However this is probably outweighed by the reduced need for margins given that less is
guaranteed by the with-profits annuity. [½]
[Maximum 5]

11.5 There is a lot of information to take in before part (i) is reached. This is obviously given for a
reason and can be turned into easy marks.
Part (i) says ‘state’, so we probably need around six elements. Try to give the most important ones
first.
Each of the elements can then be turned into a risk for part (ii).
(i) Principal elements
 unit growth rate – split by fund

 mortality rates – split by age and sex and into medical (underwritten) and non-medical
(not underwritten) lives
 expenses – split into initial, renewal, claims and investment expenses
 expense inflation
 withdrawal rates
 new business volumes – split by distribution channel
 mix of new business – by age, sex, fund, premium size, level of cover chosen.
(ii) Risks
Low unit growth
This will mean low management charges, so there is a risk that these will be insufficient to recoup
the expenses incurred. [1]
On death before age 60, the guarantee of x% of the single premium paid is more likely to bite, so
the cost of providing it may prove higher than expected, leading to a loss. [½]
The guarantee on surrender after age 60 is more likely to bite if there has been low unit growth,
so may be more costly than expected. [½]
The benefit on death after age 60 is the value of units. If unit growth has been poor or there is a
sudden fall in unit price, then the benefit may be perceived as poor value for money. This may
cause bad publicity and affect new business levels. [1]
A contract could theoretically have a lower benefit on death than on surrender, after age 60, if
unit performance has been poor. So, if the surrender value is higher than the value of units at any
time after age 60, the policyholder may be very tempted to surrender. [1]
If the value of units at age 60 is below x% of the single premium paid, there will be a sudden fall in
sum assured which may well go against PRE. [½]

Mortality
The risk is that mortality will be higher than that assumed in the pricing basis. The risk is greatest
for cases with x = 300, since a minimum of three times the single premium will be paid out on
death before age 60. [1]
Lives choosing x = 100 are not medically underwritten, so there is a risk that mortality rates are
worse for these lives than expected. On the other hand, people in very poor health are unlikely to
take out this policy purely for the sake of the ‘free’ cover, because they will only be getting back
the same money that they paid in to the policy in the first place. [1]
High expenses
There is a risk that expense levels will be higher than expected, leading to reduced profits. [½]
This may be compounded by low new business volumes, causing fixed overheads to be spread
over fewer policies than expected. [½]
Expense inflation may also be higher than allowed for in the pricing basis, leading to an erosion of
profits. [½]
However, this may be offset to some extent by higher nominal investment returns and hence
higher fund management charges. [½]
Variability of charges
Charges to cover mortality and expenses are variable in ‘certain circumstances’. This will allow
some of the risk to be passed on to policyholders. [½]
Depending on the exact nature of these circumstances, there is a risk that the company cannot
increase charges to reflect higher than expected costs. [½]
Even if charges were fully variable, the extent to which the company can increase charges is
limited by PRE and large increases are likely to lead to higher selective withdrawals and lower new
business volumes. [1]
In any case, there will be a delay between the occurrence of the increased costs and the
implementation of increased charges, because of the time taken to recognise the deterioration in
the experience. [½]
There is a risk of more withdrawals than expected at times when the surrender value is greater
than the asset share. This may occur very early on in the contract or after age 60, if the guarantee
is biting. [1]
If the value of units is less than the surrender value at any time after age 60, the financial
incentive to surrender will significantly increase the withdrawal risk. [½]

Low average premium size
Both the 1% initial charge and the 1% management charge represent a cross-subsidy between
large and small policies, so there is a risk that total charges make an insufficient contribution to
per-policy costs if the average policy size is lower than assumed when pricing the contract. [1]
High new business volume
There may be an initial capital strain when this policy is sold, as reserves on day 1 are likely to
need to be at least equal to the value of the units (99% of premium) which is likely to be higher
than the day 1 asset share (premiums less expenses). The company has a risk to its supply of
capital (and therefore its supervisory solvency) if much higher than expected volumes of sales
occur. [1]
[Maximum 11]

Risk (3)
Syllabus objective
 mortality rates
 persistency
 competition
 counterparties
 fraud
(The item in italics is covered in this chapter.)

0 Introduction
In this chapter we consider the importance of assessing the combined effect of all the individual
risks in aggregate, and not just separately. We will then see how risk assessments will affect the
decision-making process. Finally we look at a consequent effect of risk – the company’s credit
rating – and see why the credit rating itself can be used as a focus for an insurer’s risk objectives.
This chapter introduces some aspects of Subject SP2 that we will cover in more detail later
(especially modelling and risk management), and also forms a pivotal position in the actuarial
control cycle. You should therefore not worry too much if you do not follow every last detail of
this chapter on your first reading. However, the ideas covered here are some of the most
important of the whole course, so it is vital that you understand the content fully by the time you
are ready to sit the exam.

1 Aggregation and concentration of risk

It can be seen from the previous two chapters that a life insurer is exposed to many
different risks.
Depending on the mix of products sold, the target market and investments held, the
company may also be subject to concentration effects, both in terms of individual risks and
overall types of risk.
Whilst it is important to understand each risk individually, the potential financial impact on
the insurer will depend on how these risks emerge in the future and on how they relate to
each other. An assessment of the overall risk needs to take into account the level of
diversification between the risks.
This impact is best assessed by carrying out financial projections with the greatest insight
being provided if stochastic methods are adopted, incorporating probability distributions of
the key risks. This will allow for correlations between the parameters. Running the
projection model many times to test sensitivities and different scenarios will produce
results that enable the insurer to assess the inherent risk in the business.
The modelling methods involved are fully covered later in the course, particularly in Section 2 of
Chapter 14 and in Section 4 of Chapter 15.
Assessing solvency in light of the risks is covered in Section 3 of Chapter 15.
Useful detail of the modelling involved for a specific application (developing an investment
strategy) is covered in Section 4 of Chapter 28.
1.1 The aims of the insurance company

The management of a life insurance company will have a number of aims. Of greatest
relevance in this context is the desire to:
 maximise (or optimise within an acceptable level of risk) the profits of the company,
whether these go to shareholders or with-profits policyholders
 maximise (or optimise within an acceptable level of risk) the return that the company
achieves on its available capital.
These are usually complementary ambitions, but the board will want to be sure that the
capital in the company is being used to the best advantage, particularly where it needs to
decide between a number of competing uses of that capital.
Later in the course we shall discuss the relative merits of aiming to maximise the present value of
profit or the percentage return on capital. As the Core Reading says, these will usually lead to the
same conclusions and business decisions, but this is not inevitable.
Question
Suggest briefly any ways in which underwriting and reinsurance might be counter-productive with
regard to the company’s aims.

Solution
Both activities cost money, so the more carefully the company underwrites policies, or the more
reinsurance cover it takes out, the less the expected profit will be, all other things being equal.
1.2 Assessing the risk

Risk needs to be looked at in terms of the effect it has on the above aims.
The overall insurer risk can be measured against the above aims but also allowing for:
 the capital and other resources available to the insurer
 the cost of failing to meet the public interest need, as usually expressed in
insurance supervisory legislation, for the company to avoid insolvency
 the cost of failing to meet the requirements of any other applicable legislation.
Question
Describe how the availability of capital will affect the assessment of risk.
Solution
The more capital available the more emphasis there is likely to be on measuring the return on
capital. Risks that have high severity may be tolerated in this situation provided they have low
enough frequency (ie the company may tolerate the risk of making a large loss if this is adequately
offset by the potential for increasing the return on capital). This kind of strategy might be
deemed too risky if capital is scarce.
It would clearly be disastrous for a company to be declared insolvent by the regulators. This is a
potential consequence of risk with which an actuary must be particularly concerned when giving
advice to the company. While supervisory insolvency is a critical business risk for the company,
the policyholders will almost certainly receive most or all of the fair value of their policies,
depending on the particular criteria that are applied by the supervisors to define insolvency.
Whether the policies can continue in force, perhaps administered by a different insurer, will
depend on the circumstances of the case.

1.3 The effect of risk on management decision making

The above is all to do with determining an appropriate management strategy that satisfies the
company’s desired trade-off between risk and return (whether measured as return on capital or
as profit). This might alternatively be expressed as:
 finding the appropriate risk-return combination
 maximising expected return subject to an acceptable degree of risk.
Question
As part of its profit distribution strategy, a UK or US life insurance company will have to make a
decision about the proportion of total benefits, payable under its with-profits policies, that it
would like to have in terminal (bonus or dividend) form.
Explain the conflicts between risk and return that arise in making this decision.
Solution
The more terminal bonus payable, the greater a company’s investment freedom, and the lower its
investment risk. This should enable a company to invest in more risky assets, and thereby achieve
a better return for its policyholders (and, where applicable, its shareholders).
One conflict arising is that a higher terminal bonus proportion has the effect of passing more of
the investment risk to the policyholders, by making less of the benefits guaranteed. The company
then has a marketing risk: if the underlying (risky) investments do perform badly, then policy
payouts will reflect this and policyholders will be disappointed. This could damage the company’s
reputation and have significant adverse effects on new business sales and on persistency rates.
Another conflict is that the larger the terminal profit distribution, the more delayed the overall
distribution of the profits becomes. This is particularly an issue for shareholders under both
methods, and for policyholders receiving cash dividends under the US system. To the providers of
capital, a delay in profit distribution means a reduction in the return on capital, all else being
equal.
One of the practice questions at the end of this chapter is an example of how decision making is
influenced by risk.

2 Credit failure
An insurance company’s credit rating is an external assessment of the aggregation of the risks we
have been discussing in this and the previous two chapters.
Question
Explain why a downgrading of credit rating is a risk to the insurance company.
Solution
It would lead to adverse publicity. This would mean:

 new capital would be harder to raise and the cost of capital would increase
 people would be less likely to take out new policies with the insurance company, and
there would be increased rates of discontinuance amongst existing policyholders.
Both of these factors would make it much harder for the company to operate a successful and
profitable business.
The insurer will also wish to arrange its business practices and overall risk profile such that
it minimises the possibility of being downgraded (or indeed leads to the credit rating being
upgraded).
A downgrading of a company’s credit rating is a risk because it would lead to:
 adverse publicity
 greater difficulty and cost in raising additional capital in the market.
Consequently:
 the range of profitable activities may be constrained
 policyholders may be less likely to maintain or purchase policies with the insurer.
(This is basically the answer to the previous question.)

Chapter 12 Summary
Aggregation of risk
The combined effect of all the individual risks of loss must be considered, allowing for the
correlations and interactions between them. Model-office projections will be used to assess
the overall risk, typically using stochastic modelling and sensitivity testing.
Risk of loss has to be measured in terms of the effect it may have on the aims of the
company. These aims are:
 to maximise profits, and/or
 to maximise return on capital.
Risk must also be assessed in relation to:

 the company’s capital and other resources
 its impact on supervisory solvency
 the cost of failing to meet any other legislative requirements.
Risk and decision-making

The insurance company will make its strategic decisions by balancing the need for risk
control against the achievement of the profit aims of the company.
Credit failure
An insurer’s credit rating may be downgraded if the company does not appear to be
controlling risk to the satisfaction of the investment market.
A downgrading is itself a risk to the success of the company because it can lead to:
 adverse publicity, thereby causing marketing difficulties
 increased difficulty in raising capital, thereby increasing cost and/or limiting the
company’s ability to follow capital intensive projects.



12.1 Comment on the following statement, including suggesting any additional factors that might need
Exam style
to be taken into account when assessing the aggregate risk of a life insurer.
‘One way of measuring aggregate risk for a life insurance company is to look at each individual risk
separately, and determine which of these risks has the largest adverse financial impact. This
largest risk can then be taken as a measure of the company’s aggregate risk.’ [7]
12.2 A proprietary life insurance company transacting term assurance business is reviewing its
Exam style
premium rates.
The revised premium rates are not competitive.
Discuss possible reasons for this and courses of action open to the company if it wishes to offer
more competitive premium rates. [9]



12.1 First, the risk that has the largest potential financial impact (ie largest severity) may not be the
company’s largest risk. [½]
Risk is the product of {incidence} ¥ {severity} so, for example, a risk with high severity but a very
low probability of incidence may constitute much less of a risk than one with moderate severity
but relatively high incidence. [1]
On the other hand, it can be argued that the company is likely to be more vulnerable to rare but
very severe events, so such risks should not be ignored, but continually monitored so that more
drastic action can start to be taken should the probability of incidence increase. [1]
A real-life example of this was where gradually changing investment conditions led to many
guaranteed annuity options costing life insurance companies a great deal of money, in the UK,
from around the start of the current century.
However, only to consider this rare-but-large risk as the total risk would also be wrong, as the
company could well be losing money from the higher-frequency-but-smaller-severity risks, if it
paid no attention to them. [½]
The solution is therefore to consider all of the individual risks at the same time, taking into
account the probability of incidence, though this is not the whole picture. [½]
Many of the individual risks interact with each other, either in a way that increases their
combined risk when present together, or in a way that reduces it (possibly even to the extent that
their risk in combination is less than the risk from any one of them present alone). [½]
So we also have to look at the interactions between the individual risks when considering the
aggregate risk. [½]
Finally, the given definition is in a vacuum. An event will only be a risk if its occurrence makes the
company fail to achieve one or more of its aims as a result. [½]
The key aims would generally be:

 to maximise profits [½]
 to maximise return on capital. [½]
These will often go together, but the company will also want to ensure that the actions it takes
will not seriously jeopardise the return on capital. [½]
For example, the company could take some action that would increase total profits by a small
amount, but only by using a very large amount of capital, thereby creating a (possibly significant)
fall in the return on capital.

The assessment of aggregate risk would also depend on:

 the amount of available capital, as the more capital available, the more emphasis there
should be on the risk to the return on that capital [½]
 the company’s supervisory solvency position, which, if threatened, will cause all activity to
be focused on minimising the risk of making losses (which is very different from
maximising the chance of making profits) [½]
 any other legislative requirements (eg compliance with relevant codes of practice), of
which failure could have a significant impact on the company’s ability to do business. [½]
[Maximum 7]
12.2 Consider each of the principal elements of the pricing basis:
Mortality
Experience might be heavier than that of competitors, due to (eg): [½]

 weaker underwriting [½]
 different target market [½]
 different distribution channel [½]
 no smoker / non-smoker split. [½]
So, could:
 strengthen underwriting [½]
 change target market [½]
 change distribution channel [½]
 introduce differential rates for smokers and non-smokers to avoid anti-selection. [½]
Could take a less conservative view in setting the mortality assumption, if this can be justified by
evidence, eg from own data, industry or reinsurers, … [1]
… or by offering cover on a reviewable basis. [½]
Expenses
May be higher than those of competitors. [½]
Could look to reduce commission (may be paying above market rates). [½]
Reduce sales costs by more efficient marketing and increasing salesforce productivity. [½]
Try to improve admin efficiency and reduce overheads. [½]
May be able to achieve lower price if total profits are improved by increased volumes. [½]
Withdrawals
Experience may be worse (heavier) than that of competitors. [½]

Try to improve persistency through improving sales methods and training of sales staff. [½]
Withdrawal rates may differ due to different target market and/or distribution channel, so could
try to change these. [1]
Profit target
Profit requirement could be too high relative to the market, so reprice at reduced level of target
profitability. [½]
Consider how much contribution is required from this business to overheads; reducing the
required contribution will reduce price. [½]
Perhaps reduce prices and simply accept lower profitability (or even losses), if the company feels
that other reasons for offering competitive term assurance rates justify this: eg the need to
attract sales of other types of (profitable) business, and/or to provide a complete (competitive)
product range. [1]
Other
It may be possible to reinsure the business more efficiently, to make the business more profitable.
Reinsurance can be used to reduce capital requirements, and sufficient reduction in the overall
cost of capital could allow the company to reduce its premium rates. [1]
Don’t worry if you didn’t get this last point on reinsurance: the uses of reinsurance will be fully
covered later in Chapter 25.
[Maximum 9]

End of Part 2
What next?
Chapters 8 to 12.
of your revision.
… ‘learning and revision’ products
Face-to-face and Live Online Tutorials – If you haven’t yet booked a tutorial, then maybe now
is the time to do so. Feedback on ActEd tutorials is extremely positive:
‘I find the face-to-face tutorials very worthwhile. The tutors are really
knowledgeable and the sessions are very beneficial.’
‘The online tutorial was just wonderful and a very good tutor. The delivery
was very good and the sound was very clear. For my first online tutorial I
was very impressed.’
You can find lots more information in our Tuition Bulletin, which is available on our website at
www.ActEd.co.uk.

SP2-13: Unit pricing Page 1
Unit pricing
Syllabus objective
2.3 Describe the principles of unit pricing for internal unit-linked funds.

Page 2 SP2-13: Unit pricing
0 Introduction
Unit-linked products were discussed in Chapter 4. In this chapter we examine unit pricing. This
means how the units backing unit-linked products are priced. We are not looking at the related
subject of how unit-linked products themselves are priced, which involves the determination of a
charging structure.
The underlying concepts are all simple and should follow intuitively from the nature of a unit-
linked contract. However some of the material on the mechanics of unit pricing may at first seem
complex. This is partly due to the frequent use of the words ‘bid’ and ‘offer’ with meanings that
appear to change constantly. If you start to feel perplexed then you might find it useful to consult
the chapter summary for a quick clear-cut view.

1 Unit-linked products revisited

We start with a very brief recap of unit-linked products:
 A unit-linked benefit is determined by the performance of units. This performance is itself
determined by the performance of the underlying assets.
 Unit-linked products may be used for savings and protection.
 The cost of any protection element is usually met by explicit charges.
 Expenses are often met by a variety of explicit charges, some of which will normally be
variable.
 Unit-linked business can be very capital efficient for the company.
In this chapter we are concerned only with the first aspect: how the price of a unit should be
determined.
At its simplest, pricing a unit is no more than assessing the value of assets attributable to one unit
in the fund.
For example, if a fund has assets worth $200 and 10 units, the unit price is $20.
However various complications may arise:

 if the company is creating new units – because it will have to buy more of the underlying
assets
 if the company is cancelling existing units – because it will have to sell some of the
underlying assets.
We shall see how these issues affect the value placed on a unit.

2 What is an internal unit-linked fund?

The fundamental idea behind unit-linked business is that the benefits to the policyholder are
determined by the investment performance of specified types of asset. To do this, policyholders
are sold units in an internal linked fund. These units are, in effect, a share of a specified set of
assets.
An internal unit-linked fund consists of a clearly identifiable set of assets, for example
equities, property, fixed-interest securities and deposits. The fund is divided into equal
units consisting of identical sub-sets of the fund’s assets and liabilities.
This division is notional, but it is an excellent way of envisaging what is going on.
Responsibility for unit pricing rests with the company, subject to any relevant policy
conditions.
For example, suppose a company has an internal linked fund that aims to track the FTSE 100. It is
the responsibility of the company to price the units of the fund in accordance with the assets
therein; it is not the responsibility of, for instance, the body responsible for calculating the FTSE
100 index.
Some companies create more units in their internal linked funds than is strictly necessary
to cover the corresponding unit liabilities. This excess, or ‘box’, for any fund can be useful
in its management, depending on the systems used by the company. However, it is also
common for companies to match as closely as possible the number of units in issue with
that allocated to unit-holders.
This excess of units facilitates the management of the fund, and so is also often referred to as the
‘management box’ or ‘manager’s box’. There will normally be some flux of policyholders
surrendering units, while other people want new units. It is often better for the company to
manage this day-to-day flux by holding onto some units as the owner of those units.
Example
A life insurance company has a unit-linked fund with assets of £850,000. There are 1,000 units,
and so each unit has a value of £850. But only 970 units are ‘owned’ by policyholders; the
remaining 30 are owned by the company in the form of a management box.
A new policyholder pays a premium that entitles them to 15 units. Rather than buy the
underlying assets to create the extra units, the company sells to that policyholder 15 units from
the box. This brings the box down to 15 units.
The next day several policyholders surrender altogether a total of 145 units. The company
decides to keep 35 of those units, adding them to its box, and therefore has to cancel the
remaining 110 units by selling the underlying assets. It does not want a box of more than 50
units.
Question
Explain the risk to the company of maintaining a management box.

Solution
The risk here is that the value of the underlying assets goes down, so the value of the units that
the company is holding for its own account decreases.
For this reason, companies that use a box will normally keep it as small as practical given the
degree of unit flows that arise from their policyholders.
Unless otherwise indicated, it will be assumed that the company aims to match units. This
assumption does not affect the underlying principles.
Details of tax considerations in unit pricing are not required in Subject SP2.
The effect on the unit price of unrealised capital gains and losses is discussed in Subject SA2.

3 The basic equity principle of unit pricing

There are usually no statutory or other regulations on pricing units of an internal fund.
Policy documents will normally state how the company will determine unit prices, but this is
often done in very general terms. For example, the policy may specify how to determine
maximum prices for allocating units and minimum prices for surrendering units. Despite
the generality of such statements, policy wording should follow from the general principles
given below and not vice versa.
Strictly, a company needs to determine unit prices only when transactions are taking place.
Prices at other times have no relevance to the fund, apart from use in measuring fund
performance.
However, the company would normally want to value units much more frequently – typically daily
– in order to inform policyholders of the value of their benefits (through publication, or in
response to telephone or internet queries).
For the holder of a unit the only relevant prices are those at which the unit is allocated and
subsequently redeemed. In theory, the movement in price between these two events should
depend only on the performance of the assets backing the unit and the charges deductible
under policy provisions. It should not, therefore, be affected by the creation or cancellation
of other units, otherwise cross-subsidies between unit-holders will arise.
This leads to the basic equity principle of unit pricing for an internal fund, namely that:
 the interests of unit-holders not involved in a unit transaction should be unaffected

by that transaction.
This is the most important sentence in the chapter. We shall see that this means that the creation
and cancellation of units, and buying and selling units, should not give rise to a change in the net
asset value per unit.
As will be explained below, in practice the creation or cancellation of other units can affect
the price of a unit. Subject to this however, a company will wish to follow the above
principle in its pricing.

4 Appropriation and expropriation prices

These are the prices at which the insurer creates and cancels units respectively.
4.1 Definitions
Appropriation price
Consider first the situation where the company is creating units.
The basic equity principle dictates that existing unit-holders should be in the same position
after the transaction as before. This can only be achieved if the amount of money put into
the fund for each new unit appropriated, ie created, is such that the net asset value per unit
is the same after as before the appropriation.
This amount of money – per unit – is known as the ‘appropriation price’.
It can be defined as the price at which the company will create a unit. In other words, it is
the amount of money that the company should put into the fund in respect of each unit it
creates, to preserve the interests of the existing unit-holders.
We can think of the appropriation price as reflecting the best price at which the underlying assets
can be purchased, taking account of relevant dealing expenses.
Expropriation price
Similar considerations apply when a company is cancelling units. In this situation
continuing unit-holders should be unaffected by the unit cancellation. This can only be
achieved if the amount of money the company takes out of the fund for each unit
expropriated, ie cancelled, is such that the net asset value per unit is the same after, as
before, the expropriation.
This amount of money – per unit – is known as the ‘expropriation price’.
It can be defined as the price at which the company will cancel units. In other words, it is
the amount of money that it should take out of the fund in respect of each unit it cancels, to
preserve the interests of continuing unit-holders.
We can think of the expropriation price as reflecting the best price at which the underlying assets
can be sold, taking account of relevant dealing expenses.

4.2 Calculation process
Appropriation price
This can be calculated as follows:
 market ‘offer price’ value of the assets held by the fund
 plus the expenses that would be incurred in the purchase and any stamp or other
duty payable in respect of such a purchase
 plus the value of any current assets, such as cash on deposit or investments sold
but not yet settled
 less the value of any current liabilities, such as investments purchased but not yet
settled or loans to the fund
 plus any accrued income, such as interest income from fixed-interest securities and
deposits, net of any outgo, such as fund charges.
 less any allowance for accrued tax, if applicable.
This gives the net asset value of the fund on an ‘offer basis’. Dividing by the number of
units existing at the valuation date, ie before any new units are created, gives the
appropriation price.
The term ‘offer basis’ means that the underlying assets being bought are being purchased at their
‘offer price’. The concept of the ‘offer basis’ is discussed fully in the next section.
Expropriation price
The process is very similar to that for determining the appropriation price, the main
difference being that the starting point is the proceeds that would be received from selling
the assets in the fund.
This requires that the investments of the fund are valued on a market ‘bid basis’ and that
the expenses that would be incurred in the sale are deducted. Other adjustments are as for
the appropriation price.
Again, the result is the net asset value of the fund but this time on a ‘bid basis’. Dividing by
the number of units existing at the valuation date gives the expropriation price.
So here ‘bid basis’ means that the underlying assets are being sold at their ‘bid price’. The
concept is discussed fully in the next section.

5 ‘Offer basis’ and ‘bid basis’

It is a general actuarial principle that the basis used for a valuation is determined by the
purposes to which it is to be put. Section 4 shows there are two situations that need to be
considered when pricing units.
1. The first situation is when the marginal transaction involves the creation of units.
This requires that the fund be valued on an ‘offer basis’, with the amount of money
put into the fund being equal to the net number of units being created multiplied by
the appropriation price. This is called pricing on an ‘offer basis’.
Question
Explain the significance of the word ‘net’ in the above paragraph.
Solution
The point here is that requested unit purchases can be paired off with unit sales, so that rather
than buy one unit for one policyholder and sell one unit for another policyholder, the company
simply transfers the unit from one to the other. So the company only has to create (or cancel)
units equal to the net result of ‘all unit purchases less all unit sales’.
This also explains the use of the word ‘marginal’ in the Core Reading.
The time period involved for this netting would normally be 24 hours; eg look at all purchases and
sales requested from noon yesterday to noon today, calculate the net result, and then if the
management box cannot cope, create / cancel the necessary units.
The offer basis can be thought of as that appropriate for an expanding fund. So overall the life
company is creating units even if some policies are maturing or being surrendered.
2. The second situation is when the marginal transaction involves a cancellation of

units. Here the fund is valued on a ‘bid basis’, with the amount of money taken out
of the fund being equal to the net number of units being cancelled multiplied by the
expropriation price. This is called pricing on a ‘bid basis’.
The bid basis is the converse of the offer basis: it is the basis suitable for a contracting fund.
To summarise:
 If the fund is expanding then, overall, the company is a net creator of units. Creating units
will require buying the underlying assets, and so the assets are worth their offer price in
the market. So the basis is called an ‘offer basis’.
 If the company is a net canceller of units, where the underlying assets will be worth their
bid price in the market, the basis is called a ‘bid basis’.

6 Offer and bid prices
6.1 Introduction
Under both the offer basis and the bid basis, there will be offer and bid prices. The offer price is
the price at which units are offered for sale to the policyholder. The bid price is the price at which
units will be bought from the policyholder. These are often different because of the common
contract design feature of the ‘bid-offer spread’.
Just because, for instance, the fund is expanding and the pricing basis is an offer basis does not
mean that there are no policyholders who want to sell their units (eg due to surrenders). We
would need a ‘bid price of units’ calculated on an offer basis in order to deal with such
policyholders.
So there will be four potential prices:

 offer price, offer basis
 bid price, offer basis
 offer price, bid basis
 bid price, bid basis.
Question
State when a company would use the offer price on a bid basis.
Solution
The combination of offer price and bid basis would be seen with a policyholder paying a premium
to buy units from a contracting fund.
The Core Reading in Sections 6.2 and 6.3 sets out processes for calculating the various prices. This
is then illustrated by a diagram and numerical example in Section 6.4.
6.2 Offer and bid prices when pricing on an offer basis

Consider the situation where a company determines unit prices daily and for a particular
fund on a particular day the number of units being allocated exceeds that being
surrendered.
As this requires a net creation of units in that fund, pricing should be on an offer basis with
offer and bid prices being derived from the appropriation price. The company could use an
unadjusted appropriation price for dealing with policyholders, but for commercial and
practical reasons there are two types of adjustment that may be made.

Initial charges
When units are allocated to policyholders, companies may want to make an additional
charge as a contribution to meeting other management expenses and any commission
payments and as a contribution to profits. This initial charge is an addition to the
appropriation price.
The offer price could then be taken as equal to the appropriation price plus the initial charge
and the bid price as equal to the appropriation price.
This initial charge may also be referred to as a ‘bid / offer spread’.
Rounding
It is normal to quote prices to a certain number of decimal places. This could be done by
rounding the offer price up and the bid price down, for example, which rounds in favour of
the insurance company. Alternatively, rounding the offer price down and the bid price up
rounds in favour of the customer.
Actual practice varies, although there has been a general trend away from systematic
rounding against the customer.
6.3 Offer and bid prices when pricing on a bid basis

Consider next the situation, again with daily pricing, where for a particular fund on a
particular day the number of units being surrendered exceeds that being allocated.
As fund units must be cancelled, pricing should be on a bid basis, with offer and bid prices
being derived from the expropriation price. The rest of the process follows exactly that for
deriving offer and bid prices when pricing on an offer basis, but using the expropriation
price.

6.4 Numerical example

The diagram below gives one hypothetical example of how the various different prices might
compare. An explanation is given on the next page. Prices are in pence per unit.
Offer price
Offer basis 107.4
Appropriation
price 102.0154 Bid price
Offer basis 102.0
Offer price
Bid basis 101.0
Expropriation
price 95.9342
Bid price
Bid basis 95.9

These prices were derived from the following example method:
Offer basis
For an expanding fund, the prices are calculated as follows:

 take the market offer price plus expenses (assume this value is 102.0154p per unit) – this
is the appropriation price
 round to the lower 0.1p
 bid price on an offer basis 102.0p (expanding fund)
 multiply by 100/95 (assume a 5% initial charge)
 take the answer to the higher 0.1p
 offer price on an offer basis 107.4p (expanding fund).
Bid basis
Similarly, for a contracting fund:

 take the market bid price less expenses (assume this is 95.9342p per unit) – this is the
expropriation price
 take the answer to the lower 0.1p
 bid price on a bid basis 95.9p (contracting fund)
 multiply buying price by 100/95
 round to the higher 0.1p
 offer price on a bid basis 101.0p (contracting fund).
6.5 Offer basis or bid basis?

In Sections 6.2 and 6.3 above, the notes suggest that companies decide the pricing basis
taking account of just that day’s cashflow.
That is, on days when the company is a net creator of units, it will price on an offer basis, and on
days when it is a net eliminator of units it will price on a bid basis.
However, in practice, companies are more likely to use a ‘broad equity’ approach under
which the basis is only changed if there is a significant cashflow movement against the
existing basis eg if there is a significant net cash inflow for a fund currently being priced on
a bid basis. This approach provides broad equity between different unit-holders and
reduces price volatility.
Many companies will use a management box to minimise changes to the basis. For instance,
suppose we have a net ‘order’ for one day of a sale of 2,300 units, and the company has a box of
13,000 units for that fund. If the company has been pricing on an offer basis for the last few days
then it would probably add the units to the box (ie buy the units for its own account), rather than
redeem them. However, this would be subject to the box not then exceeding some pre-agreed
limit.


Chapter 13 Summary
The basic idea behind the unit price is that the value of the pool of assets is divided by the
number of units.
The life company may run a management box to enable it to meet the day-to-day needs of
policyholders without having constantly to create and cancel units.
In pricing units the fundamental rule is the basic equity principle: The interests of unit-
holders not involved in a unit transaction should be unaffected by that transaction.
Pricing bases
When the company needs to create a unit, it will pay the appropriation price. This is based
around the best price at which the asset can be purchased – ie the market offer price of the
underlying assets. When cancelling a unit it will realise the expropriation price, based on the
market bid price of the underlying assets.
The company may be looking at an expanding fund or a contracting fund. When the fund is
expanding the company will be a net creator of units, and so will value the fund on an offer
basis. The converse situation is with a contracting fund, where the company will be a net
canceller of units, and will value the fund on a bid basis.
In practice, a ‘broad equity’ approach is used, whereby the pricing basis is only changed if
there is a significant cashflow movement against the existing basis.
In selling units to policyholders, the company will use an offer price. This will differ from the
bid price at which it buys units back from the policyholders by the initial charge (also called
bid / offer spread). At this stage the company will also make some rounding decisions.
Bid and offer

The words ‘bid’ and ‘offer’ are used in three different situations:
 to refer to the buying and selling prices of assets
 to refer to the buying and selling prices of units
 to refer to the basis used to price units.



13.1 Explain how units are priced in a linked fund.
13.2 (i) Explain what is meant by a ‘management box’, in the context of unit pricing. [1]
Exam style
(ii) Discuss briefly the advantages and disadvantages of holding one. [3]
[Total 4]



13.1 Unit pricing
In general unit pricing will follow the basic equity principle: that the interests of unit-holders not
involved in a unit transaction should not be affected by that transaction.
The pricing will depend on whether the company is creating or cancelling units in the fund.
This situation might be looked at daily or, more likely, over a longer period to avoid changing the
basis too frequently.
If the company is creating units then it values units on an ‘offer basis’.
Under an offer basis, offer and bid prices for units are derived from the appropriation
price.
The appropriation price is the amount of money that the company should put into the fund in
respect of each unit created in order that the interests of existing unit-holders are not
affected.
It can be calculated as the net asset value of the fund on an ‘offer basis’ divided by the number of
units in the fund (prior to the creation of the units concerned).
This net asset value is the offer price value per unit of the fund’s underlying assets, adjusted in
respect of the fund’s current assets and liabilities, accrued income, and acquisition dealing
expenses.
The expropriation price is the amount of money that the company should remove from the fund
in respect of each unit in order that the interests of existing unit-holders are not
affected.
Calculated as net asset value on a ‘bid basis’ divided by number of units ...
… where net asset value is the bid price value of units, adjusted for current assets and liabilities,
accrued income, and sale dealing expenses.
If the company wants to avoid switching from one basis to the other too frequently then it will
probably need to create a ‘management box’.
Given the basis (offer or bid), the company will calculate the unit price as:
 offer price: the appropriation price (on an offer basis) or expropriation price (on a bid
basis), plus initial charges, rounded up
 bid price: the appropriation / expropriation price rounded down.
The above prices will need to be rounded. Some companies round the offer price up and the bid
price down. However, it is now increasingly common for companies to round to the nearest
decimal place.

13.2 (i) Management box
The ‘management box’ for a fund consists of units that have been created but are not owned by
policyholders, at any point in time. These units are therefore owned by the life insurance
company itself. [1]
(ii) Advantages and disadvantages
The primary advantage of the management box is that it enables the company to maintain a
steady offer or bid basis, rather than having to change too frequently the basis on which it prices
units. [1]
This achieves ‘broad equity’ between different unit-holders (as the basis is only changed if there is
a significant cashflow movement against the existing basis) and avoids unnecessary and artificial
volatility in published unit prices. [1]
By using a box, unnecessary creation and cancellation of units is avoided. [½]
The main disadvantage to the company is that it will be exposed to the investment risk of holding
a large number of units for its own account. [1]
Although with careful management it is a potential source of additional profit. [½]
In addition there may be problems with control – setting down guidelines as to how large a box is
maintained, what to do in the event of severe market movements or large unit purchases /
encashments. [½]
[Maximum 3]

SP2-14: Models (1) Page 1
Models (1)
Syllabus objectives
4.1 Describe the main features of a life insurance model.
4.1.1 Outline the objectives and basic features of a life insurance model.
4.1.2 Compare stochastic and deterministic approaches.

Page 2 SP2-14: Models (1)
0 Introduction
This and the next chapter cover the subject of modelling. In this chapter we look at the objectives
and basic features of a life insurance model, before looking in detail at specific examples of
modelling in Chapter 15.
Constructing a model and the setting of assumptions (that will be used for the model) are
interrelated processes and generally should be considered together. The chapters on setting
assumptions (Chapters 17 and 18) are therefore very relevant to this topic.
Any modelling investigation is essentially a variation on a common theme, namely a projection of

the company’s balance sheets and profit and loss accounts.

1 Objectives and basic features of a life insurance model
1.1 The prime objective

The prime objective in building a model is to enable the actuary advising a life insurance
company to give that company appropriate advice so that it can be run in a sound financial
way. Models will therefore be used to assist in the day-to-day work of the company and to
provide checks and controls on its business.
Considerable actuarial judgement may be required in the choice of the model to be used as
well as the selection of inputs, including deciding on the number and type of the underlying
model points.
1.2 Basic features of a model

The requirements and basic features of a generic cashflow projection model are described
in Subject CP1.
Question
State the operational issues to be considered when constructing a model.
Solution
 The model should be adequately documented.

 The workings of the model should be easy to appreciate and communicate. The results
should be displayed clearly.
 The model should exhibit sensible joint behaviour of model variables.
 The outputs from the model should be capable of independent verification for
reasonableness and should be communicable to those to whom advice will be given.
 The model must not be overly complex so that either the results become difficult to
interpret and communicate or the model becomes too long or expensive to run, unless
this is required by the purpose of the model.
 The model should be capable of development and refinement.
 A range of methods of implementation should be available to facilitate testing,
parameterisation and focus of results.
 The more frequently the cashflows are calculated the more reliable the output from the
model, although there is a danger of spurious accuracy.
 The less frequently the cashflows are calculated the faster the model can be run and
results obtained.

1.3 Basic features of a life insurance model

All life office models include some element of cashflow and profit projection. Given this very
basic commonality we can describe a number of reasonably distinct variations. These are
distinguished both by their purpose and structure: as might be expected, the design of any model
should be closely linked to its function. The four ‘types’ are:
1. Single policy profit test model. This projects the expected cash and profit flows from a
single policy from the date of issue. It is a key model for pricing and product design.
2. New business model. This projects all the expected cash and profit flows arising from
future sales of new business. It is useful for assessing future capital requirements for the
new business, and the overall return on capital achieved from future sales.
3. Existing business model. This is a cash and profit flow projection from all the existing
business a company has in force at a particular time. This is important both as a means of
assessing the intrinsic value of the existing business (the embedded value), and also for
testing the solvency of the company’s existing business.
4. Full model office. This is essentially the sum of (2) and (3). This model is of fundamental
importance in assessing the impact of future management decisions, of all kinds, on the
future financial development of the company.
These four models differ almost entirely on the basis of which policies are included in the model.
The models will also tend to differ in terms of the focus of their output, which of course is related
to the functions of the model. For example, for the profit test the focus is on the annual
emergence of profit, whereas users of the full model office will be primarily (but not solely)
interested in projections of the company’s supervisory balance sheet and the projected
distributions of profit.
The following sections refine the features of a generic projection in the context of these types of
life insurance models.
Projecting cashflows and profit

In the specific context of life insurance business, the model needs to allow for all the
cashflows that may arise, which will depend on the nature of the contract(s), in terms of
premium and benefit structure and any discretionary benefits such as non-guaranteed
surrender values or bonuses.
By discretionary benefit we mean any benefit whose level depends on the decision of the
company, rather than, for instance, the guaranteed sum insured which is defined in the policy
conditions. For example, in a model involving bonuses, the model will need to incorporate an
algorithm to ‘decide’ what bonuses should be awarded given the projected investment results.
This algorithm will probably make reference to asset share, in which case the model would need
to calculate projected asset shares for policies.

Allowance for supervisory reserves and solvency capital

The model also needs to allow, where appropriate, for the cashflows arising from any
supervisory requirement to hold reserves and solvency capital.
Here we can distinguish between what we might refer to as physically real cashflow, and notional
cashflow. The real cashflow includes, primarily: premiums and investment income as positives
(from the company’s perspective); payments to policyholders, commission to agents, expenses
and tax as negatives.
If there is a supervisory requirement to hold reserves (this can be safely paraphrased as ‘always’)
then the life insurance company will need to fund the establishment of these reserves. This is
done by contributing money to reserves from the cashflow or, initially, from the company’s free
assets or other source of capital; this item is the ‘increase in reserves’ and is a negative from the
company’s perspective. At maturity or earlier claim the reserves will be released to help pay the
appropriate policyholder benefit; this will be a ‘decrease in reserves’ (or ‘release of reserves’) and
will be positive. These flows are examples of notional cashflow, ie where the flow does not
involve a physical exchange of money. The real and notional cashflows together combine to form
the profit flow. The investment income from the supervisory reserve will also be included as a
positive contribution to the profit flow.
The total profit arising over the year, can be calculated as:
(At 1  Vt 1 )  (At  Vt )  (At 1  At )  (Vt 1  Vt )
where Vt is the total supervisory reserve at the beginning of year t.
The investment return on the reserves is effectively included in (At 1  At ) , which includes the
investment return on all the assets. The right hand side of the formula is showing the calculation
in terms of the total cashflow (including investment returns) less the increase in reserves.
(The Core Reading is simply saying that the model will need to include an allowance for the
supervisory reserving requirements in its projection.)
In addition there might be a supervisory solvency capital requirement to cover. This works in the
same way as the reserves above: just as the policy cashflow must fund the establishment of
certain reserves which are then released at the end of the policy’s life, so the policy cashflow
might also need to fund the establishment of solvency capital. The required solvency capital
would then be included in the value of Vt in the above formula.

Allowance for interactions

The cashflows need to allow for any interactions, particularly where the assets and the
liabilities are being modelled together.
It is absolutely vital that the asset and liability models be completely consistent. To achieve this
the model should be dynamic, meaning that the asset and liability parts are programmed to
interact as they do in reality. A good example of such interaction is the effect of investment
returns on declared bonus rates. For example, understating future bonus distributions in the
model, relative to assumed future investment returns, could give a very over-optimistic projection
of the company’s future solvency.
Another very important example of such interaction is the way in which the supervisory reserves
will need to be consistent with the projected investment conditions. In particular, the valuation
rates of interest assumed by the model for this purpose will be a variable, projected by the model
itself, and should be the interest rate(s) that the company would realistically be expected to
assume, at that future time, if the investment conditions projected by the model were actually to
occur.
A further dynamic link, related to this, concerns the response of the company’s investment
strategy to changing conditions. In particular, should the model project worsening solvency
margins (ie smaller free assets in excess of supervisory requirements), the model could
realistically incorporate the management’s response to this scenario: this may, for example,
involve switching to more closely matching assets.
Incorporating dynamic links is important for all modelling, but it is particularly vital when
modelling stochastically (see below). A deterministic model can have its assumptions tweaked to
reflect realistically the dynamic responses required for a given deterministic (single) scenario, but
this would be impossible stochastically where variables are changing yearly and ongoing
responses need to occur automatically as each simulation is run.
Allowance for guarantees and options

Where the business being modelled includes health options, for example an option to effect
a new term assurance contract without providing further evidence of health, the potential
cashflows from such options need to be allowed for.
The implications of these options and guarantees for the supervisory reserves should also be
allowed for in our models.
(We describe the costing of such options later in the course.)
The ability to use stochastic models and simulations needs to be allowed for, where
appropriate, in order to assess the impact of financial guarantees such as maturity
guarantees or guaranteed annuity options.
The valuing of financial guarantees is also covered more fully later in the course. However, the
example we described in the second risk chapter (Chapter 11) gives a reminder of how stochastic
models are used here.

1.4 Choosing between stochastic and deterministic approaches

The model as described so far may have been created entirely along deterministic principles –
that is, for any important variable, the actuary will assume some single value and proceed with
that. This will not give adequate information about the resilience of the company to, for instance,
asset value fluctuations given a certain level of guaranteed benefit.
With a stochastic model, resilience to investment variation could be studied by varying the market
value and reinvestment yield of assets according to some appropriate probability function. Each
run of the model based on one such sequence of market value movements and reinvestment
yields will give an answer of ‘always solvent in the projection period’ or ‘insolvent in year x’.
Doing, for example, a thousand such runs to give a thousand such answers will give a very clear
idea of the company’s resilience.
Withdrawal levels and mortality levels might also be modelled stochastically.
Often, a stochastic approach is preferable (when modelling) as it has the following desirable
features, as compared with a simpler deterministic approach:
 The method allows a probability distribution to be assigned to one or more of the
unknown future parameters.
 A positive liability can be calculated where a deterministic approach might otherwise
produce a zero liability, eg where the cashflow includes a minimum guaranteed
benefit.
This will be the case when we are considering the valuation of financial options (such as
guaranteed annuity options – see Chapter 23).
 The future parameters may be assumed to vary together as a dynamic set, which is
particularly useful for modelling with-profits business.
Stochastic modelling does have some disadvantages:

 time and computing constraints – so stochastic modelling work might be done with a very
simplified version of the model, possibly making it unrealistic
 the sensitivity of the results to the (deterministically chosen) assumed values of the
parameter(s) involved.
This second point can therefore lead to spurious accuracy: the great complexity of the model may
give the appearance of highly precise answers, but if we have little confidence in the parameter
values then we can have little confidence in the answers being correct. It’s the old saying: rubbish
in equals rubbish out.
The circumstances in which a deterministic approach may be appropriate include:

 The actuary is satisfied that similar results could be obtained as if a full stochastic
projection was used, eg the possible outcomes form a symmetric distribution and
information is only required on the expectation, or if a specific scenario is being
tested within a simple cashflow model.
 A quick, independent test is required to see that the results of a stochastic
projection are reasonable.

A single deterministic result – using average assumptions – together with a series of further
deterministic calculations on amended assumptions, may in simple cases be used to
provide upper and lower bounds on the corresponding stochastic result.
This is essentially the process of sensitivity testing, which we describe more fully at the end of
Chapter 15 (and where we also revisit the circumstances under which deterministic or stochastic
approaches would tend to be used).
Question
State with reasons whether a deterministic or a stochastic model would be used for the following
purposes:
(a) Assessing the estimated loss to a portfolio of term assurance policies if there was an
influenza epidemic as serious as that in 1918.
(b) Estimating the expected value and variance of term assurance claims next year.
(c) Estimating the cost of adding a minimum maturity value to a unit-linked product.
(d) Estimating the probability of insolvency due to an unforeseen AIDS epidemic.
(e) Calculating the bonus earning capacity of a particular block of with-profits policies.
(f) Estimating the free assets required to support the cost of smoothing with-profits bonus
distributions in future years.
Solution
(a) Probably deterministic, as only one answer is required.
(b) Stochastic, as a variance is involved, which necessitates a distribution.
(c) Stochastic, as we are costing a financial guarantee that has a positive cost when unit
growth rates are low, and a zero cost when unit growth rates are high. A deterministic
approach might easily lead to a zero value for this guarantee, even when its true expected
value is positive.
(d) Stochastic, as a variance will be needed to assess the likelihood of not being able to meet
a specific fixed target.
(e) Probably deterministic, because bonus rates generally reflect average future profits.
(f) Stochastic, as it is essential to model the size of the fluctuations in profit or loss from year
to year, as the capital is specifically needed in order to cover these fluctuations.
A deterministic or closed-form approximation approach may be needed within a stochastic

projection to avoid having a ‘nested’ stochastic model (see dynamic solvency testing in
Chapter 15).

For example, we have seen above that it may be preferable to calculate the value of liabilities
which include options and guarantees using a stochastic model. However, if we then want to
project the liabilities in many stochastic simulations we will end up with simulations nested within
simulations. To avoid this complexity we may prefer to value the liabilities using a simpler
approach, eg the Black-Scholes equation is a closed-form alternative to valuing options using
simulations.
1.5 Calibration of stochastic models

For a stochastic model in which the economic assumptions vary, there are different
approaches to the setting (‘calibration’) of these parameters. The most common are as
follows:
 Risk neutral (also known as ‘market-consistent’) calibration – which in many

countries is typically used for valuation purposes, particularly where there are
options and guarantees. The focus of these calibrations is to replicate the market
prices of actual financial instruments as closely as possible, using an adjusted (risk
neutral) probability measure.
The first step in a market-consistent calibration is to choose a number of financial

instruments (usually derivatives) for which the price is known. A model is then built that
can project the cashflows from these instruments in a range of scenarios. The parameters
are then chosen in such a way that the average present value of the cashflows from the
modelled simulations is sufficiently close to the known market price.
The idea is that if the model can closely reproduce the observed prices of quoted assets,
then the model should also provide market-consistent values for unquoted asset or
liability cashflows.
 Real world calibration – typically used for projecting into the future, for example for
calculating the appropriate level of capital to hold to ensure solvency under extreme
adverse future scenarios at a given confidence level. The focus of these
calibrations is to use assumptions which reflect realistic ‘long-term’ expectations
and which consequently also reflect observable ‘real world’ probabilities and
outcomes.
With the real world calibration we determine the model parameters using our
expectations of the future. These assumptions are then used to project the values of the
assets and liabilities under each stochastic scenario.
To see the difference between the two calibrations, consider two investments: bonds with a
market price of 100 and equities with a market price of 100.
If we used the risk neutral calibration, then the average present value of the cashflows from our
model’s simulations would be 100 for each investment, ie the model has replicated the observed
market prices.
However, if instead we used the real world calibration, we would expect that on average the
simulated cashflows from the equities would be higher than the bonds (as this is what usually
happens in real life).


Chapter 14 Summary
Different types of model
 profit test model
 new business model
 existing business model
 full model office
It is important to know how the various models would be used, as described in Chapter 15.
Basic features of life insurance models

 Involves projecting cashflows; need to allow for all cashflows that may arise,
including discretionary benefits.
 Cost of setting up supervisory reserves and required solvency capital needs to be
allowed for in order to calculate profit flows.
 Proper allowance must be made for guarantees and options: it is likely that a
stochastic modelling facility will be necessary for this, particularly for financial
guarantees.
 Allow for interactions (dynamic links) and correlations between variables.
Stochastic versus deterministic

A stochastic approach may be preferred, because:
 we need to value financial guarantees and options
 we would wish to see the likely distribution of outcomes, not just a single estimate
 the interaction between variables can be explicitly included, enabling the effect of
the interactions to be assessed
 we need to estimate a probability.
Deterministic modelling can be used:

 with sensitivity testing, in order to get an approximation to a stochastic result
 where the result obtained would be very similar to (or, if not, more prudent than) a
stochastically produced result
 as a check on a stochastic model.
Stochastic models also suffer from high sensitivity to the chosen parameter values, with the
risk of spurious accuracy. So we should only use stochastic modelling when the variable can
be reliably modelled by a well-defined probability distribution.
The parameters of a stochastic model can be set using either a risk neutral (market-
consistent) calibration or a real world calibration.



14.1 Outline the basic features of an actuarial model to be used to determine the annual premium
Exam style
rates for a new life insurance product. [7]



14.1 This will be an individual policy model. [½]
The model will project the expected cashflows arising in every future policy year, for any specified
model policy. [½]
The cashflows would be projected using appropriate (realistic) assumptions for the relevant
demographic, financial and economic variables. [½]
The assumptions need to be consistent with each other. [½]
The model will enable the projected supervisory profit to be calculated for each policy year, by
deducting the expected increase in reserves from each year’s cashflow, and adding the interest
earned on the reserves over the year. [1]
The reserves would include the solvency capital that the company is required to hold in respect of
this business. [½]
The model may need to operate stochastically in order to value any financial options or
guarantees. [½]
The model must also be capable of including the expected cost of any health options provided
under the policy. [½]
In order that a premium can be obtained, the model should allow the net present value of the
future profits to be calculated, by discounting expected profits at the appropriate risk discount
rate. [½]
Other relevant profit measures should also be produced (eg discounted payback period, internal
rate of return). [½]
The model must also allow changes to assumed premium rates, charging rates and charging
structures to be made easily. [½]
In order to be able to test the profit to variations in future experience, the model would need to
allow easily for:
 changes to any of the assumptions (including supervisory reserving bases) [½]
 changes to any of the features of the model policy (eg different ages, policy terms, etc).
[½]
As with any model, it needs to be:

 valid, rigorous and well documented [½]
 not unnecessarily complex [½]
 capable of independent verification. [½]
[Maximum 7]





using the course.

Models (2)
Syllabus objectives
 capital management
4.1 Describe the main features of a life insurance model.
4.1.3 Explain the use of sensitivity analysis.
4.2 Demonstrate the different uses of actuarial models for decision-making purposes in
life insurance, including:
 pricing products
 projecting solvency
(Only part of syllabus objective 5.1.1 is covered in this chapter.)

Page 2 SP2-15 Models (2)
0 Introduction
We now turn to considering some of the uses of the models we described in Chapter 14.

1 Pricing
1.1 Determining premiums or charging rates

Here we examine how the model can be used for profit testing work on individual products,
thereby assisting in product pricing.
The model can be used to determine a premium, or charging, structure for a new or existing
product that will meet a life insurance company’s profit requirement.
The life company should be continually monitoring the validity of premium rates on existing
business. They will need to be re-examined if the expected future experience has changed from
that foreseen when the premium basis was originally devised.
Embedded value work should also indicate any products that need to be re-priced. This is
discussed under ‘profitability of existing business’ below.
A number of model points will be chosen to represent the expected new business under the
product. In the case of an existing product, the profile of the existing business, modified to
allow for any expected changes in future, can be used to obtain the model points. For a
new product, the profile of any similar existing product combined with advice from the
company’s marketing department would be used.
The existing business used for model point selection is likely to be that written recently and then
adjusted for any expected changes in the mix of business in the near future.
Question
A relatively new actuarial student has tested the profitability of an existing contract on currently
expected parameter values but has used the last ten years of new business as a basis for model
point generation. They argue that this gives a larger and hence more valid data base from which
to group policies.
Describe how a significant influence on product profitability is likely to have been misrepresented.
Solution
The main risk here is that the average assumed policy size will be wrong (too low in a non-zero
inflationary environment). Given a certain level of fixed per-policy administration expenses,
product profitability will be understated.
It could be argued that the ten years of data could form a better basis if the student were to allow
explicitly for inflation. This is true as far as the average premium for all policies is concerned.
However it will misrepresent reality if the premiums for certain age groups or ‘term groups’ (or
any other sections of the data) have changed relative to the other new business premiums.
For each model point, cashflows will be projected, allowing for reserving and solvency
capital requirements, using a set of base values for the parameters in the model.

Holding solvency capital for a policy will impinge significantly on the rate of return earned on the
capital required to write a policy. This arises from the assets underlying the solvency capital
earning less than the rate of return required to be earned by the policy (ie the risk discount rate).
This difference is often referred to as ‘cost of solvency capital’.
Example
Consider a single premium policy with premium 100. The acquisition expenses and the difference
between premium and valuation basis are such that the reserves are 100 and the capital required
to write the business is 10 (pre-required solvency capital). The policy earns a return of 12% pa on
this capital. The assets supporting the policy reserves earn 6% pa net of tax.
What is the impact of holding solvency capital set at 4% of reserves? (Assuming that the reserves
conveniently remain at 100 for the policy duration.)
The capital required in order to write the business is now 14. The impact on the rate of return
depends entirely on what the solvency capital assets earn:
 If they earn 12% then the rate of return of the policy net of cost of solvency capital
remains 12%. (However if the company can consistently earn the same rate on normal
investments as it can on writing life business then there is no point in continuing writing
life business.)
 If they earn 6% (net of tax) then the total rate of return is 10.3%. (An investment of 10
earning 12% plus an investment of 4 earning 6% make an investment of 14 earning
10.3%).
 If they earn 4% (net of tax) then the total rate of return is 9.7%.
In practice the situation will often be as in the second scenario in the example above, where
solvency capital assets earn a similar gross rate to the assets underlying the basic policy reserves
but that income is then fully taxed.
The net projected cashflows will then be discounted at a rate of interest, the risk discount
rate, which is discussed further in Chapter 17. It may, for example, take into account:
 the return required by the company, and
 the level of statistical risk attaching to the cashflows under the contract, ie their
variation about the mean as represented by the cashflows themselves.
In other words, we would calculate the present value of the future profits generated by the
model, which will be the expected present value of the future profits (from a single model point)
and which is usually just referred to as the net present value.
Net in this context means that all positive and negative cashflows are combined to give the profit
for each time period of the projection. These profits should also be net of appropriate taxation.
Hence the company’s required return is also net of tax.
In theory, a separate risk discount rate should be applied to each separate component of
the cashflows, as the statistical risk associated with each component will be different.

We will discuss this idea further in Chapter 17.
The premium (or charges) for the model point can then be set to produce the profit required
by the company.
This implies a charging structure such that each model point satisfies the profit criterion. For
example, this might be to set the premium (or charges) such that the net present value of the
product is equal to, say, some desired proportion of the initial commission payable under the
contract.
We will now look in detail at the various profit criteria that can be used for pricing.
1.2 Profit criteria

A profit criterion is often a single figure that tries to summarise the relative efficiency of
contracts with different profit signatures. By applying a profit criterion to different
contracts with different profit signatures and ranking the results in order, it may be possible
to say with confidence which contract makes most efficient use of a company’s capital.
The profit signature of a contract is the sequence of profits over time from inception to
termination. It would normally be presented graphically. The profit signature itself is the
‘definitive guide’ to a product’s profitability but is unwieldy to use (especially when comparing
different products), and to present.
Among the profit criteria that could be used are:

 net present value, expressed in different ways
 internal rate of return
 discounted payback period.
Net present value

Discounting the profit signature at the risk discount rate produces a ‘net present value’.
Given a choice between the future cashflows from two different investments, economic
theory states that an investor should choose the one with the higher net present value. This
choice is optimal, and cannot be bettered. Another way to put this is that the first priority
for the managers of any company is to maximise the net present worth of the company.
This implies that net present value is the best profit criterion to use, and that if any other
profit criterion disagrees with it a company should go with the net present value.
This important result depends on several assumptions, including that:

 there is a perfectly free and efficient capital market
 when two risky investments are compared, each is discounted at a risk discount rate
appropriate to its riskiness.
Question
Explain how the second point translates into the world of a life insurance pricing actuary.

Solution
This means that different products should be valued using different risk discount rates
appropriate to their inherent riskiness.
There are also some practical points to bear in mind when using net present value,
including:
 It is subject to the law of diminishing returns. If it were not, then a company that
could sell one policy with positive net present value could sell an unlimited number
of policies and increase the value of the company without limit.
This can be illustrated by thinking about how it might apply in a real world context. Suppose a
company is selling a product with a net present value of £100. What is to stop it selling this to the
entire population of the country? Or selling more and more of the same product to existing
policyholders? The problem is this: after the product has been sold to all those who actively
wanted it, the company then has to seek out other customers and persuade them to purchase it.
This will increase sale costs and reduce the net present value of the policies. Eventually, as the
market becomes more and more saturated, the sale costs increase to the extent that the net
present value of new policies is zero.
 It says nothing about competition. There is no point in designing a contract with a

high net present value if it cannot be sold.
This is true, but it is also true about the other profit criteria here. Another way to think about this
is to say that we only want to study the profitability of saleable contracts.
How should net present values be expressed?
The net present value is by itself telling us relatively little – for instance, we could double the net
present value of some specimen contract by doubling the premiums of the model points used in
the profit test (in fact, if some of the expenses involved are fixed, the net present value should
more than double). In order to use net present value results to compare different product
designs, it is therefore necessary to find some measuring standard by which the products’ results
can meaningfully be compared.
Having said that, looking at the net present value in isolation can indicate one very important
thing.
Question
State what important aspect is indicated by the net present value in isolation.

Solution
The sign of the net present value will immediately show if the cashflow has met the risk discount
rate target. If the net present value is positive (or zero) then the rate of return on the cashflow
exceeds (or is equal to) the risk discount rate.
One approach is to express net present values in a way that reflects the effort that would be
expended on selling a policy. One such measure is the amount of initial commission that
rewards the salesperson (if applicable).
This ties in with the intuitive approach that the sales motivator, commission (or whatever sales
costs apply to the distribution channel), should be in line with the profitability of the contract that
is being sold. This will mean that both insurance company and salesperson are rewarded in
proportion to the amount of effort the salesperson makes. It also means that, provided all
products are priced to the same criterion, the insurance company’s profits will still be in the same
proportion to the effort expended.
However, the market for any insurance product is finite, and any one company can probably
capture only a small share of that total market. The policyholder measures the cost of
insurance in terms of the premiums that he or she pays, and industry trade associations
often measure the size of the market in terms of the premium income of insurance
companies, among other things. Therefore, another useful measure of profit is in terms of
the premium income, since this relates to the size of the market.
The net present value can, therefore, be expressed as a percentage of the present value of
the premiums that will be paid under the policy. All other things being equal, if a company
can sell the same volume of premiums, ie capture the same market share, for a higher net
present value then it should aim to do so.
An advantage of this method is that it enables the company to focus its efforts on increasing its
market share, in the knowledge that profitability will be appropriately increased as a result.
Question
‘When measuring net present value in relation to premium size, we use total discounted future
premium income. So when measuring net present value in relation to sales effort, we should use
total discounted future commissions rather than just the initial commission.’

Solution
In theory the statement is correct. However it is much easier to use just the initial commission
rather than performing any discounting. The loss in ‘accuracy’ is very slight for the following two
reasons:
 Normally, the initial commission is much greater than renewal commission. So

commission payments subsequent to the initial commission are much less significant than
are the post-initial premium payments relative to the initial premium.
 When using premium as a measuring standard, discounting future premiums gives a

relatively steady relationship of ‘net present value / discounted future premiums’ as term
varies. The initial commission will itself normally reflect the term of the policy. Thus a
longer term policy with more premiums from which to generate profit will also have a
larger commission. The relationship ‘NPV / initial commission’ should therefore be
reasonably stable with varying term.
Internal rate of return

This is defined as the rate of return at which the discounted value of the cashflows is zero.
A company will usually prefer a contract that has a higher internal rate of return. However,
the internal rate of return does not always agree with net present value.
The last statement means that a cashflow with a higher internal rate of return than another
cashflow will not necessarily have a higher net present value than that of the other cashflow.
Net present value may be more reliable in some cases, for example:
 If there is more than one change of sign in the stream of profits in the profit
signature, the internal rate of return will not usually be unique.
 As noted above, the net present value can be related to useful indicators of the
policy's worth to the company, in terms of sales effort or market share. There is no
way to do this with the internal rate of return.
 If a policy makes profits from the outset, then the internal rate of return may not
even exist. The net present value always exists, however.
This last point is true, but it is unfortunately rare that a policy will make profits from the outset.
So as a rule of thumb contracts will always have an internal rate of return. The most important
feature here is the fact that the net present value can be related to other measures, while the
internal rate of return cannot.
Discounted payback period

The discounted payback period is the policy duration at which the profits that have emerged
so far have present value zero, ie it is the time it takes for the company to recover its initial
investment with interest at the risk discount rate.
A company with limited capital might prefer to sell contracts with as short payback periods
as possible.

Question
A five-year term assurance contract comes in two forms: single premium and annual premium.
Explain which, if either, of these designs is likely to have the shorter discounted payback period
(DPP).
Solution
The single premium version is likely to have a significantly shorter DPP.
With the annual premium version, the very high initial expenses (mainly initial commission and
underwriting) will almost certainly exceed the first year’s premium. In theory, loadings included
in the valuation premiums could offset this initial loss by reducing the valuation reserve; however,
supervisory reserving requirements may prevent this (by not allowing us to hold negative
reserves). So there will be a large loss (strain) on day 1, that will only be repaid as the prudent
reserves are released as profit later in the term.
For the single premium version, all the expense loadings are delivered on day 1 and should cover
the actual initial expenses incurred straight away. The supervisory reserve is bound to be
positive, so we will not have to suffer any additional strain of making a theoretical negative
reserve to be at least zero (though some strain – and hence extension of DPP – is likely due to the
prudence required by the reserving basis).
The difference between the two types will be less if the commission structures are different. For
example, if the annual premium version pays lower initial commission than the single premium
version, perhaps replaced by some amount of level renewal commission, then the DPP for the
annual premium contract will not be as long as before.
There will be little difference in DPP on account of benefit payments, even though these will on
average be payable at the same point during the policy term under both contracts. Although the
single premium policy delivers its benefit loadings all on day 1, the company will have to set up
(prudent) reserves to cover the claim outgo at the same time, thereby delaying the emergence of
any profit until after the benefits have been paid out, as with the annual premium version.
The discounted payback period is of most importance for longer-term policies, where the initial
capital strain is large, and the potential variation of the discounted payback period is that much
greater.
The discounted payback period will not usually agree with the net present value because it
ignores all the cashflows after the discounted payback period.
That is, if comparing two sets of cashflows, that with the shorter discounted payback period will
not necessarily have the greater net present value.

Which criterion?
Of the three criteria examined, the net present value is the most frequently used. Its advantages
over the internal rate of return have already been mentioned. The discounted payback period
does not by itself contain much information.
A common approach would be to use the net present value (expressed in terms of initial sales
costs) as the prime criterion, and also to make reference to the discounted payback period.
Hence, for a given net present value, we would choose the product design which had the shortest
discounted payback period. The discounted payback period is therefore more use as a criterion in
product design than in product pricing, though in practice the two things (ie design and pricing)
would very often be performed at the same time, using the same models.
1.3 From model points to premium rates

It is possible for the desired level of profitability to be reached in aggregate (scaled up for
the expected new business mix and volume), without requiring every single model point to
be profitable. However, if certain model points are unprofitable, the aggregate profitability
of the business is then exposed to changes in mix (as well as volume) of the contracts sold.
An excellent way of looking at the aggregate profitability of a new (or re-priced) product, is to look
at the projected impact on the company’s embedded value following the changes (see Section 2
for a definition of embedded value). A full model office would be needed for this projection, as
future new business (both volumes and mix) will have a significant effect on what the future
embedded value might be. The company will wish to project the progress of the embedded value
for several years, and compare this with what could happen were different management
decisions to be taken. Outcomes would, of course, need to be thoroughly tested for sensitivity to
different experience, especially to different volumes and mix of business, as mentioned in the
above Core Reading. (We return to the subject of sensitivity testing in Section 4, below.)
Once acceptable premiums (or charges) have been determined for the model points,
premiums (or charges) for all contract variations can be determined.
For instance the life company might use the model point cashflow approach for pricing
endowment assurance policies with terms of 5, 10, 15 and 20 years for a variety of ages (say 20,
30, 40, 50) and premium sizes (£20, £50, £150, £250, £500). Intervening terms are then derived
by interpolation. The company ends up with a set of premium rates that should satisfy the profit
criterion for every term, without the need to profit test all combinations of age, premium and
term.

1.4 Marketability
The premiums (or charges) produced need to be considered for marketability. This might
lead to a reconsideration of:
 the design of the product, so as either to remove features that increase the riskiness
of the net cashflows, or to include features that will differentiate the product from
those of competing companies
 the distribution channel to be used, if that would permit either a revision of the
assumptions to be used in the model, or a higher premium (or charges) to be used
without loss of marketability
 the company’s profit requirement
 whether to proceed with marketing the product.
The above considerations would also include re-examination of the assumed expenses involved,
notably the acquisition costs (eg commission), the marginal administration costs and the required
expense contribution to fixed overheads.

The net cashflows in respect of the model points, appropriately scaled up for the expected
new business under the product, will be incorporated into a model of the business of the
whole company.
‘Scaled up’ here means multiplied up in order that the individual model point results represent
the expected new business. A full model office is, of course, required for this purpose.
The company will need to assess the amount of capital required to write the product.
The actuary can further assess the impact in capital management terms (either on a
regulatory or economic capital basis) of writing the product, by observing the modelled
amount and timing of cashflows. This may, if capital is a problem, lead to a reconsideration
of the design of the product to reduce, or amend the timing of, its financing requirement.
Both the new business and full office models will be useful for this purpose. The capital
requirement should also allow for any supervisory minimum solvency capital needed.
Question
State at least three disadvantages of writing capital intensive business.

Solution
All other things being equal, capital intensive business will give a lower rate of return. (But this
answer does not really count since the question implied ‘at parity of rate of return’.)
It could lead directly to a problem with supervisory solvency.
Even if solvency is not immediately threatened, the reduction of capital could weaken the
company’s resilience to falls in the value of assets.
It will reduce the company’s apparent financial strength (often measured simply as free assets in
relation to the reserves).
It represents an opportunity cost, in that the company could use the same amount of capital to
write more business which is less capital intensive.
It could lead unexpectedly to a problem with supervisory solvency if the company sells much
more of the business than anticipated.
Putting too many eggs in one basket – or rather too small a number of baskets. In theory, the
company will be less at risk to adverse experience if it uses capital to finance a large number of
non-capital intensive policies than if it uses it for a smaller number of more capital intensive
policies.
1.6 Assessing the return on capital

The net cashflows for the model points described in Section 1.5 can be grossed up for the
expected new business and used to assess the amount of capital that will be required to
write the product. This can be performed either on a regulatory or economic capital basis.
One-off development costs can be added to the extent that they have not already been
amortised and included in the expense cashflows used.
This gives the total capital requirement, which can be compared with the profits expected to
emerge from the product to determine the expected return on that capital.
This is equivalent to the work performed on one model point as described in the preceding
sections, but here a group of model points is examined in order to quantify the capital
requirement and the return on that capital for all of the production expected from that contract.
A new business model will be used for this.

2 Profitability of existing business

This section discusses the measurement of the total expected future profits from the policy
portfolio in force, using an ‘existing business’ model. This is done in order to calculate the
embedded value of a life insurance company. The embedded value may be defined as:
the value of the future profit stream from the company's existing business together with the value
of any net assets separately attributable to shareholders.
Embedded values are discussed in detail in Chapter 18. In this section we look only at the
mechanics of quantifying the value of the expected future profits from the in-force policy
portfolio.
To assess the profitability of existing business, the full policy data set may be used (ie the
modelling may be done on a policy by policy basis).
Calculating profitability individually for each policy should be more accurate than using model
points, but it will be more costly in terms of time and resources.
Alternatively, model points may be chosen to represent the existing business. In this case,
the model points used for any previous assessment may form a starting point, modified to
allow for changes due to the new business taken on by the company and the business
going off.
Another and more common approach is to redo the model point generation process from scratch
based on the policy portfolio as it appears now. This will be much less prone to error than
adjusting previous model point files.
The suitability of any model points used should be checked. Where, as will usually be the
case, the profitability of the business is being assessed at the same time as the supervisory
reserves are calculated, one check is to use the model points to determine the supervisory
reserves and then compare this value with the published value.
For each policy or model point, the present value of projected cashflows can be obtained as
in Section 1.1. The discounting would be done using an appropriate risk discount rate, with
similar considerations to those described in Chapter 17.
The risk discount rate to use for this purpose would normally be less than that used for product
pricing, or in a new business model. This is because some of the risks are reduced, in particular
those due to new business volume and mix.
Totalling across all policies (or totalling the scaled up results of each model point) will then
give the expected profit from the existing business.
The discounted value of these future profits is often referred to as ‘present value of future
profits’. Its prime use is in conjunction with the net asset value of the company to give the
embedded value. It is of limited use in isolation, although when broken down between product
lines it may be of some interest. Simply checking that the present value of future profits for a
product is positive rather than negative will give an indication that nothing is going badly wrong
with the financial management of that product.

Question
Explain why the present value of future profits (PVFP) for existing business would be expected to
be positive.
Solution
The PVFP for a model point that represents existing business should be positive. There are two
ways of looking at this.
First, retrospectively. At inception the PVFP should be zero or positive, otherwise the company is
knowingly writing unprofitable business. Immediately after inception significant acquisition and
management expenses will be incurred, and so the PVFP immediately after having paid for these
should be strongly positive. It will then gradually reduce to zero at maturity. So it is always
positive.
The second way of looking at the question is prospectively, with reference to the supervisory
solvency requirements. The PVFP for any existing business at time t will start with assets
allocated to the business to fully cover the supervisory solvency requirements. The solvency
requirements will take a prudent approach (eg by making prudent assumptions in the reserving
basis and/or by adding solvency capital).
The PVFP is then the discounted value of the future profit flows allowing for the release of
reserves and solvency capital over time. These profit flows will be projected using the (usually
best estimate) assumptions in the embedded value basis. We should expect positive profit flows
in the remaining years as the experience assumed in the embedded value basis should be better
than that assumed in the prudent reserving basis. Alternatively, if reserves are calculated on a
best estimate basis, the expected future experience in the embedded value basis should equal the
reserving assumptions, so that positive profits emerge as the solvency capital is released.
The above argument assumes no huge changes in the embedded value basis since the contract
was priced. If our expectations of future experience have become more pessimistic, the PVFP
could easily become negative.
It can be useful to look at the present value of future profits at the following levels of breakdown:
 by product
 by product classes (eg group vs individual)
 by distribution channel
 by subsidiaries, if appropriate
in order to compare their relative contribution to the company’s profitability.

3 Assessing solvency
A fundamental requirement for any life office is that it has enough money to meet its future
outgo, both from existing business and from new business that it may sell in the future. In current
money terms, this means establishing that the insurer holds enough reserves to cover the future
liabilities on its existing business, and enough extra money (free assets, or capital) to cover
everything else (such as the cost of smoothing bonuses and the cost of writing future new
business).
An insurer should assess the amounts and types of capital it needs given the amount of its
liabilities and the types of risks inherent in those liabilities. Given that the liabilities span a
long period of years, it is necessary to project the liabilities into future years, allowing for
any new-business plans, so as to assess the capital requirements. Effectively, ongoing
solvency needs to be tested, allowing for management actions (such as changes in bonus
and investment policy) where appropriate.
3.1 Measuring solvency

The solvency at a point of time of a life insurance company can be measured by comparing
the value of its liabilities with the value of its assets. For this purpose, there are two main
ways in which the values of the assets and liabilities can be determined:
 supervisory (or regulatory) values – these would be the values as determined for
supervisory reporting purposes
 economic values – these would be calculated on the basis of the expected
experience of the company or on a ‘market-consistent’ basis.
Refer to Chapter 20 for further discussion of supervisory and economic values.
In some jurisdictions, the two approaches may be equivalent or very close.
This would be the case if supervisory valuations are calculated on a market-consistent basis.
3.2 Static vs dynamic solvency testing

Some insurance regulators require a solvency projection as part of the regular supervisory
submission.
However, others may only require a determination of solvency to be made at a particular

point of time, ie a static solvency test. This will not enable the company’s actuary to assess
the ability of the company to withstand future changes in both the external economic
environment and the experience of the company. To do this requires a dynamic
assessment of future solvency, frequently referred to as dynamic solvency testing.
This will involve the projection of the company’s revenue account and balance sheet
forward for a sufficient period of years so that the full effect of any potential risks may
become apparent.
So, while the insurer’s current supervisory solvency valuation may show a surplus of assets over
liabilities, a projection of how that surplus might change in the future may reveal that it is likely to
run out or become negative in the next five to ten years, say. This knowledge should allow the
company enough time to change its future plans (eg for marketing, investment, or bonus
distribution) to avert this disaster.

A full model office, with a realistic representation of the liability portfolio by model points, is
required for this purpose.
For dynamic solvency testing, consideration needs to be given to the projection basis. The
projections could be done deterministically using expected assumptions, combined with
assumptions with margins to test the effect of adverse future experience.
Dynamic solvency testing is an important part of a life insurance actuary’s toolkit. When
performed deterministically, the aim is to see how future changes in long-term (or expected)
values of the experience parameters might affect the long-term security of the company. Hence
we might investigate the effect that a gradual fall in investment returns might have on the
projected balance sheet, or how the position will look if mortality rates improve at faster or
slower rates than in the base projection. This kind of information can be extremely useful for
informing management decisions.
An alternative approach is to use stochastic assumptions with simulation to assess the

level of probability of adverse circumstances occurring.
It must also be decided whether the projections should take into account only the existing
business of the company or existing business plus expected future new business.
Including new business is likely to provide the most useful assessment of ongoing
solvency, unless it is intended that the company will cease to write new business.
Question
Explain why including new business gives a more useful assessment of future solvency than using
just existing business.
Solution
There are two reasons:

 Realism – unless it is imprudent to do so, it makes sense to project in accordance with
what the company is actually going to do, so if the company is not intending to close to
new business then do not project assuming closure to new business.
 Prudence – due to the capital strain involved in writing new business, it will normally be
more prudent to assume new business continues, thereby testing the company’s ability to
finance future new business strain without becoming insolvent.
Question
Outline the disadvantages of incorporating new business into this projection model.

Solution
The disadvantages of incorporating new business into the projection are:

 levels of new business are difficult to predict accurately
 the mix of new business by product, premium size, age and term will also be difficult to
predict accurately
 the products themselves which the company will be selling in a few years’ time are not all
known ‘now’
 some actuaries might argue that, from the point of view of per policy expenses, it is more
prudent to assume closure to new business.
So, new business projections are subjective and can cause volatility when reconciling
projections from one period to the next.
It would be possible, using stochastic assumptions and simulation, to assess the

‘probability of ruin’ of a life insurance company. This could be done in the context of the
dynamic solvency testing described above. It would require the assessment of the
probability that at some point in the future the life insurance company would become
insolvent, either on a supervisory values basis or on an economic values basis.
If the supervisory valuation basis is stronger than the economic basis, then the company will be
more concerned about the probability of insolvency on a supervisory basis. This will usually be
the main risk criterion when performing full model office or existing business projections for the
purposes of decision making.
The future projected balance sheets may themselves require stochastic calculations. This
leads to a ‘nested’ stochastic model (ie one which requires stochastic projections within
each stochastic simulation). This can present significant modelling challenges and might
necessitate the use of approximations such as closed-form solutions (eg the Black-Scholes
formula) or proxy models (which are described further in Subject SA2).
A stochastic solvency projection can involve many thousands of simulations to project the assets
and liabilities into the future. However, to calculate the liabilities at a given point in time in a
particular simulation may require many thousands of further simulations (eg stochastic modelling
may be required to place a value on any guarantees or options at each valuation date). This leads
to the problem of simulations within simulations that we mentioned in Chapter 14. The
simulations to calculate the value of the options and guarantees at time t are ‘nested’ within the
simulations that project forward the liabilities for every year.
The computing power available may not be able to cope with the necessary thousands of
simulations each involving thousands of simulations (particularly given the time constraints for
producing the results). So approximations, such as the Black-Scholes formula to calculate the
value of any guarantees within the liabilities, may be required to speed things up.
An alternative assessment of the probability of ruin in respect of the existing business only
could be made by projecting forward in the simulations just the revenue account of the
company and seeing whether it ran out of assets before the last contract went off the books.

Because the projected value of the assets at any time is simply the accumulation of the net
revenue (cashflow) up to that time, then this is simply a matter of projecting the accumulated
assets to the point where the last contract has terminated. Should these projected assets be
positive, the company is (then) solvent; if they are negative, the company is (then) insolvent. The
proportion of all simulations which lead to insolvency is then an estimate of the probability of
actual, or absolute ruin with respect to the existing business.
It is only possible to calculate a probability of absolute ruin for a fund in which there is no new
business, or for which new business ceases at some future date.
Examining a closed fund and considering the probability of actual ruin would probably be the best
way to assess the impact of cashflow mismatching (which we refer to in the later chapter on
investment, Chapter 28).
3.3 The need for capital and the role of the estate
Finally in this section we review why the insurer needs capital, and briefly from where it arises.
Managing a company’s capital properly is a key aspect of the insurer’s overall ability to manage its
risks. The returns that the providers of capital ultimately receive are also significantly affected by
the company’s capital management strategy. As we have seen, capital and supervisory solvency
are very closely related concepts, and so the modelling we have been discussing in this section is
highly relevant.
Question
Explain what is meant by the estate of a life company.
Solution
The glossary defines the estate as (usually) ‘the excess of the realistic value of its assets over the
realistic value of its liabilities’, ie assets less liabilities on a realistic basis.
So this definition implies an internal realistic determination of solvency.
However, in any consideration of the issues affecting the estate – such as how much new business
the company can write – the company will always consider statutory solvency as well. Thus the
ultimate constraint is really supervisory free assets.
The fundamental reason why a life insurance company needs capital is so that it can
withstand adverse, often unexpected, conditions. In turn, the amount of capital will be
reflected in the company’s ability to:
 write new business, that is to meet the development costs of new contracts and the
capital requirements that arise from writing new business
 adopt a less restrictive investment policy so as to meet the investment aims of the
company.

Question
Describe the other advantages or uses of capital and the key disadvantage.
Solution
Capital can also be useful to:

 smooth surplus distributions – if the bonus declaration method allows any degree of
subjective decision by the actuary, then we might want to smooth bonuses by paying out
more than asset share when returns have fallen, and less than asset share when returns
have risen; to do this will require some capital
 reduce the need for reinsurance
 smooth dividend payments to shareholders
 allow the company to seize any profitable business opportunities as they present
themselves (eg buying some other company).
The key disadvantage of capital is that if it is not being efficiently used then it will lower the rate
of return earned by the company as a whole. For example, if the company has some capital which
is just sitting earning market yields then other capital will need to earn more than the risk
discount rate in order to give a total return of at least the required risk discount rate.
This capital may come externally, for example by the shareholders subscribing more
capital, but more typically it will come from within the company itself, that is from the
company’s free assets. As regards new business, the free assets will be drawn upon as
new contracts are developed and written and then will be augmented as the new contracts
themselves generate profits.
Question
‘In the case of a proprietary life company, the shareholders do not normally finance new business
strain because the capital can be provided by free assets.’
Discuss this statement.

Solution
Consider first without-profits business. Although the company’s free assets are physically
providing the finance on a day-to-day basis, in reality it is the shareholders who are financing the
strain. However, this is all implicit – the shareholders are in effect providing capital in that every
£1 of new business strain will reduce shareholder earnings in that year by £1. This will be
acceptable if the company can earn the shareholders’ required return on that £1 of capital.
Clearly we do not mean that the company will resort to a rights issue every time it wants to write
a new policy, or even launch a new product.
In the case of with-profits business where all surplus (in particular, expense surplus) goes to the
with-profits policyholders, the financing of new business strain is in effect being done by existing
with-profits policyholders. Again this is done implicitly: every £1 of new business strain will
reduce the surplus to be distributed by £1. Thus the surplus which eventually accrues in respect
of any with-profits policy should be reduced by some ‘cost of capital’ to represent that policy
paying something in respect of new business strain back to those who originally financed it.
Generalising further, we can see that financing is done by those who own the future profits of the
company, whether they be shareholders, with-profits policyholders, or the two combined. It is
equitable that they see some return on their ‘investment’ in new policies.
The ‘default option’ for a with-profits fund where all profits are distributed to policyholders,
should be that the free assets are left unchanged in real terms by the issue and eventual maturity
of a policy, compared with the situation if that policy had not existed. However, if this had always
been the case, then many present-day old insurance companies would not have the substantial
free assets which in fact they do. Historically such companies have kept more profits from
with-profits policies than was strictly necessary to repay the initial capital outlay on writing those
policies.
Question
Explain how a life insurance company would finance a new project requiring a large amount of
capital.

Solution
Ideally, the company would finance the project with existing free assets.
If this is not a possibility (because free assets are insufficient):

 In the case of a proprietary company, a rights issue will be a means of raising a large
amount of capital.
Otherwise, the options available would be:

 to sell off some assets – in this context the company might consider selling a subsidiary, or
 to raise loan capital – although regulation might make this difficult.
If desperate, a mutual could demutualise and then raise capital on the stock exchange.
By including new business in the dynamic solvency testing described above, the adequacy
of the internal capital in allowing the company to fulfil its new business plans can be
assessed. Similarly, by combining this with the asset-liability investigations covered in the
investment chapter (Chapter 28), the ability to invest in accordance with the aims of the
company can also be assessed.

4 Sensitivities
The results from the models used in Sections 1, 2 and 3 depend on the choice of model
points and the values assigned to the parameters in the model.
4.1 Sensitivity to choice of model point

If an adequate set of model points has been chosen, it should not be necessary to test for
the effect of model point error. However, situations may arise where a less than ideal
number of model points must be used, in which case the effect of a different choice should
be assessed.
This refers to the sensitivity of the model to model point selection. Normally the model will be
based on a large number of model points, enough to eliminate any significant model point error.
Often, if a large number of runs are required to test thoroughly various parameter sensitivities,
then the model might be recreated with a much smaller number of model points. In this case the
sensitivity to model point error is more likely to be an issue.
4.2 Sensitivity to parameters

The effect of mis-estimation of the parameter values can be investigated by carrying out a
sensitivity analysis. This involves assessing the effect on the output of the model of
varying each of the parameter values. When doing this, any correlation between different
parameters should be allowed for.
This mis-estimation refers to actual future experience being different from the parameter
assumptions. Since it will almost always be the case that the two are different, sensitivity testing
is very important in indicating the range within which we think the ‘best estimate’ (of our results)
will probably lie. In other words, we are assessing the range of possible outputs which might
‘reasonably’ reflect the uncertainty associated with the estimation of each parameter.
Allowing for correlation simply means testing under a scenario where two (or possibly more)
parameters are adjusted together, reflecting how they might move together in reality. For
instance, in a low future investment return scenario we would reduce the expense inflation
assumption correspondingly. Sensitivity testing in which several parameters are changed in a
consistent way is called ‘scenario testing’.
An important aspect of sensitivity testing is that it allows us to compare the relative financial
impact of the uncertainty that is associated with each parameter estimate. This is of vital
importance both in pricing and product design. For example, if a particular parameter is highly
uncertain, then we would try to design our product to be as financially insensitive as possible to
variations in that parameter. If that were to prove impossible, then it may be appropriate to
include higher margins for that assumption than for other, less uncertain, assumptions.
4.3 Sensitivity testing when pricing

The results from the sensitivity analysis will help, in the case of a model used for pricing, to
assess what margins need to be incorporated into the parameter values.

For instance, a without-profits whole life product has been priced on a yield assumption of 5% pa
but the sensitivity tests indicate a loss if the investment return drops below 4% pa. After
consulting with an investment colleague, the actuary attributes a probability of 0.1 to this
happening in the next ten years. They therefore decide to lower the investment return
assumption to 4.5% pa; sensitivity testing against this then shows reasonable results.
If the product profitability is overly sensitive to any factor, the results may indicate the need to
redefine the product or take some other measures. For instance, if it is too sensitive to increasing
withdrawal rates then a reduction in surrender values should be considered; if too sensitive to
mortality then the reinsurance programme could be revised.
Such sensitivity testing is also useful as a simple qualitative check that the model is functioning
correctly as far as the dependencies on those parameters are concerned. For instance, increasing
withdrawal rates should reduce discounted future profits (unless the company operates an
extremely penal surrender value basis).
4.4 Sensitivity testing for business in force

In the case of models used to assess return on capital and profitability of existing business,
and supervisory valuation, the results from the sensitivity analysis will enable the actuary to
quantify the effect of departures from the chosen parameter values when presenting the
results of the model to the company.
Question
List the parameters that the actuary would normally want to test in this way.
Solution
The parameters that would likely be varied in the sensitivity tests would be:
 mortality rates
 surrender rates
 paid-up rates
 future expense levels
 future expense inflation
 investment yield (income component)
 investment yield (capital gain component).
4.5 Alternative ways of allowing for risk

Where a probability distribution can be assigned to a parameter, it may be possible to
derive analytically the variance of the profit or return on capital. A sensitivity or scenario
analysis can be carried out, such as at certain confidence intervals of the distribution. This
again helps in assessing margins or in quantifying the effect of departures from the chosen
parameter values when presenting the results of the model to the company.

The variance of the profit could also be assessed by using a stochastic model for the distribution
of the parameter, and then assessing the variance of the profit or return on capital using Monte
Carlo simulation.
Question
The profit from a particular savings product has been shown to be very sensitive to withdrawal
rates, particularly at early durations of the policy when the asset share is often negative.
Explain whether a stochastic or deterministic model would be used for withdrawal rates, in order
to assess the future risks to the company from this cause.
Solution
The problem with predicting withdrawal rates is that they are heavily influenced by economic and
commercial factors. For example, withdrawal rates can be influenced by:
 economic conditions, especially those affecting employment
 media publicity
 comparison with competitors
 selling practices.
Significant events (eg economic recession) occurring in the past lead to distortions in the past
experience data, whilst future occurrences of such events are almost impossible to predict. It is
therefore difficult or impossible to devise a future probability distribution for withdrawal rates
with any degree of confidence, which makes a stochastic approach doubtful.
The best approach would probably be a deterministic model for withdrawal rates, testing the
effect of a whole range of possible outcomes (particularly at early policy durations).
The main difficulty with using a stochastic model, therefore, is in assigning a probability
distribution to the value of any parameter, and hence in estimating the parameter risk associated
with any particular case. A sensitivity analysis is a pragmatic, transparent and informative way of
getting a handle on the parameter risk, without the difficulty, subjectivity, and perhaps spurious
accuracy of deciding upon some arbitrary probability distribution to represent the uncertainty.
Nevertheless, let’s not forget that a stochastic model can be invaluable:

 when we are trying to assess the impact of guarantees
 when the variable of interest does have a reasonably stable and predictable probability
distribution (eg investment returns in a developed economy under stable economic and
political conditions)
 for indicating the effect of year-on-year volatility (random fluctuations) on risk
 for identifying potentially high risk future scenarios (eg by tracing the sequence of events
that have led to the worst simulated outcomes).

Chapter 15 Summary
Need for models
The actuary advising a life company will require models to assist with:
 product pricing
 assessing return on capital
 assessing capital requirements
 assessing the profitability of existing business (including the present value of future
profits on the existing portfolio)
 developing an appropriate investment strategy
 projecting the future supervisory solvency position
 any other work involving financial projections.
Cashflow pricing
Use a single policy model on individual model points.
Find price by finding a premium (or charges) that satisfies the company’s profit criteria.
Profit criteria
These are usually based on single figure functions of the profit signature. Three commonly
used functions are:
 net present value (which can be expressed in different ways)
 internal rate of return
 discounted payback period.
Net present value

The net present value of a profit signature is calculated by discounting it at the risk discount
rate. Economic theory implies that net present value is the best profit criterion to use.
However it is dependent on the risk discount rate being appropriate for the inherent risk.
Net present value is normally expressed in relative terms such as:

 in proportion to initial sales costs
 in proportion to total discounted premium income.


This is defined as the rate of return at which the discounted value of the cashflow is zero. It
suffers from some disadvantages in comparison with the net present value as a profitability
criterion:
 it might not exist
 it might not be unique
 it cannot be related to other indicators such as sales cost or premium income.

This is defined as the policy duration at which the profits that have emerged so far have a
present value of zero.
It is a useful reference for companies eager to recoup their initial capital investment in as
short a time as possible.
Which criterion?
The most widely used profitability criterion is net present value. In addition the discounted
payback period may be a useful measure if capital is particularly scarce, and this can assist in
designing capital efficient products.
Other pricing considerations

Price must also be considered for:
 marketability
 whether the company has sufficient capital to finance expected sales volume
 whether the return on capital for the business as a whole is satisfactory.
Model office and new-business models can be used to assess the last two points.
It is also essential that all prices are sensitivity tested at all stages of modelling.
Profitability of existing business

The future profits projected from the existing business model can be discounted at an
appropriate risk discount rate to give the expected present value of future profits. This plus
the shareholders’ share of the net assets forms the embedded value of the company.
Assessing solvency
Solvency is measured by comparing the value of assets against the value of liabilities, on a
supervisory or economic basis. If the supervisory basis is stronger then supervisory solvency
becomes the predominant criterion.

Projecting solvency
Future solvency will need to be projected in order to measure the impact of future variations
in experience not allowed for in the supervisory reserving basis.
The modelling of such future variations will need to be dynamic.
Solvency projections may also allow for management actions (such as changes in bonus and
investment policy) where appropriate.
The projection can be done on either a deterministic basis or a stochastic basis, and may be
on an ongoing basis or consider just the existing portfolio.
The need for capital

A life company needs capital:
 to withstand adverse, often unexpected, conditions
 to write new business, including meeting product development costs and financing
new business strain
 to allow a riskier investment strategy than strict matching would dictate.
This capital is normally provided on a day-to-day basis by the company’s free assets but in
effect the capital is being provided by with-profits policyholders, by shareholders (in a
proprietary company), or both.
Sensitivity testing
Results from the models will need to be looked at in conjunction with sensitivity tests to
show the vulnerability of the results to unexpected future experience.
Outcomes will be analysed for sensitivity to variations in:

 model point assumptions (if applicable)
 parameter values.
Sensitivity testing, on a deterministic basis, can be used to help determine the margins that
may be necessary in a basis.
Such margins in each parameter assumption can be a way of allowing for risk in a pricing
model, as an alternative to risk margins in the risk discount rate.
Sensitivity analysis can help to determine the variance of profit, or of the return on capital,
for any business being modelled.
Alternatively stochastic simulation techniques can be used to assess profit variance.



15.1 A life insurance company is to price a regular premium unit-linked endowment assurance product.
List the items that need to be included in the profit test model for this purpose.
15.2 A firm of consulting actuaries has developed a life office modelling system and is trying to sell this
system to a small life insurer in an overseas country where life office modelling is unheard of.
Exam style
(i) Outline the points the consultancy would put forward to promote the sale. [8]
(ii) Describe the disadvantages of buying the product from the insurance company’s point of
view. [3]
[Total 11]
15.3 An actuary works for a small life insurance company which is unwilling to invest in any ‘off-the-
Exam style
shelf’ model office software. Rather than construct a program themselves, the actuary has
decided to ‘buy’ from a consultancy the monthly cashflows expected from the major product
types sold by the actuary’s company for a small number of specimen existing business and new
business model points. The consultancy will assume a set of central parameters defined by the
actuary in line with the company’s experience.
The actuary says that this will allow the insurance company to do work on profit testing,
profitability of existing business and financial projections during the next 12 months, after which
time they will buy revised cashflows in line with any possible new parameterisation or other
changes.
Discuss the advantages and disadvantages of this approach. [9]
15.4 Describe the key factors that influence the adequacy of an actuarial model, used for projecting
cashflows to assess the profitability of a life insurance contract. [7]
Exam style

15.5 A medium-sized life insurance company writes with-profits business in a Continental European
country. Its most recent model projections show that, if it continues to write current products at
Exam style
current new business levels, the company will have problems meeting the solvency capital
requirement in three years’ time.
(i) Describe the checks that could be made to ensure that these results are reliable. [6]
The chief executive does not believe the projections, and says that the model is based on model
points rather than real policies and that the parameter assumptions could be wrong.
(ii) Outline the points that should be made to justify the modelling work in the light of these
worries. [7]
The chief executive is still not completely convinced and so suggests waiting two years to see how
the situation develops before taking any action.
(iii) Outline the points that could be made to persuade the chief executive otherwise. [3]
One option for tackling the solvency problem would be to change the product range.
(iv) Explain the changes that might be considered, and how the model might assist with this
work. [7]
[Total 23]

15.1 To project unit reserves:
 allocated premium
 charges (outgo)
 unit growth rate.
To project cashflows:
 premium (income)
 less cost of allocation
 charges (income)
 interest
 surrender costs (in excess of unit funds)
 death costs (in excess of unit funds)
 expenses
 commission
 tax.
To project non-unit reserves:

Future cashflows: as above (but using prudent assumptions if required by regulation).
To project profit:
 cashflows
 increase in non-unit reserves
 increase in required solvency capital.
15.2 (i) Promoting the sale
At a product level:
Assist with product design – test effect on profitability of differing designs and features. [½]
Likewise for capital efficiency. [½]
Assist with product pricing – can be used to set premiums to meet the given profitability
criteria. [1]
Measure sensitivity of profit to variations in experience – hence indicating possible danger areas
where action could be taken (eg redesign, reinsurance). [1]
Measure sensitivity of profit to variations in controlled parameters such as bonus levels, renewal
commission levels, surrender value scales, thereby helping in the setting of these parameters. [1]

At a company level:
Measure the profitability of existing business (eg calculate embedded values). [½]
Do this by distribution channel or product class (or cost centre?). [½]
Measure the effect on profitability of existing business of alterations to discretionary benefit

scales. [½]
Measure the profitability of the new business actually written. [½]
Assess the total capital requirements of a new product launch, and eventual return. [1]
Use for projections of the life company, giving the following:

 general information (revenue account, balance sheet, for ‘business plan’) [½]
 check on future solvency [½]
 assist in strategic decision making (eg impact of new commission levels or reinsurance
program) [½]
 assist in expense budgeting. [½]
Possible check on statutory valuation results. [½]

[Maximum 8]
(ii) Disadvantages
The cost of acquiring the package. [½]
Time and resources (actuarial time tailoring the model to the insurance company’s precise
circumstances, extra computer hardware, help on related issues such as number crunching to
generate model points). Are all the data that are needed to run the model readily available? [1]
Possible errors in the results due to the complexity of the model. [½]
Sensitivity of results to correct parameter choice. [1]
Results may be regarded suspiciously by management because it’s a new and mysterious concept
full of ‘guesses’ about the future. [½]
[Maximum 3]

15.3 Advantages of the suggested approach
It will save a lot of time and money. [1]
The cashflow results will probably be more reliable being produced by the model office
‘manufacturers’ than by a newcomer to the system. [½]
The actuary will be able to do produce some useful results:

 financial projections [½]
 capital requirements and return on capital [½]
 profitability (present value of future profits) of existing portfolio [½]
 some other miscellaneous work (eg quantify total expense charges on new and existing
business for expense budget). [½]
It should also be possible to test the sensitivity of these results to differing levels of new business,
or (a bit messier) on different risk discount rates. [½]
Disadvantages of the suggested approach
All of the above work will be based on a very limited number of model points. There will
therefore be significant potential for model point error. [1]
There is a problem if the mix of age / term / gender / premium size of new business changes much
from that initially assumed. It may not be possible to represent future new business adequately
in model office projections (absence of suitable model points), nor to assess the value of new
business actually written. [1]
The actuary will be unable to test sensitivity to parameter variation. [1]
It will not be possible to revise the parameters to reflect altering experience assumptions
throughout the year (which might be important in a time of moving interest rates, for
instance). [1]
The cashflows will not be of much use for product development purposes, unless products being
developed have a structure identical to that of some existing product which has been
modelled. [1]
They will not be of much use for pricing, because it will not be possible to revise the premiums in
the model until the desired profitability criterion is achieved. [1]
The actuary will not be able to do any stochastic work. [½]
In general, anything requiring flexibility and changes will be impossible. [½]

[Maximum 9]

15.4 Comment
The key words here are ‘describe’ and ‘adequacy’. ‘Describe’ is a rather broad term, and it is often
difficult to decide what should be included. Clearly it means you should say what something is, or
what it does. Whether you should also say how and/or why it does is less obvious, and the extent
could differ between questions. One guide to the ‘depth’ of your description would be the number
of marks available. Another guide would be that you might include some ‘how’ or ‘why’ in order
to relate your answer fully to the context of the question.
Here there are only seven marks, and there are easily seven different factors listed in the Core
Reading that you could include. Hence you should keep largely to the ‘what it does’ interpretation.
Nevertheless, we have included a few ‘whys’ and ‘hows’ in our solution, where we feel some
justification for including a factor seems necessary, given the particular context.
Another guiding principle to answering any question, is to prioritise the points you are going to
make in order of importance and relevance to the question. The more important and/or relevant
the point, then the proportionately more time and space you should devote to it. Putting your
points in order of importance (most important first) also gives the examiners confidence that you
really are on top of this subject.
Understanding what is meant by ‘adequate’ is vital to answering this question. We thought it a

good idea to begin our answer by actually stating what we believe adequacy actually means in
this context, and then the subsequent description of factors can be written to this interpretation.
If you use this approach it has the great advantage that, even if your interpretation of the word
differs from that of the examiners, you could well earn credit for your answers in the context of
your interpretation (equivalent to getting method marks in a numerical question). In other words,
if you’re not sure quite what a word means, state what you think it means, and go on from there.
Throughout your answer, keep referring to the purpose of the model. In other words, keep your
answer fully focused on the context of the question at all times. Do not just regurgitate the Core
Reading (which is general to all models): the examiners will take a very dim view of this.
If you can think of other relevant factors (ie not just those in the Core Reading) then do include
them. And do not include irrelevant factors (this again signifies to the examiners that you are not
thinking, but regurgitating).
A final point to note with this question is that it might be considered slightly ambiguous. What,
exactly, is meant by ‘contract’? It could be simply one individual policy (with a given age, term,
sum assured, etc) or it could mean a whole family of policies of one particular product type. The
use of the words ‘policy’ or ‘product’ would have avoided the ambiguity. Here it is probably
advisable to interpret the word in its broadest sense (in that way you are less likely to miss marks)
and this is what we have done. (If the question had said ‘policy’, the part of the answer which
related to model points would not have been included.)
Solution
This model will be adequate if it effectively assesses the (future) profitability of a life insurance
contract. [½]

To be effective the model will need to be sufficiently realistic, that is the extent to which the
model simplifies reality should not be so great as to lead to an inappropriate decision being made
as a result of using the model. [1]
The various factors that influence the effectiveness of the model are as follows:
 Does the model produce the required output? In this case we require a stream of profit
flows to be output from the model. [½]
Where possible, we would like the output to be in a form that can easily be checked for
reasonableness, … [½]
… and that is easy to assimilate by users of the model. [½]
 The timing of the profit flows are important: the model should encompass the whole term
of the contract … [½]
… and profit flows should be calculated at appropriate intervals: eg monthly or yearly. [½]
The smaller the interval, the more accurate should be the output. [½]
 The model must incorporate, as parameters in the model, all the various factors that can
influence the future profit stream. [½]
The impact of each parameter on the model output must be realistic. [½]
 The values chosen for the parameters of the model must be appropriate to the company,
the contract in question, and to the current and projected economic and commercial
environment. [1]
 The model points chosen (eg age, sum assured, policy term) must be a realistic reflection
of the mix of business for the contract concerned. [½]
 The model must include all the items of cashflow that the contract could generate. [½]
This includes all possible benefit payments (eg on death, surrender, maturity), expenses,
taxation, and investment returns. [½]
It must also include expected cashflows that could arise under any options provided
under the policy, eg where guaranteed terms for conversion or renewal are available. [½]
 The model must allow for the impact of the supervisory reserves and any required
solvency capital on the resulting profit flows. [½]
The reserves will defer the emergence of profit over time, although investment returns
from the investment of those reserves should also be allowed for in the model. [½]
The deferment of profit is particularly important for assessing the value of the profit
stream (at the risk discount rate) and for assessing the consequent return on capital. [½]

 The model must allow sufficiently for any interactions between components of the model.
For example, the assumed surrender and reserving assumptions at any projection point in
the model should be related to the projected conditions at that time. [1]
 In assessing the profitability of a life insurance contract it will be necessary to assess both
the expected profitability and its likely variability. [½]
In order to assess variability the model may need to have the facility to be used
stochastically, at least for key variables, particularly where the contract provides financial
guarantees. [1]
As a bare minimum the model must be capable of deterministic sensitivity testing. [½]
 Where the model needs to be used stochastically, the assumed distributions for the
random variables should be realistic. [½]
[Maximum 7]
15.5 (i) Checks on the results
Compare model output against the most recent statutory valuation. Check the basics: sums
insured, premiums, reserves. [½]
Check that unmodelled products are only an insignificant part of the portfolio, and that some
adjustment has been made to the detailed model output in respect of these. [1]
Check the data from which the model points have been derived. [½]
Check that the parameters have been input correctly. [½]
Check that the model behaves as would be expected when the parameters are varied. [½]
Re-examine the parameters – are they really best estimates or are they too pessimistic? [½]
Is the model dynamic where it needs to be, in particular is there a realistic relationship between
assets and liabilities? Is the bonus declaration programmed sensibly for a given yield? [1]
Are new business volumes sensible? Or is an exceptional year of new business being projected
forward? Any model will give a solvency capital warning after a prolonged sales explosion. [1]
Check last year’s model projection of the year to date against reality. Is it reasonable; are the
variations justifiable? [1]
Check the solvency capital requirement calculation within the model. Does it look sensible? [½]
Finally, do the numbers look reasonable? Perform some back of envelope checks, or project one
model point (for in-force business) plus one model point for each year’s new business on a
spreadsheet. Are the results of the right order of magnitude? [1]
[Maximum 6]

(ii) Justifying the model
Model points
Emphasise that they are not some theoretical creation. They are derived from the in-force policy
portfolio. [1]
The derivation is not overly simplistic – explain that from 285,000 policies there are 5,200 model
points (for example) – data handling routine ensures correct averaging, totalling etc. [½]
They have been checked by reconciliation between model and reality. [½]
In deciding on that level of grouping, grouped results were tested against non-grouped results
and found no significant difference (1%?). [½]
It is common practice to use model points. [½]
The projection can be redone on a policy-by-policy basis if the company will foot the bill
(computing time, computing resources, more actuarial time). [½]
There will be some very minor errors due to the model point grouping but they could work both
ways – so perhaps in three years’ time the company will be covering eg only 75% of the solvency
capital requirement rather than the 76% that the model is suggesting. [½]
Parameter error
It is important to say that the parameters are probably wrong. No-one can predict the future with
100% accuracy. The key thing is that the model needs to look forward, and so sensible
assumptions about the future are required. [1]
On average (over time) experience is expected to be close to these assumptions. [½]
The parameters represent best estimates of interest, expenses, mortality, withdrawals, new
business growth, etc over the next three to five years. [½]
Many of the parameter choices are the results of thorough investigations (eg mortality,
expenses). [½]
The relationship between the parameters is in some cases more important than the parameters
themselves, and that relationship can be justified. [½]
Given that the parameters are likely to be wrong, sensitivity testing is also performed. This gives a
range of results within which we can be (say) 95% sure that the future will lie. [1]
The model can be run on sensible alternative parameters corresponding to any reasonable
scenario that the chief executive wants modelled. [½]
[Maximum 7]

(iii) Wait two years?
This would not be a good idea. First, because the ‘three years’ was only a best estimate: perhaps
it could be two or four years – look at the sensitivity results. [1]
Secondly, many of the measures for dealing with the problem are not instantaneous, but will take
some time to start to work – so the company needs to start now in order not to be restricted to
the few instant remedies that might remain. [1]
Thirdly (minor points):
Are the regulatory authorities, or the market, going to be looking for stronger solvency? If they
are looking for (eg) 150% cover of the solvency capital requirement then the company will be
‘unofficially’ insolvent much more quickly. [½]
Are the free assets getting so low that the investment strategy might be seriously constrained if
action were delayed any longer, to the detriment of eventual investment return? [½]
[Total 3]
(iv) Changing the product range
Existing products need to be redesigned and repriced to be more profitable and less capital
intensive. [1]
Use the model to identify the most unprofitable and the most capital intensive products. Both of
these categories (not necessarily identical) should be looked at. [½]
Unprofitable products:
 Use profit testing to determine premiums that satisfy the profit criterion for these
products. [½]
 Are these premiums competitive? If not, the company may need to look at other aspects
(expense cross-subsidy? reduce acquisition expenses?) in order to keep premiums
competitive, or redesign the product structure. [½]
Capital intensive products:

 Look at redesigning the products (removing guarantees? defer bonus distribution?
introduce unit-linked?). [1]
 Is there any scope for repricing upwards? How far can this go before becoming
uncompetitive? [½]
Then, after examining these product / pricing changes, re-run the model office projection to see
the total effect, allowing for expected (probably different) new business volumes. Is solvency still
a problem? [1]

If so, and there has been as much redesign / re-pricing as the market will permit, start considering
other measures. For example: [½]
 Is there anything that can be done on the discretionary benefits side to help? eg ensure
that surrender values and pay-outs on maturity are not greater than asset share. [1]
 Look at expenses. Management expenses – can these be reduced? This is the biggest
immediately controllable aspect. Acquisition expenses – can these be reduced? – or can
the commission scale be changed (lower acquisition, greater renewal – this will use capital
less quickly)? [1]
 Place a cap on sales. Stop selling immediately the more capital intensive products. [½]
 Cease writing new business? [½]
[Maximum 7]





using the course.

SP2-16: Product design Page 1
Product design
Syllabus objectives
2.1 Demonstrate an understanding of life insurance product design, including:
2.1.1 Describe the factors to consider in determining a suitable design, in terms of

premiums, benefits and charges, for a life insurance product.
2.1.2 Determine a suitable design for a product in a given situation.
2.1.3 Discuss the relative merits of different product designs.

Page 2 SP2-16: Product design
0 Introduction
This chapter examines at a conceptual level the factors relevant to the design of a new product.
In the SP2 exam, students may be asked to apply the factors outlined in this chapter to set
an appropriate product design, or to comment on the merits of a proposed product design.
Before studying these factors it is useful to take a step back in order to put these concepts in an
appropriate context.
Question
Explain why an actuary might be looking at a possible new product design.
Solution
Awareness of a gap in the existing product range. This may result from simply looking at another
company’s products, or product ideas in other markets. It may also result from a change in the
market, such as:
 the market awareness of a new product need or product feature
 the availability of new financial instruments making a new product feasible (eg the various
guaranteed index bonds created since the early 1990s made possible by index futures)
 a legislative or fiscal change rendering possible, or more financially attractive, a new
product (eg the flood of pensions policies in Western European countries over the last
decade of the 20th Century, very often due to tax breaks).
Awareness of an inadequacy in an existing product, such as:

 insufficient profitability – given both the increasing sophistication of profit-testing
techniques and the need to satisfy increasingly demanding shareholders, it is possible that
an existing product deemed profitable ten years ago does not satisfy current criteria
 inefficient use of capital – the same comment applies
 inappropriateness of premium basis / charging structure in the light of:
– demographic shifts
– a revised investment return outlook (hence without-profits immediate annuity
rates might be re-rated monthly in respect of the interest guarantee, and
re-examined yearly in respect of the mortality basis)
– a revised expense experience.
You were not expected to answer in such detail. The important things here are the general
reasons for why a product might need to be created or modified.
Given these reasons for examining a new product design, the various factors explained below
should all appear very natural. The setting of appropriate values for product pricing is dealt with
later in the course.

1 Factors in product design

This section sets out the factors that an actuary will consider when determining the
suitability of a proposed contract design, not all of which will apply in every situation.
The factors which apply to a completely new product will be equally relevant in assessing both the
continuing validity of an existing product, and the appropriateness of any proposed modification
of existing products.
1.1 Profitability
A company will want to ensure that the premiums charged for non-linked contracts will be
sufficient to cover the benefits to be provided and the expenses in most foreseeable
circumstances, and provide a profit margin.
At a general level, a life insurance product can be broken down into three components: savings,
protection and administration. This brings into focus the true nature of a product: it is the
provision of certain services to the policyholder by the insurance company. What is important is
that the product be profitable in its entirety, but ideally every product should be profitable in
each of these dimensions:
 Savings: the company will want to make a profit on investment income.

For non-linked business the actuary will therefore need to set the interest rate in the
premium basis at an appropriate level given the (prudently) expected investment
environment.
 Protection: the premium basis mortality assumptions should cover the risk involved.
This risk will be gradually changing (and sometimes not so gradually – witness the
preoccupation in the 1980s with an AIDS explosion), and this future change should be
allowed for if appropriate.
Question
Give an example of a product where an allowance for expected future mortality change is
essential.
Solution
The best example is a deferred annuity contract where the mortality rates post-vesting are
guaranteed at policy inception. The total period from inception to the later payments of the
annuity will often exceed 30 years, during which time mortality is likely to improve significantly.
The problem also exists, but to a much lesser extent, for immediate annuity business.
Significant mortality improvement is not of course exclusive to annuitants, but for assured lives
with normal life cover any such improvement will enhance product profitability and so does not
need to be allowed for (other than for reasons of competitiveness).

 Administration: the expense loadings must cover the marginal acquisition and
administration costs involved.
Further to this, the company will require that its total expenses be covered by total
expense charges (subject to any conscious cross-subsidies). Each policy should therefore
also contribute to the fixed costs of the life company.
For unit-linked contracts it will want to ensure that overall the charges will be sufficient to
cover the expenses to be incurred, and provide a profit margin.
The above points about savings, protection and administration are also relevant to unit-linked
contracts. However the risk to profitability is lower because investment risk is largely borne by
the policyholder and the mortality (or other contingency if relevant) charges may be variable at
the company’s discretion.
The most critical influence on profitability will therefore be the adequacy of the expense charge.
This itself may also be variable. But, for most product types, undercharging for expenses for one
year is likely to impinge on profitability more than a slight undercharging of the mortality risk for
the same time period.
1.2 Marketability
The benefits offered need to be attractive to the market in which the contract will be sold.
Innovative design features may make a contract more attractive as may the addition of
options and guarantees. The charging structure under a unit-linked contract needs to be
attractive to the potential market and consideration needs to be given to whether the
charges should be guaranteed. More generally, it needs to be considered what guarantees
should be given with regard to premium rates.
In the context of marketability, one would consider how understandable the product is. ‘Bells and
whistles’ will only be a good thing if the typical target customer understands and appreciates
them. This aspect is very sensitive to the distribution channel intended to be used for the product
– see further comments on distribution channels below.
1.3 Competitiveness
A company will not want the structure and level of the charges under a unit-linked contract
to depart too far from those of competitors, but this depends on how it will market the
contract.
The prime influence on competitiveness will be the level of the expense charges. However some
distribution channels will be more price sensitive than others.
For instance, in the very price-sensitive distribution channel of brokers, companies have
sometimes set charges for certain products that are only slightly greater than those required to
cover the marginal costs, thereby making an apparently insufficient contribution to fixed
overheads. This is then justified on the grounds that there is still some contribution to expense
overheads, as opposed to the zero contribution which would result from not selling an
uncompetitive product with a ‘correct’ contribution built in.

What really matters to the company is the total contribution to overheads (and profit) that results
from selling the product as a whole, not per single policy. This total contribution can be
represented broadly by:
 total contribution  per-policy contribution  number of policies sold

where the per-policy contribution is the profit from a single policy net of marginal costs (but
ignoring overhead costs).
The company’s total profit can then be thought of as:
 
 total profit     total contribution  overheads  .
all products 
So we need to find the level of charges that will maximise the ‘total contribution’ from a given
product. In a very competitive market this might be achieved only by having very low per-policy
margins thereby ensuring sales. With less competition, higher per-policy margins will be possible
without jeopardising volume.
The same applies to the premiums under without-profits non-linked contracts.
So here the whole premium is an issue from the competitiveness point of view. Thus the interest,
risk and expense components of the premium basis must all be set at a feasible level. The relative
importance of these depends on the exact contract involved. For term assurance the mortality
basis will be critical and the interest rate assumption is not very important; the amount of
expenses loaded for as a fixed charge will be important. For single premium endowments the
interest rate is critical, as is the amount of expenses loaded for as a proportion of premium.
Question
Explain why the premium rate is not crucial to competitiveness for UK-style with-profits contracts.
Solution
Policyholders from different companies with identical investment, expense and mortality
experience, and distributing the same proportion of surplus, will receive the same eventual
benefits irrespective of slight variations in the premium level. This is because the bonus
distribution will release to policyholders the excess of (for instance) investment profit over and
above the guaranteed level implicit in the rates, so that a lower sum assured will be associated
with higher bonus rates, all else being equal.
However the guaranteed benefits will be different, especially early on.
Looking at the question another way, a with-profits policyholder simply wants to save a certain
amount per month (or whatever), rather than have a particular sum insured. At parity of
investment, mortality and expense experience, a given sum of premiums should lead to the same
final benefit.

However, there are two ways in which the premium rate (that is, the level of guaranteed sum
assured relative to the premium) can have an effect on the final maturity payout. First is that,
with a higher sum assured, the death benefit at early durations will be higher. Hence the asset
share will grow less quickly, as more is paid out on death, which means that the maturity amount
should be lower. This actually leads to an inverse relation between the maturity proceeds and the
guaranteed sum assured: the higher the guaranteed sum assured, the lower the maturity benefits
are likely to be.
Secondly, a policy having a higher initial level of guarantee implies that the reserves near the start
of the contract will be higher. This gives the company less investment freedom, a lower expected
return from the assets, and hence a lower expected maturity value. Hence this will also lead to
the inverse relationship described above.
Don’t worry too much about this second aspect now – it will be covered later in the course.
The level of with-profits policy premiums may influence competitiveness in the case of the
revalorisation or contribution bonus distribution method.
The extent to which this is true will depend on:
 What happens to expense and mortality profit? For instance, if the company retains all
expense profit then the level of expense loading will have a significant influence on
premium size. This cannot be justified to customers, as it could be if eventual expense
profits were to be distributed back to policyholders.
 Market maturity: How much are potential policyholders made aware, and how much do
they then understand, of the importance of long-term investment performance and the
importance of what proportion of this is distributed?
 Investment characteristics of the market: If all companies invest in a fairly uniform way
(which is true of many Continental European markets, where portfolios include an
extremely high proportion of government debt) then there will be little to distinguish
between them, even for the potential policyholder who appreciates the impact of
investment performance. In this case the main basis for comparison may be premium
size, in conjunction with the proportion of profits allocated to policyholders.
Question
Explain whether the premium level is likely to be more significant for the revalorisation or
contribution method.

Solution
The premium will be more significant for the revalorisation method, where normally only
investment profit – or a certain proportion thereof – is released to policyholders, compared with
the contribution method where portions of all of the three major sources of profit (investment,
expenses and mortality) are released.
1.4 Financing requirement

Unless the company has substantial capital resources, it will want the benefits and charges
to be designed to minimise its financing requirement. There is more scope under
unit-linked products to adjust the design to achieve this than under non-linked products.
This is a particularly important criterion for new (or recently established) life insurance companies
because of the small amount of free assets that they have available to finance new business. As a
result, these new companies often sell just unit-linked products.
As far as minimising the financing requirement is concerned, unit-linked products offer the
following advantages over non-linked products:
 no (or few) expense guarantees
 no (or few) mortality guarantees
 no (or few) investment return guarantees
 possibly a smaller supervisory solvency capital requirement.
The actuary will use model office techniques to project the financial situation with and without
the new product under sensible sales assumptions to assess the company’s ability to finance the
product, and whether the return on capital is adequate.
For non-linked products there is little that can be done at the design stage to reduce the capital
strain.
Question
Explain how the capital strain of a non-linked product can be reduced.

Solution
By reducing:
 initial acquisition expenses
 initial administration expenses
 valuation strain, via:
– increasing the valuation interest rate (are there better-earning assets in respect of
that product class?)
– reduction of guarantees.
The difficulty is that these aspects concern the environment in which the product is sold rather
than anything intrinsic within the product design per se. The most immediately controllable
element in a product design context will be the commission scale (or similar, depending on the
precise distribution method). Even this may not be easily alterable for one product.
As you will see in Chapter 24, another method of reducing capital strain is through financial
reinsurance.
1.5 Risk characteristics

Consideration will need to be given to the acceptability of the level of risk associated with a
proposed contract design. The level of risk that may be acceptable will depend upon the
company’s ability or willingness either to absorb risk internally or to reinsure or hedge it.
The company may be prepared to accept the mortality risk in full, as it is a well-studied and
quantified risk both at industry level and at company level. However, for a company entering a
new market for that company (perhaps using a new distribution channel) the mortality parameter
risk would still be relatively large. If the company is faced with a large parameter risk it could do
one or several of:
 offer the contract in unit-linked and/or reviewable form to avoid a long-term rate
guarantee
 reinsure a large part of the risk
 incorporate very ample margins in the premium rates
 offer the contract as an additional ‘rider’ benefit rather than stand-alone.
However, the ability to hedge investment risk will be an important consideration in the design of
financial guarantees.
1.6 Onerousness of any guarantees

The company will need to consider the onerousness of any guarantees, for example the
level of any guaranteed surrender values under a non-linked contract.

Question
List at least five types of guarantee.
Solution
Possible guarantees are:

 surrender values
 interest
 investment return (bonuses) accrued to date
 expenses
 mortality (or other risk factor)
 annuity option
 renewability without medical evidence.
Offering guarantees results in two problems:

 possibly having to suffer a cost that was not fully expected
 probably (depending on the supervisory reserving regime) having to reserve for this
possibility from the outset – thereby increasing the capital strain of the product.
The ‘generosity’ of a scale of guaranteed surrender values will be measured by reference to asset
share. There may be a problem at early policy durations when the asset share is very small, or
even negative. The onerousness of guaranteed surrender values can be quantified as their impact
on product profitability under various conditions. This can be assessed by modelling the product
under various asset fall scenarios and high withdrawal rate scenarios. Unfortunately, these two
aspects are not independent as an asset fall situation would often be associated with an increase
in withdrawal rates.
A more fundamental guarantee is the investment return guarantee for without-profits non-linked
business. Guaranteeing a sum assured or annuity benefit derived from a premium formula based
on a yield of x% pa means that a guaranteed return of x% pa is guaranteed. (Apologies if we have
stated the obvious.) For without-profits savings business this will be an important consideration.
For with-profits business the interest rate assumption in the premium basis will usually imply a
low investment guarantee initially. However, the amount of investment guarantee will increase
throughout the lifetime of most with-profits policies. For example, the addition of bonuses under
the ‘additions to benefits’ method causes the amount of guaranteed benefit to increase with
policy duration, which would not happen under a without-profits policy.
The onerousness of any guarantee is a function not only of how likely or how big an adverse
change in some variable might be, but also on the financial impact that the change in the variable
might have.

Hence a life company may not find its guarantees so onerous if it can reduce the financial impact
of those guarantees. In the case of investment guarantees, the company can reduce the financial
impact by immunising itself against these guarantees by the choice of suitable investments. For
instance, with immediate annuities it is theoretically possible to invest so as virtually to eliminate
the investment risk (as we mentioned earlier in the course), though there are many difficulties
(such as a shortage of suitable investments) that can prevent this happening in practice. For
savings policies with significant guaranteed surrender values, some investment in very secure
assets (such as short-dated government bonds) can reduce the surrender risk. These issues are
discussed in more detail later in the course.
Question
List at least three factors which will contribute to the onerousness of an investment guarantee.
Solution
The onerousness of an investment guarantee will depend on:

 size (how much is being guaranteed)
 the period of time for which it holds
 the significance of reserves for the contract (thus a relatively high yield assumption for
term assurance business will not be the time-bomb it could be for savings business)
 the volume of business to which it applies
 the capital available to the company to absorb the initial reserving cost and the possible
eventual real cost.
1.7 Distribution channel

The company will need to consider the sophistication of the target clients of a particular
distribution channel.
The distribution channel the product is sold through will impact on the likely level of wealth and
financial sophistication of prospective policyholders as well as affecting the amount of advice that
they receive. We discussed these factors in detail in Chapter 8.
For instance, a life company would not attempt to sell anything complicated by direct marketing
methods.
The distribution channel involved will have a fundamental influence on what product is required,
how it should be structured, and how it should be priced.
If intending to sell the same product through more than one channel, the company will need to
decide whether to apply the same pricing to all channels, or to price differently for different
channels (an approach known as ‘dual pricing’).

1.8 Sensitivity of profit

Under a unit-linked contract, it may be possible to have a well-matched charging structure
that reduces the sensitivity of the contract's profitability to variations in future experience.
There is considerable overlap between this and the point on the onerousness of guarantees
(ie Section 1.6), as they are closely related: products are more likely to be sensitive to particular
types of risk the more onerousness the guarantees are.
The important variables that might impinge on profitability are:

 investment return
 mortality, or other contingency if relevant
 expenses, including expense inflation
 withdrawal rates.
The product design for a unit-linked contract can minimise the sensitivity of profitability to
adverse experience of these factors by:
 Investment return: if there are no investment guarantees (for example, where payment at
maturity is simply the value of the unit fund, with no specified minimum value) then most
of the investment risk is borne by the policyholder.
 Mortality: make the charge for this variable at the company’s discretion.
 Expenses: make the charge for this variable at the company’s discretion.
 Withdrawal rates: don’t offer any guaranteed surrender values.
 Matching: try to match income (the charges) with outgo (expenses and benefit costs) as
closely as possible by duration, especially with regard to the initial expenses.
Likewise it may be possible to redesign non-linked contracts to reduce the exposure of profit to
unpleasant variations in future experience (for instance by not offering, or at least minimising, any
guaranteed surrender values). The sensitivity of profitability to high withdrawal rates could also
be reduced by using a commission system whereby there is some ‘clawback’ (return) of
commission for early withdrawals.
Non-linked with-profits contracts could be redesigned if there are any forms of guarantee
associated with the bonus distribution method employed.
Example
A company operating in a Continental European country uses the revalorisation method of profit
distribution, giving 90% of all investment income to policyholders. It is currently selling products
priced on a 3% pa interest guarantee. Long-term bond yields are now around 5% pa. The
company is worried about the impact on profit of rates moving lower (even if above the 3% pa
guarantee).

The company currently takes as interest profit 0.5% pa (the 10% of investment income not
distributed applied to a 5% pa yield). If yields move down to 4% pa then this profit becomes
0.4% pa, and as yields move towards 3% pa it reduces to zero. The company requires an interest
profit of 0.5% pa to cover investment expenses.
It could reduce exposure to the risk of not covering investment expenses by altering the definition
of profit distributable to policyholders from 90% to ‘90% or {yield – 0.5%} if lower, subject to the
3% pa guarantee’. This method has been adopted by some companies to anticipate this problem.
1.9 Extent of cross-subsidies

A company needs to decide on the extent of any cross-subsidies between, for example,
large and small contracts. The marketing advantage of a simple premium or charging
structure may conflict with a desire to avoid cross-subsidies.
This is an issue which tends to involve expenses, in two ways.
First, given a very competitive market which focuses on expense charges, to what extent can the
company subsidise an expense overrun (ie expenses greater than expense charges) from interest
(and to a lesser extent mortality) profits?
Secondly, to what extent should policies cover their marginal per-policy administrative and sales
costs (in other words, costs that are incurred by a policy regardless of how big or small the policy
is)? Setting a policy charge which fairly covers that cost for policies individually may prove
uncompetitive for small policies. In this case the company may set a lower fixed level of per-
policy charge, recouping the difference over the whole of the policy portfolio by a corresponding
increase in the charge expressed as a proportion of premium. So in effect small policies will be
under-priced, large policies will be overpriced. This is a conscious cross-subsidy: large policies
subsidise the administration of smaller policies.
Ideally even small policies should cover their own marginal administrative and sales costs; there
might then be a cross-subsidy in that they make very little contribution to the fixed expenses of
the company compared with the contribution from larger policies. This is the intention behind
using a fixed policy fee in the product design, and setting the policy fee equal to any marginal per-
policy administrative costs.
Question
‘Cross-subsidy is just a form of mutuality, because it involves pooling of experience.’

Solution
At first glance the concepts of mutuality and cross-subsidy seem identical, both being based on
the idea of sharing, although cross-subsidy has negative connotations and mutuality positive
connotations to some.
The difference is that mutuality involves sharing of risks (for instance the risks of death, or of
investment volatility), while cross-subsidy is the sharing of non-risk elements – here, expenses.
1.10 Administration systems

The system requirements of a new product may limit either the benefits to be provided or
the charging structure to be adopted.
For example if the company’s computer policy administration system cannot cope with
administering a waiver of premium option, the product should not have one (unless it is so
important that it is worth spending money on enhancing the system).
The issue of compatibility with the administration system can also be extended to cover the
aspect of simplicity: it is in the interests of the administration system, policyholders, agents /
brokers and the company’s staff that the product be simple. Thus any complications must be
warranted by some significant advantage in terms of the factors discussed in this chapter.
1.11 Consistency with other products

The company may wish to ensure that the charging and benefit structures of a new policy
are at least similar to any existing business.
Question
Solution
The key reason is that a major change will result in significant systems development, which will
take time.
There are benefits in terms of saving time and cost with such things as training administration and
sales staff, printing marketing literature and so on.
There is also a possibility that a design, which appears much more attractive or favourable to
policyholders, may seem unfair to existing policyholders and may lead to some dissatisfaction and
possible marketing risk. For example, it may cause many policyholders to surrender their policies
over a short period of time and the company may be unable to recover all of its expenses from
them (particularly fixed expenses).

1.12 Regulatory requirements

A company must adhere to any regulatory requirements, eg maximum (capped) charges,
treating customers fairly.
These should be taken into account in product design.
Question
Explain why it is important to treat customers fairly.
Solution
Reasons include:
 to meet what would be considered to be normal standards of professional ethical
behaviour
 because it may be a regulatory requirement, or one recommended by professional
guidance
 to foster ongoing trust and good relations with policyholders, the media, and the public in
general, which in the long term will help to generate future sales.

2 Interaction of product design factors

These factors are not necessarily independent, in that meeting one may prejudice the
meeting of another, and so a compromise will usually need to be reached. Also, the factors
are not necessarily mutually exclusive.
It is the job of the ‘marketing actuary’ to design a product, with the aim of giving the ‘optimal’
compromise between the factors set out in this chapter.
Question
List pairs of factors that are likely to be difficult to reconcile.
Solution
The following conflicts of interest are likely to occur:
Profitability
Financing requirement Marketability

Risk characteristics versus Competitiveness
Cross-subsidies (possibly)
Onerous guarantees
Sensitivity of profit

3 Product design example

This section provides an illustrative example of the potential issues outlined in Section 1,
with respect to the design of a product in a particular scenario.
3.1 Example
A life insurance company is considering entering the unit-linked pensions market in a
particular country. It is considering two possible contract designs.
Both designs pay the value of units at retirement or on earlier death or surrender.
The two designs differ only in the charges paid by policyholders:
 Design A has a low allocation rate in the early years of the contract and a monthly
policy fee payable by cancellation of an equivalent number of units. This fee is
variable, but is guaranteed not to increase by more than the rate of increase in
national average earnings in the territory concerned. There is also an annual
fund-based charge expressed as a percentage of the value of units.
 Design B has an annual fund-based charge expressed as a percentage of the value

of units and no other charges.
Note that the allocation rate is the percentage of each premium allocated to purchasing
units.
The charges under both designs have been set so that they each achieve the same level of
profitability using the company’s required rate of return on capital. Neither design includes
a surrender penalty. All assumptions used in the pricing are the same for both designs.
Describe the factors that the company would consider in deciding between the two designs.
3.2 Solution
Profitability
As outlined above, the profitability of both contract designs is the same.
The company should carry out sensitivity tests on the major assumptions to assess the
company’s exposure to changes in experience. The design with the lower sensitivity will be
preferable.
Design B will be highly sensitive to investment returns, as the only income for Design B is
from the fund-based charge and this depends on the size of the fund.
Competitiveness and marketability

The main feature affecting the marketability of both options is the benefits provided to the
policyholder. The relevant benefits are the surrender and maturity values.
The simplicity of design and comprehensibility of literature will also be relevant features.

In order for the two designs to have the same profitability, Design B would probably have
the greater annual fund-based charge, unless Design A had higher unit allocations in later
years.
Design A has a lower surrender value in the early years but higher later surrender values
and a higher maturity value than Design B. This is because the effect of the lower annual
fund-based charge eventually outweighs the initial low allocation of units and the fixed
expense charge.
The preferable design will depend on the importance attached to surrender terms at short
durations.
The charging structures used by other companies will need to be considered; it might be
easier to sell a contract that is more clearly in line with other companies. Conversely, a
different structure may lead to a marketing advantage.
Financing requirement
Design A will have a lower capital requirement at outset and will therefore have a lower cost
of capital. As both achieve the same profitability, Design A will be able to afford a higher
maturity value than Design B.
If the life insurance company’s capital is limited, the higher capital requirement of Design B
will be an important issue.
Design A has low allocation in the early years and so the unallocated premiums can be used
to repay the initial capital quicker than in Design B. As a result Design A is likely to have a
shorter payback period.
Risk characteristics
Mortality is likely to be of negligible importance.
As costs are recovered more slowly in Design B, it is more exposed to losses from early
surrenders, when initial costs have not been recouped. Alternatively, an active transfer
market would make them more important.
The investment risk is borne by the policyholder in both cases, but there is a second-order
expense risk, particularly for Design B. The expense risk depends on whether the annual
fund-based charge is guaranteed or not. If it is, then both designs are exposed to expense
risk. Design A is less exposed to expense inflation risk as the charge is more closely
matched to the expenses.
Extent of cross-subsidies
Unless the allocation rates of Design A or the annual fund-based charge of Design B are
dependent upon case size, both will be susceptible to cross-subsidy. Design A will be less
exposed as the policy fee is more closely matched with the renewal expense.
Administration systems
The design that most easily fits on the administration system would be preferred.

Consistency with other products

Similarly the design that fits most closely with other products offered by the company will
be preferred.
Regulatory requirements
The company should consider if there are any relevant legal or regulatory restrictions – for
example, could there be any restrictions on early exit which invalidate the better termination
values of Design B.
Final decision
For the company to make the final decision all the above factors would be considered.
One approach to making a final decision would be to give each factor a relative weighting.
Each design could be assessed by determining the total score for each contract, with the
design with the highest weighted score being preferred.

Chapter 16 Summary
In designing or reviewing any product the following factors need to be considered:
 profitability
 marketability
 competitiveness
 financing requirement
 risk characteristics
 onerousness of any guarantees
 distribution channel
 sensitivity of profit
 extent of cross-subsidies
 administration systems
 consistency with other products
 meeting regulatory requirements.
Such considerations will need to reflect the nature of the existing product range and the
distribution channel involved. All other things being equal, simplicity in product
design is preferable to complexity.
These factors are neither independent nor mutually exclusive. Sometimes they will be
difficult to resolve.



16.1 In a particular country, until two years ago all products were designed and priced according to
Exam style
rules and bases laid down by the State regulatory authority. Two years ago government
legislation liberalised the market, allowing companies to design and price products as they wish.
A proprietary life insurance company in that country has been slow to react to the changes and
has been losing market share.
The company has decided to add unit-linked endowment assurances and immediate annuities to
its product range.
Comment on how competition may affect the design and pricing of these products. [7]
16.2 This question relates to the same product launch that featured in Practice Question 11.3. The
details of the question are repeated below for convenience.
Exam style
A life insurance company has successfully been selling a unit-linked single premium savings
contract through insurance intermediaries. The company is now considering introducing a new
unit-linked regular premium savings contract in order to appeal to an even wider range of
investors.
The following design is being considered:

 fund management charge: 1% per annum
 minimum annual premium: £2,000
 minimum term: ten years
 surrender penalty: 0.5   x  10  y  % of the face value of the units at the

date of surrender, in the first ten years only (x = entry age,
y = age at surrender)
 all charges are guaranteed for the duration of the contract
 the face value of the units is payable on survival to the end of the term or on earlier
death.
The risks that the insurance company faces from launching the product have already been
considered.
Discuss the other considerations that the company would need to take into account before
deciding on the product design. [12]

16.3 A medium-sized American company writes primarily with-profits business using the contribution
Exam style method of profit distribution. The marketing manager has drawn the product development
actuary’s attention to a competitor’s whole life product which seems to offer, compared with the
company’s own product, the following three advantages:
 lower premiums
 higher dividend rates on its policies in the previous year
 the possibility of doubling the sum insured every five years without medical evidence.
(i) Discuss how the competitor might be able to achieve these differences. [10]
The marketing manager wants the actuary to redesign and reprice the company’s whole life
product to match this competitor.
(ii) Outline the factors that should be taken into account in trying to meet this request. [9]
[Total 19]

16.4 An established life insurance company currently issues a range of both conventional and
Exam style
unit-linked products. It now intends to launch a new regular premium unit-linked savings
contract, designed to help people meet their needs for retirement planning.
Under the policy, premiums are automatically linked to the national average earnings index, but
otherwise cannot be varied by the policyholder except by making the policy irrevocably paid up.
A wide range of investment fund links will be available; the policyholder can switch between
funds at any time, but no more than one switch is allowed each year without charge. The
policyholder can choose to take the policy benefits at any age between 55 and 75. On death
during the policy term the full bid value of units will be paid.
Two versions of the contract are planned. Under the standard version of the contract, at the time
of retirement the full bid value of the units will be used to purchase an annuity from the insurance
company, according to the company’s annuity rate basis current at that time.
The other version additionally provides a guarantee that the annuity conversion rate applied to
the retirement fund will not be less favourable than a rate specified from outset in the policy
conditions. The guaranteed rate will vary according to the chosen retirement age and the gender
of the policyholder.
Switching between versions will not be permitted once the policy has started.
The following standard charges apply to both versions:

 bid-offer spread: 5%
 annual fund management charge: 0.75% pa
 an initial charge of 15% of the first year’s annual premium, or £150 if larger, deducted by
cancellation of units at the start of the policy
 an additional 1% pa fund management charge will be levied for the annuity rate
guarantee.
All charges are guaranteed.
(i) Discuss the factors that would determine whether or not this product will be
marketable. [13]
The company is considering whether the additional 1% pa fund management charge is

appropriate for the annuity rate guarantee, both in terms of the structure and size of the charge.
The standard version of the product adequately meets the company’s profit criteria.
(ii) Describe how an assessment of the product would be carried out for this purpose,
assuming that all the company’s relevant data are made available. Your answer should
include a description of the model that might be used. [17]
[Total 30]



16.1 The structure of the new products must be similar to that of existing products in the market. For
instance, if all other endowment assurance products offer an enhanced benefit on accidental
death then the company’s product should do so as well. [1]
On the other hand, if the company is deliberately seeking to exploit some particular market niche
then it might want very different products. [½]
If competitors’ products all offer certain guarantees, then the company should aim to offer these
guarantees also. [1]
Investment guarantees are especially significant for unit-linked business. [½]
The commission (or other sale costs) should be similar to that paid by competitors. [1]
The premiums must not be much greater than those of the competition’s related products
(although this is not critical for unit-linked products). [½]
For unit-linked business it is more a case of expense and mortality charges not exceeding
significantly those of competitors’ rival products. [1]
The extent to which premium size (and expense / mortality charging) is important depends also
on the distribution channel(s) involved. [½]
Competitive pressures will be greatest for distribution channels that enable applicants to compare
products from several insurance companies, ie insurance intermediaries and price comparison
websites. [½]
In considering expense charges, the company must consider how competitors’ products treat the
issue of large premiums v small premiums – what sort of expense cross-subsidies are common in
the market? [½]
It will need to offer a similar choice of funds to that available from other companies (unless it is
intending to produce a very distinctive and different product compared to the market). [½]
The fund return will be very important. [1]
This might mean that, if the company does not have any funds whose past performance it can
proudly point to, it will need to consider basing the product on some external funds. [½]
[Maximum 7]
16.2 Marketability and distribution
Does it meet a real need for savings, so that brokers would recommend it? [½]
Having a surrender penalty would seem to be an unattractive feature for a savings contract. [½]
Does the charging structure lead to good value for money for the policyholder? [½]
Probably not, due to the cost of the guarantees and not brilliant capital efficiency. [½]

Are the fund links attractive, and is there enough choice? [½]
Do policyholders value the guarantee of the charging rate, or would they prefer a cheaper
product without the guarantee? [½]
The product is probably easy to understand, compared with some unit-linked designs. [½]
Are the limits (premium size, policy term) too restrictive? [½]
Is it appropriate to use the same distribution channel as for the single premium business? [½]
Are the commission rates offered to brokers sufficient reward for selling the business? [½]
Competitiveness
How does the product compare with competitors’ products? Are the features (guaranteed
charges, surrender penalty, minimum premium / term, choice of fund links) attractive relative to
the competition? [1]
Are brokers going to recommend it in preference to others? [½]
How do the charges compare (in amount and type)? [½]
Can it be promoted on the strength of the track record of fund performance from the existing
single premium product? [½]
Profitability
Does the product include adequate margins in order to generate the required return on capital?
[½]
Probably yes, as safety margins should have been included to cover the risk from the
guarantees. [½]
Cross-subsidies
Is the company happy with the extent of cross-subsidy occurring within this product? [½]
There is cross-subsidy by size (big subsidises small) and by term (long subsidises short). [1]
Price differentials could be introduced if the company wanted to reduce the extent of
cross-subsidy. [½]
Consistency
The company needs to ensure consistency with the existing product, particularly in terms of
charging rates. [½]
Must compare what policyholders would get back from each policy after one year, after having
paid the same premium: can any differences be justified to the policyholder? (eg it would seem
unfair to make the new product significantly more generous than the old one). [½]

The impact of the new product on the profitability of the single premium business must be
considered, eg withdrawal rates might rise and/or new business volumes might fall for the single
premium product. [1]
Does the company have adequate systems in place to administer the product? [½]
The company already has a single premium product so this should be OK. [½]
May have to upgrade the system to cope with the processing of regular payments and with the
surrender penalty, but these would probably not cause too many problems. [½]
Staff might need to be suitably trained to deal with the new product. [½]
Sensitivities
How sensitive are future profits to variations in future experience? [½]
The company would need to perform some sensitivity tests, including whole office projections of
future embedded value and statutory solvency allowing for different new business volumes and
differences in other key assumptions. [1]
Fairness
(Fairness in relation to the existing product has already been discussed.)
The guaranteed charges may end up seeming unfair if experience generally improves. [½]
The surrender penalty might infringe requirements for fairness if it has the appearance of a
‘concealed’ charge. Full disclosure and explanation of the charge at the time of sale would be
necessary in order to avert such accusations. [1]
Alternatively, a more transparent charging structure may need to be considered. [½]
Regulatory requirements
The company must ensure that the product design meets any other relevant legislative
requirements and constraints. [½]
[Maximum 12]
16.3 (i) How the competitor might be achieving the differences
Lower premiums
The lower premiums result from better expected experience in the premium basis. [½]
This may be justified or not justified (in which case it would simply be aggressive pricing). [½]
For example, the competitor might price the product on a marginal basis as a loss leader (or a
‘zero-profit’ leader). [½]

The competitor could have set the expense and mortality elements in the premium basis to be
low, and the yield assumption to be high, ie to have low margins. [½]
This would mean that, at parity of experience, they release less surplus than other companies
with larger margins (such as the company itself). [½]
The following are possible reasons why the better experience assumptions might be justified.
Mortality
The other company could appeal to a more healthy sub-set of the population. [½]
It could have stricter underwriting. [½]
Is like being compared with like (eg taking care not to be comparing non-smoker rates against
‘universal’ rates)? [½]
Expenses
The competitor company might be more efficient. [½]
It might have lower acquisition costs. [½]
It might price for shareholders with a lower rate of return requirement. [½]
Interest
Does the competitor have better investment management? [½]
Does the competitor have more free assets allowing riskier and higher yielding investments? [½]
Recent good luck? [½]
Higher dividends
The fact that it is only one year to justify suggests a fluke. [½]
The most volatile element in the contribution basis is likely to be the investment return (assuming
a mature company so expenses are not volatile, and having a large portfolio of policies so
mortality is not volatile). [½]
It could easily happen that in one year the competitor’s investment return is so much higher that,
even if the guaranteed premium basis yield is higher, the distributed yield is still higher. [1]
(For example, the competitor guarantees 5% and achieves 10%. There is a gap of 5 percentage
points of which they distribute 90%. Result: a 4.5% dividend. The company itself guarantees 4%
and achieves 8%, generating a gap of 4 percentage points of which it distributes 90% to give a
3.6% dividend.)
The proportion of surplus that the competitor distributes might be greater. [½]
The competitor may smooth its dividend payouts differently. [½]

If they smooth less, this is likely to exaggerate the above effect of a one-off investment profit on
one year’s results. [½]
The company may pay a higher terminal dividend than the competitor. [½]
This will mean that it pays out less annual dividend than the competitor, but could be paying
greater amounts over the term of the policy as a whole, … [½]
… due to the greater investment freedom it should have. [½]
Guarantee of future insurability
It is probably a risky marketing ploy. The competitor may have costed for it and increased the
mortality loading appropriately but if their premiums are lower then it is not very likely. So it is
probably not costed for adequately. [1]
[Maximum 10]
(ii) Factors in redesigning and repricing
A sensible approach is to see if the other product can be matched in all respects excluding the
future insurability guarantee. If it can, the actuary will then go on to consider to what extent they
can offer the same guarantee.
First, what is the risk discount rate? It must give shareholders their required rate of return with
sufficient probability. [1]
The actuary will need assumptions for mortality, expense and withdrawal experience founded on
recent investigations. [1]
Profit test the competitor’s premiums on the company’s own experience basis and risk discount
rate. [½]
This will need to use their dividend rates. [½]
Is the product profitable? [½]
If so then the actuary can redefine the expense, mortality and interest elements of the premium
basis to duplicate the competitor’s premiums. [½]
Is it acceptable in terms of capital use? [½]
Need to redo the projection with this new pricing structure, check that there is no problem as
regards capital (ie that there is no problem with solvency compared with before). [½]
But does the company want to introduce this product? There could be problems of inconsistency
with other products (especially if bonus rates need to be altered). [½]

Consider whether the company needs to introduce the product: [½]

 has it suffered a loss of new business production due to this competition? If so, does it
want to compete on price alone? [½]
 is the company’s own product more attractive in other ways (eg due to higher terminal
dividends, or having greater smoothing of dividends)? [½]
The second point above might provide alternative ways of competing. [½]
Now look at the insurability guarantee. It is very generous. Policyholders could multiply their
sums assured by a factor of 16 over a 20-year period. [½]
There would be significant risk of selection against the company. [½]
At a qualitative level the company would not want to offer this. [½]
Quantitatively, the actuary can assess the cost by modelling the extent to which it is exercised and
the mortality of those exercising it. This can be a deterministic or (preferably) stochastic
model. [½]
It will probably be a significant cost. Given that cost, is it appropriate to offer it for free? Profit
test again inserting a notional cost for the guarantee. [½]
The guarantee may become unprofitable, in which case the company cannot offer it for free. [½]
What is the effect on premiums of loading for it? They may now be uncompetitive. [½]
In which case the company does not want to offer that guarantee as it stands. If it is a significant
marketing factor then can the company offer something less costly? [½]
For example, limit the increases to be the lower of the change in a published price inflation index
or 10% per annum. [½]
[Maximum 9]
16.4 (i) Factors affecting marketability
Does it meet the customers’ needs?
The product meets the fundamental need to provide an income after retirement. [1]
The importance of this is affected by government provision at retirement, and by people’s

propensity to save. [½]
Provided the annuity pays income for the whole of life, this should protect the individual after
retirement from the financial risks from living longer than expected. [½]
The wide choice of retirement dates, without penalty, should be attractive. [1]
Unit-linked structure might appeal to risk-tolerant investors, as they should benefit from good
investment returns. [½]

Unit-linked may not appeal to all, as it carries significant personal investment risk (the unit-fund
value itself is not guaranteed in any way). [½]
Choice and flexibility of fund links may be attractive to some. [½]
The complexity of the product, particularly the need for the policyholder to take an active interest
in investment decisions (which he or she may not fully understand), may put off all but the most
financially sophisticated clients. [1]
The charging rate guarantees should be attractive. [½]
The guaranteed annuity rate could provide an attractive financial safeguard, though this depends
on the scale of the guaranteed rates offered and whether there is any realistic likelihood of the
guarantee biting in practice. [1]
It is useful for retirement savings products to be flexible with regard to contributions (premiums)
so as to reflect variable earnings: this product is particularly inflexible in this regard. [1]
However, the inflation-linking of premiums is an attractive feature … [½]
… although an individual’s salary increases are unlikely to be consistent with the growth in
national average earnings (they are likely to be higher than average at younger ages, and lower
than average at older ages). [½]
The apparent prohibition of transferring the policy to another provider, whether before or at
retirement, would seem to be very restrictive, and may even be illegal. [½]
Making the policy paid up will not be attractive to someone who might want a few years’
premium holiday in the future. [½]
The death benefit may not meet policyholders’ individual requirements for financial protection on
death: it may be too much or too little. [½]
The policyholders’ retirement provision may be vulnerable to long periods of being unable to
work through sickness, as there is no premium waiver benefit. [½]
Is the product competitive?
Successful marketing will depend on the product being competitive. [½]
Is the product similar to competitors’ in terms of charging rates and charging structure? [½]
For example, the initial charge for policies with premiums of much less than £1,000 pa may look
expensive. [½]
The possible lack of fund-transfer facility makes the company’s current and future expected
position in the annuity rate league tables very important. [½]
How do the features of the policy compare (eg premium and benefit flexibility)? [½]
Are the levels of guarantees similar? [½]

Does the company have a competitive track record of investment returns on its other unit-linked
policies? [½]
How is the product to be marketed, and to whom?
Success will very much depend on whether the product is successfully targeted at the right
market, whether the sales channel chosen is appropriate, and whether it is properly advertised to
both customers and intermediaries. [1½]
Marketing success will also depend on the general standing of the company in the marketplace as
a whole, and on its current and recent record of financial solvency. [1]
Whether it has been involved in any bad (or good) media publicity will also influence sales. [½]
[Maximum 13]
(ii) Assessing the suitability of the additional 1% fund management charge
Often exam questions draw on ideas from throughout the course. To answer this part we need to
use ideas from the modelling and product design chapters. It would be easy to lose marks here by
only covering one of these topics, although the question does ask us to include a description of the
model which should help. It’s always a good idea to pause to think whether we can broaden our
answer by considering a range of topics.
Profitability
Determine a range of suitable model points, to reflect the expected mix of business. [½]
These would need to differ by gender, age at entry, chosen retirement age, chosen unit fund link
and policy size. [1]
Construct a single-policy cashflow projection model that could be used for each model point. [½]
Unit fund values would need to be projected for each year, using the guaranteed charging
structure, which will require assumptions about unit growth rates and inflation rates (for
premium growth) as further inputs. [1]
Annual cashflows would be calculated, for which further assumptions are required (mortality
rates, paid-up rates, expense and expense inflation rates, and non-unit interest rates). [1]
First calculate net present values (NPVs) for each model point for the standard policy, including
only the standard 0.75% pa fund management charge. [½]
Initially need to project on the (possibly prudent) supervisory reserving basis to establish the level
of non-unit reserves required each year, if any. [1]
Also need to project the expected future solvency capital requirements. [½]
Then project the expected profit flows using the (realistic) pricing basis, allowing for the
supervisory non-unit reserves and solvency capital, and discount at the risk discount rate to
obtain the NPV. [1]
Next rerun the model including the cost of the guaranteed annuity option. [½]

This will require a stochastic model of investment yields and inflation rates at the projected
retirement date. [½]
The model chosen would be defined in terms of an assumed probability distribution and its
associated parameter values. [½]
Links would be included in the model so that the investment and inflation variables behave
realistically in relation to each other. [½]
The model will then need to be able to calculate the normal rates of annuity conversion that
would be applied at retirement, for the given simulated investment yield and inflation rate. (The
model effectively has to reproduce the annuity pricing exercise that the company would need to
perform in those simulated conditions.) [1]
In those simulations that produce worse conversion terms than guaranteed, the additional cost of
providing the guaranteed rates would be calculated as at the projected retirement date. [½]
A large number of simulations would be run, and an appropriate estimate of the cost of the
guarantee obtained from it. [½]
The unit fund and cashflow projection models would now incorporate the additional 1% pa fund
management charge. [½]
The model would again have to be run first on the supervisory reserving basis, so that the non-
unit reserves can be established. [½]
These reserves (together with any required solvency capital) must enable the company to meet
the estimated cost of the guarantee at retirement, based on a prudent assessment. [½]
This could be found as the upper (say) 97.5th percentile of the distribution of simulated outcomes
of the guarantee cost. [½]
The new reserves and solvency capital are then incorporated into the premium basis cashflow
model as before, now including a realistic value of the expected cost of the guarantee at
maturity. [½]
This realistic value would be taken as the average of the simulated outcomes from the stochastic
model. [½]
The NPV is then recalculated, and compared with that previously calculated (for the same model
point) without the guarantee. [½]
Cases that show reduced NPVs would indicate undercharging for the guarantee. [½]
It may also be necessary to run the whole projection model stochastically, because the projected
cost of the guarantee will be a function of the fund size at maturity, which will itself be a function
of projected asset returns. The correlation between high fund growth prior to retirement and
investment yields at retirement will also need to be included. [1]

The NPV would be recalculated for each simulation, and the average taken. This answer would be
used if it differed markedly from the NPV found deterministically (including the cost of the
guarantee). It would be necessary to ensure that the assumed stochastic investment model
produces the same long-term average unit-growth rates as assumed for the deterministic
model. [1]
Levels of cross-subsidy
Comparing the results over the range of model points would indicate whether the additional
charging rate ought to vary by entry age, retirement age, etc, or whether an alternative charging
structure might be appropriate (eg have a monetary policy fee, or express as a proportion of
premiums). [1]
Rerunning the models to find appropriate charging rates for alternative charging structures could
be done to assess whether any alternatives might be more suitable. [½]
Alternatively the company might consider the level of cross-subsidy implied as acceptable or even
desirable (eg for marketing reasons). [½]
All results should be sensitivity tested, using different assumptions, including different parameter
values for the probability models. [½]
An important sensitivity to check will be mortality, where a different mortality basis at the time of
retirement could significantly affect the guarantee cost. [½]
Alternative ways of charging for the guarantee might lead to different sensitivity. [½]
This might include the possibility of allowing the charge to be reviewable in the future. [½]
Capital efficiency
Consider whether an alternative structure might be more capital efficient. [½]
Marketability and competitiveness
An alternative structure might be more marketable. [½]
Competitors’ charging rates for similar guarantees should be checked, if they exist. If the rates
appear uncompetitive, choosing a different charging structure could avoid unfavourable
comparisons. [1]
Consistency with other products
Check whether the charging structure is similar to those on any of the company’s other products
that offer a similar guarantee, so as to avoid any unfavourable comparisons being made. [½]

Check that either the company can administer the different versions of the policy properly,
without errors, or is able to develop the necessary systems and processes within the required
timescales and budget. [½]
[Maximum 17]

End of Part 3
What next?
Chapters 13 to 16.
of your revision.
… ‘learning and revision’ and ‘revision’ products
Online Classroom – As an alternative / addition to tutorials, you might consider the Online
Classroom to give you access to ActEd’s expert tuition and additional support.
‘Please do an online classroom for everything. It is amazing.’
Flashcards – These are available in both paper and eBook format. One student said:
‘The Flashcards were useful for highlighting what we needed to memorise

and for revision on-the-go.’
You can find lots more information, including samples and demos, on our website at
www.ActEd.co.uk.

SP2-17: Setting assumptions (1) Page 1
Setting assumptions (1)

Syllabus objectives
(Only part of Syllabus Objective 5.1.1 is covered in this chapter)

Page 2 SP2-17: Setting assumptions (1)
0 Introduction
The next two chapters of the course deal with the construction of a basis. By basis, we mean the
set of assumptions that would be used in an investigation.
This process usually requires the actuary to make fairly long-term assumptions for each
parameter in the assumption set. In this part of the course we explain how the actuary can
decide on what is a suitable assumption for each parameter. The key part of the process is the
determination of what future experience is expected. A second aspect is what other factors may
need consideration: for instance, the extent to which margins against adverse future experience
are required. What these other factors are will depend greatly on the application, or purpose, of
the basis. Here we deal with three different applications: pricing, reserving and ongoing
profitability measurement.
A basic methodology for setting assumptions

Deriving assumptions for future experience follows a common framework.
1. Investigate the historical experience and make best estimates of the parameters from
that experience. These estimates will be appropriate in the context of the historical
conditions and circumstances that applied at the time of that experience, including the
condition of the commercial and economic environment.
2. Consider what the conditions (including the commercial and economic environment) will
be like in the future period for which we are making our assumptions.
3. Determine what the best estimates of our assumptions will be, given the expected future
conditions. The historical parameter estimates will be used as a starting point, (when
possible or appropriate), but the effect of the different circumstances of the future,
compared with those of the past, must be allowed for. In other words, how will the
parameter values change given the changed conditions?
4. The extent to which we would rely on the experience data, and the extent to which we
allow for other factors, including judgement, depends on the credibility and relevance of
the data, and how predictable the parameter is.
5. The best estimates may need to be adjusted in order to include a margin for prudence.
The size of any margin incorporated into an assumption will depend on the purpose for
which the model is required, and the degree of risk associated with the parameter.
In the first of these chapters on basis construction we look at all the demographic and financial
aspects involved in deciding on a pricing basis. In Chapter 18, we see how all these issues may be
affected by the precise purpose of the basis.
The factors involved are always considered at a generic level; precise knowledge of suitable
assumptions (eg the mortality assumption appropriate for term insurance policies in Germany) is
not required.

1 Pricing life insurance contracts

In Sections 1.1 to 1.6 we are referring to how we would obtain best estimates for our
assumptions. The subject of what margins to consider is covered in Section 1.7, which further
relates to the profit requirements covered in Section 1.8.
1.1 Mortality
Mortality rates are changing over time and the good news is that they are getting lower each year
in most parts of the world. So, we will consider the mortality assumption in two parts:
 the base mortality, ie the initial rate of mortality
 the mortality trend, ie how the rate of mortality changes over time.
Base mortality
In the context of the SP2 examination, the main demographic assumption that is used to
price a life insurance contract relates to mortality rates.
This part of the course assumes the ‘ideal case’ of an actuary having complete freedom to decide
on the most appropriate assumption. However, it will often be the case that the regulatory
authorities will constrain, perhaps severely, this choice in a pricing context.
The values assigned to these rates should reflect the expected future experience of the lives
who will take out the contract being priced.
The use of the word ‘reflect’ rather than ‘equal’ is important. As we shall see frequently in this
section, the expected future experience is the starting point in putting a value on a parameter.
We would then consider what margins might be appropriate.
The expected future experience of the policyholders will depend crucially on three things:
 the target market for the contract – this will be dependent on the distribution channel
involved
 the underwriting controls applied (or not applied)
 the expected change in the experience since the time of the last historical investigation to
the point in time at which the assumption will on average apply (typically we would be
looking about ten to fifteen years into the future).
These values will most likely be based on an adjustment to rates from a standard mortality
table.
Even a company with a very large body of suitable experience would probably not construct its
own tables, but would use the mortality data to adjust a standard table. This is both to save
resources, and to protect against the errors which might easily arise with an inappropriate
graduation (especially for ages where there is little data).
If the company has adequate data, the adjustment would be derived by analysing the
company’s own experience for the type of contract concerned. Alternatively, the experience
of a similar class of business could be used as a substitute.

This alternative may provide appropriate data, depending on the similarity between the contracts
concerned. For example, suppose we want to price a new with-profits whole life contract.
With-profits whole life and long-term with-profits endowment assurances may have similar
mortality experience. However, move a little further (eg to a pure endowment) and different
underwriting procedures and target markets may make experience much more significantly
different. Move even further (eg to a term assurance, where the customer needs differ greatly) and
experience will almost certainly be very different.
The data would relate to an appropriate period of years, such that the volume of data is
adequate, but excessive heterogeneity due to trends over time is not introduced.
There is a conflict here between the desire to have a large pool of data, and a requirement not to
look at significantly different generations of lives. A ‘generation’ in this context could be just five
or ten years. If the data include lives who enjoy different mortality because of different
generations then the mortality implied by the data may be misleading. This will be so if
differently aged lives have seen different mortality improvements from one generation to the
next. So mortality investigations are normally based on three or four years of data.
The analysis would divide the data into relevant homogeneous groups, subject to adequate
levels of data being retained within each cell.
Question
List the homogeneous groups that are most likely to be used for a term assurance mortality
investigation.
Solution
Homogeneous groups are likely to be derived by categorising according to:

 age (not necessarily by year of age – at extremes of the age spectrum five-year groups
might be sensible)
 gender
 medically underwritten (normal life) vs medically underwritten (abnormal) vs not
medically underwritten
 smoker vs non-smoker
 possibly some form of broad grouping by occupation and/or geographical area.
If the company has insufficient data to produce reliable results, or has no appropriate data
at all, industry sources, such as the Continuous Mortality Investigation reports and working
papers produced in the UK, or life reinsurance companies, would be used instead.
Question
Explain why a company would not just use the latest standard mortality table.

Solution
Standard mortality tables are based on the experience of a large number of companies. They
therefore represent an average experience which will not be suitable for individual companies
except by chance.
They are never wholly up to date. For instance, the most recent UK tables are typically based on
experience data that are between five and fifteen years old.
There might not be standard tables available suitable for the specific contract that is being priced.
Industry sources might be useful if they cover types of contract for which no standard table exists.
They may be good for showing trends, since trends in the company’s own data might be put down
to statistical variation. However, industry statistics are not based purely on the policyholders of
the specific company and so will not be 100% suitable.
Another potentially useful source of data might be population mortality statistics. In fact in many
countries it will be the first point of reference. Even in a country well furnished with good quality
‘assured lives’ data, population mortality could be useful for showing trends if it has been
re-examined at regular intervals in the past.
Question
Discuss the advantages and disadvantages of using reinsurers’ statistics.
Solution
Reinsurers have access to the mortality experience of many direct writers. Sometimes, theirs may
be the most relevant data available. This is most likely to be the case for a recently established
life company which will have little data of its own, or in the case of a new or substantially revised
contract.
However, reinsurers’ data suffer from the disadvantage that they relate to a large number of
companies. Each of these companies will have their own target market and underwriting
procedures.
In some circumstances, the reinsurer may have little or no suitable data.
The reinsurer will almost certainly want something other than gratitude for help in determining
the assumptions for a contract. If the insurer wished to reinsure anyway, then there may be no
real loss. However, if reinsurers’ technical assistance forces the insurer to reinsure more than it
would otherwise choose to do, the cost of that extra reinsurance is a problem.
If the adjusted rates are to apply to a class of lives which is expected to have a different
experience from that to which the analysed data relate, then further adjustments may need
to be made. This situation could arise due to a change in target market, distribution
channel, or the basis of underwriting and accepting lives.

The class of lives concerned is almost always going to have a different experience from that
underlying the data, just due to the influence of time.
Mortality trends
As we have already noted, the assumptions need to reflect the expected future mortality of
the relevant lives and therefore consideration should be paid to the expected changes in
rates over time.
This is a particular issue for annuities where increased longevity is a risk, so it is important
to project expected future improvements.
Question
Explain why it is more important to include projected changes in mortality where ‘increased
longevity’ is a risk.
Solution
This is because, for most of the developed world at least, the expectation is that longevity will
increase over time.
Similarly, for contracts that pay significant death benefits, any expectation of increasing mortality
(eg due to AIDS) should be included in the basis.
For some contracts, future changes may be reasonably ignored (although only after
consideration) – eg a single premium savings contract.
It should also be noted that allowing for expected trends in mortality rates in pricing is more
important if the premiums are guaranteed (non-reviewable).
That is, it is more important that we get the mortality assumption correct when there is no chance
of correcting our mistakes later.
There are several possible approaches to determining future rates of mortality

improvement, which are often used in combination:
 expectation approaches
 extrapolation approaches
 explanatory approaches.
Expectation approaches involve expert opinion and subjective judgment to specify a range
of future scenarios.
So the mortality assumption could be set by asking a number of experts for their opinion, eg by
asking what they think future life expectancy might be.
An advantage to this approach is that it can implicitly include all relevant knowledge (including
qualitative factors). However, expectations are subjective and can be subject to bias.

Extrapolation approaches are based on projecting historical trends in mortality into the
future. Such methods also require some element of subjective judgement, for example in
the choice of period over which such trends are to be determined.
Looking at trends in smoking rates in the UK is a good example of this. Smoking rates (and
resulting mortality rates) fell during the 1970s and 80s, levelled off in the 1990s, but started to fall
again in the 2000s. Extrapolating this trend into the future is far from straightforward and the
introduction of a smoking ban in public places throughout the UK, together with the increase in
vaping as an alternative, makes projection highly subjective.
Explanatory, or process-based, projections attempt to model trends in mortality rates from a

bio-medical perspective. These projections are only effective to the extent that the
processes causing death are understood and can be mathematically modelled.
Four major causes of death are:

 circulatory diseases, eg heart disease, stroke
 cancers
 respiratory diseases
 infectious diseases.
The absolute and relative rates of mortality due to these causes have changed markedly over the
past 90 years and an understanding of how these changes will continue into the future could help
with mortality projections. Since the mortality from each of these causes has had a different
pattern in the past, simply looking at rates overall will mask these underlying patterns.
For example, the use of statins to reduce cholesterol has had a big impact on the incidence of
circulatory diseases recently, but wouldn’t be expected to affect the other causes of death.
In some countries, particularly the more developed economies, observed mortality

improvements can vary significantly by year of birth or ‘cohort’.
This is often called the ‘cohort effect’, whereby mortality improvement rates appear to depend on
a person’s year of birth. For example in the UK, lives born between 1925 and 1945 have
consistently experienced higher mortality improvements year by year than the generations born
either side of them.
Question
Explain why this last fact was particularly significant for insurers in the 1990s.
Solution
Lives born between 1925 and 1945 typically hit retirement age between 1990 and 2005. With
mortality being significantly lighter for this group than for previous retirees, annuities became less
profitable during the 1990s, before the ‘cohort effect’ was widely accepted.

The causes of the cohort effect are open to interpretation, but an important contributory
factor may be social changes such as smoking behaviour.
Another suggestion is that this generation has benefited most from increased availability of
medical care and advances in treatment.
Thus, deterministic projections might allow for different mortality improvement rates
according to year of birth (as well as by age and gender).
So, for example, our rate of male mortality at age 80, q80,y , depends on the year of birth y . As
mortality rates have been falling over time we expect that:
q80,1920  q80,1940  q80,1960
That is, we expect that those born in 1940 will have lower rates of mortality at age 80 than those
born in 1920. Similarly, we expect that those born in 1960 will have lower rates of mortality at
age 80 than those born in 1940.
However, the cohort effect suggests that people born in 1940 have experienced higher rates of
mortality improvement than the people born in 1920 or 1960. So we expect that people born in
1940 will have much lower rates of mortality at age 80 than people born in 1920. However, we
expect that people born in 1960 will have only slightly lower rates of mortality at age 80 than
people born in 1940.
The period over which the ‘cohort effect’ is assumed to remain (before reducing to zero) is
also open to judgement.
We suggested above that lives born between 1925 and 1945 have consistently experienced higher
mortality improvements year by year than the generations born either side of them. However,
the Core Reading is suggesting that it is not clear how long the cohort effect lasts for, eg perhaps
the higher mortality improvements only apply for ages below 70.
Question
Explain why it is so difficult to know whether the cohort effect will apply to mortality at age 80 or
even 90 for lives born between 1925 and 1945.
Solution
The problem is that we will have to wait until 2020 to be able to observe the mortality of 80 year
olds who were born in 1940. We will have to wait even longer to observe the mortality of 90 year
olds who were born in 1945. So we cannot be certain whether the cohort will continue to
experience mortality improvements at higher ages.

Multi-factor predictive modelling techniques (eg using generalised linear models) are
increasingly in use. These techniques can combine internally held customer data (if there is
sufficient volume) with external drivers, allowing for correlations and interactions between
them.
We will see in Chapter 30 that the existence of ‘big data’ may allow a more accurate estimation of
a customer’s mortality by considering the data that we hold specifically about them, eg an
insurance company owned by a bank may have a considerable quantity of data on a customer’s
spending habits based on their credit and debit cards.
Stochastic mortality projections have become more widely used, such as the Lee-Carter or
P-spline method. These generate many different improvement rate scenarios, but can be
challenging to calibrate. One approach is to start by understanding the drivers of possible
future change and considering events relating to the major causes of death that may lead to
substantial future reductions.
You do not need to know how these models actually work for the purpose of the Subject SP2
exam. However, you do need to know that these models are used to project mortality trends so
that we can derive the values of qx ,y for each age x and each year of birth y .
One of the reasons that these models are so hard to calibrate is that we have many more
parameters than a model for the base mortality as we also need to model the trends. For this
reason, we often need to calibrate the models to national statistics to ensure that the volume of
data is sufficiently large to give credible results.
1.2 Investment return

The value assigned to this parameter will be affected by:
 the significance of the assumption for the profitability of the contract, which will
depend on the level of reserves built up and the investment guarantees given
This point is also important when it comes to considering the size of the margins required for the
investment assumption (see Section 1.7 below). However, the first step in considering what the
investment assumption should be is to establish how important the assumption is for the results
of the modelling we are performing. If, for example, there is very little sensitivity to the
investment assumption, then the actual value chosen for the parameter would not be critical.
The two key factors which lead to sensitivity to the investment assumption are the size of the
reserves built up (relative to the cashflow, for example), and the investment guarantees given.
The larger the reserves, the greater the proportion of total cashflow (and profit) that arises from
investment income, and hence the greater will be the sensitivity to changes in the investment
return. The higher the investment guarantee the greater the care needed over setting the level of
the assumption. It is possible for the choice of investment assumption to be critical, and yet the
contract may be subject to very little investment risk (provided the correct investment
assumption has been made). The following question includes an example of this.

Question
Comment on the importance of the investment return assumption for the following contracts:
(1) regular premium without-profits endowment assurance
(2) single premium without-profits endowment assurance
(3) regular premium without-profits term assurance
(4) single premium without-profits term assurance.
Solution
(1) Regular premium without-profits endowment assurance
Very important: the reserves are large and there is likely to be a significant investment guarantee
with respect to the maturity benefit.
(2) Single premium without-profits endowment assurance
Critical: the reserves are much larger than the regular premium case, and the guarantee is
(potentially) the same. (The need for margins may not be as great as for (1), because we might be
able to invest in matching assets and therefore secure a known investment yield when the policy is
taken out.) Nevertheless, it remains of vital importance to get the investment assumption correct,
that is, equal to (or lower than) the investment yield that the single premium will secure on the day
that it is invested.
(3) Regular premium without-profits term assurance
Not very important: reserves are small so that profits will be insensitive to changes in investment
return.
(4) Single premium without-profits term assurance
Very important: a much greater dependence on investment return than in the previous case, due to
the single premium. However, this is not as important as in (2) because the term of the liabilities
will not be as long (the reserves will decline towards the end of the term assurance policy while they
will increase towards the end of an endowment assurance policy).
 the extent of the investment guarantee given under the contract – this will affect the
types of assets in which the premiums from the contract will be invested
The more onerous the guarantee, the more cautious the life company should be in its asset
selection. This caution should then be reflected in the investment return assumption (because
the expected return will differ according to the assets selected – as described under the next
bullet point).

Question
Comment on the level of guaranteed investment return for the following contracts:
(2) single premium with-profits endowment assurance (using the additions to benefit
method).
Solution
High: the financial losses that would be incurred by not achieving the guaranteed rate of return
(ie the premium basis rate of interest) over the term of the policy will be high.
(2) Single premium with-profits endowment assurance
Low: while in construction the policy is the same as (1), much of the total future benefit is
determined by future bonus distributions, which can be reduced in future if investment returns are
worse than expected.
 the extent of any reinvestment risk and the extent to which this can be reduced by a
suitable choice of assets – the less important the reinvestment risk, the less account
needs to be taken of future investment yields
Reinvestment risk refers to the uncertainty of the return which can be obtained from investment
in the future. The best estimate of the return available from future investment may well differ
from the best estimate of the return available from investment now, and may also differ
according to how far into the future the investment is made. The overall best estimate
investment assumption will reflect the expected balance between the expected future and
current investment yields.
If the real cashflow is positive over a future period, then the company will need to buy assets.
The more that such future investment is expected to occur, the more the investment assumption
should reflect the expected future investment returns. Even if the real cashflow is predicted to be
negative in the future, mismatching of assets and liabilities by term may still mean that assets
have to be sold and bought in the future, so that the future reinvestment yields will still be
important. However, the greater the degree of matching, the less influence future investment
rates should have on the investment return assumption.
Of course, when we are pricing a new contract then all investment is to be made in the future.
Nevertheless we will have a much better idea of what investment returns are likely to be at the
point of sale of the contracts we are pricing (which will take place in the near future) as compared
to much later in the terms of those contracts. The expected time between changes to the
company’s pricing basis for a contract will therefore also be important in establishing the best
estimate investment assumption to use.

Question
Comment on the exposure to future investment conditions in the following contracts, and on the
ability / desire to remove that exposure:
(1) regular premium without-profits endowment assurance
(3) regular premium without-profits term assurance
(4) single premium without-profits immediate annuity contract
(5) regular premium with-profits endowment assurance
(6) single premium with-profits endowment assurance.
Solution
(1) Regular premium without-profits endowment assurance
Total, cannot match until late in term.
Low, but still some exposure to reinvestment risk, hard to get rid of totally (eg due to reinvestment
of future investment income).
(3) Regular premium without-profits term assurance
Limited exposure.
(4) Single premium without-profits immediate annuity contract
Can be removed by matching.
(5) Regular premium with-profits endowment assurance
Total, would not wish to remove exposure, because we want to maximise returns.
(6) Single premium with-profits endowment assurance
Could be limited, but in practice will want more exposure to future investment conditions for the
same reason as (5).

 the intended investment mix for the contract, as affected by the above, the current
return on the investments within that mix and, where appropriate, the likely future
return.
This is, of course, a very important point: consider the likely mix of assets which will back the
contract in the future, investigate the returns that such assets are yielding now (and in the past),
and attempt to predict the returns that will be obtained from the future asset mix bearing in mind
the impact of future changes to the economic environment, in particular.
The intended investment mix for the contract will not be derived from looking at the contract in
isolation, as it will be affected by the level of free assets available to the company to support
writing the business. This is because the extent of matching necessary to control the investment
risk will be less when there are more free assets, hence affecting the mix of assets used.
Market consistency
If a market-consistent approach is used, either deterministically or stochastically, then the
expected investment return can be set as the risk-free rate, irrespective of the actual
underlying assets held (this is the ‘risk neutral’ calibration approach).
So if we are pricing an immediate annuity, we could use a deterministic model to obtain a

market-consistent value by discounting using the risk-free rate appropriate to the term of each
cashflow. The actual assets we intend to hold are irrelevant.
However, if the stochastic approach is adopted, then the investment return volatility and
correlation assumptions do remain dependent on the actual underlying asset type(s).
So if we are pricing a maturity guarantee on a unit-linked fund, we could use a stochastic model to
obtain a market-consistent value. We would calibrate the expected investment return to the
risk-free rate, but would calibrate the volatility and correlation assumptions to the assets backing
the unit fund.
Allowing for taxation

If the fund is subject to tax on its investment return, then the appropriate tax treatment of the
investment return must be allowed for. This will involve applying the appropriate rates of tax to
the different components of the investment return, for example to income and to capital gains.
Different rates of tax may also need to be applied to investment income from different asset
types (for example, from fixed-interest securities or from equities).
When considering what future tax rates to assume, the best we can really do is to use current
rates, allowing for any future changes which we know are going to happen.
Allowing for future bonuses

When we described with-profits policies earlier in the course, we described how much higher the
premiums were for a with-profits policy than for a without-profits policy for the same basic sum
assured. Essentially what the company has done is to make some sort of assumption about the
future bonuses that it intends to declare, and included their value in the premium rate.

One way of achieving this is by assuming very conservative assumptions in the premium basis,
which would usually be done through a reduction in the interest rate assumption. Alternatively
we could use more realistic assumptions but include an explicit allowance for future bonuses in
the premium rate.
1.3 Expenses and commission

The parameter values for expenses should reflect the expected expenses to be incurred in
processing and subsequently administering the business to be written under the product
being priced.
The actuary will want to reflect the incidence of the following marginal expenses:
 initial acquisition (eg initial commission and related sales costs)
 initial medical underwriting
 initial administration
 renewal administration
 renewal reward to sales channel (ie renewal commission, or similar)
 investment
 withdrawal / paid-up expenses
 claim / maturity administration (often known as termination expenses).
The expense loadings should also contain some contribution to the company’s fixed overheads.
These will be any expenses not covered in the above marginal categories, but the distinction
between marginal and fixed is subjective. For instance the property costs of the ‘new business
admin’ department could be considered as part of the marginal initial administration expenses, or
as part of overall fixed expenses.
Usually there would be no loading for the administration of withdrawals or policies becoming
paid-up; instead, the cost would be allowed for in setting the appropriate discontinuance terms.
The values will be determined after analysing the company’s recent experience for the type
of business concerned. The result of this analysis will be a division of the expenses by
function, as appropriate, and possibly by whether the level of expense is expected to be
proportional to the level of premium or benefit, or can be expressed as an amount per
contract.
The mechanics of an expense analysis are covered in Chapter 30.
Question
Outline how the above eight types of expense (plus the contribution to overheads) should be
allowed for in a premium basis.

Solution
Initial acquisition
Load for commission exactly as it occurs – almost certainly a percentage of the first year’s
premium.
Related sales costs (ie costs of sales process attributable to the sale of that one policy) would
probably be as a percentage of premium for consistency with commission. Any sales staff are
likely to devote their time and attention in accordance with the size of the policy.
Initial medical underwriting and administration
Partly as a flat average cost per policy, and partly as a percentage of benefit size, in year 1.
Renewal administration
A flat (plus inflation) average cost per policy in years 2 onwards (or from month 2 onwards).
Renewal reward to sales channel
Load for this exactly as it occurs – almost certainly a percentage of premiums from the second
year onwards.
Investment
This will occur as a percentage of reserves. A deduction from the interest rate would be a simple
way of allowing for this.
Withdrawals / paid-up policies
These probably wouldn’t be loaded for in the premium basis. They would be allowed for when
specifying discontinuance terms.
Claim / maturity administration
Load as a fixed (plus inflation) per-policy cost occurring at a suitable time.
Contribution to overheads
What can a policy reasonably contribute to overheads after having financed its own marginal
costs? What it can reasonably contribute will be some proportion of premium, even if we rightly
argue that the overheads are invariant by policy size (implying some flat fee approach).
The above indicates how these types of expense should be loaded for. How would they be loaded
for in reality? The main difference will probably be a reduction in the amounts loaded for as flat
policy fees, re-costing them as proportional to premium for reasons of competitiveness.
In the context of expense loadings, the words fixed and flat are used rather loosely. They refer to
expenses that do not vary by premium size but they are not necessarily fixed or flat in the full
sense of the word. For example, future administration expenses should increase with inflation.

If the company has insufficient recent experience to provide meaningful results, or suitable
recent experience is not available, the parameter values may be based on a similar type of
business and, if this is not available or not reliable, on any industry data or data from a life
reinsurance company.
More commonly, the expected expenses would be derived from the construction of a suitable
expense model. This will simply be a projection of the staff structure and associated overheads,
such as buildings, systems etc. The output from that will be taken in conjunction with expected
new business volumes (and perhaps the model office output to give current and expected in-force
volumes and assets under management) to give suitable per-policy or per-premium costs. These
results would then be tested for reasonableness by comparison with industry averages.
Variations of this technique will be used for studying (in order of complexity):
 new product launches
 new distribution channels (or major extensions to existing channels, or major new sales
campaigns)
 new life insurance companies.
1.4 Dealing with per-policy expenses

In the solution to the previous question, we included per-policy expenses (those that do not vary
by policy size) explicitly as such in the basis. Part of the basis is also the average benefit size
assumed for the particular policy. The pricing process will usually lead to a premium rate,
expressed per unit of benefit, such that for the average sum assured the correct per-policy
expense loading will be made.
This means that, whilst we might have thought of the expense loading as being per policy, it will
vary in practice in proportion to the size of the policy actually taken out.
Question
Explain why this might be a problem, even if the average policy size assumption turns out to be
realistic.
Solution
Large policies are subsidising small policies. The degree of overpayment by large policies could be
so great as to make them very poor value for money, going against principles of equity and quite
possibly making them uncompetitive.
We now discuss possible ways of correcting for this inequity, where it occurs.
An area of risk relates to how to incorporate into the charging structure the expenses which
do not vary by size of contract.

The following are among the ways of coping with this risk:
 Individual calculation of premium rates or charges.
With this method, the actuary simply calculates the premium or charges for the contract using the
actual benefit level, age, term, etc for the policy in question, loading explicitly for the per-policy
expenses. This would probably only be done for the large cases: very small policies priced on this
basis would end up paying a very large percentage of their premium as expense loading – the
premium rate would then seem unreasonably large.
 Policy fee addition to the premium – or deduction from regular benefit payments –
for non-linked contracts, or charges that match the per-policy expenses for
unit-linked contracts. This approach may need to be modified if the policy fee or
charges would be uncompetitive.
This is the same problem as we mentioned for smaller policies using the first method.
The idea is that the policy fee would represent the per-policy expense costs, and the premium
rate would only cover the other, proportionate, expenses. However, the policy fee may need to
be reduced from its theoretical level in order to make the total premiums for small policies more
competitive (and, perhaps more importantly, more reasonable). But to avoid not covering
expenses, the reduction will need to be compensated for elsewhere.
So the actuary will reduce the policy fee to the level required on grounds of reasonableness
and/or competitiveness, given where the company wants to be in the market. The contribution
to expenses determined as a percentage of premium will then need to be increased. A model
portfolio would be used to ensure that the total expense contribution received by the company,
for the assumed volume and mix of business, would be the same as what it would have been
using the theoretically correct premium rates and policy fee.
However such modifications would lead to increased risk for the company.
Question
Describe the risk that is referred to here.
Solution
The risk here is that of then selling more smaller-sized policies than anticipated (either more in
absolute terms, or as a proportion of the new business portfolio). If this happens the company
might not recoup the marginal expenses. In addition the contribution to fixed overheads would
be smaller than necessary.
 ‘Sum assured differential’. For non-linked contracts, different premium rates are
charged according to which band the benefit requested falls into. For unit-linked
contracts, different charges, for example allocation rates, are applied according to
which band the premium payable falls into.
The same idea can also be applied with equal effect (in fact perhaps more efficiently given the aim
here) to premium size on non-linked contracts.

Example
In Italy, one method of allowing for expenses that do not vary by size of contract has been to give
a discount for large premiums.
For instance, for regular constant premium endowment assurances:

 The first €X of each year’s premium is applied at normal rates to give the benefit insured.
 All premium in excess of that €X is applied at a discounted rate to give the benefit insured,
the discount extending up to 10% for policies with longer terms.
Of the above three ways of dealing with expenses that are invariant by size of contract, the most
common approach is the second method (policy fee addition). However the approaches are not
mutually exclusive: when the third approach is used it is also often combined with some form of
flat policy fee.
The parameter values for any commission will be set by the company and are likely to take
account of the rates of commission normally paid in the market in which the company
intends selling the product.
These commission assumptions should take into account any intended extra payments made to
salesmen, brokers etc on achievement of certain high production targets.
The treatment of expenses may be affected by the tax regime applicable; a life company taxed on
‘I –E’ can allow for this tax in its expense assumption, by netting the expenses down
appropriately, but this treatment must be consistent with that of its investment return (by
assuming all of the investment return is also fully net of tax).
1.5 Inflation of expenses

An inflation assumption will be necessary to apply to loadings in respect of renewal
administration and policy termination expenses. These will be affected primarily by earnings
(rather than price) inflation, because the majority of an insurance company’s costs are staff costs.
The following may be considered when setting the value of the inflation parameter:
 current rates of inflation, both for prices and earnings
 expected future rates of inflation
 the differential between the return on government fixed-interest securities and on
government index-linked securities, where such exist
 recent actual experience of the life insurance company or industry.
The third approach is really just one way of arriving at the expected future inflation of the second
point. Also it is only approximate – it will be out by the extent of any inflation risk premium
implicit in the price of the government fixed-interest bonds.
The inflation assumption must be consistent with the future investment income assumption.

When pricing a contract, as we are here, we have to consider two distinct aspects of expense
inflation:
 the inflation of expenses during the term of a future new policy, from the issue date to its
termination date
 the inflation of all expenses between ‘now’ (the date at which we are setting premium
rates to be used in future) and the dates at which the future new policies are actually
issued.
As an example of the second point, we might set our per-policy initial expense assumption at
£250, but for a policy issued in one year’s time it perhaps ought to be £260, for one issued in two
years’ time it should be £272, and so on.
So we need to make the overall future expense assumptions appropriate to the future period over
which we would expect the pricing basis to be used. In the case above, for example, if we
expected our basis to apply for the next four years, say, we might fix our initial expense
assumption at £275, for all policies to be issued over the four-year period.
1.6 Persistency
If deriving premiums from a formula approach (as opposed to the profit testing emerging
cashflow approach) it is unlikely that any allowance will be made for withdrawals. This is one of
the weaknesses of the simple formula approach. However the resultant premium rates will need
to be profit tested, and the basis for the profit testing will need to incorporate suitable
persistency assumptions, ie the profit test will be performed assuming that some proportion of
policyholders withdraw each year.
The persistency (full withdrawal, partial withdrawal and paid-up rates, where applicable)
assumptions should reflect the expected future experience in respect of the contracts that
will be taken out.
They will be based on an analysis of the company’s recent experience. Ideally, this should
relate to the contract being priced, but if no such experience exists or the available data are
inadequate, then the experience under any similar contracts would be analysed.
If the company does not itself have adequate data, there may be industry-wide experience
that it could use.
The results of such analyses should be assessed to see if they have been affected by
special factors such as an adverse economic situation in the country.
An interesting example of such a special factor is that of mutual life companies demutualising. As
soon as rumours of the imminent demutualisation start to appear, with the possibility of windfalls
to policyholders, withdrawal rates decrease. The company may then suffer a ‘backlog’ of
withdrawals immediately after distributing any such windfall.

Question
Explain how a deterioration of the economic situation would affect discontinuance rates for the
following contracts:
(i) regular premium term assurance
(ii) single premium whole life assurance
(iii) regular premium endowment assurance.
Solution
The two things to consider in a scenario of economic deterioration are

 whether people can still afford to keep the policy going
 whether they would get any cash from surrendering the policy.
So looking at the three specific cases:
(i) Regular premium term assurance
The premiums are fairly low so there should be very few policyholders who now find themselves
unable to continue paying premiums.
There is almost certainly no surrender value available, so there is no incentive to surrender.
So we would not expect the economic downturn to affect discontinuance rates significantly for
this contract.
(ii) Single premium savings
There are no further premiums to pay, so affordability is not an issue.
On the other hand there will probably be a significant surrender value.
So we would expect some increase in surrender rates.
(iii) Regular premium endowment assurance
The premiums are probably a non-trivial portion of policyholders’ incomes. Some policyholders
will now be unable to pay these premiums.
There will probably be significant surrender values available for all policies greater than a couple
of years old.
So we would expect a big increase in discontinuance rates.
If the rates are to apply to a class of lives that is expected to have a different experience
from that to which the analysed data relates, then adjustments may need to be made. This
situation could arise due to a change in the benefits being offered or target market or
distribution channel.

Question
Suggest changes to benefits for a regular premium whole life contract that might lead to reduced
persistency.
Solution
The following would lead to reduced persistency:

 an increase in discontinuance terms (especially surrender values)
 a decrease in bonus rates (or a decrease in expected bonus rates).
And for unit-linked business:

 reduced fund performance (or a reduction in expected fund performance)
 an increase in charges
 removal of / variation in options or guarantees provided.
Question
Outline ways in which a change of distribution channel might affect persistency rates.
Solution
Each of the following can be associated with a change in distribution channel:

 A change in who initiates the sale can affect persistency rates (persistency rates are likely
to be higher if the client initiates the sale).
 The sales practice may be different: for example, where clients have been put under more
sales pressure to buy a policy, or to take a larger policy, then persistency rates are likely
to be lower than where sales have been more strictly based on need.
 Sales of policies made without proper information being gathered about the client’s
needs, can lead to inappropriate policies being sold through ignorance (though this is
again a reflection of poor sales practice).
 Different levels of financial sophistication of the client base can lead to different
perceptions of the value of a contract, and could lead to different persistency rates.
 Different target markets may encompass different levels of affluence and economic
prosperity. There may then be differences in withdrawals that arise through economic
hardship.
Persistency rates are significantly influenced by economic and commercial factors, which are
notoriously difficult to predict. Future persistency rates are therefore subject to considerable
uncertainty and so it is essential to explore the sensitivity of the company’s profits to variations in
the future persistency experience.

1.7 Margins
The assumptions derived in previous sections are estimates of the expected values for the
parameters.
Where a cashflow model is being used to price a life insurance contract, the risk to the
company from adverse future experience may be allowed for through:
 the risk element of the risk discount rate
 using a stochastic approach
 assessing what margins to apply to the expected values.
Refer to Chapter 15 for a reminder of how adverse future experience may be allowed for in
pricing a life insurance contract.
If a formula model is being used for pricing, the first two approaches above are not available
and the risk of adverse future experience would be allowed for by taking margins. However,
such a model does not help the actuary to quantify what these margins might be and they
must use judgement based on past experience.
The risk allowance is discussed further in the following section.
The important point from the above is that, by including a margin or margins somewhere in the
basis, the risk from adverse future experience is reduced. The basis actually chosen for pricing,
inclusive of margins, will fix the level of risk the company will be subject to once the product is
issued at that price.
Let’s consider each ‘method’ of the Core Reading in reverse order.
Under the third approach, all margins could be incorporated in the assumptions for each
individual parameter; and the risk discount rate would then (in theory) be the risk-free rate.
Hence, for example, our assumption for the future investment return might be written as:
i   i   m
where i  is the best estimate of the future annual investment return, m is the margin and i  is
the interest rate assumption in the pricing basis, inclusive of margin. Once the product is in issue,
the company risks making less profit than it anticipated if i (the actual rate of return achieved) is
lower than i  . So the larger the value of the margin m, the lower the probability that i  i  and
the lower the risk of making a particular loss.
Under the second approach, we would assume best estimates for the parameters and a risk-free
discount rate. At least one of these parameters would take a range of values from a probability
distribution, rather than a constant value, and so the pricing process would give a range of
possible charges or premiums. The company could choose which price from this range of
possibilities it would actually charge (eg the 60th percentile). The risk to the company will be that
the actual experience is in line with one of the more extreme outcomes of the distribution.
Finally, under the first approach we might have priced assuming best estimates for the individual
parameters, and included margins for risk by assuming a higher risk discount rate. The company
will now make less profit than it requires if i is sufficiently low that the return on capital is less
than the required rate of return.

These alternatives can lead to the same result (in effect, the same premium), but they are just
different ways of expressing the margin and the resulting risk. In each case the risk of loss occurs
once the margin is ‘used up’, whether that be the margin in the individual parameter assumption,
the percentile chosen, or the risk premium in the risk discount rate.
Although the primary worry is that of adverse experience compared with assumptions, the
parameter and model risks are really ‘two-sided’. For instance, looking at the issue of parameter
risk in the context of the mortality rate for pricing a term assurance contract:
 if the mortality is under-estimated the company is at risk of a substantial loss
 if the mortality is over-estimated the company is at risk of not selling the product.
However, here it is worse to under-estimate mortality than to over-estimate it. It is better not to
sell very profitable business than it is to sell loss-making business; if we are selling badly rated
business we may accumulate a lot of it before realising it is badly rated and hence loss making,
whereas if our rates were too high we could quickly adjust them. Here the actuary will need
margins to guard against the risk that the mortality is under-estimated (ie assume mortality rates
which are higher than best estimates).
The extent of the margins depends heavily on the purpose of the basis concerned. Here we are
concerned with pricing, so the margins will also be influenced by the need to have competitive
premiums. Nevertheless, the margins must reflect the risks involved, and hence the size of the
margins must crucially depend on:
 the degree of risk associated with each parameter used
 the financial significance of the risk from each parameter.
All the things we have discussed in earlier chapters about risk should therefore be brought to a
consideration of the margins required for pricing (or, in general, for other purposes for which a
basis is required). Remember that the same principles will apply whether we are allowing for the
margins through the risk premium in the risk discount rate (which would be the normal approach
when pricing), or through adjustment to the individual parameter assumptions. We describe the
assessment of the risk discount rate in detail in the next section.
1.8 Discount rate
Risk discount rate

We now turn to the derivation of a suitable risk discount rate to use in our pricing models. As we
discussed in an earlier chapter, a key aspect of the risk discount rate will be the return required by
the shareholders on the capital they invest in the insurance company.
It is reasonable to suppose that the owners of a life insurance company decide where to
invest by comparing the returns offered by different companies, relative to the risks which
are run, and are able to move their capital from one company to another if they wish.
An investor will demand a higher expected rate of return from a risky investment than from
a safe investment to compensate the investor for the risks of default, commercial failure and
so on. This is often expressed by saying that the investor requires a ‘risk premium’ over
and above the expected returns on safer investments.

It is useful to borrow from economic theory the concept of the ‘risk-free’ asset, namely an
asset which offers a certain return, free from all risk of default.
Investing in a life insurance company is certainly not the same as investing in the
economist’s idealised risk-free asset. Investors will demand an expected rate of return
equal to the risk-free rate plus a risk premium. Such a rate of return is called a ‘risk
discount rate’. The question is: What risk premium is appropriate to compensate for the
risks of investing in a life insurance company?
The capital asset pricing model

The capital asset pricing model (CAPM) has been widely used by stock market investors to
answer exactly this sort of question. The idea behind the CAPM is that a well-diversified
portfolio of shares cancels out the risks of investing in individual shares and leaves only
the unavoidable risks of investing in the stock exchange.
A suitable proxy can be chosen to represent risk-free assets, such as short-term deposits
issued by a stable government, and the average risk premium that the diversified portfolio
of shares has yielded over the risk-free rate, over a period of time, is estimated.
The short-term deposits here are just one example of an asset which is almost risk free.
Alternatives are very short government bonds, or long-dated inflation-proof government bonds.
It will be impossible to find any assets that are perfectly risk free.
The above paragraph of Core Reading is explaining how we would quantify the risk of investing in
equities. That risk is measured as the difference in return comparing ‘equities’ in a general sense
with some risk-free asset. Take a well-diversified portfolio of shares (eg a suitable index) to
represent general equities, and short-term government stock to represent the risk-free asset.
Compare the returns from these two over a period of time to iron out any random fluctuations.
The question can then be asked as to how risky a company’s shares are compared with the
diversified portfolio. The result of the CAPM is that the proper risk premium for any share is
in proportion to its beta.
At this point, a brief reiteration of some CAPM theory is perhaps called for.
If the market and the investors satisfy certain conditions, such as a perfect market with perfect
information, the CAPM will explain, amongst other things, the relationship between risk and
return.
It is possible to derive the following formula which expresses the expected return on any asset in
terms of the market return. This is given by:
E i  r   Em  r   i
where: Ei is the expected return on asset i (eg our company’s shares)
r is the return on a risk-free asset
Em is the expected market return (eg the stock market)
i is the ‘beta factor’ of asset i (defined as the covariance of ri and rm divided by

the variance of rm).

So this equation specifies the relationship between the risk premium for an asset (the risk
premium is {Ei – r}) as proportional to the market risk premium {Em – r}. The factor of
proportionality is i .
The beta factor i can be thought of as a measure of the riskiness of the asset relative to that of
the market. A value greater than 1 implies, when the market is rising, that the asset’s value will
increase more than the market average; and conversely, when the market is falling, that it will
decrease in value more than the market average.
This beta factor can be estimated by analysing the historic returns on the asset in question in
conjunction with market returns. However, the beta that is correct ‘now’ or in the future will not
necessarily be equal to that of some recent period.
This process identifies the systematic risk, or market risk, of the asset. It takes no account of the
specific risk – eg in the case of equities, the risk specific to the issuing company which can be
eliminated by sufficient diversification.
It might seem almost counter-intuitive that we are taking account of just the systematic risk, not
the specific risk. Why is the (significant) risk of investing in a specific company not allowed for?
The reason is that the life company’s shares are assumed to be just a small fraction of the
investor’s portfolio – the CAPM does not reward specific risk. So the CAPM will quantify just the
systematic risk.
The shareholders’ required return on capital is also a function of the availability of capital. The
harder it is for the company to raise capital, the greater will be the returns the company will have
to offer in order to raise the amount required, and the greater the returns will have to be to keep
the shareholders happy.
The CAPM is just one example of how the market might assess the shares of a company.
The point to note is that it is not up to the actuary alone to decide what an appropriate rate
of return is for shareholders. The actuary will need to make some assumptions, but the
market is the final judge.
Deciding upon a risk discount rate

So what does an actuary actually do in order to determine the risk discount rate to use in a
particular pricing model for a particular product? So far we have the CAPM, which can give us the
value of Ei : the rate of return that shareholders expect from investing in our company to
compensate for the risk involved. However we cannot simply use this overall rate as the risk
discount rate in our pricing models. The following Core Reading explains why not.
Not all projects are equally risky. The life company should view itself as an investor like any
other when it considers the riskiness of a new product, as in the long run the profits
emerging from the whole company are the profits emerging from the products that it sells.
A change in the mix of business, for example away from old and ‘safe’ contracts towards
new and innovative contracts, would change the market’s evaluation of the company’s
riskiness.

So if we propose to launch a new product with innovative design features, this will change the
overall market risk premium demanded by the company’s shareholders, and the only way to
satisfy this demand is for this product to generate a higher return on capital. In other words, the
risk discount rate used needs to be higher, and basically the higher the risk, the higher the risk
discount rate we should use.
The following are among the features that can make a product design riskier, viewed as an
investment:
 lack of historical data
 high guarantees
 policyholder options
 overhead costs
 complexity of design
 untested market.
Question
List the product features that might lead to a lower risk discount rate.
Solution
Basically, the opposite reasons to those mentioned as risky:

 benefits based on contingencies about which there is a lot of reliable data
 no (or not very onerous) guarantees
 no policyholder options
 low overhead costs (since this reduces the risk posed by low business levels)
 simplicity of design (and so of understanding, communicating, pricing, and administering)
 using a tried and tested market (reliable source of business)
 the ability to vary charges and/or review premiums.
So one way to derive the risk discount rate would simply be to take the ‘risk-free’ rate (which we
can assess by looking at the yields on short-term government bonds) plus an allowance for the
riskiness of the product being sold.
It is not easy to assess these risks, and it is even harder to say what effect they should have
on the risk discount rate.
Let’s have a look at what we might try and do, in theory. We are going to ignore the CAPM result
for the moment – we are instead trying to assess the risk discount rate (for an individual product)
from the ‘bottom-up’.

We can approach this by thinking of the rate of return on a product as a random variable R. Given
a certain variance Var[R] , R should be such that the probability of it falling below some reference
level l is small (ie the probability is below some desired risk level k). Such l could be equal to, or
related to, a reference point such as the risk-free return.
For instance, consider a market where short-term government bond yields are 6% pa, and the
shareholders stipulate a return equivalent to 6% pa risk free. Then one possible reasonable
interpretation of the above would be that we require that R be greater than 6% pa with a
probability of 0.95.
Having decided what sort of R is required, we can engineer the product to give a return that
satisfies the desired criteria by choosing ‘arbitrarily’ an appropriate risk discount rate r* to use in
the profit testing. This risk discount rate r* must be set equal to E[R ] . Then, by definition, if the
product’s profitability is satisfactory when measured using that discount rate r*, the rate of return
r of the product will satisfy our criteria.
For instance, expanding on the above example: we require a probability of 0.95 that R exceed
6% pa. Now assume we know that Var[R] is such that (assuming for reasonable convenience a
normal distribution) Pr  R  6%   0.95 only if E[R ] is equal to 9.3% pa. We therefore use 9.3% pa
as the risk discount rate.
Assessing Var (R) – referred to below as the statistical risk – is not easy. For example, it will need
to reflect parameter uncertainty as well as divergence due to random fluctuations.
However, the level of statistical risk could be assessed using one of the following methods:
 in some situations analytically, by considering the variances of the individual

parameter values used
Thus quantify Var[R] by considering the variance of the important parameters – such as
interest, mortality, persistency, and expense inflation – given the way in which they
influence the projected profit.
 by using sensitivity analyses with deterministically assessed variations in the

parameter values
Sensitivity analysis tells us to what changes in our assumptions (eg increase or decrease)
our profits are most susceptible. If this was, say, an increase, the idea is then to find the
increased parameter value that we assess as corresponding to a 5% probability level.
(This is a subjective assessment.) So suppose we judge that 140% of our best estimate
mortality rate assumption would correspond to this 5% probability level. We then see
what happens to the return on capital when we use this mortality rate assumption: is it
greater than the minimum acceptable (the 6% pa of the example above)? If not, then the
premium rate or charges need to be increased.

 by using stochastic models for some, or all, of the parameter values and simulation
Here we would vary the important parameter values stochastically according to their
assumed probability distribution, re-computing the rate of return for each new scenario.
Doing this, say, 1,000 times will then give a very good idea of the variance of the rate of
return. This method is the numerical equivalent of the first (analytical) approach.
 by comparison with any available market data.
In theory, a separate risk discount rate should be applied to each separate component of the
cashflows, as we mentioned when we discussed product pricing in Chapter 15. Technically, we
should be applying these separate discount rates to the different profit flows, as we do in the
following example.
Example
Suppose we wanted to adopt this approach for the three parameters mortality, persistency rates
and investment return. If we have a good bedrock of historical data from our portfolio, and the
current portfolio is reasonably sized (say at least 50,000 policies), then we have a good idea of the
expected mortality experience and there should be a small variance in mortality rates. So we
would use a relatively low risk discount rate for the mortality profit.
Our historical persistency rates have been analysed equally well and, compared with mortality,
we notice more variation of persistency rates year by year; also we know that they can be quite
sensitive to economic swings. So we would use a higher risk discount rate for valuing the
persistency profit.
Examining recent investment returns shows great variability. We therefore would use an even
higher discount rate for valuing the investment profit.
In practice applying a separate discount rate to each component of the cashflow would usually
not be done. One reason is from the point of view of simplicity. Also, it would be tricky, both in
terms of time and data requirements, to analyse accurately the variability of the different
cashflow components.
Finally we need to consider how we can bring together the two approaches – that of using the
shareholders required return (eg using CAPM) and the statistical approach described above.
If we consider that the overall required return Ei is a true reflection of the riskiness of investing in
the company as a whole, then projects that are less risky than average will require a lower risk
discount rate than Ei , while projects that are more risky than average will require a higher risk
discount rate than Ei . We can assess this relative riskiness using the statistical approaches
described above.
Of course, as the projects develop and lead to actual profits (or losses) then the nature of the
riskiness of the company as a whole could change, thereby changing Ei . This is not inconsistent
with our methodology, as the change in Ei will also reflect the changing mix of risk discount rates
that we have used for pricing the individual projects themselves.

Also, if Ei has not changed much since the last time a product was priced (or repriced), and the
product concerned has not changed its risk profile fundamentally, then it would be logical to use
the same risk discount rate as was used before for pricing (or repricing) the product. So, it should
be possible to take the most recently used risk discount rate for a product, and then adjust it as
necessary to reflect changes to the statistical riskiness of the product, to the overall required
market return Ei , or to both.
Question
Describe how a risk discount rate could be determined that reflects the expected level of
statistical risk, for a given new product, including the use of the following:
 deterministic sensitivity analysis
 stochastic simulation.
It can be assumed that the shareholder’s required return on capital is known.

Solution
The following steps could be performed.
(1) Calculate some trial premiums for the product, based on best estimate assumptions of
future experience, and using the shareholders’ required return as the risk discount rate.
(2) Calculate the return on capital generated by these premiums, on the basis of a range of
different deterministic scenarios for the future experience. The scenarios used should
include adverse outcomes that have a feasible likelihood of occurring, for example with a
5% probability. (The actuary would need to devise these scenarios using his or her
judgement, based on historical experience, and on analysis of all the factors likely to
affect the experience in the future.) Should the return on capital obtained fall below
some required minimum level (such as the risk-free rate of return), then the premium
should be increased until this criterion is satisfied.
Alternatively, should the return on capital be consistently much in excess of the required
minimum level, then there may be justification in reducing the premium. (The company
would have to consider whether the product concerned represented a ‘lower than typical’
level of riskiness for the company in order to justify this – we have to keep in mind that
the overall required return reflects other factors than just statistical risk (eg the
availability of capital)).
Use these new premium levels in the next step.
(3) Repeat step (2), only this time modelling the variables stochastically. Calculate the
frequency (out of many simulations) with which the return on capital falls below the
required minimum rate. If this frequency is too high (say more than 5%), then the
premiums could be further raised until a 5% frequency is obtained. (Or premiums could
be reduced if the frequency is too low). Use these new premium levels in the next step.
(4) Calculate the return on capital obtained from any new level of premium produced from
steps (2) or (3), using best estimate experience assumptions. This new return on capital
can then be used as the new risk discount rate. This new risk discount rate now takes
account of the levels of statistical risk for the product.
Steps (2) and (3) could equally be performed in reverse order.
If we were repricing an existing contract it would be sensible to start with the current risk discount
rate at stage (1). Account would also need to be taken of any changes in the shareholders’
required rate of return since the repricing, as well as to the statistical elements.
It is unlikely that this level of detail would be required in an exam solution. The purpose of this
question is to show you how some of the bits to do with assessing the risk discount rate could be
brought together to form a methodology that could be carried out in practice.
Also remember that it is not an exact science, and there are other approaches to devising the risk
discount rate that would be valid in practice.

In reality, the assessment of the risk, and the conversion of that into a risk discount rate, is a very
inexact science. The most important points to remember are that:
 The risk discount rate must be higher than the risk-free rate, hence changes in market
rates of interest should cause risk discount rates to change.
 The margin between the risk-free rate and the risk discount rate should attempt to reflect
all sources of risk in the product.
 The risk discount rates used for pricing different products should reflect the relative risk
of those products.
 The overall return earned on the whole capital of the company, generated from its entire
business activities, needs to meet the required rate of return (cost of capital).
 The risk discount rate is simply a number that is used as a criterion in profit testing. It is
set to ensure that the rate of return from the product is satisfactory, given the inherent
variability of the return and the required return on capital.
Profit criteria
It is necessary to decide on the profit criteria that need to be met by the price charged for a
product, given all of the preceding assumptions. These have already been discussed in
Chapter 15, but you should bear in mind that the chosen profit criterion is one of the assumptions
that defines a particular pricing basis.
Market-consistent valuation
An alternative approach is to perform a market-consistent valuation of the cashflows. The
‘risk neutral’ approach to such a valuation effectively involves the use of risk-free interest
rates for the discount rates, which are generally modelled as being term dependent.
It is then likely that some form of margin would be included in respect of other parameters
used in the pricing, such as expenses, persistency and mortality, to allow for the inherent
risks in their estimation.
For example, we will see in Chapter 20 how to calculate a risk margin using the cost of capital
approach.

2 Consistency
One of the guiding principles in creating a valid basis is the principle of consistency. The
assumptions need to be considered in their totality, not in isolation. Consistency means ensuring
that we make realistic allowance for the way that variables behave together, where correlations
exist between them: sometimes the relationship between two parameters will be more important
than the absolute values of those parameters. (This is similar to the idea of including dynamic
links between variables in our models, as discussed in Chapter 14.) Particular concerns will be the
following.
Investment return and inflation

The long-term relationship between the rate of expense inflation and the rates of return on
different types of asset must be valid. (However sometimes we do not need to make assumptions
about future investment conditions – when there is no reinvestment risk.)
Tax
If the company is taxed on ‘I–E’ type taxation then the investment return assumptions and the
expense assumptions should both reflect the same tax treatment.
In profit testing the product, the profits should be measured net of tax.
New business and expenses

If developing a new product, then the development expenses will need to be recouped, and the
product should be designed to make an appropriate contribution to the company’s fixed
expenses. The amount of loading for these purposes should be such that, given the anticipated
volume of sales of the product over its expected lifetime, the total loadings received will at least
recoup the development costs and pay a fair share towards the company’s overheads. Provided
the new business assumptions do not turn out to have been too optimistic, the product should
make additional contribution to profit. The new business assumption must itself reflect the
expected competitive position of the product in the market over the period.
Question
Explain how there can be a circularity problem in the above.
Solution
The future volume is a function of competition, which might itself be a function of price. The
price is a function of the expense loading, which is itself a function of future volume. So future
volume is a function of future volume.
(In order to get round this we could assume some bands of price, say, across which future volume
is insensitive, so as to home in on a final answer.)

Investment return and bonus loading

These must be consistent in situations where bonus needs to be loaded for specifically.
Question
A life insurance company sells with-profits business, using the additions to benefits method for
profit distribution. Bonus rates have been stable for several years.
The company wishes to issue a new series of with-profits products in the future. The premium
rates have been calculated on the basis of best estimate assumptions for future experience,
except for the future bonus rates which have been assumed to be 50% of the rates currently
being declared by the company on its existing business.
Explain the main effects this might have on the company.
Solution
The new premium rates might be cheaper than for the products it has been selling up to now, so
volumes of business might increase. However, this depends very much on other factors of which
we are not aware, in particular how competitive the company is in terms of its maturity payouts
(which should be a much bigger driver of sales volume than price).
What seems more probable is that the bonus earning capacity of the new business will be lower
than that of the existing business, so that the company will only be able to afford gradually
decreasing bonus rates on the whole fund as more and more new policies are issued over time.
This will also produce inequity between old and new generations of policyholders (which would
persist over many years), assuming the company decides to pay an average rate to the two
classes of business even though their contribution to profit might be quite different.
So either the company has to live with this inequity over a substantial period, or it has to declare
separate bonus rates to old and new generations, which is messy and contrary to the usual
principles of broad equity of distribution adopted for this distribution method.
Unless the change in future bonus rates is intentional, then the above mess is the result of making
inconsistent assumptions in the premium basis.
Persistency rates and investment return

Persistency rates can be affected by the general economic climate, which will be one of the major
factors behind the investment return assumption. Persistency may also be affected by bonus
levels, discontinuance terms and renewal commission levels.

Other products
The basis should normally be consistent with that of related products. In fact the basis of some
closely related product might form the start point for any new product’s basis. For example the
basis for a new regular premium decreasing term assurance might take as a start point the
company’s current regular premium level term assurance product.
Question
Explain what might happen if the company doesn’t use a consistent basis for the above product.
Solution
The company will end up selling two fairly similar products, at the same time, of which one must
be expensive relative to the other. If the market is competitive (it probably is for term assurance),
this might affect consumers’ buying and discontinuance behaviour:
 more of the cheaper product might be sold relative to the other
 holders of the more expensive product may lapse their existing policies and take out new
cheaper ones, to the disadvantage of the company
 policyholders could feel aggrieved at the perceived inequity, with resulting bad publicity
for the company, and ensuing marketing risk
 the company’s profits will become more sensitive to the type of business sold (ie its
profits will be at an increased risk from changes in the mix of new business).

Chapter 17 Summary
In constructing a basis the actuary will consider the future experience expected for the
contract concerned, and modify it as appropriate given the purpose of the basis. The guiding
principle in constructing any basis is consistency. This means that the relationships between
the assumptions in a basis are realistic, reflecting the correlations that exist between
variables.
The assumptions needed will involve:
Mortality
An assumption is required for both the base mortality and the mortality trend.
The actuary will consider the company’s recent experience of the contract, or related
contracts. This will allow appropriate adjustment to be made to standard tables. Industry
data and reinsurers’ data may be useful, especially if the company has little relevant data.
Adjustments should allow for differences between the lives for which the assumptions are
being set and the lives from which the data is obtained. There may be differences in target
market, distribution channel or underwriting approach.
An allowance should also be made for expected changes in mortality rates over time using a
combination of:
 expectation approaches
 extrapolation approaches
 explanatory approaches.
In some countries observed mortality improvements can vary significantly by year of birth or
‘cohort’.
Investment return
This will be set bearing in mind the significance of reserves for the contract, the extent of
investment guarantees, the importance of reinvestment and the intended asset mix for the
contract. The impact of taxation must be considered.
Expenses and commission

Expenses will be loaded for in accordance with the results of the most recent expense
investigation. Some expense modelling work might be entailed for new ventures. In
deciding on the split of expenses between ‘per policy’ and ‘per £ premium’, some cross-
subsidy might be necessary for competitive reasons.
Expense inflation
This will depend on expected future earnings inflation. It must be consistent with the
investment return assumption.

Persistency
The assumption will take as a start point the most recent investigation for that contract or
related contracts, or, in the absence of suitable data, industry statistics. Some adjustment
might be necessary. The commercial and economic environment is important.
Margins
Margins will be necessary to guard against adverse future experience. However competitive
pressure forbids too much prudence in a pricing basis. Risk may be allowed for through the
risk discount rate, stochastic modelling or margins in the parameter assumptions.
Risk discount rate

Those who provide capital to the life company expect a return on their capital of:
 that which they could obtain from a risk-free asset
 plus a ‘risk premium’ to compensate for the volatility of return.
This return on capital demanded by the investors is the risk discount rate. The actuary will
need to price products to meet that rate if the company as a whole is to earn that rate.
The risk discount rate can be determined by quantifying the risk premium appropriate for
the company. One way of doing this is by using the capital asset pricing model. The market
availability of capital should also be taken into account.
The market’s view of the risk discount rate is not necessarily the most apt for any given
product or project, as different products (for example) will have different levels of riskiness.
Factors that affect riskiness include:
 lack of historical data
 high guarantees
 policyholder options
 overhead costs
 complexity of design
 untested market.
Statistical risk may be assessed analytically, using sensitivity analysis, or using stochastic
modelling.
Profit criteria
Premiums are set to satisfy specific target measures, for example a function of net present
value (as discussed in Chapter 15). The specified criteria are part of the pricing basis.

If a market-consistent approach is used then the expected investment return can be set as the
risk-free rate (regardless of the actual assets held). However, the investment return volatility
and correlation assumptions depend on the actual assets held.
The discount rate would also be set as the risk-free rate.
A margin is likely to be included in the other parameter values to allow for the risk in their
estimation.



17.1 The adequacy of premium rates is being reviewed for a range of contracts.
Exam style
Discuss, for each of the following contracts:
 the extent to which margins against adverse contingencies are required
 how and where in the basis the margins might be applied, with justification.
(a) Twenty-year without-profits term assurance.
(b) Immediate annuity. [8]
17.2 A life company issues group life policies, which are sold as single premium term assurance
Exam style
policies, to employees of a company.
Outline the factors that the life company would take into account in determining the basis of the
mortality element of the premium rates for this product. [3]
17.3 Outline the general principles that should be taken into account when setting the mortality
Exam style
pricing assumptions for a new life assurance product. [6]
17.4 It has been suggested that a life company should use industry statistics to estimate persistency
Exam style
rates appropriate for pricing future product launches rather than analysing its own experience.
(i) Explain why persistency rates are important in product pricing. [3]
(ii) State reasons why the suggestion above may be inappropriate. [4]
[Total 7]
17.5 The role of Group Finance Actuary has just been established in a large international life insurance
company. One of their responsibilities is to ensure a sound and consistent approach to product
Exam style
profitability. Hitherto, different subsidiaries of the company have priced products to give an
internal rate of return equal to local government long bond yields plus 2% pa.
(i) Discuss the advantages and disadvantages of the existing approach. [5]
(ii) Describe what other measures of profitability the new Group Finance Actuary might
introduce, and their advantages and disadvantages. [7]
(iii) Describe how, by use of the capital asset pricing model, the Actuary might select a
suitable risk discount rate to use in profitability measurement. [6]
(iv) Describe how the measures discussed in part (ii) and the selection of the risk discount rate
in part (iii) could be used in a way that would ensure consistency of profit measurement
between the various different subsidiaries. [3]
[Total 21]



17.1 (a) Twenty-year without-profits term assurance
A contingency margin is desirable. [½]
It may be incorporated in a higher mortality assumption than expected. Also, depending on the
market, some allowance for future mortality deterioration due to AIDS might be appropriate. [1]
Maybe also assume high lapses early on when the earned asset share is negative, and low lapses
later on to guard against the effect of selective withdrawal (ie only the fit lapsing). [1]
There is little point assuming a low interest rate for regular premium term assurance. It would
make little difference to the premium rate. If the contract were single premium (unlikely) a low
interest rate would guard against reinvestment risk although even here there will normally be
little if any reinvestment. [1]
The necessity of implicit margins depends on whether or not there is an explicit profit margin: if
there is, the need for and scope for further margins is reduced. [½]
Some or all of these margins could alternatively be incorporated into the risk discount rate, if
used (by making it higher) … [1]
… or by modelling key variable(s), eg mortality, stochastically and choosing an appropriately

prudent percentile from the range of outcomes. [1]
In practice, the scope for margins will be limited by competition. [½]
(b) Immediate annuity
A contingency margin is desirable. [½]
A lower mortality assumption than expected is one method. This may guard from greater than
expected improvements in mortality or a change in that part of the population taking out
annuities. [1]
A lower interest rate than currently available will guard from reinvestment risk to the extent that
there is such risk. (Reinvestment risk will only be significant for an increasing annuity.) However,
since the business can often be fairly closely matched, the size of the interest rate margin will not
normally need to be large. [1]
A further risk is an increase in tax rate. A suitable margin here is to assume a higher tax rate or a
lower interest rate. [½]
Some or all of these margins could alternatively be incorporated into the risk discount rate or by
the use of a stochastic approach, as for the term assurance. [½]
As with the term assurance, the need for implicit margins depends on any explicit profit criterion
and the scope for margins is limited by competition. [½]
[Maximum 8]

17.2 Most group life policies are short term, often one year, so a discussion of mortality trends is
unlikely to be appropriate in this case.
The main factor will be the company’s view of mortality over the future period. This in turn is
based on:
 occupation – is the scheme ‘managerial’, ‘white collar’ or ‘blue collar’, or a mixture? [½]
 location – some parts of the country exhibit higher overall mortality than others [½]
 level of underwriting – especially free cover limit [½]
 scheme experience – if credible [½]
 whether take up is voluntary or compulsory – compulsory schemes give less scope for
anti-selection [½]
 any need for margins over best estimates – this is likely to be greater the longer the
guarantee period. [½]
[Total 3]
17.3 The values assigned to the parameters should reflect the expected future experience of the lives
who will take out the contract. [½]
In particular consideration should be given to:

 target market [½]
 distribution channel [½]
 underwriting controls. [½]
Separate assumptions will be required for the base rate of mortality … [½]
… and the mortality trend. [½]
The base rate of mortality should allow for the expected movement in mortality between the
time of the last experience investigation and the point in time at which the new contract will be
sold. [½]
The assumptions should be based on an adjustment to rates from a standard table, if this is
available. [½]
The adjustment could be derived by analysing the company’s experience of a similar class of
business. [½]
An assumption should then be made as to how these base mortality rates are expected to change
over time. [½]
The data should relate to an appropriate number of years, such that the volume of data is
adequate … [½]
… but excessive heterogeneity due to trends over time is not introduced. [½]

Split the data into relevant homogeneous groups, … [½]
… subject to having a credible level of data in each cell. [½]
If insufficient data is available then industry or reinsurers’ data should be used instead. [½]
Population mortality might be particularly useful for projecting trends, … [½]
… although it is less relevant as it covers lives that don’t have insurance. [½]
Adjustments may be required to allow for changing or differing target market or underwriting
procedures. [½]
If there is great uncertainty surrounding the correct rates to use then an allowance can be made,
eg through a higher risk discount rate than normally used or a significant margin in the
assumptions used. [1]
[Maximum 6]
17.4 (i) Why persistency rates are important in pricing
If a life company pays a benefit upon surrender that is higher than asset share, the company will
make a loss on that individual policy. The same will happen on policies that pay no surrender
benefit when asset shares are negative. [1]
Similarly, paying a benefit which is less than asset share will give rise to a profit. [½]
Persistency rates are also very important for projecting future in-force volumes. For example,
lower persistency rates would mean that less profit would be expected from the portfolio later in
the policy term, as fewer policies would still be in force. [1]
Fewer policies, as a result of withdrawals, also increases each policy’s share of the company’s
overhead expenses, increasing the required per-policy expense assumptions. [1]
When a company is pricing its products, it will want to quantify the profit that it expects to make
as accurately as possible. Hence the company will want to allow for persistency experience in its
models, and to do so as accurately as possible. [½]
Assumptions affect price, and inaccurate assumptions can therefore adversely affect
competitiveness and hence sales volumes. [½]
[Maximum 3]

(ii) Why the suggestion might be inappropriate
The company will not have the same experience as the industry average, due to differences in
such things as: [½]
 sales method [½]
 products or product features [½]
 surrender values paid [½]
 quality of service [½]
 past investment performance. [½]
Also industry statistics may be wrong, unavailable, out of date or not presented in sufficient
detail. [2]
The effect of heterogeneity in the industry data can give rise to spurious trends in experience
observed over time. [½]
[Total 4]
17.5 (i) Existing approach
It is a simple measure. [½]
Compatible with shareholders saying ‘we want a return of at least x%’. [½]
Easily comparable with other forms of investment. [½]
The comparison basis of long bonds is consistent by term with the investment that the
subsidiaries are making in financing new life insurance contracts. [½]
It is a reasonable way of relating to local investment conditions rather than imposing any absolute
figure (eg ‘IRR must be at least 10%’). [½]
The 2% margin should represent the risks of writing the relevant business. So a constant figure is
not good because: [½]
 different types of business have different risk characteristics [½]
 different subsidiaries will have different risk characteristics. [½]
The IRR has several disadvantages:

 It might not be unique. [½]
 It might not exist. [½]
 It is difficult to relate to other measures (eg premium income). [½]
[Maximum 5]

(ii) Other measures of profitability
Profit signature
This contains all the information about future profits. But difficult to use in that it is not a simple
single-figure criterion. Difficult to present, or to use in comparing different products. [1]
Net present value
= profit signature discounted at the risk discount rate. [½]
A good criterion: given a choice between two different cashflows (ie two different products),
economic theory dictates that the investor should choose that with the higher net present
value. [1]
However this assumes:

 a perfectly free and efficient capital market [½]
 that the NPV is calculated using a discount rate which correctly reflects the inherent
riskiness of the product. [½]
Also a problem with the law of diminishing returns (otherwise if a product has NPV>0 the
company could write infinite amounts of it and become infinitely rich). [½]
It is not a very simple measure to present to non-technical people. [½]
By itself it tells us very little. [½]
Although if NPV>0 then we know that the return is greater than the risk discount rate. [½]
Need to use it in conjunction with some measure, such as:

 income (eg total discounted future premiums) [½]
 ‘capital outlay’ (eg initial commission). [½]
= policy duration at which profits to that date have present value zero. [½]
It is a useful means to compare products if capital is a particular problem. [½]
It will often not ‘agree’ with the net present value as a means of deciding between two different
cashflows because the discounted payback period ignores cashflows subsequent to the
discounted payback period itself. [1]
General comments on all of these methods
All of the above measures (including IRR) ignore competition – thus the company could end up
with a product with a highly attractive NPV but which it is unable to sell. [½]

All of the single-figure measures involve, by definition, substantial loss of information compared
with the information contained in the original profit signature. [½]
[Maximum 7]
(iii) Selecting a risk discount rate
The capital asset pricing model result which we need is:
E i  r   EM  r   i [1]
where:
Ei is the expected return on asset i (eg our company’s shares)
r is the return on a risk-free asset

EM is the expected market return (eg the stock market)
 i is the ‘beta factor’ of asset i which is a measure of the riskiness of the asset compared
with the market (>1 is more risky). [2]
Those 3 marks are also available for a verbal summary instead.
Start by identifying some risk-free asset – or an approximation to such – possibly long government
bonds. [½]
Need to identify the risk margin on equities, ie the market risk EM  r  . Do this by studying the
relative return of the equity market and the risk-free asset over a reasonable past period. [1]
Also use historic data to derive the company’s beta factor. [½]
Use this as the best guess for what the beta factor is now / will be over the next few years,
although it may not remain constant. [½]
Plug into the equation to give the required result, Ei . [½]
This will need further adjustment to reflect the specific (different) riskiness of any particular
product. [1]
If using the CAPM in a situation where the company does not have shares quoted on the stock
market, then use the shares of some similar company (or companies). [½]
[Maximum 6]
(iv) Ensuring consistency
Use the same method selected as per part (ii) for every subsidiary. [1]
If using the net present value then measure it in the same way (eg as a percentage of initial
commission) for each subsidiary. [½]
The method will make some reference to a risk discount rate. Better consistency is achieved by
using appropriate risk discount rates for different subsidiaries. [½]

Use the CAPM in each territory to give a risk premium appropriate to each territory. [½]
The risk-free assets used should all be of the same type (eg always use government index-linked
bonds). [½]
[Total 3]





using the course.

Setting assumptions (2)

Syllabus objectives
5.1.2 Describe the principles of setting assumptions for determining liabilities.

5.1.3 Explain why the assumptions used for supervisory reserves may be different
from those used in pricing.
5.1.4 Outline the principles of setting assumptions for determining embedded
value.
 calculating embedded value.

0 Introduction
Chapter 17 looked at the assumptions required for pricing purposes.
In Section 1 of this chapter we look at how the choice of a basis may differ for different purposes
including published accounts, supervisory reserves, internal management accounts, and
embedded values.
In Section 2 we look at how an embedded value can be used to place a value on an insurer’s
existing business and the basis that may be used to do this.
In Section 3 we look at the importance of consistency for the choice of bases described in this
chapter.

1 Valuing life insurance contracts: liabilities
1.1 Overview
Broadly, liabilities can be determined as the present value of future claims plus expenses
(including commission) plus taxes (if appropriate), less premiums. The greater the margins
that are employed in the assumptions, the more prudent the calculation. It should be noted
that, all else being equal, decreasing the discount rate increases the degree of prudence for
a positive reserve (the opposite being the case for a negative reserve, where this is
permitted).
Published accounts
The assumptions used to determine the liabilities to be shown in an insurer’s published
accounts should be consistent with the legislation and accounting principles governing the
preparation of those accounts in the country concerned. Matters to be considered include
whether:
 the accounts are to be prepared on a going concern basis or a break-up basis
 the accounts are required to show a true and fair view
 reserves are required to be assessed as best estimates or on some other basis, and
precisely how the terms used are to be interpreted.
Supervisory reserves
If separate accounts are required as part of the process of supervision of solvency, the
rules governing the preparation of those separate accounts may or may not be the same as
those that apply to the published accounts. They may, for example, need to be prepared on
a going concern basis or on a break-up basis. Reference should be made to those rules
and any guidance that may have been issued as to their interpretation.
A going concern basis assumes that the company continues to issue new business into the future.
A break-up basis will assume new business ceases, either immediately or at some point after the
valuation date. Break-up can take a number of forms, eg:
 existing business continues to be managed by the company as a closed fund
 the liabilities (in-force policies) are transferred to another insurance company who will
then administer them (and meet all future claims) until run-off.
If the valuation is for supervisory purposes (ie demonstration of solvency) then there will almost
certainly be some constraints on the assumption decisions. These might be specific constraints
(eg the valuation interest rate must equal the premium basis interest rate) or more general
(eg use any basis, subject to the final result not being lower than it would have been under
Ministerial basis X).
In some jurisdictions, the regulations require that the assumptions must be prudent, and this is
the approach we will follow in this chapter. However, many jurisdictions have moved to a
market-consistent approach, which we will consider in Chapter 20.

Internal management accounts

The principles to be followed in preparing internal management accounts are matters to be
discussed and agreed within the insurer. The aim is likely to be to produce expected values
of the future experience, based on realistic assumptions.
So these will be examples of ‘best estimate’ reserves, which we discuss in Section 1.2, below.
Embedded values
Embedded values were introduced in Chapter 15. An embedded value may need to be calculated
for published accounts or internal management purposes. We will discuss embedded values
further in Section 2.
1.2 Best estimate reserves

This method will generally be used when management wishes to have a best estimate of the
company’s financial performance. For example, this might arise in circumstances where
the insurer is to be sold or where directors wish to reward key staff for their specific
contributions to the overall growth of the company.
The reward must be related to a realistic measure of the profit that may be attributed to the
efforts of individual staff, and realistic profit measurements require policy reserves to be valued
on a best estimate basis.
Here assumptions will be derived based on the actuary’s best estimate of future realistic
experience, without the incorporation of any prudential margins. As will be noted in
Chapter 20, the basis derived may be closer to that used for new business pricing.
Question
Discuss whether the effect of underwriting should be allowed for in the basis for best estimate
reserves.

Solution
For the mortality pricing assumption, we are estimating the future mortality of a life at the start of
an insurance policy, when he or she will have just been accepted through the underwriting
process. The life will be initially ‘select’, experiencing lower than ‘ultimate’ mortality for the first
few policy years.
When we calculate reserves, we are dealing with policies that are already in force. In theory,
there will be some policyholders who took out their policies very recently, who will still be in the
select mortality period and who will therefore be subject to some reduction in mortality, relative
to ultimate, early on. However, for simplicity (and provided it errs on the side of prudence) this
complication would normally be ignored and all lives would be treated as ‘ultimate’.
The mortality of these ‘ultimate’ lives is still somewhat different from that which would result if no
underwriting had initially been applied. The main reason is that not all the impairments that are
initially screened out can be developed by initially healthy policyholders over time (eg congenital
diseases).
See Subject CM1 or CT5 for a fuller explanation of select and ultimate mortality.
All experience items will be allowed for explicitly (expenses, persistency, investment return,
mortality). There would be no elimination of negative reserves.
A cashflow method would estimate, possibly for individual policies or possibly with some
grouping, the amounts that need to be held now to meet future claims (including options),
expense and tax outgo against the expected premium and investment revenue inflow.
Alternatively, for non-linked business, a gross premium formula reserving method could be used,
again with explicit best estimate assumptions for all variables and permitting negative values.
(This method is described in Chapter 19.)
Question
Explain how a gross premium formula reserve is likely to be unrealistic, and hence possibly
inferior to a cashflow approach, for valuing non-linked products for this purpose.
Solution
It does not take account of discontinuance. Losses or profits from discontinuance will affect the
actual value of the business to the company, and will lead to error unless such profits and losses
are expected exactly to cancel out.
1.3 Market-consistent valuations

These will be considered in Chapter 20. See also related comments in Sections 1.2 and 1.8
of Chapter 17.

2 Valuing life insurance contracts: embedded value

Embedded values are often used as a more realistic assessment of the value of an insurer. The
embedded value not only recognises the value of any assets in excess of the reserves, but also
recognises the value to the shareholders of future releases of the margins in those reserves.
Embedded values can be included as supplementary information in published accounts, or can be

used for internal management purposes.
2.1 Calculation of embedded value

Embedded value is the present value of future shareholder profits in respect of the existing
business of a company, including the release of shareholder-owned net assets.
It can be calculated as the sum of:
 The shareholder-owned share of net assets, where net assets are defined as the
excess of assets held over those required to meet liabilities.
These assets may be valued at market value or may be discounted to reflect ‘lock-
in’, for example if they are required to be retained within the fund to cover solvency
capital requirements.
These net assets may have to be invested cautiously to ensure that the solvency capital
requirements are met. As a result, they will be expected to achieve a lower return than
could be achieved with a less constrained investment strategy, and so they will be worth
less than their market value to the shareholders.
Example
An insurer is required to hold solvency capital of 100 in respect of a particular cohort of policies.
This capital is assumed to be locked in until all these policies mature in ten years’ time.
The solvency capital is invested in assets earning only 4% pa. However, the embedded value is
calculated using a risk discount rate of 5% pa.
The discounted value of the assets backing the solvency capital is then given by:
1.0410
100   90.874
1.0510
So these assets may be included in the embedded value at a discounted value of 90.874 rather
than their market value of 100.

 The present value of future shareholder profits arising on existing business.
The process of determining this amount is similar to performing a profit test, bearing
in mind that some elements will not be applicable (eg new business expenses).
The future cashflows on the existing business are projected and used to estimate the
future shareholder profit. These future profits are then discounted back at an
appropriate rate to determine their present value.
The calculation of the present value of future shareholder profits may differ for different
types of business. For example:
 conventional without-profits business: the present value of future premiums plus

investment income less claims and expenses, plus the release of supervisory
reserves
 unit-linked business: the present value of future charges (including surrender

penalties) less expenses and benefits in excess of the unit fund, plus investment
income earned on and the release of any non-unit reserves
 with-profits business: the present value of future shareholder transfers, for example
as generated by bonus declarations.
For with-profits business, the most common ways to calculate shareholder transfers are:
 as a proportion of the cost of bonus (as mentioned in the Core Reading above)
 as a charge deducted from the asset share.
Typically under the cost of bonus approach the with-profits policyholders receive 90% of any
surplus distributed and the shareholders receive 10%, ie the shareholders receive one-ninth
(10/90) of the cost of bonus. For example, if a reversionary bonus is declared such that the
policyholders’ reserves increase by 180, the shareholders would receive 20.
The insurer may decide to hold back part of the surplus (perhaps to pay a terminal bonus at a
later date). In this case the shareholders would have to wait until a bonus was declared to receive
their transfer, so the embedded value is based on the timing of bonus declarations rather than on
the timing of the emergence of surplus (under the cost of bonus approach).
Any excess of assets over the reserves may also need to be split between with-profits
policyholders and shareholders. For example, the shareholder-owned share of the net assets may
be taken as 10% of the net assets within the with-profits fund.
However, if charges are deducted from the asset shares of with-profits policies, then the
embedded value calculation would be simpler and similar to the unit-linked business described
above.

For without-profits business, the present value of shareholder transfers is effectively the
release of any margins within the supervisory reserves relative to the assumptions used
within the embedded value calculation.
If the assumptions used to calculate the supervisory reserves (such as the expected rate of return
on assets), were exactly the same as the future experience assumptions used to calculate the
future cashflows in the embedded value calculation, then the profit emerging in each year would
be zero. This is because the reserves plus the expected investment income would be of exactly
the right amount to cover the expected net outgo in each future year.
Therefore, the extent to which the two bases are different will lead to a profit or loss in each year
within the calculation of the present value of future profits. Because the embedded value
calculation is generally on a more realistic basis than the reserving basis, the reserves plus the
expected future investment returns should be more than enough to cover the estimated future
outgo, leading to the expectation of positive future profits.
Question
Describe, by considering the two components of an embedded value, the mortality risk that a
company faces on:
(a) A portfolio of regular premium without-profits whole life contracts taken out thirty years
ago by men then aged thirty.
(b) A portfolio of regular premium without-profits whole life contracts taken out in the last
year by men aged sixty.
Solution
We also considered this situation in Chapter 2. We stated that there was a risk of capital strain if
more deaths occurred than expected because the death benefit paid would be greater than the
supervisory reserves (and any solvency capital requirement) held. Hence the net assets
component of the embedded value will fall, because the assets paid out (the death benefit) is
greater than the release of the reserves.
However, there is also a risk of a loss of the future profits expected to emerge on this block of
business. Each existing contract will contribute an amount to the embedded value equal to the
present value of the release of margins in the supervisory reserves relative to the assumptions
used within the embedded value calculation. That is, we expect that the current reserves plus
future premiums will be more than enough to cover the claims and expenses. If more deaths
occur than expected in the embedded value projection basis, then these policies can no longer
contribute their margins to the future profits.
So if more deaths occur than expected, the company can see a reduction in both components of
the embedded value.

In case (a), the reserves should be quite large, so deaths will cause a relatively small reduction in
the net assets. The loss of future profits should also be relatively small as there are fewer years
left in which margins can be released (profit has already been released from the first 30 years of
the contract).
In case (b), the reserves will be quite small, so each additional death can cause a substantial
reduction in net assets. The loss of future profits will also be large – in fact all the expected profit
will be lost as the policyholder has not survived the first year.
It is important that the reserves used in the determination of net assets are consistent with
those used in the determination of the present value of future profits.
Tax is allowed for within the calculation as appropriate.
Question
A block of without-profits endowment business has two years left until maturity. The cashflows
(premiums less expenses plus investment return less claims) expected at the end of the first and
second years are 15 and –100 respectively.
The current reserve is 100. The insurer expects to hold a reserve of 110 at the end of the first
year after allowing for the cashflow received at that time. The insurer expects to earn 4% pa on
its reserves.
The insurer has assets of 120.
Calculate the embedded value of this block of business using a risk discount rate of 8% pa.
Solution
The embedded value has two parts (net assets and present value of future shareholder profits).
The net assets are the excess of the assets over the liabilities. Hence, the net assets are 20.
The shareholder profit in each year will be the cashflow plus interest on the reserves less any
increase in the reserves.
At the end of the first year, the cashflow is 15, the interest on the reserves is 4, and the reserves
increase by 10. Hence, the shareholder profit is 9.
At the end of the second year, the cashflow is –100, the interest on the reserves is 4.4, and the
reserves have decreased by 110 (the reserve is zero after the payment of the claims at the end of
the policy). Hence, the shareholder profit is 14.4.
The present value of the shareholder profit at a risk discount rate of 8% pa is then 20.7.
The embedded value is the sum of the net assets and the present value of the shareholder profit,
and so in this case equals 40.7.

2.2 Appraisal value

The starting point in valuing a life insurance company for sale or purchase will be the
embedded value. However, an important element of the price is likely to be goodwill,
corresponding to the estimated profits expected from future business.
An embedded value considers the present value of future profits from the in-force business
only. However, the company should be able to use its expertise and brand to add value to the
business through sales of profitable new business in the future too – this will be represented by
the goodwill element of the purchase price.
The sum of embedded value and goodwill is generally known as ‘appraisal value’. Further
consideration of goodwill is beyond the scope of this Subject.
2.3 Assumptions
The key determinant in deciding on an embedded value (EV) basis is the purpose for which
the EV is being calculated.
We will actually need two different bases in order to calculate an embedded value: a reserving
basis and a projection basis.
We need to know the reserves now in order to calculate the net assets (ie the assets less the
reserves) and we need to know the reserves in the future in order to calculate the present value
of future shareholder profits (as the change in reserves is part of the profit calculation).
Regardless of the purpose for which the EV is being calculated, we would always calculate the
reserves using the supervisory reserving basis.
We also need a projection basis that projects the company’s experience into the future. This basis
will be used in the calculation of the present value of future shareholder profits, and has no
impact on the shareholder-owned share of the net assets. The company may wish to use a
different projection basis depending on the purpose for which the EV is being calculated. It is this
projection basis that we discuss in the remainder of this section.
If an embedded value is being prepared for published accounts or internal management

accounts, then the principles to be applied in setting the assumptions are as outlined in
Section 1.1 above.
The calculation of embedded value for internal management accounts is likely to be carried out
on a best estimate basis. Given that the purpose of calculating an embedded value is to give a
more realistic value of the insurer, then a best estimate basis is likely to be used for published
accounts also, although regulations may require the introduction of a more prudent basis.
If an appraisal value is being prepared as a sale value, then it is likely to be based on

realistic assumptions without margins. A purchase value may be based on cautious
assumptions that include margins.
As a result, the purchase value will be lower than the sale value. The final price agreed for the
business will depend on the relative bargaining power of the buyer and seller.

2.4 Allowing for risk

EV calculations need to include an appropriate risk margin to allow for the unpredictability
of profit emergence for life insurance business.
The purpose of an embedded value is to give a more realistic assessment of the value of the
insurance company to its shareholders. We might believe that a best estimate of the profit in the
tenth year is 100, but this will not be worth as much to the shareholders as a risk-free cashflow of
100 in the tenth year. So we need to reduce the embedded value in some way to reflect the risks
of the business.
The rate of interest used to discount the future streams of surplus will traditionally reflect
the risk inherent in these amounts. It should be noted that increasing the discount rate
normally increases the degree of prudence.
So one possible way of allowing for risk is to use a risk discount rate that is higher than the risk-
free rate. The more risky the future profits, the higher the risk discount rate should be.
A stochastic or market-consistent approach may be adopted.
A stochastic projection would highlight the range of possible values that future profits might take.
Under the market-consistent approach we would use the risk-free rate as the discount rate. We
would assume that investments earned the risk-free rate (regardless of whether we intend to
invest in the risk-free assets or not). A risk margin would be deducted from the embedded value
to reflect non-investment risks, eg by using the cost of capital approach described in Chapter 20.
2.5 What is the difference between a best estimate valuation and an

embedded value?
The two methods are trying to focus on something slightly different and work in different ways,
although they are both based on realistic assumptions.
For a best estimate valuation the present liability is calculated for each policy using realistic
assumptions. The sum over all policies is compared with the total asset value to give a measure
of realistic solvency. This gives a measure of policyholders’ benefit security, but the change each
year in the retained profit (ie the change in the value of the assets less the value of the liabilities)
can also be used to give a measure of profit as an extra piece of information.
An embedded value calculation will consider the cashflows across the portfolio in each time
period, rather than the present value of cashflows for each policy. The focus of an embedded
value calculation is shareholder profit.
An embedded value takes full account of the cost of capital in the value of the shareholder
transfers, because each year’s transfer has to be made from profits that arise in excess of the
supervisory reserves held. (So an increase in the supervisory reserves required will postpone the
emergence of profit for the shareholders, at least for without-profits business, and thereby give a
reduced present value when discounted at the risk discount rate.)

Example
A simple numerical example may help clarify what’s going on here.
 liability = 100 payable with certainty in one year’s time
 current asset value = 105
 supervisory reserving basis = 3% pa interest
 realistic (best estimate) basis = 6% pa interest
 shareholders’ required return on capital = 10% pa
 mortality and expenses: ignored.
(We will assume that the company has no with-profits policyholders, so that all profits will go to
the shareholders.)
To assess the present value of the company’s profits using a best estimate valuation method we
take:
(Value of assets)  (Realistic reserve)
So:
100
Realistic reserve =  94.34
1.06
Realistic profit calculation = 105  94.34  10.66
To assess the same item using embedded value methodology we take:
(Net assets)  (Present value of future shareholder profits)
where:
(Net assets)  (Value of assets)  (Supervisory reserve)
Now:
100
Supervisory reserve =  97.09
1.03
Net assets = 105  97.09  7.91
Expected future profit (in one year’s time) = 97.09  1.06  100  2.91
(An alternative way of looking at this is as a cashflow of –100, plus interest on the reserve of 6% of
97.09, plus a release of reserves of 97.09.)
So:
2.91
Embedded value profit calculation = 7.91   10.56
1.1

So the two methods give similar, though slightly different answers. The difference is the cost to
the shareholders of deferring their profits due to holding reserves – ie the cost of capital.
(For example, if there was no cost of capital – which would be the case if the shareholders’
required rate of return was equal to the 6% they could get from their invested reserves – then the
embedded value profit would have been:
2.91
7.91   10.66
1.06
– ie the same as the other method.)
So, while both methods give the current value of the company’s (retained and future) profit at a
specified date, only the EV method allows for the shareholders’ cost of capital. This makes the EV
method more useful as a measure of the value of the business to the shareholders, and this is the
reason why it has become more important.

3 Consistency
3.1 Consistency in setting a valuation basis
Previous basis
The start point in selecting a valuation basis will be the basis used for the previous valuation. Any
deviation from that must be justifiable.
However, it would be appropriate to change the basis if new evidence was available. For
example, if the gross redemption yield from government bonds had changed, it would be
appropriate to modify the investment return assumption in a similar way.
Changing the basis for a supervisory valuation can have a significant impact on the published
financial results of the company. A weakening of the basis may require justification to the
regulatory authorities that there is some real justification and that it has not been done just to
accelerate the distribution of profits to shareholders or to improve the reported solvency
position. A strengthening of the basis may require justification to the fiscal authorities that it has
not been done to delay profits and hence delay the payment of tax.
Underlying these points we can see an important conflict of interest that the company has to
manage when setting its reserving basis. To reserve over-strongly delays surplus emergence,
which will reduce the returns obtained by the providers of capital. To reserve too weakly, whilst it
will accelerate profits and increase returns, will weaken the company’s financial security and
would increase its insolvency risk. Finding the right balance between these will partly hinge on
educating the capital providers to believe that their interests are best served by a solvent
company (ie that financial security is just as much a part of shareholder value as is the size of the
current dividend stream).
Question
Discuss the risk to the company’s financial security that can be brought about by strengthening
the reserving basis.
Solution
The increase in capital strain per policy issued means that the company is more likely to become
insolvent on the supervisory basis. This assumes that volume of business and capital provision
remain unaltered.
If this were a significant threat, then the company would either have to reduce sales or increase
capital provision, either of which would have the effect of reducing future returns on capital.
Turning customers away would also send out very negative messages to the market, and may
deter intermediaries from recommending the company to customers in the future.

The one-off profit or loss caused by any basis change between successive valuations should be
separately identified in the accounts.
Pricing basis
The pricing basis will have allowed for the cost of setting up supervisory reserves, in the pricing
models used. These reserves will have been based on some assumed reserving basis. The price
determined from this process will then have allowed for the cost of capital, assuming that the
actual reserves for the contract would indeed be set according to the reserving assumptions in
the pricing model. If the actual reserves are, for example, based on more prudent assumptions
than allowed for in the pricing model, then the cost of capital will be higher than that assumed
and the policy will be less profitable than expected. There will also be more capital strain than
expected. Hence it is essential that the assumed reserves in the pricing basis are consistent with
the company’s actual reserving basis for the contract.
Question
Assume that the pricing basis is consistent with the reserving basis, in the way described above.
Outline the further problems that could arise for a company that decides to strengthen its
reserving basis.
Solution
Increased capital strain will increase the cost of capital, so the premiums / charges will have to be
higher to maintain the return on capital.
This will make a price-sensitive product less competitive in the market, leading to a (potentially
devastating) marketing risk (inadequate sales).
Supervisory valuation vs internal valuation

Although not needing to be close, the internal valuation basis should be considered in relation to
the supervisory valuation basis.
3.2 Consistency in setting an embedded value basis
Previous basis
The start point in deciding on a suitable basis will be the basis used for the previous embedded
value calculation. Any differences will immediately cause some movement in the embedded
value. Thus any such change must be justifiable, especially if the embedded value is being used to
report externally on a company’s real worth.
However, if best estimate assumptions have changed since the previous embedded value
calculation, a basis change may well be appropriate and the effect of that change on the
company’s value should be reported.

Pricing basis
The embedded value basis is likely to be more ‘best estimate’ than the pricing basis. However the
two should be looked at side by side. Any differences will immediately lead to embedded value
movements on writing new business different to those implied in the pricing basis.

Chapter 18 Summary
Deciding on a valuation basis
The basis for a valuation should reflect:
 expected future experience
 margins to ensure adequacy of reserves
 legislation / regulation (for published reserves)
 the need for consistency.
The basis for a valuation will depend on whether the accounts are to be published or are
only for internal use.
Any published accounts may be subject to legislative constraints, in particular regarding:

 whether a going-concern or break-up basis is to be used
 whether the accounts have to be true and fair
 whether assumptions should be best estimate or include margins.
Published valuations for the purpose of demonstrating solvency may be subject to different
rules from those governing other kinds of published accounts.
Internal accounts are usually based on best estimate assumptions.
Supervisory reserves (especially for solvency purposes) may require prudential margins or
may be calculated using a market-consistent approach (see Chapter 20).
Calculation of embedded value

Embedded value is the present value of shareholder profits in respect of the existing
business of a company, including the release of shareholder-owned net assets.
It can be calculated as the sum of:

 The shareholder-owned share of net assets, where net assets are defined as the
excess of assets held over those required to meet liabilities.
 The present value of future shareholder profits arising on existing business.
The value of the shareholder profits may be calculated as follows:

investment income less claims and expenses, plus the release of solvency reserves
 unit-linked business: the present value of future charges (including surrender
penalties) less expenses and benefits in excess of the unit fund, plus investment
income earned on and the release of any non-unit reserves
 with-profits business: the present value of future shareholder transfers, for example
as generated by bonus declarations.

The appraisal value is the sum of the embedded value and the goodwill (ie the estimated
profits from future business). The appraisal value may be used when an insurance company is
to be sold.
Setting an embedded value basis

The expected future experience is usually taken as best estimate, unless more prudence is
needed for the particular purpose for which the accounts are required (eg published).
Risk can be allowed for by using a risk discount rate that is higher than the risk-free rate, or by
deducting a risk margin.


18.1 (i) Define the embedded value of a life insurance company. [2]
Exam style A proprietary life insurance company has written conventional with-profits, conventional without-
profits business and unit-linked business over a number of years.
(ii) Describe the calculation of the embedded value for this company. (You do not need to
include details of modelling approaches, data or assumptions.) [7]
(iii) Discuss how the assumptions used to calculate the embedded value are likely to compare
with those used to calculate any supervisory reserves. [14]
[Total 23]



18.1 (i) Embedded value
The embedded value (EV) is the present value of future shareholder profits in respect of the
existing business of an insurance company, plus the release of shareholder-owned net assets. [2]
(ii) Calculation of the embedded value
This can be calculated as the sum of (1) and (2) below.
(1) Shareholder share of net assets
Net assets are defined as the excess of assets held over those required to meet liabilities. [½]
These assets may be valued at market value … [½]
… or may be discounted to reflect ‘lock-in’, for example if they are required to be retained within
the fund to cover solvency capital requirements. [1]
(2) Present value of future shareholder profits on existing business
The process of determining this amount is similar to performing a profit test, …. [½]
… bearing in mind that some elements will not be applicable, eg new business expenses. [½]
The calculation differs in detail for the different types of business:

investment income less claims and expenses, plus the release of supervisory reserves [2]
 unit-linked business: the present value of future charges (including surrender penalties)
less expenses and benefits in excess of the unit fund, plus investment income earned on
and the release of any non-unit reserves [2]
 with-profits business: the present value of future shareholder transfers, for example as
generated by bonus declarations. [1]
It is important that the reserves used in the determination of net assets are consistent with those
used in the determination of the present value of future profits. [½]
Tax is allowed for within the calculation as appropriate. [½]

[Maximum 7]
(iii) Assumptions
The EV calculation will require experience assumptions (to project the business forward) as well
as reserving assumptions. [½]
The reserve values, including the required solvency capital, incorporated within the future profit
projections, should be calculated on a basis that is consistent with that used in practice, allowing
for any expected changes in conditions in the future that could affect the company’s reserving
basis. [1]

The experience assumptions used for the EV calculation will depend on the purpose for which the
EV calculation is needed. [½]
If the EV is required for internal purposes only, then the assumptions are likely to be the expected
values of the future experience, calculated realistically (ie without margins). [½]
The assumptions for an EV that are required for published accounts will reflect the purpose of
those accounts. For example, if they are required in order to present a true and fair view, they
will again be close to best estimates. [1]
If the EV is to be used as a basis for a sale price of the company, the basis could be prudent (on
behalf of the buyer) or optimistic (for the seller). [1]
The assumptions used for the supervisory reserves depend on the supervisory regime. [½]
For example, in some countries, reserves are set up on a relatively realistic basis, ie with relatively
small margins from the expected values, but there is a requirement for a substantial level of
solvency capital determined using risk-based capital techniques. [1]
In other countries, reserves are set up on a relatively prudent basis, ie with relatively large
margins, but with a relatively small solvency capital requirement. [½]
Whichever approach is adopted, the combination of reserves plus solvency margin should ensure
that the insurer holds enough assets to cover the liabilities on its existing business with a high
probability. [½]
We now consider how prudent assumptions for the supervisory reserves will differ from the best
estimate assumptions that would generally be used for the EV basis.
In all cases the margin for prudence will depend on the extent of any guarantees provided by the
product involved, and so will tend to vary by product type. In general, the margins will be highest
for without-profits business and for the guaranteed elements of with-profits business, lowest for
unit-linked contracts, and somewhere between the two for any discretionary components of
with-profits business (eg future bonuses). [1]
Mortality
The margin for prudence will generally depend on the class of business. For example, for
assurances the assumed mortality should be higher than best estimates, while for annuity
business the mortality rates assumed should be lower than best estimates. [1]
As part of that margin, for assurances it will be important not to understate any likely future
worsening of mortality rates, whereas for annuity business future mortality improvements should
not be understated. [1]
Expenses and inflation
Both bases will require assumptions for future renewal and termination expenses, and future
renewal commission payments. [½]
The commissions are known amounts on existing contracts, so should require no estimation. [½]

The only exception might be an allowance for commission clawback, which might be modelled for
the EV (but not for the reserves if surrenders are ignored). [½]
For the reserves, both the level of future expenses and their inflation would be higher than best
estimate. [½]
Withdrawals
The allowance for future withdrawals in a reserving basis will depend on the supervisory regime.
In some countries, future withdrawals will be ignored unless including them would increase the
reserves. [½]
For the experience basis, assumptions for surrenders, paid-up policies and lapses will be made,
split by product and duration. [1]
Investment returns
The rate of interest used in the calculation of the reserves should be chosen prudently (ie lower
than best estimate), having regard to the yields on the corresponding assets backing the liabilities.
[½]
Regulation may restrict the size of the yield allowed, eg by basing it only on the income and total
returns. [½]
In the EV calculation, two assumptions are required: one to project assets, and a risk discount rate
to discount future profits. [1]
The projection assumptions would be closer to best estimate and reflect the total returns that are
expected. [½]
The risk discount rate should reflect the rate of return required by shareholders from the
business, allowing for the risks associated with the projected profits in the EV basis. [1]
Alternatively, the risk discount rate could be set equal to a risk-free rate, with the risks listed
above being allowed for elsewhere within the EV basis. [½]
The risk discount rate will be chosen to be consistent with the investment and inflation
assumptions. [½]
There is no equivalent parameter to the risk discount rate for a reserving basis. [½]
Tax
Allowance for tax should be made on shareholders’ profits in the EV, but this is not needed in the
reserves. [1]
Any other tax assumptions (eg on investment income) are likely to be similar for both the reserve
and EV calculations. [½]

Future bonus rates on conventional with-profits policies
The allowance for future bonuses in reserving calculations is often approximate and may be
implicit in the method being used. [½]
For the EV basis, a realistic projection of future reversionary and terminal bonuses would be
assumed, which would be consistent with the investment return assumptions. [1]
[Maximum 14]

SP2-19: Supervisory reserves and capital requirements (1) Page 1
Supervisory reserves and

capital requirements (1)
Syllabus objective
4.4 Discuss the determination of supervisory reserves and solvency capital requirements
for a life insurance company.
4.4.1 Describe how supervisory reserves and solvency capital requirements may be
determined, including:
 non-unit reserves.

Page 2 SP2-19: Supervisory reserves and capital requirements (1)
0 Introduction
In this and the next chapter we deal with the determination of supervisory reserves and solvency
capital for a life insurance company.
We will begin in this chapter with an overview of the main reserving methods. You should already
be familiar with much of this material from Subject CM1 or CT5, although there is some new
material too (eg on the calculation of negative non-unit reserves). It is important that you are
confident in using these reserving methods as you may be asked to use them in the exam, not
only in reserving questions, but also when calculating surrender values (Chapter 21) and
determining alteration terms (Chapter 22).
In the next chapter, we will consider the assumptions to use when calculating these reserves,
eg we will look at how we can make these reserves market consistent, and we will consider the
need for additional solvency capital over and above these reserves.

1 Background
1.1 Purposes of reserves

There are two fundamentally different purposes for valuing the assets and liabilities of a life
company:
 to demonstrate solvency to the supervisory authorities
 to investigate ‘the truth’ realistically.
The demonstration of solvency to the regulators will involve a minimum valuation standard. The
reason for having such a standard is to ensure that a life company is capable of meeting all of its
guaranteed liabilities. Given that future experience is uncertain, the valuation standards may
require that prudent assumptions are used in the valuation so as to make the chance of failing to
meet liabilities acceptably small.
However, in some countries (including in the EU) there has been movement towards best
estimate assumptions, with additional solvency capital held to make the chance of failing to meet
liabilities acceptably small.
The important thing to note is that it is the sum of the reserves and solvency capital that matters
and that in total they need to be prudent. The necessary prudence may be contained in the
reserves and/or the solvency capital in accordance with the supervisory regulations of the country
concerned.
A realistic valuation to quantify the ‘true’ situation of the life company may be done for internal
management purposes. Typical uses would be:
 to help to determine the long-term sustainability of profit distribution rates (such as
bonuses) and hence to help determine current bonus declarations
 to help determine the realistic profitability of the company for the information of
shareholders (etc) and management
 almost anything else to do with the life company’s financial management.
1.2 Mechanics of reserve calculation

In this chapter we will focus on the gross premium approach to valuation. A gross premium
valuation can be performed using either a formula approach or a cashflow approach. However,
other methods are used in various parts of the world, eg you will have seen the net premium
valuation method in Subject CM1 or CT5.

2 Gross premium valuation method

The Core Reading glossary definition is:
… a method for placing a value on a life insurance company’s liabilities that explicitly
values the future office premiums payable, expenses and claims, with the latter possibly
including future discretionary benefits.
So the reserves for any policy are:
Present value of expected future claims

 Present value of expected future expenses
 Present value of expected future premiums
2.1 Gross premium formula method

The gross premium valuation method can often be expressed as a formula. For instance, for a
regular premium without-profits endowment assurance, the formula would be
SA.A  RE.
a  P .
a
x t:nt| x t:nt| x t:nt|
where: SA = sum assured

RE = renewal expenses
P = office premium.
Usually contracts are priced so that the present value of the premiums exceeds that of the claims
and expenses (eg in order to make a profit). So the reserves should theoretically be negative just
before the payment of the first premium.
Once the contract has begun, we no longer need to reserve for initial expenses or the first
premium. The initial expenses are often much higher than the first premium paid on a regular
premium policy, and this leads to reserves becoming even more negative.
However, any prudence in the reserves (relative to the pricing assumptions) may lead to reserves
being positive even at policy commencement.
For a simple regular premium non-linked without-profits policy, the cost of the benefit payments
will generally increase during the policy term, but the premiums received are level. Renewal
expenses may be assumed to be level but, more realistically, to increase with inflation each year
(for example, by calculating the expense annuity value at rate of interest (i  f ) (1  f ) , where i is
the valuation rate of interest and f the annual rate of inflation). Therefore, the value of future
benefits plus expenses will almost always exceed premiums (except very early on in the contract),
giving a smooth progression of reserves, which start negative and finish with a value equal to the
final benefit payment.

2.2 Gross premium cashflow method

Alternatively, a gross premium valuation can be calculated using a cashflow approach, ie where
the net cashflows are calculated for each future point in time.
Question
List the advantages of calculating reserves using a cashflow approach rather than a formula.
Solution
A projected cashflow approach to calculating reserves has the following advantages:

 It can identify whether net cashflows in any particular period are positive or negative
(which can be important as we will see in calculating non-unit reserves).
 The method can allow more easily for withdrawals.
 It can more easily cope with complex charging and benefit structures (eg unit-linked).
 It can more easily cope with charges and benefits which depend on future assumptions
(eg bonus rates will change over time in response to investment returns, smoothing and
PRE).
 It is easier to incorporate assumptions that vary over time, including stochastic
assumptions.
 The risk discount rate can take account of the term structure of interest rates.
 It is easier to allow for the impact of reinsurance.
 Tax can be allowed for more appropriately.
A cashflow approach can be used for any type of contract, but is particularly useful for valuing
unit-linked contracts. In this case we refer to the gross premium cashflow reserve as a non-unit
reserve as described in the next section.
2.3 Non-unit reserves
Background
In the case of unit-linked business, the cashflow to the company is the income from charges less
the outgo in terms of expenses and claims in excess of the unit reserve. The complexity of these
contracts – in particular the greater variety of patterns of cashflow that they produce – make a
cashflow approach essential. The pattern of income and outgo will not necessarily be ‘smooth’ as
was the case for the endowment in Section 2.1.
The structure of a unit-linked contract is such that the company’s liability is denominated
partly in terms of units and partly in monetary (ie non-unit) terms.

Question
One example of a non-unit liability is administration expenses.
List other liabilities that might be included here.
Solution
Other aspects which might need to be considered here are:

 mortality costs if the death benefit is in excess of the unit reserve
 withdrawal costs if the guaranteed surrender value exceeds the unit reserve
 the future cost of other guarantees (eg minimum maturity benefits related to premiums
paid)
 the future cost of policyholder options (eg option to convert the maturity payout to an
annuity on guaranteed terms).
This leads to a requirement for both a unit and non-unit reserve.
The non-unit reserve is also often referred to (in the UK) as the sterling reserve.
The Core Reading definition of a unit reserve in the glossary is:
… part of the reserve that a life insurance company needs to set up in respect of its unitised
contracts. The unit reserve represents its liability in terms of the units held under the
contracts.
Question
Describe how the unit reserve would be calculated for a unit-linked policy.
Solution
The unit reserve will simply be the number of units multiplied by their ‘bid’ value (ie the price at
which the life company is contractually obliged to buy the units off the policyholders).
The non-unit reserve is the present value of the excess of non-unit outgo (eg expenses,
benefits in excess of the unit fund) over non-unit income (eg charges, unallocated
premiums). A discounted cashflow method is used.
If the present value of the non-unit income exceeds that of the non-unit outgo, then the
overall value would be negative. It may or may not be permitted for a negative non-unit
reserve to be held, depending on the regulatory regime.
The precise details of the calculation will depend on the purpose of the valuation and any
regulations that govern it. We will now cover in more detail the calculation of non-unit reserves
for prudential valuations and best estimate valuations.

Prudential valuation
Under a regulatory regime which requires mathematical reserves to be prudent (rather than
best estimate), the non-unit reserve is typically defined as the amount required to ensure
that the company can pay claims and meet its continuing expenses without recourse to
further finance.
In these circumstances, to calculate the non-unit reserve it is necessary to consider the

year-by-year (and, at outset of the contract, possibly the month-by-month) incidence of the
various components of the non-unit cashflows to determine if and when a non-unit reserve
is required.
The company should project forwards its non-unit cashflows on the reserving basis. This
may need to be performed on a policy-by-policy basis.
The non-unit reserve can then be calculated as follows:

 The calculation process starts with the last projection period in which the net
cashflow becomes negative.
 An amount is set up at the start of that period which is sufficient, allowing for earned
investment return over the period, to ‘zeroise’ the negative cashflow.
 This amount is then deducted from the net cashflow at the end of the previous time
period.
 The process continues to work backwards towards the valuation date, with each
negative being ‘zeroised’ in this way.
 When the process has been completed, if the adjusted cashflow at the valuation date
is negative then a non-unit reserve is set up equal to the absolute value of that
negative amount.
Example
A simple example of the above process is called for. Consider the following series of projected
future cashflows (from the company’s perspective), which are all expressed as amounts at the
end of each year per policy in force at the start of the year:
Year 1 2 3 4 5
Cashflow –3 10 2 –4 1
The last negative cashflow is the –4 at time 4. So the reserve needed at that point (and for the
whole of the preceding year, ignoring interest) is 4.
One year earlier the reserve will be (4 – 2), ie 2.
One year earlier than that the reserve needed is (2 – 10) ie –8, but this is less than zero, so the
reserve will be set to zero.
One year before that the required reserve is (0 – (–3)) ie 3, which is therefore the reserve required
at the valuation date (ie at time 0).

In practice we would allow for interest and survivors. We would proceed as follows:
At the start of year 4, we need to hold the present value of the reserves that will be needed at
time 4, for each policy in force at the start of the year. So the reserve required will be:
3V  4v
Assuming interest at 4%, then 3V  3.85 .
So, looking forward again for a moment, this means that if the company holds reserves of 3.85 for
each such policy in force at the start of year 4, then the company will have just enough money to
meet the expected shortfall of 4 at the end of the year.
At the beginning of the previous year (year 3), we have to hold enough money to pay for the
reserves required at the end of the year, for those who survive to the end of the year. But we are
also going to receive a contribution of 2 towards meeting this cost, at the end of the year. If the
age at the valuation date is x, and (ap)x t is the probability of surviving between x  t and
x  t  1 , then the reserve required will be:
2V  3.85(ap)x 2  2v
If we assume (ap)x 2 = 0.99, then 2V  1.74 .
The year before that, the reserve will be 1V  1.74(ap)x 1  10  v which is, of course, negative, as
before, so would be set to zero. And so on.
Hence, while we often give examples ignoring interest and survival (because this makes it much
easier to see what’s going on), in real life the insurer would certainly not ignore these things. We
have provided the above detail in this case to leave no doubt about what should be done. You
should also recognise this from your earlier CT series courses.
Best estimate valuation

Alternatively, non-unit reserves can be calculated for a best estimate valuation.
In such circumstances, the calculation would value all future non-unit cashflows, ie it would
not disregard cashflows occurring after the last projection period in which there is a net
outflow, and there would be no other restrictions.
Question
Calculate the best estimate for the non-unit reserve required for each year in the example above,
ignoring interest and survival.

Solution
The non-unit reserves can be calculated by starting with the last cashflow and working backwards
as before. However, we must allow for all cashflows – both positive and negative.
The last cashflow is the 1 at time 5. So the reserve needed at that point (and for the whole of the
preceding year, ignoring interest) is –1.
One year earlier at time 4, the reserve will be (–1 – (–4)), ie 3. Note that this is lower than the
reserve of 4 than we obtained for the prudential valuation in the example.
At time 3, the reserve will be (3 – 2), ie 1.
At time 2, the reserve will be (1 – 10) ie –9.
And for the first year the reserve will be (–9 – (–3)) ie –6, which is therefore the reserve required
at the valuation date (ie at time 0).
There is actually no need to work backwards in this way for a best estimate calculation. The
reserve can be calculated more quickly as (–1)  (the sum of all future cashflows). So the initial
reserve is (–1)  (–3 + 10 + 2 –4 + 1) = –6 as before.
In the solution to the question above, there are negative reserves for some of the years.
Under a best estimate (or market-consistent) valuation, it would generally be the case that
negative non-unit reserves can be held.
The idea of a negative non-unit reserve might be difficult to grasp. Firstly, why does it happen?
Basically, a negative reserve will occur whenever the value of future positive cashflows outweighs
the value of the negative cashflows. This can often occur for unit-linked contracts if future
charges are expected to exceed future expenses and other costs.
So, what does this mean? If a contract has a negative reserve it means that we are treating the
contract as an asset (ie more money is going to come in through future charges than go out
through future costs). This sounds a reasonable approach for a best estimate valuation as the
company realistically will make a gain from the contract.
In the following section we will look at negative non-unit reserves in more detail in the case of a
prudential valuation.
2.4 Negative non-unit reserves

We will now look at negative non-unit reserves in the case of a prudential valuation.
Subject to certain conditions, it may be permissible to hold a negative non-unit reserve

under a contract where future non-unit income is expected to be more than sufficient to
meet future non-unit outgo.
The negative reserve represents a ‘loan’ from other contracts which have positive non-unit
reserves. The ‘loan’ will be repaid by the emerging future profits from the policy for which
the negative non-unit reserve is held.

When will negative non-unit reserves be held?

The primary constraint will be whether or not they are allowed by local regulation.
Regulations may specify that:
 The sum of the unit and non-unit reserve for a policy should not be less than any
guaranteed surrender value.
 The future profits arising on the policy with the negative non-unit reserve need to
emerge in time to repay the ‘loan’.
 After taking account of the future non-unit reserves, there are no future negative
cashflows for the policy, ie there should be no future valuation strain.
 In aggregate, the sum of all non-unit reserves should not be negative.
The positive reserves for the fourth bullet point could be from anywhere else in the company’s
business, not just from unit-linked policies. However, regulations might state that the negative
reserves cannot be offset against the unit reserves.
In addition, does the company want to hold negative non-unit reserves? The answer may well be
yes, because they will reduce the total reserve under a contract (ie the sum of unit and non-unit
reserves), and hence improve the capital efficiency of the product.
When will a company be likely to use them? They can be used, in theory, whenever there are
future positive cashflows that the company would like to take advance credit for.
With unit-linked designs, this is most likely to occur when, for example, a level allocation loading
has been deducted from the regular premiums to recoup the heavy initial expenses. This will
produce a large negative cashflow in the first policy year (because the company has just paid its
initial expenses), followed by excess positive cashflows in future years.
Negative non-unit reserves can then be used to deduct, from the liabilities, the expected present
value of these future positive cashflows. In other words, we are taking advance credit for the
value of these future cashflows (ie the future contributions to the initial expenses that the
company expects to receive) so that the new business strain is reduced.

Calculating negative non-unit reserves

A general algorithm for calculating non-unit reserves is as follows:
(1) Project the expected future non-unit cashflow from the policy, ie income from charges
less outgo.
(2) Identify the last (most distant) cashflow (whether positive or negative).
(3) Set the reserve as an amount needed to meet that cashflow at that point in time (even if
the cashflow is positive, set the non-unit reserve as a negative amount).
(4) Check that the total reserve (ie unit plus non-unit) is greater than the surrender value
(ie unit reserve less surrender penalty).
(5) Move back to the next previous cashflow, discount the reserve and then subtract from
the reserve the new cashflow at the earlier time period. Repeat step (4).
(6) Carry on repeating the process, working backwards over time to the valuation date.
(7) This will give the required non-unit reserve.
Example
Consider the same series of projected end-year future cashflows:
Year 1 2 3 4 5
Cashflow –3 10 2 –4 1
Now we are told that there is a surrender penalty of 5 during the 1st year, 4 in the 2nd and so on,
so the penalty is 1 in the 5th year.
The last cashflow is 1 at time 5. So the reserve needed just before that point (ignoring interest)
is –1. Since there is a surrender penalty of 1, we can hold a non-unit reserve of –1 at the start of
that year (ie at time 4).
One year earlier (at time 3) the reserve will be –1 – (–4), ie 3.
One year earlier (at time 2) the reserve will be 3 – 2, ie 1.
At time 1 the reserve will be (1 – 10), ie –9. However the surrender penalty is only 4 during that
year, so the non-unit reserve will be set to –4.
At time 0 the reserve will be (–4 – (–3)) ie –1 (which is OK since the surrender penalty is 5 during
the first year).
Question
Calculate the profit emerging at each of times 0, 1, … 5 in the above example, assuming we hold
the non-unit reserves that we have just worked out and ignoring interest and survival.

Solution
The reserves required are calculated in the example which precedes this question in the course.
The profit at each point is calculated as:
(Reserve available)  (Cash flow)  (Reserve required)
and the reserve available at the beginning of the year is equal to the reserve required at the end
of the previous year.
The calculations are summarised in the following table, and give the profit emerging at the end of
each year.
Year 0 1 2 3 4 5
Reserve available 0 1 4 1 3 1
Cashflow 0 3 10 2 4 1
Reserve required 1 4 1 3 1 0
Profit 1 0 5 0 0 0
Negative non-unit what?

Although the idea of a negative non-unit reserve may seem a little strange at first, it is really just a
way of taking credit in advance for future positive cashflows that we expect on the contract.
However, for a prudential valuation we need to be careful not to take too much credit for future
positive cashflows because, if future cashflows are smaller than we expected, then we could end
up making a future loss. This means that our negative non-unit reserves should not be too big (in
absolute value). So a prudent approach could be calculated by assuming that future positive
cashflows are lower than best estimates, that the rate of interest used to discount them should be
higher than best estimate, and that survival rates are lower than best estimates.
Note how this differs from the calculation of prudent positive non-unit reserves, for which a lower
than best-estimate discount rate would be used.
Bear in mind that positive reserves are to meet future negative cashflows, so negative reserves
are held against positive cashflows.
Also bear in mind that there should be adequate surrender penalties on the contract to ensure
that, if the policyholder withdraws, the value of the future cashflows is not lost.

2.5 Features of the gross premium method

Looking now at the gross premium method at a generic level, whether we are using the simple
formula approach or the more complex cashflow approach, it is possible to identify various salient
features:
 an explicit allowance is made for expenses
 an explicit allowance can be made for bonuses
 the future premiums valued are the actual (‘office’) premiums expected
 any differences between the pricing and valuation bases will immediately be taken as
profit or loss
 reserves may initially be negative for non-linked business, partly due to initial expenses
and partly due to capitalising the expected future profit
 the reserves tend to be quite sensitive to changes in basis.
The fourth feature is a particularly important one: essentially the method will take credit for the
present value of all the future profit that the company will make, if the future reality is as
assumed in the reserving basis. Hence, if the valuation basis is more lenient than the premium
basis (for example, if a realistic valuation was being carried out), then the company would ‘realise’
all the future profits that it would expect to make on account of the future premium being larger
than it needed to be according to this basis. On the other hand, were the reserving basis to be
more prudent than the premium basis, then the company would be capitalising all the expected
future losses that would arise if the reserving basis were to be borne out by reality.
We will return to the sixth feature at the end of the next chapter, where we will see that active
valuation approaches can be sensitive to changes in market conditions.
Question
For an internal valuation of a life insurer’s non-linked business, it has been suggested that a
realistic gross premium valuation method is appropriate, because (realistically) we do wish to
include the value of all the profits we expect to receive from the future gross premiums.
Comment on the validity of this suggestion.

Solution
It is certainly appropriate to value the future gross premiums that we expect to receive, if we wish
to assess the realistic value of the liabilities. (Consistent with this, we would include any
calculated negative reserve values, rather than assume all such negative reserves were zero.)
However, we do not expect to receive all of the gross premiums – we should allow for those
policies that terminate early due to death or withdrawal. While the gross premium formula
method allows properly for mortality, it ignores future withdrawals. In this way, the gross
premium formula method may not fully satisfy the requirements of an internal realistic valuation,
but explicit allowance could be made by using the gross premium cashflow approach instead.

Chapter 19 Summary
Liability valuation methods
Gross premium valuations can be carried out using a discounted cashflow approach or a
formula approach.
A discounted cashflow valuation approach:

 can be used for determining reserves for non-linked policies
 must be used for determining non-unit reserves for unit-linked policies.
Under the gross premium method:

 an explicit allowance is made for expenses
 an explicit allowance can be made for bonuses
 the future premiums valued are the actual (‘office’) premiums expected
 any differences between the pricing and valuation bases will immediately be taken
as profit or loss
 reserves may initially be negative for non-linked business, partly due to initial
expenses and partly due to capitalising the expected future profit
 the reserves tend to be quite sensitive to changes in basis.
Non-unit reserves
Positive non-unit reserves may be required to eliminate future negative cashflows in a
prudential valuation.
Negative non-unit reserves may be used to take advance credit for future excess positive
cashflows, in order to reduce capital strain in a prudential valuation.
A best estimate valuation of the non-unit reserve would value all cashflows and there would
generally be no restrictions on holding negative non-unit reserves.
Certain regulatory constraints may be imposed on the use of negative non-unit reserves in a
prudential valuation, such as:
 The sum of the unit and non-unit reserve for a policy should not be less than any
guaranteed surrender value.
 The future profits arising on the policy with the negative non-unit reserve need to
emerge in time to repay the loan.
 After taking account of the future non-unit reserves, there are no future negative
cashflows for the policy ie there should be no future valuation strain.
 In aggregate, the sum of all non-unit reserves should not be negative.



19.1 A life company currently writes unit-linked life insurance.
Exam style
(i) State the policy data requirements for the valuation of the business. [4]
(ii) Describe how the supervisory reserves would be calculated. [8]

[Total 12]



19.1 (i) Policy data requirements
The question says policy data, and so general assumptions such as mortality tables and expenses
are not required. Similarly, unit prices are not elements of policy data. Clearly you should tailor
your answer to a unit-linked contract and ensure that you cover what is needed for both the unit
reserve and the non-unit reserve.
Unit reserve
The required piece of information is the number of units split by investment fund. [1]
Non-unit reserve
The policy data required are:

 policy number
 date of entry
 date of birth
 gender
 current premium
 maturity date
 premium payment term
 premium frequency
 amounts of any risk benefits.
[½ mark per point]
[Maximum 4]
(ii) Performing the supervisory valuation
Unit reserve
Calculated as number of units  price of units, summed over all funds in which the policy is
invested. [1]
Non-unit reserve
Under a regulatory regime which requires mathematical reserves to be prudent (rather than best
estimate), the non-unit reserve is typically defined as the amount required to ensure that the
company is able to pay claims and meet its continuing expenses without recourse to further
finance. [1]
So the non-unit reserve is calculated so that there are no future negative net cashflows on the
policy, on the valuation basis. [1]

This is ensured by projecting forward the non-unit cashflows on the policy. This will also require
projection of the unit fund, assuming that there are unit related charges. [1]
Non-unit income will come from charges such as unallocated premium, regular administration
fee, fund management charge and risk benefit charges. [1½]
Non-unit outgo will consist of items such as expenses and claims in excess of unit funds. [1]
By working backwards from the last negative non-unit cashflow it is possible to ensure that, net of
changes in reserves, the net cashflow is negative in no future time period. [1]
Any positive non-unit reserve needed now to achieve this must be held in addition to the unit
reserve. [½]
If this leads to a negative non-unit reserve at the valuation date, this can only be held if the
regulatory requirements are met, eg overall reserve positive and at least as large as any surrender
value available. [1½]
Under a regulatory regime which requires mathematical reserves to be best estimate, the
calculation of non-unit reserves would value all future non-unit cashflows, ie it would not
disregard cashflows occurring after the last projection period in which there is a net outflow, and
there would be no other restrictions. [1]
negative non-unit reserves can be held. [½]
[Maximum 8]

Supervisory reserves and

capital requirements (2)
Syllabus objectives
4.4 Discuss the determination of supervisory reserves and solvency capital
requirements for a life insurance company.
4.4.1 Describe how supervisory reserves and solvency capital requirements may
be determined, including:
 market-consistent valuation.
 Value at Risk (VaR) capital assessment.
4.4.2 Discuss the interplay between the strength of the supervisory reserves and
the level of solvency capital required.
4.4.3 Compare passive and active valuation approaches, including the valuation
of assets.
5.1.3 Explain why the assumptions used for supervisory reserves may be
different from those used in pricing.

0 Introduction
In this chapter we continue our work on supervisory reserves and capital requirements.
The supervisory authority’s primary concern will be to ensure that insurance companies have
sufficient assets to cover their liabilities with a high degree of certainty. This could be achieved
by:
 requiring insurance companies to hold reserves calculated on a prudent basis
 requiring a minimum level of solvency capital to be held in addition to reserves calculated
on a best estimate basis
 requiring a combination of prudent reserves and solvency capital to be held.
In this chapter we will consider setting reserving assumptions using either a prudent or a
market-consistent approach, and we will consider the need for additional solvency capital over
and above these reserves.
We can summarise the supervisory solvency requirements in the diagram below:
Free capital
Available
capital Required
capital
Value of
assets
Value of
liabilities
An insurer’s available capital is the excess of its assets over its liabilities. The required capital is
the minimum amount of capital that must be held to satisfy the supervisory authorities. So to be
solvent, the insurer needs to have available capital in excess of this required capital, ie its free
capital needs to be positive.
In this and the previous chapter we cover a variety of different valuation methods. The suitability
of each method depends on a number of factors (different actuaries and regulators have different
views on which factors are the most important).
At the end of this chapter we allocate these valuation methods into two broad groups:
 passive valuation approaches – where the reserves are relatively insensitive to market
conditions and the assumptions are updated infrequently
 active valuation approaches – which are based more closely on market conditions and
where the assumptions are updated more frequently
and discuss their relative merits.

1 Reserving vs pricing assumptions

In some countries it is standard practice to calculate premium rates using prudent
assumptions and then to use the same assumptions to calculate the reserves for
supervisory purposes. This is particularly suitable for with-profits contracts as surplus will
then emerge from actual experience being better than that assumed in the prudent
assumptions.
The approach is less justifiable in the case of without-profits contracts, although it may still
be used.
Question
Explain why there is a problem here for without-profits contracts.
Solution
Without-profits business is likely to be priced using relatively small margins for competitive
reasons. As a result, the pricing basis is unlikely to be sufficiently prudent to use for supervisory
reserving. In contrast, with-profits business can be priced with large margins, as any future
surpluses can be returned to policyholders through bonuses.
In other countries, it is standard practice to calculate premiums using assumptions that

broadly reflect expected future experience, with the risks to the company being allowed for
mainly through the risk discount rate. If the supervisory regime in that country requires
prudence in reserving assumptions then it would not, in that case, be appropriate for the
same assumptions to be used both for pricing and for reserving.
That is, if we used the same assumptions for reserving as we used in pricing, then the reserves
would not be prudent enough for supervisory purposes.
Question
Explain why it would not be appropriate to use the same assumptions for reserving as have been
used for pricing, and allow for risk with an appropriate risk discount rate.
Solution
For a reserving basis, the discount rate has to be the investment return assumption, because a
reserve is an assessment of the amount of assets needed to meet future liabilities, and as such is
dependent on the return that those assets will earn in the future. The risk discount rate is a
measure of the return that the shareholders require on their capital, it has nothing to do with the
return that the company can earn on its assets, and hence is irrelevant to a calculation of the
reserves.

However, some regimes (including in the European Union) have adopted the use of best
estimate or market-consistent assumptions in base reserves. In that case, the underlying
basis could be the same for both pricing and reserving, but additional allowances for risk
would be needed for each.
In Section 2.3 we will see how a risk margin can be added to the best estimate of the liabilities to
allow for risk.

2 Market-consistent valuation
2.1 Market-consistent methodology

As described in Subject CP1, a ‘market-consistent’ approach (sometimes referred to as a
‘fair valuation’) can be taken when setting reserves, with assets also being valued at market
value.
In theory, a market-consistent value of a liability is the price that someone would charge for
taking responsibility for (ie ownership of) the liability, in a market in which such liabilities are
freely traded. In the usual absence of such a market, an approximate approach has to be taken.
To determine the liabilities, future unknown parameter values and cashflows are set to be
consistent with market values, where a corresponding market exists.
For example, the expected cashflows on a conventional without-profits immediate annuity,

allowing for mortality, could be projected. We can calculate a market-consistent value for the
immediate annuity by valuing the cashflows as the maturity proceeds of a series of zero coupon
‘risk-free’ bonds of matching terms.
One way of calculating the value of the liability is then to work out the current total price of the
assets (the zero coupon bonds) that need to be bought in order to generate these cashflows
exactly. In other words, we devise a portfolio of assets whose cashflows exactly replicate those of
the liabilities, and then the current market price of this ‘replicating portfolio’ becomes the
market-consistent value of the liabilities.
Equivalently, we could produce the (same) market-consistent valuation by discounting the

cashflows at current risk-free rates of interest.
Assumed future investment returns are based on a risk-free rate of return, irrespective of
the type of asset actually held, and the discount rates are also based on risk-free rates.
A financial economist would argue that the value of the liabilities is the same regardless of
whether we actually back our immediate annuities with the risk-free asset or not because the
choice of assets has no impact on the size of benefits paid.
Of course, if we invested in risky assets we would hope that they would give a higher return, but
they also have higher risk. We will allow for this risk when we consider solvency capital in
Section 3.
‘Risk-free’ rates may be determined based on government bond yields or on swap rates.
We would actually use different discount rates for cashflows at different durations, if this were
necessary to reflect current market yields. For example, if the yield curve slopes upwards, we
would use a higher ‘risk-free’ rate to discount cashflows that occur in ten years’ time than in two
years’ time.
It would generally only be appropriate to use swap rates if there is a sufficiently deep and
liquid swap market in that country, where ‘deep’ means that the market is of sufficient
capacity that large trades would not materially affect the prices.

An interest-rate swap is a contract where one party agrees to pay a series of payments based on a
fixed interest rate in exchange for a series of payments made on a variable rate of interest
(usually the prevailing interest rate on short-term deposits in the market).
The fixed interest rate in the swap agreement will vary depending on the term of the swap. For
example, the agreement might be to pay a fixed rate of 2% pa for a two-year swap, but 3% pa for
a ten-year swap. We can then use these swap rates to derive risk-free rates as an alternative to
using the yield curve of government bonds.
Whether the risk-free rates are based on government bond yields or swap rates, it may be
appropriate to make a deduction to allow for credit risk.
The market-consistent approach is based on risk-free rates, but swaps are not entirely risk free as
one of the counterparties might default (although some form of collateral arrangement will
reduce this risk substantially). Even government bonds are not entirely risk free, eg we have seen
many countries lose their AAA credit rating in recent years.
Market-consistent valuation methods have been increasing in importance in some

countries.
2.2 Illiquidity premium

It may be possible to derive the market-consistent discount rate from corporate bonds rather
than government bonds. At first sight this might appear to be an odd thing to do, because
corporate bonds are more risky than government bonds. Firstly, corporate bonds usually have a
higher probability of default than government bonds. Secondly, corporate bond prices tend to be
more volatile than government bond prices (ie corporate bonds are less liquid), which can be a
problem if we want to sell the bond before its maturity.
Corporate bonds typically have a higher yield than risk-free (eg government) bonds, where
this reflects both the greater default risk and the relative illiquidity of such assets. The
latter contributes the ‘illiquidity premium’ to the yield.
Investors will require a higher expected return on an asset to compensate them for any additional
risks, eg the higher probability of default on a corporate bond compared to a government bond.
The illiquidity premium (often confusingly called the liquidity premium) is the extra return that
investors require to compensate them for the risk of greater price volatility of corporate bonds.
Even under a market-consistent approach, it may be possible in some jurisdictions to take

credit for the illiquidity premium and thereby discount liabilities at a higher yield than the
risk-free rate within the supervisory valuation.
For example, the yield on government bonds might be 4% pa, but the yield on corporate bonds
might be 7% pa. We might believe that the market requires an extra yield on corporate bonds of
2% pa to reflect default risk and 1% pa to reflect illiquidity. Some jurisdictions would then set the
market-consistent discount rate at 5% pa (the 4% pa ‘risk-free’ government bond yield plus the
1% pa illiquidity premium).

Question
Describe the impact on an insurance company’s solvency position if the supervisory authority
changed its rules to allow the insurance company to take credit for the illiquidity premium
described above.
Solution
The insurance company’s solvency position would improve.
Under the old rules, the insurance company would have to discount its liabilities at the risk-free
rate of 4% pa.
Under the new rules, the insurance company could add in the illiquidity premium so that it
discounted its liabilities at 5% pa. The higher the discount rate, the lower the value of the
liabilities, and hence the better the solvency position.
This would normally be restricted to long-term predictable liabilities for which matching
assets can be held to maturity. Since the insurer is not exposed to the risk of changing
spreads on such assets (although is still exposed to default risk), it may be permitted to
increase the risk-free discount rate accordingly.
The above paragraph is important, because it explains why we can take credit for the illiquidity
premium.
If we match our annuity cashflows with bonds of appropriate term, then we do not care about
any subsequent price volatility as we will be holding the bonds to maturity. So in a sense, the
extra return on these bonds from the illiquidity premium is a ‘free lunch’ because we are not
exposed to the liquidity risk.
However, the insurance company is still very much exposed to the risk that the bonds default, and
so we shouldn’t include the default risk premium in our risk-free discount rate.
The Core Reading tells us that we can normally only use the illiquidity premium in this way for
long-term predictable liabilities. The cashflows on a large block of immediate annuities will be
reasonably predictable if we have a reliable estimate for future mortality rates. However, other
classes are much less predictable because of the possibility of withdrawal. We could hold bonds
to match the expected death and maturity payouts from a conventional without-profits
endowment assurance, but we would be exposed to the price volatility of these bonds if we
subsequently had to sell the bonds to pay a surrender value.
Where this practice is permitted (within a market-consistent supervisory valuation regime),

there would normally be strict rules about how and when it can be applied.
For example, the supervisory authority may restrict the types of contract for which an illiquidity
premium is allowed (as described above). It might also restrict the investment strategy, eg there
may be a requirement that bonds are held to maturity.

2.3 Risk margin

The approach described in Section 2.1 is strictly only appropriate for cashflows that are of known
and certain amounts.
Question
Explain why the approach described in Section 2.1 will not actually give the correct
market-consistent value for the immediate annuity liabilities described earlier.
Solution
This is because the annuity cashflows are not of certain amounts, but are instead dependent upon
how many (and precisely which) of the current annuitants survive to each future date. In other
words, the cashflows are subject to uncertainty due to mortality.
It may be difficult to obtain a ‘market-consistent’ assumption for some elements of the

basis, such as mortality, persistency or expenses, for which there is not a sufficiently deep
and liquid market in which to trade or hedge such risks.
In some cases it may be possible to use some market-consistent indicators for setting the
assumptions. For example:
 the mortality assumption could be determined from the risk premium rates quoted by
reinsurance companies
 the expense assumption could be determined by reference to expense agreements
available in the market, eg from third party administration companies
 with-profits endowment assurances may be traded as investment vehicles, so there may
be a current market price for these products as a whole.
However, there are a number of reasons why the examples given above might not be used in a
market-consistent valuation. For example, it is unlikely that third party administration companies
or reinsurers would provide a quote unless the insurance company actually intended to use their
services and the traded endowment market is unlikely to be considered to be sufficiently deep
and liquid.
The starting point for such assumptions would be a best estimate basis.
So with uncertain cashflows, the approach is to project their expected future amounts and then
discount these as if they were known and certain, according to the principles in Section 2.1. So, in
the case of the immediate annuity cashflows, we would first need to multiply the future annuity
payments by some assumed survival probabilities, ie using an appropriate mortality assumption.

Question
If we projected cashflows using best-estimate assumptions for these variables, the value we
would obtain for the liabilities would not be a correct determination of the market-consistent
value.
Explain whether the value obtained would be higher or lower than the actual market-consistent
value.
Solution
Using best-estimate mortality assumptions would produce a value that was lower than the
market-consistent value. This is because a purchaser of the liability would require additional
compensation, for the possibility that the liability should turn out to be more expensive than
expected.
It is then likely that a risk margin would be included in respect of these assumptions, due to
the inherent uncertainty. This risk margin would reflect the compensation required by the
‘market’ in return for taking on those uncertain aspects of the liability cashflows.
This could be done by adding a margin to each such assumption.
Question
Explain whether we should use (appropriately) higher or lower mortality rates than best estimate
in order to produce a market-consistent valuation for the immediate annuity example.
Solution
We should assume lower mortality, as it is greater longevity that will increase the cost of these
liabilities to our notional purchaser.
Alternatively, an overall reserving margin in respect of these risks could be determined

using the ‘cost of capital’ approach.
In summary, the ‘cost of capital’ method for determining the additional risk margin is as follows:
 firstly project the required capital at each future time period (ie the amount required in
excess of the projected liabilities)
 multiply the projected capital amounts by the cost of capital
 discount using market-consistent discount rates to give the overall risk margin.

This is discussed in more depth below:
This involves first projecting forward the future capital that the company is required to hold
in respect of these risks, where this is determined at the end of each projection period
(eg year) during the run-off of the business. The projected capital is determined according
to the relevant regulatory basis, and is the amount required to be held in excess of the
projected liabilities at each period.
So we calculate the amount of capital we are required to hold in excess of our market-consistent
estimate of the liabilities.
These projected capital amounts are then multiplied by a cost of capital rate. This rate can
be considered to represent the cost of raising incremental capital in excess of the risk-free
rate, or alternatively it represents the frictional cost to the company of locking in this capital
to earn a risk-free rate rather than being able to invest it freely for higher reward.
The providers of capital (eg shareholders in a proprietary company) could have invested their
money in shares earning 9% pa say, but instead they are backing the capital requirements with
bonds earning 3% pa say, which implies a cost of capital of 6% pa.
There may also be other frictional costs, such as tax, which also affect the cost of capital.
The cost of capital rate may, for example, be determined as the excess of the weighted
average cost of capital over the risk-free rate, and so may be term dependent. In some
calculations it is a specified fixed rate (eg 6% per annum in Solvency II, the regulatory
reporting environment applicable to the European Union).
The product of the cost of capital rate and the capital requirement at each future projection
point is then discounted, using market-consistent discount rates, to give the overall risk
margin:
kt  Ct
Risk margin = 
t  1  rt t
where kt is the cost of capital ‘charge’ for time t, Ct is the required capital at time t and rt
is the risk-free interest rate for maturity t.
So kt would be 6% for the example mentioned above.
For some chosen regulatory frameworks, the projection of required capital can be relatively
straightforward, for example where it is defined as a fixed percentage of reserves.
However, for others it can be complicated if the calculation of required capital itself requires
projections, stochastic modelling and/or application of correlation matrices.
We will see how a correlation matrix can be used in Section 3.3.
It is complex enough to use the above approaches to calculate the solvency capital required now.
However, the cost of capital approach requires us to project the capital required Ct at each
future point in time too.

In such situations, a simplified approach can normally be used. This might involve
selecting a driver (eg reserves or sum at risk) which has an approximately linear
relationship with the required capital or its components.
By ‘sum at risk’, we mean the difference between the benefit and the reserve held.
The initial capital requirement can be expressed as a percentage of that driver, and the
projected capital is then approximated as the same percentage of the projected values of
the driver.
If we observed that the capital requirements tended to be approximately 10% of the reserves,
then we would set Ct to be 10% of the reserves in the calculation of the risk margin above.
In practice, more sophisticated methods using a combination of drivers and correlations

may have to be used.
It is highly unlikely that our simple model based on the reserves will be accurate enough.
However, we may get a reasonably good fit by considering a combination of drivers, eg by using
the projected values of the reserves, sum at risk, interest rates, inflation and so on.
These methods are sometimes described as proxy models or light models as they are a simpler
alternative to the full model used to calculate the current solvency capital. We will return to
these models in more detail in Subject SA2.

3 Solvency capital requirements
3.1 Purpose
It is usual for insurance supervisors to require that a life insurance company maintain at
least a specified level of solvency capital. This solvency capital can be seen as providing
an additional level of protection to policyholders.
Question
Suggest two broad areas of risk where such solvency capital protects policyholders.
Solution
It will protect policyholders against:

 the reserves being underestimated (ie adverse future experience relative to the reserving
basis assumptions), and
 a drop in asset values (including individual asset defaults).
The level of solvency capital required under regulation may be specified as a formula, or it
may be based on a risk measure such as Value at Risk (VaR).
A simple formula might set the solvency capital as 3% of the reserves to cover a fall in asset values
and 0.3% of the sum at risk to cover adverse mortality experience.
We will cover the Value at Risk approach in more detail in Section 3.3.
3.2 Interplay between reserves and solvency capital requirements

In considering the adequacy of the reserves that have been set up, it is important to do this
within the context of the solvency capital requirements and not in isolation. Similarly, the
adequacy of the solvency capital requirements cannot be looked at in isolation of the
reserving requirements.
Practice on the relative balance between the two components varies between countries and
regulatory jurisdictions. In some, reserves are set up on a relatively realistic basis, ie with
relatively small margins from the expected values, but there is a requirement for a
substantial level of solvency capital determined using risk-based capital techniques.
The aim of risk-based capital is to set aside an extra amount of capital where the amount is
appropriate to the extent of the risks involved.
In other countries, reserves are set up on a relatively prudent basis, ie with relatively large
margins, but with a relatively small solvency capital requirement which is not specifically
related to the risks borne by the company.

3.3 Value at Risk approach

An example of a risk-based solvency capital requirement approach is the use of a VaR
measure, normally expressed at a minimum required confidence level (eg 99.5%) over a
defined period (eg one year).
So the insurer calculates the amount of capital it needs so that its assets will exceed its liabilities
in one year’s time with a probability of 0.995, say.
For example, a VaR of £10 million over the next year with a 99.5% confidence level means that
there is only a 0.5% probability of losses being greater than £10 million over the next year.
Probability
0.5%
-10 0
The supervisory balance sheet (which would typically be on a market-consistent basis for
this type of approach) is subject to stress tests (or ‘shocks’) on each of the identified risk
factors, at the defined confidence level and over the defined period. The (market-
consistent) surplus is then recalculated at the end of the period.
The supervisory balance sheet compares the assets and liabilities of the insurer. The excess of the
assets over the liabilities is the surplus or available capital of the insurer.
Free capital
Available
capital Required
capital
Value of
assets
Value of
liabilities
The insurer determines an adverse outcome (or shock) for the next year that it considers to be so
severe that an even more extreme outcome would only occur with probability 0.005 (ie only 0.5%
of the time). For example, the adverse outcome might be a fall in the value of the stock market of
35% over the year.

The insurer then projects its assets and liabilities at the end of the year under this shock. The
risk-based capital is then the amount of capital required to be held at the start of the year, that
will ensure its assets are no smaller than its liabilities at the end of the year, on the assumption
that the shock happens. Any excess of the insurer’s available capital over its required capital is
called free capital or free surplus.
For example, consider an insurer with assets currently worth 100 and liabilities of 80, so the
current surplus is 20. A stress test leads to revised values of assets of 85 and liabilities of 76, so
the surplus is now only 9 and the capital required is 20 – 9 = 11. So we have assets of 80 backing
the reserves, 11 backing the capital requirement, and 9 that is free surplus.
Other possible methods of calculation may exist, such as the ‘run-off’ method, which looks
at the amount of capital needed at outset to ensure a firm’s ability to cover its liabilities until
the last policy has gone off the books, allowing for suitable stresses to the risk factors.
So instead of projecting over one year, the insurer projects its assets and liabilities until the last
in-force policy has become a maturity, death or surrender. New business is ignored. The insurer
determines the amount of capital that needs to be held at outset so that assets will exceed
liabilities with an appropriate probability.
Applying stress tests to each different risk factor gives a capital requirement for each
separate risk in isolation.
It would be usual to calculate separate shocks for mortality, expenses and so on. In this way the
insurer could calculate that it needed say £100 million of capital to cover 99.5% of mortality risks
and £20 million of capital to cover 99.5% of expense risks. It may have calculated these figures by
shocking mortality rates by 1% and then a separate calculation which shocked expenses by 10%.
An alternative would be to determine a single scenario with 99.5% confidence which involved
simultaneously shocking mortality, expenses, investment return, withdrawals and so on.
However, this alternative approach is currently too computationally difficult to perform, so the
separate stress test approach is used instead.
In order to arrive at an aggregated capital requirement reflecting all risks, these separate
stress tests need to be combined in a way which reflects any diversification benefits that
exist between the various risks (ie the degree to which individual risks are correlated).
The insurer could have calculated the amount of capital to hold as the sum of the capital amounts
under each of the separate stress tests. So it could hold £120 million in respect of mortality and
expenses. However, this would be too much as it is unlikely that the mortality and expense
shocks would happen at the same time, ie these risks are not perfectly correlated so the insurer
can hold a smaller amount of capital to reflect the diversified portfolio of risks that it is exposed
to.
This may be done through the use of correlation matrices (noting that, under the extreme
event conditions being tested, correlations may differ from those observed under ‘normal’
conditions) or by copulas, for example.

Using a correlation matrix approach, the aggregated capital requirement can be calculated as
follows:
Aggregated capital requirement   Corrij  Capi  Cap j

i j
where:
Capi = capital requirement under risk i
Corrij = correlation between risks i and j.
The following table shows a simplified example of a correlation matrix for an insurance company
selling term assurance business. So for example, the correlation between mortality risk and
expense risk is 50% in this case. In practice, the correlation matrix will include many more risks.
Mortality
Expense
j
Lapse
Mortality 100% 0% 50%

Lapse 0% 100% 50%
Expense 50% 50% 100%
Question
Calculate the aggregated capital requirement using the table above, if the separately calculated
capital requirements are as follows:
 capital requirement for mortality risk = £100 million
 capital requirement for lapse risk = £10 million
 capital requirement for expense risk = £20 million.
Solution
1002  102  202
Aggregated capital requirement  2  0.5  100  20
2  0.5  10  20
 £112.69 million

As the risks are not 100% correlated with each other in the example above, the aggregated capital
requirement is less than the sum of the capital requirements for each separate risk. This
demonstrates the benefits of diversification amongst the risks.
It should also be recognised that a combination of a certain subset of events happening at

the same time, with an overall probability level of 1 in 200 (for example), may produce a
higher capital requirement than combining all of the individual capital requirements for
separate 1 in 200 events using a correlation matrix.
This is caused by the ‘non-linearity’ and ‘non-separability’ of individual risks, the latter
referring to the ways in which risk drivers interact with each other. Separate allowance
needs to be made in the capital requirement calculation for these effects.
The linearity property means that the capital required is a linear function of the risk drivers. For
example, if a fall in interest rates from 5% pa to 4% pa leads to a £10m increase in required
capital, then the linearity property might imply that a fall in interest rates from 5% pa to 3% pa
would lead to a £20m increase in required capital.
However, in practice we see non-linear relationships between the capital required and the risk
drivers. For example, consider an endowment policy with an option that an annuity can be
purchased at a guaranteed rate of 3.5% pa interest. If interest rates fell from 5% pa to 4% pa the
liabilities might increase by only a modest amount. However, if interest rates fell from 5% pa to
3% pa, the guarantee would bite and a large number of policyholders would exercise the option,
leading to a very substantial increase in liabilities (and hence a very large increase in the capital
required to withstand this event).
The concept of non-separability refers to situations where if two events happen together, the
combined impact is worse than if they had happened separately. For example, consider longevity
risk and expense risk for annuities. If annuitants live longer than expected and per policy
expenses are higher than expected, there is an additional combined impact due to having these
higher than expected expenses payable for longer than expected. If we considered the two risk
factors separately, we wouldn’t be allowing for this additional combined effect.
So, as the risks in practice show non-linearity and non-separability, and these effects generally
lead to higher capital requirements, the capital requirement calculated using a correlation matrix
will be too low unless it is adjusted upwards to allow for these two effects.
Typically, stochastic models are used to quantify the capital requirements in relation to
economic risks. The probability distribution used should properly reproduce the more
extreme behaviour of the variable being modelled, both in the size of the tail of the
distribution and, where appropriate, in the path taken during the simulation period.
So, if we are calculating the VaR for investment risk at the 99.5% confidence level, we need to be
able to determine an investment return that is so low that a lower return happens with a
probability of only 0.5%. However, we may also need to know how the investment return will
have changed over time, eg if investment returns are high for several months before falling
sharply then bonuses on with-profits policies may have initially been quite high, which will make
the subsequent fall in asset values even more severe for the insurer.

For capital requirement projections, a ‘real world’ asset model would typically be used (see
Chapter 14 Section 1.4) and this should be arbitrage free.
We use a real world calibration as we want to investigate the true probability of events occurring
when we consider the confidence level in the VaR calculation.
It is generally appropriate to calibrate such models with reference to actual historic

parameters, but advanced techniques may be required to ensure appropriate fit to the tail of
a distribution, to ensure that the distributions do not understate the frequency of more
extreme outcomes.
This is very difficult to do in practice and a discussion of the techniques used is beyond the scope
of this course. The key thing to remember is that although we may be able to obtain a
distribution that has a good fit to the data around the mean, this is not the part of the distribution
that we are interested in to calculate a VaR. A 99.5% confidence level is equivalent to looking at a
1 in 200 year event, and it is difficult to extrapolate such events from only a few decades of actual
data.

4 Active and passive valuation approaches
4.1 Passive valuation approach

A passive valuation approach is one which uses a valuation methodology which is relatively
insensitive to changes in market conditions and a valuation basis which is updated
relatively infrequently.
A change in market conditions can affect both the value of the assets and the value of the
liabilities (eg if the valuation interest rate is derived from the yield available on assets). However,
we will see below that under some valuation methods the value of the assets and/or the value of
the liabilities change very little with market movements.
Under a passive valuation approach, the assumptions for mortality and expense inflation would
rarely change.
An example would be the net premium valuation approach for liabilities (as described in
Subject CM1), which is fairly insensitive to yield changes due to the net premium also being
recalculated under the new assumptions.
Question
Describe how a net premium valuation would be performed.
Solution
The glossary for Subject SP2 gives the following description of a net premium valuation:
This is a method for placing a value on a life insurance company’s liabilities that involves
calculating a present value of the contractual liabilities and deducting the value of future net
premiums, allowing in each case only for mortality and interest.
The key point here is that the calculation ignores expenses. So a net premium valuation is the
expected present value of the benefits less the net premiums.
The net premium is the premium, calculated on the valuation assumptions and payable
under the same conditions as the office premium, that will provide the contractual benefits
offered at the commencement of the policy.
The key point here is that the calculation makes no reference to the actual premium paid. Instead
a net premium is calculated that ignores expenses and is calculated using the valuation
assumptions (ie it does not use the pricing basis assumptions).
Question
State a general formula for the net premium reserve for a regular premium without-profits
endowment assurance, defining all the terms used.

Solution
The net premium reserve at time t for a regular premium without-profits endowment assurance
with constant premiums would be:
S. A  P.
a
x t:nt| x t:nt|
where: S = sum assured

A |
x:n
P = net premium  S.

a |
x:n
x = age at outset
n = contract term
Question
Explain why the net premium valuation is relatively insensitive to changes in the valuation interest
rate compared to a gross premium valuation.
Solution
If the valuation interest rate decreased (say from 4% pa to 3% pa) then the value of the assurance
factor A and the annuity factor 
a would both increase. However, the impact would
x t:nt| x t:nt|
be greatest for the assurance factor (as the benefits are paid later than the premiums and so are
discounted for longer). So we would expect the reserve to increase for a gross premium valuation
(assuming the reserve was positive).
However, for a net premium valuation we would also need to recalculate the net premium using
the new valuation interest rate. A decrease in the interest rate would increase the net premium
P . So the present value of the premiums would increase by more under a net premium valuation
approach, and the increase in premiums would be closer to the increase in the benefits. So for a
net premium valuation the impact of a change in the valuation interest rate is less than for a gross
premium valuation.
Under passive approaches, assumptions may be ‘locked in’. That is, they remain
unchanged from those used when that policy was first written and the liability for it first
determined. It may be a requirement, however, that non-economic assumptions are
updated if experience worsens, in order to recognise the related loss and the need for
higher reserves at that time.
So the same valuation interest rate may be used throughout the term of the policy (ie the interest
rate is locked in). This might be considered acceptable because any increase (decrease) in interest
rates would decrease (increase) both the value of the assets and the value of the liabilities. If
assets are chosen that are well matched to the liabilities the true solvency of the company will be
unaffected by market changes.

However, non-economic assumptions such as mortality may have to be varied over time because
if experience deteriorates the cost of benefits will rise without any corresponding change in asset
values.
For the valuation of assets, an example of a passive approach would be the use of historic
cost or ‘book value’, possibly with amortisation (or ‘write-down’) over time.
By using a book value approach to valuing assets, we are ignoring the impact of changes in market
prices. This might be suitable if the valuation of liabilities is also largely ignoring market
conditions, eg if the valuation interest rate is locked in or if a net premium valuation is used.
Solvency capital requirements may be determined using a simplified approach such as

holding a prescribed percentage of mathematical reserves.
So what are the advantages of using a passive valuation approach?
Passive valuation approaches:
 tend to be more straightforward to implement
 involve less subjectivity
 (to the extent that they are used for accounting purposes) result in relatively stable
profit emergence.
Question
Describe the key disadvantage of a passive valuation approach.
Solution
A passive valuation is at risk, by its very definition, of becoming out of date. It is relatively
insensitive to changes in market conditions and has a valuation basis which is updated relatively
infrequently.
For example, if the stock market crashes then the book value of the assets will be overvalued
relative to their value if sold in the market today. Similarly, the net premium valuation is
relatively insensitive to changes in interest rates (as we described in the solution to the previous
question).
If the valuation basis is changed infrequently then it may not have taken account of important
trends, eg rising expense inflation or deteriorating claims experience.
So there is a danger that a passive valuation approach provides a false sense of security.
Management may fail to take appropriate actions in response to emerging problems until too
late, because the solvency position hasn’t appeared to change.

4.2 Active valuation approach

In contrast, an active approach would be based more closely on market conditions, with the
assumptions being updated on a frequent basis.
An example would be the use of market-consistent valuation approaches for both assets
and liabilities, and a risk-based capital approach to solvency capital requirements.
The market-consistent valuation approach was described in Section 2 and a risk-based capital
approach was described in Section 3.3.
What are the benefits of using an active approach?
Active valuation approaches are more informative in terms of understanding the impact of
market conditions on the ability of the company to meet its obligations, particularly in
relation to financial guarantees and options.
What are the disadvantages?
However, results are more volatile under such approaches. Under adverse equity market
conditions (eg stock market crash), an active valuation approach using risk-based capital
would indicate that higher capital requirements are needed. In order to reduce this
requirement, companies would need to sell equities – which itself could exacerbate the
market conditions. There is also systemic risk, as this would be the case for all life
insurance companies at the same time. Therefore it may be the case that regulators include
amendments to the valuation approaches under such conditions, to avoid this situation.
This is not just a theoretical risk. This is exactly what happened following the financial crisis of
2007/2008. As market prices fell (particularly for assets such as corporate bonds and equities) the
solvency of banks and insurers also fell. To protect their solvency many of these financial
institutions sold their risky assets and replaced them with safer assets such as government bonds.
This meant that there was a huge supply of risky assets but very little demand, and so the market
price slumped to a level far less than what many analysts would have regarded as the true value.
Question
Outline other disadvantages to an active valuation approach.
Solution
An active valuation approach is likely to be more complex than a passive approach. So the
calculations could take longer to perform and be more costly.

4.3 Combinations
The overall valuation approach may instead be somewhere between the two extremes,
including elements of each.
This can result in a greater mismatch between assets and liabilities, and hence greater
profits / losses or changes in free surplus when market conditions change. For example, an
approach which uses a net premium valuation for liabilities and market value of assets
could experience greater volatility of results than one which uses a market-consistent
valuation for both.
Question
Explain the last sentence above.
Solution
If the assets and liabilities are well matched, then the true impact of market movements on the
company’s solvency should also be quite small. Under an active valuation approach, any change
in the value of the assets should be reflected in a corresponding change in the value of the
liabilities (eg if interest rates fall then both valuations should rise), so the solvency position should
be little changed. Under a passive valuation approach, the value of the assets and liabilities might
be little changed by market conditions, so again the solvency position should be little changed.
However, if a combination of active and passive approaches is used then solvency might appear to
change dramatically (when in reality it hasn’t changed). A change in interest rates would change
the market price of assets but would have little impact on the net premium valuation of liabilities.

Chapter 20 Summary
Reserving assumptions vs pricing assumptions
It is common in some countries to price prudently and then to define the reserving basis as
the pricing basis.
This may be suitable for with-profits business as it can allow profit to emerge gradually and
appropriately to the distribution method. It is less appropriate for without-profits business.
In other countries premiums may be calculated using broadly best estimate assumptions
(and allowing for risk through a risk discount rate). In this case the pricing assumptions
cannot be used if the regulations require prudent reserving assumptions.
In other regimes there has been a move towards using best estimate or market-consistent
assumptions in reserving. In this case the basis for pricing and reserving could be quite
similar.
To determine a market-consistent value of liabilities, future unknown parameter values and
cashflows are set so as to be consistent with market values, where a corresponding market
exists. Market values are also used for assets, if market prices exist.
Future investment returns are based on a risk-free rate of return, irrespective of the type of
asset actually held, and the discount rates are also based on risk-free rates.
Risk-free rates may be based on government bond yields or on swap rates. It may be
appropriate to make a deduction to allow for credit risk.
Under certain conditions, it may be possible to take credit for the illiquidity premium
available on corporate bonds and thereby discount liabilities at a higher yield than the risk-
free rate.
It may be difficult to obtain a ‘market-consistent’ assumption for some elements of the basis,
such as mortality, persistency or expenses, for which there is not a sufficiently deep and
liquid market in which to trade or hedge such risks.
It is then likely that a risk margin would be added to the best estimate of the liabilities. This
risk margin would reflect the compensation required by the ‘market’ in return for taking on
those uncertain aspects of the liability cashflows.
This could be done by adding a margin to each such assumption or by using the ‘cost of
capital’ approach.

Solvency capital requirements

Insurance supervisory authorities normally require that life insurance companies establish a
certain amount of solvency capital. This is to protect policyholders from a company
reserving too little for their liabilities, and from the harmful effects of asset volatility.
Supervisory reserving needs to be considered in conjunction with the solvency capital
requirements, and vice versa.
The relationship between reserves and solvency capital requirements varies between
different countries and regulatory jurisdictions. Normally it will be one of two cases:
 strong reserving, with a small solvency capital requirement
 weak reserving, with a large solvency capital requirement
where weak reserving means a basis close to best estimate.
The level of solvency capital required under regulation may be specified as a formula, or it
may be based on a risk measure such as Value at Risk (VaR).
The VaR can be calculated by subjecting the supervisory balance sheet to stress tests on each
of the identified risk factors, at the defined confidence level and over the defined period.
The aggregated capital requirement combines the separate stress tests to reflect any
diversification benefits that exist between the various risks. This may be done through the
use of correlation matrices or by copulas.
Passive and active valuation approaches

A passive valuation approach is relatively insensitive to changes in market conditions and has
a valuation basis which is updated relatively infrequently.
An active approach is based more closely on market conditions, with the assumptions being
updated on a frequent basis.
Passive approaches tend to be easier to implement, involve less subjectivity and result in
relatively stable profit emergence. Active approaches are more informative in terms of
understanding the impact of market conditions.


20.1 Outline the cost of capital approach of determining the risk margin for a market-consistent
Exam style
valuation. [6]
20.2 (i) Define what is meant by a market-consistent valuation, in the context of a valuation of a
Exam style
life insurance company’s liabilities. [1]
(ii) Explain how a market-consistent value of the liabilities might be calculated for a portfolio
of conventional without-profits immediate annuity policies. [6]
[Total 7]
20.3 A life insurer sells term assurances.

Exam style
Discuss the relative merits of using a passive valuation approach rather than an active
approach. [6]
20.4 A life insurer sells level immediate annuities.

Exam style
It has the choice of the following assets to invest in:
 a large issue of low grade, long-term, fixed-interest corporate bonds, issued by a large
domestic multinational company
 a small issue of high grade, long-term, fixed-interest corporate bonds, issued by a small,
private local company
 long-term, fixed-interest government bonds.
The corporate bonds provide a similar expected yield, which is significantly higher than the
government bond yield.
(i) Discuss which bonds would be most appropriate to back the annuity liabilities. [6]
(ii) Describe how the choice of bonds would affect the discount rate used to value the
liabilities in a market-consistent valuation. [4]
[Total 10]

20.5 A life insurance company sells term assurances with term of three years. Premiums are payable
annually in advance. The sum assured of 50,000 is payable at the end of the year of death.
The pricing basis uses the following best estimate assumptions:

Mortality: 0.1% per annum
Interest: 3% per annum
Expenses: initial: 70 incurred at outset
renewal: 5 incurred when second and third premiums are received
claim: 80 incurred when the death benefit is paid
Profit loading: 30
The cost of capital associated with the solvency requirements has been ignored in the pricing
calculation for simplicity.
(i) Show that the premium for this policy is 86.26.
The regulator believes that the following assumptions would represent a prudent estimate of
future experience:
Mortality: 0.13% per annum
Interest: 2% per annum
Expenses: renewal: 6 incurred when second and third premiums are received
claim: 90 incurred when the death benefit is paid
The regulator is prepared to allow companies to calculate reserves on a best estimate basis if they
hold additional solvency capital calculated as 0.091% of the sum at risk.
The contract has just been sold, the premium has been paid and the initial expenses have been
incurred.
(ii) Calculate:
(a) the asset share
(b) the supervisory reserves calculated on the best estimate basis used in pricing
(assuming there is no solvency capital requirement)
(c) the supervisory reserves calculated on the regulator’s prudent basis (assuming
there is no solvency capital requirement)
(d) the sum of the supervisory reserves calculated on the best estimate basis and the
solvency capital requirement.
(iii) Comment on the results of your calculations above.

20.1 Determining the risk margin using the cost of capital approach
Projection of capital
The cost of capital method involves first projecting forward the future capital that the company is
required to hold at the end of each projection period (eg year) during the run-off of the business.
[½]
The projected capital is the amount required to be held in excess of the projected liabilities at
each period. [½]
Calculation of the projected capital might require complex projections, stochastic modelling,
correlation matrices and/or copulas ... [½]
... and should be determined according to the relevant regulatory basis. [½]
However, regulations may permit a simplified approach to be used ... [½]
... which might involve selecting a driver which has an approximately linear relationship with the
required capital ... [½]
... so that the projected capital can then be approximated as a percentage of the projected values
of the driver. [½]
Cost of capital charge
These projected capital amounts are then multiplied by a cost of capital rate. [½]
This rate represents the cost of raising incremental capital in excess of the risk-free rate ... [½]
... or alternatively it represents the frictional cost to the company of locking in this capital to earn
a risk-free rate rather than being able to invest it freely for higher reward. [½]
It can be calculated as the excess of the weighted average cost of capital over the risk-free
rate. [½]
It may be dependent on the time period being considered ... [½]
... or it may be a specified fixed rate (eg 6% per annum in Solvency II). [½]

Overall risk margin
The overall risk margin is then calculated as:

kt  Ct
 risk margin =  [½]
t 1  rt t
 where kt is the cost of capital charge for time t [½]
 Ct is the required capital at time t and [½]
 rt is the risk-free interest rate for maturity t. [½]

[Maximum 6]
20.2 (i) Market-consistent valuation
In theory, a market-consistent value of a liability is the price that someone would charge for
taking responsibility for (ie ownership of) a liability, in a market in which such liabilities are freely
traded. [1]
(ii) Conventional without-profits immediate annuities
First project the net liability cash outgo that would be expected in each future year from the
existing portfolio of policies. In this case, the cashflows will include annuity payments and
expenses. [1]
Assuming the cashflows are certain and guaranteed amounts, the market-consistent value is then
found by discounting the cashflows at current risk-free rates of interest. [½]
The risk-free rate of interest to discount a guaranteed cashflow due in t years’ time would be the
current market redemption yield on a secure (probably government-backed) zero-coupon bond of
the same term. [½]
Alternatively, determine the portfolio of such zero-coupon bonds that could be owned at the
present time that would produce exactly the same future cashflows as the liabilities. (This is
referred to as the ‘replicating portfolio’ of assets.) [½]
The current market value of this portfolio is then the required market-consistent valuation. [½]
(The two approaches are identical.)
The market values of the replicating portfolio of assets (or equivalently the risk-free interest rates)
may need adjustment for any short-term anomalies in the market at the present time. [½]
As the annuities are reasonably long-term predictable liabilities (assuming that the portfolio is
large enough to remove random fluctuations) … [½]
… then (depending on the purpose of the valuation and any regulations) it may be appropriate to
take credit for the illiquidity premium and thereby discount liabilities at a higher yield than the
risk-free rate described above. [½]

However in reality, the liability cashflows are not certain amounts, but are instead subject to
uncertainty. [½]
In this example, uncertainty is due to mortality, inflation, and any other factor that could affect
the real level of future expenses (including future volumes of in-force policies). [½]
If we projected cashflows using best-estimate assumptions for these variables, the value we
would obtain for the liabilities using the above approach would be lower than market value. [½]
This is because a purchaser of the liability would require additional compensation, for the
possibility that the liability should turn out to be more expensive than expected. [½]
To allow for this uncertainty, we would need to include ‘appropriate’ margins in the expense,
inflation and mortality assumptions, compared with best estimate. [½]
Alternatively, an overall reserving margin in respect of these risks could be determined using the
‘cost of capital’ approach. [½]
Some market-based indicators for some of these parameters might be available, and would help
determine the suitable level of margin to use. [½]
For example:
 the difference between the current market yields of equivalent fixed-interest and index-
linked bonds gives an indication of the market’s expectation for future inflation [½]
 there might be firms who provide third-party administration services for life insurers,
whose fees for taking on the admin for the portfolio would be the market value of the
expenses. [½]
But a liquid market for mortality risk does not exist, so some form of risk margin would need to be
added to the best estimate assumption to reflect the compensation required by the market for
taking on the uncertainty. [½]
[Maximum 6]
20.3 Compared to active valuation approaches, passive valuation approaches:
+ tend to be more straightforward to implement [½]
+ tend to involve less subjectivity, … [½]
… eg a net premium valuation is relatively insensitive to the choice of valuation basis [½]
+ tend to result in relatively stable emergence of accounting profit [½]
+ may be less likely to be subject to procyclicality and systemic risk: [½]

 a passive valuation of assets (such as historic book value) would be little affected
by volatile asset prices [½]
 an active approach (such as market value) may result in the need to sell risky
assets after a fall in prices, which could lead to further price falls [½]

– are at risk of becoming out of date, … [½]
… and hence management might fail to take appropriate actions in time; … [½]
… in particular, they might fail to take account of important trends, … [½]
… such as increasing mortality rates due to an epidemic … [½]
… or increasing inflation of expenses [½]
– may provide a false sense of security when the reality is that market conditions have
changed significantly [½]
– tend to be less informative in terms of understanding the impact of market conditions on

the ability of the company to meet its obligations ... [½]
… particularly in relation to financial guarantees and options … [½]
… although it is unlikely that a term assurance policy would include any financial options
and guarantees, ... [½]
… and it is unlikely to be exposed to the equity market … [½]
… although it may be exposed to price volatility if it holds corporate bonds. [½]

[Maximum 6]
20.4 (i) Appropriate matching bonds
The annuities are level, and all three of the bonds are fixed interest, so they all provide a good
match by nature. [½]
Immediate annuity payments will be highest in the short term (when the majority of policyholders
are still alive) but will also include some very long term payments … [½]
… so all three of the bonds should be appropriate to match the longer term payments … [½]
… but bonds with a range of short and medium terms will also be required. [½]
The yield margin between the government and corporate bonds reflects the higher risk of
default … [½]
… and lower liquidity of the corporate bonds. [½]
The government bonds should have the lowest risk of default, … [½]
… so these may be the preferred option, especially if the insurer is risk-averse. [½]
The low grade bonds have the highest risk of default, … [½]
… so these may be deemed inappropriate. [½]

The high grade bonds should have a more acceptable level of default risk, … [½]
… and the greater return provided by corporate bonds may make these appealing enough to
choose over the government bonds. [½]
If the insurer has matched its assets and liabilities (with bonds with a variety) of terms, then it
would not be expecting to sell the bonds before maturity, and so liquidity is not an essential
feature of the assets. [½]
The government bonds should be the most marketable / liquid. [½]
The large issue of low grade company bonds should also be fairly marketable / liquid. [½]
However, the small issue of small, private company bonds are likely to be considerably less
marketable / liquid, … [½]
… and so a significant part of their yield will be to compensate for this illiquidity. [½]
As liquidity is a feature that is not needed by this insurer if it is well matched, these bonds may
therefore be appropriate, … [½]
… as the insurer will be able to benefit from the higher yield without taking on a significant level
of risk. [½]
However, if the insurer only invests in these three long-term bonds, then it will need to sell some
of the bonds before maturity to pay benefits in the early years … [½]
… so at least some of assets should be held in the more marketable and liquid government bonds.
[½]
[Maximum 6]
(ii) How the choice of bonds would affect the discount rate used for the liabilities
In a market-consistent valuation, the discount rate would usually be the risk-free rate obtainable
on government bonds regardless of the choice of assets. [½]
However, an illiquidity premium can sometimes be included in the discount rate used for the
liabilities to take credit for the illiquidity premium in the yield on the assets held. [½]
A higher discount rate will lead to a lower value of the liabilities, … [½]
… and so a better solvency position. [½]
If the insurer has chosen to hold the government bonds, it will not be able to include an illiquidity
premium in its discount rate. [½]
If the insurer has chosen to hold either of the corporate bonds, then it may be able to include an
illiquidity premium, … [½]
… if this is permitted by regulation / legislation. [½]

It is only generally appropriate to include an illiquidity premium for long-term, predictable

liabilities … [½]
… which will be the case here if the insurer has a large enough portfolio to remove random
fluctuations from the experience. [½]
The corporate bonds have similar yields overall, however, they are made up quite differently: [½]
 the multinational company’s bonds have a high risk of default, but are very marketable /
liquid, … [½]
… so the majority of the yield margin (above the government bond yield) will reflect the
default risk [½]
 the small private company’s bonds have a lower risk of default, but are very unmarketable
/ illiquid, … [½]
… so the majority of the yield margin (above the government bond yield) will reflect the
illiquidity risk. [½]
Therefore, if the insurer held the multinational company’s bonds, then the scope for including an
illiquidity premium would be small, … [½]
… whereas if the insurer held the small private company’s bonds, then a significant allowance may
be made in the discount rate for illiquidity. [½]
[Maximum 4]
20.5 (i) Premium
We calculate the premium P by equating the expected present value of the premiums with the
expected present value of the claims, expenses and profit loading.
  
P 1  0.999 v  0.9992 v 2  50,080  0.001 v  0.999 v 2  0.9992 v 3 
 
70  5 0.999 v  0.9992 v 2  30 @3%
2.91061 P  141.51804  79.55307  30
P  86.26
The question has made no reference to lapses, so we have ignored them here (although in practice
an allowance for lapses would be made).
(ii)(a) Asset share
The premium of 86.26 has just been received and the expenses of 70 have just been paid
(assuming that the actual expense experience was the same as the best estimate). So the asset
share is 16.26.
We would often have a negative asset share at the outset of the policy.

(ii)(b) Supervisory reserve (best estimate basis)

V  50,080  0.001 v  0.999 v 2  0.9992 v 3 
  
5 0.999 v  0.9992 v 2  P 0.999 v  0.9992 v 2  @3%
V  141.51804  9.55307  164.80961
V  13.74
The reserve is negative as is often the case when calculating a best estimate reserve for a term
assurance.
(ii)(c) Supervisory reserve (prudent basis)

V  50,090  0.0013 v  0.9987 v 2  0.99872 v 3 
  
6 0.9987 v  0.99872 v 2  P 0.9987 v  0.99872 v 2  @2%
V  187.54903  11.62673  167.15368
V  32.02
The reserve is positive as is usually the case when calculating a prudent reserve.
(ii)(d) Supervisory reserve (best estimate basis) plus solvency capital
The best estimate reserves are  13.74 as calculated earlier.
The solvency capital requirement is:
SCR  0.00091  50,000   13.74  
SCR  45.51
So the total solvency requirement is 31.77 (45.51  13.74).
(iii) Comments on results
When the company writes this policy its assets increase by 16.26 (ie the asset share) and its
reserves decrease by 13.74 if we use best estimate assumptions. So the surplus of the company
(assets over liabilities) increases by 30 due to writing this policy. So if we calculate reserves on a
best estimate basis (perhaps for accounting or internal management purposes) we can see that a
profit of 30 occurs at outset (which is exactly the profit we priced for).
However, the situation is quite different if we calculate reserves on a prudent basis. When the
company writes this policy its assets increase by 16.26 (ie the asset share) and its reserves
increase by 32.02. So the surplus of the company (assets over liabilities) decreases by 15.76 due
to writing this policy. The company now has new business strain, ie the premium of 86.26 is not
enough to cover the initial reserves of 32.02 and the initial expenses of 70.

The total solvency requirement is approximately 32 under both the prudent reserving approach
and the best estimate plus solvency capital approach (of course we cheated in our choice of basis
to make sure that this was the case). So both approaches give the same level of protection to the
policyholders (they both require that the company sets aside an extra 45.5 over and above the
best estimate cost of paying the liabilities) and have the same new business strain.

SP2-21: Surrender values Page 1
Surrender values
Syllabus objectives

alteration terms.

Page 2 SP2-21: Surrender values
0 Introduction
We have already described the various forms of discontinuance (and withdrawal) in Chapter 1 –
have another look at this again if you need reminding before going any further. In this chapter we
discuss how surrender values are calculated. In Chapter 22 we show how we can go about
calculating revised terms for policy alterations, such as increases to benefit levels, or reductions in
future premiums.
We use some of the equation of value formula techniques for valuing benefits that you will have
studied in Subject CM1 or CT5: you may wish to revisit some of this material if it is a long time
since you studied it or you are not familiar with it from your day-to-day work. The gross premium
reserve covered in Chapter 19 is also relevant.
Sections 1 to 5 consider conventional without-profits contracts. Section 6 briefly covers unit-

linked surrender values (these were also covered in Chapter 4).
Finally, both the Core Reading and the ActEd text use the term premium basis with particular
connotations. The phrase is used to refer to a basis that includes implicit margins.

1 Surrender values for conventional without-profits contracts

A policyholder may wish to surrender a policy, ie terminate it early in return for an
immediate cash payment (the surrender value).
To assist in setting surrender values, it is useful to consider both the retrospective asset share and
the prospective reserve at the point of surrender.
As described in Chapter 1, when the surrender value is more than the asset share, the company
makes an overall loss on the policy. For this reason, where possible, surrender values will be less
than the asset share at the time of surrender.
The most important reserve is the supervisory reserve.
Question
Explain whether the surrender value would be expected to be bigger than, or smaller than, the
supervisory reserve.
Solution
The answer depends on the rules that apply in the country concerned. However, in many
jurisdictions the supervisory authorities require an insurance company to hold at least the total
amount of surrender value in reserve so that it can cover the (real) possibility of mass surrender
occurring over the short-term future.
If a conventional without-profits policy is surrendered there will no longer be a need to hold

a reserve for that policy, and a surrender value can be paid. Reserves for whole life and
endowment assurances increase (towards the sum assured) with increasing policy
duration, but those for term assurances are always very small compared to the sum
assured. This explains why, while surrender values are usually paid for whole life and
endowment assurances, they are not usually paid for term assurances. (The non-payment
of a surrender value for term assurances also helps with potential losses from the
anti-selective effect of surrenders and allows cheaper premiums to be offered.)
We have already covered this point fully in Chapter 2. So let’s see what else you can remember.
Question
The two main reasons for having no surrender value under term assurances are low asset shares
and the cost of selective withdrawals.
State three other reasons.

Solution
The possible reasons are:

 to recoup losses on early lapses (when the asset share is negative) by making some profit
on later lapses (when the asset share is usually positive)
 asset shares are quite volatile, so it would be difficult to devise a surrender value scale
that would treat policyholders fairly in relation to this
 asset shares can be negative at later durations – and/or are likely to be decreasing
towards the end of the policy – but it would be hard to sell the idea of decreasing
surrender values to policyholders, who may become disgruntled as a result.
Question
Discuss whether it would it be reasonable to offer surrender values for immediate annuities.
Solution
This would be very difficult. A policyholder is most likely to want to surrender an immediate
annuity if he or she thinks that they might die in the near future. So if we give surrender values to
these, the average mortality experience of those who remain will improve, which could be a big
cost to the company. So in general, life companies have not offered surrender values on
immediate annuity policies. If surrender values were offered they would need to be underwritten
in some way (giving lower surrender values to those in poor health).
Another reason is that surrender might not be allowed by legislation. For example, tax privileges
are often granted for certain kinds of policy that achieve certain ambitions for the government, in
particular policies that provide a long-term annuity income in retirement so as to relieve some of
the financial burden on the State. If such annuitants were allowed to take lump-sum withdrawal
benefits, people could use the policies as savings vehicles, benefit from the tax concessions, but
not further the government’s aims to encourage private pension provision. For this reason, such
surrenders have often been forbidden.
This chapter describes methods of calculation of surrender values for conventional

without-profits products.

2 Principles for surrender values

In determining how to calculate surrender values, the following principles should be
considered.
Surrender values should:

 take into account policyholders’ reasonable expectations
 treat both surrendering and continuing policyholders equitably
 at early durations, not appear too low compared with premiums paid, taking into
account any projections given at new business stage
 at later durations, be consistent with projected maturity values
 not exceed earned asset shares, in aggregate, over a reasonable time period
 take account of surrender values offered by competitors (and possibly also auction
values, where available)
 not be subject to frequent change, unless dictated by financial conditions
 not be subject to significant discontinuities by duration
 not be excessively complicated to calculate, taking into account the computing
power available
 be capable of being documented clearly
 avoid selection against the insurer.
These principles are covered in more detail below.
2.1 Policyholder expectations

The amount paid on surrender is normally at the discretion of the insurer. Legislation may exist to
ensure that policyholders are treated fairly and that policyholders’ reasonable expectations (PRE)
are met.
There is no generally accepted definition of PRE, but they will be influenced by, for example,
the past practice of a company and any literature issued by it.
Discontinuance at short durations

At short durations, it is natural for a policyholder to compare the surrender value with the
premiums paid, or even premiums plus some interest. However, the asset share at such
durations will usually be less than this. A prospective policy value based on best estimates of
future experience is likely to be even smaller.
In the first year, and sometimes beyond that, both the asset share and the prospective value
might well be negative. If the asset share is negative then clearly the company cannot avoid
making a loss. However, it may feel obliged to accept a loss, or at the least a reduced profit, on
surrender up to several years into the contract so that the surrender value paid does not appear
too low compared with premiums paid.

One way of coping with this situation is to try to recoup the cost from later surrenders, ie penalise
later surrenders to pay early surrenders more than they warrant. Another and rather draconian
way is not to offer surrenders at all for some initial period (eg the first two years), which may in
fact appear to policyholders less harsh than offering some very small value.
Discontinuance close to maturity

Where a benefit will become payable on the maturity of a contract, the policyholder will expect
that a surrender value payable prior to maturity will be consistent with this. So, the surrender
value payable at the end of each policy year should progress smoothly into the maturity value at
the end of the contract.
For without-profits business, this will be achieved by applying a surrender value derived from
prospective policy values.
2.2 The asset share

The Core Reading in this chapter refers to earned asset share which we will abbreviate to asset
share throughout the ActEd text. The two terms can be used interchangeably for exam purposes.
One of the starting points in considering what surrender value can be offered is the asset share of
the policy. The calculation of asset shares was covered in Chapter 5.
The asset share represents the money that the company has really accumulated in respect of any
policy, unlike the supervisory reserves, which represent how much money the company must hold
in respect of a policy.
The asset share does not define the unique value the company could afford in all cases since it
could give some policies a little more than this value and others a little less. However, in total it
should represent the maximum the company can afford to pay out measured over a reasonable
period of time.
Basing the surrender value closely on an asset share implies distributing accrued profits (or losses)
to the policyholder. This is because the asset share contains all of the accumulated profits (or
losses) from the policy to date.
Averaging over time can be achieved in two ways:

 First, the company might do its asset share calculations only once per year (for instance)
for reasons of practicality. In this case it is averaging over the year, albeit implicitly.
 A second way is to use a smoothed asset share as the basis for surrender values. The
company then needs to decide what time period to smooth over. For example, does it
want to smooth individual months’ asset share values so that the total impact of
smoothing over 12 months is zero, or does it want to smooth individual years’ asset share
values so that the impact over perhaps five years is zero?

2.3 New business disclosure and other competitive considerations

Insurers may provide prospective policyholders with illustrations of possible surrender values at
various durations. In some countries this is a regulatory requirement. Such illustrations must
then be borne in mind when eventually determining the surrender value.
The financial press may publish from time to time tables showing the surrender values offered by
different life insurance companies. Each company will need to decide to what extent it should be
influenced by any, possibly adverse, comparisons resulting from the use of such tables.
It is quite reasonable for a company to want to be seen to offer competitive surrender values as
well as competitive maturity terms.
More generally, it could be embarrassing for a company if any comment on surrender values in its
literature conflicted with figures shown in financial press surveys or actually quoted to
policyholders.
The auction value

The auction value of a policy is the value it would fetch if the policyholder were to transfer it as an
ongoing policy to someone else (eg Chapter 3 mentioned the possibility that policyholders could
sell their annuity to the secondary market in this way). Such transactions are usually dealt with by
specialised brokers.
This value has the advantage of being an independent assessment and is perhaps more likely to
result in a value which the policyholder would have to accept as fair.
However, it would generally be unsuitable from the company’s point of view because:
 The underlying assumptions for auction values would probably be different from the
company’s own assumptions. For example, they may take a more optimistic view of the
future benefits under the contracts than the company would consider appropriate.
 The value may fluctuate unpredictably, and may be difficult to determine without actually
offering the policy up for sale.
If the policyholder is dissatisfied with the surrender value quoted by the company, they can
always attempt to obtain a higher value in the second-hand market.
2.4 Change in surrender scales over time and by duration

For practical reasons, scales of surrender values should not change too frequently unless financial
conditions would dictate this.
Surrender scales should not contain discontinuities by duration to maintain equity between
policyholders who surrender on either side of the discontinuity. For example, it would be unfair if
the surrender value one week after a policy anniversary was significantly higher than for the week
before the anniversary (after allowing for any premium paid on the anniversary).

2.5 Ease of application

The surrender terms should not be excessively complicated to calculate, taking into account the
computing power available.
By surrender terms, we mean a simple formula, or table of applicable factors, which will allow
mass production of surrender values by administrative staff, probably using some computer
application. This is likely to require some simplification and smoothing of the ‘real’ situation. This
smoothing implies some cross-subsidy. The actuary must ensure that such smoothing is not
overly detrimental either to policyholders or to the company.
2.6 Selection against the company

The surrender terms should minimise the risk of policyholders selecting against the company. In
fact the potential cost of such selection is sometimes so high that it is better not to offer
discontinuance terms at all. This is typically the case with immediate annuity business.
Another opportunity for selection against the life company may occur after a market crash,
especially for single premium business. To prevent this, the company would need to revise terms
immediately after any such movement in the markets.
2.7 Surrender and re-entry

The company should ensure that surrender terms, when looked at in conjunction with the current
premium rates for that product, do not make it advantageous for the policyholder to discontinue
his or her policy and then take out a new policy. This could be thought of as a special case of
selection against the company, since the policyholder is taking advantage of a particular
circumstance to cause the company to suffer costs it would not otherwise have, in respect of
setting up the new contract.

3 Methods of calculation
The calculation of surrender values is illustrated using the example of a whole life
assurance for a sum assured of S with annual office premiums of G payable m times a year.
The notation used is:
x age of policyholder at date of issue
I initial expenses in excess of those occurring regularly each year
e level annual expenses
f normal claims expenses.
Surrender values for endowment assurance policies can be determined in a similar way.
3.1 The retrospective method

Provided sufficient information and computing facilities are available, a company may
decide to keep an up-to-date earned asset share value on file for each policy and use that to
deduce a surrender value. However, this is not usual practice for without-profits business.
It might therefore decide to calculate a retrospective reserve using a formula and parameter
values chosen to produce acceptable results at the durations required. The starting point
for the basis (mortality, interest, expenses) would be the experience of the policy.
Allowance should also be made for the costs of surrender, C, ie
Dx
Dx t 
( m )  SA 1  I  ea
Ga
x : t x : t
(m )  fA 1
x :t x: t  C
Question
Outline the disadvantages of this asset share method for without-profits business.
Solution
Basically for without-profits, no profit is retained by the insurer. Also, it could be inconsistent
with the maturity value, negative early on and hard to calculate.
This is explored in more detail in Section 4.
The ‘up-to-date asset share for each policy’ may be held, like a unit account, on a policy by policy
basis. However, it could also be based on an accurate value for specimen policies.
See Section 4.1 for a further general discussion of the retrospective method.

3.2 The prospective method

For without-profits contracts, this is the value of future benefits and expenses, net of future
premiums due, using estimates of future expected investment returns, expenses and
mortality experience of the surrendering policyholders. Costs of surrender, C, should also
be allowed for:
(m )  fAx  t  Ga
SAx  t  ea (m )  C
x t x t
See Section 4.2 for a discussion of the method.
Question
Calculate the surrender value appropriate for the following policy, using the prospective method
and the basis defined below:
 a without-profits endowment assurance contract for a 45 year old with five years to
maturity
 sum insured £20,000, payable immediately on death
 premiums £1,733, payable annually in advance
 renewal expenses at 6% of premium
 surrender expenses of £50.
Basis and assumptions:

 mortality as under the AM92 table
 only healthy lives surrender, ie they are select lives at the date of withdrawal
 interest rate 4% pa.
Solution
The surrender value will be
20,000 A  0.94  1,733  

a  50
[45]:5| [45]:5|
where
 D  D
A  1.02  A[45]  50 A50   50  0.82260
[45]:5|  D[45]  D[45]
 
D50

a  
a[45]  a50  4.6172

[45]:5| D[45]
giving a result of £8,880.

4 Analysis of the methods
4.1 The retrospective method

The retrospective value will represent the earned asset share at the date of surrender or an
estimate thereof. Therefore, it would represent the maximum that the company could pay
without making a loss. At early durations it should not look too unreasonable compared
with the premiums paid.
However, the retrospective method is likely to produce negative values (and hence zero
surrender values) for regular premium contracts at very short durations, depending on the
relationship between premium, initial expenses and commission.
So the asset share will only look reasonable (assuming it’s positive) if policyholders find the
deduction for initial expenses acceptable.
For without-profits contracts it has a number of other disadvantages including:

 It does not say anything about the profit the company would have made if the
contract were not surrendered. Hence it is not easy to ensure equity either with
continuing policyholders or with any shareholders.
 Except by chance, the surrender value will not run into the maturity value.
That is, as duration in-force approaches the full term, the surrender value will only by
chance approach the maturity value.
Developing the idea in the first bullet point, the amount of profit made if a policy is not
surrendered is the difference between the final maturity value and the final asset share. If the
asset share is paid as the surrender value, no profit is made. If the asset share minus a certain
amount (eg £100) is paid on surrender, a profit of that certain amount (£100) is made. However,
the present value of this profit may be more or less than the present value of profit that would
have been made had the contract not been surrendered.
Because the method does not take into account future benefits and expected future
experience, it could produce surrender values that are significantly different from a realistic
prospective value. The latter is likely to be the approach used to calculate auction values.
The problem here is that if the values are very much less than auction values, the company may
get a name for giving poor surrender values in the market.
Whether the method produces values that compare well with those of competitors depends
on the method – and probably more importantly the basis – that they use.
If the company adopts a retrospective method to calculate some or all of its specimen
surrender values in product disclosure literature, then policyholders would reasonably
expect it to do the same thing for its actual values.
The most complex component of the method is obtaining the necessary historic information
to build up earned asset shares or to determine suitable parameters if a formula is used.
The asset share will not tend to the maturity value if the guaranteed amount at maturity is
greater than the asset share.

4.2 The prospective method

If a realistic basis is used with this method, it will produce a surrender value that represents
what the contract is worth to the company. In the context of without-profits contracts, it
enables the company to quantify how much profit to retain and hence maintain equity with
continuing policyholders and any shareholders.
There is no guarantee that the surrender values produced will not consistently exceed the
earned asset share. Moreover, depending on the basis used, the method could produce
surrender values at early durations that look distinctly unreasonable from the
(surrendering) policyholder’s point of view.
For example, a realistic basis for a without-profits contract may produce very low or negative
surrender values early in the term of the contract.
Question
Explain how a prospective calculation could produce an early value that appears distinctly
unreasonable from the company’s point of view.
Solution
An example would be a calculation with a very low interest rate. This could give large values far in
excess of the asset share. (Although such a calculation would probably not be realistic.)
For without-profits contracts, the surrender values will run into the maturity value.
The prospective method is more likely than the retrospective method to produce
comparable surrender values to those available at auction and for comparable competitors’
contracts, since it is the method that is normally used. However, the assumptions used
with the method – rather than the method itself – are likely to be the more important factor
in any such comparisons.
Question
Explain why auction values are likely to be calculated using a prospective method.
Solution
Auction values are set by external organisations and people without access to the confidential
information necessary to calculate an asset share.
Also, a prospective buyer will be interested in the future return on the policy, rather than what
someone else has paid into the policy, in the past.
Unless the company decides to use a complicated basis, it is relatively easy to operate
since it does not require any knowledge of what has happened in the past.

5 Calculation of values
5.1 Choice of method

A table of surrender values by policy duration is usually a blend of the retrospective and
prospective values, subject to a minimum value of zero.
In the early years, a company will need to pay particular attention to the actual expenses
incurred. Other factors are of lesser importance in the short term. A retrospective
calculation is likely to be the more natural approach.
Premiums paid are also, of course, critical to the amount of early asset shares. This means, for
example, that large cases, or annual premium cases, will support positive surrender values earlier
than small cases or monthly premium cases.
In theory the retrospective approach could be used throughout, but if the calculation had to
be done by devising some independent formula, and setting parameter values for it, then it
is likely to become increasingly difficult to find the right combination of factors to produce
values which run into the maturity value.
After the early years, a prospective method is more convenient since it is only necessary to
value future benefits, premiums and expenses. The main difficulty with a prospective
calculation is deciding on the appropriate rate of interest. This makes the approach difficult
to apply at short durations – and hence long outstanding term – because even very small
changes in the rate of interest will have a significant effect on the surrender value.
5.2 Retention of profit

For the purpose of the discussion here it will be assumed that a company has decided to
use a retrospective method at early durations and then merge this into a prospective
method using a suitable basis.
The profit retained by the company is the excess of the earned asset share over the
surrender value paid.
If the surrender value is taken as the retrospective value, the company will not be retaining
any profit and it will probably not want this to continue for too long.
If it adopts a prospective approach then the amount of the profit depends upon the
relationship between the assumptions used for the surrender value and those implicit in the
calculation of the office premium. Suppose in this example, that the allowance for profit is
contained solely in margins in the assumptions used to calculate the premium. The profit
retained can then be specified as:
(EAS – SV  ) + ( SV  – SV  )
where: EAS = earned asset share

SV  = prospective surrender value using same assumptions as used
to calculate the office premium
SV  = prospective surrender value using surrender value basis
assumptions.

The first part – (EAS – SV  ) – represents the profit that has been made to date.The second
part – ( SV  – SV  ) – represents the capitalised value of the profit that will arise in future
from the differences between the premium rate assumptions and the prospective surrender
value assumptions.
If these surrender value assumptions exactly represent future experience, then the total
profit retained will be the same as if the contract had not been surrendered. If they are the
same as the premium basis assumptions, then the company only retains the profit it has
made to date.
By a suitable choice of basis – between its best estimate of future experience and its
premium basis assumptions – the company can adjust how much profit it retains in
accordance with its corporate aims.
One possible approach is to use a blended basis that starts with the premium basis
assumptions near to entry and runs into the best estimate assumptions nearer to maturity.
How soon it runs into the best estimate assumptions will depend upon how quickly it feels it
can start retaining the same profit as from a non-surrendered contract.
However, care must be taken that the basis chosen does not lead to a lapse and re-entry
option arising. Checking that the surrender value does not exceed the earned asset share
should avoid this.
The concepts explored above are crucial for understanding a number of issues to do with policy
values, profit and surrender values, so should be read and reread carefully. The following pages
include various items to consolidate the technical points above.

The graph below illustrates a number of the key ideas from the previous page.
14,000
12,000
Final
profit
10,000
Sum assured
8,000 Asset share
Realistic basis
6,000
£ Premium basis
4,000 Profit A
2,000 Profit B
-2,000
-4,000
Duration
It shows the behaviour in certain circumstances of three things for a without-profits endowment
assurance with a sum assured of £10,000: the asset share, a realistic prospective value and a
premium basis prospective value.
The difference between the asset share and the premium basis value gradually grows because of
the margins in the premium basis, which is why the Core Reading refers to this as a sensible
measure of ‘profit made to date’. (We have called this ‘Profit A’ in the diagram.) It is one such
measure, although we have seen other measures such as the change in embedded value.
‘Profit B’ is the rest of the profit expected on the contract (in present value terms). If the
company uses a prospective premium basis (basis with margins) then it gives up this ‘future
accruing profit’.
The graph shows well why a realistic prospective value is a poor way of calculating early surrender
values, although it does also show that even the asset share may not be enough very early on,
when it may be negative.
A suitable sequence of surrender values might, as indicated in the Core Reading, start close to the
asset share or even a prospective value on premium basis assumptions (subject to avoiding the
very low early values) and then move towards the realistic prospective calculation late on. Such a
sequence is shown below.

14,000
12,000
10,000
Sum assured
8,000 Asset share
Realistic basis
6,000 Premium basis
£
Surrender value
4,000
2,000
0
1
-2,000
-4,000
Duration
Here, in addition to the lines shown on the previous graph, we show a possible sequence of
surrender values, the linked sequence of asterisks.
The ideas behind the sequence of surrender values are that initially, no surrender value is
allowed. The surrender value is then subject to a minimum of premiums paid.
It is then increased in line with the asset share until the company deems it reasonable to start
making a profit.
Finally, the amount of profit taken is increased until close to maturity, the surrendering contract
gives as much profit as the maturing contract.
Question
It might be argued that the graph above shows an unusually long period of negative asset share.
State the types of regular premium savings policies that are likely to have a relatively short period
of negative asset share.
Solution
For larger policies (initial expenses less significant) or shorter terms (lower proportionate
commission), the period of negative asset shares will be shorter than for smaller policies or longer
terms.

As mentioned in the commentary to the first graph, the graphs and the ideas expressed in the
Core Reading apply most directly only in very specific circumstances. These are that the realistic
assumptions actually correspond exactly with reality as it unfolds. If this demanding condition is
not met, things get more complex.
Consider situations 1 to 4 for regular premium without-profits endowments.
Situation 1: Interest rates rise suddenly late on in the term of a without-profits endowment.
Matched asset share drops. However, realistic prospective value on new higher
best estimate interest rates also drops by enough to retain same profit.
(Unchanged) value on original premium basis now unsuitable.
Situation 2: Interest rates fall suddenly late on. Matched asset share rises. Realistic
prospective value rises.
Situation 3: As 2 except invested in cash. Mismatched asset share doesn’t rise. Can’t afford
to pay realistic prospective value.
Situation 4: Interest rate change early on. Probably can’t be matched yet using conventional
assets so profit will probably be changed by the event.
Situation 5: Early interest rate change for single premium policy. Can be matched even very
early on.
5.3 Determining a basis for retrospective value

If a retrospective method is used in the early years the company will need to examine its
actual past experience for all the relevant factors, which will include investment earnings,
expenses, mortality and (possibly) tax.
For regular premium contracts, it may not follow past experience exactly – particularly with
regard to investment earnings – so as to smooth the value.
Question
The Core Reading suggests that smoothing investment earnings is more likely for regular premium
contracts than for single premium contracts.
Solution
Policyholders are more likely to be exercising financial selection against the company when
purchasing a single premium contract. It should therefore be longer before a company considers
letting them benefit from investment smoothing.

Something other than past experience may be used for other reasons. For example, an
investment return lower than actual might be used to create a retrospective value lower than the
‘true’ asset share if the company wanted to retain some profit. (This assumes the asset share is
positive). Alternatively, expenses lower than actual might be deducted if the ‘true’ asset share
were to give an embarrassingly poor surrender value. Competitive considerations are also likely
to be important here.
5.4 Determining a basis for prospective value

Assumptions will be needed for interest, expenses, expense inflation and mortality. It might
also be necessary to make assumptions in respect of tax.
Interest
This will be the most important assumption.
Companies will probably cover their without-profits liabilities with fixed-interest

investments, perhaps chosen to give a reasonable matching by term, assuming suitable
securities exist for that purpose. Hence, a company might consider the current weighted
average redemption yield of suitable securities to be its best estimate assumption.
It may also consider the interest rate used in the premium basis if it wishes to use a blended
basis.
Expenses
The company’s most recent expense investigation will indicate the level of renewal
expenses, which may well be the same as those used to set current premium rates. It is
unlikely that any margins would be added to the assumption as these would increase the
surrender value.
Allowance needs to be made for any renewal commission as paid.
Allowance also needs to be made for the expenses involved in surrendering the policy.
Inflation
The inflation rate will probably be chosen to be consistent with the investment return
assumption. The real return anticipated on index-linked government stock will give an
indication of what might be a suitable margin below the full interest rate assumption.
Mortality
The mortality basis chosen should reflect the future expected mortality of those
policyholders who are surrendering. It would be reasonable to assume that such people
will experience lighter mortality than those who do not, since anyone conscious of their ill-
health would not wish to discontinue their life cover. However, in practice there is little
evidence of such selection.
(This is not surprising really – life companies are not in a position to monitor the mortality of
surrendered policyholders.)

A company could either make some explicit allowance for extra-light mortality or use its
up-to-date experience and rely on other margins in the basis for any difference which might
have emerged had those policies not been surrendered. Select mortality could be used, but
the nature of the selective process here is not the same as that underlying a standard
mortality table.
For most assurance contracts that have surrender values, the mortality assumption will not
in any case have much effect upon the resulting values.
Crucially, contracts like term assurance, where mortality would be critical, do not normally have
surrender values.

The terms for surrender and alteration of unit-linked contracts are normally specified at the
outset.
For unit-linked business, the surrender value will typically be the bid value of the units less
any surrender penalty which applies. The surrender penalty may be a percentage of the unit
value or a percentage of the premium, but will be designed to recoup initial expenses.
Any surrender penalty tends to be relatively large at early policy durations, decreasing over time.
Question
State under what circumstances the following would be expected:
(i) a unit-linked policy with no surrender penalty
(ii) a unit-linked policy with a surrender penalty expressed as a monetary deduction.
Solution
(i) No penalty where there are high initial charges under the policy.
(ii) Monetary penalty where there are reduced level allocation rates (< 100%) for several
years.
The general reasoning behind the calculation of these surrender penalties has been covered
already in Chapter 4. We have also discussed the need for a surrender penalty in association with
negative non-unit reserves in Chapter 19.

Chapter 21 Summary
Conventional without-profits surrender values
The factors to consider in determining without-profits surrender values are:
 policyholders’ reasonable expectations
 treating surrendering and continuing policyholders equitably
 at early durations, premiums paid and illustrations given to policyholders
 at later durations, projected maturity values
 asset shares
 competition and auction values
 avoid frequent changes
 avoid discontinuities by duration
 ease of calculation
 capable of clear documentation
 potential selection against the insurer.
Without-profits values should produce adequate profit.
Surrender values can be calculated retrospectively or prospectively.
The retrospective method is most suitable early on for without-profits contracts.
The prospective method is most suitable for without-profits contracts after the early stages.
Also, a prospective method may in fact be used for other situations to achieve results tested
against a retrospective calculation.
Unit-linked surrender values

For unit-linked business, the surrender value will typically be the bid value of the units less
any surrender penalty that applies. The surrender penalty may be a percentage of the unit
value or a percentage of the premium.



21.1 A life insurance company is currently reviewing its surrender values for regular premium
Exam style
without-profits endowment assurances.
(i) Describe the most important objectives of the company when determining surrender
values. [5]
(ii) Propose with reasons a suitable approach to calculating surrender values. [8]
(iii) Describe:
 the problems caused by a sharp increase in interest rates in the context of
surrendering policies
 how the approach recommended in part (ii) would cope with such problems. [5]
[Total 18]



21.1 (i) Objectives when setting surrender values
The objectives of a surrender value basis are to:

 minimise the risk of loss due to surrenders [½]
 make an appropriate contribution to profit, if possible and consistent with the other
objectives [½]
 be fair to the leaver and to continuing policyholders [½]
 meet policyholders’ reasonable expectations [½]
 be competitive (with other companies and possibly auction values) [½]
 compare favourably with premiums paid [½]
 blend smoothly into maturity values [½]
 be consistent with other benefits (eg paid-up values) [½]
 be not subject to frequent change [½]
 avoid significant discontinuities by duration [½]
 be reasonably simple to calculate [½]
 be easy to document clearly. [½]
[Maximum 5]
(ii) Approach to calculating surrender values
Surrender values would normally be set with reference to three values:
Asset share
The company would not want to pay more than asset share. [½]
In fact it will want to pay less in order to capitalise the profits accrued to date. [½]
To avoid volatile surrender value terms, need to consider a smoothed asset share. [½]
This is essentially a retrospective policy value calculated on a smoothed actual experience basis.
Using this as an upper ceiling protects the company from making an overall loss from the
contract. [½]
The main reason for loss will be heavy initial expenses compared with those implicit in the
premium basis. [½]
And perhaps also investment returns to date below the premium basis guarantee. [½]

Prospective realistic reserve
Using the prospective realistic reserve allows the company to capitalise all the profits it is
expecting to make on the contract, both past and future. [½]
This should more than compensate for losses on the remaining portfolio from the deleterious
mortality effect of selective withdrawals. [½]
Should be on a best estimate basis with no margins. [½]
However, capitalising future profits may be considered too unfair to surrendering policyholders,
so a basis nearer to the premium basis may be preferred. [½]
Surrender administration expenses
The company will want to be sure that the cost of administering the surrender is borne by the
surrendering policyholder, not by the company or by other policyholders. [½]
Overall
These values will be used as follows:
Use retrospective values at early durations, blending into prospective values at later durations. [½]
In either case, the surrender expenses would be deducted. [½]
This value may then need to be modified in respect of the following:

 For surrenders at short duration we might need to pay more than indicated by the above
calculation in order to compare more favourably with the premiums paid … [½]
 … and to avoid disappointing policyholders and negative publicity. [½]
 Should also allow for policyholders’ reasonable expectations, in particular as regards any
illustrative surrender values quoted at the sales stage. [½]
 Competitive considerations – if the company wants to compete on surrender values
(eg because the financial press focuses on these) then it may need to increase surrender
values. [½]
 For surrenders near maturity, need to ensure that surrender values are in line with the
maturity value (should be OK if prospective value is used at this point). [½]
 Need to ensure that surrender values are consistent with paid-up values. [½]
The basis, including such modifications, needs to be expressed in a way that will be easy to apply,
probably using a computer package. [½]
[Maximum 8]

(iii) Increase in interest rates
Earned asset share may fall as the market value of assets will fall. [½]
This will lead to loss / less profit unless surrender values are appropriately reduced. [½]
Loss from surrender and re-entry may now occur as current premium rates may be lower. [½]
Provided assets and liabilities were matched, … [½]
… the realistic prospective method should avoid loss by valuing future benefits at the new higher
market-related interest rate. [½]
(This will only be effective if the company can change its surrender value basis at this time, … [½]
… bearing in mind the aim also not to change basis too frequently, the admin cost of the change,
and policyholders’ reasonable expectations.) [½]
The realistic interest rate should be higher than the interest rate assumed in new premium rates
(which would include a margin for reinvestment risk), thereby avoiding the surrender and re-entry
option. [½]
If the company is mismatched due to being invested too long, and interest rates were to rise, then
the earned asset share would fall by more than the prospective value, even if the latter were
calculated at the new proposed higher interest rate. [1]
If this were the case, the company might want to reduce all surrender values further to try to
recoup the loss it made. [½]
Finally, these problems may be made worse if there are more surrenders caused by the increase
in interest rates eg because of inability to keep up premiums in a situation of economic
recession. [1]
[Maximum 5]





using the course.

SP2-22: Alterations Page 1
Alterations
Syllabus objectives

alteration terms.

Page 2 SP2-22: Alterations
0 Introduction
This chapter covers alterations other than surrender. The ‘simplest’ such alteration is making a
policy paid-up. Some companies allow much more complex alterations, eg changes in term, sum
assured or even a change in type of policy.
Most of this chapter is devoted to conventional without-profits alterations. Section 1 looks at the
different types of alteration that policyholders might request. Section 2 introduces some
principles for carrying out policy alterations. Section 3 describes two methods of performing
policy alterations – the most important of these is the equating policy values method. Section 4
looks at possible basis choices for the equating policy values method.
Section 5 takes a brief look at alterations to unit-linked policies.

1 Alterations of conventional without-profits contracts
1.1 Paid-up contracts

Making a policy paid-up is one example of an alteration that can be made to an existing
policy. As an alternative to the payment of a surrender value, a policyholder can receive a
paid-up value or paid-up sum assured.
This option would be taken by policyholders with regular premium policies who do not want to
continue paying premiums, but still want some eventual benefit. Once a policy has become paid-
up, it becomes to all intents and purposes just another single premium policy.
For conventional without-profits contracts, the terms and conditions of the original policy
remain unchanged, except that the sum assured is reduced to take account of the fact that
no more premiums will be paid. In effect, the value of the policy is being used as a single
premium to purchase an insurance policy at the date on which the original premium
payments ceased.
However, we would not want to charge another round of initial expenses to the policy.
Two considerations might make the bases used for the calculation of paid-up values slightly
different from the bases used for surrender values. Firstly, the costs of making a policy
paid-up may be different from those of paying a surrender value. Secondly, because the
policyholder continues to have a policy in force, the effect of mortality selection may be less
than when policies are surrendered.
Question
Explain whether the mortality assumption used in calculating a paid-up value should be higher or
lower than that used for calculating a surrender value, for a whole life policy.
Solution
Mortality selection should be more intense for the surrender than for the paid-up policy. We
would therefore expect the surrender value basis to use a lighter mortality assumption than the
paid-up basis.
1.2 General alterations

Most life insurance policies are long-term contracts. While at the outset a policyholder may
have had a clear idea of the risks they faced and purchased life insurance policies to cover
these risks, over time a mismatch between the cover provided by these policies and the
current risks faced by the policyholder may evolve. To remove this mismatch,
policyholders may wish to purchase additional policies and/or to alter current policies.

Some examples of common alterations are to:
 change the term of an assurance (this would include changing a whole life
assurance to an endowment assurance)
 alter the sum assured
 alter the premium payable.
While the life insurance company quotes the terms and conditions for a proposed alteration,
it is the policyholder who chooses whether to accept this quotation. It is, therefore,
important that the terms offered are consistent with those offered for other alterations
(eg surrender values, paid-up values) and with the current terms offered for new policies.
There are links between special alterations and the more commonly occurring alterations.
For example:
 a reduction in the term of a policy to zero is equivalent to surrendering a policy
 a reduction in the sum assured of a policy so that no future premiums are required
is equivalent to calculating a paid-up sum assured
 increasing the sum assured is analogous to keeping the original policy and
purchasing an increment policy at current premium rates.
These links are examples of ‘boundary conditions’ which we discuss further in Section 2.2.
This chapter describes possible methods of calculation of paid-up sums assured and other
altered benefits for conventional without-profits products.

2 Principles for without-profits business
2.1 Paid-up sums assured

The overriding principles are fundamentally the same as those that we discussed in relation to
surrender values. So most of the things we talked about in Chapter 21 are also relevant to paid-
up calculations. In addition, we have to consider the issue of consistency with other benefit
payments, which chiefly means the surrender value.
Question
Explain why consistency between surrender values and paid-up values is important.
Solution
The most obvious reason is fairness. It would seem unfair, for example, to treat a surrendering
policyholder more generously than one who is making their policy paid up. All the marketing
implications relating to unfairness therefore apply here.
It also avoids possible problems with selection. For example, if the bases were inconsistent, it
might be possible for a policyholder to make their policy paid up, then surrender immediately
afterwards, and as a result obtain a better surrender value than if they simply surrendered their
original policy in the first place. This would impact unreasonably on the insurer’s profits.
In determining how to calculate paid-up sums assured, the following principles should be
considered.
Paid-up sums assured should:
 be supported by the earned asset share at the date of conversion, on the basis of
expected future experience
 at later durations, be consistent with projected maturity values, allowing for

premiums not received
 be consistent with surrender values, so that the surrender values before and after
conversion are approximately equal.
Question
Explain why the Core Reading particularly emphasises the need for consistency between paid-up
values and normal maturity values.

Solution
This is because the paid-up sum assured and the full sum assured are defined in exactly the same
‘currency’ – ie they are payable at exactly the same times and under exactly the same
circumstances. The fact that they are directly comparable makes it essential that the differences
between them appear reasonable to policyholders.
In addition, the company’s expected profit from a paid-up policy should be consistent with that
expected from a non-paid-up contract.
The use of the word consistent is important here – it does not necessarily mean equal. It could be
argued that the company should not expect to take the profit that it would have accrued from
future premiums if the policyholder is not going to pay those premiums. This is in contrast with
the (more reasonable) taking of future profits from premiums already paid.
2.2 General alterations

The key principle is that the terms after alteration should be supportable by the earned
asset share at the date of alteration to avoid the company making a loss.
In other words, the altered policy should not need more than the future premiums after the
alteration plus the current asset share to meet the benefits and expenses of the post alteration
policy.
Ideally the profit expected from the contract after alteration should be the same as that
before, or alternatively the same as the expected amount had the policy been written
originally on its altered terms.
The next section of Core Reading develops a concept sometimes referred to as ‘boundary
conditions’. The idea of a boundary condition is that some alterations are rather similar to other
alterations, or to not altering the policy at all, or to taking out a new policy. Given that this is the
case, the terms offered for one alteration should be consistent with the terms offered for another
similar alteration. They should also be consistent with other ways in which the policyholder could
achieve the same result. Boundary conditions are often a useful way to test proposals on how to
alter a particular policy.
The purpose of the following question is to encourage you to think about boundary conditions
before the Core Reading explanation.

Question
State the most relevant boundary condition in each of the following cases:
(i) term assurance altered to endowment assurance
(ii) endowment assurance premium reduced to a quarter of its current level
(iii) endowment assurance sum assured quadrupled
(iv) endowment assurance term reduced to two years, 15 years in out of 35
(v) endowment assurance term reduced by two years, 15 years in out of 35.
Solution
(i) New policy (because the policy value of the term assurance is likely to be small)
(ii) Paid-up policy
(iii) New policy
(iv) Surrender
(v) Alteration to itself (ie minor change).
Other principles include:
 Surrender can be viewed as the limiting case of a reduction in policy term.

Therefore, the company will wish that, as the outstanding term tends to zero, the
premiums charged look consistent with the difference between the surrender value
and the maturity value.
 A conversion to paid-up status can be viewed as the limiting case of a reduction in

sum assured. So, apart from any differences in expenses to be incurred, the
premium after alteration should approach zero as the sum assured approaches the
paid-up sum assured.
 If the benefits are to be increased, this should be on terms consistent with the
additional premium which would be charged for a new policy with a sum assured
equal to the proposed increase. If the policy term is to be extended, the terms
should reflect the current premium basis so far as the period of extension is
concerned.
Question
A without-profits endowment assurance is to have its term extended.
Comment on whether the premium would be expected to rise or fall.

Solution
Probably fall, since the same amount of benefit is payable later.
However, for a small increase in term, expenses may outweigh this effect.
For larger increases in term, this means that the policyholder retains the same life cover for a
reduced premium. Theoretically, this gives scope for mortality selection against the company.
 Any methods adopted should be stable in that small changes in benefits should
result in small changes in premium, if expenses of alteration are ignored. In other
words, an alteration method should ideally reproduce the existing terms if a policy is
altered to itself (ignoring the impact of the cost of the alteration).
A method that significantly failed to do this would be likely to produce large changes in
premium for small changes in sum assured, for example. (Of course, an alteration of a
policy to itself is just a theoretical concept.)
 The terms offered after alteration should avoid the option of lapse and re-entry.
Question
State why a company wants to avoid lapse and re-entry.
Solution
Lapse and re-entry increases costs, especially if commission is paid.
Lapse and re-entry also gives rise to the possibility of selection against the company.
 Any increase in benefit may be subject to additional evidence of health, depending

in part on the scale of the alteration and when it occurs in the policy’s lifetime.
 The costs associated with carrying out an alteration should be recovered.
So, in assessing an alteration method, the principles we might judge it against include:
 affordability
 consistency with boundary conditions, eg surrender, paid-up, new policy
 stability
 avoidance of lapse and re-entry
 fairness in terms of extracting a suitable amount of profit from the altered policy
 ease of calculation and of explanation to the policyholder.
Section 3 looks at two alteration methods and assesses them against these principles.

3 Methods of calculation
In this section we will look at two methods that can be used to calculate alteration terms. The
method of ‘equating policy values’ is the most accurate. But first of all we will look at a more
approximate method that is sometimes used to simplify the calculation of paid-up values.
3.1 Proportionate paid-up values
The method
For without-profits endowment assurances, the paid-up value may be calculated as the
basic sum assured multiplied by the ratio of the total number of premiums actually paid to
those originally payable throughout the total term.
Proportionate values will only ever be approximately ‘correct’, as we discuss below. Its simplicity
is of reduced value with the computing power available today.
Meeting the principles

 Proportionate paid-up values are usually too high at short durations, because they
do not allow for the high initial expenses. At medium durations, on the other hand,
they tend to be too low because no allowance is made for investment earnings.
Early premiums buy a larger proportion of the sum assured than later premiums, because
of the time to maturity and the effect of compound interest. At most durations this
means that a proportionate value understates the true value of the policy, although it will
tend to the maturity value as the duration of making the policy paid-up approaches the
maturity date. However, early on it is the initial expenses that dominate, causing the
‘true’ paid-up sum assured to be negative (ie a paid-up value should not be offered).
 The method is unlikely to be consistent with surrender values.
Question
A particular regular premium endowment produces a positive asset share after two years. At this
point a surrender value is offered.
(i) State the duration at which it would be possible to offer a paid-up value for the policy
without suffering a loss.
(ii) Explain your answer.

Solution
(i) Later than two years.
(ii) With a paid-up policy, renewal expenses continue (unlike with a surrender). This means
that a particular policy will not support a paid-up value until some time after it supports a
surrender value. If, for example, the present value of renewal expenses is £500 and the
annual premium is also £500, it will be at least a further year before the policy supports a
paid-up value.
 The method does, however, have the virtue of being a very simple one to apply and
explain to policyholders.
3.2 Equating policy values
The method
The value of the contract before alteration, on a prospective or retrospective basis, can be
equated to a prospective value after alteration that takes into account the requested
changes to the terms of the contract. The method can be used for any type of alteration,
including conversion to paid-up status.
Calculation of paid-up sums assured

The example of the whole life assurance from Chapter 21 can be used to illustrate the
calculation of a paid-up sum assured. Denoting the paid-up sum assured at policy duration
t as (PUSA)t and the surrender value as (SV)t , the paid-up sum assured can be determined
using the equation of value:
(m )  fAx  t  (SV )t

(PUSA)t Ax  t  ea x t
The level annual expenses from the original contract are denoted as e , and the claims expenses
as f .
Strictly, e should be the renewal expenses associated with paid-up policies, but in practice
this distinction is rarely drawn.
However, we must make sure that we do not charge paid-up policies for any future renewal
commission, which could make a big difference.
Paid-up sums assured for an endowment assurance can be determined in a similar way.

Question
In the previous chapter, the prospective method was used to calculate a surrender value of
£8,880 for a without-profits endowment assurance for a 45 year old with five years to maturity
under a contract with sum assured £20,000 (payable immediately on death), premiums £1,733
payable annually in advance, renewal expenses at 6% of premium and surrender expenses of £50.
Calculate the paid-up sum assured on the following basis:

 mortality as under the AM92 table
 only healthy lives surrender, ie they are select lives at the date of withdrawal
 interest rate 4% pa
 paid-up admin expenses are £25 pa, payable at the start of each year.
Solution
The new sum assured S  will satisfy the equation:
S A  25 
a  8,880
[45]:5| [45]:5|
which solves to give:
8,880  25  4.6172
S   £10,655
0.82260
Question
Calculate the paid-up sum assured in the previous question again, but now assume that paid-up
lives experience ultimate mortality from the paid-up date, rather than select.
Solution
We have to recalculate both the surrender value and the paid-up value on the new basis.
The ‘surrender value’ is no longer a surrender value in the strict sense, because the amount is not
being paid out as cash. Instead, we can think of this as this policy’s fair share of the money
available at the conversion date, which can be used to buy the paid-up benefit. As we are
assuming the lives are ultimate, then we should recalculate the amount based on an ultimate
mortality assumption.
So:
SV  20,000 A  0.94  1,733  

a  50
45:5| 45:5|

where:
 D50  D50
A  1.02  A45  A50   0.82264
45:5| D45
  D45
D

a a45  50 
  a50  4.6159
45:5| D45
giving a (notional) surrender value of £8,883. Then the paid up sum assured is:
8,883  25
a45:5
PUP   £10,658
A
45:5
Clearly this difference is so small that it would be ignored. However, differences could be
significant if more intense selection effects were assumed (eg if there was much more difference in
mortality between the different groups of lives involved).
Calculation of terms for an alteration

For general alterations, the principle of the calculation is that the reserve for the policy held
before the alteration should equal the prospective reserve of the altered policy plus the
costs of alteration, C.
The example of the whole life assurance from the last chapter where the sum assured after
the alteration is S  , secured by a premium of P  and future renewal expenses at the date of
alteration are assumed to be e  , is again used. Then the equation of value is:
(m )  e a
Value before alteration = S Ax  t  P a (m )  fAx  t  C
x t x t
From this equation we would be looking to find out the value of S (or P ) for a given value of P
(or S ).
The formula assumes that future claim expenses, f , are unchanged after the alteration, which
may not necessarily be the case. So in general, we could put f  for f in the above equation.
By ‘value before alteration’ we mean the policy value (provision) calculated for the policy, at the
alteration date, according to its original sum assured, premium, and policy conditions. Let’s call
this value tV (with no prime). So our equation of value can be written:
tV  tV   C
where:
tV   S  Ax t ax(m
 P  )
ax(m
t  e 
)
t  f  Ax t

ie tV  is the policy value of the contract, on the same date, calculated in its ‘new’ form. So the
equation is:
old policy value  new policy value  alteration expense

Another way we can look at the equation of value is to rearrange it (taking premiums to the other
side) as follows:
tV  P  
ax t  S  Ax t  e  
a x t  f  Ax t  C
that is:
old policy value  value of new premiums

 value of new benefits  value of new expenses  alteration expense
or:
EPV of income  EPV of outgo

In this second form, it is easy to see how the ‘old’ policy value is being used as a single premium
which, in addition to the future premiums receivable, will pay for the (new) policy benefits.
Alterations to satisfy any given specification for the values of S  and P  can be determined
using this relationship.
Similar methods can be used to develop expressions for alterations to endowment

assurances.
Question
Calculate the revised premium which should be payable in respect of an increase of £3,000 to the
sum assured for the endowment assurance of the previous two questions, using the same basis as
in the surrender value calculation:
 AM92 Select mortality
 interest rate 4% pa
 future renewal expenses of 6% of the current (original) premium.

Solution
We have already calculated the surrender value of 8,880 (which is the old reserve minus the
surrender expense).
So the new premium P will satisfy:
23,000 A  0.06  1,733  

a  P 
a  8,880
[45]:5| [45]:5| [45]:5|
which solves to give P   £2,278.41 .
This assumes that the surrender cost already deducted in the derivation of the surrender value is
adequate as the expense allowance for administering the policy alteration.
This example shows how an alteration can be consistent with a surrender value basis.
Things to do in an equating-policy-values calculation question in the exam

Equating policy values
Start with an equation of the form:
(Old policy value) = (New policy value) + (Alteration expenses)
Prospective or retrospective reserve
The question should state the basis and method by which the policy value calculations are to be
performed. If the method is not stated, or at least reasonably implied by the information given,
then it would be usual for a prospective value to be required.
Renewal expense assumption
Unless otherwise stated, expenses expressed as a percentage of premium will relate to the
premium that applies to the reserve calculation (new or old) concerned.
Ultimate or select mortality
Again, unless otherwise stated, any policy value calculated at time t years through an existing
policy will be calculated assuming mortality appropriate to lives of select age [ x ]  t . So, for all
cases where AM92 mortality is assumed, reserves calculated for t  2 will be calculated using
ultimate mortality (irrespective of whether the lives were select or ultimate lives at age x).
The following question will illustrate these points.

Question
A policyholder took out a whole of life policy for a sum assured of £50,000 exactly ten years ago.
They were aged 42 at entry, premiums were payable monthly throughout life, and benefits were
payable at the end of the year of death.
Immediately before the start of the 11th year of the policy, the policyholder wishes to alter their
policy to an endowment assurance for the same sum assured, maturing in 13 years’ time.
(i) Show that the original monthly premium is £57.41.
(ii) Calculate the new monthly premium that should be payable under the revised policy.
Basis:
 Mortality: AM92 Select
 Initial expenses: £250
 Renewal expenses: 4% of each monthly premium except the first
 Alteration expense: £50
 Interest: 4% pa
Premiums are calculated using the equation of value approach.
Solution
(i) Original monthly premium
The monthly premium is P where:
(12)
[42]
12Pa  50,000 A[42]  250  0.04  12P 
a (12)  1 
 [42] 12 
where:
(12)  11 11

a[42]  a[42]   19.560   19.102
24 24
A[42]  0.24770
Therefore:
50,000 A[42]  250 12,635

P   £57.41
(12)
0.96  12
a[42]  0.04 220.095

(ii) New monthly premium
The ‘old’ policy value is:
(12)
10 V  50,000 A52  0.96  12  57.41  
a52
(We use ultimate mortality because the select age is [42]  10 , which takes us out of the select
period of the AM92 Select mortality table.)
Now:
A52  0.35238
(12) 

a52  a52  24
11  16.838  11  16.380
24
So:
10 V  50,000  0.35238  0.96  12  57.41  16.380  £6,785.87
To calculate the new monthly premium P we equate:
6,785.87  12P
a (12)  50,000 A52:13  0.04  12P
a (12)  50
52:13 52:13
(We assume that renewal expenses are now 4% of the new monthly premium P .)
Now:
11  D65  11  689.23 
a(12)  
 a  1    10.106   1 
52:13 52:13 24  D52  24  1,256.80 
 9.899
A52:13  0.61130
So:
50,000 A52:13  50  6,785.87 23,829.13

P    £208.96
a (12)
0.96  12 114.03648
52:13

Meeting the principles

 The method will produce consistent surrender values immediately before and after
alteration if the same methods and assumptions are used as for calculating
surrender values.
 For an extension of term or increase in benefit, use of the current premium basis to
calculate the before and after alteration policy values would ensure consistency with
the terms for new contracts. It is unclear whether other bases would.
 There will be consistency between the terms for alterations, surrender values and
conversions to paid-up status, if the same bases are used.
 Assuming the same basis is used for the before and after policy values, the method
is stable.
Question
Explain what is meant by an ‘unstable’ alteration method.
Solution
A minor alteration leading to a big change in premium, or sum assured, or whatever is changing.
 It will not necessarily avoid lapse and re-entry and the company would need to
check that the premium being charged after alteration is not more than what it would
charge for a completely new contract.
 Provided that the policy value before alteration is not greater than the earned asset
share and the basis for the policy value after alteration is not weaker than a best
estimate basis, the alteration terms should be affordable.
In general, the approach of equating policy values can be made appropriate in almost all
circumstances provided appropriate bases are chosen.

4 Determining the basis for the equating policy values method

In order to set the alteration terms using the equating policy values method, we need to
determine a set of assumptions, eg investment return, mortality, expenses. No assumptions are
required to use the proportionate paid-up approach.
4.1 Expected profit from altered without-profits contracts

The total profit expected from an altered contract depends on the relationship between:
 the method and basis for calculating the policy value before alteration, which
determines the profit ‘released’ at the time of alteration, and
 the method and basis for calculating the policy value after alteration, which
determines the profit that is expected to emerge over the remaining term of the
contract.
The profit ‘released’ at the date of alteration will be:
 the full expected profit under the unaltered contract, if a realistic prospective value
is used for the policy value before alteration
 no profit at all, if an earned asset share is used for the policy value before alteration
 something in between, if a prospective value using a basis incorporating margins is

used for the policy value before alteration.
This profit is not the same as the accounting profit or the embedded-value profit.
The profit expected to emerge, from the date of alteration, over the remaining life of the
altered contract will be:
 no profit at all, if a realistic prospective value is used for the policy value after
alteration
 profit corresponding to the margins in the assumptions, if a prospective value using

a basis incorporating margins is used for the policy value after alteration.
Depending on the profit that the company wishes to retain, it will choose an appropriate
combination of bases.
Chapter 21 contained graphs showing how retained profit is related to the surrender value basis
chosen. It might be helpful to have another look at these before continuing, as they also help to
illustrate the profit implications of the basis choices described above.
Question
The sum assured on a without-profits endowment assurance is being doubled and the
outstanding term reduced from 15 to 10 years.
Explain whether equating prospective policy values on a realistic basis is a suitable method / basis
of alteration.

Solution
The effect of the method is that expected profit is unchanged as using a realistic basis for the
policy value before alteration takes the full expected profit from the original contract, and using a
realistic basis over the remaining policy term takes no further profit. Given that the policy has
increased in size, it is likely that an increase in profit would be required and the method is
therefore not likely to be ideal.
Finally, a useful way to summarise the issues on profit for the equating policy values method is as
follows:
 The left-hand side is like a surrender value.
 The right-hand side is like a new contract.
 To find the total expected profit from the altered contract, add the two bits together.
Here, ‘left-hand side’ and ‘right-hand side’ relate to the policy values calculated before and after
the alteration.
Question
Discuss whether there is ever anything wrong with equating policy values on the premium basis
for without-profits alterations.
Solution
What ‘premium basis’? Using the original premium basis is highly unpredictable since conditions
may have changed a lot since the start of the policy.
For without-profits alterations, using the current premium basis, the method is often a good
choice. Using the current premium basis ‘before’ will normally extract some profit, and using the
current premium basis ‘after’ will be expected to produce some more profit after the alteration.
If the policy increases in size, expected profit will normally rise; if it reduces in size, expected
profit will normally fall. This will often seem reasonable.
4.2 Assumptions
The assumptions used would likely be those implicit in the pricing of a new contract issued
at the alteration date to provide the benefits before alteration and the benefits after
alteration. The basis may be prudent (with margins) or realistic (without margins).
We discussed the different ways that an allowance for risk could be taken in the pricing basis in
Chapter 17.
The choice between Select and Ultimate in the mortality assumption is likely to depend on
whether medical evidence is required before the alteration is proceeded with.
Assumptions are also often required for alteration expenses.

4.3 Selection
A theme running through much of the material in this chapter, but not addressed in detail, is the
theme of selection. When considering the terms to offer in a particular alteration, or whether to
allow a particular alteration at all, it is important to consider whether the company is at risk of
financial or mortality selection by the policyholder.
A number of examples of potential selection are discussed briefly below to help you think about
issues of selection and to recognise them in an exam situation.
Situation 1: Extension of term of term assurance. Fairly obvious scope for mortality selection
(although alterations are rarely allowed to term assurance because the policy
values are low, and because they are sold as ‘cheap to administer simple
policies’).
Situation 2: Extension of term of without-profits endowment. More subtle issue of mortality

selection. This will decrease the premium and therefore increase the sum at risk
in the short term, as well as give rise to a sum at risk later than expected.
However, not such a major issue and full underwriting less likely.

When converting to paid-up status, the units attaching to the contract at the date of
conversion will remain attached, possibly after deduction of any penalty that applies.
Let’s assume, for the moment, that this Core Reading also applies to alterations other than just
conversions to paid-up status.
Question
‘I can understand why a penalty applies on surrender, but it seems a bit mean to apply one if the
contract is remaining in force.’
Discuss this comment and how the company might approach the subject of ‘alteration penalties’.
Solution
The aim of this kind of penalty is to recoup that part of the initial expenses that otherwise
wouldn’t be recouped. So to the extent that future charges on the altered policy continue to
recoup initial expenses, a penalty isn’t required.
However, the expected present value of charges might be less on the altered policy than on the
original policy. The company will need to decide how it wishes to deal with this. Some possible
approaches are:
 Apply an ‘alteration penalty’ at the time of alteration, equal in value to the expected
difference in future charges. Care would be needed here, since applying a penalty might
itself alter the value of future charges (eg unit charges would apply to a smaller unit fund).
 The company might decide to accept a degree of cross-subsidy. If some alterations will
lead to increased charges (to recoup initial expenses) and others to reduced charges, then
the overall position might be close to neutral. In this case the company will be open to
the risk of a change in the profile of alterations. In any case, the company is likely to be
accepting expenses cross-subsidies on its policies anyway.
 If the likely volume of alterations is small, the company might decide that the negative
marketing effect of penalties more than outweighs the advantages of recouping the
‘correct’ expenses on each policy.
One of the attractions of unit-linked business to policyholders is its flexibility. Applying a penalty
whenever this flexibility is used might dramatically reduce a product’s marketability.
For unit-linked contracts it is also common only to allow contractual alterations, although
more flexibility may be allowed. For policies that provide pensions, increases and
decreases in premiums and changes in retirement date would normally be allowed. Some
unit-linked whole life policies can be highly flexible, with premiums and sums assured that
can be increased or decreased as required.


Chapter 22 Summary
Without-profits alterations
In assessing an alteration method, the principles we might judge it against include:
 affordability
 consistency with boundary conditions, eg surrender, paid-up, new policy
 stability
 avoidance of lapse and re-entry
 fairness
 ease of calculation and of explanation to the policyholder.
For without-profits alterations, one of the key themes is to extract a suitable amount of
profit from the policy after alteration. The method of equating policy values before and
after the alteration can usually be used to produce satisfactory (or the least unsatisfactory)
terms for the alteration. Often, using a current premium basis before and after is a good
starting point.
Paid-up values can also be calculated using the proportionate method.
When performing alterations, particular note should be taken of the risk of financial or
mortality selection against the company.
Unit-linked alterations
For unit-linked contracts, the units attaching to the contract at the date of converting to
paid-up will remain attached, possibly after deduction of any penalty that applies.



22.1 A life company is reviewing the basis it uses for altering non-linked without-profits annual
Exam style
premium endowment contracts to paid-up status.
(i) State the boundary conditions that would typically need to be satisfied. [3]
(ii) Give a formula that could be used to derive the paid-up value by equating the realistic
prospective value of the annual premium policy with a realistic prospective value of the
paid-up policy, defining any terms used. [3]
[Total 6]
22.2 Discuss in general terms the following procedures for performing alterations to without-profits
Exam style
policies. You should assume that the company in question calculates premiums for
without-profits business on a basis that it expects will produce profit.
(i) Equate policy values before and after the alteration on a realistic prospective basis,
except for a realistic allowance for the cost of alteration. [4]
(ii) Equate policy values before and after the alteration on the original premium basis for the
altered contract. [5]
[Total 9]
22.3 A life insurance company uses prospective methods and the following basis for calculating its
Exam style
surrender values and alteration terms:
 Mortality: AM92 (lives assumed to be Select at policy outset)
 Interest: 4% pa
 Expenses: Renewal: 2% of each premium
Alteration or surrender: 120
A without-profits endowment assurance policy was issued exactly fifteen years ago to a life who
was then aged exactly 40. The sum assured under the policy is 250,000, payable at the end of the
year of death or at the end of the term of 25 years, and the annual premium is 6,200.
(i) Calculate the surrender value payable, if the policyholder surrenders the policy just before
the premium now due. [2]
Instead of taking the surrender value, the policyholder asks to convert the policy into an
endowment assurance maturing in five years. The sum assured is to remain the same.
(ii) Calculate the revised premium. [2]

[Total 4]

22.4 A life insurance company allows its without-profits endowment assurance policyholders to extend
Exam style
the terms of their policies on request. The new premium payable after such an alteration is
calculated by equating present values before and after the alteration on the current premium
basis.
Discuss the implications of this approach for the company’s profit on these policies and any other
factors the company should consider in assessing the method’s suitability. [9]

22.1 (i) Boundary conditions for setting paid-up policy values
This question is similar to a past exam question, in which the examiners commented that few
students mentioned ‘profit’ as a consideration. The profit retained by the company is fundamental
to alterations, and we believe that most students do realise this. However, they obviously didn’t
say so in the exam.
The paid-up value must be consistent with the surrender value and run smoothly into the
maturity value where the contract becomes paid up late in the term. [1]
The company should aim to leave the profit from the contract unchanged, although this might not
always be possible. [1]
The surrender value after the change should be consistent with the surrender value before. [½]
The paid up value should be consistent with policyholders’ reasonable expectations, ie consistent
with any sales literature or any previous quotation or the figures disclosed at the point of sale. [1]
The paid-up value should be supported by the earned asset share at the date of conversion, on
the basis of expected future experience. [½]
[Maximum 3]
(ii) Formula for calculating paid-up values
At the date of conversion to paid up, the following equation can be used:
SA  A a (12)
+ R   P 
a  PUSA  A + R 
a (12) [2]
x t:nt| x t:nt| x t:nt| x t:nt| x t:nt|
where:
P = annual premium (before conversion)
SA = sum assured (before conversion)
PUSA = paid up sum assured
R = level of monthly renewal expense before conversion
R = level of monthly renewal expense after conversion
t = duration
n = original policy term. [1]
The renewal expenses might be marginally lower for the paid-up policy as no premiums need to be
collected.
[Total 3]

22.2 Part (i) explicitly mentions alteration costs, therefore it allows correctly for this item. Part (ii)
doesn’t.
(i) Equate realistic policy values
May be suitable if the size and term of the new policy are not too dissimilar from the old policy. [1]
Using the current realistic value takes out full expected profit for the current policy. Using a fully
realistic basis for the new policy takes out no further profit. [1]
The method may not be suitable if the size of the policy is being substantially reduced. It may be
too mean. For example, if very early on, a policy is reduced substantially in size, the new
calculated premium may be larger than the office premium for the new small policy. The limit of
this process is the early surrender, where we know that a realistic prospective approach
sometimes gives unsuitable values (often gives negative values, in fact). [1½]
It may also be unsuitable if the policy is being substantially increased in size. Here, the approach
may be too generous. The method keeps expected profit the same as it was, which may not be
what the company expected if the policy double / triples etc in size. [1]
[Maximum 4]
(ii) Equate policy values on the original premium basis
May be suitable, provided that bases haven’t changed too much. [½]
Under these circumstances, the policy value of the current policy on the original premium basis
extracts a proportion of profit that increases as duration increases (sometimes called ‘accrued’
profit). Using the same basis for the new policy would produce further expected profit if the basis
were still suitable now. The method copes well with an increase in the size of the policy: using the
premium basis for the new policy will extract more profit. It also copes well with a decrease in
the size of the policy. [2]
The problem comes if conditions have changed substantially since the policy was
issued. [½]
Let’s say the policy was issued when we were happy guaranteeing 6% pa for a regular premium
policy. Let’s say we are now only happy guaranteeing 4½% pa. If there is a substantial increase in
the size of the policy, the premium calculated using the old basis may be much too low. [1]
Conversely, if the old basis was on 4% pa interest and we were currently prepared to guarantee
5% pa, and the policy size increased substantially, the new premium calculated on the old basis
might be far too high. It might even be higher than a new premium calculated at 5% pa for the
new contract. [1]
There is no mention of adjusting for initial expenses in the premium bases to the extent that
these are not required, nor of the costs of the alteration. It is therefore unlikely that expenses
have been correctly allowed for. [1]
[Maximum 5]

22.3 (i) Surrender value
The surrender value is:

SV  current prospective reserve  alteration expense
 250,000 A  0.98  6,200 

a  120
55:10| 55:10|
 250,000  0.68388  6,076  8.219  120

 120,911
[2]
(ii) Revision of premium
Let P be the new annual premium, then:
120,911  250,000 A  0.98 P 

a
55:5| 55:5|
[1]
So:
250,000  0.82365  120,911
P  18,917
0.98  4.585
[1]
[Total 2]
22.4 Implications for profit
Use of the current premium basis to determine the policy value before the alteration will extract
the ‘accrued’ profit for the period up to the date of alteration, for policies that were sold on this
premium basis. [½]
For policies sold on a previous premium basis, the profit extracted will depend on how the basis
differs from the current basis. [½]
For example, if the premium basis has been strengthened, less profit will be extracted than if the
original premium basis were used. [½]
Use of the current premium basis to determine the policy value after the alteration will give
expected profit over the remaining life of the contract corresponding with that from current new
policies. [½]
This combination should result in an ‘appropriate’ amount of profit for the company, assuming
that new business terms are profitable. [½]
The company should check, however, that the policy after the alteration is supportable by the
asset share at the alteration date in order to avoid making an expected loss. [½]

Other factors
The calculation of the present value of the policy after the alteration should be adjusted for initial
expenses and commissions (which are not payable in the case of an altered policy) … [½]
… otherwise the terms may be too harsh. [½]
Similarly, the alteration expenses should be deducted from the present value before the
alteration (or added to the present value after the alteration) … [½]
… otherwise the terms may be too generous. [½]
The company should consider the risk of anti-selection by policyholders. [½]
Under the new contract, the sum assured will be paid out at either the same time as before (if
death occurs before the original maturity date) or later (on survival beyond the original maturity
date). [½]
This means that either the revised premium will be lower than before … [½]
… or, if the premium stays the same, the sum assured will be increased. [½]
Therefore, those in very poor health would gain by extending the term of their policy, because
either they would be paying a lower premium and expect to receive the benefit at the same time
as before … [½]
… or, for the same premium, they would receive a higher sum assured on death prior to the
original maturity date. [½]
To deal with this risk, the company could:

 require a declaration of continuing good health [½]
 underwrite all alterations and modify the terms offered if necessary [½]
 use a basis with higher mortality than in the premium basis to calculate the policy value
after the alteration, for all alterations. [½]
The first two options will not be available if the policy terms allow extensions with no further
medical underwriting. [½]
Use of the current premium basis will ensure consistency between the terms offered on
alterations and new policies. [½]
However, it is not clear that the resultant premium will always avoid potential problems with
lapse and re-entry. [½]
Using the same basis for the present values before and after the alteration will make the method
stable, ie a small increase in term will result in only a small increase in premium. [½]
[Maximum 9]

End of Part 4
What next?
Chapters 17 to 22.
of your revision.
… ‘revision and rehearsal’ products
Revision Notes – Each booklet covers one main theme of the course and includes integrated
questions testing Core Reading, relevant past exam questions and other useful revision aids.
One student said:
‘I found the revision booklets very helpful as they grouped all past exam
questions by topic and the answers were bullet pointed into key points so
it was easy to check how many of the points you get.’
You can find lots more information, including samples, on our website at www.ActEd.co.uk.





using the course.

ST2-23: Cost of guarantees and options Page 1
Cost of guarantees and

options
Syllabus objectives
1.1.3 Outline typical guarantees and options that may be offered on life insurance
products.
4.3 Demonstrate methods of determining the cost of options and guarantees.
4.3.1 Describe the use of stochastic simulation and the use of option prices to
determine the cost of an investment guarantee.
4.3.2 Describe the assessment of the cost of simple mortality options.

Page 2 ST2-23: Cost of guarantees and options
0 Introduction
In this chapter we shall consider investment guarantees and mortality options. Examples of
investment guarantees are found in both unit-linked and non-linked business. Guarantees may
relate to surrender and/or maturity values. Surrender guarantees apply at more than one point in
time and hence have the additional uncertainty of when they may be taken up.
There is a chance that the asset share underlying the policy does not reach the level guaranteed
on payout. Another way of looking at this is that guarantees may constrain the investment policy
that an insurance company would otherwise follow. Hence they carry a cost, and so the
guarantee should be priced appropriately. We will look at some ways in which the costs can be
evaluated.
We will also consider mortality options. These are situations where the policyholder can choose
to extend the term or increase the level of cover at normal premium rates but without providing
further medical evidence. To the extent that the option might be exercised by someone in poor
health, the insurance company will bear a cost: the difference between the ordinary premium
rate granted under the terms of the option and that which would have been granted had the life
been underwritten.

1 Investment guarantees
1.1 Examples
Traditional life insurance contracts transfer the mortality, expense and investment risk from
the policyholder to the life insurance company.
Investment-linked (including unit-linked) contracts leave the investment risk with the
policyholder. The attractions of these policies can be enhanced if part of the investment
risk is transferred to the insurance company by the contract including, for example:
 guaranteed minimum maturity value
 guaranteed minimum surrender value
 guaranteed annuity option.
For investment-linked business, the guaranteed minimum maturity and surrender values referred
to above are stated in absolute money terms, and guaranteed annuity options are as described
for conventional policies below.
With-profits contracts transfer investment risk to the policyholder, but generally to a lesser
extent than for unit-linked contracts (due to smoothing). They typically have some level of
guaranteed benefit, which is equivalent to an investment guarantee, eg the basic sum
assured plus attaching bonus for conventional with-profits business under the additions to
benefits approach. Each of the above investment guarantees may be available on
with-profits as well as on investment-linked contracts.
The attractions of traditional policies can also be enhanced if the company provides
guarantees of investment performance. For example, an option could be provided to
convert the maturity value of a without-profits endowment assurance into an immediate
annuity on guaranteed terms.
For example, the policyholder takes the cash maturity value but can choose to buy an immediate
annuity from the company at a rate guaranteed at policy inception (perhaps 30 years earlier).
This will be an attractive option if the policyholder is worried that annuities may become more
expensive when he or she retires.
In each case the policyholder is protected from a ‘downside’ risk.
1.2 Implications for the insurance company

The risk assumed by the life insurance company is that at specified times in the future the
‘backing’ assets will be insufficient to meet the guarantees.
This risk is particularly difficult to control if the policyholder has a choice over whether to exercise
the option. For example, a guaranteed annuity option gives the policyholder a choice of taking
cash or an income at maturity. The company will not know which option the policyholder will
choose and hence which option it should match in investment terms.
Additionally, if the guarantee relates to the surrender value, the company will not know the
specified time at which assets must cover the liability (ie the surrender value) as the time of
exercise will be under the control of the policyholder.

Question
Suppose that an insurance company offers guaranteed annuity payments on its without-profits
deferred annuity contract, but invests to meet the open market cash option.
Explain whether the company is at risk from low or high interest rates at retirement.
Solution
The company is at risk from low interest rates since the annuity may cost more to buy than the
cash available.
The reverse is true if it invests to be able to pay the annuity. In this case, the risk is that interest
rates are high and that the annuity the company has invested to be able to pay is worth less than
the cash alternative.
If the company has control over the investment policy, (eg traditional contracts with
guarantees), there is conflict between investing to meet the guarantees and investing for
maximum performance.
In other words if the company invests the assets backing a with-profits contract to meet a
minimum guarantee, then all the policyholder will receive is this minimum return. The
policyholder will miss out on the prospect of any out-performance and might just as well have
bought a without-profits contract.
If the company chooses not to invest to match the guarantees, it must include the cost of
the guarantee in the original pricing basis. If the company has no control over the
investment policy (eg under a unit-linked endowment assurance), it must include the cost of
the guarantee in the original charges to the extent that the guarantee will not be matched.
Whether the company mismatches through choice or necessity it will need extra funds from
increased premiums or charges in order to offset its risk.
The risk to the company will depend on the outstanding term of the policy. In general, the longer
the timeframe involved then the greater the chance that things might go wrong compared with
our current forecasts. For instance, offering a guaranteed annuity rate for five years’ time will be
less risky than offering the same rate for twenty years’ time.
However, the picture is more complex for conventional with-profits business. Here, a long
timeframe may give the company time to act on profit distribution levels in the event of
eg disappointing investment returns. The company’s management has much less scope for
manoeuvre with a short-term policy or a policy that is approaching maturity. If things go wrong
there is much less time for corrective action, eg reducing reversionary bonuses, to take effect.
For conventional with-profits policies, the higher the amounts of regular distributions of profit
made (like reversionary bonus), the more onerous the guarantee becomes.

Question
Suppose that an insurance company offers a guaranteed surrender value that can be exercised on
an annual basis.
Explain why the company cannot match the guarantee by assuming (on the basis of past
experience) that a certain percentage of policyholders surrender each year, ie why it cannot just
hold cash for these amounts and hold the remainder in appropriate assets to meet the maturity
or death benefits payable to the rest of the policyholders.
Solution
The reason that the company cannot do this is that more or fewer policyholders may surrender in
the future. Withdrawal rates are very variable, particularly if interest rates and market values
vary. For example, if interest rates rise and market values fall policyholders will see the
opportunity to surrender and reinvest on attractive terms. In effect they are selecting against the
company.
1.3 Valuing an investment guarantee

The liability created by an investment guarantee is the excess of the guaranteed amount
(eg guaranteed maturity value of an endowment, or the fund needed to purchase the
‘guaranteed annuity’ at current market rates), over the cost that would have been incurred at
the time in the absence of the guarantee. The policyholder should only choose to exercise
the option to take up this guarantee if it is in-the-money (which can alternatively be
expressed as ‘if it bites’), ie is financially advantageous.
As discussed earlier, such guarantee costs will be the result of the insurance company being
unable or unwilling to invest the premiums to match the guarantee and subsequently losing out
as a result of its mismatched investment strategy. The insurance company will need to honour its
guarantee whenever this situation occurs and hence must hold sufficient funds to meet its
commitments. It will therefore seek to charge an additional premium to meet the liabilities
created by the guarantee.
The value of these liabilities can be determined using:
 option-pricing techniques
 stochastic simulation of investment performance.
For example, in order to be sure of meeting a guarantee to pay a sum of 1,000 increased in line
with cash returns, an insurance company would need to invest in cash. However, investing in
cash may not be allowable under the terms of the contract, eg a unit-linked contract, or it might
reduce unacceptably the return that the company could provide to the policyholder. The
company therefore needs to charge an extra premium to reflect the extra sums it may need to
pay out under the guarantee.

There are two main ways in which this extra premium can be assessed:
 Option-pricing / market valuation techniques. These assess the extra premium by looking
at the market price of a derivative that the insurance company could acquire to mitigate
its risk.
 Stochastic simulation. The extra sums likely to be needed under the guarantee can be
modelled by simulating a range of investment scenarios.
Use of option prices / market valuation techniques

The options incorporated into life insurance contracts are analogous to options traded in
the market place. A guaranteed minimum maturity value corresponds to a (European style)
put option on the investment funds at an exercise price corresponding to the maturity
guarantee.
Suppose that a unit-linked endowment assurance with a single premium of 1,000 has a
guaranteed minimum maturity value of 1,200. Suppose also that at maturity, the assets
underlying the policy are worth 1,150.
At maturity the office will have to sell the underlying assets in the marketplace for 1,150, and
then pay 1,200 to the policyholder. Compare this with the situation where the office had granted
the policyholder a European style put option on those assets with an exercise price of 1,200,
ie the office had given the policyholder the right to sell the underlying assets to the office for
1,200 on the maturity date. The office could then resell them for 1,150, ending up in exactly the
same position as before (ie with a loss of 50).
Hence, if a put option exists in the market for the same (or similar) terms, then the market price
for that option provides the company with an estimate of the expected present value of the
guarantee under the policy.
A guaranteed minimum surrender value corresponds to a similar American style option or a

series of options with different exercise prices which match the guaranteed surrender
values.
If a policyholder is given a guaranteed minimum surrender value, then the policyholder may have
the option to receive the guaranteed surrender value at any time. So the situation is very similar
to the maturity guarantee described above, except that the policyholder can exercise the option
whenever he or she wishes to do so. This is equivalent to the office granting the policyholder an
American put option, which can be exercised at any time up until the maturity date. A European
style option can be exercised only at the maturity date itself.
A guaranteed annuity rate corresponds to a call option on the bonds that would be
necessary to ensure the guarantee was met, ie at an exercise price which generates the
required fixed rate of return. Alternatively, it can be mirrored by an option to swap floating
rate returns at the option date for fixed rate returns sufficient to meet the guaranteed
annuity option (a ‘swaption’).
As described in the earlier subjects, a call option is the option to buy an asset at a future date at
the exercise price.

Question
An insurance company offers guaranteed annuity payments on its without-profits deferred

annuity contract, but invests to match the open market cash option.
Explain how the insurance company could protect itself from the annuity rate guarantee by:
(i) purchasing call options on bonds
(ii) purchasing interest rate swap options.
Solution
(i) Buying call options
At the maturity (retirement) date of the policy, the insurance company will have cash available
equal to the open market cash option. If interest rates are now lower than used in the
guaranteed annuity rates, then the bonds that the company needs to buy to match the annuity
payments will cost the company more money than the cash available from the policy maturity.
A call option essentially fixes a maximum purchase price for the bonds at the exercise date. So
holding the option means that the company will not need to pay more money (than it can afford)
for the bonds if interest rates do fall below the guaranteed level.
(This means that the current market price of the call option represents the expected cost of the
annuity rate guarantee provided by the company.)
(ii) Buying interest rate swap options
For this the company would buy options that would permit it to undertake ‘receive fixed / pay
floating’ swaps as from the retirement date, where the fixed rate in the swap agreement was
equal to the interest rate in the guaranteed annuity basis. The option would only lead to the
swap being undertaken if bond yields had fallen below the guaranteed rate at the retirement
date.
Thereafter (assuming the option is exercised) the insurance company would have to keep its
assets as cash, so that it could cover its obligation to ‘pay floating’ under the swap agreement.
Meanwhile the swap agreement will pay the insurer the income shortfall on its (cash) returns,
making this up to the level of income required to cover the guaranteed annuity payments.
(So this time it is the current market price of the necessary swap options that represents the
market valuation of the annuity rate guarantee.)

Question
Suppose the insurance company in the previous question has instead invested to match the
future annuity option, and holds bonds.
State, with justification, the type of derivative that the company would now look at in order to
assess the cost of the option it is providing under its contract.
Solution
The company could look at the price of a bond put option.
Without a put option in place, the company is at risk if interest rates rise, its bonds fall in market
value and the policy is surrendered for cash at retirement (or indeed at some earlier time). If it
has a bond put option, it can sell those same bonds at a guaranteed price (or the put option will
increase in price, which comes to the same thing).
It is difficult to ensure that the whole investment fund corresponds to a single option traded
in the market. However, an approximation is possible by using options written on market
indices for equities and bonds.
At the date of policy issue all guarantees will normally be expected to be out-of-the-money,
ie they will have no intrinsic value because current market rates are sufficient to meet the
guarantees. However, they will have a time value which is the result of the views of many
investors (‘the market’) of the present value of the likely future costs of the option.
Although current market rates of interest are sufficient to meet the guarantee, a number of
investors are sufficiently pessimistic about the prospect of future adverse rate changes to be
prepared to bet that the guarantee may bite at some date in the future. The amounts they will
pay for this ‘bet’ determines the present value of the option – and this is what we use to estimate
our expected present value.
Thus, the market price of a suitable option produces a way of costing a financial option or
guarantee incorporated in a life insurance policy.
It is important to be clear on what is going on here: we are talking about how we can use the
current market prices of traded ‘guarantees’ to provide a market-based valuation of the
guarantees we may be providing under our insurance policies. (This is actually just an example of
a market-consistent valuation, as described in Section 2 of Chapter 20.) We are not implying that
these derivatives have to be or are purchased by the company: this is an entirely separate matter,
which would depend on the company’s investment strategy in this regard.
It is possible that a guarantee will not be out-of-the-money. For example, current yields
might be so low that a life insurance company would be happy to provide a guarantee at a
future date based on a higher yield.

For example, we might be providing a guaranteed annuity option at retirement (which may be in
many years’ time) guaranteeing a yield of, say, 3% pa. Current yields might be lower than this (so
the guarantee is in-the-money at the present time), but the company might still offer the
guarantee if there is an extremely high chance of it being out-of-the-money by the time of
retirement.
Stochastic simulation
By projecting forward the value of the assets using a stochastic investment model and comparing
this with the sum payable under the guarantee, the insurance company can measure the extent
to which additional costs will be incurred under a range of investment scenarios.
A stochastic model of rates of return on investments is used to simulate the future price of
assets. The assumptions underlying the model must be carefully evaluated to ensure that
they correspond to the company’s planned investment strategy. A large number of
simulations is needed in order to obtain reliable estimates.
Key assumptions will be the probability distribution used to model the investment return and the
mean and variance. Typically 5,000-10,000 simulations might be used.
Example
A company is issuing a five-year unit-linked savings bond with a single premium of 1,000 and
guaranteed maturity value equal to the premium rolled up at 3.5% interest pa. The unit-linked
fund is to be invested in cash.
How could we cost the guarantee using stochastic techniques?
One approach is as follows.
We need to model the returns earned by the cash in the unit-linked fund, so we need to decide
on how to model future interest rate movements. In this example, we have assumed that each
year’s interest rate is equal to the previous year’s rate, but with a random drift which is normally
distributed with standard deviation of 1% (and mean drift zero). We kick off with current rates
programmed in at 5% pa.
Suppose that we have a random number generator, which is able to generate random numbers
that are uniformly distributed in the range 0 to 1. (You may have a function like this on your
calculator.) We start by generating say five numbers from this distribution:
0.115 0.023 0.864 0.212 0.816
We now use the inverse distribution function method to find the corresponding values from the
standard normal distribution:
z1   1 (0.115)  1.2 z2   1 (0.023)  2.0 z3   1 (0.864)  1.1

z4   1 (0.212)  0.8 z5   1 (0.816)  0.9

Our random interest rate in Year t ( It say) can now be simulated using:
It  It 1  0.01 Zt
where Zt is the standard normal random variable we have just simulated above.
Now, noting that I0  0.05 , we get the relevant simulated interest rate in each of the following
years:
I1  0.05  0.012  0.038 I2  0.038  0.02  0.018
I3  0.018  0.011  0.029 I4  0.029  0.008  0.021
I5  0.021  0.009  0.03
This will give a run of interest rates for the next five years. We then repeat the process, say
10,000 times. For each run, we can calculate the projected unit fund of the policy by
accumulating the premium, less charges, at the simulated rates of return. For this example we
have taken initial expense charges of 50 and ignored any other deductions from the unit fund.
We can then compare the unit fund against the guaranteed minimum maturity value of 1,000
accumulated at 3½% pa (ie 1,188).
Doing say 10,000 runs we can then look at the results:
Run Interest Interest Interest Interest Interest Unit fund Guarantee Cost
year 1 year 2 year 3 year 4 year 5
1 3.8% 1.8% 2.9% 2.1% 3.0% 1,086 1,188 102
2 7.5% 7.0% 6.8% 5.2% 4.7% 1,285 1,188 0
3 5.1% 6.3% 6.9% 6.4% 7.4% 1,297 1,188 0
4 4.2% 4.9% 4.7% 4.6% 3.5% 1,177 1,188 11
… … … … … … … … …
9,999 5.0% 3.9% 3.5% 3.4% 5.4% 1,169 1,188 19
10,000 5.6% 5.7% 7.2% 7.1% 6.1% 1,292 1,188 0
Total – – – – – – – 142,184
The total costs for the 10,000 runs here amount to 142,000, giving an average result of 14.2 per
run. So the cost of the guarantee, which we need to load into the premium, is 14.2.
The principle of stochastic modelling is that the simulated average cost is an estimate of the
expected value (of the cost).
Question
Explain why it is only an estimate

Solution
The average is only an estimate because there may be random error (there are only a finite
number of simulations), and there may also be errors in the structure of the model, and in the
values of the parameters we have used.
For some guarantees the liability if the guarantee is taken is fixed, eg maturity guarantee,
and for some guarantees the liability will depend on future market conditions,
eg guaranteed annuity rates. In the latter case, factors influencing the value of the liabilities
as well as assets will need to be simulated.
So we will need to project the market bond yields as at the retirement date, in each simulation,
which will allow us to compute the cost of the annuity liabilities in retirement. This simulation
should be consistent with the assumed performance of the assets over the same timeframe.
If the guarantee relates to a with-profits policy, the modelling will also need to include
assumptions about the level of future profit declarations. The assumptions for each simulation
need to be consistent with the simulated asset performance.
The company will need to make assumptions about future rates of exercising options,
which would take into account expected policyholder behaviour and the size of the
guaranteed amount relative to the alternative benefit (eg asset share).
So, for example, a guaranteed minimum surrender value on a unit-linked policy is more likely to
be taken when the guaranteed amount is greater than the alternative benefit based on the value
of the units. Similarly, a with-profits policy is more likely to be surrendered if the guaranteed
amount is greater than an alternative benefit based on the asset share. However, it would be
quite simplistic to assume a policy would always be surrendered in these situations. Policyholders
might value the maturity benefits more highly or expect the guarantee to become even more
valuable at a later date.
The present value of the liability can be determined by discounting the simulated cost of
exercising the option at a suitable rate. Repeated simulation will generate the probability
distribution of the present value of the cost of the option. The company can then charge a
premium having a present value which reflects the ‘market cost’ of providing that
guarantee; this could be the expected (ie average) simulated cost plus a margin.
Question
In the earlier example, the guarantee applied only at maturity. Suppose that the company offered
the same 3½% pa guarantee for surrenders at any time over the five years to maturity.
(i) Explain how the cost of this surrender guarantee would compare with that of the maturity
guarantee.
(ii) Explain how the method of the above example could be modified to cost this guarantee, if
it applied to surrenders at the end of each year.
(iii) Explain how the method should be modified to set an appropriate premium, if the policy
were conventional without-profits rather than unit-linked.

Solution
(i) The cost of the guarantee would be greater since the guarantee could bite at any time in
the five-year period if the unit fund is less than 1,000 rolled up at 3½% pa.
The cost will be further increased by policyholders choosing to surrender more often
when the cost of the guarantee is highest (this is called financial selection).
(ii) We will need to allow for the policyholder to exercise the guarantee the first time that it is
in-the-money. Thus for each run we compare the unit fund of the policy at the end of the
year with 1,000 rolled up at 3½% pa. The cost incurred for each run will be the difference
between the guarantee and the unit fund on the first occasion that the guarantee is
greater.
The average cost per run would then be calculated as for the maturity guarantee.
(iii) The guarantee for a without-profits policy is the basic sum assured. One approach to
costing the guarantee would be to run the model with stochastic investment returns and,
for a given level of premium, calculate the asset share at maturity allowing for expenses
and the required profit deduction. The asset share for each run is compared with the sum
assured and the loss calculated. The premium should then be varied until the probability
and amount of loss are acceptable.

2 Mortality options
2.1 Examples
It is common for life insurance policies to include options to:
 purchase additional benefits without providing further evidence of health at the

normal premium rates (for a life of that age) at the date on which the option is
exercised
 renew a life insurance policy, eg a term assurance at the end of its original term,
without providing additional evidence of health
 change part of the sum assured from one contract to another, eg from term
assurance to endowment assurance.
2.2 Implications for the insurance company

The terms and conditions under which the option can be exercised need to be clearly set
out in the original policy. Sometimes an option can only be exercised at fixed points of
time, eg at the end of every five years of a twenty-year term assurance or at any time
providing a qualifying event has occurred, eg the birth of a child, or taking up a new job at a
higher salary. The extent of the option will also be specified, eg the additional sum assured
cannot exceed the original sum assured.
The terms and conditions for the exercise of the option are designed to reduce ‘selection
against the office’, ie an excess of lives in poor health using the options to obtain large
amounts of life assurance at premium rates that do not reflect their expected mortality.
This is just another example of anti-selection, such as we have already discussed elsewhere in the
course.
Question
Explain why restricting the points of time at which the option can be exercised reduces the risk of
anti-selection.
Solution
The policyholder cannot then immediately exercise the option on discovering that they are a poor
risk, eg when they have just been diagnosed with a life-threatening disease.
The cost of an option is the value of the excess of the premium that should, in the light of
full underwriting information, have been charged for the additional assurance over the
normal premium rate that is charged. For some lives the option will have no cost.
If a life, who is in good health and who would be expected to satisfy normal underwriting
requirements, exercises the option, then the option will generate little or no additional
costs. The exercise of the option by lives in poor health will generate considerable
additional costs.

The total expected additional costs of an option depend on the health status of those who
choose to exercise the option, and the proportion of lives who choose to exercise the
option.
In general, the smaller the proportion who exercise the option, the worse will be the
subsequent mortality experience of those exercising the option. If a substantial proportion
exercise the option, then their subsequent mortality experience will on average be less
extreme.
So the cost of a mortality option is (roughly speaking) the product:
proportion of lives exercising option  average health of lives exercising option
Question
Explain which is worse for the company: every single person with worse-than-select mortality
taking up the option, or just the small number of very high risk lives taking up the option.
Solution
The first is worst – if everyone with higher than select mortality takes up the option.
This means that the maximum possible total amount of gain has been made by the policyholders.
This will include the cost to the company of insuring the few very high risk lives at standard
premiums, plus the costs of insuring the other medium-high risks on the same terms.
Some factors affecting mortality options are:
 The term of the policy with the option. The longer the term, the longer the policyholder
will have the option, and the more likely it is that, at some time, his or her health will
make the option appear worthwhile.
 The number of times the policyholder gets the chance to exercise the option, eg every
five years, on every policy anniversary or at any time whatsoever.
 Conditions attaching to exercising the option such as limiting the size of the option or
restricting the choice of contracts available under the option.
 The encouragement given to policyholders to exercise the option. As discussed in the

Core Reading, if take up of the option is low it tends to be only those who have most to
gain who exercise the option. As explained in the previous question, this can be a good
thing, as it could keep the total cost of the option low.
On the other hand, encouraging more of the healthy lives to exercise the option will not
cause any additional expected loss, and should contribute to the company’s total profit as
the company will essentially be issuing lots of new policies to lots of good risks, which
should be a profitable proposition.

Publicising the option more widely can achieve greater take up by healthy lives, but care
should be taken that the benefit (from future profits) is not outweighed by the risk of
attracting a bigger proportion of the loss-making high risk lives from taking up the option
as well.
 The extra cost to the policyholder who exercises the option. If the option involves a steep
increase in premiums, then the healthier lives might shop around to try to get the same
cover cheaper elsewhere. This means that the company will lose out on the potential
profit that these policyholders would have generated.
 Selective withdrawals. A healthy life may cancel a ten-year renewable term assurance
policy after two years because he or she discovers that the cover without mortality
options is much cheaper. The company has not in this case collected the option loading
from this person for very long, but is still left with the unhealthy lives who will exercise
the option to the cost of the company.
This is important to understand. Withdrawal by healthy lives before the option date
reduces the income (extra premiums) that we would have expected to receive from the
whole risk pool. If only healthy lives withdraw, then the cost of the option (which
depends only on unhealthy lives) will be little altered. This loss of income therefore
constitutes a significant risk in the management of option costs.
Question
Explain why a life insurance company might wish to restrict the choice of contracts available
under the option.
Solution
The risk of mortality selection against the company is more severe if the policyholder can choose
a term assurance contract rather than an endowment assurance under the option, for example.
So the company might wish to restrict the availability of contracts such as term assurance, in
these circumstances.
2.3 Valuing a mortality option

Mortality options are normally valued using cashflow projections. These cashflows would
include the additional benefits expected to be payable under the option and the additional
premiums expected to be received in relation to these benefits, to the extent to which the
option is assumed to be taken up.
The additional premiums would be based on the expected premium rates that would be
charged to standard lives for the additional benefit, as at the option exercise date.
So when the policyholder exercises the option they will pay exactly the same premium as a new
policyholder that has just passed underwriting on that day, ie they will be charged a premium
based on the select mortality tables in use at that time.

The projections should also allow for any additional expenses incurred in the administration
of the option.
For example, expenses would be incurred if the company writes to each policyholder to remind
them of the option before the exercise date. Expenses will then be incurred in processing the
policyholder’s request to exercise the option.
If the purpose of the valuation is pricing the option (rather than setting the liability) then
allowance should also be made for the additional reserves that should be held, both before
and after exercise.
If we price the option using cashflow projections we will need to calculate the profit vector each
year. As outlined in an earlier subject, the profit vector is given by the premiums, less claims, less
expenses, plus investment returns, less increase in reserves.
Question
Explain why a policy that contains a mortality option is likely to need much higher reserves than
the equivalent policy without an option.
Solution
The existence of the mortality option increases the mortality risk. For example, an option to
extend the term increases the time for which the insurer is exposed to risk. Similarly, an option to
increase the sum assured increases the amount that is exposed to risk. The reserves need to be
increased for the expected cost of the increased claims.
However, reserves will also need to increase to reflect the uncertainty in this expected cost.
Alternatively, additional solvency capital will be required to cover these risks.
Before the option is exercised, the company is exposed to the risk that policyholders will
selectively choose to exercise the option if their health is worse than for a select life.
Even after the option has been exercised, the company is still exposed to considerable
uncertainty. As there is no further underwriting, it has no way of knowing the health of the lives
that exercised the option. Reserves will need to be calculated on the assumption that a level of
anti-selection has taken place.
Valuing a mortality option therefore requires extra assumptions, including the:
 probability that the option will be exercised, at each possible exercise date
 additional benefit level that will be chosen, if this is at the discretion of the
policyholder
 expected mortality of the lives who choose to exercise the option
 expected mortality of the lives who choose not to exercise the option
 additional expenses relating to the option.

Option take-up rates

The model may assume that all eligible policyholders will take up the option, and that the
maximum additional benefit will always be taken.
You may see this approach (ie using the assumption that everyone exercises the option) described
as the ‘conventional method’ in past exam questions.
The assumption that everyone will exercise the option is obviously unlikely to be borne out in
practice, but the method does have the advantage of simplicity.
If there are many possible dates on which an option may be exercised, or there is a choice
from several alternative options, the model may assume that the worst option from the
financial point of view of the company is chosen with probability one.
Alternatively, the model may use more sophisticated take-up rate assumptions which vary
by exercise date or by alternative option. These would ideally be based on experience.
You may see this approach (ie making an assumption about the proportion of policyholders that
exercise the option) described as the ‘North American method’ in past exam questions.
This method uses a more realistic assumption for the option take-up rate than the first approach
suggested above. However, it may be difficult to obtain the data to make suitable assumptions
(historically this data was only available in North America, hence the name).
Mortality rates
Typically, due to anti-selection, the expected mortality of lives who take up the option will
be heavier than that of those who do not.
For example, the mortality of those who exercise the option may be assumed to be a higher
percentage of the base mortality table.
For example, we may assume that those that exercise the option will experience mortality of 150%
of the base mortality table.
Alternatively, an age loading may be applied (eg a policyholder of age x may be assumed
to experience mortality of age x  5 years).
It may instead be assumed that the mortality experience of those who take up the option will
be the Ultimate experience which corresponds to the Select experience that would have
been used as a basis if underwriting had been completed as normal when the option was
exercised. This would be consistent with an assumption that all eligible policyholders take
up the option.
So this would be the approach to use for the ‘conventional method’ described above.
Note, there should be a link between the assumed option take-up rates and the assumed
mortality rates.
It may be assumed that the lives who do not take up the option will continue to experience
the same level of mortality as would have been assumed without the existence of the
option. However, this would mean that the average mortality for all lives has been assumed
to be more than the base mortality assumption, since those taking up the option are
assumed to experience higher mortality than this level.

For example, if we assume that lives that take up the option are ‘ultimate plus ten years’ and the
lives that do not take up the option are ‘ultimate’, then the average of their mortality must be
heavier than ultimate.
However, we would expect that the average mortality of all lives (both those that take and do not
take the option) would be ultimate. To address this, the following alternative approach could be
used.
An alternative assumption would therefore be that the mortality of those who do not take up
the option is such that average mortality for all lives remains at the expected ‘base’ level.
The assumed mortality of those who do not take up the option would then be lower than this
base level.
For example, under this alternative approach, we could assume that the lives that take up the
option are ‘ultimate plus ten years’ and that the average mortality will be ultimate. If the
proportion that take up the option is 25% say, then we could determine a mortality assumption
for the lives that do not take up the option, q , as follows:
q x  0.25q x 10  0.75q 
The assumed mortality of those who do not take up the option, q  , will be lower than ultimate
level. For example, we may find that q   q x 3 .

Chapter 23 Summary
Investment guarantees
Examples of investment guarantees include:
 guaranteed minimum maturity values for both unit-linked and non-linked
endowment contracts
 guaranteed minimum surrender values for both unit-linked and non-linked contracts
 the ability to convert a lump sum into an annuity or vice versa on guaranteed terms.
An insurance company needs to be able to model the investment guarantee in order to

quantify the extra liabilities that it will incur when the guaranteed amount exceeds the
earned asset share. Liabilities can be assessed using:
 Option-pricing techniques: the value of the liabilities will be similar to the cost of a
derivative which covers a similar guarantee or option to that which the company is
offering.
 Stochastic simulation: the extra sums likely to be needed under the guarantee can be
modelled stochastically by running a simulation of investment returns thousands of
times.
Mortality options
Many life assurance contracts contain options whereby the policyholder can choose to
extend the term or increase the level of cover at normal premium rates, ie without providing
further medical evidence. To the extent that the option might be exercised by someone in
poor health, the assurance company will bear a cost: the difference between the ordinary
premium rate granted under the terms of the option and that which would have been
granted had the life been underwritten.
Mortality options can be valued using cashflow projections. Assumptions are required for:
 option take-up rate
 benefit chosen
 mortality of those that exercise the option and those that do not
 expenses relating to the option.



23.1 A life office offers a without-profits endowment assurance policy. At the maturity date of the
Exam style
policy, a policyholder may effect a whole life without-profits policy at normal premium rates
without medical evidence, for a fixed sum assured.
(i) Outline how cashflow projections could be used to calculate the extra premium to be
charged for this mortality option. [3]
(ii) Describe how the cost of the option is impacted by the proportion of policyholders that
exercise the option. [3]
[Total 6]
23.2 A deferred annuity policy will pay a guaranteed amount of £20,000 annually in advance from age
60. Alternatively, at age 60, the policyholder may choose to take a lump sum payment of
equivalent value, which must be used to purchase an annuity from another insurance company at
that time.
The lump sum is guaranteed to be the higher of:

 the equivalent value calculated according to market rates of interest available at the time
of conversion
 £250,000.
The company expects interest rates to have the following probability distribution at the time of
this policyholder’s 60th birthday:
 P(i = 4% pa) = 0.6
 P(i = 6% pa) = 0.4.
Calculate the expected cost of this option to the insurance company as at age 60, assuming that:
 the company bases its annuity rates on AM92 ultimate mortality
 policyholders always take the lump sum option when it is financially advantageous for
them to do so.

23.3 A life office has just sold a single premium deferred pension policy to an individual aged exactly
45. This policy guarantees to pay a cash sum of £200,000 on their 60th birthday, which must be
used to buy a whole of life annuity at that time. The policy also carries an annuity option whereby
the policyholder can elect to receive a pension of £15,000 per annum payable monthly in advance
from the same date, until their death.
The office invests the single premium such that the value of related assets on the policyholder’s
60th birthday will be normally distributed with a mean value of £250,000 and a standard deviation
of £50,000. It also believes that the annual interest rate, i, which will be available on the
policyholder’s 60th birthday is a random variable where
 i = 0.04 with probability 0.25
 i = 0.06 with probability 0.50
 i = 0.08 with probability 0.25.
The distribution of the value of assets on the policyholder’s 60th birthday is independent of this
annual interest rate, i.
The basis used for setting the annuity rates is:

 mortality: AM92 ultimate
 interest: i per annum
 expenses: nil.
Calculate the probability that the value of assets on the policyholder’s 60th birthday is less than
the cost of providing the annuity benefit, assuming the policyholder is alive at age 60.
23.4 An insurer is proposing to offer a two-year single premium without-profits term assurance policy
Exam style
to lives aged 50 exact. The basic sum assured is for £50,000.
The policy carries an option whereby, in exchange for an additional premium paid at the outset,
the policyholder can choose to increase the sum assured to £75,000 at the end of the first year of
the policy, on normal premium rates and without showing evidence of health.
In all cases the death benefit is paid at the end of the year of death.
The calculation basis for the additional option premium is as follows:

 normal mortality = AM92 Ultimate (assumed for normal premiums, and for all lives while
not taking the option, and for non-takers after the option date)
 mortality of policyholders choosing the option, following the option date = 200% of AM92
Ultimate
 proportion of policyholders surviving to the end of Year 1 who choose the option = 30%
 interest = 7% pa
 ignore expenses and the cost of setting up reserves.
Calculate the additional single premium payable at the start for a policyholder who requires the
option. [8]

23.1 (i) Method
The cashflows arising from the whole life contract would be projected. [½]
These cashflows would include the sum assured payable on death … [½]
… and the premiums payable throughout the term of the whole life contract, … [½]
… to the extent to which the option is assumed to be taken up. [½]
The premium for the whole life contract would be based on the expected premium rates that
would be charged to standard lives … [½]
… as at the maturity date of the endowment assurance. [½]
The projections should also allow for any expenses incurred, including the administration of the
option. [½]
As the purpose of the valuation is pricing the option (rather than setting the liability) then
allowance should also be made for the additional reserves that should be held, both before and
after exercise. [½]
The cost of the option calculated above would then be respread over the premiums payable by all
policyholders over the term of the endowment assurance policy. [½]
[Maximum 3]
(ii) Impact of the proportion that exercise the option on the cost of the option
If a life, who is in good health and who would be expected to satisfy normal underwriting
requirements, exercises the option, then the option will generate little or no additional costs. [½]
The exercise of the option by lives in poor health will generate considerable additional costs. [½]
So the total cost of the option depends on the health status of those who choose to exercise the
option … [½]
… and the proportion of lives who choose to exercise the option. [½]
In general, the smaller the proportion who exercise the option, the worse will be the subsequent
mortality experience of those exercising the option. [½]
If a substantial proportion exercise the option, then their subsequent mortality experience will on
average be less extreme. [½]
The maximum possible cost occurs when every person with mortality worse than the standard
rate takes up the option. This will include the cost to the company of insuring the few very high
risk lives at standard premiums, plus the costs of insuring the other medium-high risks on the
same terms. [1]
[Maximum 3]

23.2 If interest rates are 4% pa, then the normal lump sum would be:
@4%
20,000
a60  20,000  14.134  282,680
As this is higher than the guaranteed amount, then there is no cost to the company of meeting
the guarantee.
If interest rates are 6% pa, then the normal lump sum would be:
@6%
20,000
a60  20,000  11.891  237,820
The cost of the guarantee is then:
250,000  237,820  12,180
So the expected value of the cost of the guarantee is:
 E C i   P(i)  0  0.6  12,180  0.4  £4,872

i
23.3 We have to find out the probability that, at age 60, the value of the assets will be less than the
(expected present) value of the liabilities.
Define A as the random variable denoting the value of the assets at age 60. So the question tells
us that:
A ~ N(250,000,(50,000)2 )
Define L as the random variable denoting the value of the liabilities at age 60. If there were no
annuity option, then this value would always be £200,000.
This is because the company would calculate the annual amount of annuity as (for example)
200,000
, which has an expected present value of:
(m)@ i
a60
200,000 (m)@ i
 
a60  200,000
(m)@ i
a60
The actual expected cost of the annuity will be higher than this if the expected present value of an
annuity of £15,000 pa (payable monthly in advance) and discounted at rate i is greater than
£200,000, because in this case the guaranteed annuity would then be payable. This will happen if:
(12)@ i 200,000

a60 
15,000
@i 200,000 11
ie: a60 
   13.7917
15,000 24

@4% @6%
From the Tables we see that 
a60  14.134 and 
a60  11.891 . Since discounting at a higher
@8% @6%
rate of interest reduces the present value, it follows that 
a60  a60 . So the guarantee will
only bite if i  4% .
We can now work out the values of L, conditional on the value of i . These are:
For i  0.04 :
(12)@ 4%  11 
L  15,000 
a60  15,000  14.134    £205,135
 24 
For i  0.04 (that is, for i  6% or i  8% ):
L  £200,000
as explained above.
The probability of interest, that is the probability that A  L , is therefore also conditional on i .
We shall write this conditional probability as:
PA <L i
So:
P  A  L i  4%   P  A  205,135
 205,135  250,000 
  
 50,000 
   0.8973   1    0.8973 
 0.1848
Similarly:
P  A  L i  4%   P  A  200,000 
 200,000  250,000 
  
 50,000 
   1  1   1 
 0.15866

Finally, our overall (unconditional) probability that A  L is:
P  A < L i  P i 
i
 P  A < L i  4%   P  i  4%   P  A < L i  4%   P  i  4% 
 0.1848  0.25  0.15866  0.75  0.1652 .
23.4 We start by drawing up a life table for those who don’t take up the option, based on AM92
Ultimate. Using a radix of 100,000, we get:
x lx dx qx
50 100,000 250.80 0.002508

51 69,824.44 196.137 0.002809
52 69,628.30
where l51 excludes the 30% of the survivors at age 51 who then take the option. The number
who take the option is 29,924.76. [2]
We now calculate the corresponding table values for those who take the option, using 200% of
AM92 mortality (in this table, l51 is the number of people taking the option from the previous
table):
x lx dx qx
51 29,924.76 168.12 0.005618

52 29,756.64
[1]
We now calculate the premiums payable by those not taking the option. The initial premium paid
at the start of the policy is:
P0  50,000 [0.002508v  (1  0.002508)  0.002809v 2 ]  239.563 [1]
The premium paid at the start of the second year by the option-takers is:
P1  25,000  0.002809v  65.63 [1]

The total value of the benefits is:
PV  50,000  0.002508v  50,000  0.00196137v 2  75,000  0.0016812v 2
 312.98
[1]
The total value of the premiums is:
239.563  65.63  0.2992476v  257.92 [1]
So the value of the option (premium) is 312.98  257.92  £55.06 [1]

[Total 8]





using the course.

SP2-24: Reinsurance (1) Page 1
Reinsurance (1)
Syllabus objectives
3.2.2 Describe the different types and structures of reinsurance.

Page 2 SP2-24: Reinsurance (1)
0 Introduction
In this chapter and the next we see how a life insurance company can help to control its exposure
to risks (in particular mortality risks), and help with its financing, using reinsurance.
We begin in this chapter with a description of the main types of reinsurance contract. In the next
chapter we consider the reasons for buying reinsurance and the factors the insurer should
consider when determining how much and what type of reinsurance to buy.
Reinsurance is an arrangement whereby one party (the reinsurer), in consideration for a

premium, agrees to indemnify another party (the cedant) against part or all of the liability
assumed by the cedant under one or more insurance policies, or under one or more
reinsurance contracts.
An insurer that issues policies directly to the public is often referred to as the direct writer. The
direct writer may therefore transfer (cede) part of its liability for these contracts to a reinsurer,
which may further cede part of its liability to a second reinsurer, and so on. The first of these
transfers would be called a cession, while subsequent transfers may be called retro-cessions.
An insurance company that acts as a reinsurer may also be a direct writer in its own right. In the
main, however, the bulk of reinsurance cessions (that is, from direct writers) is made largely with
specialist pure reinsurance companies. These companies combine their reinsurance services with
the provision of valuable technical assistance to their direct-writing clients.

1 The main types of reinsurance contract

In this section we cover the main types of reinsurance contract:
 original terms
 risk premium
 excess of loss, eg:
– catastrophe
– stop loss
 financial
 facultative and obligatory.
Some of these headings overlap, as we shall see below.
1.1 Original terms (coinsurance) reinsurance

In the UK and Ireland this method is usually referred to as ‘original terms’ reinsurance,
whereas in most other countries it is known as ‘coinsurance’.
This method involves a sharing of all aspects of the original contract.
Hence, the premium is split between the insurer and reinsurer in a fixed proportion and any claim
is split in that same proportion. This will generally mean that the reinsurer shares in full the risks
of the policy, including the risks of investment and early lapse.
The cedant will provide the reinsurance company with the premium rates (known as retail
rates) it is using for the class of business it wishes to reinsure.
This is so that the reinsurer can check that the insurance company’s premium rates are adequate.
The reinsurance company will determine the rates of reinsurance commission it is prepared
to pay to the cedant for the business.
Question
List the main factors that the reinsurer would consider when deciding on how much reinsurance
commission to offer.

Solution
The factors it might consider include:

 the profit it expects to make from the business
 the risks that it is taking on (for example, how uncertain the future claim experience is
likely to be)
 how much it wants to obtain the business, taking into account competition from other
reinsurers.
Factors on which the above also depend include:

 the expected future mortality experience
 the extent of uncertainty associated with the future experience and with the insurer’s
pricing basis
 the expected level of the reinsurer’s other expenses
 the cost of the reinsurer’s capital
 the reinsurer’s profit criteria, including its required return on capital
 the insurer’s premium rates
 the quality of the insurer’s underwriting
 the likely persistency of the insurer’s business
 the extent of any recovery of reinsurance commission that would be paid back when a
policy lapses
 the extent of any profit-sharing there may be between the two parties.
In determining the reinsurance commission, the reinsurer will have regard to its likely future
experience and its knowledge of the quality of the cedant’s underwriting. It will also want to
make a profit on the business.
The reinsurance commission will usually cover – in respect of the reinsured portion of the
policy – the commission that has been paid by the cedant and part or all of the cedant’s
other initial expenses.
The reinsurance commission therefore determines the overall net cost of the reinsurance contract
for the cedant (the higher the commission, the less the cost). The insurer will then have to decide
whether it wants to accept the terms that the reinsurer is offering, or whether it should be
looking to place its business with a different reinsurance company (who may be offering higher
commission and/or be more attractive in some other way).
Question
Outline other ways in which a reinsurer might be more attractive to the insurer in this context.

Solution
The reinsurer might:

 have greater financial strength and so be less likely to default
 offer broader coverage, for example, to accept higher sums assured or older-age
applicants
 offer a profit-sharing arrangement, or a more generous one.
Specifying the amount to be reinsured

There are two ways in which the amount to be reinsured can be specified:
 individual surplus – the reinsured amount is the excess of the original benefit over
the cedant’s retention limit on any individual life
 quota share – a specified percentage of each policy is reinsured.
A company is likely to adopt a mixture of the above, retaining for itself a percentage of each
policy up to a maximum retention.
Example
Suppose the direct writer has a policy with sum insured $100,000.
With a straight individual surplus arrangement, supposing the direct writer has a retention of
$40,000 (it is prepared to accept up to $40,000 of ‘risk’ on that policy), the reinsurer would
reinsure $60,000.
With a simple quota share arrangement – say 30% – the reinsurer would reinsure $30,000.
Consider now a combined arrangement, assuming a 30% quota share subject to a retention of
$40,000. Looking at the quota share element, the direct writer would retain $70,000 (70% of the
sum insured). But the $70,000 is subject to the maximum retention of $40,000. So the company
retains only $40,000, and reinsures $60,000.
The critical difference between the two forms is that for quota share the same percentage (in the
example above 30%) is used for every policyholder, no matter how big the sum assured. For
individual surplus it is the fixed retention limit that is the same, so for a policy with a sum assured
of £200,000 the £40,000 would be fixed and hence 80% of the risk would be ceded.
1.2 Risk premium reinsurance

Here the reinsurer does not share in the office premium of the ceding office but charges a
specific premium for the risk, which may be level over the term of the policy or may vary
annually with the probability of claim.
The cedant reinsures part of the sum assured or the sum at risk, ie the excess of the benefit
payable over the reserve, on the reinsurer’s risk premium basis.

The risk premium is very similar to the risk charge for additional death benefits under unitised
contracts. The risk premium will essentially reflect the probability of claim over the coming year
(or month, if appropriate) multiplied by the amount of benefit covered by the reinsurer. So the
risk premium for a policy for policy year t could be calculated as:
RPt  qRx t Bt
where qRxt is the (reinsurer’s) risk premium rate and Bt is the reinsured amount at time t.
This risk premium will pay for the reinsurance for the given year (or month), at the end of which
another risk premium will be payable for the next year (or month), and so on. The risk premiums
will therefore change in amount each year (or month).
Question
Give possible reasons why the risk premiums change over the term.
Solution
There are a number of possible reasons:

 If the reinsurance is based on the sum at risk, then the sum at risk is likely to have
changed over the period, leading also (possibly) to a change in the reinsured amount at
risk.
 The risk premium rate (ie premium per £ of reinsured benefit) will change because it will
be based on the (older) age of the policyholder. (Age-related changes only tend to
happen on policy anniversaries.)
 If the reinsurance pricing basis is not guaranteed, then the risk premiums may change as a
result of the reinsurer carrying out a pricing review.
The reinsurance company determines its risk premium rates by assessing the likely
experience of the business it is to reinsure and then adding expense and profit margins.
So, in our formula, the risk premium rates qRxt will include loadings for expenses and profits.
Alternatively, these loadings could be added separately, ie as a per-policy amount (which would
be particularly appropriate for the reinsurer’s fixed expenses).
It may or may not guarantee these rates for the term of the policy.
Whether or not the terms of the reinsurance (including the risk premium rates to charge) are
guaranteed for the term of the policy, will depend on the individual agreement made between
the two parties. (This agreement is called the reinsurance ‘treaty’ – more about this later.) It is
likely that the insurer would wish to have similar rate guarantees with its reinsurer as it is
providing to its policyholders. So, if the reinsurance is in relation to a term assurance contract
that has been issued on fully guaranteed terms, then it is most likely that the insurer will wish to
pay reinsurance premiums that are fully guaranteed for the whole term of the contract.

There may also be some form of profit participation, whereby a share of any profits to the
reinsurer from the portfolio reinsured are paid back to the direct writer.
With a risk premium structure, changes in the insurer’s premium rates will not necessarily
require changes in reinsurance rates. It therefore gives the insurer greater freedom to
respond to competitor changes in premium rates.
In comparison, with an original terms structure the premium received by the reinsurer will change
whenever the insurer changes its premiums. The reinsurer may therefore impose limits on the
extent that the insurer can change its premium rates (or may change the reinsurance terms,
eg the rate of reinsurance commission).
Under a ‘net level premium arrangement’, the reinsurer spreads the risk premiums so that
they are level over the term of the contract. This method has the advantage that the insurer,
if desired, can simply load the reinsurance charges to obtain the premium payable by the
policyholder.
The approach applies mostly to protection products.
Question
An insurer sells term assurances in a competitive market.
Explain why a net level premium arrangement might be more suitable than original terms
reinsurance.
Solution
Competition in retail insurance markets will tend to bring down the price of insurance products. If
margins become too thin then the original terms approach may even leave the reinsurers with
loss-making premiums. The result is that insurers will find it increasingly difficult to obtain the
reinsurance protection that they need, at least on an original terms basis, so they end up having
to increase their premiums to satisfy their reinsurers. This puts a significant constraint on an
insurer’s pricing strategy.
The risk premium approach separates the insurer’s and reinsurer’s premium rates, and also
simplifies the insurer’s pricing process. So, a direct writer can decide to charge more keenly-
priced premiums to its policyholders if it so wishes, while still able to obtain the necessary
reinsurance (on the reinsurer’s terms).
As the risk premium is a level amount, the reinsurer’s premium rates are known in advance, so
the insurer can price its products with full information. As the insurer will always be looking to
make a profit long-term, then it is likely that the insurer will not diverge too far from the
reinsurer’s rates for its own basis anyway. And competition within the reinsurance market should
ensure that the reinsurance premium rates are themselves quite keenly priced in the first place.

Specifying the amount to be reinsured

As stated in the Core Reading at the start of this section, the amount to be reinsured can be based
on either the full sum assured or just the sum at risk. Where the sum at risk is used, the reserves
used to work out the sum at risk will be specified in the reinsurance treaty. These will not
necessarily be exactly the same as the insurer actually holds, but will be of the correct order of
magnitude.
Non-temporary life insurance contracts, such as whole life and endowment assurances, will
usually be reinsured on a sum-at-risk basis, whereas for term assurance contracts the reinsurance
will usually be based on the full sum assured.
Question
Explain the likely reasons for this.
Solution
Here we have to think about the likely reason for risk premium reinsurance – ie to reduce the
insurer’s exposure to adverse mortality experience for the best possible cost.
Under sum-at-risk reinsurance the insurer retains the full reserves under its contracts. The
reserve does not constitute any risk to the insurer from mortality, as the reserve needs to be
available whether a policyholder dies or survives during the year. So, reinsurance that covers any
part of the insurer’s reserves is a waste of money, at least from the view of covering mortality
risk. Also, such reinsurance would allow the reinsurer to set up its own reserves for its share of
the risk, and so would allow the reinsurer to earn investment profits at the expense of the insurer.
Hence, it is (usually) in the insurer’s best interests to retain 100% of the reserve if it can do so.
(The exception to this would be if the insurer wishes to share the investment risk as well as the
mortality risk with the reinsurer.)
On the other hand, sum-at-risk reinsurance is somewhat more complex to administer, which
implies greater expense.
The reason for the differences between the two product types therefore relates to the relative
size of the reserves. For term assurances, the reserves are tiny compared with the size of the sum
assured, so that reinsurance benefits and risk premiums will be very similar whether or not the
sum-at-risk approach is used. So the advantages of the sum-at-risk basis would not justify the
extra hassle involved.
On the other hand, endowment and whole life assurances develop significant reserves relative to
the sum assured. Here, using the sum at risk makes a significant difference to the cost of the
reinsurance, and so would usually be preferred.

The part to be reinsured can be on an individual surplus or a quota share basis.
This is identical to the case for original terms reinsurance.
For individual surplus, the excess (if any) of the sum at risk or sum assured above a specified level
is reinsured.
For quota share, a specified percentage of the sum at risk or sum assured is reinsured.
Question
Sketch the development of reserves, risk retained and amount reinsured for a regular premium
without-profits endowment assurance policy, under an individual surplus arrangement based on
the sum at risk.
Solution
Individual surplus, based on sum at risk
The following graph shows the development of the direct writer’s retention and the amount
reinsured for a twenty-year without-profits endowment assurance policy of sum assured 100 and
retention 40:
100
90
80
70
60
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Reserve Risk retained Reinsured

Question
Sketch the development of reserves, risk retained and amount reinsured for a regular premium
without-profits endowment assurance policy under a quota share arrangement based on the sum
at risk.
Solution
Quota share, based on sum at risk
The following graph shows the development of the direct writer’s retention and the amount
reinsured for a twenty year without-profits endowment assurance policy of sum assured 100 with
a quota share of 60%:
100
90
80
70
60
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Reserve Risk retained Reinsured

Example
Imagine a policy with sum assured £100,000. The direct writing company has a retention limit of
£40,000 of sum at risk. Consider the situation when reserves are £5,000, when they are £30,000
and when they have increased to £80,000.
With individual surplus, the direct writer retains £40,000 of risk in the first two instances. So the
total retention is £45,000 (when reserves are £5,000) passing £55,000 of risk to the reinsurer, and
£70,000 (when reserves are £30,000) ceding £30,000 of risk. Then, when reserves are £80,000,
the remaining sum at risk is £20,000, which is below the direct-writing company’s retention limit.
The company therefore does not pass any risk on to the reinsurer.
What would happen under quota share? Assuming a percentage of 50%: initially, when reserves
are £5,000, the sum at risk is £95,000 and the direct writer reinsures 50% of this (ie £47,500),
keeping £52,500 of sum insured. When reserves are £30,000, the direct writer cedes £35,000 and
retains £65,000. With reserves at £80,000, the direct writer cedes £10,000 and retains £90,000.
We’ll briefly summarise the key differences between the risk premium and original terms
reinsurance methods, as these are particularly important for life insurance.
Original terms (coinsurance) reinsurance
The insurance company sets the premium, and the reinsurance premium is in direct proportion to
this. The amount of commission agreed with the reinsurer sets the price of the reinsurance
arrangement, and is usually very significant.
Level risk premium reinsurance
The reinsurer sets a level premium for its share of the risk, based on its share of the full sum
assured. The insurer then calculates its own premium rates in the knowledge of the reinsurance
premiums it will be paying. The reinsurance commission is usually not significant.
Variable risk premium reinsurance
The reinsurer sets the reinsurance premium rate. The risk premium operates as a recurring single
premium, with each premium covering the immediate period of risk. The reinsurer may cover a
part of the full sum assured or a part of just the sum at risk. The risk premiums will vary from
period to period due to changes in the sum at risk, age of policyholder, or as a result of rate
reviews. The reinsurance commission is usually not significant.
For all three methods, the split between the reinsurer’s and insurer’s shares of the benefits can
be specified using either quota share or individual surplus arrangements.

1.3 Excess of loss reinsurance

Excess of loss reinsurance can be enacted on a risk basis, where the reinsurer pays any
loss on an individual risk in excess of a predetermined retention.
This kind of reinsurance would be used where the loss (claim amount) at risk under a single policy
is of unknown amount until the loss actually occurs. At first sight, therefore, it would appear that
most individual life insurance contracts, such as term assurances, would not require excess of loss
reinsurance cover.
Question
Explain whether this is true.
Solution
An individual term assurance has a known claim amount, should the claim occur. The kind of
protection afforded by excess of loss reinsurance is therefore entirely provided by individual
surplus reinsurance, where, by knowing the reinsurer’s share of the risk in advance of the claim,
the insurer will pay a premium related directly to that share. Excess of loss would therefore not
be used for an individual policy.
However, insurers with term assurance portfolios (for example) can run into additional difficulties
if they incur an exceptionally large number of claims, even where the insurer’s retention under
each individual policy may be capped at an acceptable level.
For example, a sudden ‘run’ of claims could occur as a result of a localised accident involving
several of the insurer’s policyholders at once. This could happen where an insurer has a very
concentrated target market by territory, occupation, or most importantly, where it insures groups
of individuals who all work for the same employer, for example. An excess of loss reinsurance
contract could then cover the total claims occurring in the portfolio as a result of such an event,
and which would not be known in advance of the losses occurring.
However, excess of loss can also be enacted on an occurrence basis, where the aggregate
loss from any one occurrence of an event exceeds the predetermined retention.
This is exactly the situation that we have just described in the solution to the question above.
Most excess of loss reinsurance contracts last one year and need to be renegotiated each year.
The insurer will pay a risk premium to cover the reinsurer’s expected loss, allowing for expenses
and profit margins, at the start of the year of cover. The premium basis will include margins to
reflect the uncertainty of the reinsurer’s total claim amount.
Excess of loss is an example of non-proportional reinsurance, because the reinsurance premiums

paid are not in proportion to the actual claim amount paid by the reinsurer. For example, the same
reinsurance risk premium will have been paid whether the claim paid by the reinsurer turned out to
be £1m or £10m. Under individual surplus, on the other hand, the reinsurance premium would be
approximately ten times higher for a reinsured benefit of £10m compared with £1m. So both risk
premium and original terms reinsurance are examples of proportional reinsurance types.

An excess of loss treaty may be organised into different levels, or lines. The cedant may
retain the first £Xm of losses. The first £Ym over the £Xm will have one price. The next £Zm
over £(X+Y)m will have a different price, and so on. Different reinsurers may then take
different proportions of each line.
In life insurance, multiple lines of excess of loss reinsurance are mainly found for group life
policies.
Within lines, the reinsurer may also pay only a percentage of the loss actually incurred.
So, in the above example, the reinsurer may pay, say, 90% of the claims in excess of £Xm but less
than £(X+Y)m, and 80% of the claims in excess of £(X+Y)m but less than £(X+Y+Z)m. The insurer
would have to pay all other claims, including the excess of total claims over £(X+Y+Z)m, if
applicable.
Question
Determine the amount of claims paid by both insurer and reinsurer when X  2; Y  3; Z  5
(assuming the percentages above apply), if total losses covered by the treaty amount to:
(a) £4m
(b) £8m
(c) £12m.
Solution
(a) Total claims £4m
The claim is between £ Xm and £Ym .
The reinsurer pays 90% of  4  X  , ie 0.9  2  £1.8m .
The insurer pays the balance of £2.2m.
(b) Total claims £8m
The claim is between £( X  Y )m and £( X  Y  Z )m .
The reinsurer pays 90% of Y and 80% of  8  ( X  Y )  , ie 0.9  3  0.8  3  £5.1m .
(c) Total claims £12m
The claim is above £( X  Y  Z )m .
The reinsurer pays 90% of Y and 80% of Z, ie 0.9  3  0.8  5  £6.7m .

The main types of excess of loss reinsurance used in life insurance are catastrophe and
stop loss reinsurance.
Catastrophe reinsurance
The aim of catastrophe reinsurance is to reduce the potential loss to the cedant due to any
non-independence of the risks insured.
Question
State, with justification, the type of life insurance contract where this protection against non-
independence of risks is particularly useful.
Solution
Catastrophe insurance will be particularly important for group business, since each group contract
will involve perhaps hundreds of employees of one company who will work in the same physical
area. Some accident, such as an explosion, would probably affect a large number of these
employees.
The cover is usually only available on a yearly basis and must be renegotiated each year.
Question
Describe the advantages of this annual renegotiation to the reinsurer and to the direct writer.
Solution
This annual re-negotiation will allow the reinsurer to vary premiums (for instance, if the premium
is expressed as ‘some factor × sum at risk’, that factor could be changed) if it is felt that the risk
posed by the direct writer’s portfolio has changed. Such a change would often be made in the
year after any claim by the direct writer in respect of the catastrophe cover.
This freedom should result in lower rates, to the advantage of the direct writer.
In addition the direct writer will have the freedom to shop around different reinsurers.
The reinsuring company will agree to pay out if a ‘catastrophe’, as defined in the
reinsurance contract, occurs. There is no standard definition of what constitutes a
catastrophe, but typically to qualify there needs to be a minimum number, say five, deaths
from a single incident with the deaths occurring within a specified time, say 48 hours of that
incident.
The sort of events that might give rise to a catastrophe payment are industrial disasters (an
individual life company might have a group life policy covering a particular employer) or perhaps a
train crash (which is less likely to affect an individual insurance company sufficiently to count as a
catastrophe, unless the company has a high proportion of its business in one geographical
location).

The reinsurance contract will also specify how much the reinsuring company will pay if a
catastrophe occurs. Typically, this might be the excess of the total claim amount, net of any
amounts already reinsured, over the cedant’s catastrophe retention limit. The reinsuring
company’s liability in respect of a single catastrophe claim would be subject to a maximum
amount and any amount above this would revert to the cedant. There is also usually a
maximum amount of cover per life.
Question
Suggest possible reasons for there being an upper limit on cover provided.
Solution
The reinsurer will want an upper limit because it will not want unlimited liability.
In addition the life insurance company may itself want to limit cover in order to reduce the cost of
reinsurance.
Legislation / regulation may impose the existence of such an upper limit.
The cover would usually exclude certain risks such as war risks, epidemics and nuclear
risks. However, separate catastrophe covers may be available for any excluded risks.
War and nuclear risks may also be excluded from the original life insurance policy.
These risks are extremely unpredictable with potentially massive volatility, and so require special
consideration. The premiums for the cover would themselves be quite volatile and subject to much
bigger margins than cover for other causes. This is why they are better covered separately, if
required.
Stop loss reinsurance

Stop loss reinsurance means that the reinsurer pays the aggregate net loss over the
predetermined retention for a portfolio over a given period, usually a year. In this way, the
portfolio’s loss to the cedant in any period is capped.
As with catastrophe reinsurance, the reinsurer’s liability would be limited to a specified maximum
amount, and the reinsurance would need to be renegotiated annually.
1.4 Financial reinsurance

This is often shortened to ‘Fin Re’.
Features of financial reinsurance

A wide variety of financial reinsurance contracts have been developed and used, although
all were devised primarily as a means of improving the apparent accounting or supervisory
solvency position of the cedant.
One feature of a financial reinsurance contract is that it tends to involve only a small
element, if any, of transfer of insurance risk from the cedant to the reinsurer.

Rather than transferring insurance risk (which is the aim of traditional reinsurance), financial
reinsurance aims to manage the insurer’s capital position.
Financial reinsurance is not effective under accounting or supervisory regimes where credit
can already be taken for the future profits and/or where a realistic liability must be held in
respect of the loan repayments. With the increased global movement towards realistic
reporting bases, as described in Chapter 20, these arrangements are likely to become less
common.
Risk premium reinsurance

The risk premium reinsurance method is one type of arrangement that has been associated
with a financing arrangement, whereby the reinsurer relieves the cedant of part of its new
business financing requirement. A straightforward loan from the reinsurer would not
achieve this, as the cedant would usually have to add the amount of the loan to its liabilities.
So, the ‘loan’ is usually presented as a reinsurance commission related to the volume of
business reinsured. The ‘repayments’ – spread over several years – are added to the
reinsurance premiums. The reinsurer takes into account the expected lapse experience of
the portfolio of reinsurances in determining the loan repayments.
Essentially all that is happening is that the reinsurer has increased its risk premium rates in future
years. In return for charging the company higher premiums in the future, the reinsurer is able to
offer the insurer a substantial initial commission. The ‘loan’ received over a period is therefore
equal to the total initial commissions payable under the arrangements over that time. As far as
repayments are concerned, the company simply agrees to pay more for its reinsurance protection
than it would normally have done.
Question
Explain which type of traditional (ie non-financing) reinsurance would achieve the same result.
Solution
An alternative method is to use original terms reinsurance. Here the reinsured proportion of the
full office premium becomes the reinsurance premium. In return for the large loadings that the
reinsurer receives through these premiums, a substantial total reinsurance commission would
usually be payable back to the insurer, producing a similar effect to that described above for risk
premium. Original terms is actually a more natural method for this purpose, as the premiums do
not need to be increased ‘artificially’ in order to cover the commission payments.
Contingent loan
This type of Fin Re is sometimes referred to as surplus relief or asset-enhancing reinsurance.
This alternative approach makes use of the future profits contained in a block of new or
existing business.

Regulation will usually require the insurance company to set aside assets (to cover reserves and
solvency capital requirements) that are well in excess of the realistic value of the liabilities. Over
time, this excess is expected to be returned to the company as profit.
These profits are often called the VIF as they relate to the value in force, ie the value of the profits
from the in-force business.
In this instance, the reinsurer again provides a loan to the cedant, but, as the repayment of
the loan is contingent upon the stream of future profits being generated by the business,
the cedant may not need to reserve for the repayment within its supervisory returns
(depending on the regulatory regime).
So, should the VIF profit not occur at all, the reinsurer would simply have made a capital gift to
the insurer equal to the amount of loan. If the VIF profit does emerge, then the reinsurer will
receive all of the specified loan repayments.
Compare this with a normal, non-contingent, loan. In the normal case, the assets would be
increased by the amount of the loan, but then the liabilities would have to increase as well to
show the expected cost of the future loan repayments. So, there would be no improvement to
the amount of surplus shown on the balance sheet.
With the contingent loan, the insurer does not have to show as a liability the value of the future
loan repayments, as these will only materialise should the VIF profit materialise (which is not
shown as an asset).
Question
In relation to contingent loan financial reinsurance:
(i) State, with a reason, whether the contingent loan repayments would be expected to be
higher or lower in amount than the equivalent non-contingent repayments.
(ii) Explain why little change would be expected to the realistic embedded value of the
insurer if this type of financial reinsurance is used.
(iii) Explain how the company’s risk would be expected to have changed as a result of this
reinsurance, with respect to the company’s realistic embedded value.

Solution
(i) The contingent repayments will be higher, because not all of them may be made.
(ii) The assets would be increased by the amount of the loan, which would increase the
embedded value. On the other hand, the expected present value of the profits will be
reduced on account of the loan repayments. We would expect the two to be broadly
equivalent in value, so that overall there should be little change.
Looking slightly more deeply, the loan repayments will include a profit margin for the
reinsurer, which should reduce the expected future profits from the portfolio for the
insurer. On the other hand, the value of the future repayments is reduced in the
embedded value calculation by discounting them at the risk discount rate, realistically
reflecting the possibility that that they will not be paid. Once again, this points to the
exchange of cash for profits to be broadly consistent, especially if we were to value
everything in a market-consistent way.
(iii) Without this reinsurance there is a risk that the VIF profit, which is included in the
embedded value, would not materialise. With this reinsurance, part of that profit has
been converted into current cash, making it certain. The company’s profits are therefore
subject to less uncertainty than before, so there is less risk.
1.5 Facultative and obligatory reinsurance

The term ‘facultative’ applied to the cedant’s part of the agreement means that it is free to
place the reinsurance with any reinsurer. Similarly, so far as the reinsurer is concerned,
facultative means that it may accept or reject the reinsurance offered. The term ‘obligatory’
indicates the removal of this freedom of action.
The agreement between cedant and reinsurer may be:
Not necessarily formalised, though treaty

(a) facultative / facultative
documents might be prepared anyway
(b) facultative / obligatory Formalised agreements in a treaty

(c) obligatory / obligatory between the two parties
In the first column of the table above, the first word refers to the situation from the direct writer’s
perspective, the second word denoting the reinsurer’s position. Thus ‘facultative / obligatory’
means that the direct writer can choose whether or not to reinsure the risk; the reinsurer is
obliged to accept the risk if the direct writer decides to cede it.
A type (a) agreement might not be written under treaty, although to avoid having to make up
treaty documents at the time these may be prepared in advance.
The treaty is a legal contract governed by contract law, which will include international trade
and shipping agreements and a large body of case law.

Arrangements (b) and (c) above might seem potentially risky to the reinsurer, if the reinsuring
company is obliged to accept whatever the direct company throws at them. However, the treaty
will only apply to certain specified products, and will only have been accorded when the reinsurer
is satisfied about the type of business that the direct writer is taking on (ie satisfied about the
direct writer’s underwriting procedures). It may be that the direct writer is taking on some
unattractive business – for example, small whole life policies sold via direct marketing with no
medical underwriting. This is not a problem for the reinsurer if it is aware of this and can
therefore charge appropriate reinsurance premiums (or pay appropriately low reinsurance
commission in the case of original terms reinsurance).
The great advantage of treaties where the reinsurer’s acceptance is obligatory is that the direct
writer can write large policies without having to refer back to the reinsurer – a process which
might impede the sale of such policies due to the delay in acceptance, quite apart from the
associated administrative hassle.


Chapter 24 Summary
Reinsurance involves a direct writing company ceding business to another life insurance
company, the reinsurer. The reinsurer will have no contact with the policyholders.
Original terms (coinsurance) approach

Original premiums and benefits are proportionately shared. Reinsurance commission is
significant and determines the price of the reinsurance.
Level risk premium approach

The reinsurer sets a level premium rate. The insurer can price the contract taking this into
account.
Variable risk premium (recurring single premium) approach

The reinsurer sets the reinsurance premium rates. Risk premiums change from year to year,
and the rates may or may not be guaranteed for the policy duration. Reinsurance benefits
are based on a share of the full sum assured or sum at risk.
All the above types may be quota share or individual surplus. Quota share is a constant fixed
proportion reinsured; individual surplus has a maximum retention per life.
Excess of loss – non-proportional reinsurance types

Catastrophe reinsurance – shares in the total claims above a threshold from multiple claims
from a single event. Renewable annually. Covers non-independent risks, eg group life
insurance. There may be multiple lines for group business.
Stop loss reinsurance – covers the excess of all aggregate claims in a year over a threshold,
up to a maximum.
Financial reinsurance
Financial reinsurance can help the cedant to relieve part of its new business financing
requirement. It can be structured like a loan, receiving either a lump sum or reinsurance
commissions with repayments incorporated into the reinsurance premium or paid out of
future profits.
Facultative and obligatory

Reinsurance is usually codified by treaty. Facultative implies freedom of action, obligatory
means removal of this freedom.



24.1 Describe the aim and main features of catastrophe reinsurance from the point of view of a ceding
Exam style
life insurance company. [4]
24.2 Describe the various ways in which a life insurance company can manage its capital position using
Exam style
reinsurance. [7]
24.3 (i) Describe the main ways in which risk premium reinsurance arrangements can vary. [5]
Exam style A life insurance company has asked a reinsurer to quote for risk premium reinsurance of a
standard whole life assurance contract. The terms it has been offered are on a guaranteed rate
basis and include a profit participation clause. With this, the reinsurer’s profit as defined by a set
formula is returned to the direct writer at the end of every year (unless it is negative, in which
case it is carried forward to be offset against future profits). The profit formula is:
Profit = reinsurance premium × 90% – reinsurance commission – reinsurance claims paid.
(ii) Discuss the advantages and disadvantages to the direct writer of these terms. [5]
[Total 10]
24.4 A new life company is planning to enter the without-profits life assurance market for the first
Exam style
time.
(i) List the types of reinsurance cover available to reduce the company’s exposure to large
claims on individual policies. [2]
(ii) Outline how the sum assured payable to the policyholder is apportioned between the
direct writer and the reinsurers for each of these types of reinsurance. [3]
(iii) Suggest, with brief reasons, the most appropriate type or types of reinsurance cover to
satisfy each of the following aims.
(a) To reduce new business strain from conventional term assurance policies.
(b) To reduce claim fluctuations from conventional without-profits and with-profits

whole life assurance policies.
(c) To reduce new business strain and claim fluctuations from unit-linked insurance
policies with high levels of death benefits. [8]
[Total 13]



24.1 The aim is to help protect the ceding company from large adverse fluctuations in claim costs, in
order to help stabilise the company’s profit flow and ultimately to protect the company from
ruin. [1]
Catastrophe reinsurance cover relates to the aggregate losses arising from a single event (a single
industrial accident, plane crash, etc), and is therefore designed to cover the risks arising from non-
independent claims. [1]
This is most likely to occur for group life assurance contracts, where there is a relatively high
probability of more than one member being killed at the same time by an industrial accident, for
example. [½]
The reinsurance terms would include a specific definition of what constitutes a catastrophe claim
event (for example, when n or more death claims occur within h hours of the causal incident
occurring). [½]
The reinsurer will pay the excess over a certain threshold amount of loss resulting from a single
event, subject to a maximum limit. There would also usually be a maximum payment with respect
to any individual life involved. [1]
Claims payable under the contract would be assessed after any other individual reinsurance
recoveries had been made. [½]
The reinsurance premium would need to be renewed and renegotiated on a yearly basis. [½]
The cover would normally exclude war risks, epidemics and nuclear risks. [½]
[Maximum 4]
24.2 Contingent loan
This approach makes use of the future profits contained in a block of new or existing business. [½]
The reinsurer provides the insurer with a cash loan. [½]
In return, the insurer will repay the loan to the reinsurer over a number of years. [½]
The repayments are contingent on the profits emerging from the business, and will not be paid if
the profits do not materialise. [½]
The assets in the company’s supervisory balance sheet will be increased by the amount of the
loan. [½]
However, in some regulatory regimes the future loan repayments will not be included in the
liability value shown in the balance sheet. [½]
The net effect is then to increase the value of the free assets in the company’s balance sheet, so
increasing capital. [½]

However, this form of financial reinsurance is not effective under accounting or supervisory
regimes where credit can already be taken for the future profits and/or where a realistic liability
has to be held in respect of the loan repayments. [1]
Risk premium reinsurance can be used to reduce new business strain. [½]
The reinsurer provides a ‘loan’ to the insurer in the form of reinsurance commission based on the
volume of business reinsured. [½]
The loan repayments are spread over a number of years as additions to the reinsurance
premiums. [½]
The reinsurer takes into account the expected lapse experience of the portfolio in determining the
loan repayments. [½]
So, in a similar way to the contingent loan, the assets are increased by the reinsurance
commission, but the liabilities may not need to increase to cover the additional reinsurance
premiums. [½]
Again, this form of financial reinsurance is not effective under accounting or supervisory regimes
where a realistic liability has to be held in respect of the loan repayments (ie the additional
reinsurance premiums). [½]
Original terms reinsurance
Original terms reinsurance will reduce new business strain on individual policies as they are
sold. [½]
A quota share basis will usually be used as this means that the benefits of the reinsurance will
apply to all contracts (both big and small). [½]
The large initial commission will enhance the company’s free assets, by capitalising the future
expense loadings that the insurer is passing to the reinsurer through the reinsurance
premiums. [1]
Further reduction in strain may be achieved if the reduction in net reserves exceeds the
reinsurance premium, though the extent of this will depend on regulation. [½]
[Maximum 7]
24.3 (i) Variations in risk premium reinsurance
The cession can be based on either the whole sum assured or on the sum at risk (the difference
between the sum assured and the policy reserve). [1]
Risk premiums can be paid as recurring single premiums, varying in amount each year to reflect
changes in risk and (sometimes) amount reinsured. [½]
When the reinsured sum assured is constant each year, risk premiums may alternatively be set as
level regular payments, ie a net level premium arrangement. [½]

The share of risk can be determined on either a quota share or surplus basis. [½]
Under the surplus method, the company’s retention remains at a constant amount. If the cession
is based on the sum at risk, the amount reinsured will reduce as policy reserves increase. This
could lead to the reinsurance running off before the end of the policy term for contracts such as
endowment or whole of life assurances. [1]
With the quota share method, the proportion of sum assured or sum at risk retained by the direct
writing company is held constant throughout the policy term. [½]
The reinsurance might be done on guaranteed rates, ie once a policy has been set in force and the
reinsurance initiated, all future reinsurance premiums in respect of that policy are defined as per
the treaty. [½]
Alternatively, the rates might not be guaranteed, and the reinsurer could vary them year by year
(or at less frequent intervals). [1]
The reinsurance could be on a ‘facultative / obligatory’ basis, ‘facultative / facultative’ or

‘obligatory / obligatory’, depending on whether the direct writer’s cession, and the reinsurer’s
acceptance, of the risk on any policy covered by the treaty is compulsory or optional. [1]
[Maximum 5]
(ii) Advantages and disadvantages of the proposed terms
The terms are guaranteed (ie fixed) so, if the mortality experience is lighter than originally
anticipated, the direct writing company will be paying more reinsurance premiums than it really
needs to. [½]
Although as far as new business is concerned it could renegotiate the treaty. [½]
This is, however, offset by the profit sharing arrangement, which is essentially an alternative to a
renewable rate structure, as both profits and losses are shared. [1]
It protects the direct writer from paying excessive premiums, overall. [½]
It protects the reinsurer from receiving excessively insufficient premiums. [½]
If there is some reinsurer’s profit, this is passed back immediately (net of expenses etc) to the
direct writer, while if there is a loss the direct writer waits before ‘paying’ extra premiums to the
reinsurer via a reduction in later profit participation payments. So the timing aspect works in the
direct writer’s favour (assuming positive interest rates). [1]
However the profit participation gives the reinsurer a 10% expense margin – this will probably be
a great deal more than their expenses (ie it is really a profit margin). [½]
Could only establish if this were penal or generous by reference to any competitors’ terms
available. [½]

The profit participation clause does represent a certain cost to the reinsurer, which is presumably
reflected in the premiums – if reasonably sure about the future mortality experience the insurer
might prefer to have lower premiums and no profit participation. [1]
[Maximum 5]
24.4 (i) Types of reinsurance to cover large claims on individual policies
Original terms or risk premium. [½]
Quota share or individual surplus. [½]
For risk premium only:

 cession based on full sum assured or sum at risk [½]
 level or recurring risk premiums payable (only the full sum assured version can be subject
to level risk premiums). [½]
[Total 2]
(ii) How apportioned between direct writer and reinsurer
Quota share based on full sum assured
An agreed proportion of each sum assured is taken by the reinsurer. The proportion is constant
for all policies within an agreed block of business. [½]
Quota share based on sum at risk
The direct writer retains the reserve. The sum at risk is apportioned such that the proportions of
sum at risk retained and reinsured stay constant at their initial levels. The proportion reinsured is
the same for all policies. [1]
Individual surplus based on full sum assured
The direct writer has a retention limit. It retains the whole sum assured on a policy where the
sum is below the retention limit. Where the sum assured is greater than the retention limit, the
direct writer retains a sum assured equal to the retention limit and the reinsurer takes any excess
over the retention limit. [1]
Individual surplus based on sum at risk
The reinsurer takes the amount of the sum at risk in excess of the direct writer’s retention, with
the direct writer also retaining all the reserve. The retained part of the sum at risk is kept
constant at its initially chosen level, with the reinsurer taking any excess. If the initial retained
amount of sum at risk eventually exceeds the total sum at risk, then the policy is fully retained by
the direct writer. [1]
[Maximum 3]

(iii)(a) Reduce new business strain from conventional term assurance
Probably best to use original terms quota share, with large initial reinsurance commission. [½]
Choose original terms as this will help with reducing supervisory reserves (in proportion to the
amount reinsured – regulations permitting – and is simpler than risk premium to administer). [½]
Choose quota share as new business strain is proportionate to total volume of business sold, and
is not affected by the size of individual policies. [½]
Alternatively, might use the net level risk premium arrangement, with increased initial
commission payments for increased risk premiums, again on a quota share basis. [1]
(iii)(b) Reduce claim fluctuations from conventional whole life policies
Risk premium individual surplus would do for both with- and without-profits forms, using a sum at
risk basis. [½]
The reserve is not at risk and becomes significant in size later in the policy term; basing the
reinsurance on the sum at risk only is much cheaper (for same amount of protection) than full
sum assured methods. [½]
Individual surplus would be chosen, as it is important not to retain more than a maximum
absolute sum at risk per policy, regardless of the total sum at risk under the policy. [½]
Could use original terms with deposit back for with-profits, as the total benefits (including any
bonuses added to benefits) will be reinsured as well as just the basic (initial) sum assured.
Deposit back is then needed, to protect the reinsurer from the investment risk of having to
replicate the direct writer’s bonus rates. [1]
(iii)(c) Unit-linked with large death benefit
Risk premium individual surplus based on the sum at risk is the obvious choice for reducing claim
fluctuations. [½]
This is because it allows the insurer to retain its full unit reserves (which do not constitute any
mortality risk), it keeps the insurer’s risk concentration below a maximum level and is
administratively straightforward for unit-linked. [1½]
Would need quota share risk premium with high initial commission to control new business strain
(higher risk premium rates would be charged to repay the initial commission over time). [1]
This will also help to reduce the company’s exposure to adverse claim experience. [½]
The insurer will end up with a combination of both types of reinsurance, which might look like (for
example):
‘40% of sum at risk plus balance of risk over £100,000’
and the same risk premium and commission rates would be used throughout. [½]
[Maximum 8]





using the course.

Reinsurance (2)
Syllabus objectives
3.2.1 Describe the purposes of reinsurance.

3.2.3 Discuss the factors that should be considered before taking out reinsurance.

0 Introduction
In this chapter we conclude our discussion of reinsurance.
In Section 1 we look at the reasons why an insurer would want to buy reinsurance. In Section 2
we look at the factors the insurer should consider to determine the amount and type of
reinsurance to buy. We apply these ideas in Section 3, where we consider a number of example
situations and discuss the reinsurance that would be appropriate in each case.

1 The reasons for reinsuring

Reinsurance allows an insurance company to transfer risks from its balance sheet on to
that of the reinsurer in a way that reduces the volatility and uncertainty of its future results.
Insurers require capital to hold risks, so by changing the amount of risk they retain, the
capital required by the insurer can be managed. The risks transferred can be insurance
risks (eg mortality), market risks, operational risks and capital planning risks, depending on
how the reinsurance treaty is constructed.
Apart from potentially using reinsurance to raise capital (covered in Section 1.4 of
Chapter 24), a life insurance company could use it to:
 limit the amount paid on a claim
 limit total claims payout
 reduce insurance parameter risk*
 reduce claim payout fluctuations*
 receive technical assistance*
 reduce new business strain*
 increase profits, return or risk-adjusted return on capital (by reducing risk cost or
increasing volumes)*
 reduce overall capital requirements by using a reinsurer’s capital*

Reinsurers may have lower capital requirements due to risk diversification or
regulatory position. This can result in overall lower capital requirements for the
industry and lower reinsurance prices.
 separate out different risks from a product

Products contain multiple risks, eg an immediate annuity primarily contains
investment and longevity risks. Reinsurance can disaggregate these risks, allowing
the cedant to optimise its risk management and capital requirements.
 allow aggregation of risks that the cedant cannot manage on its own, so allowing
manufacture of product lines.
For example, if the cedant had a large concentration of risk in a certain geographical
location then the cedant could reinsure this and still be able to write the business.
*These reasons will now be explained in further detail.
In doing so, we will consider the type or types of reinsurance that might be suitable in each case.

1.1 Parameter risk

There is a risk that the level of claims may be different from expected. This may be due to
incorrect pricing, underwriting failures, fraudulent activities, etc.
The ‘expected’ level of claims refers to the long-term expected (or average) rates of mortality
assumed in the company’s pricing basis. Even without any claim fluctuations, if the assumed
mean mortality assumptions turn out to be an underestimate of the actual experience, then the
company will make mortality losses. These losses are therefore described as arising from
parameter risk or, equivalently, pricing risk (ie where the insurer’s pricing basis turns out to be too
optimistic).
Quota share reinsurance could be used to share the parameter risk with the reinsurer,
although in practice the reinsurer would review the premium rates and the insurer’s control.
The reinsurer isn’t going to be happy to pick up the cost of serious failings in the insurance
company’s pricing and controls (eg underwriting controls and fraud detection). So the reinsurer
will use its expertise to help the company to improve its pricing and controls. However, even
after improvements have been made, some parameter risk will still remain and quota share will
allow this risk to be shared between the company and the reinsurer.
Question
Explain why individual surplus reinsurance would not be as useful for this purpose.
Solution
Imagine a scenario in which a company sold a large number of policies (such as term assurances),
all of which were for sums assured less than the company’s maximum retention limit. The
company would end up with no reinsurance. However, it might still have a very large parameter
risk (eg claims might be running at twice the rate that was assumed in the pricing basis, causing
large losses on the (large) portfolio).
If the company had used quota share (eg 70%), then it wouldn’t matter whether it sold large or
small policies, it would know that 70% of any adverse claim experience would be paid for by the
reinsurer, rather than by itself.
1.2 Claim payout fluctuations

For a cedant, one of the main ways in which reinsurance can help reduce risk is by reducing
the variance of the cedant’s claim costs.
The variance relative to the mean can be high because:

(a) there are a small number of contracts for very high levels of cover
(b) the lives insured are not independent risks.
The first of these points is often described as a ‘high concentration’ of risk.

If we had a million identical and independent policyholders with sums assured of $1 each, then the
variance of the total claim cost would be:
1m 1m
 Var 1  D    (1)2 Var  D   1,000,000  Var (D)
i 1 i 1
where D is the random number of deaths from one policy.
If, on the other hand, we had either:

 one policyholder with a sum assured of $1,000,000, or
 one million policyholders with sum assured $1 each but who would all die at the same
time, or not at all,
then the variance of the total claim cost would be:
Var 1,000,000  D   (1,000,000)2  Var  D 
while the mean claim cost would be the same in all three cases. This illustrates the potential
(though extreme) effect that these factors can have on the company’s mortality risk.
From the law of large numbers, the variance of claim payments will be high if there is a small
number of contracts. For this reason reinsurance against claim payout fluctuations is particularly
important for small companies.
For reason (a), the methods of reinsurance that could be used would be either the original
terms (coinsurance) or the risk premium method, usually on an individual surplus basis.
To avoid high risk concentration, and yet to maximise profits, the company would seek to reinsure
a high proportion of its very large risks, but none at all of its small cases. This is precisely what is
achieved using the individual surplus method.
For reason (b), the methods of reinsurance that could be used are catastrophe, stop loss or
other types of excess of loss reinsurance.
For example, catastrophe reinsurance will generally only pay out when more than a certain number
of claims occur as a result of a specified single event occurring, so it is specifically targeted at paying
non-independent claims.
(Catastrophe and stop loss reinsurance are technically both types of excess of loss reinsurance – as
described in Section 1.3 of Chapter 24.)

So, reinsurance will reduce the fluctuations in claim costs. Large fluctuations in claims costs can be a
bad thing for a number of (closely related) reasons:
 They may make the life company insolvent. This is an appreciable risk for small
companies, who therefore have an unavoidable need for reinsurance. For many larger
companies, however, mortality fluctuations will not plausibly threaten solvency. War
might do so, although the policy may have a war exclusion.
 They may reduce the excess of the value of the company’s assets over its liabilities (ie its
free assets) below the level desired by the company.
 They may reduce in some years the rate of return on free assets below the level desired
by the company.
 They may cause fluctuations in dividends to shareholders.
Reinsurance is not the only way to avoid the effects of payout fluctuations. Other options are:
 To set up a ‘payout’ or ‘mortality fluctuations reserve’. This is really just a pot of money
that the company sets aside. If claims exceed the expected level in any year, the company
will take the difference from the reserve; when claims fall short of the expected level, the
difference (a profit) will be added to the reserve. We see later how the two options of
reinsurance and establishing a mortality fluctuations reserve can be compared in an
attempt to decide how much reinsurance to arrange.
 To decline (turn down) high sum insured proposals. Generally, companies wish to avoid
doing this, and so one reason to reinsure is to avoid having to decline any business (and so
help sell more business).
1.3 New business strain

A cedant can use reinsurance to reduce the financial risk associated with new business,
either through an increase in its available capital or through a reduction in its financing
requirement.
Original terms reinsurance (usually on a quota share basis) or financial reinsurance can be
used (if the latter is effective under the applicable regulatory regime).
Financial reinsurance of an existing block of business will improve the balance sheet of the
company immediately, thereby increasing the available capital. So, more new business may be
written before the resulting new business strain leads to an unacceptably low solvency position.
Even without any specific financing arrangement, any original terms reinsurance (eg 40% quota
share) will mean some of the liability is passed to the reinsurer (40% in this case). This in turn
means that the company will need to hold lower statutory reserves (here it would be up to 40%
lower, although there may be limits on this imposed by the regulators). So, the new business
strain is reduced by around the same proportion. More significantly, the large initial reinsurance
commission will make a contribution to the company’s assets at the time of sale, effectively
discounting the future profits that will be tied up in the (large) reinsurance premiums.

Example
To illustrate how this might work, imagine a young company selling a twenty-year endowment
assurance policy. Suppose that initial expenses are 120% of the first year’s premium, renewal
expenses are 5% of subsequent premiums, and that statutory reserves amount to 85% of all
premiums paid (ignore interest).
The cashflow (premium – expenses – increase in reserves) on writing a $1,000 premium policy is:
Year 1 – $1,050
Years 2+ $100
The capital strain from this is too great for a young company to be able to write significant
amounts of new business without a huge amount of capital to finance the new business strain.
A reinsurer offers an original terms treaty, offering the following commission:

Year 1 Initial commission = 150% of reinsurance premium
Years 2+ Renewal commission = 10% of reinsurance premium
If the company were to cede 50% of business on this basis, the net cashflow would become:
Gross cashflow Monies to Change in Net cashflow

reinsurer reserves
Year 1 – $1,050 – $250 – $425 – $375
Years 2+ $100 $450 – $425 $75
where for instance in year 1 the company cedes $500 of premium, and so ‘passes’
$500 – 150% × $500 = – $250 to the reinsurer; and the company now has to establish reserves
equal to only 85% × $500 = $425.
It should be clear from this example that, with this arrangement, the new business strain is greatly
reduced, although at the expense of future profits.
Reducing new business strain means that more new business can be written for the same amount
of capital.
Quota share reinsurance is ideal for this purpose, as it gives the insurer greatest control over the
amount of financing it will receive in relation to the volume of business (new premium income) it
sells.
Question
Explain why individual surplus would be less useful here.

Solution
New business strain is a function of the total initial expenses, initial commissions, premium
income, initial supervisory reserves and initial solvency capital requirement relating to the new
business sold, whether this is made up of many individual small policies or fewer individual large
policies. If the company used individual surplus reinsurance for financing, then the amount of
capital support it would obtain would be a function of the mix of new business by size (which is
actually irrelevant to the total amount of strain).
The direct writer’s decision on how much business to cede will result from considering the
company’s future solvency position as projected using a model office.
Risk premium reinsurance can also be used to reduce new business strain, again using a quota
share basis. The strain is reduced by negotiating high initial reinsurance commission, in return for
higher risk premium rates.
1.4 Technical assistance

Reinsurance companies may have a considerable degree of expertise on such matters as
underwriting, product design, pricing and systems design.
This is particularly important when a cedant intends to launch a new line of business. It will
find that it can receive technical assistance from its reinsurer until it has built up its own
expertise.
A reinsurer can give support for existing lines in areas such as underwriting.
Even for more established companies, assistance from reinsurers on underwriting is sometimes
useful. This is especially the case in unusual underwriting situations where the reinsurer may
have much wider experience (eg of an unusual type of extra risk) than even a substantial direct
writer. This often results even in major companies reinsuring ‘unusual risks’, not because such
risks would give rise to greater mortality fluctuations, but because the direct writer is relying on a
reinsurer to assess the basic level of the risk (and is thus willing to share the hopefully profitable
contract with the reinsurer). The same arguments apply when a company develops a new
contract – all the policies will in effect be ‘unusual’ for a while.
In effect, by getting expert help on underwriting and pricing, the company is reducing the extent
of its parameter risk (the mortality assumptions used in the pricing basis are likely to be closer to
the actual average future experience of the company). The reinsurance that follows will further
dilute the remaining parameter risk by sharing exposure with the reinsurer.
Under this scenario, original terms reinsurance (with a high quota share reinsured) would
be the most likely reinsurance to be used.
This is as we discussed in Section 1.1 above.
The high quota shares will generally be reduced over time, as the insurer’s experience and
expertise develops.

1.5 Cost reduction

Due to various reasons (including different capital requirements, diversification benefits,
different taxation and different assessment of risks) a reinsurer may be able to price the risk
at a lower cost than the cedant.
For example, a reinsurer may have a lower supervisory capital requirement for writing certain
kinds of business than a direct writer. This means that the reinsurer can write the business more
cheaply than the insurer can, all else being equal, so it may be able to pass on some of those
profits by offering more favourable terms for its reinsurance. In this way, it is possible for both
insurer and reinsurer to make a profit from the reinsurance arrangement.
This is often referred to as ‘regulatory arbitrage’. Similarly, tax arbitrage can occur, where the
reinsurer is taxed more favourably for a certain line of business than the direct writer.
Question
Explain how ‘diversification benefits’ can also lead to cheaper reinsurance terms.
Solution
The reinsurer is likely to have a much greater spread of risks than any individual cedant, in terms
of the numbers of individuals covered, geographical area, original sales channels, and so on. This
will reduce the reinsurer’s random fluctuations risk relative to any individual insurer, and the
quantity of data should mean that parameter risk is less of an issue for the reinsurer as well. This
will mean that the reinsurer will not need to build in such large risk margins into its reinsurance
premium rates, which again might make it more profitable for a direct writer to reinsure this
business than to retain it.

2 Considerations before reinsuring

In this section we describe some of the considerations that an insurer will take into account
before deciding:
 whether or not to use reinsurance
 which type(s) of reinsurance to use
 how much reinsurance to use.
2.1 Cost of reinsurance

The reinsurer intends to make a profit as well as meet its cost of capital and expenses.
These costs will reduce the expected absolute level of profit for the cedant.
So an important reason why insurance companies do not reinsure all their risks is that it costs
money. The direct writer must balance the increased cost of reinsurance against the increased risk
of loss from adverse experience if less reinsurance were to be used.
In essence, we have the recurrent problem of balancing risk and return. The more a company
reinsures, the lower the risk of adverse claim fluctuations but the lower (usually) the expected
return. (However, remember that on some occasions we can actually increase our expected
return using reinsurance – see Section 1.5 above.)
Question
Explain why life insurance companies do not just use a mortality fluctuations reserve to reduce
their claims volatility, given that reinsurance costs money.
Solution
The problem is that setting up a mortality fluctuations reserve requires capital. The company
might not have free assets to spare, given other requirements. However, the amount concerned
is not likely to be large relative to other items in a company’s balance sheet. If the company does
have the capital, allocating that capital to this use costs money – because that capital will then
only earn the rate of return on the underlying investments, but the shareholders want their
capital to earn the risk discount rate. We can therefore quantify the cost of this capital as being
the difference between what it earns and what the shareholders require – this difference will
typically be around 3-4% pa (assuming that investment income on the reserve is not taxed –
otherwise the difference is likely to be far greater).
Quite apart from the non-zero cost of this approach, it also provides only limited protection
against adverse claims experience. Suitable reinsurance arrangements may be able to provide
protection against much worse claims volatility for a smaller cost.

However, the risk reduction caused by reinsurance may leverage up the return or
risk-adjusted return on capital.
Reducing the risk inherent in a stream of future profit flows will increase their value to the
shareholders by:
 requiring the insurer to hold less capital, so that the return on capital (for a given amount
of profit) will increase (this is what the Core Reading means by ‘leveraging up the return’)
 implying a lower risk discount rate for the future profit stream, making them more
valuable to the shareholders (so the ‘risk-adjusted return on capital’ has increased).
2.2 Retention limits

The retention limit is the maximum amount of risk retained by the cedant on any individual
risk.
General factors to consider when setting the retention limit include the:
 average benefit level for the product and the expected distribution of the benefit
 company’s insurance risk appetite
 level of the company’s free assets and the importance attached to stability of its free
asset ratio
 terms on which reinsurance can be obtained and the dependence of such terms on
the retention limit
 level of familiarity of the company with underwriting the type of business involved
 effect on the company’s regulatory capital requirements of increasing or reducing

the retention limit
 existence of a profit-sharing arrangement in the reinsurance treaty
 company’s retention on its other products
 nature of any future increases in sums assured.
The factors listed above will give rise to very different numerical answers for different companies
and different product lines. In answering the following questions, you will gain a feel for what
retention limits are used in practice.
Question
Suggest the type of company that might have the following retention limits for individual term
assurance business:
(i) £500,000
(ii) £75,000
(iii) ‘there are no reinsurances’.

Solution
(i) A medium to large company.
(ii) A small company.
(iii) Only a very large company would consider no reinsurance at all.
Question
A UK life company has a £500,000 retention for individual life business and a £50,000 retention
for group life business.
Suggest possible reasons why it chose such a large difference between these two retentions.
Solution
There are many possible reasons including:

 It may have less experience in group life.
 A lower retention on group business may be a good way of avoiding catastrophe risk.
 The group portfolio may be smaller.
There are a number of different possible approaches that a cedant can use to determine the
level at which the cedant should set its retention limit.
(1) Stochastic simulation – reinsurance only

Set the retention limit at such a level as to keep the probability of insolvency (or ruin
probability) below a specified level. An alternative approach is to aim for a probability that
the loss in any one quarter or year does not exceed a proportion of the earnings of the
business, if that is how the cedant determines its risk appetite.
This can be done by using a stochastic model for projecting claim rates and a model of the
business, so that claims can be projected forward together with the value of the company’s
assets and liabilities. By using simulation, a retention level can then be determined such
that the company stays solvent, or earnings stay above a certain level, for 995, say, out of
1,000 runs.
Question
(i) State the most important variable that needs to be modelled stochastically in this process.
(ii) Describe the most important unknown parameter value or values to estimate correctly for
this purpose, ie the model assumption(s) to which the derived retention level will be most
sensitive.
(iii) Outline the aspect of the model points to which the derived retention level will be most
sensitive.

Solution
(i) This method will only be useful if we model the numbers of deaths stochastically.
(ii) The important unknown here is therefore the variance of the number of deaths at each
age / gender / duration etc grouping. This will depend on the mean mortality rate and the
random fluctuations variance for each group. The random fluctuations variance will be a
function of the mean mortality rates (the Binomial model) provided the assumptions hold
about lives being identical and independent. However, we might also wish to assume a
larger random fluctuations variance to the extent that we believe that lives are not
homogeneous or independent.
The expected (mean) mortality is also unknown (ie subject to parameter error). Hence it
is important also to model the mean mortality rate stochastically, and both the mean and
variance of its distribution are key parameters in the determination of the retention level.
The greater the total variability of the mortality experience, the greater the probability of
the total mortality cost exceeding any given limit – and the greater the probability of
insolvency. Clearly, the greater the overall variance assumed, the more reinsurance will
be indicated by the model for a given probability of ruin.
(iii) This question part is asking which aspect of the model points would need to represent the
existing and future new business most realistically in order to obtain valid results.
The key element of the model points for determining retention levels is the distribution of
the sums assured. In other words, we need to have a model which realistically mimics the
proportion of new (and existing) business with sums assured in the different size ranges.
It should follow from considering the above questions that a company will be moved to reinsure
more:
 the less certain the company is about future mortality experience
 the lower is the acceptable probability of future insolvency
 the greater the variance of the sum assured distribution.
This is what we would intuitively expect.
Phrased in terms of retention limits, more reinsurance means lower retention limits.
As we saw with investment variability, here the size of the company’s free assets will have a major
bearing on the reinsurance requirements as far as claims volatility is concerned. In fact, many
large companies have sufficient free assets to be able to dispense with risk premium reinsurance
for normal policies, and will reinsure only exceptionally large policies (eg £1 million sum at risk)
and against catastrophes.
(The size of the free assets will also obviously affect the company’s financing reinsurance
requirements – the more capital a company has, the less it needs capital assistance from
reinsurers.)

(2) Stochastic simulation – reinsurance with fluctuations reserve

Consider the total of:
(a) the cost of financing an appropriate mortality fluctuation reserve; and
(b) the cost of obtaining reinsurance – the reinsurer naturally incorporates an expense
and profit loading in its reinsurance terms, and the ceding company incurs
administrative expenses.
As the retention limit increases, (a) will increase and (b) will decrease. A retention limit can
then be adopted that minimises the total (a) + (b).
To calculate (a), the simulation approach discussed above would probably need to be used
to determine the reserve that the company needs to hold.
Summarising the procedure for determining suitable retention limits and then the optimal
balance between a mortality fluctuation reserve and reinsurance, one approach is:
1. Decide on some criterion for claim volatility beyond which the company cannot go. For
example, it might want to have only a 1% chance that the net loss from claims is at least
$25m.
2. For differing retention limits, having sounded out reinsurers on terms available for the
business, model the function ‘{total claims net of reinsurance} less {total risk premiums
net of reinsured risk premiums}’. This modelling will be done stochastically, varying the
mortality experience.
3. The function is therefore:

X  [C (g)  C (r )]  [P(g)  P(r )]
where C(g) = gross claims

C(r) = claims recovered from reinsurer
P(g) = risk premium available (from policy) to insurer
P(r) = risk premium paid to reinsurer.
The criterion would therefore be that Pr( X  25m)  0.01 .
4. Look at the results of this modelling to choose the retention limit that will satisfy the
criterion.
5. The company could use this as a retention limit (if the cover is available in the market at
an acceptable price). However, it could also check whether this protection can be done
more cheaply using a mortality fluctuation reserve.

6. To do this it might assume that some of the cost of the risk premium reinsurance is
instead going to be spent on financing a mortality fluctuation reserve. The cost of holding
a reserve of size M is equal to:
M( j  i)
where j is the expected rate of return from the company’s capital, and i is the expected
rate of return from the assets that will back the reserve. This follows from the fact that,
by tying up some of its capital in a mortality fluctuation reserve, the company is unable to
use that capital to finance other ventures, from which it would have expected to earn a
return of j. Instead, it will earn an expected return of i, hence the difference is the cost
(ie loss of expected return) to the company over one year.
7. If the company decides to redirect, say, 60% of the reinsurance risk premium to the
mortality fluctuations reserve, then (if j  9% pa and i  7% pa, say) the amount of
mortality reserve purchased (for one year) is:
0.6  P(r )
M  30 P(r )
(0.09  0.07)
The company now has only 0.4 P(r ) left with which to buy reinsurance, so it will have to
have a somewhat higher retention level in order for the cost to be only 0.4 P(r ) .
Hence the company has exchanged reinsurance for a mortality fluctuation reserve, at
parity of cost.
8. Now model the distribution of X under this new arrangement, noting that there is now a
higher retention level (so the claims recoveries from the reinsurer will be lower) and X is
now calculated as:
X  [C (g)  C (r )]  [P(g)  P(r )]  M
 [C (g)  C (r )]  P(g)  29P(r )
where C (r ) is the claims recovered from the reinsurer at the new level of retention.
9. Compare the protection offered under this new construction against that offered by the
previous arrangement, ie recalculate Pr( X  25m) . If this probability is less than the 1%
previously obtained, then using a mortality fluctuation reserve (to the extent assumed) is
cheaper than using reinsurance, and would therefore be the preferred strategy.
10. Try this for other levels of reinsurance / mortality fluctuation reserve. Then decide on
which combination offers the most protection for a given cost. If the degree of protection
is increased by the use of a mortality fluctuation reserve, then the actual amounts of
mortality fluctuation reserve and retention level that are necessary to meet the desired
ruin criterion, with lowest cost to the insurer, can be determined (probably by trial and
error).

(3) Financial economics approach

Another possible approach is based on the theory of efficient investment frontiers, and
looks at reinsurance as an asset class that allows the firm to optimise its risk and reward
trade-off.
This approach allows the insurer to identify those ‘asset’ portfolios (including reinsurance), which
cannot be bettered in terms of either reducing risk for no reduction in return, or increasing return
for no increase in risk.
Students should be aware of this approach, which is otherwise beyond the scope of
Subject SP2.
2.3 Counterparty risk

The cedant retains liability to the policyholder for the benefits even if the reinsurer becomes
insolvent and cannot meet claim payments as they become due. This is an example of what
is known as a counterparty risk (or credit risk).
We have already come across this in Chapter 11 where the insurer’s risks were considered.
The greater the counterparty risk the less valuable the reinsurance will be to the insurer, and the
less will be the inclination to take out the reinsurance, all else being equal.
Question
Suggest how an insurer might reduce its reinsurance counterparty risks.
Solution
By reinsuring with many different reinsurance companies, and so diversifying the risk.
Deposits back
To manage counterparty risk in certain countries, the supervisory authority may require the
reinsurer to collateralise or ‘deposit back’ its share of the total reserve under a reinsured
contract with the cedant.
This would only generally apply to original terms and net level premium arrangements, as level
reinsurance premiums are paid in both these cases.
Question
Explain why deposits back generally only apply to original terms and net level premium
arrangements.

Solution
Under these arrangements, the reinsurer would have to set up reserves out of its earlier
premiums to cover future shortfalls from its later premiums, as with any other long-term
level-premium insurance contract. Hence, for these arrangements, a reserve exists that may need
to be deposited back.
Under normal risk premium arrangements, where the reinsurance premiums operate as recurring
single premiums, the insurer holds all of the policy reserve anyway and so there will be no
reserves to deposit back.
Where reserves are deposited back, the reinsurer simply has to make up (the reinsured share of)
the shortfall between the reserve and the sum assured on a death claim.
Question
Explain what the reinsurer would have to pay on a maturity claim of an endowment assurance
contract that is reinsured in this way.
Solution
It should have to pay nothing. This is because, by the maturity date, the policy reserve should be
equal to the benefit payable, so there will be no shortfall to make up.
Even where it is not a supervisory requirement, deposits back are sometimes done so that
the cedant gets the benefit of reinsurance, whilst at the same time being able to maintain a
reserve for the whole contract and hence maximise the funds it has to invest.
This therefore makes the arrangement generally more profitable to the direct writing company, as
it will retain all of its potential investment profit.
The arrangement can also be an advantage to the reinsurer. For example, on with-profits
business, an original terms arrangement would normally leave the reinsurer with a significant
investment risk, because it would have to match the insurer’s bonus rates on maturity claims. By
depositing back reserves it will avoid this investment risk.
Question
State one other class of business, other than with-profits, for which depositing back under original
terms reinsurance works to the benefit of the reinsurer in a similar way.

Solution
Unit-linked contracts, where the insurer will be required to hold the whole of the unit funds at all
times.
Also, in the case of a large international reinsurer, depositing back may avoid any problems with
having to invest in an unfamiliar market.
2.4 Legal risk

Reinsurance is usually governed by a treaty between the ceding company and the reinsurer.
This was mentioned in the previous chapter.
There are usually many clauses to be negotiated to cover numerous contingencies and
risks, each one potentially impacting the price of the reinsurance. Due to the operational
risks involved and the impact of disputes where contracts have not been finalised,
regulators around the world are increasingly focused on ensuring that reinsurance treaties
are complete and signed.
It is also in the direct interests of the insurer to make sure all its reinsurance treaties are legally
watertight, in order to avoid these problems, including the risk that reinsurance recoveries would
not be recoverable from the reinsurer to the extent expected.
2.5 Type of reinsurance

The type of reinsurance (eg original terms, risk premium, excess of loss, financial
reinsurance), and the way in which the amount reinsured is specified (eg individual surplus
or quota share), will depend on:
(1) the reason the ceding company is using reinsurance
(2) the reinsurance costs
For example, in recent years the unrestricted form of catastrophe cover has become
extremely expensive in many markets, and the terms have become increasingly
restrictive. This has occurred because reinsurers have become better at assessing
the true extent of tail risk by modelling more extreme events, such as pandemic ’flu
or terrorist actions.
(3) the type of business
For term assurances, original terms reinsurance or risk premium reinsurance may
be used.
For original terms, a level reinsurance premium set at the outset will be charged. Under
the risk premium method, the reinsurance premium rates may be level over the term of
the policy or may vary annually to reflect the changing probability of claim with age (using
guaranteed or annually reviewable mortality rates).

It may prove difficult to obtain original terms reinsurance on with-profits business as

the reinsurer would be obliged to follow the cedant’s bonus rates, though reinsuring
the guaranteed elements or options attaching to with-profits business is becoming
increasingly common in some markets.
Reinsurance is not normally sought on individual immediate annuity business,

unless it is a very large case. The possible impact on mortality profits of
unexpectedly light mortality of annuitants should however be considered and
reinsurance arrangements to cover this risk are available. Reinsurers can also
provide advice on such things as impaired life annuities and smoker / non-smoker
rates.
Catastrophe reinsurance can be very important for group business as there is an

accumulation of risk.
For unit-linked business, risk premium reinsurance is normally used.
(4) the legal conditions applying
(5) the forms of reinsurance coverage on offer in the market.

3 Examples
There follows four illustrations of the reasoning behind the choice of reinsurance method in
some specific cases.
3.1 A new unit-linked company with limited capital resources

The need here is likely to be for a low retention – to reduce any mortality risk whilst the
portfolio is small – and a method of reinsurance which helps to reduce any new business
strain, thereby enabling the company to grow faster and reach critical mass without the
financial limitations imposed by employing just its own capital resources.
The company will probably wish to build up its retained premium income as quickly as
possible and so will want to use risk premium reinsurance. In any case a unit-linked
company would not normally use original terms reinsurance as the reinsurer would need to
match the unit liability under the reinsured part of the contracts, and if the link is to the
ceding company’s internal fund it would not be able to do this directly. There are ways
round this problem using deposits back, but in general linked companies will avoid original
terms reinsurance.
Also, even more so than for with-profits business, there is little point in reinsuring unit-linked
business on original terms. Unreinsured, the assets and liabilities are matched. Why reinsure
them? (In a sense, unit-linked business is already ‘reinsured’. The investment risk has been
passed from the life insurance company to the policyholder.)
The risk premium reinsurance could be combined with a financing commission

arrangement to help reduce the new business strain, if such an arrangement effectively
reduces capital requirements under local regulations.
Risk premium reinsurance is a particularly natural choice for unit-linked business because it can
be administered very like the mortality charging arrangement in place for policyholders, and
should be part of the same overall administration system.
Question
‘A small unit-linked company such as this would be expected to have a highly capital-efficient
design for its unit-linked products, such as a nil initial allocation period. So it should, in fact, have
very little need for a financing arrangement.’

Solution
It is true that the company will probably have a capital-efficient design, but this does not mean
that financing is not valuable. Even with a nil initial allocation period, a monthly regular premium
contract will produce only one month’s premium at outset. Something that helps to deal with
even a short-term cashflow problem could be valuable to such a company.
Furthermore, the company may not want to make the design as capital efficient as is theoretically
possible, for marketing reasons.
The reinsurance could be on a quota share basis or a relatively low retention individual
surplus basis.
The company may also consider obtaining catastrophe reinsurance, although the general
market capacity has reduced, and price increased since the attacks on the World Trade
Center in New York on 11 September, 2001.
3.2 A large, established and financially strong mutual life insurance company
mainly transacting with-profits business
As the company has healthy free assets and writes mainly with-profits business, it will be
able to absorb a fairly high level of mortality risk and will probably be able to finance new
business from its own resources. It may therefore not purchase any reinsurance.
Question
Suggest two ways in which this company is able to absorb mortality risk, if it does not purchase
any reinsurance.
Solution
The company could absorb mortality risk in one of the following ways:
 it is financially strong, so it will have free assets that could pay the death benefits
 it sells with-profits business, so it could reduce bonuses to reflect the mortality losses.
If it offers guaranteed death benefits in excess of asset share and does not wish to reduce
bonuses in the event of a significant increase in mortality experience, it may arrange risk
premium reinsurance with no financing commission and with a relatively high retention limit
(since it will not want to give profit away to the reinsurer).
A mutual will want to maximise the profits that it can pass on to its with-profits policyholders.
However, any reinsurance that it buys will result in some of the profits being passed to the
reinsurer. So it may choose to buy risk premium reinsurance to protect bonuses from mortality
risk, but its financial strength means that it has no need for financing commission.
The company may seek some catastrophe cover.

3.3 A large proprietary company which is expanding rapidly and whose free
assets are declining
The company will be looking for a form of reinsurance which helps reduce its financing
requirement. If effective under the relevant regulatory regime, it could opt for a financing
reinsurance agreement based on its large in-force portfolio. Or it may be able to use risk
premium reinsurance with financing commission, with the retention set such that the
company obtains the required reduction in new business strain, but without giving away too
much profit to the reinsurer.
In addition, it would probably seek catastrophe cover in view of its reducing free assets.
Of course, reinsurance is not the only possible answer to the company’s declining free assets
during this period of rapid expansion. The company will have other options for raising capital and
will have to compare their pros and cons with the use of reinsurance.
Proprietary companies will face scrutiny from the stock market on measures such as return
on capital employed. Reinsurance can be used to reduce the capital employed, creating
leverage, and so increasing the firm’s own return on capital. This can be achieved by stop
loss reinsurance or (if effective) some form of financial reinsurance.
Reinsurance can also help to smooth profit emergence.
3.4 A company writing large amounts of group life insurance business

The major risk here is adverse fluctuations in mortality experience and there will also be a
catastrophe risk.
If the company is large, it will probably adopt an individual surplus arrangement, with a high
retention, on a risk premium basis. Otherwise, it might prefer to reinsure on a quota share
basis.
Experience refund agreements may be used.

Chapter 25 Summary
Reasons for reinsuring
 limiting claim payouts (single or total)
 reducing parameter and claim payout fluctuations risks
 financing new business strain – use financial reinsurance and/or quota share
 obtaining technical assistance
 benefiting from regulatory or tax arbitrage opportunities
 reducing costs
 separating out different risks from a product
 allowing aggregation of risks that could not otherwise be written.
Considerations before reinsuring

 cost of the reinsurance
 counterparty risks – increase number of reinsurers and/or use deposit back
 legal risks
 type of reinsurance
 amount of reinsurance (choosing retention limits).
Reinsurance costs money. The life insurance company should examine other ways of
reducing claims fluctuations (such as by using a mortality fluctuation reserve) and avoiding
capital problems on new business, and compare their costs to arrive at the most efficient
solution.
The company will seek to adopt an optimal strategy on the basis of achieving the desired risk
level with minimum cost (which is equivalent to achieving maximum returns for the required
level of risk). A stochastic model office, including assumptions that allow the mean mortality
rate as well as random mortality fluctuations to be modelled stochastically, would be used.
The results would also be sensitive to the assumed distribution of sums assured in the model
points.



25.1 Describe the principal ways in which a reinsurer may assist a new unit-linked life company,
Exam style
indicating the advantages and disadvantages of each. [9]
25.2 A proprietary life insurance company transacts large amounts of term assurance business in a
Exam style
highly competitive market.
Outline the main reasons why the company might wish to reinsure part of the term assurance
business. [7]
25.3 A small mutual life company, which has hitherto transacted mainly regular premium savings
Exam style
business, proposes to introduce a ten-year renewable convertible term assurance under which
both premiums and sums assured will increase by 10% per annum simple. The conversion option
allows conversion to a whole of life or endowment assurance without evidence of health and the
maximum age at conversion is 60. The product will be marketed by placing advertisements in
national daily newspapers.
State the factors that should be taken into account when recommending a reinsurance retention
limit for this product. [6]
25.4 A small proprietary life insurance company issues only individual policies, all on a with-profits
Exam style basis, and distributes profits using the revalorisation method. Only the savings profit is
distributed to policyholders.
(i) Explain why reinsurance is particularly important for this company. [5]
(ii) Describe, with reasons, the types of risk premium reinsurance that the company might
have in place. [6]
(iii) Describe any other type of reinsurance that the company might need in addition to this,
again giving reasons. [4]
[Total 15]
25.5 A small mutual life insurance company writes both individual and group life insurance business.
Exam style
The company is currently reviewing its reinsurance arrangements.
(i) Describe the reasons why the company may be expected to require reinsurance and the
type(s) of reinsurance that may be appropriate in each case. [10]
(ii) Describe a theoretical approach that could be used to determine suitable retention limits
and outline other factors that the company should take into account for this purpose. [6]
[Total 16]



25.1 Protecting the company against mortality risk
Risk premium reinsurance, covering the excess sum at risk of any guaranteed death benefits over
unit fund values, would be used to reduce mortality risk. [1]
Using a surplus basis will limit the amount of claim payments to a certain maximum retention.
This means that the variance of claim outgo is reduced and so there is less chance of large claims
making the company insolvent. [1]
It will also give the company more stable and predictable profits. [½]
This is particularly important for a new company, which will initially have a small portfolio. [½]
Being a unit-linked company it will not have large free assets to absorb fluctuations. [½]
Some quota share reinsurance may also be used along with the surplus method, to share some of
its parameter risk with the reinsurer. [½]
This risk will arise because the insurer will have no past data of its own for pricing its new
products. [½]
The need for high quota shares may be reduced to the extent that the charges under the new
products will be reviewable (as they almost certainly will be). [½]
The company may also seek catastrophe cover, although if it writes no group business, or has a
low individual retention, this may not be needed. [½]
Reducing new business strain
Reinsurance will reduce new business strain through the payment of financing commission to the
direct writer. This could be a key advantage because capital may be limited and the company will
wish to expand rapidly to maximise profits. [1]
Reinsurance would also reduce any supervisory capital requirement, although the company will
probably seek to design products that will minimise this anyway, if possible. [1]
Technical assistance and advice
The reinsurer may give advice to a new company in a wide range of areas, eg: [½]
 underwriting standards and claims control
 product design
 levels of risk benefit charges
 design of administration systems. [1]
For a new company with little experience this could be vital. [½]

However, the disadvantage is that the reinsurer will expect ‘payment’ in the form of reinsured
business. It also means some loss of control over the terms the company can offer. [1]
The main disadvantage of using reinsurance for these purposes will be the sacrifice of expected
profit to the reinsurer. Also, there is an increased credit risk – ie from the reinsurer becoming
insolvent. [1]
[Maximum 9]
25.2 Reasons for reinsuring term assurance
It can protect the ‘real’ solvency of the company (or risks to it). This may be especially welcome
given any uncertainty about mortality. [1]
This will then:

 allow capacity (ie enable company to write more business), and
 allow acceptance of very large risks and a wider range of impaired lives. [1]
It can flatten out mortality fluctuations. Even for an established company, it protects against
unwanted ‘blips’ in the emergence of surplus. [½]
Reinsurance can reduce new business strain through financing arrangements. [1]
It can reduce solvency capital requirements (unlikely to be critical unless the company is recently
established). [1]
The reinsurer can provide advice on underwriting and impaired lives. The company writes a large
amount of this business, and so will be experienced in underwriting. However, the reinsurer may
still be able to provide expert advice on impairments where there is relatively scarce data. It
could also provide temporary support (eg if a senior underwriter left, was ill etc). Also, the terms
on which the reinsurer will accept business are a good check on the premium rates. [2½]
If the reinsurer is subject to less onerous solvency requirements than the direct writer, then this
might result in the direct writer being able to offer lower premium rates to policyholders. [½]
The reinsurer might accept the business at a loss (but probably not for long, and that isn’t the way
to develop good long-term relationships with a reinsurer). [½]
It is important to tailor your answer. It would not look good if you merely commented that
reinsurers could give advice, without qualifying your comment by stating that a company with lots
of experience will only need advice for non-standard situations.
[Maximum 7]

25.3 The factors are:

 the mix of business by size of initial sum assured, bearing in mind that this will be double
by the end of the term, and potentially much bigger for contracts that are renewed [1]
 the level of underwriting [½]
 the types of reinsurance contract available [½]
 the cost of reinsurance and how this varies with different retention limits [½]
 the company’s experience in this market [½]
 the free asset ratio [½]
 the cost of capital (ie the returns required by the with-profits policyholders for the use of
their capital) [½]
 modelling investigations, which may be used to determine the optimal balance between
retention level and amount of claim fluctuation reserve used, if any [1]
 the size of the company and how well it can cope with fluctuations in claim experience [½]
 the level of new business strain for the contract [½]
 the expected volumes of business [½]
 whether policyholders can convert at any time and whether they can increase the sum
assured when they do convert [½]
 the expected age of the applicants, which affects the variance of claims outgo. [½]
[Maximum 6]
25.4 (i) Why reinsurance is important
The company needs reinsurance to protect itself from adverse claims experience. [½]
All insurance profit falls to the shareholders; the company cannot reduce bonuses to reflect poor
mortality experience. [1]
This company is likely to require reinsurance for all purposes due to its small size, which may
indicate that there are only limited free assets available as an alternative cushion against
reinsurable risks. [1]
Reinsurance will help to reduce the ultimate risk of ruin. [½]
It will also allow the company to write more business, … [½]
… so increasing potential profits. [½]
Some types of reinsurance enable the company to write policies with large sums assured, so that
marketability will not be compromised. [½]
The company may well receive advice from its reinsurer about how to underwrite its policies,
especially with regard to dealing with impaired lives (for which its own experience will be very
limited). [1]

As the company is small, its experience data will be limited, so the company may well need the
reinsurer’s advice on pricing and reserving issues. [½]
The company may need to enter into a reinsurance treaty in order to obtain this advice. [½]
Shareholders prefer smooth dividends; reinsurance helps to smooth profit emergence. [½]
[Maximum 5]
(ii) Risk premium reinsurance
A proportion of the sum at risk at each policy duration will be reinsured. [½]
The reinsurer will charge a premium each year, … [½]
… which broadly reflects the expected cost of covering the reinsured sum at risk for the year plus
appropriate loadings for expenses and profit. [1]
The risk premium rates may or may not be guaranteed for the whole policy term, once a policy
has been issued. [½]
The company may use either or both of the following structures:
Individual surplus method [½]
The amount reinsured is the residual sum at risk (if any) after deducting the company’s fixed
retention. [½]
The company is likely to use individual surplus to reduce its concentration of risk, … [½]
… which will prevent any one policy having a retained sum assured bigger than the chosen
retention level, thus reducing the risk from claim fluctuations. [½]
Individual surplus allows the company to write larger sums assured. [½]
Quota share method [½]
The same proportion of the sum at risk is reinsured under all policies, and at all durations. [½]
Quota share might be required in order to reduce the company’s total exposure to parameter risk
– ie the risk that the mortality assumption in the premium basis is too optimistic. [½]
This is an issue despite the fact that policies are with-profits, because none of the mortality risk is
shared with the policyholder. [½]
Quota share may also be used if the company is using the reinsurance for financing. [½]
For this purpose, the reinsurance would be arranged to have a high initial reinsurance commission
payment, … [½]
… in return for higher risk premium rates payable by the insurer. [½]
[Maximum 6]

(iii) Other types of reinsurance
Catastrophe reinsurance [½]
It is possible that the insurer considers itself at risk from multiple death claims arising from a
single event, such as some major physical disaster. [½]
This type of reinsurance can reduce the risk from such non-independent claims. [½]
The reinsurer will pay the excess (up to a maximum) of the total claims due to a single event over
a certain minimum threshold. [½]
Premiums and cover would probably be negotiated on an annual basis. [½]
The need for this type of cover is probably limited as the company has no group business. [½]
The company may take out financial reinsurance to increase its available capital. [½]
It can use a contingent loan, in which the reinsurer provides a cash loan in return for repayments
that are contingent on future valuation profits. [½]
This allows the insurer to take advance credit for its ‘value-in-force’ profit in its balance sheet,
which would otherwise be tied up in a prudent valuation basis. [½]
However it is not effective under accounting or supervisory regimes where credit can already be
taken for the future profits … [½]
… and/or where a realistic liability has to be held in respect of the loan repayments. [½]
[Maximum 4]
25.5 (i) Reinsurance arrangements
Reduce impact of adverse claims experience [½]
Claims fluctuations will lead to volatile profits, making it harder to smooth profit distributions. [½]
This can also lead to losses that could potentially threaten solvency, especially for a small
company like this. [1]
Fluctuations can arise from randomness in the mortality experience, made significantly worse if
individual lives are covered for large sums assured. [½]
This source of risk is best covered using individual surplus reinsurance, as only sums assured in
excess of an agreed retention limit would be reinsured. [1]
Original terms or risk premium approaches could be used. [½]
Adverse claim costs can also arise if the company has priced its contracts using mortality
assumptions that are too low. [½]

Any original terms or risk premium method can reduce the exposure to this risk, though some
quota share and/or reducing retention reinsurance would be expected here, as there would be
more control over the proportion of the risk shared. [1]
Protect against non-independent claims [½]
Most likely to occur for group contracts where several (non-independent) death claims can be
caused by one single event (eg accident in the workplace). [1]
Catastrophe reinsurance cover would be appropriate for this, specifically total death claims arising
due to any single event in excess of an agreed threshold level. [1]
Several lines of cover may be necessary for the group business, if the most extreme outcomes
need to be covered. [½]
Alternatively, stop loss reinsurance would cover adverse losses on the whole portfolio. [1]
Provide capital finance [½]
This could be important for this company, which may have limited capital, especially if it wished
to expand. [1]
Could be achieved by:

 quota share reinsurance, either risk premium or original terms, provided a large
reinsurance commission was payable [1]
 a contingent loan, where the reinsurer pays a loan in return for repayments made
contingent on future supervisory profits from the business (provided this is effective
under the regulatory regime). [1]
The required minimum level of solvency capital may be reduced by having reinsurance. [½]
Technical assistance [½]
A small company may lack experience data and expertise in a number of key areas. [½]
Assistance is typically with underwriting, pricing and product design. [½]
This can be associated with any type of reinsurance, … [½]
… but quota share commonly features, as this reduce the insurer’s parameter risk. [½]
Accepting large risks [½]
Without reinsurance there would be a limit to the size of individual risk the company could take
on. [½]
Use individual surplus, as before. [½]
Treaties enable large risks (up to a maximum) to be accepted automatically. [½]
Helps the company achieve its marketing aims. [½]

[Maximum 10]

(ii) Retention limits
Working out the retention limit
A simulation approach might be adopted. [½]
This would involve the following steps:

 Decide upon some criterion for claim volatility beyond which the company cannot go.
The aim may be to keep the probability of insolvency below a specified level. [1]
 Combine a stochastic model for expected claims rates together with a model of the
business, to project forward claims, assets and liabilities. [1]
 Use a simulation method to determine a retention level such that the company stays
solvent for say 995 times out of 1,000 runs. The model will need to be run lots of times at
differing retention levels. [1]
 Sensitivity analysis may also be carried out, varying the mortality distribution. [½]
Other factors
 The company could check to see if the protection might be provided more cheaply using a
mortality fluctuation reserve. [1]
 This would be done by minimising the total of the cost of financing this mortality
fluctuation reserve and the cost of obtaining reinsurance. [1]
 The company could alternatively use the theory of efficient frontiers to optimise its risk
and reward trade-off. [½]
 Any proposed arrangements would need to be practical and simple. [½]
 The cost of the reinsurance available would affect the retention level decided upon. [½]
 Take into account the expected volume and mix of business expected over the duration of
the proposed reinsurance treaty. [1]
 The result should be checked for reasonableness against the market norm. [½]
[Maximum 6]





using the course.

SP2-26: Underwriting Page 1
Underwriting
Syllabus objectives
3.3 Demonstrate the application of underwriting as a risk management technique.
3.3.1 Outline the purposes of underwriting.

3.3.2 Describe the different approaches by which underwriting is applied.
3.3.3 Discuss the factors that should be considered when determining the level of
underwriting to use.

Page 2 SP2-26: Underwriting
0 Introduction
Underwriting is the process of consideration of an insurance risk. This includes assessing
whether the risk is acceptable and, if so, setting the appropriate premium, together with the
terms and conditions of cover. It may also include assessing the risk in the context of the
other risks in the portfolio.
Underwriting helps a life insurance company to control the quality of lives accepted for insurance.
Question
‘There is no such thing as a bad risk, only an insufficiently priced risk.’
Discuss this statement.
Solution
The statement is true for most unhealthy lives – the life insurance company will want their
business, albeit subject to their paying an appropriate premium. However, it is not true for
seriously impaired lives. There are three reasons for this:
 It is much more difficult to rate them accurately.
 The risk of death may be so high that there is little hope of recouping the initial expenses
involved.
 The appropriate premium might be so high that there is no realistic hope of selling the
policy.
In this case, the appropriate decision would be to decline the application for insurance.
The insurance company will generally aim to accept as many policies as possible at appropriate
premium rates. Given this, we can see that the company is exposed to two fundamental risks:
 that the premium rates are not appropriate for the lives concerned, and
 that the premium rates permit selection against the company.
To ensure that premium rates are appropriate for the lives concerned, the company will break
down the lives into groups homogeneous with respect to mortality (or whatever contingency is
relevant) and price accordingly. This would be done at the product development (and subsequent
review) stage.

For instance, a typical grouping for term assurance would be:

 by age (graduating by year, or possibly down to quarter-year)
 by gender
 by smoking / non-smoking habit
 by broad occupational group
 by health.
As mentioned in Chapter 9, regulation may restrict the rating factors that can be used. For
example, in the EU it has not been permitted to use gender as a rating factor since
December 2012.
The first four factors listed above are all easily identifiable (barring occasional dishonesty in not
confessing to smoking). The last aspect, the health or medical status of any life, is more difficult
to assess. Most underwriting work is concerned with this aspect: to ensure that all lives are
grouped by broadly similar medical condition, and then to rate these lives accordingly.
As far as selection is concerned we can see that, the more we categorise lives into such groups,
the more we prevent selection against the life insurance company. For instance, by having a
smoker / non-smoker differential, we prevent smokers taking advantage of rates which would
otherwise (if uniform by smoking habit) be too low. However, selection against the company by
unhealthy policyholders is more difficult to counter, because the health of a potential
policyholder is difficult to assess. One of the aims of the underwriting process is to prevent such
selection against the company by identifying abnormal risks.
However, in attempting to counter the risk of selection, the company must ensure that the
control procedures themselves do not discourage business or cost more than they save.

1 Managing risk
The main purpose of underwriting is to manage risk. This can be done in the following
ways:
 It can protect a life insurance company from anti-selection. Being slightly different
from the market standards may lead to an accumulation of anti-selection.
So, an insurer that charges the same (average) premium to both non-smokers and
smokers is inviting smokers to take out policies so as to benefit from the cheap rates that
they are getting. But if all other insurers charged different rates for smokers and non-
smokers, then the anti-selection effect would be much worse.
Question
Explain why the anti-selection effect would be much worse in this case.
Solution
Now the insurer would be the only company offering aggregate rates for smokers, and because
those rates would be much cheaper than any other company’s smokers rates, nearly all smoker
applicants would be expected to apply for insurance with this insurer.
On the other hand, every other company will be charging much cheaper rates for non-smokers
than this insurer (because its rates are based on the average of smoker and non-smoker
mortality), so it will receive almost no applications from non-smokers. The result is that the mix
of lives that it insures will be hugely biased towards smokers, contrary to the assumed mix, and
on its assumed premium rates it would therefore expect to make heavy losses.
(This is an example of a mix of business risk, caused by the existence of a cross subsidy in the
product pricing: allowing non-smokers to subsidise the higher claim costs caused by smokers.)
In general, whenever an insurer is more relaxed in its process of risk classification than
other insurers in the market, then it will open itself up to increased anti-selection.
 The underwriting process will enable a life insurance company to identify lives with
a substandard health risk for whom special terms would need to be quoted.
 For the substandard risks, the underwriting process will identify the most suitable
approach and level for the special terms to be offered.
By approach, we mean a decision as to whether or not to reject the proposal, and how to
price for the risk, if accepted.

 Adequate risk classification within the underwriting process will help to ensure that
all risks are rated fairly.
This is another way of saying that adequate risk classification will prevent policyholders
from profiting from any heterogeneous groupings. By grouping lives appropriately
(eg distinguishing office workers from manual labourers) such selection is avoided.
Clearly no broad grouping can ever achieve this completely, but the groups must reflect
known mortality experience.
For example, we might be intuitively sure that actuaries have higher mortality than
accountants due to the dashing and rakish life style involved, but in the absence of any
quantifiable experience there is little point introducing such a grouping into our premium
structure. Having said that, some actuaries might argue that a ‘best estimate’ educated
guess is preferable to ignoring the distinction.
 Underwriting will help to ensure that actual mortality experience does not depart too
far from that assumed in the pricing of the contracts being sold. Underwriting is the
process that aligns the risks written by a company with the pricing performed by the
actuary. It is therefore critical for the pricing actuary to understand how the
business is being underwritten by the company’s own underwriters and how this
compares to competitors’ underwriting standards and pricing.
He or she must then ensure that those lives, whose mortality experience underlies the
company’s standard premium rates, have the same average health as enjoyed by the lives
being accepted by the company at those premium rates.
Alternatively, an insurer can achieve the same result by designing its underwriting
approach to correspond with its premium basis. Usually, in practice, insurers would
consider the two things hand-in-hand.
 Financial underwriting will help to reduce the risk from over-insurance.
We discuss this in Section 2.3 below.
Question
State how important underwriting is for each of the major life insurance product types.

Solution
Product Underwriting is …
Term assurance Critical
Whole life assurance Very important
Endowment assurance Important if regular premium,
not very important if single premium
Pure endowment Irrelevant
Immediate annuity Not important, unless impaired life annuities
In theory, for annuity business a company should want to underwrite against good lives just as it
would underwrite against bad lives for term assurance. For practical reasons it would not
normally do this.

2 The process
Given the various objectives of underwriting, as described above in terms of the risks that we
want to circumvent, what is the most practical method? We can break underwriting down into
four distinct areas:
 medical underwriting – the medical assessment of a potential policyholder’s health
 lifestyle underwriting – the influence of sporting and hazardous leisure pursuits on the
risk, and the extent to which the lifestyle might increase the possibility of contracting
dangerous diseases such as AIDS
 financial underwriting – assessing the financial situation of a potential policyholder
 claims underwriting – although the phrase might seem a contradiction, we can do some
underwriting at the claims stage.
Question
Describe the ‘underwriting’ that the company can do at the claims stage.
Solution
For death claims, it can check that the claim is true (not a fake death), and that the details
declared at policy proposal were correct. For instance, if a declared ‘no dangerous hobbies’
policyholder dies from a parachuting accident, the insurance company might decide not to pay
out. This might lead to bad publicity – so at the very least it would want to recoup the difference
in premiums involved.
Here we consider in detail medical and financial underwriting.
2.1 Medical evidence

Where a life insurance company has mortality risk under a contract, it will obtain evidence
about the health of the applicant to assess whether he or she attains the company’s
required standard of health and, if not, what their state of health is relative to that standard.
For contracts where there is only a longevity risk, evidence of health could also be
obtained, if the life insurance company intends to offer different terms according to the
health of the applicant. Such differentiation by health is increasingly common, particularly
in the UK and North America, as competition forces increased risk differentiation.
The aim here has as much to do with improving the marketability of the contract and increasing
fairness, as to controlling risk. For example, an applicant for an annuity who is in excellent health
(and hence a bad longevity risk) is hardly likely to be declined acceptance. On the other hand, a
person who is in poor health would not consider applying for an annuity unless some special rates
were to be offered. In this latter case, the insurer would grant a higher level of annuity payment
on the expectation that the payments would continue for a shorter time. An example would be
an annuity contract where better rates were to be offered to smokers than to non-smokers.

Question
Describe the risks of offering preferential annuity rates to smokers.
Solution
The risk here is that the policyholders deemed to be unhealthy improve in health – ie in this
instance, smokers giving up smoking.
There is also the risk of non-smokers or light smokers declaring themselves as heavy smokers
when they apply for an annuity.
There is also a moral dimension to this, because the company is rewarding people for indulging in
an injurious habit. The company would certainly need to consider whether any proposed
underwriting practice was ethical before proceeding. (This relates to the control cycle: here we
have an example of ethical (ie professional) considerations possibly affecting the decision.)
Medical evidence can be obtained from the following sources:

 questions on the proposal form completed by the applicant
 reports from medical doctors that the applicant has consulted
 a medical examination carried out on the applicant by a doctor, nurse, paramedic or
pharmacist
 specialist medical tests on the applicant.
Questions on the proposal form completed by the applicant

The potential policyholder will give details such as:
 height and weight
 smoking and drinking habits
 current health and details of any current treatment being received
 personal medical history (any major illnesses or operations)
 family medical history (check for hereditary ailments like heart disease)
 occupation and potentially dangerous pastimes.
The extent to which these answers are reliable depends on the culture of the country concerned
and also on the distribution method. For instance it is more likely for applicants to answer
honestly if filling out the proposal form with a broker present, than it is if they are filling out a
postal direct marketing form.
Question
Explain why some dishonesty in compiling the proposal might not be a problem.

Solution
Dishonesty is not a problem for the whole portfolio if the premiums allow for it. This will happen
automatically if premiums are based on the company’s mortality experience for its own
policyholders, and they showed the same degree of dishonesty in answering their original
proposal form questions.
So dishonesty will only be a problem if

 it is increasing, or
 the company’s rates are based on some other group of lives (eg from reinsurers’ data)
and underwriting procedures were stricter for these lives.
In addition, dishonesty will be less of a problem if it can be picked up at the claims stage.
Obviously, it is not a big problem for contracts where the sum at risk is low, for which there would
be very little underwriting and very little mortality risk.
Ideally, compiling the proposal form will be the only underwriting done. This is very cheap for the
company. The only problem is what to do if some of the answers suggest a non-standard risk, or
if the sum assured is so high that the company wants to be more certain about the risk involved.
So, normally the life insurance company would start off with the proposal, and if some of the
answers suggest any problem areas, or if the sum assured is very high, then some of the other
measures would be resorted to.
The limits on sum assured, above which some real medical evidence is required, are known as
‘medical limits’. They will reflect the balance which the company wants to make between the
cost of obtaining such evidence, and the cost of the risk of not having it.
Reports from medical doctors that the applicant has consulted

If the life insurance company has access to these then they represent a cost-effective way of
learning most of the applicant’s medical history. However, in countries where people do not have
a single doctor to whom they refer all queries (such as the General Practitioner doctor in the UK)
the idea is not so useful. In addition, there may be difficulties with privacy of records, data
protection and so on.
A medical examination carried out on the applicant

This would typically be a basic half-hour examination covering many of the points in the proposal
form (height and weight, discussion of past illnesses etc) but allowing a doctor to check for any
obvious problems. Blood and urine tests may feature at this stage.
It is a costly option. There is also the risk of discouraging healthy applicants if resorted to too
much. It is therefore normally only used in the event of unsatisfactory answers on the proposal
form, or a high sum insured, and if no reports from previously consulted doctors are available.

Specialist medical tests on the applicant

This is the most intensive and expensive option. Such tests might be indicated as necessary by
the basic medical examination described above. Some of these tests might also be automatically
required for very high sums insured on temporary life assurance contracts. The tests themselves
include things such as blood tests, electrocardiograms and chest X-rays.
Medical practitioners have to be paid for the services they provide, and each of the services
stated are progressively more expensive as we go down the list (doctor’s report – medical exam –
specialist tests).
Apart from the proposal form, each of the above will involve the life insurance company
incurring additional expense. Hence, the extent to which they are used in a particular case
depends on the extent of the loss that the life insurance company will make if it
mis-estimates the state of health of the applicant. Thus the level of medical evidence
required normally increases for higher levels of benefit.
The more information an insurer obtains about each applicant the greater the proportion of the
high-risk applicants that will be identified. This means that more (extra) premiums will be
received for the same policyholders, but at a higher underwriting cost. The company needs to
judge the level of underwriting required such that the (marginal) extra premiums it receives for
the (marginal) effort it makes are greater than the (marginal) costs involved.
Question
Explain for which types of products the life insurance company might impose no underwriting
requirements.
Solution
For any product with a zero sum at risk, underwriting is normally irrelevant, unless there is a
significant possibility of deaths leading to non-recovery of initial expenses. The latter can happen
at early durations of many types of regular premium policies if the death benefit (however
defined) is higher than the asset share at the time of death. This would be the case, for example,
with a deferred annuity with return of premiums with interest on death prior to vesting: at early
durations the death benefit could be considerably higher than the asset share. (This is the same
argument as for persistency risk, as described in the earlier parts of this course.)
As we have already touched on, the company also wants to avoid losing any business from using
an underwriting stance that is too exacting. For this reason, it would normally look at market
practice in setting the medical limits at which the various tests are triggered. The risk of selection
against the company is more relative than absolute, in the sense that the company is most at risk
of being selected against if its underwriting procedures or premium structure are weak with
respect to the market, rather than weak with respect to some medical / actuarial ideal.

2.2 Other evidence

Besides the medical state of health of the applicant, other factors that can affect the
mortality risk need to be investigated, namely any risks associated with the applicant’s:
 occupation
 leisure pursuits
 normal country of residence (and its health care environment).
Question
An applicant for term assurance lives in an undeveloped African country.
Suggest ways in which this will affect claim frequency compared with living in a developed
country.
Solution
The expected level of claim frequency may be significantly affected by:

 climate
 diseases
 availability and quality of medical facilities
 levels of violent crime.
Some life insurance companies also attempt to use socio-economic factors.
2.3 Financial underwriting

The life insurance company may gather information on the sums assured applied for by an
applicant to ensure that:
 The premiums payable by the person are affordable. This requires the life insurance
company also to gather information on the income of the person, which is used to
control the persistency risk. This level of underwriting will tend to be imposed only
on large levels of sums assured.
 The person concerned is not trying to commit fraud. An indication of this could be
having higher levels of sums assured than could be justified by the applicant’s
current circumstances. The simplest level of such underwriting will involve
aggregating the total sums assured across the proposed new policy and any
existing policies held by the insurer. A more complicated level of underwriting,
involving aggregating the total sums assured across all insurers, may be attempted
where there is a suitable database to which all insurers contribute.
A standard check in the case of term assurance would be the ratio of sum assured to the salary of
the applicant. The company needs to ensure that the applicant’s proposed policy is in keeping
with the normal financial needs and lifestyle of such a person. This is to reduce the risk of policies
being taken out with ultimately fraudulent intent, or some other motive such as suicide.

2.4 Claims underwriting

A life insurance company may also decide to ‘underwrite’ at the claims stage, for example to
ascertain whether the claim is subject to an exclusion clause, or if there is suspicion of non-
disclosure (although obtaining evidence to substantiate this can be difficult).
For example, a claim might be rejected if death occurred due to a dangerous activity such as
motorsport that had been specifically excluded in the policy wording.
The insurer might also reject a claim on death from a smoking-related disease if the policyholder
had claimed to be a non-smoker on the application form. However, the insurer would still have to
pay the claim if there was a small possibility of a non-smoker contracting the disease.
2.5 Interpretation of the evidence

The evidence obtained needs to be interpreted in terms of the standard of health required
by the life insurance company.
The proposal form will initially be reviewed by clerical staff. They will look through the answers,
converting the numerical fields (such as height and weight, or units of alcohol consumed) into
‘OK / not OK’ by reference to some simple tables prepared by the company’s specialist
underwriters.
Most proposals will be ‘OK’ on all fronts, and can then be processed immediately. Those
proposals for which queries have been raised will be passed on for further consideration.
This will be done by specialist underwriters employed by the life insurance company, who
will make use of:
 any doctors specifically employed by the life insurance company for this purpose
 underwriting manuals prepared internally or by the major reinsurance companies.
For basic medical evidence and disclosures on an application form, specialist underwriting
software is proving increasingly efficient at assessing information and making rating
decisions. For moderately complex and financial underwriting cases, experienced
professional underwriters are required.
2.6 Specification of terms

Applicants who fail the financial underwriting test would usually be declined.
Before declining, the insurer may ask the policyholder for additional information, eg an
explanation of the need for the large sum assured. However, if an attempt at fraud is suspected
then any further information will be treated with caution. The insurer may decide that any loss of
goodwill associated with declining a proposal is a small price to pay for reducing the risk of fraud
and avoiding the costs of further investigations.
Applicants whose state of health reaches the required standard can be offered the life
insurance company’s normal terms for the contract. Other applicants will be offered special
terms, unless their state of health is such that the company will not accept them on any
terms, in which case they will be declined, at least temporarily.

In some countries legislation may not permit the refusal of a life insurance policy on health
grounds. For instance, in New Zealand a life insurance company may not refuse cover to an
applicant, although the company may charge any actuarially justifiable premium.
The company will normally rely on reinsurers’ and industry data to assist in rating abnormal lives,
since the company’s own experience for any particular condition will normally be very small.
A temporary refusal of acceptance is in effect a deferral of the underwriting decision to a later

date. This could apply whenever the applicant’s level of risk is too uncertain at the present time
to assess it with adequate precision, but a future time may be foreseen when the risk level could
be clearer. An example would be a person who has recently undergone an operation, and whose
prognosis is currently unclear. In a few months’ time it may be possible to assess the person’s
mortality risk with much more confidence; hence a deferral of any decision to this later date
would be the correct current decision to take. During the deferral period, of course, no cover
would be provided by the company.
The main ways in which the special terms can be specified are as follows:
 An addition may be made to the premium which would have been charged to an
applicant who did meet the required standard, this addition being commensurate
with the degree of extra risk.
 A deduction may be made from the benefit which would have been paid to an
applicant who did meet the required standard, again being commensurate with the
degree of extra risk.
Question
Explain which of the above two approaches is likely to be the more suitable for the following
contracts:
(a) regular premium endowment assurance
(b) term assurance.
Solution
(a) Regular premium endowment assurance: the product is a savings vehicle, where
policyholders generally decide what premiums they can afford to pay, and then see what
eventual benefit they get at maturity. If this is so, then an approach which involves
reducing the death benefit (only) would probably be more suitable. It also means that
policyholders who do, in the end, survive to the end of the term (and hence have not
contributed in any way to any mortality loss), receive the same benefit for the same
premium as an unimpaired life. This seems very fair, and appears to meet most
policyholders’ needs from this contract. A variation would be to impose a reducing debt,
where the debt reduces by level amounts over the term of the policy, reaching zero by the
maturity date. On the other hand, if for any reason the policyholder needed to have
some specific amount of sum assured on death, (eg to repay a house mortgage), then an
increase in premiums would be more appropriate.

(b) Term assurance: the policyholder normally decides on some necessary level of cover, and
then pays the required premiums. So it will be more appropriate to increase premiums,
rather than to reduce the cover from what the policyholder originally wanted.
 An exclusion clause may be added to the contract, which excludes payment of

benefit claims that arise due to specified causes.
The exclusion clause will normally be the least preferred option, since if a policyholder requires
cover for some reason, such as providing an income for a widow(er) and family, a policy with an
exclusion clause would not be meeting fully the policyholder’s insurance need. In fact, the policy
is providing no cover for what is almost certainly going to be the most likely cause of death, and it
is very likely for adverse publicity to follow the enforcement of exclusion clauses in practice.
Another significant difficulty with exclusions is that they are very difficult to enforce at the time of
claim. Officially recorded causes of death can be vague or ambiguous, or even deliberately
misleading in order not to invalidate the insurance claim. A good example is the practice of
applying an AIDS exclusion clause under a life insurance policy. The physical causes of death of
people with AIDS are often diseases that non-AIDS sufferers can get – various kinds of cancer, for
example. If the death certificate only mentions the cancer, and not the reason for the cancer
(ie AIDS), then the claim would be payable. (In theory the company might be able to get round
the problem by carrying out its own post-mortem inspections, but this might not be legally
permissible.) Hence exclusion clauses are unlikely to be completely effective in practice, so that
an extra premium would still probably be necessary as well as any exclusion clause that is applied.
An alternative to applying special terms might be to offer the applicant a slightly different policy.
For example, it might offer a contract with a reduced term. However, this might not be
acceptable to the policyholder, as it might not meet the needs of the policyholder and/or the
premiums may not be affordable.

2.7 Underwriting flow chart

We can summarise the above process diagrammatically as:
Policy proposal form:

 medical questions
 lifestyle questions
 financial questions
Check by clerical staff:

 standard answers to questions?
 financial underwriting check?
 sum assured vs medical limits?
Further scrutiny by
Policy issued
underwriters
normally
Proposal
rejected
Policy issued but

perhaps at special Medical exam and/or
rates (perhaps also doctor’s report
reinsured)
On claim,
further checks Specialist
medical tests

The exact form of such a flow diagram will vary from country to country and company to
company, and will vary by product and sales method for an individual company.
Question
List steps that the underwriter might take after reading reports from medical doctors that the
applicant has consulted.
Solution
After examination of the report the underwriter will decide whether:

 to reject the proposal
 to accept at normal rates
 to accept with some special rating
 to ask for completion of a medical examination
 to ask for some specialist medical tests.
Question
Suggest arrows that might be missing from the above diagram, including the events they would
represent.
Solution
Most companies would probably have an arrow going from ‘check of proposal form by clerical
staff’ to ‘medical exam’ and also to ‘specialist medical tests’, corresponding to the action to be
taken when the sum insured exceeds the relevant medical limits and the extra testing is always
required.
There might also be an arrow from ‘check of proposal form by clerical staff’ to ‘proposal rejected’,
if the company always wants to exclude specific classes of risk (eg residents in AIDS-stricken
countries, or keen pot-holers).

3 Determining the level of underwriting to use

In setting the overall level of underwriting, the insurer would need to ensure that the
increased volumes that result from more relaxed underwriting outweigh the costs of
increased anti-selection and any increased costs of obtaining reinsurance. The main
benefits would be increased profits, reduced underwriting expenses, and the increased
attractiveness of the product to the distribution channel concerned.
Conversely, a company could usually improve its mortality experience by increasing the amount
of underwriting it does. For example, in the extreme, it could require medical examinations and
further specialist tests on every proposer.
Question
Explain why it will not do this.
Solution
There are two key costs in doing so:

 the actual expenses of the underwriting
 the discouragement to new business, if most potential applicants can obtain insurance
without all that trouble by going to a rival insurer.
The potential benefit of more underwriting may be limited. In the vast majority of cases the tests
will just confirm what the company was confident of from the proposal form or PMAR – that the
applicant is in good health.
The relative costs would be particularly high for small sums assured.
So the big message is: underwriting is justified to the extent it pays for itself in improved
experience. Companies will therefore review their underwriting arrangements regularly to check
that their underwriting procedures and medical evidence limits are still delivering this.
The following factors will need to be considered in an analysis of the appropriate level of
underwriting to use:
 The expenses associated with the level of underwriting proposed. Underwriting

involves a number of costs, such as salary of underwriter and medical reports,
which will increase the overall costs to the company.
The analysis should include the costs involved in obtaining further evidence at
lower sums assured based on information disclosed in the application form.

 The extent, and the financial significance, of any potential anti-selection risk. If
anti-selection is going to occur, it will present itself to the insurer with the least
stringent underwriting requirements within the peer group offering that product.
Therefore, offering the same product with less underwriting may mean that an
insurer attracts a disproportionate share of the anti-selection risks (the ‘Sentinel
effect’).
This is known as the Sentinel effect since the presence of underwriting guards against
anti-selection in a similar way to a sentinel being posted to keep guard.
Question
One of the main costs of underwriting is the fact that it discourages people who are healthy from
applying for insurance.
Suggest ways by which the risk of anti-selection could be reduced for a life insurance contract
offered with no underwriting.
Solution
Methods include:
 low sum assured
 made available to recipient of letter (and spouse, perhaps) only
 no death benefit payable within, say, two years
 target at unsophisticated market
 limited time available in which to apply after a mailing or advert
 underwrite at claim stage (although, for life assurance, this is not a popular strategy in the
UK)
 exclude pre-existing conditions
 maximum age at entry.
 The interaction between the level of underwriting and the potential level of sales –
the greater the level of underwriting the lower the mortality costs, but the expenses
will rise. Also, people may be more inclined to take out contracts where there is a
low level of underwriting. Furthermore, less underwriting leads to quicker
processing of new business proposals. See Section 3.1 below for further discussion
of this point.
 Claims underwriting may deter people from taking up a contract, due to the
uncertainty as to whether a claim may be accepted.
 The effectiveness of the proposed underwriting – benefit exclusions may be difficult

to police, or limiting medical evidence may make it difficult for staff to underwrite
effectively. Non-disclosure also makes underwriting less effective.

 The more detailed the underwriting, the greater the homogenisation of risk that may
be achieved.
Question
(i) Explain what is meant by ‘homogenisation of risk’.
(ii) Comment on whether it is a good thing.
Solution
(i) This means that applicants are classified, for rating purposes, in such a way that the risks
for applicants classified together are similar.
(ii) The more ‘cells’ we have in this classification the more accurate our rating can be for each
individual, which is a good thing. However, more accuracy means that more information
is required from the proposer, and so we are back to our familiar cost-benefit question of
whether this greater accuracy actually pays for itself.
 The impact of regulation, which might constrain the level and/or type of underwriting
that is permitted, eg the use of genetic testing.
For example, in the UK consumers do not have to disclose any past genetic test results,
for cover up to specified limits (eg £500,000 for life insurance). Above these limits
information on past results may be requested for approved tests only.
 The interaction between underwriting level and terms offered by the company’s
reinsurers (see Section 3.2 below).
With more stringent underwriting, the level of claims should be more predictable and so
reinsurers are likely to include lower margins when assessing the terms they are prepared
to offer. Also, if a direct writer tightens up on underwriting and so expects a more
predictable level of claims, it will require less reinsurance and so might increase its
retention levels.
 How to vary underwriting criteria by age, sum assured, target market and various
other factors.
3.1 Marketing
Underwriting is generally recognised as a barrier to sales as well as being time-consuming
and invasive. Although contrary to minimising risk, less (or no) underwriting may be used
to increase the marketability of policies. Although risk costs would be higher with such
policies the other costs may not be, making overall pricing competitive.
So, less underwriting can help sales by simplifying the selling process, making the sale easier, and
also possibly by making the premium cheaper.

Question
Give two distinct examples of how more underwriting can increase the marketability of policies.
Solution
(1) More underwriting will result in an increase in costs. However, if this is outweighed by a
reduction in claim payments then the company will be able to charge lower premiums.
The policy will then be more attractive to the customer, increasing marketability.
(2) As was mentioned in Section 2.1, underwriting may be used with annuities in order to be
able to offer enhanced terms to impaired lives, eg smokers.
The levels of sum assured, or other criteria, at which underwriting is automatically applied
may be relaxed in certain products and/or sales channels of an insurer to increase
competitiveness and hence marketability.
In the case of insurance intermediaries, the levels of underwriting adopted by insurers can
be used as a discriminator of which products, or indeed which insurers, might be
recommended to clients.
Question
A life insurer decides to relax the underwriting used for term assurances.
Explain how this might affect the premium it will charge.
Solution
It depends.
Relaxed underwriting is likely to lead to:

 a reduction in underwriting costs, reducing the theoretical premium
 an increase in anti-selection and hence mortality experience, increasing the premium
 improved marketability, which should increase sales, reducing per policy costs and
increasing total profits (assuming per policy profits are positive).
The total effect on the premium will depend on the relative size of these factors. However, even
then the company may decide to charge a different premium for, say, competitive reasons.
Certain policy options may be available where the policyholder can increase and/or extend
the length of cover without evidence of health. The increased anti-selection risk may be
allowed for by an increase in overall product charges compared to a corresponding policy
without the options.
However, it is still exposed to the risk that it underestimates the increase in product charges that
are required. The valuation of this type of option was discussed in Chapter 23.

Although the introduction of such options might make the policy more marketable, the increased
premiums that are likely to accompany such options will have the opposite effect.
3.2 Reinsurance terms

If a reinsurance arrangement is in place for certain classes of business, the reinsurer will
require a certain level of underwriting to be in place and will carry out reviews of the
underwriting carried out. The terms offered will depend on the past claims experience of
the insurer and the underwriting standards in place, together with any expected change in
that experience consequent to changes in underwriting standards.
This might be another advantage of underwriting. The more underwriting that is used, the better
the reinsurance terms that might be offered, increasing profitability (all else being equal).
It is usually the ceding company’s responsibility to maintain the specified underwriting

standards. This involves potentially considerable operational risk, as the reinsurer may not
pay claims if the underwriting was not up to the specified standard.
If a company changes the level of underwriting it uses then it will probably be required, under the
conditions of the reinsurance treaty, to inform the reinsurer. The terms of the treaty, in particular
the price, will then need to be renegotiated.


Chapter 26 Summary
How underwriting can be used to manage risk
Underwriting will:
 protect the insurer against anti-selection
 identify the lives with substandard health risk
 identify the special terms to offer to the substandard risks
 help ensure that all risks are rated fairly
 help ensure that mortality experience is consistent with the pricing basis
 reduce the risk from over-insurance (through financial underwriting).
Underwriting is particularly important for contracts that have a high sum at risk.
Medical underwriting
By obtaining various pieces of medical evidence the company will be able to gauge the
health of an applicant and hence decide on the appropriate premium rates (or possibly
decide to reject the proposal).
The evidence can be (in order of detail, and also cost):
 a medical examination carried out on the applicant
 specialist medical tests on the applicant.
Financial underwriting
The company should ensure that the proposed policy is in keeping with the probable needs
of such an insured life – in particular that the sum insured is not strangely high.
This will help ensure that:

 the premiums are affordable, controlling the persistency risk
 the applicant is not trying to commit fraud.

Special terms
Applicants whose state of health does not reach the required standard might be:
 declined
 deferred
 offered special terms, eg:
– an increased premium
– a reduction in benefit
– an exclusion
 offered an alternative policy.
Some ‘underwriting’ can also be done at the claims stage.
Determining the level of underwriting to use

Underwriting is part of a life insurer’s risk management strategy. The level of underwriting
required should balance the risk to the company of mis-estimating the risk, and the cost of
obtaining the evidence. Hence, evidence requirements normally increase with the sum at
risk – the trigger points are known as medical limits.
Harsh underwriting will tend to discourage sales but conversely relaxed underwriting may
improve marketability.
Policy options may be available where the policyholder can increase and/or extend the
length of cover without evidence of health.
Regulation may constrain the level and/or type of underwriting that is permitted eg the use
of genetic testing.
A reinsurance treaty will normally require a certain level of underwriting.
The terms offered will depend on the underwriting standards in place, and it is usually the
responsibility of the insurer to ensure the required standards are met.


26.1 (i) List the sources of medical evidence that may be used to underwrite an application for life
Exam style
insurance. [2]
(ii) Give reasons why some of these sources may not be used for all applications in
practice. [5]
[Total 7]
26.2 A life insurance company in a developed country writes term assurance business. It is deciding
how best to adjust premiums for the extra risk in the case of lives that it is willing to accept on
special terms. It is considering the following possibilities:
(a) Adding a fixed amount c to the qx at each age in the premium formula.
(b) Multiplying the qx by an appropriate fixed percentage k% at all ages, where k  100 .
(c) Calculating the premiums for someone aged x by assuming an age of x  n , with n chosen
appropriately.
Comment on which of the above would be suitable for the following risks, giving a brief
explanation in each case:
(i) The risk resulting from the practice of a dangerous pastime such as bungee jumping from
hang gliders.
(ii) The risk resulting from residence in an under-developed country.
(iii) A family medical history of heart disease.
26.3 For each of the following contracts:
(a) a term assurance
(b) an immediate annuity
(i) Describe, with reasons, at what point underwriting would take place.
(ii) Describe how the company might alter the terms offered for new business if a new fatal
disease were to emerge.
26.4 Describe briefly:
(i) the reasons for underwriting life assurance proposals
(ii) the effect that the standard of underwriting will have on the pricing of the contract.

26.5 (i) List the four main general stages of underwriting for a proposal that is not acceptable at
Exam style
ordinary rates. [2]
(ii) Discuss whether or not it would be appropriate for a life insurance company to offer a
conventional short-term temporary assurance contract with no underwriting, if, as has
been suggested, a very high premium rate per mille could be charged for such a policy,
and if the premiums could be reviewable at the discretion of the office. [4]
[Total 6]
26.6 (i) Describe four ways in which a life insurance company can obtain underwriting
Exam style
information about a person making an application for a life insurance contract. [5]
(ii) Explain how the information in part (i) would be used by the insurance company, and the
purpose that is served by this process. [6]
(iii) Explain why all four sources of information are not necessarily obtained for all applicants,
and how the company may decide upon which of the four to obtain in any particular
case. [3]
[Total 14]
26.7 A life insurance company issues conventional without-profits term assurances and with-profits
Exam style
endowment assurances through insurance intermediaries. It sells similar volumes of both classes
of business.
The company currently operates a uniform approach to underwriting all its business, seeking
additional medical information from applicants of both policy types according to identical criteria
of age, size of sum assured, health and lifestyle information obtained from the proposal form.
The company wishes to increase sales of both contracts, and is considering whether to relax the
company’s underwriting criteria.
Discuss this possible relaxation of the underwriting criteria. [7]

26.1 (i) Sources of information

 a medical examination carried out on the applicant
 specialist medical tests on the applicant. [½ each]
(ii) Reasons why these may not be used
The basic principle behind obtaining and using the underwriting information, is to enable the
company to control its mortality risks in the most cost-effective manner.
Reasons for not using all the information could therefore include:
 to save on expenses
 when the opportunity for anti-selection is negligible (eg for a standard annuity)
 in order to be competitive and to increase sales (people don’t like revealing medical
information)
 if the amount of cover is small (so the financial effect of anti-selection is minimal and
would not justify the cost of obtaining the information)
 if group cover is being applied for (as anti-selection may be lower)
 if the cover is compulsory (as anti-selection is negligible)
 if such an invasion of privacy is not culturally acceptable
 if not allowed under legislation
 if earlier sources (eg proposal forms) suggest no problems, then further investigation may
not be cost-effective
 if the product specifies that underwriting will not be performed (eg on the exercise of the
option under a renewable or convertible term assurance)
 if the information is simply not available (eg medical reports, either due to data
protection or because there aren’t such things). [½ each]
[Maximum 5]
26.2 (i) Dangerous pastime
Probably (a), a fixed addition.
This is because there is unlikely to be significant variation in the amount of extra risk with age,
and (b) and (c) will add more in absolute terms at older ages.
Any additional premium might apply only while the person continued to practice the sport.

(ii) Residence in under-developed country
(a) is again the most likely.
Many of the additional risks will not be particularly age-dependent (eg only basic medical
services, risk of being eaten by a tiger etc).
Although could make out a case that older people will be less able to deal with the consequences
of some of these things (eg less able to run away from the tiger, more vulnerable to disease), in
which case (b) or (c) would be better.
(iii) Family history of heart disease
(b) would be a strong contender here, as the risk is likely to increase with age.
(c) is also possible, as it is in fact a pragmatic and simple way of achieving a similar effect to (b).
26.3 (i) Underwriting
(a) Term assurance
Proposals will be medically underwritten at outset, in order to charge a premium that is

commensurate with the risk being accepted. There is also likely to be financial underwriting at
outset, to check that the level of cover requested is not suspiciously high.
At the time of a claim, the company will want to see the death certificate before the claim is paid.
On occasions, the company may want to query the original information supplied, but it would
usually be hard to prove that anything had been withheld.
There is often no underwriting for annuity business.
If enhanced rates are offered for unhealthy lives (eg smokers), then there will be some form of
underwriting at outset to ensure the proposer fulfils the specified criteria. In theory, a company
may wish to check periodically that the criteria continue to be met; however this isn’t really
practical.
(ii) New fatal disease
(a) Term assurance
The risk of death (and therefore payment of a claim) is now greater, so overall, premium rates are
likely to increase.
If symptoms can be detected, people showing them may be declined cover, or an exclusion clause
may be added (though this would be hard to enforce).
If the disease takes a long while to cause death, policies might be offered with an increased
premium (especially shorter-term policies).
Underwriting may be tightened up to try to identify those with a higher than average chance of
contracting the disease.

Any renewal or convertibility options might be removed.
Due to the extra risk, reinsurance arrangements will be reviewed. This may result in more costly
cover being reflected in premium rates.
People with the disease are unlikely to purchase an annuity at normal rates, since this would
represent poor value for money.
The company could offer better rates to infected people, but there would be the risk that a cure is
found, making the business less profitable than anticipated.
Mortality rates from the disease will be very uncertain, so experience should be monitored
frequently if enhanced rates are offered.
The most likely result, though, is no change, or possibly a slight improvement in rates.
Guarantee periods might become more attractive to policyholders, in light of the possibility of
contracting the disease soon after purchase. This will depend on how long it takes for the disease
to prove fatal after contraction.
26.4 (i) The reasons for underwriting
The key purpose of underwriting is to ensure that the company is charging a fair premium for the
risks that they are taking. If the premium is too low, then the company will make a loss on the
contract. If it is too high then the company will lose the business.
In particular, the company wishes to:

 protect itself against anti-selection – in particular, the company wants to identify lives in
such poor health that it should not give them cover
 classify other lives appropriately, to reduce heterogeneity – this helps to ensure a fair
premium but also makes data easier to analyse and reduces the variance of claims
 identify the level and type of loading appropriate for extra risks
 underwrite to the extent that it at least pays for itself in terms of improved mortality
experience
 combat over-insurance using financial underwriting.
(ii) Effect on pricing
The critical thing is the connection between level of underwriting and mortality experience. The
stricter the underwriting, the lighter the mortality experience.
This will also affect expenses. The stricter the underwriting, the higher the initial expenses.
Level of underwriting may also affect the level of ‘selective’ withdrawals, ie healthy lives
withdrawing and only less healthy lives keeping the contract in force.

26.5 (i) Underwriting stages for non-standard lives
The stages are:

 identification of overall class of life (eg sales channel involved)
 obtaining medical and financial evidence
 assessment of extra risk (type, intensity and incidence)
 translation of extra risk into terms to be offered to policyholder.
[½ per point, total 2]
(ii) Offer term assurance with no underwriting
This contract is not feasible. [½]
High premium rates would not guard against the massive scope for very poor risks (for instance,
very ill people) selecting these policies. [½]
Healthy lives would buy their insurance from companies that do underwrite and so have lower
premiums. [½]
There is nothing to stop brokers going round hospitals signing up the dying. [½]
A typical term assurance premium rate is say 4 Francs per annum per mille ie for 4F, we get a sum
assured on death of 1,000F. So even if the premium rises to 40F per annum per mille an
enormous loss can be made. [½]
If term assurance was offered without underwriting the insurer would face insolvency. [½]
Variable premiums do not help either. [½]
With the huge losses from death-bed proposals, the insurer would have to increase premiums by
a ridiculous level to compensate. Any policyholder in reasonable health would lapse their policy
and the insurer would not receive the increased premium. [1]
[Maximum 4]

26.6 (i) Sources of information
Proposal form [½]
Applicant has to supply information about:

 current health and medical history
 height, weight, gender, age
 contract applied for and size of benefit
 lifestyle: eg smoking habits, alcohol consumption, hazardous pursuits
 family history of disease and of early death
 occupation and income
 country of residence
 whether special terms have been applied (or offered) by other insurers in the past.
[Maximum 2 marks for at least 6 points]
Medical attendant’s report (MAR) [½]
This is a confidential report written by the applicant’s medical attendant (doctor). It gives an
account of the applicant’s medical history and an opinion about the applicant’s future health. [½]
Medical examination (ME) [½]
The applicant is required to undergo a medical examination by the company’s medical officer or
some other independent medical practitioner. This may include routine medical tests: eg blood
pressure tests. [½]
Specialist medical tests [½]
This will involve investigations about specific conditions: eg an electrocardiogram, chest X-ray,
etc. [½]
[Maximum 5]
(ii) Use of the information
The information is used to classify the applicant as to whether he or she is insurable at standard
rates, [½]
…or if special terms apply, [½]
…and in the case of the latter, what the special terms should be. [½]

Special terms include:

 charging an extra premium [½]
 reducing the benefit payable on death [½]
 postponing the decision (eg when prognosis is uncertain but should become clearer at a
later date) [½]
 issuing a contract with exclusion from particular causes of claim [½]
 declining insurance, if allowed in the country concerned [½]
 shortening the policy term. [½]
Purpose of underwriting
The aim is to ensure that the policyholder is charged a premium that accurately reflects the level
of risk presented, and that as few applicants are declined insurance as possible. [1]
Underwriting will help to ensure that the actual experience does not depart too far from the
assumptions of the pricing basis, and hence that the expected profitability will materialise. [½]
Underwriting will help reduce the financial risk from over-insurance (preventing people from
taking out policies for larger sums assured than they need). [½]
The intended result is to minimise the risk of anti-selection by high-risk applicants seeking
inappropriately cheap rates. [½]
[Maximum 6]
(iii) Why not all information for everyone?
In many cases the risk can be adequately and accurately classified on less than the total possible
information, and this therefore saves on expense. [1]
The majority of applicants can be judged to be a standard risk from the proposal form. The few
high risks that are let through at inadequate premiums can be tolerated by the fund provided
their sums assured are not too high. [½]
Determining which information to obtain
All applicants will fill in the proposal form. [½]
Adverse information given on the proposal form will automatically trigger either or both of the
MAR or ME. [½]
Sum assured levels higher than the company’s chosen threshold for a particular age and contract
type will automatically lead to MAR and/or ME even if proposal form is clear. [½]
As the ME is more expensive the automatic threshold is higher for the ME than for the MAR. [½]
Specialist examinations will be triggered when necessary to provide additional clarification about
conditions identified from any of the other sources. [½]
[Maximum 3]

26.7 Effect of relaxing underwriting
Relaxing underwriting can make products more marketable. [½]
Long and drawn-out underwriting procedures can put people off applying for life insurance. [½]
This is especially the case when insurance intermediaries are involved in the sale, as they will soon
get to know which companies go in for more or less underwriting, and will tend to recommend
those companies with the simplest procedures for their clients, all else being equal. [1]
Less underwriting should also reduce the insurer’s costs, increasing profitability. [½]
On the other hand, lower underwriting standards will increase the risk to the insurer from anti-
selection, as higher risk lives are more likely to be charged the standard rates of premium than
before. [1]
The insurer has to decide whether the increase in the average claim costs is more than offset by:
 the lower average underwriting costs per policy sold [½]
 the increase in volume of business achieved, which will reduce the per-policy cost of the
overhead expenses and generate greater total profit contribution from the business. [1]
Additionally the expected increase in mortality (and reduction in expenses) could be incorporated
into the pricing assumptions, which could alter the competitiveness of the contracts further. [½]
Differences between products
The situation could be very different for the two product types. Underwriting, anti-selection and
price are of much more critical importance for the term assurances, so there will be a limit to the
extent to which underwriting could be relaxed, if at all, for these policies. [½]
Anti-selection should be less severe for the with-profits policies, which will largely attract savers.
For these contracts, lengthy underwriting would be seen as an unnecessary hindrance to sales
and could well be relaxed without significantly increasing the mortality claims costs. [1]
Premium rates would almost certainly not be altered for the with-profits business, for which any
change in experience would normally be dealt with by an adjustment to future profit
distributions. [½]
Depending on the level of underwriting currently employed, it might even be beneficial for the
term assurance business if the underwriting criteria were made more stringent, provided this can
be accompanied by a significant reduction in price, thereby making the product more price
competitive and hence more marketable that way. This may be a more significant effect than the
increased underwriting hassle incurred. [1]
[Maximum 7]





using the course.

SP2-27: Policy data checks Page 1
Policy data checks

Syllabus objective
 policy data checks.

Page 2 SP2-27: Policy data checks
0 Introduction
The actuary will want to ensure that his or her work is founded on valid data. In addition to this
professional peace of mind, often the actuary will be the only person (or the best placed person)
to check the data at an ‘overall reasonableness’ level. For instance, staff in the systems area who
might be ostensibly responsible for data accuracy cannot be assumed to know how various
numbers ought to be behaving, even at a broad level.
A number of tests may be conducted towards ensuring that data are adequate, accurate and
complete. The actuary should ensure that all the data items required are being maintained
by the company and can be accessed in the required format.
This chapter describes assorted checks that the actuary can use to help ascertain the validity of
the data.

1 Data reconciliation checks

Where an investigation is being carried out on a regular basis, a reconciliation can be
attempted of the current data with those used for the previous investigation.
The data are first grouped in some sensible way, for example by year of entry within each
broad contract type. Using data similarly grouped relating to business that has come onto
the company’s books and gone off between the dates of the two investigations, the
following check is made for each group:
data at business business data at

previous + come onto – gone off = current
investigation the books the books investigation
The above can be used to check the following items of data:
For non-unitised business:

 number of contracts
 basic sum assured – or the equivalent benefit depending on the nature of the
contract
 office premium
 amount of any attaching bonuses (for with-profits contracts).
For unitised business:

 number of units allocated, sub-divided by unitised fund
 current premium payable
 current benefits available, eg amount of death cover.
If relevant, the reconciliation also needs to allow for items such as:
 changes in the number of units allocated arising from switches between unitised
funds
 changes in the premium payable and benefits under existing contracts.
The systems for producing the movements data must be checked periodically to ensure that
they are working correctly and that the staff involved are following the procedures laid
down.
The movements data should also be checked against any appropriate accounting data,
especially with regard to benefit payments.
For instance, the claims paid out according to the movements data should be equal to the total
claims costs in the company’s revenue account.

It is important for the movements data to be independent of (ie separately generated from) the
in-force data.
Question
Explain the preceding statement.
Solution
If the movements data were calculated as a function of the in-force data, eg:
[movements off] = [in force at start] + [new business] – [in force at end]
then nothing is being checked at all.
A company with good data validating systems will have all items recorded as they occur, with
records of movements on and off kept separately from, and in addition to, the current in-force
record. At the same time, the actual movements on and off will be input to the current in-force
record, which will therefore be subject to continual updating over the year. The reconciliation
check will then pick up discrepancies and hopefully help identify the errors involved.

2 Consistency checks
Examples of such checks for non-unitised contracts are:
 The average sum assured or premium for each class of business should be
sensible, and consistent with the figure for the previous investigation.
 The ratio of the basic sum assured (or equivalent benefit) to the premium payable for
each class of regular premium contracts should be sensible and consistent with the
figure for the previous investigation.
Question
Explain what relationship would be expected between the average sum assured of all endowment
assurances this year and last year.
Solution
Probably higher, as the policies going off the books (say 25 years old) will tend to have a much lower
sum assured than the policies taken on this year.
(The relationship between new sums assured this year and last year is not the crucial one to
consider.)
Question
Identify the data error:
With-profits endowments Sum assured $50m

Term 10 years
Annual premium $6m
Without-profits endowments Sum assured $50m

Term 10 years
Annual premium $4m
Term assurances Sum assured $50m

Term 10 years
Annual premium $1m

Solution
We should be most suspicious of the term assurance data (dubiously high premiums).
A useful rule of thumb that can be used here (and is applicable in many ‘G7’ countries) is that
with-profits endowment assurance premiums are quite frequently larger than the sum assured
divided by the term. Without-profits endowment assurance premiums will normally be smaller
than the corresponding with-profits premiums.
Examples for unitised contracts are:
 A check that the numbers of units purchased by premiums and encashed to pay
benefits are consistent with the corresponding revenue account items.
 A check that internal unit movements, for example charge encashments, are
consistent with the surplus emerging during the year.
Here, ‘charge encashments’ generally refer to monetary expense and mortality charges that are
deducted from the policyholders’ funds by the cancellation of units.

3 Other checks
3.1 Unusual values

Checks should be made for any unusual values, for example:
 very large or zero unit values under unitised contracts or benefit values for
non-unitised contracts
 impossible dates of birth or retirement ages or start dates.
As well as looking at individual values, it may also be possible to group items and look at
how well distributed they are. For example, an unusually high clustering of birth month may
represent a data input error worthy of further investigation.
3.2 Spot checks

It is good practice to compare an extract of the computer held data with the information in
the paper administration files. This can be done on a spot check basis by randomly
selecting a number of policies.

4 Analysis of surplus and embedded value profit

An analysis of surplus is an investigation that seeks to explain the reasons for the change in the
valuation result (ie the value of the assets less the value of the liabilities) between one valuation
date to another. It is discussed in more detail in Chapter 30.
A major discrepancy in the analysis compared to previous analyses may indicate a problem
with the data.
The same will also be true of large discrepancies in any analysis of the change in embedded value.
This will be covered in more detail in Chapter 30.

Chapter 27 Summary
The actuary will want to ensure that the data he or she uses are complete and accurate.
Various simple check procedures exist. These examine:

 data movements (reconcile data between start and end period)
 consistency between data items and over time (check numbers and ratios are
sensible)
 unusual values (check whether they are in fact valid or not)
 consistency between systems data and paper admin files (use spot checks)
 analysis of surplus (any unexplained items?)



27.1 The actuary of an established life insurance company has been having problems with data
Exam style
accuracy in their modelling work.
(i) Describe the checks that they could make to test the data. [6]
(ii) Explain how these problems might affect the modelling work. [2]
[Total 8]



27.1 (i) Data checks
First, a data reconciliation between ‘now’ and the previous model run. [½]
Group data into large blocks (not model points) such as products subdivided by year of entry. [½]
Test that:
Data at previous investigation
+ business come onto the books
– business gone off the books
= data at current investigation. [½]
Use this to check items such as policy numbers, premium, sum assured, units by fund (for linked
business). [½ per example, maximum 1]
While doing this check, it is also useful to check the movement data against appropriate accounts
data. [½]
Secondly, check consistency. [½]
This means checking that the averages by class (eg products by year of generation) of important
values such as premium and sum assured are sensible, … [½]
… and that the ratios of sums assured to premiums are reasonable, … [½]
… and consistent with last year. [½]
Thirdly, check for unusual values. [½]
Such as very large sums insured, or ridiculous ages, or total units held of zero. [½]
Also check for groupings in specific years or specific premium sizes. [½]
Fourthly, perform spot checks of the data held on systems against paper admin files. [½]
Finally, look at the valuation analysis of surplus. Any major discrepancies should be investigated
as they may indicate data problems. [½]
[Maximum 6]
(ii) Effect on modelling work
Inaccurate input will lead to inaccurate results. [½]
Such inaccuracy may get compounded as the projection period rolls forward. [½]
Bad data may cause inaccurate mortality / withdrawal investigation results … [½]
… and hence inaccurate parameter estimation. [½]

[Total 2]

End of Part 5
What next?
Chapters 23 to 27.
of your revision.
… ‘revision and rehearsal’ products
ASET – This contains past exam papers with detailed solutions and explanations, plus lots of
comments about exam technique. One student said:
‘ASET is the single most useful tool ActEd produces. The answers do go
into far more detail than necessary for the exams, but this is a good
source of learning and I am sure it has helped me gain extra marks in the
exam.’
You can find lots more information, including samples, on our website at www.ActEd.co.uk.

SP2-28: Investment Page 1
Investment
Syllabus objectives
3.4 Demonstrate the application of asset-liability matching as a risk management

technique.
3.4.1 State the principles of investment for a life insurance company.

3.4.2 Analyse life insurance liabilities into different types for asset-liability
matching purposes.
3.4.3 Propose an appropriate asset-liability matching strategy for different types
of liability.
 developing investment strategy.

Page 2 SP2-28: Investment
0 Introduction
In this chapter we consider how a life insurance company should invest its assets. The investment
strategy followed by a life company is of critical importance to its:
 commercial success
 financial results
 security.
As life insurers in some markets find themselves competing more and more both with other life
companies and with other savings institutions such as banks, more and more attention is paid to
the investment performance of the policyholders’ funds. Potential with-profits or unit-linked
policyholders will normally be influenced more by recent performance than by premium rates (or
charges in the case of unit-linked business). Even for without-profits business, higher expected
investment returns will allow lower premium rates. High investment returns are therefore of
great commercial importance to the life company.
High returns on assets will also be necessary to provide good financial results. Here the assets fall
into two categories – those where all or most of the return goes to policyholders (ie unit-linked
and with-profits business), and those where the shareholders profit directly from high returns
(ie without-profits business and the shareholders’ own free assets). Even where most investment
return goes to policyholders – for instance with-profits business where typically 80 to 90% of
returns may be distributed to policyholders as surplus – the life company will still want to
maximise returns for the shareholders as well as for the policyholders.
Finally, the choice of investment strategy will affect the security that the life company is able to
offer to policyholders. Most obviously, that guaranteed benefits will be met; but also that the
returns which policyholders might reasonably expect to receive will be earned.
Expressing all of the above in the most simple terms, the life company will want high returns
which are not volatile. However these two objectives conflict; investments offering higher returns
normally show relatively volatile returns. We shall see below how the life company will attempt
to balance returns and security, given the nature of its liabilities.
Before starting to consider the selection of the most appropriate investment strategy for a life
company, we revisit briefly the characteristics of the major asset types in Section 1. This should
be familiar from earlier subjects.
In Section 2 we look at the principles of investment and in Section 3 look at how different types of
asset can match different types of liability.
Finally in Section 4 we show how modelling can be used to determine an appropriate investment
strategy.

1 Asset characteristics
We can consider asset types according to the following characteristics:
 return
 security (and volatility of market value)
 marketability
 expenses
 term
 currency.
Additional points to consider would include any statutory constraints on the life company in
holding that asset, and any problems caused by the minimum size of the investment.
Of these factors, the most important aspect is the return on the asset. In considering the return,
we shall need to think about:
 how much is the expected return?
 is the return real or nominal?
 how is the return split between income and capital redemption?
 is the running yield sufficient for the investor’s needs?
 what is the variance of the return?
 how is the return affected by tax?
We now consider how the major asset types compare, using the above categorisation. Since it is
meaningless to discuss returns in isolation, we shall compare their relative returns at the end. In
addition, we shall not comment in detail on tax, since the tax treatment of individual asset types
will depend on local market legislation. However it is worth bearing in mind the following generic
implications of the tax treatment of investments:
 tax reduces returns
 tax regimes may favour investment in particular assets
 tax regimes may favour income over capital gains (or vice versa).
Government fixed-interest bonds

These offer a nominal return. They are normally issued with a coupon yield similar to market
yields, although some countries also offer zero-coupon bonds where the running yield is zero and
all of the return comes through the high redemption value. The return on these bonds is not
variable (although it is variable if not held to redemption – this is an extremely important point
when it comes to consideration of matching issues).

They are normally considered the most secure asset type, other than cash. Although the market
value of such bonds may fluctuate with the market, the eventual redemption value will be
unaffected by such fluctuations.
In most countries they are the commonest and most marketable asset type. Dealing costs tend to
be very low. The choice of term available will depend on the local market, but a typical range is 0
to 20 years.
Government index-linked bonds

These differ from fixed-interest bonds in that the coupon payments will be defined with reference
to some index or value – such as local price inflation.
The amounts and variety of index-linked bonds issued by a government are often smaller than the
amount of fixed interest, which can render the index-linked bonds slightly less marketable.
Corporate fixed-interest bonds
Question
Describe the properties of corporate fixed-interest bonds.
Solution
Corporate fixed-interest bonds should offer a return slightly better than government bonds of
equal term. That return will not be volatile. Running yields will normally be similar to prevailing
interest rates for the term concerned.
Security could be a significant problem, especially if the issuing company is not AAA-rated. The
value of a bond may fluctuate with the markets, but this will not be too important if the investor
is prepared to hold the bond to redemption.
Marketability of such bonds is often poor, and dealing costs high.
The terms available are normally similar to those found for government bonds.
Equities
Equity investment offers an income (the dividend) which would be expected to increase in real
terms. The market value of the share should also increase in real terms from the purchase date.
However the running yields on equities are correspondingly low.
There are risks attached to the income and capital value. The underlying company itself might go
bankrupt, or just perform very badly. In addition, the market value of the share will be volatile.
This volatility can be a problem even when holding the asset for long-term income (because we
might need to value it at market value to help demonstrate solvency), and obviously a problem if
redeemed (it is sold for much less than was hoped).

What is the term of an equity? By assuming that an equity is redeemed after a certain period, we
can see that the nominal term can be as long as we require. However for liability matching
purposes it’s the discounted mean term that matters, and this is finite for equities.
In some markets equity investment is very developed, and is highly marketable, although even in
such markets there will be some almost unmarketable stocks. In other markets equity investment
may not really be an option because of the size and reliability of the local market; in this case a
life company might consider overseas investment in more mature stock exchanges.
Property
Question
Describe the investment characteristics of property.
Solution
Property is normally associated with a relatively high return. The income, in the form of rent,
gives a low running yield but should eventually increase in real terms.
Property investment is normally seen as secure, although the actual income stream may suffer
occasional interruptions. The market value can vary significantly, many property markets
suffering from some form of cycle.
It is a very unmarketable investment, with significant dealing expenses. Expenses are incurred
even in just administering (ie holding) the asset.
It is a very long-term investment – as with equities, it can be thought of as a perpetuity. Unlike

equities the option of buying with the intention of selling in the short term is not practical due to
the impact of dealing costs.
Direct property investment can only be made in large chunks – so a small life company may have
problems with diversification.
Cash
By cash we normally mean money held on overnight accounts earning spot rates of interest.
It is the most secure asset type, with the least variability of value. It is obviously very liquid, and
dealing costs are almost non-existent. It offers only a relatively low return and has discounted
mean term of zero.
Comparative returns
Question
Outline how returns on the investment types described above would be expected to compare.

Solution
The normal order would be (going from high to low):

 equities and property
 corporate bonds
 government bonds
 cash.

2 The principles of investment

These can be stated as follows:
(a) To minimise risk, a company should select investments that are appropriate to the
nature, term and currency of the liabilities.
(b) The investments should also be selected to maximise the overall return on the
assets, where overall return includes both investment income and capital gains.
(c) The extent to which the appropriate investments referred to above may be departed
from in order to maximise the overall return will depend, amongst other things, on
the extent of the company's free assets and the company’s appetite for risk.
These principles can also be expressed as:
 The company should invest so as to maximise the overall return on the assets,
subject to the risk being within the financial resources available to it.
This states the fundamental approach towards balancing risk and return. The aim is to maximise
return subject to an acceptable level of risk. The level of risk a company can take on, given its risk
appetite, will depend on the financial resources of the company – ie its free assets. We consider
below how holding more free assets enables a life company to pursue a riskier strategy.
Although the insurer may decide to mismatch deliberately as a means of seeking a higher
expected investment return, the actuary should still assess the asset-liability position in
order to understand the risks being run, including assessment of the adequacy of free
assets to support the mismatch.

3 Asset-liability matching requirements
3.1 Nature, term and currency of the liabilities

The first of the three investment principles above is about reducing risk, and essentially concerns
the amount and timing of the receipts, or proceeds, arising from the assets. If it is possible to buy
assets whose proceeds are identical to the outgo of money being paid out on the liabilities, as
they occur, then the assets and liabilities would be perfectly matched and there would be no
investment risk.
Perfect matching is a theoretical concept that is usually not achievable in practice. Nor is it
normally desirable, because it eliminates all possibility of making investment profit.
Question
Outline circumstances under which perfect matching would be desirable.
Solution
Perfect (or nearly perfect) matching is desirable if the company has very low free assets such that,
without matching, the probability of ruin would be unacceptably high.
Whatever the situation for the insurance company, the degree of matching between assets and
liabilities is of critical importance for its effect on the company’s actual investment risk. We
therefore need to consider in detail the nature, term and currency of the liabilities and how these
will affect the selection of suitable assets. In this section, we make the (simplifying) assumption
that this is the only aim of the company’s investment strategy.
The liability outgo consists of:
benefit payments
+ expense outgo
– premium income.
The expected liability outgo in any year, or month, depends on the monetary value of each
of the constituents and the probability of it being received or paid out.
The asset / liability modelling described later in the chapter will incorporate any probabilistic
effects. Here it is sufficient to bear in mind that, in considering any particular contract, we can
estimate the likely future outgo in respect of uncertain elements such as mortality claims,
surrenders and expenses. Future premium income is also an area of uncertainty. For a large
portfolio of policies and given a reasonable amount of past experience, there should be no
problem in deriving reasonably accurate estimates of future net cashflows.

Nature
Following on from the above breakdown of liabilities, we can consider separately the nature of
benefit payments, expense outgo and premium income.
The benefit payments can be subdivided into four types:
 Guaranteed in money terms – this consists of benefit payments where the amount
payable is specified in the life insurance contract in money terms. It will therefore
include the guaranteed benefit payments under all forms of without-profits contracts
and the accrued contractual benefits under with-profits contracts.
 Guaranteed in terms of an index of prices, earnings or similar – this consists of

benefits whose amount is directly linked to such an index.
 Discretionary – this consists of any benefit payments that are at the discretion of the
life insurance company, so in particular the future bonus payments under
with-profits contracts.
The level of discretion will depend upon the bonus distribution method used.
Discretionary benefits also include surrender values where these are not guaranteed.
 Investment-linked – this consists of benefits where the amount is determined

directly by the value of the investments underlying the contracts, so in particular the
benefits under unit-linked contracts (or benefits linked directly to a specified
investment index).
Expense payments tend to increase. The rate of increase is not strictly comparable to the
rate of change in a price (or earnings) index, but for investment purposes it is adequate to
treat it as being so. Hence, they can be included with benefit payments guaranteed in terms
of an index of prices or similar.
The point about expense inflation is that it will be unique to a particular company, depending
largely on the increases in salaries and other related costs that occur for that company. Here we
are treating expense inflation as if it is 100% correlated with increases in an ‘index of prices or
similar’ (salaries in this case). The difference to the company’s investment decisions caused by
this assumption should be negligible.
Premium payments are usually fixed in monetary terms and can be thought of as negative
benefit payments guaranteed in money terms. The existence of contracts where the
policyholder can choose the amount of premium to pay each year does not invalidate this.
Question
Explain the last sentence.

Solution
The idea of defining liability as the value of future benefit payments less the value of future
premium payments should need no explanation. For normal regular premium policies, a large
guaranteed sum insured is purchased at outset, financed by the contractual obligation to pay a
certain amount of premium in the future.
The right approach is less clear in the case of recurrent single premium products, where the
policyholder can choose at any time when and how much premium to pay. Valuing future
premiums in this instance will be impossible. This is not a problem because, for such policies,
there is no sum insured to be financed by future premiums. So in effect, for such contracts the
value of the liability corresponds purely to what has been bought by premiums paid to date, and
the value of future premiums and the benefit that they buy can be taken as zero.
However if the policy involves some guarantee whereby the policyholder can pay future
premiums under some advantageous terms, it would be more prudent to assume a certain level
of future premium payment, deducting the value of the associated purchased liability.
As a result, the liability outgo may be split into four categories:
 guaranteed in money terms
 guaranteed in terms of a prices index or similar
 discretionary
 investment-linked.
Question
Outline the nature of the liability outgo of each of the following contracts, by breaking it down
into these four categories of guaranteed in money terms, guaranteed with reference to an index,
discretionary and investment-linked:
(i) a regular premium with-profits policy halfway through its term
(ii) a single premium without-profits annuity with annuity payments increasing at the official
national consumer goods’ price index
(iii) a unit-linked regular premium savings contract offering a minimum sum assured on death
of $100,000.
Solution
(i) Regular premium with-profits policy halfway through its term
The benefit liability will be mostly guaranteed in money terms. There will also be some
discretionary liability outgo, in respect of future bonuses. The liability for surrender values is
normally discretionary. There will be a small liability in inflation-linked terms due to
administration expenses. There will be a significant negative liability of fixed future premiums.

(ii) Single premium without-profits inflation-linked annuity
The liability as far as the benefit is concerned is guaranteed in terms of the relevant inflation
index. There is also a liability in terms of the administration expenses – but this too can be
thought of broadly as guaranteed in terms of an index (probably mostly an earnings index).
(iii) Unit-linked regular premium savings contract, minimum SA = $100,000
The liability is primarily investment-linked, corresponding to the value of the policyholder’s unit
account. There is an inflation-linked liability arising from the expenses. At any point in time there
may be a small guaranteed liability in money terms arising from any excess of the sum insured
over the value of the units held. However this should correspond to the value of the death
benefit only over the next month (or until the next deduction from the units is made to pay for
the mortality cover), and so should be very small.
Term
In thinking about term in an investment context – whether discussing assets or liabilities – we
normally use the concept of ‘discounted mean term’ (another term for this is ‘duration’) rather
than the actual nominal term. The discounted mean term of an asset (or a liability) is defined as
the weighted sum of the terms of the payments, where the weight attributed to each term is the
present value of the payment at that term.
The discounted mean term of a liability is useful in considering an appropriate investment

strategy. If we have a liability with a certain discounted mean term, then choosing an asset with
an identical discounted mean term will give us an investment that will move in value with the
liability in the event of interest rate movements (provided certain other conditions hold – see
below). We are therefore protected against the risk of such interest rate fluctuations. This is the
concept of immunisation which we discuss further below.
Question
Estimate the discounted mean terms of the following, assuming an interest rate of 9% pa:
(i) a ten-year bond with a coupon of 6% pa
(ii) an equity whose dividends increase at 3% annually
(iii) cash
(iv) a ten-year single premium endowment assurance contract (immediately after sale),
assuming a constant mortality rate of q = 0.0015 pa.

Solution
The discounted mean terms are:
(i) Ten-year bond: 7.4 years
This is calculated using:

 
0.06  1 v  2 v 2  3 v 3  ...  10 v10  10 v10
, with v calculated at 9%.
0.06  a |  v10
10
(ii) Equity: 18.2 years
First let D be the initial annual amount of dividend. This will be increasing at 3% a year. The
discounted mean term (DMT) is then:
 1.03
     
1.03  3 1.03  ...
2 3
D 1 1.09  2 1.09 1.09 
 
     
 1.03 1.03  1.03  ...
2 3
D  1.09  1.09 1.09 
 
The Ds cancel.
The expression in square brackets in the denominator represents a perpetuity valued at an

interest rate of i  1.09 1.03  1 = 0.05825. The value of this perpetuity = 1 / i  17.167 .
The expression in square brackets in the numerator represents an increasing perpetuity valued at
this same interest rate, i = 0.05825. The value of this increasing perpetuity = (1  i) i2 = 311.886.
So the discounted mean term is:
311.886
 18.2 years.
17.167
(iii) Cash: Zero
(iv) Single premium contract: 9.9 years
This is calculated using:
 
q  1 v  2 v 2 p  3 v 3 p2  ...  10 v10 p9  10 v10 p10
 
q  v  v 2 p  v 3 p2  ...  v10 p9  v10 p10
where p  0.9985 , although any broadly similar mortality will give the same answer.
The discounted mean term will reduce significantly on allowing for surrenders – eg assuming a
5% pa surrender rate and a surrender value equal to the discounted (at 9% pa) value of the sum
assured, gives a duration of 7.9 years.

Currency
Liabilities denominated in a particular currency should be matched by assets in the same
currency, so as to reduce any currency risk.
An alternative way of matching is to invest in overseas assets and then to hedge the currency risk
by use of appropriate currency options.
Question
Explain why a company might want to invest overseas.
Solution
A company would need to invest in an overseas market if some of its liabilities were denominated
in the currency of that market.
In addition, the company might want to invest overseas in order to increase diversification, or to
invest in assets not available locally, or because some country’s assets seem to offer particularly
good value compared with the home market.
As we shall see, such deliberate mismatching will only be feasible if the company is able to
allocate some free assets to cover the potential downside.
Mismatching and demonstrating supervisory solvency

A careful distinction needs to be made here between the risk of cashflow mismatching, and the
risk from short term shocks in investment conditions.
The risk from cashflow mismatching is that, over time, the income from the asset proceeds falls
short of the outgo needed to meet the liabilities, due to such things as having to buy assets in the
future at lower than expected yields, or having to sell assets (to meet outgo) at depressed market
values. These events would result from a mismatching of assets and liabilities by nature, term or
currency, and its effect would only unfold over time as the actual cashflows took place. In order
to assess the true mismatching risk of a company, a cashflow projection model would be used.
The risk from short-term asset shocks relates to whether the company would continue to be able
to meet its supervisory reserving requirements if market investment conditions were to change
suddenly. Examples might be a change in fixed-interest yields or a fall in the capital values of
equities or property. To identify this risk, the company would have to analyse its supervisory
solvency position under different assumptions of current investment conditions.
To cover either of these risks a life company may have to set up additional reserves or hold
additional solvency capital.

3.2 Effects of the nature of liabilities

We now consider in further detail the investment implications of the four liability types defined
above:
 guaranteed in money terms
 guaranteed in terms of a prices index or similar
 discretionary
 investment-linked.
Guaranteed in money terms

A company will ideally want to invest to ensure that it can meet the guarantees. This means
investing in assets which produce a flow of asset proceeds to match the liability outgo. The
term of the liability outgo will also need to be considered, along with the probability of the
payments being made, so as to indicate the term of the corresponding assets.
This form of liability applies to without-profits policies and the guaranteed component of
with-profits policies. To simplify things, we shall focus on without-profits business. Here,
premiums and benefits are fixed. With large numbers of policies, the cashflows from premiums
and benefits (including expenses) will be statistically predictable. So the life company could try to
invest in assets which have an equal and opposite cashflow.
Except for certain types of company, it will probably be impossible in practice to find assets
whose proceeds exactly match the expected liability outgo. In particular, the terms of
available fixed-interest securities are often much shorter than the corresponding liabilities.
A best match is all that can usually be hoped for.
It will often be difficult to find assets which match the required liability cashflow. For instance,
there may not be any assets with a long enough term. Or, with a block of recently written
business, premium income may exceed benefits and claims for some years, giving a net inflow in
such years. No asset type can match that (without straying into the complexities of derivatives),
and so the company will have to invest these surplus premiums at unknown future rates.
The best match that the company finds may be a poor match for some patterns of outgo. Often,
the proceeds from assets will exceed the net liability outgo. The excess will have to be reinvested
at unknown rates, making strict matching impossible. This problem can be overcome by the use
of zero-coupon bonds, but they might not exist in the market being considered.
The technique of immunisation (covered in earlier subjects) may be used for this purpose.
However, that technique is subject to theoretical and practical problems.
In an immunised position – of which there would normally be many different possibilities, for any
given set of liabilities – the company would be protected against assets not covering liabilities in
the event of interest rates moving, provided the movement were not too large.
Question
State the various problems and limitations of this approach.

Solution
The problems and limitations of immunisation theory are:

(i) it immunises against profits as well as losses
(ii) it can be difficult to immunise against real liabilities
(iii) the theory only works with small interest rate fluctuations, though in practice it is actually
usually quite robust to large interest changes
(iv) the theory assumes that when interest rates change, the same change occurs at all terms
(ie it assumes a flat yield curve)
(v) we would need to reanalyse the situation every day and change investments accordingly
(vi) assets of the required term may not exist
(vii) the timing of asset proceeds might not be known, and that of liabilities can only be
estimated (for a life insurance policy).
Points (iii), (iv) and (v) do not apply to absolute matching (though the other points do), which
explains why immunisation is really ‘second best’ compared with absolute matching.
Given these various limitations, immunisation is really only of relevance for guaranteed liabilities;
and even then, the approach described later in the chapter will normally be preferable.
Guaranteed in terms of a prices index or similar

A suitable match would be securities that are linked to the same index in which the
guarantee is denominated, if available, ideally chosen to match also the expected term of
the liability outgo. In their absence, a substitute would be assets that are expected to
provide a ‘real’ return.
Question
Discuss how good equities would be as a match for liabilities linked to a price index.

Solution
Equities are not a perfect match for this liability because their return is not guaranteed, either in
absolute or index-linked terms.
However the return does behave to some extent in line with price inflation, and so they offer a
reasonable match. This would be contingent on diversification: any one equity stock is likely to
offer a very inferior match to an inflation index (quite apart from the exposure to asset default).
The suitability of the match is also a function of how long a time period is considered. Hence over
a short time period the match will be poor, due to the high volatility of the return from equities
over the short term, while over a longer period the match will be much better, as the long term
return from equities tends to be less volatile.
Discretionary
The most important example of a discretionary liability is the non-guaranteed portion of
with-profits liabilities. In other words, the excess of asset share over the (realistic) reserve.
Another common form of discretionary liability is the surrender value (or paid-up policy value)
when it is not guaranteed by the policy contract.
The main aim of the company will be to maximise the discretionary benefits and hence the
investment strategy should also aim to do that. This means investing in assets that will
produce the highest expected return.
Although this is broadly true, an important consideration here is the expectations of policyholders
and their attitude to risk. Normally, with-profits policyholders will want returns that are high in
real terms. However they will normally prefer a medium-risk approach which is almost certain to
surpass cost of living inflation, to a high-risk approach which may give much greater returns but
could give very unsatisfactory returns. Those policyholders who want a high-return / high-risk
approach would take out a unit-linked policy, choosing appropriate specialist funds.
The recipients of the discretionary benefits will usually expect the proceeds of their
contracts to maintain their value in ‘real’ terms. Hence the assets recommended should be
those that are expected to provide a ‘real’ return.
Given these considerations, a common approach in choosing assets to match discretionary

liabilities is to invest in equities, but primarily equities in secure established companies in
mainstream sectors (what a stockbroker might call ‘blue-chip’) rather than high-return / high-risk
emerging-sectors equities.
In addition to the choice of assets in respect of the discretionary liabilities for a with-profits
contract, the choice of assets for the guaranteed liabilities might be affected by the balance
between discretionary and guaranteed benefits, which in turn is influenced by the method used
to distribute bonuses.

Example
A block of with-profits business has liabilities that are 50% guaranteed, 50% discretionary. From
what we have read so far, one choice of suitable assets might be 50% government bonds of
suitable term and 50% equities.
However the company might say, ‘The assets for the discretionary benefit are also giving us
security to cover the guaranteed benefits’, and so might choose some equities to back the
guaranteed 50% – perhaps giving a total allocation of 25% bonds and 75% equities.
In effect, the company is using the assets in respect of the discretionary benefits almost as if they
were free assets to support the investments in respect of the guaranteed benefits.
The decision on assets for discretionary benefits, and whether this will affect the investment
approach for guaranteed benefits, will depend on a number of factors:
 the bonus philosophy of the company – a company which gives low guaranteed benefits
and concentrates on terminal bonus or a significant terminal dividend may tend toward
investing primarily in equities, only worrying about matching as policy maturity
approaches
 the level of free assets – we shall see below how free assets can support a riskier
investment strategy
 the risk appetite of the company
 the published (or expected) investment strategy of the company – policyholders’
reasonable expectations will be influenced heavily by what the company states that it
does, and so the company should not deviate too much from a particular publicised risk vs
return stance
 the views of the company at any time as to the relative performance of the asset classes
concerned.
Investment-linked
The benefits are guaranteed in the sense that their value can be determined at any time in
accordance with a definite formula based on the value of a specified fund of assets (or
investment index). Clearly the company could avoid any investment matching problems by
investing in the same assets as used to determine the benefits.
The most common example of this liability is a unit-linked product. In some markets the life
company might be required to invest so as to match the underlying assets exactly. Even if not
legally obligatory, it is normal practice to do so and there would have to be very strong reasons
not to do so. One such reason would be if the company thought that it could profit from such a
deliberate mismatch (eg the units are defined in terms of asset A, the company thinks that asset B
will perform better over the next six months so the assets invested by the company to meet the
unit liabilities are switched from A to B.) We shall see below how such a position might be
allowed by a sufficient level of free assets.

Question
Describe an appropriate choice of backing investments for each of the following contracts,
assuming that no free assets are available:
(i) a regular premium with-profits policy halfway through its term
(ii) a single premium without-profits annuity with annuity payments increasing at the official
national consumer goods’ price index
(iii) a unit-linked regular premium savings contract offering a minimum sum assured on death
of $100,000.
Solution
(i) Regular premium with-profits policy halfway through its term
Suitable assets in respect of the liability guaranteed in money terms would be government bonds
of equivalent duration. The purely discretionary liability outgo will be best invested for in high
returning assets such as equities.
In other words, invest an amount equal to the (realistic) policy reserve in government bonds, and
invest any excess of asset share over that amount in equities.
(ii) Single premium without-profits inflation-linked annuity
If available, invest in an inflation-linked government bond, with term reflecting the life expectancy
of the annuitant(s). If such a bond is not available then the company could invest in fixed-interest
bonds and some inflation-resilient investment such as equities. The investments would be
weighted so as to simulate an inflation-linked bond. For instance if we anticipate the increase in
equity earnings to be equal to inflation, the mix would be 0% fixed-interest / 100% equity; and as
we anticipate equity earnings to outperform inflation, so the proportion of fixed-interest
investment will grow.
(iii) Unit-linked regular premium savings contract, minimum SA = $100,000
Invest an amount equal to the value of the unit-related liability in the underlying assets. Invest
the non-unit part in inflation-linked bonds (for expenses). In theory, invest the amount
corresponding to mortality outgo in short-term fixed-interest bonds – in practice, this might be
invested in cash for convenience.
3.3 Free assets

The existence of free assets in a life insurance company means that it can depart from the
matching strategies outlined above so as to improve the overall return on its assets and
thereby benefit its policyholders, through higher bonuses or lower premium rates, and its
shareholders (if any), through higher dividends.
It is almost always the case that assets with the highest expected return also have the
highest variance of that return.

If the assets supporting guaranteed benefits were invested in the assets with the highest
expected return, the probability that the asset proceeds will become inadequate may be too
high for comfort, ie the risk of insolvency is too great.
This assumes that only enough assets are held so that their expected proceeds will cover
the benefits. If there are free assets, they can be used as a cushion to reduce the
probability of becoming insolvent.
Example
Suppose Abingdon Life Company has guaranteed fixed liabilities worth £100 now and it needs to
choose some appropriate investment strategy. In order to increase returns (moving away from a
matching strategy) it would like to invest 50% in equities and the remaining 50% in government
bonds.
To simplify the situation even further, suppose the value of equities might fluctuate by +/- 10% in
the course of any year, and that accounting regulations allow government bonds to be valued at
book value if they are to be held to maturity (ie we can ignore downward market fluctuations on
government bonds). There is no minimum capital requirement.
What investment strategy would be possible if Abingdon Life possessed no free assets? If it were
to invest in £50 of equities, as it would like to, and £50 of bonds, then it would be likely that at
some point those equities fall in value to £45 and the total asset portfolio would then be worth
£95, making the company insolvent. In fact, with no free assets, the company cannot make any
equity investment.
What would be possible if the company had £7.50 of free assets? These free assets would enable
the company to withstand a drop in the value of the policyholders’ assets of £7.50 (if the free
assets themselves do not drop in value – otherwise we would need to allow for that movement).
Thus the company could invest in up to £75 of equities. So the company can invest the
policyholders’ fund 50% in equities as originally hoped for.
The techniques discussed in Section 4.1 below can be used to determine how much of the
free assets are needed as a cushion so as to reduce the probability of insolvency to an
acceptable level.
Clearly this use of the free assets is most appropriate in regard to the assets backing the
guaranteed benefits.
So, the greater the free assets the greater the freedom in choosing investments for guaranteed
liabilities, ie the greater the variability of return which the company can accept on the backing
assets.
It could be argued that the free assets are irrelevant where the discretionary benefits are
concerned, since a company will want to invest in the securities with the highest expected
return anyway. On the other hand, there are actuaries who would argue that, although the
benefits are discretionary, policyholders will expect to receive them and moreover will have
expectations of their level. If this argument is accepted, the company will want to make use
of some of the free assets to ensure that the probability of the discretionary benefits falling
below a particular level stays within acceptable limits.

In other words, for discretionary liabilities, the free assets can support a riskier asset profile than
the company would otherwise be prepared to implement.
Also, it could be argued that it is a reasonable use of the free assets to mismatch
investment-linked benefits if the company can expect to achieve a higher return. Note
though that if this is done, any return achieved above (or below) that on the ‘matched’
assets will not accrue to the unit-linked policyholders but to the owners of the business.
Question
In practice, mismatching of assets and liabilities for investment-linked (eg unit-linked) benefits is
uncommon.
Suggest possible reasons why this is the case.
Solution
The main reason why such mis-matching is not common is because the risk of a sizeable loss is too
great, even if such a loss can be absorbed by the free assets (and hence not threaten solvency).
The company will need to hold significant additional reserves to cover the risk from mis-matching.
This will increase the amount of capital required to write the business, increasing the cost of
capital, and this might ultimately increase the cost of the product for the policyholder through
higher charges.
Another reason is that it might not be permitted by regulation.
A final point to consider in this section is the appropriate matching strategy for the free assets
themselves. In effect they do back liabilities: as they form part of the capital of the company,
there is an obligation to provide a return to the providers of that capital. While this would
suggest investment in high-yielding assets, part of the reason for having free assets is to protect
the company from adverse experience and so at least some of the assets chosen must not be too
risky.

4 Developing an appropriate investment strategy

We now consider how a life company would use asset and liability modelling techniques to assess
the viability of any investment strategy. The logical starting point is to identify the strategy that
forms the most appropriate match for the liabilities. The following summarises the approaches
we have discussed in the previous section:
(1) Split liabilities into the categories of guaranteed in money terms, guaranteed with
reference to an index, discretionary and investment-linked.
(2) Start with investment-linked liabilities: match these exactly – ie invest in the assets
underlying the benefit determination formula.
(3) If possible, match liabilities guaranteed with reference to an index – if not possible, then
choose the nearest thing.
(4) Match the liabilities guaranteed in monetary terms with government bonds (and possibly
some corporate bonds) of suitable term.
(5) This leaves discretionary liabilities. The ideal will be to invest in equities and property for
these. This would normally be heavily weighted towards equity because of the difficulties
associated with property (expensive to deal, to administer, difficult to value, comes in
large chunks). The shares would be low to medium risk, typically major ‘blue-chip’
companies. The same effect could be obtained by index tracking (assuming that we are
not tracking the Venezuelan Tin Futures 100 Index).
(6) Having taken care of all liabilities, how to invest free assets? Here again, they would
normally be invested in equities and property. The difference compared with (5) above is
that the company can adopt a slightly riskier stance in its equity selection, and it could
invest overseas without necessarily having to hedge the currency risk. It is also common
to have significant property investment in the shape of the company’s own premises.
(7) Modify the above strategy to include sufficient cash for the company to operate on a daily
basis without needing to realise any non-cash assets.
The company now has a reasonable investment strategy to assume as a base, representing a
practical ‘minimum risk’ strategy. We now need to consider what divergences from the base will
lead to greatest expected rewards to both shareholders and policyholders, as appropriate, while
maintaining sufficient risk control commensurate with the company’s risk appetite and resources
(particularly its available capital).
4.1 Using a model office to determine an investment strategy

Using a model of the business in force, a model investment portfolio can be built up based
on the company’s proposed investment strategy and incorporating an appropriate
proportion of the free assets.
Question
Explain why the model would incorporate just a proportion of the free assets

Solution
A life company may want to dedicate only a portion of free assets to support the investment
strategy of the policyholders’ fund. This is because the shareholders may not want to use all of
their assets as ‘security’ for that fund: deliberately mis-matching the policyholders’ fund to
increase return, and using free assets as a cushion, means that those free assets might disappear
to pay for any loss arising from the mis-match. Quite reasonably, the shareholders may not think
that increasing the return on assets for the benefit of policyholders justifies the risk of all of their
free assets disappearing.
Another more prosaic reason is that some of the free assets might be inadmissible under local
statute as assets to cover reserves (or any solvency capital requirement).
What is appropriate here will be determined to some extent from the results of the
investigations below.
The liabilities and the assets would then be projected forward on assumptions that
represent expected future experience, although the company will also want to consider the
effect of variations from these.
In fact one of the most important considerations will be how the proposed investment strategy
fares in the light of significant market movements downwards.
For the assets, stochastic investment models can be incorporated to project future
investment income and changes in capital values. Inflation rate models can also be used to
project future expenses on the liabilities side.
As always in modelling work, great care will be required that the parameters are consistent – for
the assets we will need consistency between inflation, interest rates and market (capital) values,
and these should be consistent with the valuation interest rates used in the liability model. (By
‘consistent’ here we mean that the relationships between the different variables should be
suitably realistic.)
The projected liabilities and assets can then be valued at the end of each year of the
projection on the company’s supervisory basis.
This implies that the company wants to test solvency at the end of each year. With modern
computational power there should be no reason not to test this solvency continuously throughout
the projection period, to ensure that the company is always solvent.
The item of interest will be the excess of the value of the assets over the value of the
liabilities. This will need to be sufficient to cover comfortably the level of solvency capital
required by the company, which may itself depend on the investment strategy being
investigated. What is ‘comfortable’ will depend on any regulatory requirements, the nature
of the business, and the level of cover provided in other companies.
An example of ‘comfortable’ cover could be having solvency capital equal to twice the required
minimum; but exactly how much cover is ‘comfy’ will depend very much on whether the
regulatory authorities, perhaps ‘unofficially’, expect companies to achieve a certain level of cover,
the riskiness of the business in question and what is normal in the market.

What is normal in the market will be a consideration because brokers may examine the amount of
cover which insurers show in their published returns, perhaps via articles in the specialist financial
press, and use it as their main measure of a life company’s financial strength.
Using a stochastic investment model and simulation techniques, the above can be extended
to produce a statistical distribution of the amounts available each year to cover the level of
solvency capital required. From this, the probability of potential future insolvency can be
estimated for any given investment strategy.
The simulations might involve several hundreds or thousands of randomly generated ‘sample
paths’, each one representing the future solvency positions resulting from one set of randomly-
generated investment returns.
For example, a collection of sample paths might look those shown in Figure 1. The ratio of assets
to liabilities at the date of projection is 1.2.
Fig 1: Example sample paths
1.4
1.3
Assets / liabilities
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time (in years)
As can be seen, one of the paths clearly goes insolvent (assets less than liabilities). Others may
denote insolvency, depending on the extent of any required amount of minimum capital
requirement under the supervisory regime in question.

Another possible set of sample paths is shown in Figure 2.
Fig 2: More example sample paths
1.4
1.3
Assets / liabilities
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time (in years)
In Figure 2 all of the sample paths go insolvent at some point – which would be more worrying for
the company.
So, the projection of future solvency is an important component of this model. The
assessment of the supervisory solvency position, including dynamic solvency testing, was
also described in Section 3 of Chapter 15 and in Chapters 19 and 20.
For valuing the liabilities, the company’s current basis will most likely be chosen, but it
could incorporate dynamic assumptions into the simulation exercise which take into
account the simulated investment conditions.
If the valuation basis is in any way contingent on the assets underlying the reserves, the return on
those assets, or even just the expected reinvestment yields in the market, then the reserves in the
model for year y should be calculated on a basis which reflects the simulated asset situation in
year y. Keeping the valuation basis constant could give very misleading results. In essence, we
are back to the fundamental concept of consistency.
Finally, and very importantly, we need to identify some measure of success (ie of profitability)
which we can use to compare investment strategies. This is much more straightforward to
measure for a proprietary company (see below) than for a mutual. In the case of a mutual it is the
with-profits policyholders’ returns which matter most, as these policyholders are the providers of
capital. Provided, however, that other aspects of the company’s strategy or experience do not
change over time (for example, its with-profits premium rates or its new business volumes), then
some aggregate measure of the company’s distributed profit over a future time horizon should
give an objective measure of profitability for this purpose.

In general, however, the method of increasing the insurer’s expected profitability is the opposite
of the process of reducing its risk: by deliberately mismatching its assets and liabilities to take
advantage of opportunities that are expected to give rise to increased profit. There is an
important link here with the liabilities: choosing assets which produce the highest expected
returns does not necessarily produce the highest expected profit, because profit is a function of
how the liabilities, as well as the assets, behave. Hence a more rewarding strategy is, by
definition, a more risky one (because it implies less matching). This is precisely why we need to
look at the overall profit emerging from an asset-liability model as a basis for determining an
optimal investment strategy.
Summarising the above approach:

 allocate a certain amount of free assets to support the assets underlying the reserves
 perform asset / liability model projections of the company’s future assets (arising from
the starting assets defined above), and compare these total assets against the reserves
 check that the excess of these assets over the liabilities exceeds any minimum capital
requirement (or possibly, exceeds some comfy multiple of that minimum) for the entire
projection period for most (eg 99%) of the model runs
 calculate some measure of aggregate profitability
 repeat these last three steps assuming different (more or less risky) investment strategies
until the target probability of insolvency is achieved
 identify which of the possible strategies, having equal insolvency risk, produces the
highest profitability.
Question
Describe three potential weaknesses in this particular approach.
Solution
First and most obviously, there is the issue of modelling risk. The conclusions of the asset /
liability modelling process will only be valid to the extent that the model, and the inherent
parameterisation, are valid.
Secondly, there is the issue of considering only a portion of free assets. Although the owners of
the free assets might justifiably not want all of those assets to be put at risk in supporting an
unmatched investment strategy, that does not imply that the model should necessarily exclude
certain assets. For instance, we can still have a constraint along the lines of ‘we accept that 50%
of free assets will disappear with a probability of 1%’ but model the entire asset portfolio, and
interpret the results accordingly. This could give more useful results.
Thirdly, we will need to be very careful if the minimum capital requirement is not constant. In
fact it is commonly defined in many countries as a function of, amongst other things, the reserves.
In this case an approach which is less vulnerable to error would be to include the minimum capital
requirement in the asset / liability model to ensure that it is modelled correctly.

In other words, a better approach might be for the model to be a model of the whole life
insurance company but possibly with some free assets earmarked ‘not for solvency
demonstration’.
What are the decision variables here? In essence there are three elements which can be varied:
 the riskiness of the investment strategy
 the level of the free assets
 the probability of insolvency.
The basic approach that has been described above is to take the level of free assets, decide on
some acceptable risk of insolvency (eg 1%), and then determine the most risky (and hence
rewarding) investment strategy compatible with those criteria.
A minor variant of this is to take only a proportion of the free assets and determine the most
rewarding investment strategy feasible.
Question
Suggest ways in which an investment strategy could be made riskier.
Solution
For assets backing some given liabilities:

 reduce the amount of matching by nature
 reduce the amount of matching by term
 reduce the amount of matching by currency.
Looking at assets in isolation (ie assuming there are no associated liabilities):

 increase duration.
In both cases:
 move from government to corporate bonds
 move from AAA-rated corporate bonds down to something lower-rated
 for equity stock, move from an established company to a new company …
 … or from an established sector to a new sector.
A life company would never normally increase risk by reducing diversification. This is because we
are interested in varying sector risk, not specific risk (which should always be minimised).
A company would never increase risk unless it also leads to increased expected profitability.

The simulations could also be used to determine the level of free assets that the company
needs in order to support a particular investment strategy and keep the probability of
insolvency below an acceptably low figure.
An alternative, which corresponds to what a company might do in reviewing its current strategy,
is to take that investment strategy in conjunction with its actual free assets and to estimate the
probability of insolvency. This is the essence of ‘dynamic solvency testing’, which we met earlier
in the course.
For a proprietary company, the process can be extended further and the effect of the
investment strategy on future shareholder earnings can be considered. In particular, an
investment strategy can be determined that maximises shareholder income whilst keeping
the risk of insolvency sufficiently low, bearing in mind the available level of free assets.
In the case of a proprietary company we can measure profitability directly in terms of the value of
the distributions of profits made to the shareholders (ie as transfers from the projected profit and
loss accounts). This can therefore be used as a means of comparing the merits of different
possible investment strategies in conjunction with the appropriate risk measure.


Chapter 28 Summary
The investment strategy of a life company is of paramount importance to its commercial
success, financial performance and security.
The company should invest to maximise the overall return on the assets, subject to an
acceptable level of risk.
Assets should be selected to match the nature, term and currency of the liabilities, where by
term we mean discounted mean term.
The nature of the liability outgo can be considered as one of the following:
 liability guaranteed in money terms
 liability guaranteed in terms of some non-investment index
 discretionary liability
 investment-linked liability.
Assets would normally be selected to match these four components as follows:
Liability guaranteed in money terms

Immunisation theory can be used to build up a suitable portfolio of government bonds, but
there are a number of problems and limitations with this approach. Normally an appropriate
balance between risk and return would be chosen after dynamic asset / liability modelling, in
order to allow for free assets. For with-profits contracts, the amount of the discretionary
liabilities may influence the asset choice, permitting a riskier stance.
Liability guaranteed in terms of some non-investment index

The ideal is to invest in some asset based on the same index; failing that, use the most
similar asset available.
Investment-linked liability
It is normal to invest in the assets underlying the investment link concerned.
Discretionary liabilities
The ideal will be to invest in high return / high risk assets. However this will be subject to
not exceeding the degree of risk acceptable to policyholders, and not exceeding the degree
of risk allowable given the extent of free assets available.

Overall strategy
In deciding on the investment strategy, the company will need to consider any statutory or
regulatory constraints.
The overall principle is to invest away from a matched position, no more than the maximum
extent possible consistent with the desired level of risk, in order to maximise returns for
shareholders and with-profits policyholders, as appropriate.
The company will perform asset / liability modelling, to test the proposed investment
strategies, which could be used alternatively to investigate:
 the level of riskiness of investment strategy that can be supported
 the level of free assets required to support any strategy
 the resulting probability of insolvency.
There is interdependence between these three aspects.


28.1 A small proprietary life insurance company with limited free assets writes both conventional
Exam style
with-profits and unit-linked without-profits business. Surplus arising on the with-profits business
is distributed using the additions to benefits method. A non-executive director with an
investment background has queried the high percentage of fixed-interest investments backing the
with-profits liabilities and has suggested investing more in equities.
Explain why this may not be possible. [7]
28.2 A life insurance company sells immediate annuities to impaired lives with low life expectancy.
Exam style
The annuities increase in line with an index of prices. The company believes that the best
matching assets are index-linked bonds with a range of terms up to six years.
(i) Outline the factors that affect whether the insurer would mismatch its assets and
liabilities. [7]
(ii) Describe what is meant by an ‘illiquidity premium’. [2]
(iii) Discuss whether it is appropriate to allow for an illiquidity premium in the discount rate to
value the annuity liabilities. [5]
[Total 14]
28.3 A life insurance company transacts all principal types of life insurance and annuity business, and is
Exam style
reviewing the investments that it holds.
(i) Describe the main features of the liabilities of the company that might affect the
investment strategy. [7]
(ii) Discuss the advantages and disadvantages of matching the without-profits annuity
portfolio with fixed-interest securities and the extent to which this is possible. [7]
[Total 14]
28.4 A mutual life insurance company transacts only with-profits business.

Exam style
(i) Describe how a stochastic investment model would be used to determine an appropriate
asset allocation policy, indicating how the projection of liabilities would depend on the
modelled investment conditions. [10]
(ii) State the limitations of a stochastic modelling approach for this purpose. [4]
(iii) Outline the other investigations that should be carried out, and other factors that would
be taken into account, when deciding on an appropriate asset allocation. [4]
[Total 18]

28.5 Discuss the suitability of the suggested assets to match the following liabilities for a life insurance
Exam style
company with substantial free assets:
(a) Liabilities: a block of immediate and deferred annuities

Assets: a portfolio of medium-dated and long-dated fixed-interest securities
(b) Liabilities: a block of with-profits endowment assurances

Assets: a portfolio of equity shares
(c) Liabilities: the future expenses arising from a block of single premium unit-linked
contracts
Assets: the future management charges under the contracts, at a fixed rate of 1%
per annum of the linked funds (no other charges may be levied).
[9]
28.6 A mutual life insurance company writes all types of with-profits life and pension business. The
Exam style
surplus distribution method used for the with-profits business is the contribution method, with a
small final dividend.
Describe the factors influencing the choice of assets comprising the fund dedicated to the
company’s with-profits contracts. [14]

28.1 Given complete investment freedom, there would certainly be a case for greater investment in
equities. Historically they have provided better returns than government bonds and may be
expected to do so in the future. [1]
Equities also provide ‘real’ returns, ie in excess of inflation, which is what policyholders are likely
to want. [½]
However, it is also necessary to consider the guarantees given to policyholders. Initial sums
assured and already declared bonuses are guaranteed in monetary terms, and so are best
‘matched’ by assets which produce certain amounts of money in nominal terms. Fixed interest is
therefore appropriate here. [1]
Equities give an uncertain return, and if the return is low then guarantees to policyholders might
not be met. This is an unacceptable risk. [1]
In addition, as the free assets are limited, the company will have very little ability to smooth
bonuses over time, so that any years in which returns are very low (or even negative) pose a real
threat to statutory solvency. This means that it cannot invest significant amounts in volatile
assets such as equities. [1½]
If the company had substantial free assets then the risk of insolvency would be greatly reduced as
the free assets would be a ‘cushion’ against adverse experience. However, the company does not
have this and so must match guaranteed benefits closely. [1]
In fact the company may have done extensive modelling work, projecting forward its assets and
liabilities assuming different investment strategies and appropriate volatility for the different
asset types, and it may have found that increasing its equity holdings would give a real possibility
of insolvency in the medium-term future. [1]
The company must consider policyholders’ reasonable expectations. It cannot move to a ‘high
return but high risk’ position when the policyholders took out contracts on the basis of a ‘medium
risk’ promise. [1]
Also any change might affect the amount and timing of shareholders’ dividends, and this might
not be acceptable to shareholders. [½]
[Maximum 7]

28.2 (i) Factors that affect the extent of mismatching
The insurer will be forced to mismatch its liabilities: [½]

 as the term of the annuities is not known with certainty [½]
 if index-linked bonds of the right duration do not exist / are not available [½]
 if the index-linked bonds available are not linked to the same price index [½]
 if there is a lag in the index used for the bonds [½]
 if regulation specifies:
– certain assets that the insurer is required to hold [½]
– certain bonds that the insurer is not permitted to hold. [½]
In addition, the insurer may choose to mismatch its liabilities if: [½]
 it has a sufficient level of free assets [½]
 regulation does not prevent mismatching, eg: [½]
– by currency [½]
– by imposing onerous requirements to hold mismatching reserves [½]
 this does not lead to excessive solvency capital requirements [½]
 the proposed alternative assets:
– are available [½]
– provide a sufficiently high expected return to justify the extra risk taken [½]
– are sufficiently liquid [½]
– do not need a level of expertise above that of the insurer [½]
– do not have prohibitively high dealing expenses or high taxes [½]
 its attitude to risk permits such a strategy, which may depend on: [½]
– the risk management tools it has in place, eg reinsurance [½]
– its need for diversification. [½]
[Maximum 7]
(ii) Illiquidity premium
Corporate bonds typically have a higher yield than risk-free (eg government) bonds, … [½]
… where this reflects both the greater default risk … [½]
… and the relative illiquidity of such assets. [½]
The latter of these contributes the illiquidity premium to the yield. [½]
It is effectively compensation for the fact that it is normally more difficult to sell a corporate bond
for a stable and acceptable price than a risk-free bond. [1]

An illiquidity premium can sometimes be included in the discount rate used to value the liabilities
to take credit for the illiquidity premium in the yield on the assets held. [½]
[Maximum 2]
(iii) The appropriateness of allowing for an illiquidity premium in the discount rate
Allowing for an illiquidity premium increases the discount rate. [½]
A higher discount rate will lead to a lower value of the liabilities, … [½]
… and so a better solvency position. [½]
However, the insurer would only be able to include an illiquidity premium to the extent that it
holds corporate bonds to back its liabilities. [½]
If index-linked corporate bonds do not exist or are not available, then it would not be possible to
include an illiquidity premium. [½]
Equally, if the insurer deems that the default risk of holding corporate bonds is too great, then an
illiquidity premium will not be an option. [½]
Assuming that corporate bonds are held, it is only generally appropriate to include an illiquidity
premium for long-term, predictable liabilities … [½]
… for which matching assets can be held to maturity. [½]
Provided the insurer has a large enough portfolio of liabilities so that the term is broadly known,
… [½]
… then the liabilities may be predictable enough to justify the inclusion of an illiquidity premium
... [½]
… as the insurer can hold the bonds to maturity so will not be exposed to changes in their spread.
[½]
However, the market for impaired life annuities for people with very low life expectancy may not
be large enough to obtain a sufficiently large portfolio to remove the impact of random
fluctuations. [½]
The term of the liabilities (up to six years) is also probably not long enough to justify the use of the
approach, ... [½]
… because the choice of the discount rate has little impact when discounting over a short period. [½]
Overall, whether an illiquidity premium can be included is likely to be dictated by the regulatory
regime in place, … [½]
… which would also specify how and when it can be used. [½]
[Maximum 5]

28.3 (i) Main features of liabilities that might affect investment strategy
With-profits liabilities: if future profit distributions are discretionary, this will permit investment in
real assets, like equities and property. [1]
This is especially the case if the distribution method defers profit distributions to any degree (like
the additions to benefits method). [½]
Policyholders’ expectations (eg of smoothing) may restrict investment freedom to an extent, … [½]
… and so lead to some fixed-interest securities even for discretionary benefits. [½]
Guaranteed monetary benefits (including any declared attaching bonus benefits), payable on
maturity, death and surrender (if applicable), will require fixed-interest securities. [1]
Annuities in payment will be closely matched by fixed-interest or index-linked bonds, as

appropriate to the nature of the annuity benefit (fixed monetary payment or index-linked). [1]
Assets backing unit-linked liabilities will be invested according to the unit funds chosen by the
policyholders. [½]
Any liabilities denominated by foreign currencies are likely to be matched by corresponding

foreign currency assets. [½]
Other liabilities that behave similarly to an inflation index (such as expenses) will be matched by
index-linked assets … [½]
… and by assets that are expected to provide real returns (such as equities). [½]
Any shareholders’ liabilities (ie their retained funds) should be invested in real assets … [½]
… in order to maximise returns. [½]
The term of the liabilities will affect the term of the assets chosen, … [½]
… where this is relevant, eg when matching with fixed-interest or index-linked bonds. [½]
The state of the fund – whether it is expanding or contracting – will affect liquidity requirements,
… [½]
… and hence the extent of the need to hold cash or other highly liquid assets. [½]
Different taxation treatment of different assets may affect what’s chosen, … [½]
… and the tax treatment may vary depending upon which liabilities the assets are backing. [½]
Large free assets will mean more investment freedom to move away from a matched position in
pursuit of higher returns. [1]
[Maximum 7]

(ii) Matching annuities with fixed-interest securities
Advantages
Complete matching, if it were possible, would prevent loss occurring as a result of future changes
in investment conditions. [1]
Risk of default may be lower than with other asset types. [½]
New business capital requirements are likely to be minimised, as there would be no need to hold
a mismatching reserve and the reserving basis would require least margins in the investment
return parameter. [1]
Allows the company to price the contracts with only a small margin in the investment return
parameter, which could make the product more competitive. [1]
The reduced capital requirement should also make the product more competitive by reducing the
cost of capital. [½]
Disadvantages
The company would not be able to profit from future changes in investment conditions. [½]
A large number of individual holdings may be required, increasing costs. [½]
Fixed-interest securities do not provide an appropriate match for future expenses, nor for the
annuity liability if the benefit is index-linked. [1]
Long-term returns are expected to be lower than for equities and property. [½]
Reasons why absolute matching may not be possible
Fixed-interest assets are unlikely to be available at the exact durations required to produce
complete matching. [½]
This can be got round by immunisation, but this introduces its own problems (eg having to change
investments continually), and even then assets of sufficiently long duration may not be
available. [1]
Fixed-interest securities may have a spread of redemption dates, so cannot be used to match with
certainty. [½]
The duration of the liabilities is not known with certainty either, because of uncertainty in the
mortality experience. [1]
Non-government fixed-interest securities will have a default risk. [½]

[Maximum 7]

28.4 (i) Stochastic model for asset allocation
A stochastic model could be used as part of an asset-liability modelling exercise. This involves the
projection of the company’s assets and liabilities, the investment return being determined as a
random variable. [1]
The liabilities would be projected using representative policies or model points. These would
represent the in-force business and possibly also new business. For example, one point may be
used to represent all ten-year endowments of duration six years. [1½]
The projection would require realistic estimates of basis elements such as mortality, expenses and
lapses. This is likely to be deterministic, although stochastic projection of some variables is
possible. Sensitivity testing of key variables might be employed. [1½]
Projection of asset values would be carried out stochastically for different trial mixes of assets.
Statistical models and appropriate parameters would have to be chosen for the main asset
classes. For example, the company might project bond gross redemption yields, equity dividends
and dividend yields. Economic variables such as inflation might also be projected. [2]
Random error terms would be used to give stochastic returns, although this randomness might be
constrained. For example, the dividend yield in one time period might be related to the yield in
the previous period and/or a fixed value, plus a random term. This prevents the dividend yield
fluctuating wildly or drifting too far from a ‘normal’ level. [1]
Results are sensitive to the model and parameters chosen, and so these must be chosen with
care. [½]
The projections would be performed using simulations, eg 10,000 for each trial asset mix. One
could then look at various quantities, eg proportion of runs ending in supervisory insolvency,
average payouts. [1]
To decide which asset distribution is ‘best’, clear criteria must be specified, for example highest
mean free asset ratio, lowest probability of insolvency, highest bonuses. [1]
The model could allow the projection and valuation of the liabilities to ‘react’ to the modelled
investment conditions, and vice versa. [1]
For example, the liabilities of a UK with-profits company depend on bonuses, which depend on
investment returns. Bonuses could be made to respond to the modelled investment conditions.
The company would seek to use an algorithm which was consistent with its bonus philosophy, for
example in the mix of regular and terminal bonus, and on smoothing. [1½]
The supervisory valuation basis could also be adjusted to be appropriate to the projected
investment conditions. [½]
The volume of new business would also depend on investment conditions. Poorer returns may be
associated with lower future sales. A weak supervisory position could have the same effect,
because of limited capital and brokers’ concerns. In such a situation the model might
automatically switch more assets into fixed interest government stock. [1½]
[Maximum 10]

(ii) Limitations
Problems are:
 Results are highly dependent on
– choice of model
– choice of parameters. [1]
 In particular, often interested in ‘extreme values’ and these are unreliable. [½]
 Difficulties in modelling some asset classes, eg term structure of bond yields. [1]
 Computer power and time required are considerable, although this is less and less of a
problem. [1]
 To keep the model manageable, simplifications are needed, eg relatively few model
points. [1]
 The intended timescale for some models is much longer than that for which they are
often required and used. [½]
 Many available models are not good at modelling possibility of a ‘crash’. [½]
[Maximum 4]
(iii) Other investigations
Investigations might include:

 initial investigation into mix of guaranteed and non-guaranteed liabilities, to give some
guidance on mixes of assets to consider [1]
 policyholders’ reasonable expectations (PRE) as established by past practice and any
communications on investment policy, eg literature or mailings [1]
 competitors’ asset mixes [½]
 experience investigations for determining projection assumptions. [1]
Other factors to consider:

 possible future changes which might alter investment policy eg taxation changes,
increased future use of derivatives [1]
 supervisory restrictions, method of statutory valuation of assets and liabilities. [1]
[Maximum 4]

28.5 (a) Annuities
Provided that the duration and currency of the assets is the same as that of the liabilities, the
stocks will be a good match (assuming that the annuities are without-profits). [1]
However:
 There will be an investment risk on the deferred annuities, on the assumption that the
fixed-interest stock is paying an income, and no outgo is paid to the policyholder. [1]
 The duration of the liabilities is long, so stocks of an appropriate duration might not be
available. [½]
 The nature of the outgo to policyholders might not be fixed. The fixed income might not
be appropriate if the annuity in payment is increasing and the income does not match the
increases. [1]
(b) Endowments
The assets will provide an income that is fairly stable and that increases in real terms, but the
market value will be volatile. [1]
The company has significant free assets, so the company can afford to invest in volatile assets and
to smooth payouts. [½]
If the system used is the UK ‘additions to benefits’ approach, the income will provide real gains,
which will protect the investment in real terms and so should be in line with policyholders’
expectations. The income is also well matched for providing regular bonus. [1]
Under other systems of declaring bonus (revalorisation and contribution methods) these are
probably not appropriate, due to PRE and the high levels of guarantees. [1]
The suitability of volatile assets depends upon the amount (if any) of terminal bonus declared.
The higher the terminal bonus, the more appropriate the assets will be. [½]
(c) Expenses from linked contracts
To match expenses, which are similar to an inflation-linked liability, the asset proceeds need to
produce a stable real income. [½]
This depends on the underlying assets. A reasonable long-term match would be achieved if the
underlying assets were equities, property or suitably linked index-linked bonds. A good long-term
match would not be achieved if other types of assets were involved. [1]
To consider short-term matching:

 Cash will be a much better match to short-term inflation than long-dated stocks. [½]
 The income is related directly to the market value of the assets, so a short-term fall in the
value of the assets would have a big impact upon the level of the charges being taken
from the linked fund. This means that equities and property may provide a poor short
term match for these liabilities, as their short-term volatility of returns is much greater
than the short term volatility of inflation. [1½]

 The company’s expenses might not increase in line with the growth in the economy in
which the company is operating. [½]
Hence the income is not necessarily a good match for the outgo in the short term. [½]
However the company does have significant free assets, so the company may well be comfortable
with the short-term mismatch provided a better match is achieved over the longer term. [1]
[Maximum 9]
28.6 The main influences are:
Nature, term and currency of liabilities
The liabilities of the company can be considered as either guaranteed in money terms,
discretionary, inflation-linked or investment-linked. [1]
Some liabilities, eg guaranteed sums assured and any bonuses added to benefits to date, will be
guaranteed in money terms, so these would be best matched with fixed-interest stock. [1½]
The fixed-interest stock chosen should be of the same discounted mean term as the liabilities
concerned. [½]
As the pensions liabilities may be very long term, it could be hard to find an exact match. [½]
There may also be discretionary liabilities on this with-profits business, in respect of future annual
dividends and any eventual final dividend payments. These are best matched by assets offering
real growth, such as equities and property. [1½]
Dealing costs are much higher for some assets (eg property) than others (eg government bonds)
and for property it is only possible to deal in very large units. [1]
Also property is much less marketable than government bonds and equities, causing difficulties if
required to realise the assets quickly. [½]
The future administration expenses of the company are in effect inflation-linked. So there will be
a certain amount of inflation-linked liabilities, which are best matched by any bonds offering such
a link. [1]
If no such bonds exist, need to simulate them by a combination of fixed interest and equities. [½]
The investments should be in the currency of the denomination of the liabilities. [½]
Risk and returns
It is vital that, given its liability portfolio, the company stays solvent. [½]
The accumulation of deferred profit for the small final dividend should reduce the insurer’s
insolvency risk, giving it the ability to invest more freely. [1]
Subject to offering security to the desired extent, the aim of the asset portfolio is to maximise
returns. [½]

Property and equities have given higher returns historically and probably will in the future.
However this is in the long term; in the short term government bonds may outperform both. [1]
The company should consider its short-term views on different asset sectors. [½]
The company should hold a diverse portfolio of assets to minimise specific risk. [½]
The company should have a similar mix of assets to its competitors or there is a risk of falling sales
if its performance is too much out of line with the market. [1]
The asset portfolio should attempt to match the expected cashflow from the portfolio of
liabilities. [½]
This may be impossible to achieve, in which case the company should match as closely as possible
to avoid exposure to reinvestment risk (unless there is scope for mismatching – see below). [½]
Free assets
The greater the size of the free assets, the bigger the cushion that the company has against
adverse experience. Therefore the higher the level of free assets, the more risky the investment
policy that the company can afford to take. [1]
The more risky strategy will generally mean holding a greater proportion of real assets, and/or
choosing fixed-interest assets that do not match the guaranteed liabilities by term. [1]
A stochastic asset / liability model can help the company to choose the risk v return stance most
appropriate for its level of free assets and acceptable probability of ruin. [1]
It is possible that the market looks at the size of free assets, so sales might fall if the free asset
ratio falls. [½]
The free assets can be useful to smooth bonuses (if the surplus distribution system involves any
subjective judgement on the part of the actuary) and possibly to smooth the effect of claims
volatility. [½]
The regulatory authorities might require demonstration of some minimum capital requirement.
Depending on the exact derivation of this, the choice of assets in respect of this requirement may
need to take account of the nature, term and currency of the underlying liabilities. [1]
Policyholders’ reasonable expectations
Policyholders expect the company to continue to invest in line with what was stated in marketing
literature and from the history of the company. [1]
Their expectations will concern both the security of the company to meet guaranteed benefits,
and in the case of with-profits contracts to obtain the sort of returns indicated in any marketing
literature, projections etc. [1]
A sudden change in assets backing the with-profits policies will go against these expectations. [½]
[Maximum 14]

SP2-29: Further risk management Page 1
Further risk management

Syllabus objectives
 expense control
 policy retention activity
 management of new business mix and volumes
 management of options
 systematic risk assessment and management strategies.

Page 2 SP2-29: Further risk management
0 Introduction
In this chapter we continue the theme of risk management. There is quite a lot of overlap with
earlier chapters, particularly those concerned with risk and product design, so some revision of
these topics may be useful if it is some time since you have studied them. Alternatively, you
might like to treat this as a revision chapter, and use the questions as revision tests, to see how
much you can remember from before.
One of the key additional themes that is developed through this chapter is the role of monitoring,
which is then fully discussed in Chapter 30.

1 Expense control
At a company level, and in the long term, the insurer should seek at least to contain
expenses and commissions within the policy loadings built into the company’s premiums
and charging rates.
Question
Explain the logic behind this point.
Solution
The loadings for expenses and commissions, being received through premiums and charges, are
the actual amounts of income the company receives to pay for these costs. The Core Reading is
saying that, if the actual expenses and commissions are not higher than these receipts in total,
then the company will avoid making an expense loss (relative to its pricing assumptions).
The aim seems laudable (and probably obvious) enough. Achieving the aim is not so simple. The
first thing the insurer should do is to monitor the position closely.
The insurer can monitor the actual levels of expenses being incurred and can compare
these with the policy loadings. Monitoring expense experience is covered further in
Chapter 30.
Comparisons will also be made with expense ratios for competitors where possible. The
insurer will wish to ensure that its level of costs remains competitive.
Question
Explain why it is important for a company to keep its expense ratios in line with those of its
competitors, and which ratio(s) it would actually use for this comparison.
Solution
The aim of doing this is to make sure that the company can remain competitive, which means
charging prices that are at least similar to those of its competitors (assuming identical products).
This is important so that the company will be able to maintain adequate volumes of sales.
The impact of expenses on price comes from the loadings (expense assumptions) included in the
pricing basis, so the company might like to see how big its competitors’ loadings are relative to
the prices they charge, and then how its own compare.
However, information on competitors’ loadings are likely to be hard to find, but actual expense
data – at least for total business – are often publicly available. In any case, the loadings must
ultimately reflect the actual expenses incurred over the long term in order for a company to avoid
insolvency, so a comparison of actual expenses over premiums received is likely to be the best
ratio to use.

Expenses can be controlled by reducing the current cost base of the insurer to be within the
policy loadings.
This is easy to say, but not necessarily easy to do.
Question
Suggest ways in which a company might try to keep its ‘current cost base’ within its policy
loadings.
Solution
The following are some of the possible ways:

 maintain active control of staffing levels to reflect the volume of business (and the
amount of work) being achieved
(This should involve increasing staff numbers when business is prospering, and decreasing
them when volumes are falling. Given this is long-term business, most staff fluctuation
will occur with respect to employees working on new business activities, including sales
and underwriting.)
 impose budgetary constraints or targets within which individual departments are required
to manage
 ensure that the staff used are not overqualified (and overpaid) for the work they are
required to do
 ensure that cost-of-living and grade salary rises are not excessive, but consistent with the
price needed to retain the necessary quality of staff
 attempt to sell greater volumes of (profitable) business, without increasing the cost base
by as great a proportion
(This should be possible as some of the expenses should be fixed – overheads – so only
the marginal expenses would need to be increased. Changes in product design and/or
sales methods might be necessary to achieve this.)
 improve efficiency wherever possible, for example:
– increase automation and/or computerisation of tasks (so requiring less
manpower)
– streamline underwriting procedures (if this can be done without unreasonably
compromising risk)
– start doing some tasks in-house that were previously done by external consultants
– use cheaper distribution channels, such as the internet
– sell simpler products (so that sales and administration costs are lower)
 increase the loadings in the premium so that higher expenses can be covered (but this is
only possible if the competitive position will permit it).

Commissions should be controlled by a regular monitoring procedure at the level of:
 product line
 the specific salesperson or broker concerned.
Commission is an expense that should be highly controllable by the company. However, the
company could find itself under considerable pressure to raise commissions to meet (especially)
brokers’ demands, and/or to retain the necessary quality of sales staff. The point being made
here is that monitoring is necessary at each of the above levels to ensure that commission rates
do not start drifting upwards (which would most likely start to happen at the bottom –
salesperson – level).
For those distribution channels where the insurer incurs the costs of acquiring business –
for example, quotations systems – the number of quotations produced versus the actual
business acquired should also be monitored and managed.
Question
Suggest actions that the company could take if it found that the rate of sales per quotation was
falling steadily.
Solution
(So this would mean each sale was becoming more expensive.)
The company might:

 improve the information provided on the quotation – how does it compare with
competitors’ versions?
 provide additional sales support services, if this would improve the overall sales efficiency
 try a different sales method
 try altering the product design and/or the price of the product, in case these are the cause
of the decline in popularity
 etc.

2 Policy retention activity

With a view to managing the persistency risk, an insurer may seek to minimise the volume
of lapses and surrenders.
We have seen earlier in the course that withdrawals – especially at early policy durations – can
pose significant financial risks for life insurance companies.
2.1 Distribution channel management

As with expense control, one of the key elements of a successful control system is good
monitoring.
The insurer can monitor persistency rates by distribution channel and by the specific
salesperson or broker concerned.
By identifying where any persistency problem occurs, the insurer can focus its efforts on
addressing the cause of the problem where it is likely to achieve greatest improvement.
Subject to not damaging the future sales potential from the channel, the level of
commissions or other remuneration can be designed to encourage better persistency
and/or penalise early lapses and surrenders.
Question
Describe two possible ways in which commission systems can be designed to encourage better
persistency.
Solution
Two possible ways are as follows:

 Pay lower initial commission and higher renewal commission. This will give brokers more
incentive to take an active interest in their existing clients and hence reduce the
proportion who withdraw early.
 Use a commission clawback system, whereby a proportion of the initial commission paid
has to be returned to the insurance company if the policy is discontinued within a set
period (eg three years), the proportion repaid reducing the longer the policy is
maintained.
It may also be possible to identify systematic reasons for the lapses and surrenders and to
invoke suitable management strategies to avoid the trend continuing.
For example, the insurer might be providing some misleading information to the client and/or
intermediary, which results in the policy being inappropriately sold (so targeting the wrong types
of individuals, not fully meeting their needs, etc).

2.2 Customer relationship management

The insurer may seek, via any intermediary or salesperson as appropriate, to encourage
good persistency by ensuring that the product sold is suitable to meet the policyholder’s
needs. In addition, it should aim to maintain or improve the quality of ongoing
administration and contact with the policyholder.
Persistency can also be improved by type of payment method: for example, policies having
premiums paid by direct debit have higher persistency than when cheque or cash payments are
made.

3 Management of new business mix and volumes

The insurer will wish to ensure that it has sufficient capital and administrative resources to
allow the writing of new business.
In extreme cases, this might involve the actuary recommending that the directors withdraw
from sale, for a suitable period, those products that are the most capital intensive or for
which inadequate administrative resources are available.
In the normal course of events, the insurer should monitor the extent of mismatch arising
between the actual new business sold and that assumed in the original pricing of the
product.
Question
Explain why this (last paragraph) is important.
Solution
Directly or indirectly, both the volume of business and the mix of business are assumptions made
when pricing the product. Divergence from the pricing assumptions could therefore lead to
losses.
In the case of volume, an assumption for this would have been implied within the expense
assumption, where overheads would have been spread over assumed future business volumes to
determine the per-policy expense loadings. If actual volumes are not as high as assumed, then
per-policy expense losses will be made, reducing the return on capital.
An assumed volume of new business will also have been used to assess the total expected
contribution of the product to future profit, and whether the company has sufficient capital to
write the business. Lower volume than expected will disappoint profit expectations, and may lead
the company to rethink its marketing plans. Higher volume than expected could be causing
unacceptable capital strain, so future sales may need controlling. The mix of business is
important here too – the capital position will depend on the relative capital efficiency of the
business being sold. So we may need to control the mix of business we should continue to sell,
and/or redesign particular products to make them more capital efficient, in order to prevent
possible insolvency.
The size of policies is usually the other main issue regarding the mix of business. A lower than
assumed average policy size may reduce the absolute amount of loadings received towards the
per-policy expenses, whereas the per-policy expenses themselves (by definition) would not be so
affected.
Control of new business volume and mix largely comes down to marketing, product design and
pricing activities. Sales channels can also be encouraged – through reward systems – to achieve
particular sales targets, though care must be taken not to conflict with other aims (such as
persistency targets) or to be unprofessional.

Typically, it will be necessary to monitor, by product line and distribution channel, the
number of contracts and amount of premium to determine the average case size.
In addition, the following items will also be monitored by product line and distribution
channel:
 frequency of premium
 policy charges / loadings
 actual expenses incurred
 new business valuation strain.
These items are all particularly concerned with managing the company’s capital position. So, in
terms of control, we should be considering the product mix and product design factors that are
relevant to this. We will look at each of these in turn.
Premium frequency
Question
Explain the importance of premium frequency for the management of the company’s capital
position.
Solution
Premium frequency affects the extent of new business strain.
The strain is least for single premium business, as all of the initial expense loadings have been
received on day one of the policy.
Most strain is incurred for monthly premiums, as only one month’s loading for initial expenses has
been received, whereas all (or most) of the initial expenses themselves have been incurred.
This is therefore an example of the general problem described in the next subsection: of a
mismatch between the expenses incurred and the loadings received to pay for them.
Policy charges / loadings and the actual expenses incurred

Comparing the loadings with actual expenses incurred is part of what we should be doing anyway
with regard to the expense management we discussed in Section 1 above. (We would need to be
monitoring the different categories of expenses separately – ie initial (as here), renewal,
termination and investment expenses.)
However, there is an important (additional) issue with regard to the management of capital
concerning the extent to which the loadings and expenses are matched.

Question
Explain how the matching issue affects the company’s capital strain for unit-linked business.
Solution
If the charges under unit-linked policies are closely matched, by incidence and amount, to the
costs that they are designed to cover, then the supervisory reserves required at the start of the
policy will be closer to the initial asset share value. This will reduce (or possibly even avoid) the
new business strain when the policy is sold.
The extent of matching of charges is therefore a key factor in determining the extent of any strain.
We have already described another example of this – the effect of premium frequency – in the
solution to the previous question.
To reduce the mismatch between actual expenses and policy charges / loadings, and to
have lower new business strain, the insurer could:
(a) restrict or encourage certain product lines and/or distribution channels
 directly by structural change
 indirectly by remuneration arrangements, by the level of other support given

to the distribution channel, or the literature used to market to policyholders.
Question
Give examples of structural changes to policy design that should significantly reduce the
mismatch problem.
Solution
The following are possible examples:

 change from non-linked to unit-linked designs
 change from regular to single premiums
 change from guaranteed to reviewable premiums and/or charges.
The last point in the solution above improves the matching position by allowing future charges to
change in line with ongoing experience at each review date; whereas with a guaranteed product
the charges and costs could diverge and continue to do so for as long as a policy is in force.
Question
Explain how a change in distribution channel can help reduce the mismatch between actual
expenses and policy charges.

Solution
If different channels have different commission structures (eg high initial / low renewal vs
not-so-high initial / not-so-low renewal) then the incidence of the outgo will be altered, which
(with appropriate targeting) could reduce the mismatch.
(b) reprice and/or redesign contracts, including the possibility of an increased minimum
premium.
An increased minimum premium will reduce the risk of having inadequate per-policy
expenses loadings due to a lower than expected average policy size.
From this section, key points are that:

 mismatching of income and outgo can be an important factor in determining new
business strain
 suitable product design (choice of charging structure, premium frequency) can improve
cashflow matching and so reduce strain.
New business valuation strain

The key word in this phrase is valuation: so it is to do with the impact of the supervisory reserves
and solvency capital requirement on the company’s capital position.
Question
Explain how new business valuation strain arises.
Solution
Valuation strain arises when a policy is sold because the combination of supervisory reserves and
solvency capital requirements tends to place a higher value on the net liabilities than the pricing
basis. This results in the initial reserves and required solvency capital exceeding initial asset
shares (initial premium minus initial expenses) when the policy is issued, thus causing the strain.
As mentioned earlier in this section, premium frequency will affect the extent of new business
strain. Another key aspect of product design that will significantly affect the extent of valuation
strain for both unit-linked and conventional business is the amount of guarantees given. The
greater the level of guarantee, the more conservative the valuation basis has to be, and the
greater the valuation strain.

4 Management of options
Question
List two major types of options given under life insurance policies, giving one example of each.
Solution
(1) Financial options – eg guaranteed annuity option, where a policyholder has an advance
guarantee of converting a lump sum into an annuity income at a minimum conversion
rate (usually attached to a policy that provides savings for retirement benefits).
(2) Mortality options – eg the policy gives the right to purchase a further policy at a future
date, on then standard rates of premium and without needing to present further evidence
of health at that time.
The insurer will seek to adopt continual monitoring and review of the benefits of options
versus their costs.
Question
Explain what is meant by the ‘benefits’ and the ‘costs’ of options in the previous paragraph, and
why this monitoring and review is important.
Solution
These words actually have two possible interpretations, each of which is valid and useful.
(1) We should look at the benefits provided by the options to the policyholder, to see if they
meet a useful policyholder need. The cost is then the price the policyholder has to pay for
these benefits. So the question is whether the price the policyholder is paying for the
option appears to be worth the benefits provided by the option. This then has
implications for the company, which should be striving to have satisfied customers who
feel they are getting something valuable for a fair price, and this should lead to successful
sales and good persistency of its business.
(2) We should be considering whether it is worth the company’s while in offering the options
(ie what is the ‘benefit’ to the insurance company of providing options under its policies?).
Even if the option benefits meet customer needs at a ‘fair’ price, it may be doing little to
attract or maintain business for the company and, worse, the options may be causing the
company losses or be presenting an unacceptable risk of losses in the future. (So, the
‘cost’ is here referring to how much it is actually costing the insurer to provide the
options, not how much the policyholder is paying for them.)

At an objective level, the insurer should seek to monitor the charges / loadings included for
options in the product pricing with the actual costs being experienced. The analysis should
look separately at the take-up rate for the option and the profit or loss which arises once an
option is exercised.
If these assessments show that the continued availability of the option, even allowing for a
subjective assessment of the increased marketability which the option brings, carries a net
loss for the insurer then the option’s pricing should be increased, and/or its availability
should be reduced or removed altogether.
Alternatively we could change the benefits provided under the option so as to make it more
affordable.
Question
An option to increase benefits under a life assurance policy, without health evidence, has a low
take-up rate. Even so, the company is losing money on this benefit.
Discuss whether this indicates that the option should be dropped from future new business.
Solution
We have to consider whether the availability of the option is doing enough to enhance business
volumes to more than compensate for the loss incurred by option take up. For example, it may
be that policyholders really do appreciate the availability of the option (even if few end up using
it) and as such may be a big factor in ensuring sales. What competitors offer, and what
intermediaries consider important, would also be material factors here. So, it might well be that
sales would suffer significantly if the option was withdrawn from new business, thereby causing
significant loss of profit (and increased impact of overhead costs) from the loss of the normal
premium income. The losses avoided by removing the option could be insignificant compared
with this.
Also we would have to see whether it might be possible to increase charges levied for the option,
so as to reduce the losses, without significantly impairing sales.
Therefore we do not have enough information (about the impact of the option on marketability
and resulting profits) to say whether the option should be dropped or not.
The time lag between removal of an option and the impact on experience emerging must be
borne in mind.
Question
Explain why this time lag occurs and why it is important.

Solution
Options only incur the insurance company in a cost at the time when the option is exercised. So
removing options from new policies, say, will only affect the insurer’s actual option costs at such
future dates as the options would have been exercised, which might be many years in the future.
It is important from the point of view of control. If the future exercising of options that are held
by existing policyholders is expected to lead to insolvency, removing options from all new
business from now on will not prevent the insolvency occurring. (It would still be the smart thing
to do, but it would be necessary to take serious additional action about the existing options to
save the company.)
It is likely that it will only be possible (legally and in terms of treating customers fairly) to
amend the terms of new business written.
Question
Suggest ways in which an insurer could deal with an expectation of serious future losses under
options held by existing policyholders.
Solution
Possibilities are:
 set up appropriate (prudent) reserves for the future option cost (thereby retaining current
surplus and preventing future insolvency)
 match the options by purchasing appropriate financial derivatives (if appropriate).
Either method has disadvantages, but then these may be desperate times. As a further (definitely)
desperate measure, the company may try to buy the options back off the policyholder. However,
this may just worsen the situation, by making everyone aware that the company believes the
options to be valuable, with the result that everyone hangs on to them and an even higher
proportion end up exercising the option as a result.

5 Systematic risk assessment and management strategies

In this section, we set out the core actuarial role in life insurance company management.
As discussed in Chapter 12, a company should be aware of, and assess, the overall risk
profile to which it is exposed, based on an aggregation of the underlying risks that it faces,
allowing for correlation effects.
The role of the actuary in managing a life insurer’s risks should therefore be a regular
programme of advising the directors of the nature and size of the risks faced. Particular
attention will need to be given to those risks that are the most material and/or the most
sensitive to change.
Risks should be controlled by analysing and explaining their nature, costing them as far as
possible, and agreeing on management strategies for them.
In the case of sensitive risks, the actuary should model a range of long-term scenarios to
show the impact of variations to future experience, and management strategies should be
designed to protect against those risks that the insurer is willing and able to control.
A sensitive risk is one to which the company’s financial results are most vulnerable.
Deterministic and stochastic modelling methods may be used as appropriate to model the
nature of the risks concerned.
Once risk management strategies have been agreed, they should be documented and
implemented. Where the strategies consist of objective rules and procedures, these should
then be monitored on an ongoing basis. Strong governance and controls are vital.
The actuary’s role is therefore to advise the decision-making process, by identifying in particular
the risks and rewards of the various courses of action that the company might take. This has to
be an ongoing, continuous activity, evolving with the changing commercial and economic
environment and as the company’s own experience changes and is expected to change in the
future. All aspects of the control cycle, especially the monitoring experience and the updating of
assumptions, are important to ensure that the actuary’s advice remains fully up-to-date.
Monitoring is the subject of Chapter 30.
One approach to managing risks is Enterprise Risk Management. You may have come across this
already in Subject CP1 or Subject SP9.
Enterprise Risk Management (ERM) is a risk management framework which considers the
risks of the business as a whole, rather than considering individual risks in isolation. This
allows the concentration of risk arising from a variety of sources within an organisation to
be appreciated, and for the diversifying effects of risks to be allowed for. This is the
holistic, integrated approach to risk management.
So, by looking at the insurer as a whole, the management will be able to see the overall level of
risk to which they are exposed. The overall level of risk will be lower if the insurer is well
diversified, eg by writing business in many different countries. One benefit of greater
diversification is that the insurer will need to hold less capital against its risks.
Another element of ERM is the recognition that value can be added to a business through
educated risk-taking, with a strong risk management framework that better allows
companies to identify and assess strategic opportunities.

For example, the longevity risk on annuities offsets, to some extent, the mortality risk on term
assurances. So it may be beneficial for an insurer writing lots of annuities to take on extra
mortality risk by selling more term assurances.

Chapter 29 Summary
Expense control
Insurers should try to keep expenses and commissions within the loadings received to cover
them. To achieve this, the company can:
 monitor the position regularly
 monitor competitors’ expense ratios to ensure expenses are at a competitive level
 have monitoring procedures in place to pick up (and prevent) any upward slippage in
commission levels
 control staffing and salary levels to be consistent with the work required (and hence
consistent with business volume), especially for new business / sales effort
 impose budgetary constraints and targets
 attempt to sell more business without increasing the (overhead) cost base
 improve operational efficiency, eg through automation
 increase premium loadings and/or charging rates, provided still competitive.
Persistency
Insurers should aim to minimise the volume of lapses and surrenders.
To achieve this, the company should monitor experience, especially by distribution channel,
to identify problem areas.
To improve persistency, the company might:

 change distribution channel(s)
 set up alternative remuneration (commission) structures that encourage persistency
 improve sales methods so that policies are sold more strictly to meet customer
needs
 restrict premium payment methods (to direct debit, for example).
Managing new business mix and volumes

A company must ensure it has adequate capital and administrative capacity to sell its new
business.
It must monitor volumes and mix compared with pricing assumptions:

 excess volume can indicate inadequate capital
 low volume will reduce total profit and increase the per-policy cost of overhead
expenses
 lower than expected case size may uncover per-policy expenses.

It needs to monitor the mix of business from a capital efficiency point of view, especially:
 extent of mismatch between charges / loadings and expenses
 premium frequency
 valuation strain.
Volume and mix can be controlled by:

 appropriate marketing, especially with regard to remuneration and targeting of
distribution channels
 product design (eg matching charges, premium frequency, minimum premiums,
guarantees).
Management of options
Options are either financial or health related.
It is vital to monitor how much the exercising of existing options is costing, or is expected to
cost, the insurer, compared with the charges received to pay for them.
Items to monitor include take-up rate, and profit / loss arising following take up.
The decision to provide options depends on the extent to which they are appreciated by the
policyholder, and hence contribute to successful sales, and the burden of cost they place on
the company.
Control measures include:

 increase charges / loadings that are paid for the option
 alter the benefits or terms of the option
 remove the option from new business.
Option costs occur after a considerable time lag after sale, and may have the capacity of
being extremely onerous (especially financial options). Possible mitigation for existing
options include:
 appropriate reserving
 using derivatives
 buy back from policyholder.

Systematic risk assessment and management strategies

Actuaries advise the company’s decision-making process.
The key role is to analyse, explain and quantify the risks and rewards of various courses of
action.
It is important to model the potential variations in future experience caused by the most
sensitive risks.
Agreed risk management strategies should be documented, implemented, and their

effectiveness monitored continually.
ERM is an integrated approach to risk management that allows managers to understand

concentrations of risk and the benefits of diversification. It also improves the company’s
ability to identify and assess strategic opportunities in order to add value to the business
through educated risk-taking.



29.1 Outline the risks to a life insurance company’s solvency, and how those risks can be
Exam style
addressed. [15]
29.2 A life insurance company writes whole life assurance contracts on the following bases:
Exam style (a) without-profits non-linked
(b) with-profits non-linked.
Describe the methods the company has available to limit its exposure to each of the principal risks
it faces in issuing these contracts (excluding risks relating to new business). [7]
29.3 A mutual life insurance company is considering introducing guaranteed surrender values for its
Exam style
without-profits endowment assurances.
Discuss the implications for the company of this proposal. [5]



29.1 The various risks to a life company’s solvency are:
1. Experience worse than expected / guaranteed, where experience means:
 expenses
 mortality (or other contingency where relevant)
 withdrawal rates (persistency). [2]
2. New business mix deteriorates (sell more of the less profitable or capital intensive
contracts). [½]
3. New business volumes decline or explode, possibly due to actions of competitors. [1]
4. New business strain causes solvency problems (this could be seen as a special case of
expenses worse than expected and new business volume / mix worse than expected). [½]
5. Discretionary payments to policyholders too large. [½]
6. High surrenders when surrender payments exceed asset share. [½]
7. Creditors default, including reinsurer and asset defaults. [1]
8. Supervisory authority imposes special restrictions or closes company to new business due
to misconduct or insufficient financial strength. [½]
9. Systems / staff unable to cope with new business. [½]
10. Fraud and/or mismanagement. [1]
11. Regulatory changes such as a strengthening of supervisory solvency rules and capital
requirements. [½]
[Maximum 5]
Solutions to these respective risks include:

1. Remove / reduce guarantees, including mortality, interest and expense guarantees (eg by
writing variable charge unit-linked business). [1]
Reinsurance against excessive claims experience (risk premium and/or catastrophe). [½]
Underwrite to ensure policyholders’ mortality is not worse than expected. [½]
Use stochastic asset / liability modelling to find appropriate risk / return balance for
investment strategy, thereby maximising return for any given acceptable level of risk. [1]
Implement expense controls and efficiencies. [½]
Implement policy retention activities, eg flatter commission structure. [1]

Monitor experience as it develops, by specific investigations and by embedded value or

profitability monitoring on continuously refined parameters. This will allow repricing or
other action almost as soon as problems arise. [1]
2/3/4 Financial reinsurance against new business strain. [½]
2/3/4 Product design / pricing to minimise new business strain. [½]
2/3/4 Attractive products, good bonus history, high commission to keep new business volumes
up. [1]
2/3/4 Commission structure reflects desired new business mix. [½]
5/6 Discretionary payments limited to smoothed asset share. [½]
6 Surrender penalties on early withdrawal, and clawback of commission. [1]
7 Invest in diversified secure assets to protect against asset default. [½]
7 Use more than one reinsurer. [½]
8 Train sales staff to reduce risks of mis-selling. [½]
8 Target solvency capital greater than minimum supervisory requirement to guard against
random fluctuations. [1]
9 Design products with systems / staff in mind to protect against inadequacy. [½]
10 Protect against fraud by strict internal controls and checks. [½]
10 Protect against mismanagement by monitoring experience attributable to managers

(eg new business volumes for marketing manager, embedded value growth for manager
of a subsidiary). [1]
11 Maintain close links with regulators and strengthen capital position in advance of any
changes. [1]
[Maximum 10]
29.2 (a) Without-profits
Under without-profits contracts the risk is borne by the company, not the policyholder. The
company will allow for this via margins in the assumptions used for pricing. [1]
Giving the company the right to review premiums throughout the term of the policy, say every
five years, would give the company protection against all types of risk. In practice there may be
constraints (eg competition, selective withdrawals, PRE) as to how much the company could
increase its rates. [1]
Also quota share reinsurance would reduce the company’s exposure to all risks. [½]

To protect against mortality risk, the company could:

 strictly underwrite each proposal to remove sub-standard lives [½]
 use individual surplus reinsurance to protect against losses from large cases [½]
 design the contract so that the sum at risk is not too large. [½]
To protect against withdrawal (persistency) risk the company could:

 pay low, or no surrender values [½]
 reduce the level of initial commission (thereby reducing the loss on surrender). [½]
To protect against investment risk the company could:

 match with fixed-interest stocks of an appropriate term, nature and currency [½]
 remove guarantees, eg guaranteed early surrender values. [½]
To protect against expense (and expense inflation) risk the company could:
 have tight control and continual monitoring of expense levels [½]
 match the expense liability with an appropriate inflation-linked asset. [½]
(b) With-profits
Under the additions to benefits and contribution methods, the risk is passed to the policyholder
via lower bonuses or dividends. This covers all risks, such as: [1]
 mortality
 expenses
 investment and inflation
 taxation
 withdrawals (persistency). [1]
In addition the company can use the techniques that have been mentioned for without-profits
business, such as underwriting, reinsurance, paying low surrender values and matching
investments. [1]
In the case of the revalorisation method, it is normally only the investment surplus that is split
between shareholders and policyholders. All other risks are borne by the company (so exactly the
same as without-profits in respect of the other risks). [1]
[Maximum 7]

29.3 Removal of company’s discretion over amount of payout means that company is much more
exposed to investment risk. [½]
Policyholders will exercise financial selection against the company, and will be more likely to
surrender when surrender value exceeds asset share, causing loss to the company. [1]
Liabilities will be harder to match since the timing of the proceeds is unknown. [½]
Investing short leaves company exposed to interest rate fall. [½]
Investing long leaves company exposed to interest rate rise. [½]
Ability to manage the increased investment risk will be very dependent on the size of the
company’s free assets. [½]
Some protection may be afforded by investing in suitable derivatives to match the option. [½]
Supervisory reserves should rise to reflect the increased cost of providing the guarantees. [½]
This will produce an immediate valuation strain (if the guarantees affect existing policies, though
this would be very unlikely). [½]
It will produce increased new business capital strain for new policies. Amount of free assets will
determine the extent to which future new business volumes may be restricted following the
change. [1]
Capital strain may be a particular problem if the guarantees lead to an increase in popularity of
the product, so that sales volume increases. [½]
Premium rates will probably need to rise to cover the cost of the guarantee. [½]
This may make the contract less marketable if the guarantee is perceived not to be worth the
extra cost. Sales may then fall, with a risk of reducing coverage of fixed expenses. [1]
[Maximum 5]

SP2-30: Monitoring experience Page 1
Monitoring experience
Syllabus objectives
5.2 Demonstrate the relevance of experience monitoring to a life insurance company.
5.2.1 Explain why it is important for a life insurance company to monitor its
experience.
5.2.2 Describe how the actual mortality, persistency, expense and investment
experience of a life insurance company should be monitored, including the
data required.
5.3 Demonstrate the relevance of analysis of surplus or profit.
5.3.1 Give reasons for undertaking an analysis of surplus and an analysis of

embedded value profit.
5.3.2 Suggest ways in which the results of such analyses can be used.

Page 2 SP2-30: Monitoring experience
0 Introduction
In this chapter we look mainly at how the actuary will complete the control cycle loop: how to
monitor experience in order to monitor the validity of assumptions, and if necessary to revise
them; or in order to take some alternative action. As was described earlier in the course, the
decisions taken by the company in the past are only as good as the models used to formulate
them. If the assumptions are not borne out in practice, then the decisions taken will not have
been the best that could have been taken, with hindsight. Our experience investigations allow us
to revise our assumptions (and our models) so that the decisions the company takes remain as
‘best’ as possible, within the practical limitations imposed by the operation of the business. The
experience monitoring for this purpose has two strands.
First, experience can be investigated directly – for example, we want to investigate mortality so
we do an investigation of the sort described in Subject CS2. These ‘explicit’ investigations are
described here for mortality (or other contingency), persistency, expenses and investment return.
This will tell us how much, for example, our actual mortality differs from our expected mortality
experience (where the ‘expected’ experience could be the assumption used in our pricing model,
for example). We can refer to this as an ‘actual versus expected’ investigation.
Secondly, we can investigate the financial impact of any difference between actual and expected
by performing an analysis of surplus. In other words, while an ‘actual versus expected’
investigation may tell us that mortality over the last three years has been 50% higher ‘than
expected’, the analysis of surplus will tell us if the 50% higher mortality has made a big dent in the
company’s profits – in which case remedial action should be taken – or whether it has hardly
mattered.
There are other reasons for analysing the experience, which are listed in the next section.

1 Reasons for monitoring experience

Experience will be monitored as part of the control cycle to:
 develop earned asset shares
 update assumptions as to future experience
 monitor any trends in experience
 monitor actual compared to expected experience and take corrective actions as

needed
 provide management information to aid business decisions
 make more informed decisions about pricing and about the adequacy of reserves.
Question
(i) List the items of experience that should be monitored for the purpose of performing asset
share calculations.
(ii) List the areas of actuarial work that use assumptions as to future experience.
(iii) Suggest corrective actions that might be taken if monitoring shows that actual experience
has been worse than expected.
Solution
(i) Asset share calculations
Asset share calculations will require historic information on:

 mortality experience (or other contingency if appropriate)
 withdrawal experience (if we want profits / losses from discontinuances to be reflected in
the asset shares of remaining policies)
 expenses.
(ii) Future experience assumptions
Assumptions as to future experience will be required for:

 model office work (which will include embedded value work, profitability monitoring,
financial projections, asset / liability modelling for assistance in setting investment
strategy, determination of reinsurance requirements)
 product pricing
 valuations
 setting discontinuance terms.

(iii) Corrective actions
The corrective actions that might be taken include:

 re-price products
 re-design products
 change the investment strategy
 change the sales strategy – eg which products are marketed through which distribution
channel; change the sales reward (eg commission) structures; etc
 change the reinsurance strategy
 change the underwriting strategy
 change the profit distribution strategy – eg bonus rates
 re-organise the workforce and IT systems to make more efficient use of expensive
resources.
In order for appropriate action to be taken, it is therefore imperative that the company
understands the reasons for any adverse experience.
Providing management information is closely linked to the previous point (ie as covered in the last
part of the question above). However, management of a life insurance company is not just a
reactive process (responding to adverse experience) but also a pro-active one. Hence while the
company may not be suffering particularly adverse experience, it may still be able to identify ways
in which it can make its operation more profitable. Therefore it will be interested as much in its
profitable activities as it is in its unprofitable ones.
Thus, by keeping its experience in view, it may be able to identify such things as:
 profitable products
 profitable sales channels or agents
 profitable markets
 efficient sections of the business
 successful investment strategies.
In providing management information, the important question to bear in mind is why?

Management information is provided to allow management to manage the company. This is
stating the obvious, but following on from that we realise that information must be provided at a
level of detail appropriate to the management concerned in order that, as a result of the
information, they can act as necessary. The information must also be reliable – so information on
persistency, for instance, would need to be based on at least, say, six months of experience to
give credible results, and so avoid the risk of management taking decisions based on incorrect
data.

2 Data required
2.1 Data grouping

The basic requirement is that there is a reasonable volume of stable, consistent data, from
which future experience and trends can be deduced.
By ‘consistent’ here we mean that the data we use should be in a similar form, preferably
extracted from the same source, grouped according to the same criteria and equal in terms of
reliability. Examples of inconsistent data include:
 where ‘age’ for one group is calculated as ‘calendar year of death minus calendar year of
birth’, and for another is calculated as ‘age last birthday at death’
 where lives counted as smokers in one time period are counted as non-smokers in
another time period, due to a change in the definition of what is meant by ‘smoker’ and
‘non-smoker’.
The data ideally need to be divided into sufficiently homogeneous risk groups, according to
the relevant risk factors described below. However, this ideal must be balanced against the
danger of creating data cells that have too little data in them to be credible.
The problem with grouping together heterogeneous lives into single cells for analysis, is that we
would not be able to tell whether a change in some observed variable was genuine or just the
result of having a different mix of lives within the cell. For example, if smokers and non-smokers
are not separated, an increase in observed mortality rates over time might be just the result of
having a greater proportion of smokers in the more recent data samples.
So that’s why, ideally, we need to subdivide our data into purely homogeneous groupings. We
will know, then, that the interpretations we make about any observed changes in experience will
be valid.
It is important to agree the period over which the data will be collected.
One of the key decisions in conducting any investigation is the time interval to be analysed. In
order to have reasonably sized cells, we want an interval as long as possible, and normally greater
than, say, one year. However, we want to investigate recent experience, not an average of recent
and ancient; so this will place an upper limit on the time interval. This upper limit will depend on
what we are investigating; for instance an expense investigation would normally be confined to
twelve months, whereas mortality investigations would often be based on a three or four year
span.
In practice, the level of detail in the classification of the data, therefore, depends upon the
volumes of data available. It would, however, at least be desirable to separate different
classes of contract.

2.2 Big data

Life insurance companies can take advantage of technical developments in the analysis of
large volumes of data (‘big data’).
Increasingly large quantities of data are being recorded about us every day: the things we buy, the
places we go, the websites we visit. As computers get ever faster, it is possible to analyse this
data to look for patterns and predict future behaviour. For example, some internet stores are
incredibly good at predicting the goods and services people want and sending personalised
adverts for just the right thing.
For example, bancassurers (companies which offer financial services that encompass both
banking and insurance operations) hold a large quantity of information about a life
insurance customer who also has a bank account with them, including data on that
customer’s spending habits.
The bank can analyse spending patterns on current accounts or credit cards to look for behaviour
indicating low mortality rates (eg subscriptions to a gym) or high mortality rates (eg purchases of
potholing equipment). The bank can then pass this information to its insurance operation.
Insurance companies that are not part of a bank could access other sources of data. For example,
supermarkets collect large quantities of information through their loyalty card schemes –
customers that buy lots of fresh fruit might have lower mortality rates than customers that buy
lots of cream buns.
The ability to analyse and use such information to understand better the underlying risks
will become increasingly important in insurance, as more sophisticated analytical
techniques develop to facilitate this.
Risk classifications are likely to develop as a result, allowing more accurate rating for each
individual customer.
Traditionally, life insurance has been rated on factors such as age, gender, occupation and
location, and then making generalisations that all the people within a particular rating cell will act
in the same way. In the future, life insurance may be rated on factors that directly affect
mortality such as diet and exercise.
The insurer also then has greater ability to select the preferred risks.
Big data can be used in many ways. Not only can it be used as part of the underwriting process, it
can also be used to target the kinds of people the company wants to insure. For example, a bank
could analyse its data to find customers that demonstrate a healthy lifestyle (these customers are
the preferred risks that the Core Reading is referring to) and then invite only these customers to
take out term assurance at preferential rates.
Similarly, a bancassurer might wish to target customers that it expects to have a low lapse rate,
perhaps by selecting customers that have held the same account, credit card or mortgage for
many years.
Monitoring the available data may also allow the insurer to drive better experience, through
early identification of changes in individual insured risks or potentially through being able
to intervene and influence customer behaviours.

For example, if a customer’s records showed a decline in the purchase of fruit and vegetables
then the insurer could send out advice on healthy eating.
We’ve described some of the advantages of big data above, but there are also risks.
Question
Describe the risks for an insurance company that uses ‘big data’.
Solution
There is lots of potential for reputational damage, eg concerns over privacy and data protection
failures. A lot of policyholders could be upset at receiving a message ‘you didn’t buy enough fruit
this month’. As a result, sales might fall and lapses might rise.
As the use of big data develops, the regulations governing it might change, eg the regulator may
require that certain types of data are not used. There could be fines for any misuse of the data.
The data collected may be inaccurate, incomplete or irrelevant, eg the insurer will not know
whether the person has actually eaten the fruit that they bought and will be unaware of the fruit
that was bought for cash at a market stall.
Whenever complex models are used there is the risk of choosing the wrong model.
The expenses of collecting and analysing the data may outweigh the benefits of the extra
information received.

3 Analysis of experience
3.1 Introduction
In this part of the chapter we consider the various demographic and financial investigations which
an actuary might conduct:
 mortality (and other contingencies)
 persistency
 expenses
 investment return.
3.2 Mortality experience
Subdividing the data

The data (both claims and exposed to risk) would ideally be analysed, where relevant, by
 type of contract
 age
 sex
 duration from entry
 smoker / non-smoker status
 medical / non-medical status
 source of business.
We would want to analyse the experience by the above factors even if we weren’t allowed to use
them all in pricing. For example, if regulation does not permit the use of sex (or gender) as a
rating factor, we would still analyse male and female rates separately if possible in order to
calculate the unisex rate that reflects our balance of male and female lives.
We also need to be careful about the time interval considered. This is for various reasons. First,
we expect mortality to change over time due to advances in medical science, changes in cultural
attitudes, improvements in standard of living, new diseases, etc. This would lead us to consider
an interval of at most five years.
Question
Describe the other constraints on the time interval.

Solution
The time interval needs to be large enough to ensure that there is a sizeable amount of data, and
that seasonal influences do not affect the data. For instance a study of annuitants’ mortality over
a hard winter would give rather misleading results.
However the period of investigation should not be too great, because we want to be able to act as
quickly as possible on any recent changes in mortality levels, in particular as regards product
pricing. Thus looking at an investigation based on the last five years of data, for instance, will
imply a reaction time three years slower than if we were to use just two years of data.
Further classifications (eg by location, even down to postcode level) could be used if there
is sufficient data.
Occupation (or at least broad occupational group) would also be a useful factor for a mortality
investigation of temporary assurance contracts.
The process
The process for analysing experience has been examined in Subject CP1 and will not be
re-examined here.
3.3 Persistency experience

The factors by which the data could be analysed, in rough order of importance, are:
 type of contract
Term assurances have different persistency rates from with-profits endowment

assurances for example, as healthy policyholders may be more likely to lapse the
former.
 duration in force
Persistency rates (as measured over a specified period, eg month or year) are
generally lower near the start of a contract.
For example, 95% of policies in force at the start of the fifth year may be in force at the
end of the fifth year, but only 80% of policies in force at the start of the second year may
be in force at the end of the second year.
 sales method used and target market
The degree of care taken in ensuring that a suitable product is sold may depend on
the sales method and target market. The more suitable the product, the better will
usually be the persistency experience.

 frequency and size of premium
With monthly premiums there are more opportunities to stop paying premiums than
if premiums are annual; conversely, a larger annual premium may be considered
less affordable than a smaller, regular payment. A high premium relative to income
will be harder to afford than a smaller one, but it may not be considered worthwhile
continuing to pay a small one.
 premium payment method
Premiums paid in cash are more noticeable than premiums paid directly from a bank
account and so lead to lower persistency rates.
 original term of contract
 sex and age.
For example, experience tends to be worse for younger ages.
These are the factors by which persistency experience could be analysed. In practice, often only
the first three factors will be considered, in order not to end up with a large number of almost
empty cells.
The fourth and fifth points are not necessarily independent. For example, annual premiums are
more often paid in cash than monthly premiums, which are often paid directly from a bank
account. Hence annual premiums are usually more noticeable than monthly premiums for two
reasons: being in cash and being large amounts. Persistency is likely to be lower for both reasons
together.
Question
Explain how persistency rates would be expected to compare for the following sales methods:
 broker
 direct sales force
 direct marketing telephone sales in response to newspaper advertising.
Solution
Lowest persistency rates: direct sales force

‘Medium’ persistency rates: direct marketing
Highest persistency rates: broker
Broker sales will be initiated by the policyholder in response to some perceived financial need,
and the broker should be able to advise the person well on the most appropriate policy. There
should not therefore be many policyholders who later feel that they do not need the policy.
Policies sold by the direct sales force will have been generated by some insistent salesman making
contact with a potential policyholder and convincing them that they need the policy. A number of
these will eventually no longer feel the need to keep paying premiums.

The direct marketing policies should have a persistency which lies between these two.
Policyholders will have reacted of their own accord to newspaper advertising, although the
telephone response and any associated policy literature would ‘push’ some people into taking out
a policy which they did not really need, and therefore lapse later on. There may also be a number
of people who take out the policy just to obtain a free gift (eg alarm clock) and lapse their policies
after receipt of their gift.
Persistency rates in the future are also affected by, amongst other things, the:
 economic situation
 competitive situation of the product – eg the introduction of a more attractive

product can have an adverse effect
 perceived value of the product to the customer.
These would not usually be allowed for in an analysis of past experience but may be used to
understand or explain patterns in this experience.
The process
For each homogeneous group to be analysed, full withdrawal rates can be determined as
follows.
The number of contracts issued in the company’s last financial year is divided into the
corresponding number that survive in-force until the first policy anniversary to give a
first-year persistency rate. The first-year withdrawal rate can be determined as one less the
persistency rate.
Deaths and maturities should be excluded from the calculation (if material).
A similar procedure can be adopted to obtain the second year, third year, etc withdrawal
rates, by looking at the number surviving from the number of contracts, in each group, that
have their first, second, etc policy anniversary in the last financial year.
There are two possible approaches to calculating persistency rates in the second and subsequent
years. Firstly, the persistency rate can be calculated over a defined period as the number
surviving the year divided by the number in force at the beginning of the year. Secondly, the
persistency rate can be calculated cumulatively as the number surviving the year divided by the
number in force at the outset of the contract.
Question
A company sells 100 insurance contracts. At the end of the first year only 50 contracts are still in
force. At the end of the second year there are 40 contracts in force and at the end of the third
year there are 36 contracts in force.
Calculate the following:

 the cumulative persistency rate
 the persistency rate as measured over each year
 the withdrawal rate.

Solution
The cumulative persistency rate is 50%, 40% and 36% for the first, second and third years
respectively. So we see that the cumulative persistency rate falls with duration in force.
The persistency rate as measured over a year is the same for the first year, ie 50%. For
subsequent years we must divide the number of surviving policies at the end of the year by the
number in force at the start of that year. So the persistency rate for the second year is 40 / 50 =
80% and the persistency rate for the third year is 36 / 40 = 90%. So we see that in this case the
persistency rate as measured over a year rises with duration in force.
The withdrawal rates are calculated as simply one less the persistency rate over the year. So the
withdrawal rate is 50%, 20% and 10% for the first, second and third years respectively.
As the results are examined, it will be clear that for some groups there is little difference between
withdrawal rates at different durations (for instance the difference in withdrawal rates for policies
of duration twelve years compared with those of duration thirteen years), and intuitively a
difference would not be expected. For some groups the differences emerging might seem to be
statistical quirks stemming from small cell sizes (eg withdrawal rates for eighty-six year old
policyholders). As a result of these considerations, the actuary would regroup the data and
recalculate withdrawal rates for these broader groups.
After having finished the investigation itself to the greatest possible degree of valid accuracy
(ie regrouping to avoid spuriously small groups), the presentation of results for management
information purposes would probably require further consolidation depending on the level of
management involved. For instance senior management would probably want results split by
broad product class, not by every product.
In addition to analysing persistency into the groups described above (type of contract,
duration etc) the analysis may be broken down to look at the different types of discontinuance as
described below.
Analysis of policies that are made paid-up is normally done as a subsidiary part of the
withdrawal analysis.
There may also be a related analysis of partial (or income) withdrawal rates, depending on
product design.
3.4 Expense experience

One way to think of an expense analysis is to consider the desired end results of the process,
which themselves will depend on the purpose of the investigation.
For some purposes it may be sufficient simply to know the expenses in total: for example, a
company can calculate its total profit over a year without performing a detailed expense analysis.

For other purposes this would be completely inadequate, however:

 if the company uses the contribution method then it will need to analyse expenses into
policy groups so that the dividend can be properly calculated for each group
 when calculating asset shares – perhaps for determining terminal bonuses or terminal
dividends, or when assessing appropriate surrender value scales – historical expenses
have to be apportioned between different policy types
 for pricing and reserving it is essential that a policy’s fair share of the company’s costs is
established, so that the correct premiums or charges can be levied.
Ultimately, the expenses are a function of the work that the company does – ie administering a
portfolio of life insurance policies. If we want to project the company’s future expected expenses,
then we will need to determine the relationship between the company’s business – the policies it
sells and maintains in force – and the expenses incurred in doing so.
In most of these cases, therefore, the main idea is that we want to be able to attribute the
historical expenses to each policy that the company had on its books over the period during which
the expenses were incurred. In other words what was policy X’s fair share of the company’s
actual expenses over (say) calendar year Y?
But we need to be more sophisticated than this, in order to avoid significant model error at the
modelling stage. For example, we need to identify a policy’s share of the initial expenses
separately from its renewal expenses. This is vital, for example, in order to estimate the likely
capital strains of writing new business and therefore how much business we can afford to write.
Similar arguments lead to the conclusion that we need each policy’s share of the expenses to be
subdivided as follows:
 expenses relating to new business
 expenses relating to existing business
 expenses relating to terminations (ie deaths, maturities, surrenders, lapses – and possibly
for each of these separately)
 expenses relating to investment management.
For each of the first three categories, we also need to identify whether the expenses are:
 proportionate to the premium payable under the policy
 proportionate to the benefit level (eg sum assured, annual amount of annuity) under the
policy
 fixed amounts per policy.

Example
Suppose the historical expenses over one year for a particular annual premium assurance product
line are as follows (arbitrary units):
New business Existing Terminations

Per premium 500 1,000 0
Per sum assured 100 0 750
Per policy 200 5,000 250
Define:
P(new) = total new annual premiums received during year
P(exist) = total premiums received from existing policies during year
S(new) = total sums assured under policies issued during year
S(term) = total sums assured under policies terminating during year
N(new) = total number of new policies issued during year
N(exist) = average number of policies in force (excluding new business) during year
N(term) = total number of policy terminations during year
Then we would calculate a policy’s share as:
500
Initial expenses  per unit premium
P(new)
100
 per unit sum assured
S(new)
200
 per policy
N(new)
1,000
Renewal expenses  per unit premium
P(exist )
5,000
+ per policy
N(exist )
750
Termination expenses  per unit sum assured
S(term)
250
+ per policy
N(term)
After suitable adjustment for expected future changes to the experience, we could now use these
results as assumptions in our projection models.

In the above example, we have described the amounts as all relating to a particular product line.
Whether or not expense experience is subdivided by product line, or by some other classification,
depends on the nature of the business (and in particular, whether the expense contribution
differs significantly between categories and whether the volume or quality of the data justify it).
An important consideration for an expense investigation is the time period involved.
Question
Discuss the time period that should be used for an expense investigation.
Solution
It depends somewhat on the purpose and context of the investigation. In practice, companies
would normally be seeking to perform an expense investigation on an annual basis, in which case
a logical time period to use would be one year (and, of course, we are talking about the most
recent year’s experience here). If a shorter period were to be used, then any seasonally
influenced expenses will not be correctly represented (they will be either under- or over-
represented in the totals). New business figures are often sensitive to the time of year, in which
case expenses associated with selling or processing new business would be particularly distorted
by only looking at part of a year.
On the other hand, the company would not just look at its most recent year’s expenses in
isolation. It would also need to look at, say, the most recent three to five years of expense
experience, to identify and if necessary allow for unusual or ‘random’ influences, and to identify
trends that could be occurring over time. When looking over several years it would be necessary
to adjust the figures for inflation, so that each years’ costs could be compared in current money
terms, for example.
The other important consideration in an expense analysis is the subdivision between direct and
overhead expenses.
In theory, the expenses of a life insurance company can be divided into:
 direct expenses – the expenses that can be attributed directly to a product or policy
 overheads – the balance of the expenses, ie those that relate to general management
and service departments which are not directly involved in new business or policy
maintenance activities.
Direct expenses are typically variable (ie they depend on either the volume of new business
or the level of in-force business), but may be fixed to some extent.
For example, the amount of underwriting performed will depend on the number of applications
that need to be underwritten, ie it is a variable cost. However, we wouldn’t expect to hire and
fire underwriters on a daily basis depending on volumes, so underwriters’ salaries are a fixed
expense at least in the short term.

Overheads are more typically fixed costs (ie are insensitive both to the volume of new
business and to the level of in-force business), but will be variable to some extent – often in
step changes.
The cost of the head office building (rent, heating, lighting) is largely a fixed overhead. The head
office will remain the same size regardless of how many policies sold in the month. However,
large companies with lots of policies will have much larger head offices than small companies. As
a company grows it may move to progressively larger premises, so this cost tends to vary in steps.
In practice, there is not a clear dividing line between these two categories, and some
judgements will have to be made.
We therefore need to consider how each of the expense assumptions we have estimated from
the data should be divided between direct and overhead expenses. In practice, nearly all of the
overhead expenses can be considered as per policy, while the direct expenses fall into all three
groups. This means that the per premium and per sum assured categories are (very broadly)
100% direct expenses, while the per policy expenses are a mix between overhead and direct. For
some modelling purposes, particularly those involving full model office projections, it is important
to model the overhead expenses as a global assumption (ie they should be modelled
independently of the number of policies projected to be in force).
Summary: subdividing expense data

The Core Reading provides a summary of what we are trying to achieve in an expense
investigation.
Commission (if applicable) is normally excluded from the expenses being analysed on the
basis that its format is known and can be allowed for explicitly.
While we haven’t mentioned this before, it should be obvious, as this is one of the few items of
future experience which is known with certainty. Commission can therefore be loaded for
explicitly in the pricing basis (as a proportion of premium or however it is actually defined), so
that past commission costs are actually irrelevant. The expense investigation proper is therefore
concerned with an analysis of the non-commission expenses only.
For the purpose of an expense analysis, the non-commission expenses can then be split into the
categories in the following table.

Breakdown of When the expenses

Any other factors
non-commission expenses arise
Is the expense
Initial expenses Start of the policy proportional to the:
Regularly during the written or in-force?
Renewal expenses
policy term  amount of benefit
written or in-force?
Termination expenses End of the policy  amount of premium

written or in-force?
Management of the
Investment expenses
company’s assets
So the first step is to break down the expenses into initial, renewal, termination and investment
expenses. Then the first three types of expense can be further broken down into whether they
are proportional to the number of contracts, amount of benefit or amount of premium.
It is found in practice that most of these expenses are proportional to the number of
contracts written or in-force. Exceptions can include:
 marketing expenses – may be related to the amount of initial commission paid (if
applicable)
 underwriting expenses – may be related to the size of benefit.
Question
Explain the justification for these two links.
Solution
Marketing expenses: initial commission payments to agents will determine the importance of
each sale to the agent. It is reasonable that the marketing department of the life company relates
the effort and time dedicated to assisting agents in their sales to the worth of the sale to the
agent – ie the initial commission.
Underwriting expenses: the company may choose to perform more medical and financial
underwriting if the benefits are large due to the increased mortality risk.
Investment expenses are normally expressed as a percentage of funds under management

so that they can be treated as a deduction from the earned investment return.
More on this later.

Having been separated by type, the expenses will need to be further split down and
analysed into the required ‘cells’.
For example, the cells may be based on each accounting fund or each main product line of
the company. These may be further sub-divided between regular and single premium
business.
The choice of cells will vary across companies depending upon the types and volumes of
business written and the requirements of the analysis. The cells chosen should not be so
small that the analysis becomes unreliable.
The process
The next stage is to consider how to split the company’s various expense items into the
subdivisions we require. Commission, as mentioned above, is easy to do, so the remaining
process description considers expenses other than commission.
The main items of expense are:
 salaries and salary-related expenses
 property costs
 computer costs
 investment costs.
There is no definitive approach to splitting these. One possible approach follows.
Salaries and salary-related expenses

These will be the major expense of a life company.
Staff can be split into three groups:

(i) staff whose work comes entirely within a single cell of the analysis
(ii) staff whose work comes within more than one cell
(iii) other staff.
In the approach described here, we are treating all expenses which can be directly attributed to a
particular product line as direct expenses. Hence categories (i) and (ii) will be 100% direct, while
category (iii) will be partly direct and partly (in some cases 100%) overhead. But, as we said
earlier, this is not the only possible view we can take about the split between direct and overhead
expenses: it is not an exact science.
The salaries etc of staff in (i) can be directly allocated to the appropriate cell.
An example of a single cell would be ‘the initial expenses of product X which are proportionate to
the sum assured’, and an example of the staff whose costs we might attribute to this cell might be
certain members of the company’s underwriting department. However, if the analysis involves
cells for every product then it is likely that there will not be any category (i) staff, since it is rare
for staff to be dedicated to individual products.
For group (ii), staff timesheets can be used to split their salaries etc between the
appropriate cells.

Question
An administration department services two products. A staff member, who cost the company
£30,000 in salary, pension and other benefits over the previous calendar year, recorded their
activities over the same period as:
Product X Product Y
Initial 30% 40%
Renewal 20% 10%
Termination 0 0
Calculate the amounts to be attributed as expenses in relation to this person, split by product and
expense type.
Solution
The attributed expense amounts are:
Product X Product Y
Initial £9,000 £12,000
Renewal £6,000 £3,000
Termination 0 0
The work of the group (iii) staff will straddle both overheads and direct expenses. The split
between the two is likely to be made pragmatically. The direct part can be split further in
proportion to the overall split of the group (i) and (ii) staff.
For instance, some such staff – such as the company’s catering staff – would be entirely overhead.
Other group (iii) staff might spend some of their time identifiably involved with some of the cells.
For example, a computer support department responsible for dealing with systems failures and
misbehaving PCs might find that some of their work, such as repairing a crashed quotes system,
was clearly ‘new business’ while other work, such as switching on the Managing Director’s PC,
would be unclassifiable as a direct expense, and so must be treated as an overhead.
Having dealt with the direct part of group (iii), how do we deal with the overhead component?
The key word here is pragmatically. There is no rigorously right or wrong method here, and
different companies will take different approaches.

Common approaches include:

 splitting them down in proportion to the direct expenses already identified
 splitting them in proportion to expense charges from the policies (as it should be easy to
calculate total initial expense charges, renewal charges, etc)
 splitting them down in proportion to something else – eg the number of ‘new business’,
‘renewal’ etc staff they serve, if we are talking about a department such as catering or PC
support.
Question
Identify which of groups (i) to (iii) the following staff would be best placed within, including for
any group (iii) staff an indication of whether they are entirely overhead or mixed overhead /
direct:
(a) staff in the valuation area
(b) staff in the secretarial pool
(c) an actuary doing an expense investigation.
Solution
Valuation staff: group (i) – all of their work could be attributed to renewals; or group (ii), if the
analysis needs to break time down between different products; or group (iii), if we consider part
of the valuation work as an overhead cost: for example, the number of valuation staff may be
insensitive to whether there are 100 or 100,000 temporary assurance policies in force.
Secretarial staff: group (iii), mixed direct / overhead.
Actuary doing expense investigation (lucky individual): group (iii), since the work will not relate
directly to any product.
Property costs
If the company owns, as an asset of its long-term fund, any of the buildings that it occupies,
a notional rent needs to be charged to the relevant departments.
This is because, as an asset of the long-term fund, the property ought to be earning a return
(rental income) from its tenant. However, owner occupation means that no such rental is
forthcoming, so that occupation has the effect of costing the company to the tune of the annual
rental that it could have earned if the property were let. This is the ‘notional rent’: it is an
expense to the company, we need the policy expense loadings to cover it, and it must therefore
be included in the expense analysis.
Of course, if the company actually rents its office buildings, it would have a real rental cost for
each year: the treatment of actual or notional rent in the analysis would be the same.

‘Long-term fund’ is the normal term given to the policyholders’ fund in the UK. However the
same concept of charging a notional rent applies whether the property is owned within a
policyholders’ fund or as part of the shareholders’ assets.
This rent, plus property taxes, heating, lighting, cleaning costs etc, can be split, for example
by floor space occupied, between departments and then allocated in accordance to salaries.
So once property and associated expenses have been allocated to departments, they can then be
allocated to cells in the same proportion as staff expenses in that department. As with the
allocation of any overheads, this is not a uniquely correct way of doing the allocation, just one
sensible and pragmatic solution.
Computer costs
The cost of purchasing a new computer could be amortised over its useful lifetime and then
added to the ongoing computer costs. These can then be allocated according to computer
usage.
Most computer usage should be readily identifiable as belonging to one of the cells. For instance,
valuation runs would be a renewal expense while quote calculations would be new business.
Some expenses, eg a company’s internal electronic mail system, would not be readily identifiable
and would have to be split between the cells pragmatically – perhaps in proportion to the other
(known) computer costs, or allocated to departments in proportion to staff numbers and from
there to cells.
Investment costs
As stated earlier in this section, these costs (investment department, stamp duty,
commission etc) would be directly allocated to investment expenses and hence allowed for
in assessing the investment return to use for pricing etc.
(Stamp duty is a form of tax found in the UK which may be incurred on certain investment
transactions. Similar taxes might be found in other countries.)
The common approach to take with investment expenses is to represent them as a reduction in
yield. So, in doing a profit test for example, initial, renewal and termination expenses would be
modelled explicitly and investment expenses would be felt implicitly via a lower yield. In profit
testing and model office work, investment expenses could be modelled explicitly if required (as a
percentage of reserves), but the yield reduction approach may be better with an ‘equating
present values’ formula.
Question
Give another reason why it can be preferable to present investment expenses as a reduction in
yield.

Solution
In this way investment yields will be more comparable between different asset classes.
For instance, it would be misleading to compare the yields of government bonds and large
property shopping mall developments without taking into account the vastly greater investment
management expenses of the latter. So it is important to subdivide the investment expenses by
broad asset class, and obtain the appropriate reduction in yield for each asset class separately.
One-off capital costs (other than purchase of a new computer)

The expenses analysed elsewhere should exclude large one-off capital costs, which need to
be amortised over the expected useful lifetime of the item purchased. The amortised cost
may then simply be treated as part of the overheads.
If the item can be treated as an asset of the long-term fund, eg a new head office building,
the cost would not be amortised. Instead a charge, eg a notional rent, would usually be
made.
Exceptional items, which are not likely to recur, would be excluded completely from the
analysis.
Question
Explain why exceptional items should be excluded from the analysis.
Solution
The expense investigation results are going to be the starting point for setting assumptions for
pricing, reserving and embedded value work. These one-off expenses are defined as ‘non-
recurring’, but nevertheless they have each occurred once, and others like them (though not
necessarily having the same causes) are likely to occur in the future. However, their
unpredictability means that the past experience is not likely to form a very reliable basis for future
projection, and it would probably be more appropriate to allow for them through a margin in the
expense assumption, either directly or through the risk discount rate (where applicable).

Summary: the process

 We start by knowing the expenses, such as salary costs, computer costs, etc.
 We know commission, so this can be excluded from the analysis.
 We then need to subdivide the non-commission costs into the required ‘cells’ – ie into
initial, renewal, termination, investment, and whether related to per policy, premium, or
sum assured.
 Some expenses can be allocated directly to a particular cell.
 Staff expenses may need to be subdivided between cells by the use of timesheets.
 Overhead expenses need to be allocated pragmatically, eg in proportion to the split in
direct expenses.
 Property expenses can be allowed for by charging a notional rent to each department in
proportion to the floorspace occupied, then allocated to the different cells in proportion
to the department’s salary costs.
 Costs of new computer equipment can be spread over their future expected lifetimes and
then allocated to departments in proportion to usage.
 Investment costs would be subdivided by asset class and usually allowed for by a
reduction in yield for each class.
 One-off capital costs would be spread over the expected future lifetime of the item, then
just treated as an overhead.
 Exceptional expense items may be ignored in the analysis, but their future incidence may
be allowed for in margins in the future expense assumptions or risk discount rate.
Having now reached the end of the expense investigation, it would be normal to reconcile the
results (summing expenses for new, in-force and terminated policies over all policies) with the
total expenses in the published accounts.
Question
Outline other checks that could be made at this stage.
Solution
It would be normal to compare results with those of the previous expense investigation(s), both
to look for trends which may indicate the need for action, and as an immediate check. Such
comparisons should take account of expense inflation.
It could also be interesting to compare results with industry statistics, again both as a check and
to indicate possible problem areas.

3.5 Investment experience

A company will want to assess the return it is achieving on its investments. The methods
used to do this are examined elsewhere and will not be examined here.
Which basis is used for measuring the actual investment return will depend on the purpose. So, if
we wished to determine the amount of investment surplus to distribute to with-profits
policyholders using the revalorisation method, we would probably use partially written-up book
values; if we wished to compare the investment performance of two different unit fund links, we
would use market values; etc.
The experience is likely to be analysed by main asset types and may be done both gross
and net of investment expenses.

4 The analysis of surplus / profit

A company will want to analyse the surplus arising over a year, for example on its
supervisory basis, in order to:
 show the financial effect of divergences between the valuation assumptions and the
actual experience, exposing which assumptions are the more financially significant
 show the financial effect of writing new business
 provide a check on the valuation data and process, if carried out independently
 identify non-recurring components of surplus, thus enabling appropriate decisions

to be made about the distribution of surplus to with-profits policyholders
 give management information on trends in the experience of the company
 comply with regulatory requirements.
4.1 Valuation surplus arising
Question
Define the surplus arising over a year.
Solution
The surplus at any point in time is the value of the assets less the value of the liabilities.
So using the notation A for assets, V for liabilities, the surplus arising from time 0 to time 1 will be:
(A1 – V1) – (A0 – V0).
However, the word ‘surplus’ is often used to mean surplus arising. It is convention to do so, and it
is normally obvious from the context which is meant (ie surplus or surplus arising).
An analysis of surplus is a breakdown of the surplus arising over a year into elements such as
‘surplus arising attributable to investment return experience exceeding valuation expectation’.
Question
List other contributors to surplus (or loss) that might be included in an analysis of surplus.

Solution
Surplus / loss due to the difference between actual experience and valuation assumptions for:
 mortality (and other contingencies)
 expenses
 withdrawal rates
and the impact of new business.
Changes in valuation assumptions would also contribute to surplus (or loss).
You would not be expected to do an analysis of surplus for the Subject SP2 exam. However, you
should know the uses of the analysis, and your understanding of the financial dynamics of a life
company could be tested by questions relating to the surplus.
In any question on surpluses arising, always consider whether the valuation basis involved is
statutory or not. This can affect what you might expect to see in the analysis.
We now examine some of the uses listed in the Core Reading.
4.2 The financial effect of A vs E, and of writing new business

As experience differs from the assumptions in the valuation basis, a surplus or loss will arise. For
instance, looking at mortality, the mortality surplus / loss will be
expected death strain – actual death strain
where
expected death strain = expected mortality × (sum assured – reserves)
actual death strain = actual mortality × (sum assured – reserves)
(We are ignoring expenses, the effect of premiums not being paid after death, and the effect of
interest for part of the year).
For the elements interest, mortality, expenses and withdrawals, the analysis of surplus tells us:
 if actual experience has been better or worse than expected, and
 what the financial impact is of that variation.
If we calculate supervisory reserves on a prudent basis then we should always expect to see a
surplus in an analysis of surplus on the supervisory basis. So alarm bells should start to ring if any
item is negative. But we might be doing the analysis on some internal realistic basis. What
surplus do we expect to see, and how does the actual result compare with that?

For instance, again looking at mortality, we would in effect be comparing

 realistic mortality expectation, and
 actual mortality.
So we would expect to see negatives just as often as positives, especially if the portfolio of
business is fairly small. A negative result is not in itself a big worry, but would be so if it were very
large, or were to be repeated year after year.
Question
Describe the actions that might be taken if the surplus arising from a particular source is negative.
Solution
The company should consider changing the best estimate assumption for that parameter. If the
phenomenon is of significant size, and the same thing happened the previous year or couple of
years, then it is likely that the current assumption is incorrect and should be revised. This might
then indicate a need to reprice the product, if profit testing on this revised basis shows
unsatisfactory profitability.
It also might be appropriate to consider changing the appropriate valuation assumption,

especially for a statutory basis where the level of prudence in the assumption might need to be
justified to the regulators.
The company may wish to investigate the cause of the negative surplus and see whether the
experience can be improved by taking other actions, eg implementing expense controls.
One of the more interesting influences on surplus is new business.
Question
(i) Describe how surplus on a prudent supervisory basis would be affected by writing new
business.
(ii) Explain how the answer to part (i) would change if using a realistic basis.

Solution
(i) New business will normally contribute negatively to the revenue account in the year of
inception, because the premiums will need to pay acquisition costs, administration costs,
contribute to company overheads and set up the prudent supervisory reserves.
On regular premium business, this will lead to a large reduction in surplus arising; the
effect will normally be much less significant with single premium business because of
lower acquisition expenses.
(ii) If using a realistic basis, the initial reserves would be much lower, perhaps even negative.
The reserves should reflect the future profits expected from the business, which should
offset the initial costs involved (otherwise there is something wrong with the pricing
basis).
We should not therefore get any new business strain.
4.3 Valuation check

If we derive the analysis of surplus components from scratch (rather than working backwards
from the valuation result) and they total to give the surplus arising then we have performed a
valuable check on the valuation process. Even if the analysis is not done with completely
independent data, it can still highlight errors, and so has a useful checking function.
4.4 Assistance with bonus distribution

The Core Reading says that the analysis of surplus can help ‘to identify non-recurring components
of surplus, thus enabling appropriate decisions to be made about the distribution of surplus to
with-profits policyholders’.
Question
Explain why we are interested in identifying non-recurring components of surplus as opposed to

recurring components of surplus.
Solution
Some items of surplus we would expect to see – eg we would expect to see a mortality surplus
every year due to the prudence of the mortality valuation assumption. These are the items which
the Core Reading refers to as recurring. The analysis of surplus will be more interesting in
showing us things which we were not expecting – for instance a negative contribution to surplus
due to a large number of decrepit policyholders exercising a ‘renewability without medical
evidence’ option.

The answer to the above question explains how the analysis can aid bonus distribution where the
actuary has flexibility to declare bonuses – ie the additions to benefits method. So with that
bonus system, if an item of surplus was non-recurring the actuary would not want to declare it as
a regular bonus.
Question
Explain the previous sentence.
Solution
Under the additions to benefits system, it is normal to keep the regular bonus rates stable over
time, so a one-off source of surplus would not be suited to distribution in this way. It would be
better to distribute it via a special or terminal bonus. This will avoid creating erroneous
policyholder expectations.
This is generally true for both conventional and accumulating with-profits contracts.
What about other bonus systems?
With the revalorisation system, normally the actuary is not able to declare bonus levels other than
by reference to some prescribed method, so the analysis will not help with the bonus declaration.
With the contribution system the picture is broadly intermediate between the other two
methods. The non-recurring item may or may not be included in the company’s distributable
surplus for that year. Where a significant terminal dividend is being used, the company is likely to
exclude the item from this year’s distributable surplus, and to include it in the terminal
distribution, like the UK method.
The contribution method is, in fact, closely related to an analysis of surplus on the supervisory
valuation basis. The dividend formula defined earlier in the course comprises the three key
components of surplus that we have been referring to in this section: investment, mortality and
expenses.
4.5 Information on trends

Just as the mortality and withdrawal investigations described earlier would normally be done over
a period of several years, so too the valuation surplus investigations will normally be most useful
when looking at several years’ investigations together.
This will allow us to see any significant trends emerging, for instance a gradual increase in
withdrawal surplus.
Question
Describe the other advantage of looking at several years’ results.

Solution
We are less at risk of being unduly influenced by random fluctuations, and thereby this helps to
establish any observed differences as due to parameter error and hence more likely to be worthy
of remedial attention.
4.6 Analysis of embedded value profit

A company may analyse the change over a year in its embedded value (the sum of net
assets and the present value of the expected future profit from existing business).
This will allow the company to:
 validate the calculations, assumptions and data used
 reconcile the values for successive years
 provide management information
 provide data for use in executive remuneration schemes
 provide detailed information for publication in the company’s accounts or those of

any parent company, in particular the value of new business taken on by the
company.
Validation of calculations, assumptions and data

The key here is the validation of assumptions. It is common to use the same assumptions for
embedded value work as for pricing (other than the risk discount rate which is often set higher for
pricing to take account of the extra risks involved in acquiring business). It is crucial that the
assumptions used in product pricing be valid, and their continuing validity can be monitored by
the analysis of embedded value movement.
Reconciliation of successive years’ values

Due to the complexity of the models used to determine embedded values, and all of the other
opportunities for error to creep in (for instance, insufficiently representative model points), it is
essential to check the results thoroughly. One of the best ways of doing this is to reconcile the
embedded value with the previous result, ie to identify in an analysis every contribution to the
movement in embedded value. So we are really talking about a very detailed check here, which
will in turn validate the calculations and data.
The details of how the surplus arising and the change in embedded value can be analysed
will not be examined in Subject SP2, although there is further coverage in Subject SA2.
Although detailed knowledge of the analysis of embedded value movements is not required, this
would be a suitable point to go over the basic ideas of how an embedded value ‘behaves’.

Example
We shall do this by considering the movement of the embedded value of a company which has a
policy portfolio giving a profit of 10 pa for the next five years, start year free assets of 100, and a
risk discount rate of 10% pa. Assume that the assets earn 8% pa.
The present value of future profits (PVFP for convenience) is 10a5  37.9 .
So at the start of the year we have:
PVFP Free assets Embedded value

37.9 100 137.9
What happens during the year? Suppose that experience is in line with assumptions. During the
course of the year the free assets earn an investment return of 8%. In addition, profit from the
policy portfolio will flow through the revenue account into the free assets. So free assets at the
end of the year will be
100  1.08  10  118
What happens to the PVFP? It will increase by the risk discount rate, because the future profits
are now one year closer; and it will decrease by the profit being transferred from it into free
assets in that year. So
PVFPend  PVFPstart  (1  risk discount rate)  profit in year
ie PVFPend  37.9  1.1  10  31.7
This checks with what we should expect by calculating PVFPend from scratch as
10a4  31.7
So the situation at the end of the year is:

31.7 118 149.7
So the PVFP has decreased, free assets have increased and the embedded value has increased by
8.6%.
Question
Explain why we should expect an embedded value growth of around 8½% in the above example.

Solution
The PVFP will earn 10% by definition since it has been determined using a risk discount rate of
10%.
The free assets will earn 8%.
The total earnings of the two, given the relation free assets : PVFP, should therefore be around
8½%.
This is a good example of how holding capital, as opposed to using it, dilutes the return obtained
by the shareholders (ie from 10% to 8.6%).
The above example demonstrates how the embedded value will increase over time, even without
including any new business – and assuming that experience is in line with assumptions.
Question
(i) Describe the impact on embedded value if experience is not in line with assumptions.
(ii) Describe the impact on embedded value of writing new business.
Solution
(i) If experience is worse than expected, free assets will increase by less than the expected
yield (eg because there are extra claims to pay, or investment income is lower than
expected, or capital values have plummeted) and the PVFP component will increase at a
rate lower than the risk discount rate (for instance because, due to a high number of
withdrawals, there are fewer policies around generating future profits). The converse is
true if experience is better than expected.
Assumptions will be realistic (or close to realistic), so the A vs E effect should on average
be zero over time.
(ii) Overall, the embedded value (EV) will increase if the business is profitable on the EV basis.
Considering the two components of EV separately, the PVFP will increase, whilst the free
assets will normally decrease.

Example
A simple numerical example shows the impact on embedded value of writing new business.
Consider the sale of a policy with the following characteristics:

 The PVFP at inception but before incurring any initial costs is 5 (in other words, the net
present value of profits seen in a profit test is 5).
 The premium received at inception is 3.
 Initial costs and commission at inception are 10.
 There is no valuation strain.
To simplify the matter, we are ignoring the company’s existing portfolio. So before sale of the
policy we have, say:

0 100 100
The impact of new business on this will depend on two things: the future profits we expect from
the business over its lifetime, and the initial cashflow involved. Now the PVFP of the policy at
inception but before incurring the initial cashflow is 5. So the embedded value once the policy is
sold but before any cashflow occurs will be:
5 (PVFP of new policy) + 100 (free assets) = 105
ie due to selling a profitable policy, the embedded value has increased. At this stage, before any
real cash changes hands, the increase is all attributable to the PVFP component.
Then, over the next day, we receive a premium of 3, and pay initial expenses and commission
totalling 10, producing a net cashflow (CF(0)) of 7 . So the PVFP the day after inception
(PVFP(0+)) will have increased by 7, because:
PVFP = CF(0) + PVFP(0+)
ie PVFP(0+) = 5  (7)
= 12
The free assets will have gone down by 7. So the embedded value one day after inception is:

12 93 105
So the overall impact of writing new business has been to increase the embedded value, although
free assets have decreased. In effect, the free assets have provided capital to finance future
profits. However, the present value of these future profits (ie 12) exceeds the initial capital
required (ie 7), so the impact is an increase in embedded value.

Management information
The Core Reading referred to one of the purposes of the analysis of embedded value movement
as being to provide management information. The sort of information that it will give is:
 the value of new business written, normally by product
 the amount of any expense profit or loss
 the amount of any mortality profit or loss
 the amount of any withdrawal profit or loss
 the impact of free assets on embedded value growth (ie is spare capital being used
efficiently, or is it just sitting around earning market yields and hence reducing the overall
return achieved by the company?)
 the impact of supervisory solvency capital requirements on the rate of return achieved.
Question
Explain the relevance of these expense, mortality and withdrawal profits, given that we already
know them from the analysis of valuation surplus.
Solution
The embedded value basis will be much closer to reality than a prudent valuation basis, so
deviations from expected will immediately tell us something useful, compared with the valuation
analysis deviation against an extremely prudent assumption.
The embedded value deviation from expected will take account of the future loss of profits on
decrement (ie the change in PVFP due to decrements) while the valuation analysis looks only at
free assets.
For instance, if mortality is greater than expected for a block of term assurance business, there is
a reduction in embedded value because the free assets will have decreased due to the extra
claims; but the future profits will also have suffered due to there now being fewer policies than
expected left to generate future profits.
Also, the embedded value may have expenses and withdrawals modelled much more accurately
than would be possible in a valuation basis.
Data for use in executive remuneration schemes

Executive remuneration schemes should ideally be based on measures that reflect how well a
company is really doing and so reward good management actions. The change in EV over a year is
a good measure to use because it will be increased if, for example, the company has sold lots of
new policies on terms that are expected to be profitable. The change in supervisory surplus might
be a less good measure to use as supervisory surplus may be depressed by new business strain.

Information for accounts

In some countries, it has been common practice for proprietary companies to publish the
embedded value in their accounts.
Question
Explain the purpose of adding embedded value information to published accounts.
Solution
Published accounts normally misrepresent the real financial situation of a company because they
do not take any credit for the future profits expected to arise from business already written. This
is a particular problem if a company is expanding quickly, because it will be making large revenue
account losses (unless using heavy financial reinsurance) and will so seem to be in a bad way,
even if in reality it is writing a lot of good business which will generate large profits in the future.
Reporting embedded values will show a truer picture, although it is also much less prudent
because it is entirely based on the idea of taking credit for future profits.
Where this is so, companies may also want to highlight the changes in embedded value from one
year to the next. This information will allow shareholders (and financial analysts) to see what the
real drivers of company profitability are.

5 Using the results

We now consider how the results of all these investigations can be fed back into the control cycle.
The results of analysing the experience, the surplus arising and the change in the
embedded value will be used by the actuary to reassess their view of the future with regard
to the company. This may result in changes to the assumptions or models used for pricing
or reserving, or changes to the ways in which the business and its risks are managed.
Thus, we go around the control cycle loop. This will be repeated continually.
For example:
(a) updating the pricing basis
For example, a deterioration in mortality experience generally could be mitigated by

raising premium rates in the future (for new business only, unless premiums on existing
policies are reviewable).
(b) revising product design
For example, an option might be removed if experience indicated that the option was
rarely used or if the mortality of those taking up the option was very high.
(c) changing the product mix / launching new products
The company may decide to stop selling unprofitable lines of business and launch new
products that have been successful for other companies.
(d) revising the underwriting processes
For example, the level of underwriting may need to increase if mortality rates are high for
products with a significant death benefit.
(e) revising reinsurance arrangements
For example, volatile mortality experience could be smoothed using reinsurance.
(f) implementing or improving retention activity
This might improve persistency experience, but will incur costs.
(g) changing the marketing message, target market and/or distribution channel
For example, an analysis that shows poor new business volumes in one particular product
may indicate that the advertising spend should be reviewed or that the product should be
marketed in a different way.

(h) revising sales procedures in terms of training and selection of distributors, wording
and format of sales literature and the mechanics of any commission payments and
clawback
Monitoring cashflows may indicate, for example, that one particular broker has a poor
administrative process resulting in lengthy delays in premium receipt. A tightening of the
premium collection procedures might help in this case.
(i) improving the wording of policy contracts
For example, if the level of complaints is high.
(j) improving the adequacy of staffing resources (numbers and competence)
A rapid increase in new business will require an increase in the number of staff.
(k) improving systems and data recording processes
Improvements may be required to update the sophistication of the analysis, eg so that

more rating factors can be analysed.
(l) improving actuarial models
For example, the rating factors used in a pricing model should reflect the factors that appear
to be most significant in the analysis.
(m) changing the investment strategy
Volatile investment surplus may indicate that assets are poorly matched to liabilities.
(n) changing the with-profits surplus distribution approach
Changes in the surplus emerging may lead to changes in the split between reversionary
and terminal bonus and the extent that the company wishes to smooth.
(o) updating the reserving basis
The reserving basis will normally be based loosely on past experience, subject to other
constraints such as regulation.
(p) raising capital
If the surplus arising is negative, then capital may need to be raised.
(q) altering the capital allocation methodology
If experience has been poor for a particular line of business then more capital may need
to be allocated to that line.
(r) improving risk management governance and controls.
The results of the various monitoring exercises will form the management information on
which key decisions are made. The results may even change how these decisions are
made, eg a risk management committee might meet more often if results become more
volatile.

In essence, this is an iterative process. The actuary is trying to estimate how the company
will progress in the future, based on what has happened in the past. Over time, the actuary
should be able to collect more information and therefore improve the modelling of risks.
However, historical data may become less relevant and there are likely to be unforeseen
changes that mean that the process of experience monitoring is a continuous one.
Although assumptions should gradually home in on reality in the case of a static world, in the real
world the ever changing nature of actual experience will prevent the actuary’s assumptions from
getting really close to reality. Thus our assumptions will normally need to contain margins.

Chapter 30 Summary
Experience will be monitored so as:
 to develop earned asset shares
 to update assumptions as to future experience, thereby feeding back into the control
cycle
 to monitor any trends in experience
 to monitor actual compared to expected experience and take corrective actions as
needed
 to provide management information to aid business decisions
 to make more informed decisions about pricing and about the adequacy of reserves.
A reasonable volume of stable consistent data will be required for experience investigations,
preferably of sufficient size to allow credible analysis down to the level of broad product
groups split by relevant risk factors.
Experience investigations would include:

 mortality
 persistency
 expenses
 investment return.
Expense investigations
Expense investigations aim to convert known total costs such as salaries, systems etc into
initial, renewal, termination and investment costs per policy. The allocation will need to
assume that some policy costs are invariant by policy premium size, some vary by premium
size and some vary by benefit size. Many assumptions, some rather arbitrary and pragmatic,
will be needed in allocating costs down to cells. (See also the two intermediate summaries
in the course notes.)
Other sources of information

Important and useful information on experience can also be obtained from analyses of
valuation surplus, and analyses of embedded value movement.



30.1 A life insurance company is investigating the lapse experience of a portfolio of without-profits
Exam style
term assurance business, in order to provide a basis for profit testing new business.
(i) List the variables that would be extracted from the database. [5]
(ii) List the factors that would be taken into account in determining the period to be covered
by the investigation. [5]
[Total 10]
30.2 Describe how the following expenses would be allocated for product pricing purposes:
Exam style
(i) The costs of obtaining medical evidence. [2]
(ii) The costs of running an investment department. [2]
(iii) The costs of purchasing and running a mainframe computer. [2]
(iv) The costs of Head Office property owned by the policyholders’ fund. [2]
[Total 8]
30.3 A life insurance company intends to monitor the experience of a cohort of term assurance
Exam style
policies.
Outline the analyses that would be carried out on an ongoing basis. [6]
30.4 Describe the principal steps in determining the initial expense loadings to be built into the
Exam style
premium rates for a new term assurance contract. [8]
30.5 A life insurance company writes only unit-linked endowment assurances.

Exam style
Describe how the expense assumptions would be determined for the supervisory valuation. [7]
30.6 (i) Describe a suitable model for estimating the probability of supervisory ruin of a life
Exam style
insurance company which is open to new business, explaining how the model would be
used to produce the required estimates. [9]
(ii) Describe how the company would determine the investment, demographic, persistency
and expense assumptions that it would need for the model, mentioning any investigations
that would be made. [13]
(iii) Explain how the model could be used as an aid to determining the company’s future
investment policy. [4]
[Total 26]
30.7 A proprietary life insurance company that currently transacts a range of conventional business, is
Exam style
proposing to enter the unit-linked life insurance market.
Describe the procedures that should be established to monitor and control the unit-linked
business. [10]

30.8 (i) State the main reasons a company would perform an analysis of embedded value
Exam style
profit. [2]
A life insurance company has just performed an analysis of embedded value profit and identified
that the withdrawal experience on its term assurance business was higher than its assumptions.
(ii) Discuss the likely impact of the higher withdrawals on its embedded value. [7]
[Total 9]

30.1 (i) Variables to extract
Basically the company wants to know the people who left, how they left and information to calculate
the exposed to risk. Normally data would be analysed by duration and sales method.
 date contract taken out
 date of leaving if they left
 reason for leaving
 sales method
 sales branch / agency.
The company may also want to analyse the data by factors other than duration and sales method.
Hence it may want:
 gender
 date of birth or age at entry
 premium amount
 frequency of payment
 method of collection
 original policy term. [½ per point, maximum 5]
(ii) Period to be covered
There are two conflicting factors here. Enough data in each ‘cell’ are needed to make the results
credible, but the company should not amalgamate so many different groups that heterogeneity is
introduced. [1]
This holds for the dimension of time, not only the other categorisations – so the company does not
want to choose a period so long that it includes intervals that might have significantly different lapse
experiences. [1]
Therefore factors to consider might be changes in:

 underwriting procedure
 policy terms or features added / removed
 sales method or training
 premium rates
 legislation, eg disclosure. [1 per point, maximum 4]
Should also consider analysing separately time periods that are affected by unusual external
economic conditions (eg recession). [1]
[Maximum 5]

30.2 This is based on a past exam question. When answering this question, bear in mind that:
 Expense allocation is not an exact science.
 Life companies do something pragmatic and sensible for a particular situation.
 There is no unique right answer, so in the exam say something that looks reasonable.
(i) Medical evidence
These costs form part of the underwriting expenses and so should be included as new business
expenses. [1]
There will be different costs associated with different product types, so the company should
investigate the differences between different classes, and load appropriately. [½]
Also, because the amount of extra medical evidence will increase as the sum assured increases, the
loading should depend on the level of sum assured. [1]
However the company might be using a modified policy fee method, in which case this loading is
automatic. [½]
[Maximum 2]
(ii) Investment department
Investment costs will depend upon the level of funds under management. Therefore the loading
should be a percentage of funds under management. [1]
Alternatively, this could be allowed for implicitly by assuming a lower investment return, ... [½]
… giving the same result if the reduction were the same for all products. [½]
This might be more appropriate if the investment is by a third party, since the company receives a
return after deduction of the third party’s charges. [1]
[Maximum 2]
(iii) Mainframe computer
The purchase cost would be spread over the estimated working lifetime of the computer. [½]
This would then be allocated to the products that will use the system. [½]
The system cost would also need to be allocated to new business, renewal or termination expenses,
depending on the system’s use. [½]
The annual costs would be allocated by taking one year’s expenses and allocating this across the
products and adding the loading to initial, renewal or termination expenses as appropriate. [½]
The split across product and by usage might be available from computer logs. [½]
Allowance should also be made for any anticipated cost savings associated with the new system. [½]
[Maximum 2]

(iv) Head office
The property will be an asset of the fund, but will contribute no direct rental as it is occupied by the
company. The expense (of loss of rental) has to be covered by policy expense loadings, as usual. So
we need to determine a notional rent for the property. [1]
The notional rent would be divided among departments according to property usage, probably
based upon the floor space occupied, or on number of staff. [1]
This rental income would then be allocated to products and then allocated to new business,
renewals or terminations. This would probably be in the same proportion as each department’s
salary costs are allocated to products and expense type (based on timesheet analysis). [1]
[Maximum 2]
30.3 Having to cover all the possible analyses for 6 marks is a pretty tough assignment. This means
that you must confine most of your effort to the three key assumptions for this contract: mortality,
lapses and expenses. You must also be brief, which is in accordance with the command verb
‘Outline’. Key areas to think about when describing an investigation are the data (including the
period covered) and how the data will be split down.
The company will probably analyse the data annually, assuming sufficient for credibility.
Cumulative analyses may also be conducted. [1]
Mortality
Compare claims (numbers and amounts) with those allowed for in the premium basis. [1]
Split analysis by:

 age and gender
 smoker / non-smoker
 any other rating factors used (data permitting). [1½]
Lapses
Need to analyse by duration, since rates tend to vary by duration, especially early on. [1]
Divide lapses by relevant exposed-to-risk and compare with allowance in premiums. [1]
If data permit, split by sales channel. [½]
Expenses
Company expense allowance should apportion an appropriate amount of expenses to term

assurances. [½]
Need to distinguish between expenses which are per premium, per sum assured and per policy.
[1]
Also need to split initial, renewal and termination expenses. (Perhaps also investment expenses,
but reserves are small.) [1]

Compare with premium basis expense loadings, adjusted for assumed inflation. [1]
[Maximum 6]
30.4 The key ideas are:
The aim is to allocate all the company costs to policies ...
... so investigate experience over last few years ...
... making sure all costs are allocated once and once only.
In this case we are particularly interested in new business costs ...
... eg underwriting, sales and marketing, policy documentation.
Split out new business costs as much as possible ...
... some will be easy eg cost of sales literature ...
... for others use something sensible, eg split property cost by a notional rent, …
... split salaries by time spent.
Split the new business costs by product ...
... and allocate costs according to the product ...
... some will have higher underwriting ...
... and others might have higher documentation costs.
Some overhead costs might still not have been allocated …
... make sure they are allocated by a reasonable method ...
... take into account future levels of new business ...
... probably per-policy loading.
One-off costs would be treated separately ...
... maybe ignore ...
... or spread over a suitable number of years.
Any commission payments would be known and loaded explicitly ...
... also could be advertising ...
... or other specific costs to add in.
These expenses should be inflated to the midpoint of the period for which premium rates are
expected to be in force.

30.5 Expense assumptions
The aim of the exercise is to estimate the per-policy costs of administering this class of business.
[½]
Expenses are split into a number of categories, such as: initial, renewal, claim and investment.
(An initial expense assumption would not be needed for the valuation.) [1½]
The company will investigate the recent experience and try to allocate all its expenses, firstly by
policy type and then into one of these four groups. [½]
Exceptional costs may be ignored. [½]
Commission will be treated separately and loaded explicitly (although, again, up-front commission
is not relevant to the valuation assumptions). [½]
Similarly investment costs may be treated separately, especially if an external manager is used. [1]
The costs should be directly allocated to the different products, as far as possible by timesheets,
functional costing or some other method of apportionment. [1]
The expenses will be expressed as a proportion of premium, sum assured, unit-fund or per policy
as best reflects their behaviour. [1]
Fixed costs may be spread more approximately, such as by floor space or number of staff.
Alternatively they may be added as a percentage of the expense loading. [1]
This will give the basic loadings, which may need to be adjusted for prudence (depending on the
valuation regulations). [½]
The company also should consider how expenses will inflate in the future, on a prudent basis if
necessary. [1]
If there is an inflating administration fee this might reduce the need for any margins. [½]
[Maximum 7]
30.6 (i) A model for estimating supervisory ruin
Determine a suitable projection period over which to model. Given that the model will include a
projection of future new business, which is highly problematic, a five- to ten-year horizon would
probably be appropriate. [1]
Projections may be done on a policy by policy basis using the actual in-force policy data. [½]
Or model points could be used. These must reflect, by type and volume, the actual distribution of
the in-force policies of the company. [1]
Assume an appropriate new business model. Each assumed future year’s new business must
reflect, by type and volume, the expected distribution of policies to be issued by the company. [1]
Assume an appropriate existing asset distribution, reflecting the company’s existing asset mix. [½]

Assume a future investment strategy which is consistent with the company’s stated strategy and
its past behaviour in this respect. [1]
Different asset types will need to be modelled separately. [½]
Use a stochastic model for the investment return variables. [½]
May need to model other variables stochastically (eg mortality), depending on the nature of the
business being modelled. [½]
Ensure that the model reflects realistically the effect of all of its key parameters. [½]
Ensure that all significant dynamic links are included in the model, both between different asset
classes and between the assets and the liabilities. For example, the projected supervisory reserve
assumptions should be sensitive to the relevant projected investment conditions; the projected
investment strategy should be sensitive to projected level of the free assets; etc. [1½]
Project the future in force liabilities (ie policies) on the basis of the assumed demographic
experience (including new business). [½]
Project future cashflows, including investment returns, and hence the accumulated assets to the
end of each future projection year. [½]
Calculate the supervisory value of the liabilities at the end of each projection year, including any
required amount of solvency capital. [1]
As a full projection of future solvency might require nested stochastic calculations, the company
may adopt a simplified approach, eg a closed form solution or proxy model. [1]
Calculate the value of the statutory free assets, ie the difference between the projected assets
and the supervisory value of the liabilities at the end of each projection year. [½]
Running the model
Make a projection of the future statutory free assets at the end of each year using a randomly
generated set of future scenarios. [½]
Identify whether the company remains solvent in all future projection years, or the company
becomes insolvent in at least one projection year. [½]
Perform many such random simulations (eg 10,000). [½]
Calculate the estimate of the ruin probability as the ratio:
Number of simulations in which at least one insolvency occurred

[½]
Total number of simulations
[Maximum 9]

(ii) Determining the assumptions
The assumptions should be best estimates. [½]
Investment returns
Best estimates will be needed for the expected future returns on each asset type, … [½]
… the variability of those returns, … [½]
… and also the correlation between the returns from different asset classes. [½]
These would be estimated from historical experience of these asset classes, bearing in mind the
economic conditions of the time. [½]
The historical estimates may need to be amended depending upon the view taken of the future
economy and the expected behaviour of the relevant investment markets. [½]
Mortality
An investigation of recent experience would be made. [½]
Different policy classes would be analysed separately (as separate assumptions will be required
for different classes). [½]
Classes with relatively little data (or none, if these are new products expected to be issued in the
future) will have to be estimated from other sources: eg from the experience of any similar
classes; from industry experience; from reinsurer’s data; and (possibly) population data. [1]
Data would be analysed in homogeneous groups to the extent that the volume of data allowed
(eg gender, age, duration, sales channel). [1]
The historical estimates would be adjusted in line with any reasonable continuation of ongoing
trends, for each class separately. [½]
Planned changes to target markets, and to underwriting levels, would need to be allowed for in
the mortality assumptions used for new business. [½]
Allow for any expected effects of changes in claim control in the future. [½]
Consider effect of changes in the future economic environment on the mortality experience. [½]
Any other expected future changes would be allowed for (eg due to improvements in general
medical care in the population). [½]
Allow for expected effects of selective withdrawals in the future. [½]
Persistency rates
An analysis of the recent experience would be made, taking care over the choice of time period
used. [½]

The analysis would consider full withdrawal, and if appropriate also partial (or income)
withdrawal and conversion to paid up. [1]
Analysis would be performed for homogeneous groups, eg by policy class, sales channel, policy
duration, possibly age, etc. [1]
Historical one-off internal and external events would be taken into account. [½]
Future changes would be incorporated into the projection, eg due to changes in the economy,
changes in sales channel or sales practice in the future, or to changes in the target market. [1]
Expenses
An expense analysis would be performed, usually in respect of the last year. [½]
Expenses would be identified separately by policy class to the extent that they lead to different
costs. [½]
It would be necessary to identify separately those expenses which were due to acquisition,
renewal, termination and investment, … [½]
… those which were variable or fixed, … [½]
… and of the variable expenses those which can be associated with premium size, benefit size, or
per policy. [½]
Future expenses would be projected with allowance for future inflation, with the individual
components modelled appropriately (eg per-policy acquisition expenses modelled as a function of
the projected new business, etc). [½]
Fixed expenses should be modelled as a global item, which would (normally) be independent of
the number of policies issued. [½]
Future changes to expenses should be allowed for; eg due to company restructuring, taking on or
reducing staff levels, new branch networks, etc. [1]
One-off costs (eg due to buying new computer equipment or special marketing exercise) would be
separated from the main experience. [½]
Costs of future marketing campaigns, product developments, technology improvements, etc,

should be allowed for in the projection. [½]
[Maximum 13]
(iii) Determining investment policy
Estimates of the probability of supervisory ruin can be made on the basis of different future
investment strategies for the assumed liability profile (including new business). [1]
To be effective any strategy assumed needs to be responsive to the nature of the liabilities
projected to be in force, and to the projected level of future statutory free assets. [1]

A strategy can be chosen which gives an acceptably low probability of statutory ruin for a given
liability profile. [½]
The sensitivity of the ruin probability to the projected assumptions would need to be tested. [½]
This would include sensitivity to the investment model assumptions, in particular, but also to the
other demographic and financial assumptions. [½]
It would be essential to see how the ruin probability varied with different future liability profiles,
which could be projected using a range of new business assumptions. [½]
In order to establish an optimal strategy it would be necessary to use a summary measure of

shareholder value (or some suitable alternative in the case of a mutual company), for example the
discounted value of the future shareholder dividend stream. [1]
A strategy could then be found which maximised shareholder returns while keeping the ruin
probability at an acceptably low level. [1]
[Maximum 4]
30.7 Monitoring the experience
An analysis of the company’s experience would need to be carried out at least annually. [½]
Analyses would be performed for expenses, mortality, persistency, investment performance, and
new business volume and mix. [1]
Initially it will be essential to analyse the unit-linked policy experience separately from
conventional business; as the volume of business grows then it will be necessary to analyse
different subclasses of unit-linked policy separately. [1]
Comparison would be made between the actual experience, and that expected according to the
assumptions used in the company’s initial pricing models. [½]
Divergences between actual and expected would be investigated, to locate the reason for the
good or bad experience. [½]
Management would attempt to take remedial action to improve adverse experience, where this is
identified, and try to promote examples of good practice more widely. [½]
An analysis of surplus would be required to show the immediate financial impact of the various
sources of divergent experience on the company’s profits. [1]
This will provide key information to management, which will be able to focus most effort in
dealing with deviations that have the greatest financial consequences to the company. [½]
The company should ensure that its procedures are able to identify any developing trends in
experience over time. [½]

Feedback into the company’s models
Allowing for the effect of any proposed and/or implemented corrective actions, and allowing for
any expected continuation of past trends, the company may need to revise its assumptions for
future experience in its models. [1]
The company will be able to see the effect of any revised assumptions on its embedded value, so
that the impact on the shareholders’ interest can be computed. This could trigger further actions.
[1]
The pricing models will need to be rerun (including sensitivity testing) using the new assumptions,
to ensure that the current charging rates are still profitable. [1]
If and when it becomes necessary to revise charging rates, the new charges might be obtained
directly from the results of the experience analyses, especially if the charging structure is
designed to produce close matching between charges and costs. [1]
The adequacy of premiums under policies that have regular premium reviews will need to be
investigated as and when necessary. [½]
Monitoring the general commercial and economic environment
The achieved new business volumes and mix, and unit fund growth rates, would be compared
with the company’s targets and with competitors. [1]
The market would be studied carefully to ensure that the products are proving competitive, meet
customer needs, and are being sold effectively (through the distribution channel(s)). [1]
The results of expected future experience, any expected changes to the market and any proposed
changes in the product can be investigated using the company’s full model office projections. [1]
The progress of the company’s embedded value can be projected, using a range of possible future
new business (and other) assumptions, and the effect on the company’s ongoing available capital
(and hence supervisory solvency) can also be checked. Further decisions regarding the
management of the unit-linked business might follow these investigations. [1]
Reference to the control cycle is a useful way of answering this question.

[Maximum 10]

30.8 (i) Reasons for performing an analysis of embedded value profit
The main reasons for analysing embedded value profit are to:
 validate the calculations, assumptions and data used in the calculation of the embedded
value [½]
 reconcile the values for successive years [½]
 provide management information, eg to identify profitable areas and aspects of the
business, and thereby inform decisions [½]
 provide data for use in executive remuneration schemes [½]
 provide detailed information for publication in the company’s accounts, in particular the
value of the new business taken on by the company. [1]
[Maximum 2]
(ii) Impact of higher withdrawals on the embedded value
When a withdrawal occurs, the embedded value reduces due to the loss of future profits.
However, the reserve for the policy no longer needs to be held, so the embedded value will
increase by the amount of reserve released. [1]
There is no cash payment made on the withdrawal itself, so this does not influence the embedded
value. [½]
Whether a withdrawal causes an increase or decrease in embedded value therefore depends on

how the reserve value compares with the expected value of the future profits at the time of
withdrawal. [½]
Early in the policy term, reserves for term assurance business tend to be relatively small (and may
even be negative). [½]
This is because, with the majority of the policy term still remaining, the level future premiums are
close in value to the remaining expected future claims and expenses. [1]
Most of the future profit on the policy is still to be made, however, so the expected present value
of the future profits is high. Therefore a withdrawal at this time will generally produce a
reduction in embedded value. [1]
Later in the policy term, reserves are higher, because

 future claims will be higher on average in each remaining future year of the policy due to
the older age of the policyholder [½]
 average expense will be higher due to inflation [½]
 the average premium is unchanged. [½]
For the same reasons, the expected present value of the future profits will be low. [½]
The result is that a withdrawal at this time will generally increase the embedded value. [½]

We would also need to consider whether the embedded value experience assumptions should be
changed in the light of the recent experience. [½]
If the future rates of withdrawal are increased, in the year of change this will:
 reduce the embedded value at early policy durations [½]
 increase the embedded value at later policy durations [½]
for similar reasons to those discussed above.
A higher withdrawal rate assumption may also imply a greater selective effect on the mortality of
the continuing policyholders, so that the future assumed mortality rates will need to be
increased. [1]
This will reduce the expected future profits and therefore the embedded value also. The financial
impact of this could be considerably greater than the direct effect of the withdrawals themselves.
[1]
Higher withdrawals could imply a higher per-policy expense cost, as overheads will be spread
more thinly. This would also reduce the embedded value. [1]
Reserves may need to be strengthened to reflect these changes in expected future experience.
This will delay the emergence of the expected future profits and again be likely to reduce the
embedded value (assuming that the discount rate is higher than the assumed investment return).
[1]
As a result of the above assumption changes, the increase in embedded value arising from the
sale of new business during the year would also be reduced. [½]
[Maximum 7]

SP2-31: Problem solving Page 1
Problem solving
Syllabus objectives
6.1 Analyse hypothetical examples and scenarios in relation to the financial
management of life insurance companies.
6.1.1 Propose solutions and actions that are appropriate to the given context,
with justification where required.
6.1.2 Suggest possible reasons why certain actions have been chosen.
6.1.3 Assess the implications of actions within a given scenario.
6.1.4 Discuss the advantages and disadvantages of suggested actions, taking into
account different perspectives.

Page 2 SP2-31: Problem solving
0 Introduction
The examiners will expect candidates to be able to apply the knowledge and understanding
they have developed through the study of the Subject SP2 Core Reading to produce
coherent solutions and actions in relation to the financial management of a life insurance
company.
In particular students are expected to be able to combine ideas across the chapters in the
Core Reading, and apply them to scenarios proposed by the examiners.
From your studies of the earlier chapters you should now have a firm grasp of all of the issues
central to the financial management of a life insurance company. However, the course has had to
introduce these in a relatively isolated fashion. But, as the Core Reading above makes clear, you
may be asked in the exam to combine these issues together to solve the kinds of problems that
actuaries face in the financial management of a life insurance company.
In this chapter we take a look at the financial mechanics of the life insurance company,
considering how the different parts of the course come together. The material should not be
new, but it should represent some new perspectives. Thinking through the ideas in this chapter
will help to generate ideas in the exam, particularly for questions that cover a wide range of
topics.
We look at:
 a life insurance company income statement and balance sheet
 the lifetime of a policy
 the interested parties in the life insurance world
 the control cycle revisited.
The chapter then finishes with some longer practice questions to help you develop your problem
solving skills.
As this chapter discusses exam technique rather than adds new technical material, it does not
have a summary section.

1 Life insurance company accounts

One approach that can be useful to generate ideas is to consider each element of the accounts in
turn.
To ensure you understand fully where life insurance company profits come from, we consider
what a life insurance company income statement and balance sheet normally look like. The exact
form that they take will depend on local statute and practice. For instance, it is common for the
operation of the policyholders’ funds to be dealt with in a revenue account, and for profit from
that account to then pass through to the income statement (or ‘profit and loss’ account), which
adds in the impact of shareholders’ assets to give a final result for the company. Here we assume
that everything ‘happens’ in the revenue account.
1.1 The revenue account

A life insurance company revenue account looks something like:
Premiums received positive

Investment income positive
Increase in reserves negative
Benefit payments negative
Expenses negative
Commission negative
Profit / loss (pre-tax)
Tax payable negative
Profit / loss (post-tax)
Dividends to shareholders negative
Transfer to balance sheet ‘shareholders’ retained profit’
This should be intuitively very familiar. One useful application is as a way of generating ideas
about the financial impact of any action, both in terms of the individual elements that might be
affected and the ‘bottom line’ effect on profit.
Question
Describe how the sale of a new policy will affect the above items.

Solution
Writing a new policy will have the following impact:
Written premiums increase to the extent of the premium concerned (or slightly less, if we are
considering a policy sold during the year with premiums payable eg monthly).
Investment income will not change significantly: on average there will be a half-year’s interest on
the premium less expenses and benefit outgo.
The reserves will increase by some fraction (typically 85-95%) of the premium. If the statutory
valuation basis is stronger than the realistic assumption basis used in pricing (or if there is a
solvency capital requirement), the company will suffer valuation strain and the proportion
‘increase in reserves / solvency capital as a percentage of premium’ will generally be higher.
Expenses will increase, due to the underwriting, marketing and administration costs involved.
This question is about the marginal effect of selling one policy: thus the expenses from overheads
which we need to attribute to policies in other contexts (eg pricing) are not relevant here.
Commission will increase if the policy was sold through commission-rewarded brokers / agents.
Benefit payments are likely to increase slightly, but not significantly (unless the policy were a
contract of duration one year or less – which is unlikely).
The net effect is likely to be a loss (unless there was some financial reinsurance arrangement in
place).
In this case it is almost certain that no tax will be incurred in respect of this (considering the two
common tax bases, profits and ‘investment income less expenses’).
Total shareholder profits will therefore be reduced by the sale of the policy. This means reduced
from what they otherwise would have been; it does not necessarily mean reduced compared with
the previous year.
The account shown here gives the basic idea. Many variations are possible. You might find:
 There might be separate entries for ‘increase in reserves’ and ‘increase in required
solvency capital’, or the required solvency capital may be excluded from the accounts
(depending on the purpose).
 Instead of ‘increase in reserves’, there could be two entries: ‘reserves brought forward at
start of year’ and ‘reserves carried forward at end of year’. A similar split could then be
made for required solvency capital as well.
 Commission might be included in expenses (or there may be no commission at all).
 Benefit payments might be split into their components.
Question
List the possible components of the benefit payment entry.

Solution
The possible components of benefit payments will be:

 maturities
 death claims
 surrenders
 cash bonuses
 annuity instalments.
‘Waivers of premium’ (where a policy allows the non-payment of premiums under certain
conditions) might fall into this category.
 Investment income might be split between policyholders’ fund income and income on
shareholders’ assets.
 Investment income might or might not include unrealised capital gains.
 Reinsurance might be implicit in the items (so ‘premiums received’ is really ‘gross
premiums received’ less ‘premiums ceded under reinsurance’) or it might be split out.
 Surrenders and lapses might be implicit in the items (so ‘commission’ is really ‘commission
less commission recovered from early discontinuance’) or they might be split out.
One consequence of these possible variations is that, if an exam question presents any form of
revenue account, or talks about some revenue account item such as ‘benefit payments’, you need
to be aware of what each item really means. If not made clear, and it is important, you need to
specify your assumption (for instance, whether ‘profit’ is post-tax or pre-tax), or to consider both
cases. Areas to watch out for are:
 Is the figure net or gross of reinsurance?
 Is the figure net or gross of discontinuances?
 Does the figure relate to the policyholders’ fund only, or the whole company?
 What about tax?
Question
Explain how the cashflow of the company will differ from the revenue account.

Solution
At its simplest, the cashflow will be the total of all of the items excluding the increase (or
decrease) in reserves (and required solvency capital). This is because the increase in reserves is
really just a reallocation of monies inside the company, while all of the other items involve real
cash being paid to or by the company.
If accounts are prepared with assets valued at market value, then the revenue account may
include unrealised capital appreciation / depreciation but the corresponding amounts will not
feature in the cashflow.
It is also possible for the expense figure in the revenue account to include an allowance for
depreciation, while there will be no corresponding cash movement in that year (there will in fact
have been a cash payment in respect of that amount in some previous year).
Tax shown in the revenue account will be the amount that is incurred or due. Depending on
when tax is collected, this may not be the same as the tax actually paid during the period, which is
what will impact cashflow.
So the cashflow is:

premiums
+ investment income (excluding unrealised gains / losses)
– benefits
– actual expenses (ie excluding depreciation / amortisation)
– commission
– tax paid
1.2 The balance sheet

The presentation of life insurance company balance sheets seems to vary even more than that of
revenue accounts. In essence, they are extremely simple, with total assets equal to total liabilities
(including shareholders’ capital and retained profits).
We now consider the relationship between the revenue account and balance sheet. We approach
this by means of an example.
The balance sheet for Abingdon Life insurance company as at 31 Dec 2016 is:
Assets (£m) Liabilities (£m)

Share capital / retained profit 45
Reserves 114
Total assets 159 Total liabilities 159

The revenue account for the year 2017 is as follows:
Premiums written 46
Investment income 18
Increase in reserves (41)
Benefit payments (8)
Expenses (3)
Commission (6)
Profit / loss (pre-tax) 6
Tax payable (2)
Profit / loss (post-tax) 4
Dividends to shareholders (1)
Transfer to balance sheet

‘shareholders’ retained profit’ 3
Question
Calculate the 31 Dec 2017 balance sheet, and the company’s free assets at that date.
Solution
The balance sheet will be:
Assets (£m) Liabilities (£m)

Share capital / retained profit 48
Reserves 155
Total assets 203 Total liabilities 203
The mechanics are as follows:

 Share capital increases by the retained profit transferred to balance sheet of 3.
 Reserves increase by the increase in reserves.
 Total assets will have varied in accordance with the cashflow:
159 + 46 + 18 – 8 – 3 – 6 – 2 – 1 = 203
 Free assets = total assets less non-shareholder liabilities = 203 – 155 = 48
 Alternatively, free assets = shareholders’ capital = 48
 Check that the two sides balance: 203 = 48 + 155

1.3 Solvency capital requirement

In all of the above, how would any solvency capital requirement come into play? For convenience
we’ll call it the SCR.
The company might build up the SCR via the revenue account in the same way as it builds up
reserves. This is the approach that we would need to take in profit testing / model office work.
However, it is often the case that the SCR is not seen in any way in the revenue account.
Even in the balance sheet, where we might reasonably expect to see the SCR, it is common for it
not to be shown explicitly in the published accounts. However, the company will need to
demonstrate to the supervisory authority that, given the balance sheet as at a certain date, the
SCR as prescribed by statute can be covered by shareholders’ capital and retained profit (so we
would expect to see the SCR in the supervisory balance sheet).
For instance, in the example of Abingdon Life as at 31 December 2016, the company might have
needed to show that the figure of £45m for shareholders’ capital and retained profit exceeded
the SCR at that date of £6m (say).
So we may or may not see the SCR in either the revenue account or the balance sheet. The
approach used will depend on the purpose of the calculation (eg reporting profit to shareholders
or demonstrating solvency to the supervisory authorities) and the regulations that apply in that
country.
Question
Describe how the sale of a new term assurance policy will affect the balance sheet at the end of
the year.
Solution
Writing a new policy will have the following impact:
The assets will be increased by the premium coming in. However, term assurances have relatively
small premiums as there is no maturity benefit. Also, most term assurances are regular premium.
So the assets will be increased by only a small amount.
The assets will be reduced by the initial expenses and any commission paid. Underwriting
expenses could be very significant as the policy has a large sum at risk. Initial expenses of writing
the contract will also be large compared to the relatively small regular premium. So overall, the
expenses and commission will outweigh the premiums and the assets will fall.
Any claims that occur before the end of the year will reduce assets further. The contract will have
provided no assets to contribute any investment return at this stage (in fact the asset share is
negative, so any interest attributed to the contract would be negative).

If reserves are calculated on a prudent basis, then reserves are likely to increase following the sale
of the contract. The insurer may also be required to hold additional solvency capital to cover the
increased risk from selling this contract. So the total of reserves and required solvency capital will
rise.
So overall, the balance sheet will show a fall in assets and a rise in reserves / solvency capital.
Hence the shareholder capital will fall (as the assets must equal the liabilities, ie the balance sheet
must balance).
This is consistent with the question at the start of this section that showed that writing new
business could result in a loss on the revenue account.
The first and last questions in this section indicate that considering the elements of the accounts
in turn can generate a good description of the impact, in this case of writing new business, on the
insurance company.

2 The life and times of a policy

A useful way for thinking about how a life insurance company functions is to consider all of the
stages involved in the lifetime of a policy. This can also be a good way to generate ideas in the
exam.
We break it down as follows:

 the idea
 the sale
 the first day
 a typical year
 the last day
 even later days.
2.1 The idea

It can be argued that the lifetime of a policy starts with the product itself being created. This will
reflect:
 some perceived market need for the product
 a regulatory and taxation environment which makes the product both possible and
practical
 the company wanting to write such a product.
Once the company decides that it wants to write the product, the next stage will be product
design.
Question
Outline the main steps in product design and pricing.

Solution
The main steps in product design and pricing will be:

 formulate the structure of the product (with marketing)
 decide on assumptions for expenses, interest, mortality, persistency
 profit test to give premiums to meet set profitability target
 given those premiums, sensitivity test; need to adjust premiums or structure?
 compare with competition, adjust premiums? (especially expense loadings / charges)
 run model office with expected tranche of new business – are there any capital, tax or
expense problems? Is it worthwhile in terms of impact on embedded value?
2.2 The sale

Our policy now needs to be written. Or, as a marketing manager might say, ‘the policy needs to
be sold’.
The sale process will depend greatly on whether the sales channel is direct, where members of
the public are approached, or whether the initiative is taken by the public – eg they decide to go
and ask a broker about a term insurance policy. Looking at the agent / broker sales process:
The first thing will be when some innocent member of the public, Person X, thinks that they might
need a policy of a particular type. This might be pre-empted by advertising by our company, or
more likely by any other life insurance company, about such a contract. The other main avenue
would be if our prospective policyholder is doing something which entails getting insurance –
eg much term assurance will be written as a result of people taking out mortgages on house
purchase, for example.
X then decides to do something about their whim. They visit a high street broker, or the
insurance adviser at their bank or building society.
Suppose the high street broker works with a number of life insurance companies (ie they are
‘independent’ rather than ‘tied’). In that case the broker will ask X about their finances and
requirements, decide on which companies’ products are most suitable, and show X the relevant
marketing literature from those companies. The broker might indicate just one contract, which
they judge to be the most suitable for X – being careful not to be influenced by any commission
which they might receive on sale of the contract.
The broker could also produce a ‘quote’ – ie the premiums and ensuing projected benefits, and
possibly surrender values and overall yield – given X’s age and gender, and intended premium (or
required benefit). The quote could be done on the broker’s PC using software supplied by the
companies. Alternatively, there might be a web-based quotation system. In some markets, rate
tables are still used, ie the broker would calculate the premium by hand using a printed copy of
the insurer’s rates.

Question
Outline how the above procedure would differ if X had approached their bank’s insurance adviser,
and the bank has its own insurance subsidiary.
Solution
The adviser would recommend one of the bank’s products (unless there was absolutely nothing of
the right type). The quote would be run off from the bank’s computer system, or from a PC.
Assume that sensible Person X decides to take out our company’s product. Assume also that it is
a with-profits product.
First, X will need to fill in a proposal form.
Question
Outline the information required on the proposal form.
Solution
Personal information such as name, address.
Policy information such as premium and benefit intended, and any options.
Medical information such as smoking and drinking habits, current or recent illnesses, life history
of major illnesses.
Information for financial underwriting such as ‘do you have any other life policies?’.
Bank details if payment is to be by automatic deduction from account.
Next, the proposal form goes to the life insurance company. A large company might have regional
branches that would be able to deal with this.
If the answers to the medical questions raise any doubts, or the sum assured is over a certain
level, some form of extra medical evidence might be required.
Question
List the extra medical evidence that might be asked for.

Solution
The extra medical evidence could be:

 a report from X’s regular doctor
 a general medical examination
 specific medical tests such as ECG.
The company would also consider the financial condition of X to ensure that the policy was
normal given their probable needs and lifestyle.
For a whole life or endowment assurance contract, if the sum assured was particularly high the
life insurance company might need to contact its reinsurer to ensure that it was appropriately
covered.
If there are no problems, and the proposal is accepted by the company, the proposal data will be
fed into the life insurance company’s policy administration computer system and the system
should print out the policy. X will then be sent the policy (a life insurance policy is in reality a legal
contract printed on paper, signed by the policyholder and the company) and cover will commence
on the agreed date, probably subject to satisfactory payment of premium.
(The premium might be paid to the broker, rather than direct to the life insurance company.)
Question
Describe the main differences to the above process if Person X had taken out the policy via a
direct marketing telesales channel.
Solution
X sees an advertisement in the paper and rings the number.
They are given a brief description of the product on the phone, and then need to give their
address so that the company can send an information pack.
This would include information about the product, about the company and might include quotes
for X’s age / gender / premium if that was sorted out in the telephone call.
The pack might also include a proposal form for X to fill in and send off. This might include some
underwriting questions.
On receipt of this, and possibly conditional on appropriate payment and responses to

underwriting questions, the company would produce the policy and send it to X.

2.3 The first day

Here we shall consider what happens to the company’s finances at inception of the policy.
On cover commencing, the policy becomes a liability of the life insurance company. The company
therefore needs to establish suitable reserves (and if there is a solvency capital requirement, then
this will also need to be considered). The reserves may be calculated on a more prudent basis
than the premium basis (and/or the requirement to set aside solvency capital will build in a level
of prudence). If this is the case then the company incurs some valuation strain at inception –
ie the cost of reserving for the policy is greater than that assuming the (realistic) premium basis.
The company must also pay the acquisition costs of the policy. In the example of the broker, this
would probably entail the company’s agency account department crediting the broker’s account
with the commission amount (alternatively the broker may receive a fee directly from the
policyholder) – which might be something like 50-120% of premium for a regular premium
contract, 3-6% for a single premium contract, depending on the market, the company, the
product and the broker. The broker will then receive actual cash when the account is settled,
which might be done, say, monthly.
In addition the policy will have cost a certain amount to set up on the company’s administration
system, to administer the proposal / policy and to cash and invest the premium.
The above expenses (including commission) are the initial marginal expenses of the company in
respect of the policy, and it is vital that the expense loadings or charges in the contract cover
these.
However it is normal to expect every policy to contribute something to cover the other expenses
of the company – for instance the overheads of the company, or the development costs of the
contract. This contribution would normally be expressed per policy, or possibly in relation to
premium size. This will go into the ‘initial expenses per policy for contract Y = £270’ figure which
we saw coming out of the expense investigation – the product should be priced on the basis of
these suggested loadings. But the contract’s loadings might not be equal to these suggested
expenses. Perhaps because the contract was priced two years ago and the expense experience of
the company has changed since then. Or because the balance between ‘fixed per-policy
expenses’ and ‘percentage of premium’ expenses has been adjusted for competitive reasons.
So the company might make an expense profit or loss if the required contribution to current
expenses is lower or greater than that assumed in the premium basis.
Moving on from considering the expenses in isolation, what is the overall impact on the
company’s financial results?

Considering the income statement earlier, for an annual premium contract the company might
see on the first day an effect along the lines of:
Premium P
Increase in reserves 0.95  P
Expenses 0.6  P
Commission 0.8  P
Profit / loss –1.35  P
So writing that policy will have required the company to find capital of 1.35  P. That money will
have been provided by the free assets of the company.
So Person X has been fortunate: the company has been prepared to suffer an income loss of 1.35
times the premium paid so that X could have their favourite insurance policy. (Remember, this
has assumed a full expense figure including a contribution to overheads, rather than the marginal
expense involved.)
In fact, the company would need to find even more capital than this if there is some solvency
capital requirement (any weakness in the reserving basis is likely to be offset by stronger capital
requirements). For instance, if there is a requirement of 0.05P in respect of the policy then the
company has to find capital of 1.4P to write that business.
So the revenue account impact is: a loss. This will often be the case for a regular premium policy.
A single premium policy might give some loss, primarily due to valuation strain, but the size of loss
(in relation to the premium) would be much lower than that seen with a regular premium
contract.
Question
Outline the embedded value impact of writing the policy.
Solution
If the policy is profitable on the embedded value basis, the embedded value will increase.
This will normally come from a large increase in the present value of future profits and a small
decrease in free assets.

2.4 A typical year

Several years have passed since that exciting first day. What might happen in a typical uneventful
year?
Consider the financial impact of the policy on the revenue account of the company. At the start
of the year the policy will be ‘visible’ in the balance sheet, via its associated investments on the
one hand and reserves and shareholders’ retained profit on the other hand.
During the year X faithfully pays the premium. At each policy anniversary, or perhaps at
31 December, the company may increase X’s benefits from any surplus arising. That surplus will
be the result of experience having been better than that assumed in the valuation basis.
Question
The premiums coming in every year are calculated on the premium basis, so it might be thought
that the surplus arising for distribution should be the result of experience being better than that
assumed in the premium basis.
Explain why this is not the case.
Solution
As soon as the company receives a premium, the company has to set up the necessary reserves
for the policy concerned (and be able to cover any solvency capital required). For simplicity, we’ll
consider the case where the valuation basis contains margins for prudence and there is no
requirement to hold solvency capital. As the valuation basis is more prudent than the premium
basis (usually) the company will record an immediate loss. This is the capitalisation of all the
future losses the company is now expecting to make, from a policy whose premiums are
inadequate to meet its liabilities according to the reserving basis assumptions of the future
experience. The basis underlying that premium becomes irrelevant from this point: having set up
these reserves, the company will not make any (more) loss, or profit, if its experience is exactly
that of the reserving basis. Hence if the actual experience is better than its reserving basis, then a
profit will be recorded in the company’s accounts.
Assuming that the policy is a regular premium with-profits policy, how would its passage through
the year be reflected in the revenue account?

Stopping the accountants one day before bonus distribution, we might see something like:
Premium P
Investment income 0.4  P
Increase in reserves (1.2  P)
Benefits (0.01  P)
Expenses (0.04  P)
Renewal commission (0.05  P)
Surplus 0.1  P
The investment income of 0.4P may seem high, but it is earned off the back of the reserves on
that policy, which now total 6P (say). The increase in reserves will reflect both another year’s
premium, and the effect of the valuation rate of interest on the existing reserves.
The life insurance company’s venerable actuary now needs to decide how much of that surplus to
distribute to the with-profits policyholders, and how. This assumes that the surplus distribution
method permits some form of judgmental / subjective input from the actuary. In this example
the statutory regime (or it might be the policy conditions) requires that 90% of the surplus is
distributed to policyholders, 10% to shareholders. The distribution to policyholders is done by
increasing the benefit such that:
reserves after bonus = reserves pre-bonus + policyholders’ share of surplus
So the revenue account would continue:
Surplus 0.1  P
Increase in reserves
due to bonus (0.09  P)
Profit 0.01  P
Tax (0.004  P)
Profit net of tax 0.006  P
(Assuming that tax is paid on profits at 40%.)
So the shareholders make a small profit. In fact the original pricing of the contract should have
been such as to ensure that the profits that emerge in subsequent years are sufficient to repay
the initial new business strain, valuing at the risk discount rate.
What is the effect on embedded value? We saw in Chapter 30 how, over the course of the year,
the present value of future profits from this policy will decrease as profits flow through the
revenue account into free assets.

Question
Describe how the embedded value will move over the year, if experience is worse than that in the
embedded value projection basis.
Solution
If experience is worse than expected then there will be two effects:
The free assets will not increase by as much as expected. (This could equally appear as the free
assets decreasing by more than expected, or the free assets decreasing when an increase was
expected.) This corresponds to the immediate profit consequences of this experience (ie less
profit – or more loss – than expected).
The PVFP might be decreased. In the case of decrements (mortality and withdrawals) their PVFP
will no longer contain the future profits of those ‘extra’ policies which have gone off the books.
Both of these will lead to a reduction in embedded value. The effect on free assets will normally
be much greater than the effect on PVFP. However, there might be instances where the opposite
is true (eg if the bad experience in question is a high medium- to late-term lapse rate for term
assurance policies).
Furthermore, if the experience then moves the company to change the embedded value
assumption set, then the PVFP could be significantly affected.
During these years Person X may also enquire about how the policy’s bonuses work, what the
surrender value would be at such and such a date, or what the paid-up value would be if they
were to cease paying premiums.
2.5 The last day

Sadly, Person X dies. After checking the death certificate, the life insurance company sends X’s
nominated beneficiary a cheque for the amount of benefit accrued at that time. This might
include some allocation of terminal bonus, depending on the bonus allocation method.
Settling the claim will involve some termination expenses.
If the policy comes under any reinsurance arrangement then the company will get some money
back from the reinsurer. This will not happen there and then – the company will receive the
money when the reinsurer settles the relevant quarter’s account (if the reinsurance accounts are
quarterly – which is common practice).
Question
(i) Outline the impact of the claim on the company’s financial results.
(ii) Describe the impact of the claim on the company’s embedded value.

Solution
(i) There will be a loss equal to the difference between the amount paid and the policy
reserve.
(ii) The net asset value will take a loss as per the revenue account result in the previous
question – but it would be net of tax.
The present value of future profits component will suffer a small loss due to the fact that
that policy will no longer be contributing profits to the company. However, if the
experience for the portfolio as a whole is as per the embedded value projection basis,
then the present value of future profits for the portfolio will not decrease in this way.
2.6 Subsequent to the last day
Question
Suggest ways in which the policy might affect the life insurance company after that last day.
Solution
Possibilities here are:

 the company contests the claim and gets the money back in the event of fraud (eg a faked
suicide)
 the company contests the claim and gets some money back (eg extra premiums) in the
event of significant untruths on the policy proposal form coming to light
 the company tries direct selling methods on X’s next of kin, trying to sell a policy
appropriate to the recent receipt of the claim monies – eg an annuity to give regular
income, or a simple single premium savings policy
 X’s claim will appear in the company’s mortality investigation, and will eventually feed
into the assumption sets used for future modelling work.

3 Different parties in the insurance world

One way to create ideas when thinking about certain life insurance issues is to consider the
various different people, or stakeholders, who may be affected, or in any way interested, and
what their standpoints are.
We can think of the following parties as relevant:

 policyholders
 future policyholders
 shareholders
 sales channels
 life insurance company management
 life insurance company marketing department
 the life insurance company actuary
 other life insurance companies
 reinsurers
 supervisory authorities
 taxation authorities
 the government
 others.
3.1 The policyholders

Policyholders will ideally want the following, although many of these desires may conflict:
 products that meet their requirements (eg no exclusion clauses)
 products that are easy to understand
 minimum hassle (so eg no medical before being accepted for cover, preference for
premiums which will not vary in the future at an initially unspecified level)
 products that allow them to take advantage of any tax breaks
 large benefits for any given premium (so minimum loadings, good investment
performance for with-profits / unit-linked policies)
 for savings policies in particular, benefits on maturity which have kept pace with inflation
 maximum security of benefit and hence …
 … a large proportion of benefits guaranteed rather than discretionary
 … financially strong companies
 in other words … value for money.

3.2 Future policyholders
Question
Give two examples of how the desires of members of the public who see themselves as potential
future policyholders may differ from the views outlined above.
Solution
Members of the public will also want:

 a smooth sales interface (ie easily accessible agents)
 a large choice of insurance companies.
3.3 Shareholders
Shareholders will want:
 a return on their investment of at least the risk discount rate, and ideally as high as
possible
 low risk (so a low probability of the value of their shares plummeting)
 smooth supply of dividends (so low volatility of financial results)
 in theory, useful financial information (eg embedded value results as well as traditional
revenue account and balance sheet).
3.4 Sales channels

Sales channels (ie brokers, agents and direct sales forces) will want to
 maximise their remuneration
 provide a good service
 provide suitable products
 build up a portfolio of happy customers.
Question
Discuss the implications of sales channels wanting to maximise their remuneration.

Solution
Sales people will often want to sell those products that give them the highest remuneration,
rather than sell the product that is best for the customer. From a strictly financial and
quantitative point of view, their only interest in the suitability of the product for the customer is
the fact that a well-matched product should show good retention and so bring in future renewal
commissions.
If products are sold that prove not to be the best fit for customers’ needs, then persistency rates
will be low – particularly at early in-force durations. There may be reputational damage to the
sales channel and potentially also the insurance company, with corresponding implications for
new business volumes.
In reality, sales people may not be completely influenced by commission, because

 they want to offer a good service to attract business in the long term
 they may have a sincere interest in the suitability of the product for the customer.
3.5 Life insurance company management

Life insurance company management should want to:
 maximise long-term profits
 ensure that the company operates legally (ie demonstrates supervisory solvency,
complies with regulations etc).
However, in order to be seen to be doing something, they may want to:

 maximise short-term profits, and/or
 maximise new business levels (probably short-term).
3.6 Life insurance company marketing department

The life insurance company marketing manager will normally want to:
 maximise new business levels.
The view here may often be short-term more than long-term. The definition of ‘new business
level’ will affect the products being pushed: for instance, if the criterion is simply total new
business premium income then single premium business will be pushed hard, whereas if the
criterion is a measure such as ‘regular premium product new business premiums + 10% of single
premium product new business premiums’ then annual premium products will be pushed.
3.7 Life insurance company actuary
Question
List the actuary’s aims.

Solution
The actuary will want:

 to conform to statute and regulations
 to conform to professional guidance
 generally to work to a highly professional level (eg concern about policyholder equity)
 to work for the company to increase its profits, subject to not going against PRE
 possibly to maximise policyholders’ bonuses.
3.8 Other life insurance companies

Taking the perspective of any one life insurance company, other life insurance companies will
want to maximise their profits and (probably) new business levels.
There are a number of possible ways in which they might try to outperform their competitors:
 lower premiums from a lower profit margin
 lower premiums from lower expenses (perhaps from economies of scale)
 better investment performance (historic and expected)
 better administrative service
 quicker and simpler underwriting
 greater name awareness
 appeal to different target markets
 specialist (niche) products
 product designs that are easier to understand
 higher bonuses for with-profits business
 greater security of policy benefits (eg more benefit guaranteed rather than expected as
possible future bonus, other guarantees on future benefits and/or premiums)
 greater security of company (by demonstrating higher free assets)
 greater incentivisation of sales channels (often, higher commissions)
 more incentives for policyholders (free gifts, enhanced benefits, loyalty bonus).

3.9 Reinsurers
The aim of reinsurers is to maximise their long-term profit, subject to operating legally
(ie demonstrating solvency, etc). They will do this by:
 charging premiums which cover the cost of the inherent risk, plus their expenses and
some profit margin
 helping direct-writing companies reduce the cost of that inherent risk by providing
underwriting advice and assistance
 attracting direct writers by offering high commission on business ceded
 attracting direct writers by offering advice on pricing, underwriting and other areas
 attracting direct writers by demonstrating financial strength
 charging premiums which are competitive compared with other reinsurers (and
competitive with respect to the alternatives to reinsurance that direct writers might
consider)
 offering reinsurance and expertise for new risks (perhaps using their research and
experience from other markets).
In addition, in the interests of short-term profit stability, they may reinsure some of the business
that they have taken from direct writers.
3.10 Supervisory authorities

The supervisory authorities will want to:
 ensure life insurance companies act in compliance with the law
 ensure life insurance companies remain solvent in the future
 ensure that policyholders are not mistreated
 ensure that potential policyholders (ie customers) are not misled.
3.11 Taxation authorities

The taxation authorities will want to
 maximise tax income (subject to what the tax laws allow).
3.12 Government
The government’s aims are likely to be:
 to ensure that the legal framework maximises policyholder security
 to incentivise products which are thought socially desirable (eg personal pensions)
 possibly to assist policyholders of bankrupt companies
 to encourage a fair and competitive marketplace.

The government may also act in its own interests, for instance
 by forcing life insurance companies to invest in large amounts of government bonds
 by protecting the monopoly (and profitability) of any State insurer.
Question
Suggest ways in which the government might encourage a fair and competitive marketplace.
Solution
By imposing restrictions on monopolies within the life insurance sector.
By ensuring that companies present their products in a consistent way (eg all benefit projections
done on the same assumed investment return, regulation on advertising and sales processes).
By ensuring that different sectors – eg banks v life insurance v personal investments – all enjoy
equivalent regulation and tax treatment (both from the point of view of the companies and the
‘consumers’).
By encouraging cross-border freedom between countries as far as the provision of insurance

policies is concerned.
3.13 Others
The list could continue. Here are a few examples:
 other employees of the company – who will have a considerable personal interest in the
financial prosperity of the company
 investment analysts – who will be interested in the financial prospects for the company’s
shares, if the company is proprietary
 directors of any holding company
 non-life insurance companies, banks, building societies, etc – whose businesses in some
way compete with the life insurance market.

4 The control cycle revisited

It’s a useful exercise to look again at the control cycle (introduced in Subject CP1), but this time
applying all of the knowledge that we have built up in Subject SP2.
In essence, the control cycle can be thought of as the best way for a life insurance company to
operate. In any situation, the company will have some aims shaped by the business environment.
What are the risks that might impinge on these aims? Given those risks, what is the best
approach? Eventually, in the light of experience, was our perception of risks correct, and was our
selected approach the best or can it be improved?
4.1 Using the control cycle in an example

Here we shall look in some detail at an example to see how these ideas operate. The example is
presented as a possible exam question and solution. It is laid out without marks, because we
want to concentrate on the ideas involved rather than on exam technique.
Question
For political reasons, the government of Ruritania has just passed legislation allowing house-
owners to take out term assurance policies up to the value of their house tax-free (ie premiums
will be payable from pre-tax income and the benefit will be tax free), compared with the normal
treatment of life insurance policies where premiums are paid out of net income and benefits
beyond certain levels are subject to inheritance tax.
A life insurance company wants to take advantage of this to launch a new range of term
assurance products.
Describe:
(a) the risks faced by the company
(b) possible solutions to these risks
(c) how the company might then modify its chosen approach in the light of experience.
You should now spend 5-10 minutes on writing down an outline solution, just putting down the
keywords (rather than details about expense investigation mechanics etc).
Having done that, now read through the Syllabus in the Study Guide and check that you have not
missed out any areas which might affect your solution. Then read on.

Solution
This solution is written in ‘ActEd course note’ style rather than ‘standard exam solution’ style.
(a) Risks
The company’s aims will be to stay solvent, and subject to that, almost certainly to maximise
profits. What are the risks that might affect these aims? The risks here are the various areas
covered in Syllabus Objective 3.1. We shall look in detail at just three risks: mortality, expenses
and persistency.
Mortality
Fundamentally, the risk is that of actual mortality exceeding that priced for. This risk can be
broken down into the three components of model risk, parameter risk and random fluctuation.
There is the model risk that the expectation of mortality is not correctly modelled – for instance
the pricing actuary might model it as a constant when it will in fact vary over time.
There is the parameter risk that the estimate of future mortality, and hence the parameters which
define the modelled version of mortality, will not be correct.
To estimate mortality, the company will need to look at its experience with the current term
assurance contracts (and to a lesser extent related contracts). However, the policyholders eligible
for this new product may be significantly different from those in the company’s portfolio, or those
underlying any other statistics which the company may look at (industry data, reinsurers).
If the company overestimates mortality then it is at risk of not selling the product. If it
underestimates mortality then it is at risk of making a loss, or eventually facing solvency problems
if the new product becomes a major part of the portfolio. Of the two, underestimating will be
more problematic than overestimating.
Associated with the parameter risk is the risk of selection against the company, if the company
does not rate by splitting policyholders into groups reasonably homogeneous with respect to
mortality.
There is also the risk of random fluctuations – even if the estimate of mortality is spot on, claims
will fluctuate around the expected level. The extent to which claims vary from expected will
depend on the size of the portfolio.
Expenses
There is the risk that the contracts sold will not cover the associated expenses. This could come
about in various different ways.
First, the company will not know the expenses involved with this contract. They should be very
similar to that for similar contracts – for instance the current range of term assurance contracts.
But there will be some differences; for instance here the company will need to incorporate some
checks that the potential policyholders are house-owners, so the underwriting might be slightly
more involved.

Secondly, future expense inflation might be underestimated.
Thirdly, there is a risk of selling insufficient levels of the contract to recover the development
costs, or to make a sufficient contribution to the company’s overheads.
Finally, the price structure might involve cross-subsidy of small policies by large policies,
otherwise small policies could prove uncompetitive. In this case there is the risk that a greater
than expected proportion of new business is low premium, and that the overall production does
not then make a sufficient contribution to overheads.
Persistency
There are two completely different risks here. First, if most of these policyholders will have
mortgages then in the event of economic recession they may be especially hard hit and may be
even more likely than other term assurance policyholders to lapse their policies (if legislation or
whatever does not oblige them to have such policies in respect of their mortgage).
The extent to which this is true will depend on the average wealth of house owners with
mortgages, compared with the wealth of the remaining population of term assurance
policyholders.
Secondly, there is the risk that vast swathes of existing term assurance policyholders who are
house owners will be better off lapsing their current policies and taking out new policies under
this tax-efficient format.
(b) Possible solutions to the risks
Mortality
To minimise the risk of modelling or parameterising mortality risks incorrectly, the company will
need to study the experience of its related contracts. Term insurance and whole life insurance
should be a good start point, and endowment assurance should be closer than it would be with a
standard term insurance contract (because of the expected socio-economic profile of the new
policyholders).
Reinsurers’ and industry statistics will also be a useful point of reference.
The risk of pricing on the wrong mortality basis could be reduced by offering the contract as
unit-linked with a variable mortality charge, rather than conventional with a guaranteed mortality
charge.
The company should use at least all of the rating factors which most of the competition use
(eg age, gender, smoker / non-smoker, medical).
The company will need to ensure that the policyholders undergo underwriting at least as strict as
that enjoyed by the policyholders underlying the mortality investigations above. Alternatively the
mortality advantage of belonging to the socio-economic group ‘house-owners’ might be deemed
sufficient to compensate for a weakening of the underwriting procedures, eg a raising of the sums
assured at which detailed medical evidence is required.
Quota share reinsurance could be taken out to reduce the financial impact of parameter error.

To deal with the risk of random fluctuations, the company could use aggregate excess of loss
reinsurance.
The extent of reinsurance required will depend on the company’s total portfolio of similar
policies, its free reserves and the extent to which smooth financial results are desirable (eg more
important for a proprietary than for a mutual company).
An alternative to this would be to set up a mortality fluctuations reserve, but this would require a
certain amount of capital.
Persistency
The company could tackle the risk of low persistency on this product by setting premiums
assuming a high withdrawal basis, although the extent to which the company can do this is
constrained by competition.
A commission structure to agents which shifts remuneration from initial to renewal commission
may make it less likely that agents sell the product to people for whom it is not very suitable.
Not offering any surrender value will reduce the problem, compared with the perhaps
theoretically above-zero surrender values which could be paid out at some durations.
The second problem, of lapse and re-entry on current products, could be tackled in various ways.
One way would be to offer to reclassify eligible policies, charging an appropriate administration
fee. However, a realistic fee might be so high that it still leaves lapse and re-entry as the
preferable option to a non-lazy policyholder.
Another possibility would be to quantify what addition to benefit (or reduction in premiums)
would equate to the expected cost of administering a swathe of lapses and re-entries, and
offering some portion of such an increase as an incentive to continue with the policy in the
existing format.
Expenses
The risk of underestimating expenses could be reduced by making the contract unit-linked, with a
variable charge.
The risk of insufficient sales could be reduced by aggressive marketing, and a competitive pricing
basis.
The aim is to maximise the total expense contribution plus profit, which will be equal to
{number of sales} × {expense contribution and profit per sale}.
This will be almost zero for very high and low premiums, and maximised for some middle-of-the-
range premium – depending on elasticity of demand.

The risk that a cross-subsidy of small policies by large policies combined with a low proportion of
high premium sales gives rise to an insufficient contribution to the company’s overheads could be
countered by:
 setting the cross-subsidy such that all policies at least pay their own marginal expenses
 profit testing the new business portfolio on a variety of new business mix assumptions
 market research – what is the likely new business mix?
(c) Modification with experience
The company will need to monitor the experience of the contract. Such monitoring will include
looking at:
 composition of new business by age, gender, premium level
 mortality experience from mortality investigations
 persistency experience from persistency investigations
 expense experience for the product
 embedded value analysis of movement
 analysis of valuation surplus.
In the light of this experience, the actuary will then check that the pricing assumptions for the
contract are adequate. If they are not (or if they were too conservative) then the contract will be
repriced.
This could include altering, if necessary, the balance between fixed and ‘percentage of premium’
expenses in the price structure, depending on the actual composition of new business.
If persistency experience is particularly poor for any group of policyholders (eg all in one area, or
from one agency) then some further investigation would be warranted to determine the real
reason.
The company should also be monitoring the persistency experience of related contracts to guard
against the lapse and re-entry syndrome. If this looks like getting out of hand then the company
might adopt one of the measures outlined above.

4.2 Conclusion
In the above, all that we have done is:
 use the actuarial control cycle to structure our approach
 think about the context of the problem (the general business environment)
 use the specify the problem and identify the risks stage to identify potential risks to the
company’s aims, and how they might unravel in the particular situation
 use the develop the solution stage of the cycle to identify possible solutions, and how they
would apply in this situation
 use the last stage of the cycle to think about how to monitor experience
 look at the risks and solutions earlier to see how that experience will allow us to usefully
modify the approach.
This demonstrates how the control cycle is very useful as a way of generating the various aspects
to consider.

4.3 End of the course

Sadly, this is also the end of the course – although don’t forget to review the glossary.
After completing Assignment X6, you should therefore start your revision.
This should include some of the following:

 re-reading all chapter summaries
 re-reading all the Core Reading
 re-reading the glossary
 looking at each syllabus item, ensuring that you are familiar with the ideas
 doing all the practice questions at the end of the chapters
 doing the Mock Exam and having it marked
 looking at past exam questions – use them as further mock exams, and mark them
yourself using the examiners’ report or the SP2 ASET
 reviewing old assignments
 revising your knowledge of premium and reserve calculations, profit testing and
investment from earlier subjects.
Remember that the more active your revision is, the more effective it is likely to be. Thus making
your own notes from the reading is more effective than just passively reading; doing questions
under exam conditions is more effective than passively reading the questions and the associated
solutions.


31.1 A proprietary life insurance company currently issues a full range of conventional and unit-linked
Exam style
without-profits protection and savings contracts. The company is planning to launch a new
regular premium unit-linked endowment assurance contract with a very simple charging
structure. The aim is to address recent adverse media publicity about the alleged unnecessary
complexity of unit-linked contracts currently available in the market. The market is very
price-sensitive at the present time, so charges need to be brought down to a minimum.
The proposed design of the new product is as follows:
Allocation rate: 100% of all premiums are allocated to units at the ruling price;
there is no bid-offer spread.
Sum assured: The annual amount of the first year’s premium multiplied by
(55 – age next birthday at entry).
Death benefit: The sum assured, or the value of the units held under the
policy if higher.
Surrender and maturity values: The value of the units held at the time of claim.
Surrender is permitted at any age.
Fund management charge: 1.25% per annum, reviewable at the company’s discretion.
Maximum entry age: 54 next birthday.
Maximum maturity age: 75.
Premium level: The amount of premium can be chosen by the policyholder,
provided that the annual amount of any year’s premium does
not fall below the level of the first year’s premium. If this
condition is not met, then the policy has to be surrendered for
the then available surrender value.
All contracts are subject to the company’s standard underwriting procedures.
(i) Discuss the risks of the proposed product design to the company. [11]
(ii) Suggest, with reasons, possible ways in which the product design might be improved,
bearing in mind the company’s current business environment. [7]
(iii) Discuss whether a non-unit reserve would be held at any point during the term of this
contract and comment on its size. [10]
[Total 28]

31.2 A proprietary life insurance company sells without-profits immediate annuity contracts with level
Exam style
benefits. Withdrawal is not permitted under the contract. The company sells the contract
through independent intermediaries, and its annuity rates are usually very competitive.
The company keeps its annuity rates under regular review, and changes them frequently. A
commission of 1% of the single premium is paid to the intermediaries who sell the contract.
(i) State the aims that the company should be trying to achieve when deciding on the rates it
should offer for its immediate annuity contracts. [1]
(ii) Explain why the company needs to change its annuity rates so often. [4]
(iii) State the key variables that the company needs to monitor to achieve its aims, explaining
briefly why each one is important. [3]
(iv) Describe:
 a model that could be used to determine the new annuity rates, which allows
explicitly for the cost of capital employed
 how the model would be used to obtain a new set of annuity rates.
(You are not required to describe how the assumptions for the model would be
obtained.) [7]
(v) Describe briefly:

 the other model investigations that might be performed before adopting a
particular set of annuity rates
 the impact that the investigations might have upon the final pricing decision. [5]
[Total 20]

31.1 (i) Risks
Mortality
There is a significant sum at risk at young ages. [½]
Anti-selection risk exists at entry, but should be controlled by underwriting. [1]
There is a mismatch between the incidence of the charge (fund management charge) and the
expected outgo: eg sum at risk decreases with increasing duration whereas the charge increases
with duration. [1]
Leads to a significant risk from different mix of business from expected: eg if there is a greater
proportion of younger policyholders than expected when pricing the contract. [½]
Risk also of different class of life from expected, with different mortality experience. [½]
Reviewable charges should protect the company from any significant longer term mortality risk,
… [½]
… but reviewability can also lead to anti-selection risk on surrender: those unwilling to pay
increased charges would tend only to be healthy policyholders, who might then surrender their
policies, producing an increase in the average mortality of the policyholders remaining in force. [1]
Surrender (persistency)
Significant risk as the surrender value (the value of the unit fund) is likely to exceed the asset
share at early durations, a position that might persist for several years. [½]
The cause of this is the uneven incidence of expenses (high initial and low renewal), while the
charges gradually increase from a low level at the start of the contract. [1]
The forced surrender for low premium levels is probably counter-productive: it is almost certainly
better to maintain a policy paid-up and so receive some future fund management charge, rather
than receive no future charges at all. [1]
A given level of withdrawal will have been assumed in the pricing of the contract: the risk is
therefore that a higher rate of withdrawal than expected occurs. [½]
Investment
The direct consequences of poor or good investment performance is borne by the policyholder, so
there is no direct risk to the company. [½]
However, poor performance over long periods may lead to policyholder dissatisfaction and hence
a marketing risk. [½]
Poor returns will increase the sum at risk and hence the cost from mortality. [½]

Poor returns will reduce the value of the fund management charge and hence will increase the
risk of making losses. [½]
Expenses, volume and mix of business
Too few policies issued will make fixed expenses increase per policy. [½]
Two possibly unattractive features of the product design might increase the risk of inadequate
new business sales: inflexibility of the amount of death benefit provided (policyholder has no
choice); and the restriction on premium flexibility (ie unable to reduce premiums below a given
amount without forfeiting the contract). [1]
Development costs need to be recouped, so a critical minimum volume will be needed to cover
this. [½]
Smaller average policy size than expected will lead to loss, as fund management charges will be
lower, and the per policy expenses will not be fully covered. [½]
Whilst reviewable charges protect from direct risk in theory, marketing and competitive
considerations will prevent rises in charges above a certain level. [½]
Inflation
The fund management charge will increase with the growth in the unit fund, which if invested in
equities is likely at least to match expense inflation in the long term. [1]
Short-term investment fluctuations compared with inflation could lead to expense losses in
individual years (for example if asset values fall when inflation is rising). [½]
Taxation
There is a risk that tax rates will be higher than assumed in the pricing basis. [½]
Capital requirements
These could be extremely onerous because the unit fund will exceed the asset share for a
considerable period. [1]
There would then be a risk of selling too many policies and endangering solvency. [½]
[Maximum 11]
(ii) Possible solutions
Introduce surrender penalty
Introduce surrender penalty to bring the surrender value much nearer to (or even less than) the
asset share. [½]
This will much reduce the surrender risk. [½]
May also reduce solvency requirements. [½]

Match charges to outgo
Change incidence of charge so as to produce a high initial charge and low renewal charges, eg use
an initial nil-allocation period. [1]
This will have similar risk and capital benefits to surrender penalty, but with less risk (as the
surrender penalty would not apply on death), though the new design may appear to have more
charges and hence be less marketable than before. [1]
A mixture of fund related, premium related (eg bid-offer spread, allocation reduction) and per
policy charges (eg policy fee) could be used. This would match charges to the nature of the
expenses and hence reduce the risk due to the size of new business. [1]
The mortality risk could be improved by introducing a charge that is calculated as a function of the
actual sum at risk. [½]
However, all of these suggestions involve increasing the complexity of the product, which could
make the policies considerably less marketable. [1]
Other solutions
A simple way of controlling the size risk might be to introduce a minimum premium level. [½]
Could try devising a different sum assured structure that better matches the charges for mortality
emerging from the existing fund management charges. This would reduce the risk from an
unexpected business mix, and would tend to make the policy relatively more attractive at the
older ages than hitherto, thereby also leading to an improvement in mix. [1]
Could simply reduce the death benefit to make its cost less important. Underwriting may then
become less important and hence cheaper, increasing marketability and reducing costs to the
company (and to the policyholder, if passed on through reduced charges). [1]
Could remove requirement to surrender if premium falls below initial level; will improve
marketability and will help to maintain volume thereby increasing income from fund charges. [½]
Could put in a maximum value for the fund management charge (ie allow the fund management
charge to be reviewable up to a maximum ceiling). This should increase marketability without
greatly increasing the risk, bearing in mind a cap effectively exists with the existing structure due
to the impact of policyholders’ reasonable expectations and competition. [1]
[Maximum 7]
(iii) Non-unit reserves
The non-unit reserve is the present value of the excess of non-unit outgo over non-unit income.
[½]
For this contract the non-unit outgo is the expenses plus any excess of the sum assured over the
unit value payable on death. [½]
The non-unit income is the fund management charge. [½]
The expenses will increase with inflation over the term of the contract. [½]

The cost of the death benefit is likely to decrease over the term as the unit value will increase as
further premiums are paid and (hopefully) as investments grow in value … [1]
… but the reducing cost of the death benefit will be offset slightly by the increasing rate of
mortality as the policyholders grow older. [½]
The non-unit income is likely to increase over the term as it is a proportion of the unit value
(which should grow over time as described above). [½]
So we would expect the non-unit cashflows to be negative in the early years (ie the outgo will
exceed the income) … [½]
… but the non-unit cashflows should be positive in the later years … [½]
… becoming ever more positive as the assets grow. [½]
The size of the non-unit cashflows will depend on the following factors:
 a lower unit growth rate will make the cashflows less positive (or more negative) [½]
 a low premium will make fixed expenses relatively more important and so cashflows will
be negative for longer [½]
 a low entry age will affect the cost of the death benefit as the sum assured will be higher
but the probability of claiming will be lower. [1]
Under a regulatory regime which requires mathematical reserves to be prudent, the non-unit
reserve is typically defined as the amount required to ensure that the company is able to pay
claims and meet its continuing expenses without recourse to further finance. [½]
When calculating the reserve in the early years of the contract, the positive cashflows later in the
term will be ignored, … [½]
… the calculation process will start with the last projection period in which the non-unit cashflows
are negative … [½]
… and a positive non-unit reserve will be calculated so that all the future negative cashflows are
‘zeroised’. [½]
The company should project forwards its non-unit cashflows on the reserving basis. [½]
The prudent nature of the reserving basis will make the expenses and death benefits larger and
the fund management charges smaller … [½]
… and so will increase any positive non-unit reserves held. [½]
When calculating the reserves later in the term of the contract, once no more negative cashflows
are expected, it may be permissible to hold a negative non-unit reserve. [½]
Regulations might require that the sum of the unit and non-unit reserve for a policy is not less
than any guaranteed surrender value. [½]

If this is the case then negative non-unit reserves will not be allowed for this product as there is
no surrender penalty (ie the full value of the units is paid on surrender). [½]
Regulations may also specify that a negative non-unit reserve can only be held if:
 the future profits arising on the policy with the negative non-unit reserve emerge in time
to repay the ‘loan’ [½]
 after taking account of the future non-unit reserves, there are no future negative
cashflows for the policy, ie there should be no future valuation strain [½]
 in aggregate, the sum of all non-unit reserves is not negative. [½]
If a negative non-unit reserve cannot be held due to any of the above regulations, then the non-
unit reserve will be set to zero. [½]
Under a regulatory regime which requires mathematical reserves to be a best estimate valuation,
the calculation would value all future non-unit cashflows (ie it would not disregard cashflows
occurring after the last projection period in which there is a net outflow). [½]
The best estimate nature of the reserving basis will make the non-unit reserves more negative
than for the prudent reserves described above. [½]
If the contract is written to be profitable, then the expected present value of the non-unit income
should exceed the outgo. [1]
At the outset of the contract there will be a large non-unit outgo due to the initial expenses, but
no income as 100% of the premium is allocated to units. [1]
So the expected present value of the income will be considerably greater than the outgo once the
policy is in force. [½]
Hence a negative non-unit reserve could be held throughout the term of the contract. [½]
This reserve would become more negative throughout the term until the point at which the net
income first became positive (as there would no longer be years of negative income offsetting the
later years of positive income). [½]
negative non-unit reserves can be held, so there would be no requirement to set these reserves
to zero. [½]
[Maximum 10]

31.2 (i) Aims
To maximise profit while ensuring that the risks faced by the company in issuing the contracts can
be covered by the company’s available resources. [1]
Total profit is a function both of the individual profit per policy and of the volumes of business
sold, so the company will aim to keep the product competitive while maximising the per-policy
profit contribution. [1]
[Maximum 1]
(ii) Why change annuity prices so often?
Without-profits business is usually fairly simple, so competition between different companies’

products is based almost entirely on price. [½]
To remain competitive, the product has to remain one of the cheapest in the market. [½]
So the company needs to monitor competitors’ rates to ensure it doesn’t lose its competitive
position. [½]
Without-profits contracts provide the highest possible guarantee, and coupled with a very cheap
price leads to potentially high risk. [½]
One way of controlling this risk without introducing uncompetitive margins in the pricing basis is
by pursuing a matched investment strategy. [½]
This involves investing in fixed-interest securities with appropriate duration. [½]
Because annuities usually don’t require any reinvestment in the future,… [½]
…it is possible to secure an actual future yield very close to the average redemption yield of the
stocks purchased at the time the single premium is paid. [½]
The company can then price the contract on this basis, and so secure the cheapest price the
company can charge while still producing acceptable risk. [½]
If market yields fall, an unchanged annuity rate could quickly become loss-making when sold in
these conditions. [1]
Frequent changes to annuity rates are necessary because the market redemption yields are quite
volatile. [½]
[Maximum 4]
(iii) Key variables to monitor (in addition to redemption yield)
Mortality: this contract will make losses if aggregate mortality is lower than that assumed in the
pricing basis. [1]
Greater (or lesser) longevity than expected will lead to a mismatching of assets and liabilities that
will increase the investment risk, possibly significantly. [1]

Expenses: a certain amount of expenses per policy are assumed in the pricing basis. Higher
average per-policy costs than assumed (whether through inefficient operations or from higher
than expected inflation, for example) can lead to losses. [1½]
Inflation is therefore also a key variable. [½]

[Maximum 3]
(iv) Pricing model
Use a single tranche model profit test, project the in-force expected profit at time t (for
t = 0,1,2,…) to be Pt where:
P0   0.99  SP  IE  0V  ; and
Pt  t 1V (1  i )  px t 1  tV  A  Et  for t > 0
where SP = amount of single premium

IE = amount of initial expenses
tV = supervisory reserve plus required solvency capital at time t
A = annual amount of annuity (assumed paid at end of year)

Et = amount of renewal expense at time t
i = expected yield per annum from the invested assets

x = entry age [2½]
Then calculate various summary statistics, eg:

(1) Net present value (NPV) = P0   v t .t 1 px .Pt
t 1
1
(where v  and r is the risk discount rate) [1]
1r
NPV
(2) Profit margin = [½]
SP
(3) Return on capital (provided P0  0 ) = rate of discount r  that, when substituted for r, gives
a net present value of exactly zero. [½]
Profit criteria would be defined in terms of NPV, profit margin, and return on capital. [½]
Find the highest value of A which meets the profit criteria for a given value of SP. [½]
Repeat for each entry age or as necessary to produce a full range of annuity rates … [½]
… noting that the yield i may well vary by age (as the matching asset portfolio will vary by age due
to the different average policy duration implied). [½]

The profitability would be tested for sensitivity … [½]
… eg assuming different yields, mortality, expenses, expense inflation and average policy size. [1]
This indicates the range of conditions for which the price remains acceptable in terms of its
profitability (ie its tolerance) … [½]
… which is particularly important to know, since it indicates when the price will need changing
again in the future. [½]
If the price is too intolerant to changes in conditions, it might be necessary to change the
assumptions – for example by increasing the margins for adverse experience in the
assumptions. [½]
[Maximum 7]
(v) Other models
The company might look at the impact of issuing business at the new rates in the context of a full
model office projection. [½]
This would involve a model of the company’s existing business in force … [½]
… using a wide range of model points designed to represent the profile of the actual in-force, in
terms of numbers, size, ages, durations, policy types, etc. [1]
Assumptions about future new business would need to be made (for all business types). [½]
The expected future volumes of annuity business would be particularly crucial. [½]
This would need to be based on historical experience, bearing in mind economic conditions
(annuities are less popular when yields are low) … [½]
… the competitiveness of the product (by relative price) … [½]
… and the amount of marketing effort applied to selling the contract. [½]
This will give a more realistic indication of the added value from the contract (allowing explicitly
for the interaction between volume and fixed expenses) … [½]
… and also of the impact on the company’s solvency capital. [½]
The new annuity business could also be projected as a separate exercise, in order to project the
total capital required and the return on capital expected from this business in the future. [1]
The outcome of these other investigations might be to change the profit test assumptions and the
resulting annuity rates; … [½]
… it might be necessary to try a different price that, allowing for the effect of competition on new
business volumes, may lead to a higher return on capital when the totality of the new business is
considered. [½]
Other sensible comments would earn credit [Maximum 5]

End of Part 6
What next?
Chapters 28 to 31.
of your revision.
… ‘rehearsal’ products
Mock Exam and Marking – You can attempt the Mock Exam and get it marked. Results of
surveys have found that students who do a mock exam of some form have significantly higher
pass rates. Students have said:
‘I find the mock a useful tool in completing my pre-exam study. It helps

me realise the areas I am weaker in and where I need to focus my study.’
‘Overall the marking was extremely useful and gave detailed comments on
where I was losing marks and how to improve on my answers and exam
technique. This is exactly what I was looking for – thank you!’
You can find lots more information on our website at www.ActEd.co.uk.
And finally ...

Good luck!





using the course.

SP2-32: Glossary Page 1
Glossary
Syllabus objective
0.1 Define the principal terms used in life insurance.

Page 2 SP2-32: Glossary
0 Introduction
This chapter includes the Core Reading definitions with which you need to be familiar. These
definitions are an easy target for a short bookwork question. You do not have to reproduce them
word-for-word, but you must be able to express the ideas with equal precision.
The glossary is also a useful reference as you study the course, where many of the ideas are
explored in more depth. In fact, in some instances we have repeated a Core Reading definition
within the course notes so as to set it in context and explain it further.

1 Definitions
Actuarial Control Cycle

A systematic method of planning, managing, and reviewing an insurance company’s
business in the face of the risks undertaken by the insurer. Having modelled and made
assumptions about the risks, the actual experience is reviewed to add insight into the future
modelling and management of existing and new business.
The control cycle can be applied at both macro and micro levels of the insurance business.
Aggregate asset share

Asset share calculations can be carried out for individual policies, for specified product
lines or for the whole with-profits portfolio. The latter two are sometimes referred to as
aggregate asset share. This term may also be used to describe the sum of the individual
asset shares.
Anti-selection
People will be more likely to take out contracts when they believe their risk is higher than
the insurance company has allowed for in its premiums. This is known as anti-selection.
Anti-selection can also arise where existing policyholders can exercise a guarantee or an
option. Those who have most to gain from the guarantee or option will be the most likely to
exercise it.
Appraisal value
The appraisal value of a proprietary life insurance company is the sum of the embedded
value of the company and the value to its shareholders of the future profits they expect to
receive from future new business. The latter part of the appraisal value is often referred to
as the ‘goodwill’ value of the company.
Appropriation price
This is the amount of money per unit put into a unit-linked fund for each new unit
appropriated, ie created, such that the net asset value per unit is the same after as before
the appropriation.
Therefore, it is the price at which a company will create a unit.
Assets
The assets of a life insurance company are what it holds to meet its liabilities and to provide
working capital. It usually refers to the investments held by the company.

Asset share
The asset share is the retrospective accumulation of past premiums, less expenses and the
cost of cover, at the actual rate of return on the assets. The accumulation could be carried
out for a single contract or a group of contracts. It is also referred to as the earned asset
share or the retrospective earned asset share.
In the case of with-profits contracts, allowance may be made for miscellaneous profits from
without-profits contracts and from surrenders and lapses, and also for the cost of
guarantees and any capital support provided.
Bancassurer
A bancassurer is an insurance company that is a subsidiary of a bank or building society
and whose primary market is the customer base of that bank or building society.
Best estimate reserve

This reserve would be calculated using assumptions that are best estimates rather than
assumptions that contain conservative margins.
Bid price
In the context of a unitised life insurance contract, this is the price the life insurance
company uses to redeem the units it has allocated to the contract.
Bonus earning capacity

The bonus earning capacity of a block of contracts is the rate (or rates) of bonus that those
contracts can sustain over their future lifetime, on the basis of a set of assumptions with
regard to future experience.
Commission
Commission refers to the payments which may be made by a life insurance company to
reward those who sell and subsequently service its products, whether they are financial
advisers, tied agents or a direct salesforce. Typically, the amount of the commission
depends on the type and size of contract. It can be paid when a contract is taken out (initial)
and/or over the duration of the contract (renewal) as a proportion of the premium or the
fund size.
Conventional contracts
A conventional contract typically has benefits that are initially set at the start of the
contract, based on the premium paid. In the case of conventional with-profits contracts,
additional benefits may be paid on top of these base benefits if investment returns have
been sufficient.

The discounted payback period is the policy duration at which the profits that have emerged
up to that point in time have a present value of zero. Therefore, it is the time it takes for a
company to recover its initial investment with interest at the risk discount rate.

Embedded value
This is part of the appraisal value of a proprietary life insurance company. It represents the
value of the future profit stream from the company’s existing business together with the
value of any net assets separately attributable to shareholders.
Equity
This is a term that is difficult to define. In essence, it means that all policyholders are
treated fairly. That is, that some groups of policyholders do not benefit at the expense of
other groups. In a proprietary company, equity also needs to be considered between
policyholders and shareholders.
Questions of equity arise in the distribution of surplus, in the determination of variable

charges, in the determination of surrender values and alteration terms, and in unit pricing.
Estate
The estate of a life insurance company usually refers to the excess of the realistic value of
its assets over the realistic value of its liabilities.
Expropriation price
The expropriation price is the price at which a life insurance company will cancel units in a
unit-linked fund. In other words, it is the amount of money that it should take out of the
fund in respect of each unit it cancels to preserve the interests of continuing unit-holders.
Extra premium
An extra premium is an addition to the standard premium payable under a contract to cover
an extra risk.
Extra risk
An extra risk arises where a proposal for life insurance is not acceptable at standard rates.
Facultative reinsurance
Facultative reinsurance is where individual risks are reinsured as and when the ceding
company writes the policy, with terms agreed for that risk. Compare this to reinsurance
under a treaty where all policies within its scope are reinsured automatically at the terms set
out within the treaty.
Fair value reserve

This is a reserve calculated using fair value accounting principles. A ‘market-consistent’
method of valuation, which has been increasing in importance in some countries, is an
example of fair value reserving.
Financial reinsurance is reinsurance where the main purpose is to provide a capital benefit
to the ceding company, although there will still be an element of risk transfer.

Financial strength
This is often measured by the level of its free assets or estate and usually refers to the
ability of a life insurance company to:
 withstand adverse changes in experience, including those arising from investment

in higher yielding, but more volatile, assets
 fulfil its new business plans
 meet the reasonable expectations of its policyholders.
Free assets
This term is used to refer to that part of a life insurance company’s assets that are not
needed to cover its liabilities. Opinion differs as to what should be included in the
liabilities, and they may or may not include capital requirements.
Genetic testing
A genetic test is defined as a test that examines the structure of chromosomes (cryogenic
tests) or detects abnormal patterns in the DNA of specific genes (molecular tests). The aim
of these tests is to determine the certainty or increased risk of developing diseases with an
inherited component. A genetic test can be predictive or diagnostic:
 A predictive genetic test is taken prior to the appearance of any symptoms of the
genetic condition in question.
 A diagnostic genetic test is taken to confirm a diagnosis based on existing

symptoms.
Gross premium valuation

This is a method for placing a value on a life insurance company’s liabilities that explicitly
values the future office premiums payable, expenses and claims, with the latter possibly
including future discretionary benefits.
Group contract
This is a contract that covers a group of lives, where the name of the group is specified but
where the individuals within it may not be.
Guarantee (investment)
In the context of life insurance, this refers to a promise that the company will pay a
specified sum of money – or sums of money – at specified times if a specified condition is
fulfilled. The condition can be an event such as the surrender or maturity of a contract.
The term can also refer to the situation where the company guarantees the rate it will use, at
some future date, to convert a lump sum into an annuity or vice versa.
Index-linked
A life insurance contract is index-linked if the benefits are linked directly to a specified
investment index or economic index, and are guaranteed to move in line with the
performance of that index.


This is defined as the rate of return at which the discounted value of the future profit
cashflows arising under a contract is zero.
Internal unit-linked fund

An internal unit-linked fund consists of a clearly identifiable set of assets, for example
equities, property, fixed-interest securities and deposits. The fund is divided into a number
of equal units consisting of identical sub-sets of the fund’s assets and liabilities.
Responsibility for unit pricing rests with the company, subject to any relevant policy
conditions.
Lapse
A life insurance contract lapses if the policyholder terminates a contract early due to
non-payment of premiums without the company making a surrender value payment to the
policyholder. Some companies also use ‘lapse’ to describe policies that have been
surrendered.
Liabilities
The liabilities of a life insurance company are the benefits it has contractually agreed to pay
to its policyholders, plus its future expenses less future premiums.
Marginal pricing
This refers to when a company decides, for competitive reasons, to price a contract
ignoring any contribution to overheads.
The value at which whatever is being valued (eg assets or liabilities) could be exchanged in
a sufficiently deep and liquid market, between knowledgeable and willing parties in an arm’s
length transaction.
Market value reduction (MVR)

Insurance companies normally retain the right to apply a market value reduction to
surrenders or withdrawals from an accumulating with-profits contract, if the value of the
underlying assets is less than the fund value including all bonuses. The MVR is designed to
protect policyholders who remain in the with-profits fund, as its application means that the
withdrawing policyholder gets a fair share of the assets. There are normally points at which
no MVR will be applied (eg at maturity, on death, after ten years, or on partial withdrawals
below certain limits).

Matching
This refers to the relationship — by size and incidence — between the cashflows from the
assets and those from the liabilities. The assets and liabilities of a life insurance company
are said to be (absolutely) matched if the two cashflows cancel out. Otherwise, the
company is said to be mismatched. In practice, absolute matching is almost impossible to
attain, except in very special circumstances, and some degree of mismatching is therefore
inevitable and acceptable, provided it does not lead to an unacceptable probability of
insolvency for the fund or the insurer.
Mathematical reserves
In the context of supervisory reporting, the mathematical reserves consist of the value of a
company’s liabilities. The terminology might also include any explicit additional
contingency reserves, for example a mismatching reserve. However, in some jurisdictions
contingency reserves are termed ‘capital requirements’ rather than ‘mathematical reserves’.
Mismatching reserve
If the assets of a life insurance company are not matched to its liabilities, it may be unable
to meet claims as they fall due in the event of adverse future investment conditions. A
mismatching reserve is a reserve that has been set up as a contingency reserve that may be
called upon in the event of such adverse conditions.
Mutual (life insurance company)

This is a life insurance company that does not have any shareholders – it is effectively
owned by the policyholders.
Negative non-unit reserve

If projected non-unit income exceeds non-unit outgo on a unit-linked contract, it may be
possible for a life insurance company to set up a negative non-unit reserve in respect of
that contract. This may be subject to certain constraints, depending on the regulatory
regime.
Negative reserve
The valuation of a life insurance contract (using a gross premium reserve or equivalent) will
give a negative reserve if the value of future premiums exceeds the value of the benefits
plus future expenses. This means that the contract is being treated as an asset.
Net premium valuation

This is a method for placing a value on a life insurance company’s liabilities that involves
calculating a present value of the contractual liabilities and deducting the value of future net
premiums, allowing in each case only for mortality and interest.
The net premium is the premium, calculated on the valuation assumptions and payable
under the same conditions as the office premium, that will provide the contractual benefits
offered at the commencement of the policy.

New business strain

New business strain arises when the premium paid at the start of a contract is not sufficient
to cover the initial expenses, any commission payments and any reserve or additional
capital requirements that the company needs to set up at that point.
Non-unit reserve
A company will have non-unit liabilities under its unitised contracts – for example the
expenses of managing the business – for which it receives monetary payments in the form
of the future charges it extracts. If it expects that the charges will not be sufficient to meet
these liabilities at any point on a cashflow basis, it must hold a non-unit reserve to provide
for the deficiency.
Depending on the regulatory regime and any related constraints, it may be possible for a life
insurance company to hold a negative non-unit reserve where it expects that future charges
will be more than sufficient to meet the future non-unit liabilities.
Offer price
In the context of a unitised life insurance contract, this is the price a life insurance company
uses to allocate units to the contract.
Office premium
The office premium is the premium that the policyholder pays under a life insurance
contract. It is also called the ‘gross’ premium.
Option (health)
A health option is where the life insurance company gives a policyholder the right to
increase or extend the death cover under a life insurance contract at some future time or
times without further evidence of health.
Original terms (coinsurance) reinsurance

This method of reinsurance involves a sharing of all aspects of the original contract.
Overheads
The term ‘overheads’ is used to refer to that part of a life insurance company’s total
expenses which are independent of the volumes of business it is currently writing or has
already written. In practice, companies will have different views as to which of their
expenses will be overheads. A company will usually allow for its overheads in pricing by
allocating them on a per policy basis using an estimate of the number of contracts it
expects to write.

Paid-up policy
This is a regular premium policy under which no further premiums are payable and
sufficient premiums have been paid such that benefits are paid on claim even without the
payment of further premiums. It normally arises because the policyholder decides not to
pay any further premiums, in which case the company would reduce the benefits under the
contract allowing for the actual premiums paid. Under some regular premium policies,
premiums may be contractually payable for a shorter term than the term of the contract.
When the premium paying term has expired, such policies are sometimes referred to as
‘fully paid-up’.
Persistency
In a life insurance company, ‘persistency’ is used to refer to the rate of retention of policies
that is experienced by the company. If a company has ‘poor persistency’, this indicates a
high level of lapses, surrenders, partial withdrawals and/or conversions to paid-up status.
Persistency rates are typically measured over a defined period, such as a year, although
may be expressed as a cumulative figure over the period since policy inception.
Policy fee
This is an amount, usually independent of the size of benefit under a contract, included in
the office premium to cover part of a life insurance company’s maintenance expenses.
Policyholders’ reasonable expectations (PRE)

This relates to policyholders’ reasonable expectations with regard to the level of benefits –
or charges – under contracts where these are at the discretion of the life insurance
company.
There is no generally accepted definition of PRE, but they will be influenced by, for example,
the past practice of a company and any literature issued by it.
The concept of PRE is linked to the idea of treating customers fairly. In some jurisdictions,
there may be a statutory requirement placed on an insurer to meet minimum standards in
this respect.
Profit test
A profit test is a technique that involves discounting the future cashflows arising under a
contract for assessing the expected profitability of that contract. This profitability is often
measured as a percentage of the first year’s premiums (or commission, if applicable) under
the contract. Profit testing can be used to determine the premium and/or the level of
charges under a contract.
Proprietary (life insurance company)

This is a life insurance company that is owned by shareholders.
Regular reversionary bonus

This is a bonus that is declared on a regular basis, usually each year, throughout the
lifetime of a with-profits contract. Once declared it becomes attached to the basic benefits
and cannot generally be taken away.

Reinsurance
Reinsurance is the process by which a direct-writing life insurance company transfers part
of its risk under a contract to another life insurance company. This may be another
direct-writing company or a specialist reinsurance company. The reinsuring company may
in turn reinsure some of the risk with another direct-writing or reinsurance company.
Larger companies may also use reinsurance to transfer liabilities between companies within
their own group.
Requirement for capital

On a per contract basis, the requirement for capital is the amount of finance a company
needs to be able to write that contract, ie the new business strain. This can be extended to
the whole company where its requirement for capital is the finance it needs to carry out its
business plan, and could include writing new business, acquiring an existing business, or
investing in infrastructure such as new computer hardware or software.
Retention
In the context of reinsurance, a company’s retention is the amount of any risk that it wishes
to retain for itself. It will then reinsure the excess over that amount.
In the context of persistency, retention refers to the extent that a company can prevent
policies from lapsing or surrendering earlier than expected.
Return on capital
This is broadly defined as profit (before tax) divided by capital employed, expressed as a
percentage. It arises in the context of product pricing. A company will usually need to
provide capital in order to write new business. The expected return on that capital will
influence whether or not the company writes particular types of business and the price at
which it will write them. The expected level of return required will depend on the expected
levels from other uses of the company’s capital.
Risk discount rate

A risk discount rate is a rate at which future uncertain cashflows might be discounted. It
typically arises when carrying out a profit test of a life insurance contract. It represents the
risk-free rate of return that the providers of capital demand plus an amount to allow for the
risk that the profits may not emerge as expected from the contract.
It also arises in the determination of embedded and appraisal values.

In this method of reinsurance the direct writing company reinsures part of the sum assured
or the sum at risk on the reinsurer’s premium basis. The sum at risk is the excess of the
benefit payable over the valuation reserve.

Smoothing
A with-profits policy normally invests in equities and property, resulting in more variable
returns than if it were invested in lower-risk investments. Instead of paying out the exact
asset share, with-profits funds aim to even out some of the variations in investment
performance. Profits and losses are spread from one year to the next so that in total all
investment surpluses are paid out in the long term. The effect is a reduction in investment
risk for the policyholder.
As well as this smoothing of investment returns over time, life insurance companies also
smooth payouts across individual policies at any point in time. This reflects the ‘pooling’
effect of insurance and the practicalities of bonus setting.
Solvency
A life insurance company is solvent if its assets are adequate to enable it to meet its
liabilities and any solvency capital requirements that it must hold. Insurance supervisory
authorities will usually have requirements, in terms of the values a company can place on
its assets and liabilities, for the purpose of demonstrating solvency.
Surplus
Surplus is the excess of the value placed on a life insurance company’s assets over the
value placed on its liabilities. A negative surplus is usually called a strain or deficit.
Surrender value
This is the amount that would be paid out to a policyholder who cancels a contract.
Terminal bonus
A terminal bonus is a bonus that may be payable on maturity, death or surrender of a with-
profits contract. It is typically a percentage, varying with duration in force and possibly with
original policy term, of attaching regular reversionary bonuses and/or sum assured under a
conventional with-profits contract, or of the accumulated benefit (allocated premiums plus
regular bonuses added to date) under an accumulating with-profits contract.
(For accumulating with-profits business, the accumulated benefit may have other charges
deducted from the allocated premiums.)
Underwriting
This is the process by which a life insurance company decides whether an applicant can be
accepted at the standard premium or, if they cannot, on what (if any) terms.

Unitised contracts
After deducting an amount to cover part of its costs, each premium under a unitised
contract is used to buy units at their offer price. These units are added to the contract’s
unit account. When the insured event happens the amount of the benefit is the then bid
price value of all the units in the contract’s unit account. This may be subject to a minimum
amount specified in monetary terms.
The price of the units may either relate directly to the value of the assets underlying the
contract, or may be related to an investment or other index, or may be based on smoothed
asset values, perhaps with a guarantee that the price of the units will not fall.
Unitised contracts include unit-linked contracts and those accumulating with-profits

contracts that are written on a unitised basis.
Unit reserve
This is part of the reserve that a life insurance company needs to set up in respect of its
unitised contracts. The unit reserve represents its liability in terms of the units held under
the contracts.
Valuation
This is the process by which a life insurance company will place a value on its assets and/or
its liabilities.
With-profits
A life insurance contract is with-profits if the policyholder is entitled to receive part of the
surplus of the company. The extent of the entitlement is usually at the discretion of the
company.
Without-profits
A life insurance contract is without-profits if the life insurance company has no discretion
over the amount of benefit payable, ie the policy document will specify at outset either the
amount of the benefits under the contract or how they will be calculated.

All study material produced by ActEd is copyright and is sold for
the exclusive use of the purchaser. The copyright is owned by
Institute and Faculty Education Limited, a subsidiary of the
Institute and Faculty of Actuaries.
Unless prior authority is granted by ActEd, you may not hire out,
lend, give out, sell, store or transmit electronically or photocopy
any part of the study material.
You must take care of your study material to ensure that it is

not used or copied by anybody else.

addition, we may seek to take disciplinary action through the
profession or through your employer.
These conditions remain in force after you have finished using

the course.

Subject SP2: Assignment X1
2019 Examinations
Time allowed: 2¾ hours
Instructions to the candidate

1. Please:
– attempt all of the questions, as far as possible under exam conditions
– begin your answer to each question on a new page
– leave at least 2cm margin on all borders
– write in black ink using a medium-sized nib because we will be unable to mark
illegible scripts
– note that assignment marking is not included in the price of the course materials.
Please purchase Series Marking or a Marking Voucher before submitting your script.
– note that we only accept the current version of assignments for marking, ie you can
only submit this assignment in the sessions leading to the 2019 exams.
2. Please do not:
– use headed paper
– use highlighting in your script.
At the end of the assignment

If your script is being marked by ActEd, please follow
the instructions on the reverse of this page.
In addition to this paper, you should have available actuarial tables and an
electronic calculator.

Submission for marking
You should aim to submit this script for marking by the recommended submission date. The
recommended and deadline dates for submission of this assignment are listed on the summary page
at the back of this pack and on our website at www.ActEd.co.uk.
Scripts received after the deadline date will not be marked, unless you are using a Marking Voucher.
It is your responsibility to ensure that scripts reach ActEd in good time. If you are using Marking
Vouchers, then please make sure that your script reaches us by the Marking Voucher deadline date
to give us enough time to mark and return the script before the exam.
When submitting your script, please:
 complete the cover sheet, including the checklist

 scan your script, cover sheet (and Marking Voucher if applicable) and save as a pdf
document, then email it to: ActEdMarking@bpp.com
 do not submit a photograph of your script
 do not include the question paper in the scan.
In addition, please note the following:
 Please title the email to ensure that the subject and assignment are clear eg ‘SP2 Assignment
X1 No. 12345’, inserting your ActEd Student Number for 12345.
 The assignment should be scanned the right way up (so that it can be read normally without
rotation) and as a single document. We cannot accept individual files for each page.
 Please set the resolution so that the script is legible and the resulting PDF is less than 4 MB
in size.
 Do not protect the PDF in any way (otherwise the marker cannot return the script to ActEd,
which causes delays).
 Please include the ‘feedback from marker’ sheet when scanning.
 Before emailing to ActEd, please check that your scanned assignment includes all pages and
conforms to the above.

2019 Examinations
Please complete the following information:
Name:
Number of following pages: _______
Please put a tick in this box if you have solutions

and a cross if you do not:
Please tick here if you are allowed extra time or

ActEd Student Number (see Note below): other special conditions in the
profession’s exams (if you wish to
share this information):
Time to do assignment
(see Note below): _____ hrs _____ mins
Note: Your ActEd Student Number is printed on all
personal correspondence from ActEd. Quoting it will help Under exam conditions
us to process your scripts quickly. If you do not know (delete as applicable): yes / nearly / no
your ActEd Student Number, please email us at
Note: If you take more than 2¾ hours, you should
ActEd@bpp.com.
indicate how much you completed within this
Your ActEd Student Number is not the same as your exam time so that the marker can provide useful
IFoA Actuarial Reference Number or ARN. feedback on your progress.
Score and grade for this assignment (to be completed by marker):
Q1 Q2 Q3 Q4 Q5 Q6 Q7 Total
=_____%
9 12 10 11 5 6 27 80
Grade: A B C D E Marker’s initials: ________
Please tick the following checklist so that your script can be marked quickly. Have you:
[ ] Checked that you are using the latest version of the assignments, ie 2019 for the sessions leading
to the 2019 exams?
[ ] Written your full name in the box above?
[ ] Completed your ActEd Student Number in the box above?
[ ] Recorded your attempt conditions?
[ ] Numbered all pages of your script (excluding this cover sheet)?
[ ] Written the total number of pages (excluding the cover sheet) in the space above?
[ ] Included your Marking Voucher or ordered Series X Marking?
Please follow the instructions on the previous page when submitting your script for marking.

Feedback from marker
Notes on marker’s section
The main objective of marking is to provide specific advice on how to improve your chances of
success in the exam. The most useful aspect of the marking is the comments the marker makes
throughout the script, however you will also be given a percentage score and the band into which
that score falls. Each assignment tests only part of the course and hence does not give a complete
indication of your likely overall success in the exam. However it provides a good indicator of your
understanding of the material tested and the progress you are making with your studies:
A = Excellent progress B = Good progress C = Average progress

D = Below average progress E = Well below average progress
Please note that you can provide feedback on the marking of this assignment at:
www.ActEd.co.uk/marking

SP2: Assignment X1 Questions Page 1
X1.1 Compare the revalorisation approach to distributing with-profits surplus (as used in continental
Europe) and contribution method dividends (as used in North America). [9]
X1.2 A life insurance company in the Duchy of Ruritania has a portfolio of in-force regular premium
unit-linked endowment assurance contracts. The contracts are designed to provide income in
retirement, and all contracts are written to age 65 (compulsory retirement age for men and
women in Ruritania), whatever the age at which the contract is taken out. Under Ruritanian
pensions legislation:
 At least 50% of the maturity proceeds from the policy must be used to buy a pension
annuity for the policyholder at age 65. The rest may be taken as a cash lump sum.
 The initial premium chosen by the policyholder is automatically increased in line with IRRP
(Index of Ruritanian Retail Prices). The policyholder may stop paying premiums at any
time without penalty. However, the policyholder is not allowed to surrender the contract
for an immediate cash sum.
The company is allowed to revise its charges under the contract.
The contracts include a guaranteed minimum annuity rate (ie a rate for converting the
endowment proceeds to income) applying at age 65. The company’s non-guaranteed rates
mostly depend on yields on Ruritanian government stock and expected mortality rates at the date
the annuity is purchased. Under Ruritanian anti-discrimination law, annuity rates (guaranteed
and non-guaranteed) must be the same for men and women.
Discuss the principal areas of risk underlying the contracts for the life insurance company. [12]
You are not required to comment on the risks faced by the company from the annuity once it has
started to be paid, although any effect of the guaranteed minimum annuity rate should be
discussed.
X1.3 (i) Describe the types of bonus which may be used under the ‘additions to benefits’
approach to distributing surplus on conventional with-profits policies. [6]
(ii) Comment on which of the types of bonus are most suitable for distributing:
(a) income from assets
(b) capital appreciation on equities
(c) surplus arising from a one-off event. [4]

[Total 10]

Page 2 SP2: Assignment X1 Questions
X1.4 A Spanish life insurer is about to offer an index-linked contract in the following form in its
domestic market.
A single premium will be paid at outset. Premiums will be accepted in the two months leading up
to launch date, and will earn deposit interest between being paid and the launch date.
The contracts will be for a fixed five-year period from the launch date. The maturity benefit will
be the premium paid indexed by the average of the changes in the Nikkei 300 equity index in
Japan and the Standard & Poor’s 500 equity index in the United States. A ‘maturity bonus’ of 5%
of the final value will also be paid.
No guaranteed minimum return will be given, and so the benefit will not necessarily exceed the
premium paid. The insurer does not intend to invest in derivatives.
The insurer aims to cover its expenses and the cost of the maturity bonus and make a profit by
retaining the dividend income on the investments backing the contracts.
On death the benefit will be the indexed value of the premium. On withdrawal before the five-year
period is completed the benefit will be the premium paid or the indexed value, whichever is the
lower.
Describe the risks that the insurer may face as a result of writing this contract. [11]
X1.5 Explain what is meant by the term ‘earned asset share’ in the context of with-profits life insurance
contracts. [5]
X1.6 A life insurer writes two types of without-profits term assurance contract.
The first (contract A) is a regular premium ten-year contract sold to individuals. The fixed sum
assured is limited to a maximum of ten times the policyholder’s salary at the time the contract is
taken out. The larger the sum assured, the greater the extent to which the life insurer carries out
investigations of the applicant’s health before agreeing cover.
The second contract (contract B) is a group term assurance sold to employers to enable them to
provide life cover as a benefit to their employees. The contract is a collection of individual single
premium one-year term assurances. The benefit for all members is five times salary. The contract is
renewable each year, but on non-guaranteed terms.
Under contract B, the employer provides a list each year of employees covered, along with
information on salaries, ages and sexes, and a single aggregate premium is charged. The contract is
only offered where membership is compulsory for all employees. No investigations of members’
health are conducted before acceptance for cover.
Discuss how the mortality risks for the life insurance company will differ for the two contracts. [6]

X1.7 (i) Describe the key features from a policyholder’s perspective of the following contract
structures, in relation to level regular premium whole life assurance products:
(a) conventional without-profits
(b) unit-linked
(c) accumulating with-profits. [14]
An individual aged 33 wishes to take out a whole life assurance contract (on their own life) for the
benefit of their child, currently aged 4 years.
(ii) Give possible reasons why the individual is considering taking out such a policy. [3]
(iii) Discuss:
 the extent to which each of the structures listed in part (i) meet the possible
needs for the contract
 the risks that would be accepted in each case by the policyholder in taking out the
policy. [10]
[Total 27]
END OF PAPER





using the course.

2019 Examinations

1. Please:
illegible scripts
2. Please do not:



in size.

2019 Examinations
Name:


ActEd@bpp.com.
=_____%
12 6 13 12 8 14 15 80
to the 2019 exams?
[ ] Rated your X1 marker at www.ActEd.co.uk/marking?



X2.1 A life insurance company has to date sold its business only through its own salesforce. It sells only
unit-linked business, its main product being a unit-linked whole life contract. The contract is
designed to provide high levels of protection with only a small residual savings element.
Premiums are reviewable by the company every five years.
The company now wishes to sell this contract through insurance intermediaries as well as its own
salesforce. Assuming the company successfully carries out its plan:
(i) Explain why the mortality experience of the business is likely to change. [3]
(ii) Explain how the company’s actuarial pricing basis may change. [9]
[Total 12]
X2.2 (i) Explain how it might be possible for someone to buy a life insurance product that is not
the most appropriate product available to meet their needs. [5]
(ii) State the main risks run by a life insurance company of selling inappropriate products to
customers. [1]
[Total 6]
X2.3 A life insurance company that writes mainly with-profits business is performing an investigation
that will involve projecting the value of the total assets and liabilities of the company for each of
the next ten years.
(i) Describe the possible approaches to projecting the investment return on the assets. [7]
The company has used its asset-liability projection model as the basis for its decision-making.
(ii) Explain, including suitable examples, how investment risk can arise from three distinct
sources of error in the investment return assumption used in this model. [6]
[Total 13]
X2.4 A life insurance company is planning to launch a ten-year, regular premium unit-linked
endowment assurance. It plans to guarantee that the maturity value will never be less than the
total amount of premiums paid.
The marketing manager has suggested that as the company is willing to offer such a guarantee at
maturity then it could also offer a return of premiums paid on surrender of the contract at any
time, as this would make the contract much more marketable.
(i) Comment on the extent of the risk for the company in giving the maturity guarantee. [5]
(ii) Set out the main points that should be made in reply to the marketing manager’s
suggestion that guaranteed surrender values should also be offered. [7]
[Total 12]

X2.5 The insurance supervisory authorities in a particular country impose the following restrictions on
life insurance company investments:
 At least 60% of investment must be in government bonds.
 Remaining investment must be in quoted equity shares, quoted company fixed-interest,
or cash.
 No overseas investment is allowed.
 No more than 1% of a company’s total assets may be invested in the equity or corporate
fixed-interest stock of an individual company.
(i) Comment on how such restrictions would protect policyholders. [3]
(ii) Explain how such restrictions might reduce benefits to policyholders. [5]
[Total 8]
X2.6 A life insurance company currently sells only without-profits term assurance business. All the
contracts are sold through its own salesforce.
In order to diversify its product range, it plans to launch a regular premium unit-linked pension
contract. The only charge levied under the contracts will be an annual amount equal to 1% of the
value of the units.
Discuss the principal risks the company faces in relation to the new pension product. [14]

X2.7 A proprietary life insurance company started selling business seven years ago, and has so far
specialised in selling regular premium unit-linked savings products. The company has a good sales
record to date, but is looking to expand further by offering products with protection benefits, on
competitive terms.
It therefore plans, for all new sales of its monthly-premium unit-linked whole of life products, to
include the following additional benefit features:
 on death, a sum assured equal to a certain multiple of the person’s salary at the start of
the policy, or the value of the unit fund (whichever is the higher amount)
 the multiple can be any value between 0 and 5, but once chosen the amount of death
benefit cannot be varied by the policyholder
 all applications for multiples greater than 0.5 will be subject to underwriting.
Charges for the product will be the same for all policies, except for the unit allocation rate, which
will vary (between 10% and 100% of every premium paid), according to the age of the
policyholder at entry, and the sum assured applied for. The unallocated part of each premium is
designed to cover just the cost of the death benefit.
The allocation percentage remains the same for the duration of the contract, unless changed as a
result of a charging rate review by the company, which the company is allowed to do at any time.
However, the minimum allocation rate of 10% will always apply.
Premiums are reviewable on every fifth policy anniversary.
Describe the new or increased risks the company faces if it were to carry out this proposal. [15]
END OF PAPER





using the course.

2019 Examinations

1. Please:
illegible scripts
2. Please do not:



in size.

2019 Examinations
Name:


ActEd@bpp.com.
=_____%
4 7 11 14 10 12 22 80
to the 2019 exams?



X3.1 A large insurer offers a wide range of unit-linked funds. One of the actuarial teams is responsible
for calculating the unit prices for each fund each day. Someone has suggested that this job must
be very easy as the unit price is simply the market value of the assets divided by the number of
units.
State the adjustments that need to be made to this suggested calculation in order to calculate the
unit price. [4]
X3.2 A life insurance company sells mainly unit-linked contracts. The Board of this company has
agreed a merger with another life insurance company that also sells mainly unit-linked contracts.
Both companies operate a range of unit-linked funds including equity, fixed-interest and property
funds.
An actuary employed by this insurance company has been asked to provide advice to the Board
regarding the merger of the unit-linked funds of the two companies.
Describe the key considerations for the new combined company regarding the calculation of unit
prices at the merger date. [7]
X3.3 A long-established but small mutual life company is reviewing the premium rates for conventional
without-profits term assurance business.
(i) Describe the basic features of a cashflow projection model that is used to price these
contracts. [4]
The Marketing Manager has said they believe that the products can only be sold if they are costed
with overhead expenses excluded from the expense loadings.
(ii) Comment on this suggested approach. [5]
For many years the company has written ‘family income benefit’ assurances with level premiums.
Under these policies, if the policyholder dies during the policy term, a guaranteed fixed annual
income is paid from the date of death up to the end of the original policy term.
(iii) Explain why the company’s contract has a premium payment term that is less than the full
term of the policy. [2]
[Total 11]
X3.4 A proprietary life insurance company is planning to re-launch its regular premium unit-linked
endowment assurance.
Outline the steps that would be followed in designing and pricing the product. [14]

X3.5 A life insurance company sells with-profits business using the contribution method with annual
cash dividends but no terminal dividend.
(i) Describe how the company’s model office should take this bonus method into account in
modelling such business. [2]
The current product range guarantees the distribution, as cash dividends, of 80% of interest
profit, expense profit and mortality profit (all defined with reference to the premium basis). The
chief executive has asked whether this can be raised to 90% for competitive reasons, and if not
then how else the company’s competitive position could be improved.
(ii) Describe the factors that should be taken into account in analysing this request, including
how the model might be used. [8]
[Total 10]
X3.6 A life insurance company has sold a unit-linked single premium savings contract for many years.
The policyholder may choose at outset to invest in either the European equity fund or the North
American equity fund, but subsequent switches between funds are not permitted. The term of
the contract is ten years. The benefit paid on death, surrender or maturity is the bid value of
units.
To increase the marketability of the contract, it has been suggested that a guarantee will be
added so that the maturity payout will be the higher of the bid value of units or the single
premium accumulated at 1% per annum. This guarantee will be charged for by a reduction in the
allocation rate.
(i) Describe how the company would use a stochastic model to determine the charge to be
taken for providing the maturity guarantee, including how each of a risk neutral
calibration and a real world calibration of the stochastic model could be used. [8]
(ii) Suggest four modifications to this product’s design that would improve its marketability,
commenting on how each of these modifications would affect the modelling performed in
part (i). [4]
[Total 12]

X3.7 A life insurance company issues a type of annual premium unit-linked policy that is designed to
provide substantial life insurance cover at a reasonable cost. The death benefit under the policy
is equal to the unit fund value or the sum assured, if higher. On each policy anniversary during
the first ten years, the policyholder has an option to increase the sum assured by 10% of its then
current level, without evidence of health, and regardless of whether or not the option has been
exercised on previous occasions. The sum assured cannot be increased after the tenth policy
anniversary. The premium will increase, by the same proportion, whenever the policyholder
elects to increase the benefit level.
The minimum death benefit can be removed at the option of the policyholder at any time, but
cannot be reinstated once removed.
No other changes to the premium can take place during the first ten years (unless the policy is
surrendered).
The insurance company will review the level of premium at the tenth policy anniversary, and
every five years thereafter.
Charges are deducted annually from the unit fund to pay for the mortality cost. The charges are
based on an assumed mortality rate for the policyholder’s then age, applied to the then sum at
risk. The rates of mortality assumed are guaranteed under the contract for the first ten years, but
can be reviewed by the company every five years thereafter.
The other charges and relevant terms under the contract, which are fixed for as long as the
contract is in force, are as follows:
 Bid-offer spread: 5%
 Fund management charge: 1% of the unit fund per annum
 Allocation rates: Year 1 = 60%
Year 2 = 80%
Year 3 = 90%
All subsequent years = 100%
 The surrender value is equal to the bid value of the units at any time.
(i) Comment on the design of this policy under the following headings:
(a) marketability [5]
(b) the risks borne by the company [8]
(c) the financing requirements of the contract. [3]
(ii) Describe ways in which the company could make improvements to the product design, in
relation to the headings listed in part (i), noting any conflicts that may arise. [6]
[Total 22]
END OF PAPER





using the course.

2019 Examinations
Time allowed: 3¼ hours

1. Please:
illegible scripts
2. Please do not:



in size.

2019 Examinations
Name:


Note: If you take more than 3¼ hours, you should
ActEd@bpp.com.
Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Total
=_____%
15 5 6 11 15 13 14 13 8 100
to the 2019 exams?



X4.1 The Government Life Assurance Monitoring Office (GLAMO) is the regulatory authority in the
country of Megalomania. Historically, only without-profits business has been written in
Megalomania. Unit-linked business does not exist for statutory reasons, but legislation has just
been passed enabling companies to write such business.
GLAMO is in the process of reviewing the current regulations on how life insurance companies
demonstrate supervisory solvency.
The current system is:

 liabilities are valued using the gross premium formula method, an interest rate of 4% pa
and assuming industry expenses and population mortality
 assets are valued at book value (defined as the price at which they were purchased)
 policy reserves must not be lower than any guaranteed surrender value
 negative reserves are allowed but only up to a maximum of 50 million Megalomanian
dollars per company
 solvency capital is required equal to 7.5% of the previous year’s written premiums.
Discuss how effective this system is likely to be in measuring the solvency of life insurance
companies in Megalomania. [15]
X4.2 Describe the market-consistent approach to valuing an existing portfolio of term assurance
liabilities, ignoring expenses. [5]
X4.3 The valuation regulations in a particular country require that reserves are calculated using
prudent assumptions (no solvency capital is required).
Describe the problems that would be caused if the supervisory authority were to impose a
valuation basis that was too prudent. [6]
X4.4 A life insurance company is repricing its conventional without-profits term assurance product,
because the company’s sales of the product (through insurance intermediaries) have declined
significantly over the past three months.
The pricing actuary is first considering the mortality assumption that should be used in the pricing
basis for this product.
Describe the sources of data that would be expected to be available for this purpose, and how the
actuary would arrive at a choice of assumption to use. [11]

X4.5 The supervisory authority of a particular country has decided to strengthen the supervisory
solvency regulations. It has just issued the following statement:
‘In order to ensure that life insurance companies are capable of meeting their future financial
commitments as they fall due, and having regard to the uncertainties prevalent in the economy,
in the life insurance marketplace and in the field of human health, as from 31 December 2018 all
life insurance companies will be required to reduce the valuation rate of interest specified in
legislation from its current level of 6.5% pa to a new level of 4.5% pa. In addition, companies
must hold solvency capital to cover the following stresses:
(a) per-policy administration expenses increase over the next year by 25%
(b) the mortality of life assurance policyholders immediately worsens by 5%, while that of
annuitants improves by 1% pa more than previously assumed.’
Discuss what effect the changes might be expected to have on the solvency requirements and
financial management of life insurance companies operating in that country. [15]
X4.6 A life company, which commenced operating three years ago, transacts all types of unit-linked
individual life assurance, pensions and annuity business. Each year the embedded value of the
company is calculated, and the change in embedded value is reported as profit in the group
accounts of its parent company, a clearing bank. The company is about to begin the annual
calculation of the embedded value for this purpose.
(i) Describe how the embedded value will be calculated for this company. [3]
(ii) List the financial and economic assumptions required. [2]
(iii) State the principles that should be followed in determining each of the other principal
assumptions (mortality, persistency and expenses). [8]
[Total 13]
X4.7 In the country of Ruritania, insurers sell all forms of conventional without-profits contracts. In
order to protect policyholders from unscrupulous life insurance companies, the supervisory
authority of Ruritania has proposed that the surrender value of all life insurance policies be equal
to the total of policy premiums paid to date, adjusted as follows:
 Premiums are rolled up at the Ruritania Consumer Price Inflation Index for each year
concerned.
 Companies may do the calculation net of the cost of any mortality cover, calculated
assuming the mortality rates of the Ruritania Population Life Tables 1995.
 There is a deduction to allow for the life company’s expenses. This is equal to 10% of all
premiums paid to date.
(i) Discuss the implications of this proposal for life insurance companies in Ruritania. [12]
(ii) Describe the implications for Ruritanian policyholders. [2]

[Total 14]

X4.8 A life company sells 25-year without-profits endowment assurances to lives aged 30 exact. The
sum assured is 50,000. The death benefit is payable at the end of the year of death. Level
premiums of 82.20 are payable monthly.
The premium and reserving basis is as follows:

Mortality: AM92 Select
Interest: 6% pa
Expenses: Initial, 30% of the first year’s premiums, payable at the start of the contract
Renewal, 5% of all premiums, including the first, payable at the start of each year.
(i) Calculate the gross premium prospective policy value just before the start of the tenth
year of the policy. [6]
Immediately before the payment due at the start of the tenth year, the policyholder wishes to
alter the policy by increasing the sum assured to 80,000, and increasing the total term of the
policy to 30 years.
(ii) Calculate the new monthly premium payable under the policy, allowing for an alteration
expense of 150. [5]
(iii) Comment on your answer to part (ii). [2]

[Total 13]
X4.9 An insurer sells regular premium without-profits endowment assurances. The insurer is reviewing
the methodology and basis that it will use to calculate the embedded value of these contracts in
its accounts.
(i) Define a passive valuation approach and an active valuation approach. [2]
(ii) List the items that the insurer will need to choose a methodology and basis for in order to
calculate the embedded value. [3]
(iii) Give examples of how these items could be calculated using:

 a passive valuation approach
 an active valuation approach. [3]
[Total 8]
END OF PAPER





using the course.

2019 Examinations

1. Please:
illegible scripts
2. Please do not:



in size.

2019 Examinations
Name:


ActEd@bpp.com.
Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Total
=_____%
13 5 17 8 11 21 8 10 7 100
to the 2019 exams?



X5.1 A life company issues unit-linked personal pension contracts. These are written as pure
endowments, with the proceeds at maturity being available either as cash, or to purchase a
conventional without-profits immediate annuity.
The company guarantees that the annuity terms available at the specified retirement age will not
be worse than those calculated on a 4.5% pa interest rate basis and according to a specified
mortality basis.
(i) Describe how the company would determine a prudent reserve to cover the cost of the
guarantee using stochastic simulation. [5]
(ii) Outline how your answer to part (i) would differ if the policyholder could choose their
retirement age at any time, rather than needing to specify it when taking out the
policy. [1]
(iii) Describe how your answers to parts (i) and (ii) would differ if the company were costing
the guarantee for pricing purposes rather than reserving purposes. [2]
As an alternative to valuing the guarantee using stochastic simulations, the company is

considering valuing the guarantee using option pricing techniques. The value of the guarantee
will then be set equal to the price of the derivative contracts that match the payments from the
guarantee.
(iv) Outline two different derivative contracts that could be used to price this guaranteed
annuity option. [2]
(v) Outline the problems that the company might face if it tried to hedge the guarantee by
buying the derivatives in part (iv). [3]
[Total 13]
X5.2 A life insurance company, which commenced operating three years ago, transacts all types of
unit-linked individual life assurance, pensions and annuity business. Each year the embedded
value of the company is calculated, and the change in embedded value is reported as profit in the
group accounts of its parent company, a clearing bank. The company is about to begin the annual
calculation of the embedded value for this purpose.
Outline the steps that should be taken to prepare policy data suitable for the calculation
process. [5]

X5.3 An insurance company sells single premium term assurances with terms of one and two years.
The sum assured for these contracts is 10,000 and is payable at the end of the year of death.
The company uses the following pricing basis for these standard contracts:
 Mortality: 1% pa
 Interest: 3.5% pa compound.
(i) Calculate the single premiums for the one-year and two-year contracts. [2]
The insurance company also sells special single premium term assurances with a term of three
years. The initial sum assured for these contracts is 40,000. The sum assured is payable at the
end of the year of death.
For these special contracts, the policyholder has the option to increase the sum assured to 50,000
without further evidence of health. The option can be exercised either at the end of the first year
or at the end of the second year. If the option is exercised, the policyholder pays an additional
single premium calculated using the standard terms as in part (i).
(ii) Discuss whether the special contract should be underwritten. [2]
The company uses the following pricing basis for these special contracts:
 Mortality after exercising the option: 3% pa
 Mortality otherwise: 1% pa
 Proportion of eligible lives who exercise the option at time 1: 50%
 Proportion of eligible lives who exercise the option at time 2: 30%
 Interest: 3.5% pa compound.
(iii) Calculate the single premium for the special contract using the basis above. [5]
The company has undertaken a review of its premium calculation methodology. One reviewer
stated that the pricing basis only allows for mortality, option take-up rate and interest, but should
also include other factors.
(iv) State:
 the factors that should also be included in the premium calculation
 in each case, whether including the factor would increase or decrease the
premium for the special term assurance. [4]
Sales of the special term assurance have been low.
(v) Suggest possible reasons for the low sales. [4]

[Total 17]

X5.4 A small proprietary life insurance company writes only unit-linked business. It has decided to
review its reinsurance arrangements.
Outline the possible reasons for reinsurance and the types of arrangement that might be suitable
for this company. [8]
X5.5 A life insurance company writes a significant volume of group life cover benefits associated with
the death-in-service element of occupational pension schemes.
(i) Describe how the company might review its retention limit for this portfolio. [5]
The company establishes a one-third quota share reassurance arrangement with Reinsurer A. For
the balance of risk, it establishes a surplus treaty with Reinsurer B. This provides surplus cover in
excess of 80,000 up to a maximum reinsured of 600,000.
(ii) Calculate how the claim payment would be apportioned between the company and its
two reinsurers in the event of a total claim on a single life of:
(a) 90,000
(b) 290,000
(c) 1,000,000
(d) 1,200,000. [2]
(iii) Describe the reinsurance methods available to the company to protect itself against very
large claims. [4]
[Total 11]
X5.6 A proprietary life insurance company operates in a country where it is estimated that between 5%
and 10% of adults are HIV-positive or are sick with AIDS, and this figure is still gradually growing.
The company writes significant amounts of term assurance, which represents about 15% of its
business by new premium income. All its contracts are non-unitised and have guaranteed
premiums.
All of the company’s policies currently have an HIV / AIDS exclusion clause, such that a claim is
disallowed if the cause of death is found to be HIV or AIDS related.
The government of the country has just announced a law banning the use of such exclusion
clauses. It is expected to come into force within three months and will apply to in-force as well as
new business.
Discuss the possible implications for the company of this proposed law, including the actions that
it could take in response. [21]
X5.7 Describe the sources of information typically used by life insurance companies to underwrite
proposals for life insurance, and the procedures for determining the information to be obtained in
individual cases. [8]

X5.8 A life company writes a full range of protection products in the life fund. It is currently reviewing
its underwriting procedures and in particular is considering the introduction of a non-smoker
discount. Many of its competitors offer a non-smoker discount to policyholders prepared to sign
the statement ‘I have not smoked in the last year and do not intend to start smoking in the
future’.
Discuss briefly the advantages and disadvantages of each of the following possibilities:
(a) The company continues not to offer non-smoker discounts.
(b) The company offers a discount on all contracts based on the same question as its
competitors.
(c) The company offers a discount to people prepared to state that they are non-smokers,
have never smoked, do not intend to smoke, and who undergo a rigorous medical
examination to ensure that they are not suffering from any smoking related ailments.
(d) The company offers a discount on similar terms to other companies except that the
discount lapses, and the past discount is repaid to the company, if the policyholder starts
smoking in the future.
(e) As (b) but only on term assurances and whole life assurances, not endowment
assurances. [10]
X5.9 A life insurance company adopts a strict underwriting approach under a particular class of
business.
(i) Explain why it may also wish to use reinsurance to manage the mortality risk under the
business. [4]
(ii) Explain why, having entered into a reinsurance treaty, it would still wish to underwrite the
business. [3]
[Total 7]
END OF PAPER

2019 Examinations

1. Please:
illegible scripts
2. Please do not:



in size.

2019 Examinations
Name:


ActEd@bpp.com.
=_____%
9 6 12 24 10 17 22 100
to the 2019 exams?



X6.1 A medium-sized life insurance company that writes all types of business is designing a new
product.
In order to derive the expense assumptions to be included in the profit tests, a detailed expense
investigation has been carried out.
Explain how the following costs would be translated into expense assumptions for the profit tests:
(a) Overhead costs
(b) Branch office costs (staff and property)
(c) The cost of a replacement mainframe computer
(d) The building costs of a new head office. [9]
X6.2 A life insurance company writes only individual pension and term assurance products. All aspects
of new business processing are performed in a single department. The premium rates for the
stand-alone term assurance contract are due to be reviewed.
It has been suggested that the loading for initial expenses, other than commission, should be
derived by dividing the total expenses incurred by the department over the preceding three
months by the total number of pension and term assurance policies processed over this period.
Explain why this might not be appropriate. [6]
X6.3 A life insurance company is reviewing its premium rates for term assurance.
Discuss how it could determine an appropriate persistency assumption to use in the profit
test. [12]
X6.4 A small proprietary life insurance company is planning the launch of a new single premium
with-profits pure endowment pension contract, to take advantage of new legislation. This
legislation allows favourable tax treatment of premiums for policyholders with pure endowment
policies of term at least ten years and maturity at age 60.
The insurer currently has some with-profits business, which uses the revalorisation method that is
prevalent in the country concerned. However, the company decides to use the UK-style
accumulating with-profits bonus system for the new product. The death and maturity benefit
under the policy will be the value of the benefit fund, plus any terminal bonus payable, at the
time of claim.
(i) Discuss how the pricing of the product might allow for the operation of this surplus
distribution method, including how the cost of any guarantees might be incorporated. [5]
(ii) Describe the major risks to the company that arise from the adoption of this bonus
system. [11]
(iii) Discuss how the bonus system will affect the financial management of this product. [8]
[Total 24]

X6.5 (i) Explain why the financial strength of a life insurance company can be measured by the
level of its free assets. [7]
(ii) Explain why a life insurance company might not want too large a level of free assets. [3]
[Total 10]
X6.6 The Managing Director of a large life insurance company has noticed that the company’s
persistency rates appear higher than those shown in the analysis of industry persistency produced
by ALAC, the Association of Life Assurance Companies. However, the company’s own last report
on persistency rates showed that, for its major product classes, persistency rates seem lower than
those of the company’s main competitors.
The Managing Director, not trusting the company’s own results and analysis, has suggested that it
would be more prudent if the company were to use the ALAC rates for future pricing, reserving
and embedded value work.
(i) Give possible reasons for the apparent discrepancy in the two comparisons. [4]
(ii) Outline the checks that could be made to ensure that the results for the company’s own
experience are reliable. [3]
(iii) Discuss the advantages and disadvantages of the Managing Director’s suggestion. [10]
[Total 17]
X6.7 (i) State the principles of investment in respect of life insurance business. [3]
(ii) Discuss, in the light of these principles, the suitability of each of the investments listed
below for a large life insurance company with ample free assets writing a complete range
of products:
(a) long-dated domestic government stocks
(b) fixed-interest policy loans (these are loans made to policyholders against the
security of their policy value)
(c) ordinary shares (domestic market)
(d) direct investment in real property
(e) overseas ordinary shares.
You should cover the possibility that with-profits business is written in conventional form,
using either the UK-style ‘additions to benefits’ system or the revalorisation system.
[15]
(iii) Describe how your answer to part (ii) would have been different if the company had been
small, with limited free assets. [4]
[Total 22]
END OF PAPER

Assignment deadlines
For the session leading to the April 2019 exams – SP Subjects
Marking vouchers
Subjects Assignments Mocks
SP1, SP2, SP4, SP7, SP8 13 March 2019 20 March 2019
SP5, SP6, SP9 20 March 2019 27 March 2019
Series X Assignments
Recommended
Subjects Assignment Final deadline date
submission date
SP1, SP2, SP4, SP7, SP8 5 December 2018 9 January 2019

X1
SP5, SP6, SP9 12 December 2018 9 January 2019
SP1, SP2, SP4, SP7, SP8 12 December 2018 23 January 2019

X2
SP5, SP6, SP9 19 December 2018 23 January 2019
SP1, SP2, SP4, SP7, SP8 2 January 2019 6 February 2019

X3
SP5, SP6, SP9 9 January 2019 6 February 2019
SP1, SP2, SP4, SP7, SP8 16 January 2019 20 February 2019

X4
SP5, SP6, SP9 23 January 2019 20 February 2019
SP1, SP2, SP4, SP7, SP8 30 January 2019 6 March 2019

X5
SP5, SP6, SP9 6 February 2019 6 March 2019
SP1, SP2, SP4, SP7, SP8 13 February 2019 13 March 2019

X6
SP5, SP6, SP9 20 February 2019 20 March 2019
Mock Exams
Recommended
Subjects Final deadline date
submission date
SP1, SP2, SP4, SP7, SP8 6 March 2019 20 March 2019
SP5, SP6, SP9 13 March 2019 27 March 2019
We encourage you to work to the recommended submission dates where possible.
If you submit your mock on the final deadline date you are likely to receive your script back less than a week
before your exam.

Assignment deadlines
For the session leading to the September 2019 exams – SP Subjects
Marking vouchers
Subjects Assignments Mocks
SP2, SP8 21 August 2019 28 August 2019
SP1, SP4, SP5, SP6, SP7, SP9 28 August 2019 4 September 2019
Series X Assignments
Recommended
Subjects Assignment Final deadline date
submission date
SP2, SP8 5 June 2019 17 July 2019

X1
SP1, SP4, SP5, SP6, SP7, SP9 12 June 2019 24 July 2019
SP2, SP8 19 June 2019 24 July 2019

X2
SP1, SP4, SP5, SP6, SP7, SP9 26 June 2019 31 July 2019
SP2, SP8 3 July 2019 31 July 2019

X3
SP1, SP4, SP5, SP6, SP7, SP9 10 July 2019 7 August 2019
SP2, SP8 10 July 2019 7 August 2019

X4
SP2, SP8 24 July 2019 14 August 2019

X5
SP2, SP8 31 July 2019 21 August 2019

X6
SP1, SP4, SP5, SP6, SP7, SP9 7 August 2019 28 August 2019
Mock Exams
Recommended
Subjects Final deadline date
submission date
SP2, SP8 14 August 2019 28 August 2019
SP1, SP4, SP5, SP6, SP7, SP9 21 August 2019 4 September 2019
We encourage you to work to the recommended submission dates where possible.
If you submit your mock on the final deadline date you are likely to receive your script back less than a week
before your exam.

SP2: Assignment X1 Solutions Page 1
Assignment X1 – Solutions
Solution X1.1
This topic is covered in Chapter 7.
Both approaches are essentially trying to do the same thing: to distribute the profits (or some of
them) from various sources to with-profits policyholders. [½]
For the contribution method, these sources include investment return, mortality and expenses.
[½]
For revalorisation, only investment surplus is usually distributed to policyholders. [½]
In the revalorisation system, the regular distributions are given as increases to the guaranteed
benefits under the contract. [½]
The contribution method pays out profit through regular distributions, and sometimes through a
terminal distribution. [½]
Under the contribution method, the regular distribution may be given as a cash dividend to the
policyholder (perhaps as an offset to premiums). [½]
However, the dividend may also be converted into a ‘paid-up’ benefit resembling the
revalorisation bonus. [½]
Under the revalorisation method, future premiums usually increase as well as future benefits, at
the same rate as each other. [½]
Under the contribution method, future premiums do not increase. [½]
There is some smoothing of the amount of profit distributed each year under both systems,
though using slightly different mechanisms. [½]
With the contribution method, the company can exercise some discretion over how much to
distribute each year. [½]
However, the split of profits between the different groups of policyholders will be driven by a
formula. [½]
The revalorisation method uses a defined formula to distribute profits, with no discretion, but
smoothing can occur relative to market returns by distributing book value profits. [1]
With the contribution method only, any terminal payment allows profit to be held back against
adverse future experience. [½]
This makes it suitable for dealing with volatile sources of profit such as capital gains on equities.
[½]
Part or all of any contingency margin can also be released in the terminal payment. [½]

Page 2 SP2: Assignment X1 Solutions
In both systems policyholders are grouped in some way … [½]
… and there is a broad aim to give back to policyholders in each group their ‘share’ of profits
arising. [½]
In the North American system there is much more specific focus on the different sources of profit
than for revalorisation (and is in most cases irrelevant for revalorisation as investment profit is
usually the only source that is distributed anyway). [½]
There is also a greater emphasis on distributing the specific sources of profit to narrower groups
than for revalorisation, even in relation just to investment profit. [½]
[Maximum 9]
Solution X1.2
Unit-linked contracts are covered in Chapter 4. However, we will discuss risk further in Chapters
10 and 11.
The instruction in the question to ignore risks after the annuity has started to be paid does not mean
that students should ignore the risks at the time the cash or annuity option is selected. This is clearly
a significant element of the risks of the initial contract.
The major areas of risk for the life insurance company are:
Mortality [½]
There is a mortality risk because the death benefit will exceed the value of the unit fund for many
(about ten?) years. [½]
Also, the unit fund itself is likely to exceed the asset share at outset, depending on initial expenses
and contract design. [½]
The risk should largely disappear for contracts of long duration in force. [½]
In any case, the death benefit is not particularly high for longer term contracts. The contract is
fundamentally for savings. [½]
Expenses [½]
The company has the right to vary the charges under the contract and so is in theory protected
against unexpected increases in costs. [½]
This protection may be limited by competition, legislation or policyholder expectations. [½]
The indexation of premiums also reduces the risk, as the increased funds are likely to produce
higher income from charges. [½]
In fact, the company may be at risk of lower than expected inflation if it is relying on increases in
premiums to achieve profitability. [½]
Higher inflation might be associated with higher nominal investment returns. If so then any
fund-related charges would increase. [½]

This would provide some protection against inflation. [½]
Rates of ceasing premiums [½]
If more policies than expected cease paying premiums then this is likely to reduce profitability
because of reduced charges income. [½]
The company has no right to levy a penalty in these circumstances and so has little protection. [½]
The exact effect will depend on the types of charges used and how they are levied. [½]
For example, a regular monthly admin fee might be taken from premiums before they are
invested (in which case this would be lost if premiums stopped) or by cancellation of units (in
which case the charges would continue to be taken). [½]
One problem with the cancellation approach is that it can make a significant dent in the unit funds
on small, ‘paid-up’ policies.
The potential loss from policies being made ‘paid-up’ is worsened here by the indexation of
premiums. [½]
The amount lost becomes more significant, particularly if Ruritanian inflation is high. [½]
Investment [½]
The company is only indirectly affected by investment performance on the endowment, via any
fund-related charges. [½]
Also, there is a very small additional mortality risk if unit performance is poor, because this
increases the sum at risk. [½]
Guaranteed annuity rate (ie conversion terms) [½]
The fundamental risk here is that the guaranteed terms for converting cash to income prove more
generous than those the company would offer on the basis of conditions when the policyholder
retires. [1]
Specifically:
1. Investment returns assumed within the guaranteed annuity rate may be higher than
those actually available at retirement. [½]
2. Expected future mortality at the retirement date may be lighter than that underlying the
guaranteed rate. [½]
3. The actual expenses of the annuity may be higher than allowed for in the guaranteed rate.
[½]
The need to offer the same terms to men and women is not necessarily a significant extra risk, as
all life insurance companies in Ruritania are required to offer equalised rates. [½]
There is a possible risk if the company somehow ends up with an unexpectedly high proportion of
women at retirement (assuming that women have longer life expectancy). [½]

There is also a theoretical possibility that women will choose to take a higher proportion of their
benefits as income rather than cash, as the terms for conversion are more attractive to women.
[½]
Capital strain [½]
Initial reserves and solvency capital should not be too onerous as charges are reviewable … [½]
… whilst the annuity rate guarantee and any other future losses may not require significant
reserves until later in the policy term. [½]
Hence new business strain could be minimal provided a suitable charging structure for initial
expenses is employed. [½]
[Maximum 12]
Many of the points concerning the guaranteed annuity rate could equally well have been dealt
with under the headings of investment, mortality etc above. Markers should give credit for valid
points whichever structure is adopted.
Solution X1.3
(i) Additions to benefits bonus types
There are three types of bonus: regular reversionary, special reversionary and terminal. [½]
Regular reversionary
A regular reversionary bonus is declared throughout the lifetime of a contract, usually once a
year. [½]
It is an addition to the basic benefit under the contract and once declared it cannot be taken
away. [½]
The amount of the bonus can be calculated in one of three ways:

 simple – the bonus is expressed as a percentage of the basic benefit under the
contract [½]
 compound – the bonus is expressed as a percentage of the basic benefit plus any already
attaching bonuses [½]
 super-compound – the bonus is expressed in terms of two percentages: one applied to
the basic benefit and a second applied to any already attaching bonuses. Where this
method is used the second percentage is typically higher than the first. [1]
Special reversionary bonus
These are ‘one-off’ bonuses given during the term of the contract in addition to any regular
reversionary bonuses. [½]
Once added to the benefits these also cannot be taken away. [½]

Terminal bonus
Terminal bonus may be paid when a policy becomes a claim (maturity, death or sometimes
surrender). [½]
The bonus given to a particular contract may be specified in a number of different ways, for
example:
 a percentage (possibly varying by duration in force) of total attaching reversionary
bonuses, including any special reversionary bonuses [1]
 a percentage of the total claim amount (before addition of terminal bonus) with the
percentage varying according to duration in force. [1]
[Maximum 6]
(ii) Which types of bonus are suitable?
(a) Income could be distributed by any of the types of bonus. [½]
However, if it is reasonably stable and predictable … [½]
… then it is well suited to distribution by regular reversionary bonus. [½]
(b) Capital appreciation on equities tends to be volatile … [½]
… and so is best distributed via terminal bonus. [½]
A proportion of capital appreciation built up over a period could be distributed as a special

reversionary bonus. [½]
(c) If a ‘one-off’ surplus is distributed as regular reversionary bonus then this may create
expectations that cannot be fulfilled. [½]
A special reversionary bonus or terminal bonus are therefore suitable. [½]

[Total 4]
Solution X1.4
Index-linked contracts are covered in Chapter 4. We will discuss risk further in Chapters 10 and 11.
Investment [½]
The extent to which this is a risk for the insurer will depend on the extent to which it can and does
match the benefit. [½]
To do so exactly it would need to split the premiums equally between the two markets and then
buy stocks in proportion to their market capitalisation. [½]
As the constituents of the index change the insurer would have to buy and sell stocks accordingly.
[½]
If it did all this then the investment risk from the overseas markets would be eliminated (but this
would be difficult to do). [½]

In practice it will probably try to broadly match the movements of each index with a carefully
chosen selection of fewer stocks. [½]
To the extent that these move differently from the whole index there is an investment risk from
which the company may lose or gain. [½]
The company might even choose to mismatch more aggressively, for example by putting more
than 50% into the market it expects to perform better. This would be a very risky strategy. [½]
Currency
These points might equally well be made under ‘investment’.
The company is guaranteeing the movements in the index, without any mention of adjustment to
local currency. [½]
There is therefore a significant currency risk. [½]
If the company invests in the Japanese and US markets and the yen or dollar depreciates against
the euro then there is potentially a large loss from the mismatch between the liability to Spanish
policyholders and the value of the underlying assets. [1]
Expenses [½]
The company is at risk of the dividend income on its portfolio not being sufficient to cover
expenses, cover the maturity bonus, and produce profit. [1]
This should not be a major problem because dividend income is generally much more stable and
predictable than capital values over a period of five years. [½]
Income is also likely to increase in nominal value if inflation increases. [½]
However, both the above points ignore the currency risk to the income. [½]
As it is a new product launch there is also a risk that volumes will be insufficient to cover
development expenses. This is a theme that we shall pick up later in the course and so it is not fair
to first-timers to require it in the solution here.
Withdrawal (persistency) [½]
The company has largely protected itself against persistency risk by giving only the lower of
premium paid and indexed value, which will mean that on withdrawal the asset share will usually
exceed the surrender value, often by a considerable margin. [½]
It would probably lose money on early withdrawals if the markets had performed poorly, … [½]
… because it might not have recouped the initial expenses (ie not received enough dividends to
recover the initial expenses by that time). [½]
There may be a marketing risk on withdrawals late in the five-year term if markets have
performed well. [½]
A return of premium may attract ill-feeling and bad publicity. [½]

Mortality [½]
Higher than expected early deaths would cause a loss of future dividend income … [½]
… (which is likely to exceed ongoing expenses if the insurer is making a profit). This is likely to be
insignificant. [½]
Lower than expected mortality later in the term will increase the cost of the maturity bonus, ... [½]
… which is unlikely to be entirely covered by the corresponding increase in dividend income. [½]
This is also likely to be insignificant. [½]
Capital strain [½]
The initial reserves and solvency capital needed to cover the risks will probably lead to a capital
strain at entry. [½]
So the company is at risk from new business strain. [½]

[Maximum 11]
Solution X1.5
The earned asset share is a kind of retrospective policy value where the basis used is the
company’s past actual experience. [½]
For with-profits policies it is the accumulation of premiums less deductions associated with the
contract, … [½]
… plus a share of the company’s profits from its without-profits business, … [½]
… all accumulated at the actual rate of return earned on investments. [½]
Deductions include all expenditure associated with the contract(s), in particular:

 expenses incurred and any commissions paid [½]
 the cost of providing all benefits in excess of asset share, both life cover and any other
guarantees or options granted [½]
 tax (if appropriate) including any provisions made for future tax liabilities [½]
 transfers of profit to shareholders (if any) [½]
 the costs of any capital necessary to support contracts in the early years [½]
 any contribution to the free assets which will be the source of future bonuses, for
example to support smoothing. [½]
Returns from capital provided at later durations, to support other with-profits sales, may also be
added to the asset share. [½]
[Maximum 5]

Solution X1.6
Term assurances are covered in Chapter 2.
The main issues are:
Anti-selection
The anti-selection risk is probably greater on contract A. [½]
The individuals have choice over whether to seek cover and therefore some applications from
lives in poor health are likely. [1]
However, the insurer is making considerable efforts to counteract this risk and may be largely
successful. [½]
The compulsory nature of contract B removes most anti-selection risk. [½]
However, the total absence of underwriting could open the insurer to an anti-selection risk. [½]
Term and guarantees
Mortality rates are less predictable over ten years than one year. [½]
New diseases may emerge, antibiotics may stop working leading to world-wide pestilence etc. [½]
The shorter guarantees on contract B reduce the risks from mortality, all other things being equal,
… [½]
… as premiums may be revised. [½]
Size of benefit
The benefit as a multiple of salary is lower on contract B and so there is a smaller risk per individual.
[½]
Concentration of risk
There may be a ‘concentration of risk’ on contract B due to the non-independence of claims. An

industrial accident, for example, might lead to a large number of claims. [1]
Selective withdrawals
There is a risk of individual selective withdrawals on contract A but not on B. [½]
But it is possible that group schemes with particularly good experience are more likely to look
around the market for a better deal at renewal. [½]
This would be more likely if the insurer puts up rates at renewal. [½]
[Maximum 6]

Solution X1.7
Whole life assurances are covered in Chapter 2. The different contract structures (without-profits,
with-profits, unit-linked) are covered in Chapter 4.
(i)(a) Key features – conventional without-profits
This provides a fixed, guaranteed, sum assured payable on the policyholder’s death, whenever
that should occur. [1]
A surrender value would normally be payable, except perhaps at very early policy durations. [½]
The surrender value would normally not be guaranteed, although such a guarantee is possible. [½]
On premium discontinuance, the policyholder would normally be offered the alternative of

continuing the policy in paid-up form – for reduced benefits but with no future premiums
needed. [½]
(i)(b) Key features – unit-linked
Premiums would be paid into a unit fund. [½]
Charges are deducted on allocation of premiums, and also regularly from the unit fund … [1]
… in order to cover (often explicitly) the various costs incurred – such as investment costs, selling
expenses / commissions, and the cost of meeting claims. [½]
The value of the unit fund fluctuates (usually) daily, in line with a specified asset portfolio from
which the price of the units is determined. [½]
The benefit on death will be the unit fund, but this will generally be substantially increased, at
least at early policy durations, by having a minimum level of death benefit stipulated in the
policy. [1]
On surrender it is normal to pay the value of the units as a surrender value. [½]
This may be subject to a penalty, if surrender is so early that not all of the company’s initial costs
have been fully recouped by that time. [½]
There might be more flexibility about paid-up terms for (a) than (b), though the company would
reserve the right to lapse the policy if the unit fund were to reduce to zero at any point. [½]
A typical feature of many unit-linked policies is reviewability. [½]
Here the company has the right to vary charges (over all policies of a particular type) to reflect
changes in expected future experience. [½]
The policy will also usually be subject to periodic premium reviews, … [½]
… where the policyholder’s premium level may be adjusted to allow for any expected or known
changes in the future charges, … [½]
… and to reflect unit growth rates, both past and future. [½]

This usually makes total premium costs lower than for (a), … [½]
… though there is greater uncertainty of costs for the policyholder. [½]
Policies are usually flexible, in that sums assured and/or premium levels can be changed during
the policy term, … [½]
… though increases in death benefit would normally be subject to underwriting (evidence of good
health). [½]
(i)(c) Key features – accumulating with-profits (AWP)
The benefit is defined as the current value of an accumulating fund of premiums. [1]
There may be charges, in the form of deductions made from premiums, before the premiums are
added to the fund. [½]
The benefit fund will accumulate at some guaranteed rate, which is often zero. [½]
Bonus interest will also be added regularly to the benefit amount. [½]
The amounts of bonus interest are at the discretion of the company. [½]
The amount of death benefit will be the unit fund, possibly with an additional terminal bonus
paid, again at the discretion of the company. [1]
As with the unit-linked policy, there may be a minimum death benefit that would be paid, if the
normal death benefit was less than this at the time of death. [½]
Charges may be deducted from the benefit fund. These could be any of the charges described for
unit-linked, and essentially operate in an identical way except for the way in which the
accumulating fund benefit is calculated. [½]
The AWP structure may or may not be unitised. [½]
If it is, then the bonus interest can be added as an increase to the unit price, or as additional
units. [1]
On surrender of an AWP policy, a surrender penalty may be deducted from the fund benefit to
recover any initial expenses that have not been recouped by that time. [½]
There may also be a market value reduction – ie a downward adjustment to the value of the fund
benefit on surrender (only), to reflect the realistic asset share value. [1]
[Maximum 14]
(ii) Why take out a whole life policy
The first call on the money on death might be to pay for the individual’s funeral expenses, so
avoiding any financial burden on their child(ren). [½]
If the individual dies while the child is still a dependant, then the money can be used towards
paying for the costs of bringing up the child until independence has been reached. [1]

The individual will expect to pass on an inheritance to their child(ren) when they die, assuming
that they live to a reasonably old age. The policy will ensure that a substantial inheritance is paid
out to the child on the individual’s death, even if they die early and/or with only a low level of
life-time savings. [1]
There may be tax advantages to such an arrangement, eg by reducing the liability to inheritance
tax. [½]
[Total 3]
(iii)(a) Needs and risks – conventional without-profits policy
The contract provides a large sum of money when the policyholder dies, so this could meet all the
basic needs from the policy. [1]
The main risk is that the real value of the benefit is eroded by future inflation, as the sum assured
is fixed (usually) in monetary terms. If inflation is higher than expected then the child may still be
short of money (or have a lower than expected inheritance in real terms) on the parent’s
death. [1]
The benefit level is very inflexible to any changes in the policyholder’s situation: for example, to
increases or decreases in affluence (income and/or wealth). [½]
A period of unemployment, for example, might cause the policy to be surrendered, or at least
paid-up. [½]
While benefits can be increased (or restarted) in future by taking out new policies, even this is not
always possible if the individual’s health has deteriorated substantially. [1]
A further risk is that the insurer becomes insolvent in the future and becomes unable to pay the
promised benefits in full. [½]
(iii)(b) Needs and risks – unit-linked policy
A fixed minimum sum assured can meet the immediate need for a large sum on early death. [½]
The flexibility of the product means that minimum benefit levels and/or premiums can be varied,
at least to an extent, to reflect changing needs and circumstances. [½]
If the minimum death benefit is not too large relative to the premium, a point in future may arise
where the unit fund value exceeds the guaranteed sum assured. [½]
Subsequent growth in the unit fund, which could be invested in real assets, could then reduce the
risk of erosion by inflation. [1]
However, in that case (ie when the unit fund is larger than the guaranteed minimum benefit
level), the policyholder also takes on the risk of a much more uncertain claim payment. [1]
Increases in minimum benefit levels will probably be subject to underwriting requirements being
met – this is the same as for the other types. [½]

A possibly substantial risk is where the policyholder maintains too high a level of cover for too low
a level of premium. As a result, the unit fund might become exhausted by a particular age, which
will either require a very large increase in premium to restore the benefit level, or the
policyholder left with reduced cover or even no cover at all (policy lapses). [1]
(iii)(c) Needs and risks – accumulating with-profits policy
The comments are the same as for the unit-linked policy, except for where the unit fund value
exceeds the minimum death benefit. [½]
There is then no risk to the policyholder from downward fluctuations in fund value, which can and
does happen with unit-linked. [½]
Because of the guarantees inherent in the fund values, the company will need secure
fixed-interest securities to back them. This reduces the potential of the company to make real
returns on its assets, and so reduces the ability of the fund value to match inflation. [1]
On the other hand, if there is a terminal bonus the company would have more investment
freedom, enabling it to invest in some real assets and so help provide some inflation protection
for the policyholder. [½]
In terms of inflation protection, this policy type is likely to be intermediate between the other two
types. [½]
[Maximum 10]

Solution X2.1
(i) Why the mortality experience might change
The different channels are likely to hit different target markets, and so there may be
‘socio-economic selection’. [1]
In particular the clients of insurance intermediaries are likely to be better off … [½]
… and may therefore experience lighter mortality due to better education and standard of
living. [½]
On the other hand, anti-selection may be more of a worry on business from intermediaries. [½]
But the company may counter this with heavier underwriting, resulting in lighter mortality. [½]
The effect of selective withdrawals may be greater through intermediaries. [½]
In particular, the intermediary may compare rates around the market at the times of premium
reviews. [½]
More selective withdrawals would worsen mortality experience. [½]

[Maximum 3]
(ii) How the pricing basis might change
General
Overall, the pricing basis probably needs to be more competitive. [1]
This means that smaller margins must be taken in any pricing assumptions. [½]
(Alternatively, any explicit profit margin might have to be reduced.) [½]
If the company intends to have the same charges for the two channels then the assumptions
would reflect a weighted average of the two sets of expected experience. [½]
Otherwise its assumptions for its intermediary product would reflect the expected experience for
that channel only. [½]
Mortality
For the reasons given in part (i), the mortality experience on the business is likely to be different,
although one cannot be certain about how much different. [½]
Expenses
The expenses involved in writing the business will be quite different. [½]

For example:
 Commission payments to intermediaries may be different from the salary or commission
payments made to the company’s own salesforce. [½]
 Other costs of distributing and maintaining the business may differ (eg marketing to
intermediaries). [½]
 Costs of underwriting may be different (eg if assess risk more carefully, obtain more
detailed information, etc). [½]
 Volumes should be increased by the new channel and so per-policy expenses could
fall … [½]
 … although the costs of developing the new channel would have to be recouped. [½]
The level of withdrawals may vary between the two channels, … [½]
because:
 the pressure of the sale may be different [½]
 the financial sophistication of the customers of intermediaries may be greater, so that
more appropriate sales are made, giving fewer withdrawals in the first few years [1]
 sales may be more often client-initiated, also giving fewer withdrawals [½]
 there may be more withdrawals by the customers of intermediaries at premium review
dates, if the intermediaries actively review rates at those times. [1]
Unit growth rate
It is just possible that customers would tend to choose funds with different levels of risk, which
would affect unit growth assumptions. It is unlikely that in practice a company would try to allow
for this. [½]
Mix of business
The mix of business by size, age, term, etc could change. [½]
Probably the most important change would be an increase in the average policy size, which can
lead to lower charging rates for the same profitability. [1]
[Maximum 9]
Solution X2.2
(i) Selling inappropriate products
This might be a result of coercion. Intermediaries are often encouraged to sell policies in order to
enhance their own remuneration, for example where high initial commission is paid for every new
policy sold by them. [1]

The intermediary might be tempted to sell the product, even though it may not meet the person’s
needs properly, because:
 it may be the best product available from that intermediary for the purpose, and the
intermediary prefers to make a sale rather than pass the customer on to a competitor [½]
 the intermediary may have promoted the product, irrespective of more suitable ones
being available, on the basis of higher commission rewards [½]
 more suitable products were not available anywhere in the market at the time of sale,
and the policyholder preferred some cover to no cover (or was persuaded to think so). [½]
The reason could be due to lack of education of the policyholder with regard to their needs and
the possible products available to meet them. [½]
If this was coupled with a lack of advice (eg due to being sold through a direct sales channel) … [½]
… or with receiving inappropriate or wrong advice (eg due to poor quality of the intermediary or
through deliberate mis-selling) … [½]
… then an inappropriate product may have been bought. [½]
Another reason is that, despite appropriate advice being given to the contrary, and the person has
been made aware of their needs and the risks faced from the options available, they still choose
to take the inappropriate product. [½]
For example, this might occur as a result of personal greed or, more likely, an incomplete grasp of
the various risks involved. [½]
A more complex situation arises if, in the event, a product is sold that appears to meet the
perceived needs of the customer at the point of sale, but then either:
 the perception of these needs by all parties turned out to be wrong, or
 circumstances changed significantly after the sale, so making (with hindsight) the sale
unsuitable. [1]
[Maximum 5]
(ii) Risks to the insurance company
A significant risk is low persistency rates, which generally leads to losses or to a reduction in
profitability. [½]
The other main risk is to the company’s reputation, which could significantly affect the company’s
ability to sell business in the future. [½]
[Total 1]

Solution X2.3
We shall be studying stochastic modelling in more detail later on in the course, and have confined
ourselves here to what could reasonably be expected as an answer at this stage.
(i) Approaches to projecting investment returns
Essentially there are two approaches: deterministic and stochastic. [1]
A deterministic investment assumption is just a point estimate of future investment returns, … [½]
… (although this could be made to vary (deterministically) in different years of the projection). [½]
The point estimate would often represent the average return over all asset classes … [½]
… or could be done for individual asset classes separately, with averaging over the explicit mix of
assets assumed in each future year. [½]
A central assumption could be supplemented by sensitivity tests or scenario modelling. [½]
A stochastic approach involves treating investment returns as random variables. [½]
One must therefore specify a probability distribution for the investment return. [½]
This would probably be done for each asset class separately, and aggregated over the explicit
assumed mix of assets, as above. [½]
The model would be used by running a large number of simulations. [½]
The average of all the projected outcomes would be an estimate of the expected value of the
outcome. [½]
The projections, whether deterministic or stochastic, would allow for expected income and gains
in each year, perhaps separately. [½]
The model would have to deal with:

 the investment return from the existing assets, [½]
 the reinvestment of proceeds from the existing assets [½]
 the investment of future premiums on existing and new business. [½]
[Maximum 7]
(ii) Different sources of investment risk
The three types of risk are model risk, parameter risk and random fluctuations risk. [1½]
Model risk
This is the risk that the model itself is wrong in some fundamental way. [½]
For example in a stochastic model the probability distribution may have the wrong form. [½]

In a deterministic model the absence of any explicit modelling of variability may be viewed as a
model risk. [½]
Another example of model risk would be that some key determinant of investment returns is
absent from the model (eg no link to inflation). [½]
Markers should give credit (up to 1½ marks) for any suitable illustrations of model risk. The same
goes for the other types of risk below.
Parameter risk
Parameter risk is the risk that the parameters of the underlying model are incorrect. [½]
For example the mean, variance or higher moments chosen for probability distributions in a
stochastic model may be wrong, even if the form of the model is appropriate. [½]
In the case of deterministic modelling we are essentially dealing just with a mean value
assumption, and this may be incorrect. [½]
If spot rate assumptions are used, varying over time, we may incorrectly predict any trends in
these values. [½]
Random fluctuations risk
This risk arises, in the case of investment returns, from the inherent volatility and unpredictability
in the income or gains from most assets. [½]
This is most evident in equity shares and property, but is also true of the capital values of
government or other fixed-interest stock, particularly long-dated stock. [½]
Actual experience is just one of a whole range of possible outcomes, … [½]
… and is unlikely to be the same as a deterministic assumption, nor the same as the mean or
median simulation from a stochastic projection. [½]
[Maximum 6]
Solution X2.4
This topic is covered in Chapters 10 and 11.
(i) Risk from the maturity guarantee
The risk is that the unit fund performance net of charges is so poor, that the fund value at
maturity is less than the total premiums paid. [1]
The insurance company would then have to make up any shortfall. [½]
On the face of it, the guarantee is not particularly onerous, … [½]
… and probably requires only a low positive return on the underlying investments. [½]

However, the risk may be increased if policyholders are able to choose particularly risky and
volatile fund links for their policies. [½]
It will also depend on the size of the charges for expenses and any mortality cover. [½]
Whilst the company may mitigate the cost by charging for the expected value of the guarantee,
… [½]
… there is still a risk that the value is underestimated … [½]
… because the company may have an inadequate model of the investment returns, or may have
mis-estimated the parameters. [½]
The extent of the risk will also depend on the volumes of business sold, … [½]
… and how much stays in force until maturity. [½]
Because policyholders choose their maturity dates in advance, there is no risk of financial
selection occurring with respect to the maturity guarantee. [½]
However, there is a risk of financial selection on withdrawal, as healthy people are less likely to
withdraw whenever it appears that the maturity guarantee is likely to bite (ie to have value). [½]
[Maximum 5]
(ii) Reply to marketing manager about adding guaranteed surrender values
Marketability
Yes, it probably would make the contract more marketable, … [½]
…and higher sales would be good in themselves, provided that the contract is profitable and that
there is capital to write the business. [½]
However, any charges for the guarantee might make the contract less marketable. [½]
Risks
It might make the contract most marketable to those who think it likely that they will withdraw. [½]
This would be a significant risk. [½]
Early withdrawals in particular would be a risk, … [½]
… because total premiums paid would exceed the asset share and any likely ‘normal’ surrender
value at the start of the contract. [1]
As withdrawal rates tend to be highest in the first few policy years, the cost could be
significant. [½]
Another key additional risk for surrender guarantees compared with maturity guarantees is that of
active selection by policyholders. [½]
Whereas the maturity date is chosen at outset, policyholders choose when to surrender. [½]

Policyholders would be most likely to take advantage of the guaranteed surrender values following
a market crash, when the cost to the company would be greatest. [½]
However, the fact that the guarantee is ‘only’ a return of premiums paid would limit the risk. [½]
Reserves and capital requirements
Reserves and/or solvency capital will also need to be increased, especially in the first few years. [½]
(For example, reserves may need to be at least the current surrender value.) [½]
This will increase the cost of capital (ie require charges to be increased). [½]
It may also cause problems of new business strain if large volumes of business are issued. [½]
Conclusion
Conclusion is that the surrender guarantee would carry significantly more risk than the maturity
guarantee. [½]
It should not be contemplated at early durations but might be considered at later durations. [½]
Other sensible conclusions supported by arguments should receive equal credit.

[Maximum 7]
Solution X2.5
Regulatory restrictions are covered in Section 4 of Chapter 9.
(i) How restrictions might protect policyholders
Substantial investment in government bonds helps to ensure a fairly low overall default risk for
the investment portfolio, assuming a stable political and economic environment. [½]
Unquoted companies would generally have a higher risk of default and may have more volatile
prices. [½]
Because other investment must be in quoted stock these risks are lower. [½]
The risk that money will not be available when needed is also reduced, … [½]
… because quoted stocks will be more marketable. [½]
By banning overseas investment the risk of currency losses is eliminated. [½]
The restriction on investment in a single company prevents over-exposure to the performance of

that company. [½]
It should ensure a reasonably well-diversified portfolio, although there is nothing to prevent

over-concentration by sector. [½]
[Maximum 3]

(ii) How restrictions might reduce policy benefits
Overall, the restrictions impose a fairly cautious investment strategy on life insurers. This means
lower expected investment returns than with an unrestricted strategy. [1]
(We cannot be sure that it will reduce actual returns to policyholders.) [½]
Particular ways in which the restrictions do this are:

 A large part of the investments must be in the lowest risk, lowest expected return
investment, ie government bonds (safe provided there is a stable environment). [½]
 Investment in other domestic asset classes, including property and unquoted securities, is
not allowed. At some times, insurers may feel that these offer higher expected returns.
[½]
 Overseas markets may offer higher expected returns at certain times, especially where
developing markets are concerned. [½]
 There is no opportunity to benefit from favourable exchange rate movements. [½]
 1% is quite a low level for investment in a particular stock. It may prevent life insurers
from holding as much of a particular stock as they would like, … [½]
 … and could force sales from holdings of stocks that are performing particularly well. [½]
 The costs of dealing in a large number of (relatively) small holdings may also reduce
policyholder benefits. [½]
 The constraints on investment returns will feed through to reduce the expected benefits
payable to policyholders (eg on with-profits policies). [½]
[Maximum 5]
Solution X2.6
Before starting, it needs to be recognised that a unit-linked pension contract is one that is
designed to build up a fund for a pension. That is, it is essentially a unit-linked endowment, and is
unlikely to have any significant protection benefits. There is therefore little mortality risk (other
than a possible longevity risk if there are guaranteed annuity conversion rates).
Investment risk
Because all of the company’s income must arise from the fund management charge, which is
proportionate to fund value, then slower than expected growth in fund value will reduce profits,
all else being equal. [1]
This could be mitigated if the fund charge were to be reviewable, though no information is given
to suggest this. [½]
Even then, policyholders’ reasonable expectations may limit the amount of any increase
possible. [½]

Persistency risk
This arises as a consequence of the mismatching of income and outgo of this product design. One
aspect of the mismatching is timing of cashflows, in particular regarding the initial expenses (at
the start of the policy) and the charges available to meet them (the regular annual charge). The
value of the unit fund will exceed the asset share under the policy for a considerable number of
years after the policy has been issued. Assuming the value of units is available without penalty on
withdrawal, then withdrawals could cause significant losses, especially at early durations. [2]
The company would also have had difficulty in assessing the withdrawal assumption to use in its
pricing, because its past experience (of a mortality protection contract) will probably bear little
resemblance to the experience of a future savings-contract. [1]
This is because the two classes of lives involved have very different needs and behaviour,
reflecting the different contracts they choose to take out. There is therefore a significant risk
from having used incorrect withdrawal assumptions in the pricing basis. [½]
Options and guarantees
If there is any guaranteed rate of annuity conversion at retirement, then the company is at risk of
the guaranteed terms being too generous. [½]
Expense risk
There is a risk that expenses will be higher than assumed in the pricing basis, leading to losses. [½]
This risk is particularly high because the company’s past expense experience will not give much
indication of the level of future expenses: term assurances have very few admin requirements
compared with the complex systems required for unit-linked business (for example, keeping
records of unit allocations for each policy, calculating daily unit prices, etc). There is therefore a
significant risk of underestimating the level of expenses in the pricing basis. [1]
This risk is increased by inflation of the company’s ongoing (regular) expenses, … [½]
… although this should be mitigated to an extent if the fund is invested in real assets, as would be
likely for a unit-linked savings contract. [½]
Fund growth rates (and hence fund charges) should be partly correlated with inflation and this
should reduce the risk. (Other aspects of expense risk are covered below.) [½]
New business mix
This is another consequence of the mismatch between income and outgo. Some of the actual
expenses are likely to be per policy (eg most administration costs would be similar for each policy
regardless of the size of that policy), whereas the income to cover those expenses is fund related.
The smaller the policy size the less the income will be, but per-policy expenses will remain the
same. [1]
Hence the company is at risk from issuing smaller policies, on average, than it expected when it
priced the contracts. (This applies both to direct and fixed per-policy expenses.) [½]

Furthermore the fund charging rates do not vary with the duration of the policy (ie with entry age,
as the younger the new policyholder, the longer will be the average duration to retirement). [½]
To the extent that some of the initial expenses are fixed, then the shorter-term policies will be
much less likely to recoup their initial expenses than longer-term policies. [½]
The company is therefore at considerable risk of issuing a greater proportion of short-term

policies than it allowed for in its pricing. [½]
New business volume
Per-policy expenses will increase, the fewer policies that are issued. This is because some of the
expenses of the company are fixed, relative to business volume: this includes the company’s
general overhead expenses, but also the development and marketing costs associated with this
product launch, which will be considerable. These include the cost of developing all the admin
systems described above. [1]
The per-policy income will, however, be invariant in relation to business volume, so there could
be a significant risk from not selling enough policies. [½]
On the other hand, if the product is a marketing success, there could be substantial new business
strain. [½]
The product is potentially very capital intensive, eg the reserves may need to be at least as high as
the surrender value (the unit fund value) whereas the asset shares will be much lower at the start
of the policy (premium less initial expenses). [1]
This (as well as the persistency risk) could be substantially reduced if a surrender penalty is
applied. [½]
Marketing and development risk
There is a very substantial risk that the product launch is unsuccessful. [½]
The company has only sold temporary assurances to date, so it has no experience of managing a
unit-linked product, and no track record of relevant investment performance (most of its current
investments would be fixed interest, whereas most of the assets backing the unit-linked products
would be equities and property). It would therefore find it very difficult to convince a sceptical
public that its new product would provide a good vehicle for such an important personal
investment (saving for retirement), and its new business volumes could be very poor. [1]
On the other hand, it does sell through its own salesforce, so competition from other providers is
considerably reduced, and this will help sales. [½]
This does put a great deal of onus on the salesforce (and hence the company training it) to ensure
that policyholders are properly informed about the product and that appropriate expectations
about the policy are generated. [½]
This could lead to subsequent marketing risk if, for example, future performance is disappointing
and many policies withdraw (with the possibly significant losses on withdrawal described
above). [½]

There is also considerable risk that future performance will be disappointing. [½]
This is because the existing investment managers will have little expertise in managing an equity
portfolio. The company could improve its chances by recruiting suitably experienced staff, and/or
by training its existing staff, though this will be expensive and adds to the initial product
development costs. Alternatively use could be made of an external investment house, or the
policy benefits could be linked to existing external funds (such as unit-trusts). [1]
The company also needs to set up a completely new and complex administration system for the
new contracts. There is a considerable risk of this system failing and the financial consequences
(eg of errors in calculating fund values or supervisory reserves) could be devastating to the
company. [½]
Poor administration would also lead to dissatisfied customers and further increase marketing
problems. [½]
[Maximum 14]
Solution X2.7
Mortality risk
There is a new mortality risk. If mortality is higher than assumed in the pricing basis, the company
will make less profit than expected (or make increased losses). [1]
This is mitigated by the fact that charging rates are reviewable at any time. Should mortality turn
out higher than expected, then future charges could be increased. [½]
Policies on the minimum allocation rate are effectively subject to guaranteed charges until the
next premium review date. [½]
If the company needed to increase mortality charges under any policy to greater than the 90%
non-allocation rate, but could not due to the minimum, then the company would suffer increased
mortality losses from those policies until the next premium review date. [1]
Investment risk
The sum at risk is the death benefit less the value of the unit fund. [½]
So the insurer is exposed to investment risk whenever the sum at risk is positive. [½]
Many unit-linked contracts include a mortality charge that is related to the sum at risk, so that the
charge would be higher whenever unit values fell. However, for this contract the mortality charge
is taken as a percentage of the premium, so the insurer is exposed to investment risk in relation to
the death benefit (at least until the next charging rate review date).

Selective withdrawal risk
There is a significant risk from selective withdrawal. This could happen whenever charges were
increased during the policy term, as healthy lives would choose to withdraw, resulting in worse
average claims experience amongst those who remain. [1]
The incidence of this, and its financial effect, are likely to be more pronounced the larger the
death benefit multiple that has been chosen by the policyholder. [½]
Further selective withdrawal is likely on premium reviews, if premiums have to be increased. [½]
Policyholders who feel unable to pay the higher premium may withdraw from the policy, and
again it is likely to be those in best health who do so. [½]
There appears to be no provision to reduce death benefits on the premium review date, which
might have reduced the impact of this effect. [½]
The need to increase premiums at the review date may be reduced to the extent that an increase
in the unit fund over the preceding five-year period may have reduced the sum at risk. [½]
This mitigation will be least, however, for cases with the highest death benefit and/or lowest
allocation rates, which is also where the mortality risk is greatest. [½]
The reduction in allocation rate to pay for the non-unit death benefit is constant for the lifetime
of the policy, whereas the sum at risk may be reducing significantly over time as the unit fund
builds up. [½]
There may therefore be a risk of selective withdrawals when the value of the future reduction in
allocation rate exceeds the value of the life cover on the sum at risk. [½]
Anti-selection risk
There will be a completely new anti-selection risk at entry. Those who consider themselves to be
at highest risk of dying will apply for the largest death benefits. [1]
This should be controllable by the proposed underwriting … [½]
… although the company has no past experience or expertise in underwriting and runs a risk of
making significant errors of judgement … [½]
… although expected assistance from reinsurers should reduce this risk. [½]
Probably more significant is the lack of underwriting for benefits up to 0.5 times starting salary.
This could be a significant sum assured for ‘death-bed’ policyholders, who would be granted cover
without underwriting on standard terms. The anti-selection from this cause could turn out to be
completely crippling. [1]
Data risk
There is a significant data risk. While the company will have some data regarding the mortality
experience of its policyholders to date, these relate to holders of savings contracts, and the new
policies could attract very different types of individual with very different mortality. [1]

So mortality parameters estimated from the company’s past data may be very inappropriate. [½]
Using reinsurers’, industry and population data would also incur considerable uncertainty. Noting
that the basis is hoped to be competitive, there is therefore a big risk that the pricing basis
mortality rate is wrong. [1]
Volume and mix of business risks
There is a risk that the product will not prove to be popular, and the expansion in business volume
will not be achieved. [½]
The company would then lose money from its uncovered development costs. [½]
Further, if the increased volume was necessary in order to keep the costs of overheads at a
competitive level (ie so that the expense charges can be kept competitive), then a failure to
increase new business could reduce profits over all business and/or make them less competitive
by having to increase charges. [½]
If the product is competitive, then new business volumes could increase significantly. The new
death benefit could significantly increase the company’s supervisory solvency requirements for
these policies, and coupled with the increased sales could place heavy capital demands on the
company, reducing its free assets and quite possibly threatening its supervisory solvency
position. [1½]
From the above it will be seen that some policies are more risky to the company than others:
those with high death benefit multiples and/or lowest allocation rates causing greatest risk. The
company’s overall risk will therefore be dependent on the mix of business sold (so that the mix of
business is also, in effect, a risk). [1]
Policyholders may become unhappy with charging-rate and premium reviews. This may
ultimately damage the company’s reputation and adversely affect all sales for the company. [1]
Expense risk
Expenses will now be higher, from underwriting, claims management, increased experience
monitoring, reviewing charges and reviewing premiums. Charges determined for these may
prove to have been underestimated. [1]
Reinsurer default risk
It is likely that mortality risk protection will have been sought from reinsurer(s). If so, however,
there will be a risk from reinsurer default. [½]
Management risk
The company will not be familiar with the process of mortality-experience monitoring and the
associated charge and premium reviews. Failure of management to operate the systems
effectively is a significant risk. [½]
[Maximum 15]





using the course.

Solution X3.1
The description of the calculation of the appropriation and expropriation prices from Chapter 13
gives most of the ideas. Further adjustments are then required to allow for initial charges and
rounding.
The following adjustments need to be made:

 the appropriate market value is the offer price of the assets when pricing on the offer
basis [½]
 the appropriate market value is the bid price of the assets when pricing on the bid basis
[½]
 expenses and stamp duty should be added to the market value when pricing on the offer
basis [½]
 expenses should be deducted from the market value when pricing on the bid basis [½]
 the value of any current assets (such as cash on deposit or investments sold but not yet
settled) should be added to the market value [½]
 the value of any current liabilities (such as investments purchased but not yet settled or
loans to the fund) should be deducted from the market value [½]
 any accrued income (such as interest income from fixed-interest securities and deposits,
net of any outgo such as fund charges) should be added to the market value [½]
 an allowance for accrued tax, if applicable, should be deducted from the market value [½]
 if the insurer wants to deduct an initial charge through the unit pricing mechanism, then a
bid and offer price for the units is required [½]
 the resulting unit prices need to be rounded according to the stated rules. [½]
[Maximum 4]
Solution X3.2
Thinking through the steps of the unit pricing calculation as described in Chapter 13 should help
generate ideas. A consideration of any practicalities should follow.
Basic equity
When calculating unit prices at the merger date, the company should ensure that each unit-linked
policyholder has the same value of units afterwards as they did immediately prior to the merger.
[1]
Since it won’t be possible to keep all the unit prices the same, this means that the number of units
must be adjusted accordingly. [½]

The company will need to decide on the starting price for the new funds, which could either carry
on from the existing price for one of the funds or could be an arbitrary new figure, say 100p. [½]
The number of decimal places to record for both the unit price and number of units will affect
how accurately the policy values can be kept the same as before the merger. [½]
To demonstrate that policyholders haven’t been adversely affected on the day of the merger, the
company is likely to send out a unit statement to each policyholder. This would show their
holdings as at the merger date, in terms of both the old funds and the new funds, so the ‘before’
and ‘after’ positions can be reconciled. [1]
Pricing basis
The pricing basis to use for each fund will need to be decided. For each new fund, the company
should look at the past pattern of unit transactions on the previous funds and derive the overall
position of the new fund, ie whether it is expanding or contracting. If it is expanding, then an
offer basis will be used, otherwise a bid basis will be used. [1]
For some policyholders, the merger will mean a change of basis for at least some of their funds, so
the company needs to decide how to deal with this. [½]
Other aspects of the calculation
It is possible to calculate either the bid or offer price first and then derive the other one from it. If
the two sets of funds used different methods, then the company will need to decide which one to
go with, which might depend on which of the company’s unit pricing systems will be used. [½]
Rounding rules will also need to be decided. [½]
The calculation of unit prices might have happened at different times of day in the two previous
companies, eg one at noon and one at 5pm. If so, the merged company will need to decide what
time to use, which might depend on which systems are to be used in the new setup. [½]
The valuation of certain types of asset, eg property, might have been updated more frequently in
one company than in the other. This needs to be brought into line in the new company and might
involve obtaining current valuations for all the property holdings. [½]
All aspects of the calculation should be consistent with all sales and marketing literature and
other documentation issued by the company, so this should all be reviewed and updated as
necessary. [½]
Practicalities
If the two companies use different unit pricing systems then the merged company will need to
decide which one to carry on using after the merger. This decision should be based on an analysis
of the pros and cons of each system, taking into account ease of use, speed, flexibility,
compatibility with other investment and policyholder systems etc. [1]
If any unit pricing is carried out externally, the company will need to discuss whether it’s better
for this to continue for all the new funds or whether it should now all be carried out internally. [½]
[Maximum 7]

Solution X3.3
These topics are covered in Chapters 14 and 16.
(i) Basic features of a cashflow projection model for pricing term assurances
The model needs to allow for all the cashflows that may arise, … [½]
… which in this case are premiums, death benefits and expenses. [½]
It also needs to allow for the cashflows arising from the supervisory requirement to hold reserves
and solvency capital. [½]
It will be important to model the cost of capital in this case as the initial expenses and solvency
capital requirements are likely to be much larger than the first premium. [½]
The cashflows need to allow for any interactions, particularly where the assets and the liabilities are
being modelled together. [½]
However, the investment return is likely to be of little significance here, as term assurance
contracts have low reserves. [½]
However, the existence of selective withdrawals means that low persistency should be associated
with high mortality rates. [½]
The model should allow for the potential cashflows from any health options included in the
contracts, … [½]
… such as options to increase the sum assured or increase the term of the contract. [½]
It is unlikely that a stochastic model will be used in this case (as term assurances do not include
financial guarantees). [½]
[Maximum 4]
(ii) Using marginal costing
If the business was only going to cover marginal expenses then there seems little to recommend it
beyond maintaining a market presence and being able to offer a wide range of products, unless the
allowance for profit is very significant. (The importance of doing so depends on the sales
distribution method.) [1½]
However, if business volumes were large enough then this might reduce average per-policy
expenses (in, for example, the new business processing department) and allow some contribution to
covering overheads. [½]
But in this case the company must consider whether this is the best use of capital, or whether it
could be better employed in a less price-sensitive market. [1]
Whatever happens, the company must at least cover its overheads across its whole portfolio of
business. [1]

To a certain extent it is fooling itself if it allows a class of business to make no contribution. It would
be better to include some allowance for overheads but to recognise the competitiveness of the
market by accepting a lower degree of profit than on some other classes. [1]
The company should also consider alternatives for boosting sales that do not involve such
competitive pricing, such as concentrating on channels other than brokers (easier said than done),
or differentiating the product with options (which carries its own risks). [1½]
[Maximum 5]
(iii) Premium term for family income benefit policies
Because this policy has a decreasing benefit for a level premium, a point can be reached where
the value of future benefits and expenses is lower than the value of future premiums (ie the
reserve is negative). [1]
The company would then be vulnerable to lapses, having provided more cover than has been paid
for. This would lead to reduced profits. [½]
By an appropriate shortening of the premium paying term the company can ensure that benefits
are paid for before they are provided. [½]
It has the additional merit of making the asset share positive earlier in the term at the start of the
contract, reducing the risk from early withdrawals. [½]
[Maximum 2]
Solution X3.4
Steps in pricing and design
The key steps are as follows:
Step 1 – Assimilate marketing information: [½]

 age, gender, socio-economic group of policyholder
 why the product is bought
 competitors’ charging structures
 average premium, term, sum assured
 how the product is sold
 information shown on quotations
 level of underwriting.

Step 2 – Determine assumptions for: [½]

 economic basis (investment returns, inflation, tax, expenses)
 supervisory reserving basis and solvency capital requirement
 demographic factors (mortality, persistency rates)
 risk discount rate and profit criterion.
Step 3 – Determine appropriate model points: [½]

 look at age, gender, term, premium size, sum assured. [½]
Step 4 – Perform profit-test: [½]

 start with a charging structure
 project monthly cashflows (ie premiums, interest, expenses, claims, tax)
 project supervisory profit, by allowing for supervisory reserves
 allow for any supervisory capital requirement
 compare profit measure (eg net present value) against profit criterion
 repeat for different charging rates until profit criterion met.
Step 5 – Perform model office projections: [½]
Perform projections for whole company, … [½]
… showing projected assets and liabilities (to investigate capital requirements)… [½]
… and embedded values (to show profitability). [½]
This will allow for:

 a moving picture of sales volumes
 economies of scale
 impact of volumes on charging structure
 impact of volumes on mortality and underwriting
 impact of sales on other products.
[½ per point, maximum 1½]
Step 6 – Consider the risks: [½]
Perform sensitivity testing by varying the experience parameters … [½]
… and modify product design if results appear too volatile. [½]

Step 7 – Amend charging structure: [½]
Repeat steps 4 to 7 until the charging structure is found that should maximise profit, without
putting undue risk or capital requirements on the company. [1]
[Maximum 14]
Solution X3.5
This topic is covered in Chapters 15 and 16. It might also be helpful to refer back to Chapter 7 for a
reminder of the contribution method.
(i) Modelling the bonus distributions
Program in the dividend formula so that it will calculate the correct annual dividend per policy
(ie per model point) in each future projection year. [½]
It must be dynamic, the projected dividend varying with the projected actual interest, expense
and mortality experience, for each homogeneous grouping. [1]
The valuation basis for these elements will need to be included so that the model can calculate
the difference between actual and expected. [½]
It will be sensible to include, as a parameter in the model, the proportion of the surplus emerging
each year which is distributed to policyholders. [½]
[Maximum 2]
(ii) Raising the policyholders’ share of profits
Redo all the normal profitability work on these new bonus allocation rates and see what the
results are. [½]
This will include:

 profitability of existing business – is it still positive? Do some products suffer more than
others? [1]
 profit testing new business model points – are they profitable on this basis? [½]
If the new terms are expected to give a significant competitive edge then the company might
project a large increase in new business levels – what is the effect of this on expenses? The
company might now require a smaller contribution from every policy to its fixed overheads. [1]
Value of new business – if the company had written the last 12 months (or previous calendar
year’s) new business on these terms, would it have been profitable? [½]
Effect on projection – re-compute the financial projection (probably five years) on these new
terms. Does the revision in bonus terms cause any problems with capital (ie the solvency
position) or any other problems (eg profits unacceptably reduced)? [1]

Doing the above work will have given the company a good idea of how far it can go quantitatively.
Depending on the precise numbers, it will then need to consider:
 Does it want to extend these more generous terms to existing business, or restrict them
to new business only? [½]
 Does it want to apply the same new bonus rates to all products, or can some of them not
support it? [½]
 The principle of fairness must be considered if the company is going to give higher
dividend rates to some groups and not others. [½]
 If products are not profitable with the 90% bonus allocation, is it very important that
people are told that the company declares on 90%? If so, the company might want to
look at repricing to accommodate that bonus rate. [1]
 It might consider keeping the rate at 80% and introducing a terminal dividend, especially if
it could focus more on total payouts as opposed to just the annual rates. [1]
Also consider:
 What alternatives are there?
 What about 85%?
 What about moving up only for particular categories, in order to encourage the type of
business that the company wants – eg regular premium only, or premiums above a
certain level only, or policies above a certain duration only?
 What about a scale that increases from 80% to 90% as policies age, allowing the company
to quote ‘eventual 90% bonus rate’ to potential policyholders?
[½ mark per sensible suggestion, max 1]
For the various alternatives that are considered, the company will need to repeat the earlier work
of profit testing and financial projections to check that they satisfy its financial criteria and that
they do not present any problems with use of capital. [1]
[Maximum 8]
Solution X3.6
Modelling is covered in Chapters 14 and 15 – these chapters contain enough material to score full
marks in this question. Some of the points in the solution below might be tricky at this stage, but
we will return to the topic of valuing guarantees in Chapter 23.
(i) Using a stochastic model to determine charges for a maturity guarantee
A representative selection of model points would be chosen. [½]
Separate model points would be used for each fund (European and North American), ... [½]
... and possibly for different premium sizes (as any per policy charges would reduce the unit fund
by proportionately more for small policies) and for different ages and gender (as the guarantee
does not apply on early death). [½]

The unit fund would be projected, net of all charges, allowing for deaths and withdrawals. [½]
These charges would include the charges applied to the existing contract (eg fund management
charges, policy fees) as well as the reduced allocation rate to be charged for the new guarantee.
[½]
The model would be run with different allocation rates for each model point to find the optimal
solution in terms of both marketability and profitability. [½]
A suitable probability distribution for the unit growth rate variable would be chosen. [½]
This distribution may be the same for the two funds, but it would have different parameters
(eg volatility) to reflect differences in the behaviour of the backing assets. [½]
If a risk neutral (or market-consistent) calibration were being used, the parameters of this
distribution would be chosen to replicate the market prices of actual financial instruments as
closely as possible. [½]
It is most important that the model could reproduce ten-year put options on both European and
North American equities as this is the financial instrument that the guarantee resembles most
closely. [½]
The expected return would be the risk-free rate, regardless of which fund was being considered.
However, the volatility would differ between the funds. [½]
If a real world calibration were being used, the parameters of the unit growth rate distribution
would be chosen to accord with realistic long-term expectations of future investment returns on
portfolios of either European or North American equities as appropriate. [½]
A deterministic assumption would be used to model mortality. [½]
Withdrawals might be modelled stochastically such that the level of withdrawals is related to the
fund performance, eg withdrawals could be low late in the policy term if the maturity guarantee is
likely to bite. [1]
However, a stochastic model of withdrawal rates might be considered too complex or subjective,
and so a deterministic assumption could be used supported by scenario testing. [½]
The unit fund value at maturity would be simulated a large number of times using the stochastic
unit growth rates. [½]
The excess of the guaranteed value over the unit fund at maturity, where positive, is the
simulated guarantee cost for each simulation (zero cost otherwise). [½]
The guarantee is the single premium accumulated at 1% pa, ie P  1.01 

10
 1.10462P . [½]
So the stochastic model could be used to generate a distribution of the guarantee. [½]
If a risk neutral calibration were being used, the charges would be set as the average of the
discounted value of the guarantee cost. [½]

This discounting would be performed at the risk-free interest rate. [½]
If a real world calibration were being used, the charges would be accumulated in each simulation
in a similar way to the projection of the unit fund (although the charges could be invested in
different assets requiring a different set of parameters). [½]
The charges would then be set so that their accumulated value exceeded the guarantee cost in a
large number of simulations, say 90%. [½]
Sensitivity tests could then be performed on these charges, eg the sensitivity of the charges to the
choice of probability distribution or its parameterisation could be investigated. [½]
Charges for all product variations would be extrapolated from the model points. [½]
[Maximum 8]
(ii) Modifications to contract design and their impact on the modelling required
Markers, please give credit for other suggestions to modify the product’s design [½] together with
an appropriate comment on the modifications to the modelling required [½].
The policyholder might be offered a wider range of unit funds to choose from, eg property,
corporate bonds or a managed fund; ... [½]
... the charge would need to be calculated separately for each unit fund using different
assumptions for the unit growth rate. [½]
Switching might be allowed between funds after the policy is taken out; ... [½]
... the model would need to allow for the possibility of switching to a different, possibly more
risky, unit fund at a later date (which would increase the charges required from the lower risk
funds). [½]
The guaranteed growth rate could be applied to surrenders too; ... [½]
... the model would need to allow for a proportion of policyholders surrendering each year where
this proportion would be dynamically linked to the performance of the fund. [½]
The guaranteed growth rate could be applied on death too; ... [½]
... the model would need to allow for a proportion of policyholders dying each year, but this
proportion would be modelled deterministically (as mortality rates are likely to be independent of
the performance of the fund). [½]
The policyholder might be allowed to choose a term other than 10 years; ... [½]
... the charge would need to be calculated separately for each policy term that is allowed, eg 5,
10, 15 years. [½]
The guarantee charge might be deducted over time as an increased fund management charge
(rather than as a single upfront reduction to the allocation rate); ... [½]

... the model would need to be modified for this change to the charges out of the unit fund (and
into the non-unit fund) allowing for the fact that the insurer will no longer receive the full charge
on death or surrender before the end of the term. [½]
[Maximum 4]
Solution X3.7
(i)(a) Marketability
This policy should be attractive to people who are looking for long-term protection on death, who
do not necessarily want maximum cover straight away (for example, couples planning on starting
a family). [1]
The rising optional benefit would seem attractive partly as a means of keeping the benefit level in
line with inflation, … [½]
… and partly to increase the real benefit level as needs increase (eg due to increases in family
size). [½]
The removal of underwriting for these increases is also attractive. [½]
This design will make the initial cost particularly attractive compared with level premium
contracts offering equivalent cover. [½]
The policy suffers from some lack of flexibility as benefit increases are limited to ten years and
individual increases are limited to 10% per annum, which could reduce the attractiveness to some
potential applicants. [1]
The unit-linked design allows the policyholder to benefit from future good investment
returns. [½]
The facility to drop the minimum sum assured (for example, when death cover is no longer
necessary), and the ability to surrender for the unit fund value, mean that the policy can continue
as a savings contract (and into which any excess premiums over charges will already have
contributed). [1]
The other charging rates being guaranteed is also attractive. [½]
The low initial allocation rates may deter some applicants, especially if early surrender values are
quoted to them. [½]
The presence of the guarantees and options makes it likely that the charges may be relatively
high, which might also lead to a higher starting premium. [1]
This might offset some of the ‘low cost’ advantages set out above, especially if the market is
competitive. [½]
[Maximum 5]

(i)(b) The risks borne by the company
There is a risk that the option to increase the benefit is used more often than expected. This is a
risk because the value of the 10% increase in sum insured is greater than the value of the 10%
increase in premiums (ie the unit fund might not be large enough to support the higher mortality
charges required to meet the higher benefits). [1]
Anti-selection risk also exists on the optional benefit increases. [½]
It is probable that people in worse health are more likely to take their option to increase benefits.
[½]
The company cannot compensate by increasing mortality charges during the first ten years, so
significant risk arises. [½]
After ten years, the company may have difficulty in increasing charging rates as policyholders may
surrender their contracts in response. [½]
Hence the company may find it very difficult to recoup any losses incurred in the first ten years
from profits earned thereafter. [½]
Both surrender and the optional removal of the minimum death benefit are likely to be selective,
in either case only the healthy choosing the option. [1]
There is a risk that the fund management charge is inadequate to cover ongoing expenses. [½]
Inflation will contribute to this risk, though the extent of the risk will depend upon the assets
underlying the unit fund (no information is given about this). [½]
The risk is that the expenses will inflate at a faster rate than the unit-fund value (and hence the
fund management charge) will grow. [½]
There is a substantial risk of loss on early surrender. [½]
Assuming that the initial costs are recouped by the 40-20-10 allocation charges in the first three
years, any surrender in years 1 or 2 will cause a loss broadly equal to the allocation charges not
received by that time. [½]
The investment risk is apparently borne by the policyholder, but this is not entirely true. A
problem could arise over the first ten years if investment returns are much lower than assumed
when the premium was set. [½]
Lower fund growth leads to higher sums at risk, producing higher mortality charges and even
lower net fund growth after charges. It is possible that the unit fund could go negative as a result.
[½]
As there is no premium review facility for ten years, the policy could continue up to the end of ten
years and then be lapsed, having incurred significant losses in the process. [½]
As the mortality charge is paid for annually, there could be a significant adverse discrepancy
between the sum at risk used for calculating the charge, and that of the actual death strains that
arise, if unit values fall between the charging date and the date of death. [1]

It is possible that inadequate volumes of business will be sold, leading to the policy’s share of the
fixed expenses exceeding the loading for those expenses. [½]
If the average policy size is lower than expected, then the fund management charge may not be
sufficient to cover the variable per-policy expenses. [½]
This risk is particularly onerous given that all of the expense charge is linked to fund size, and that
this charging rate is essentially guaranteed for the whole of life. [1]
[Maximum 8]
(i)(c) Financing requirements
The company would need to hold a unit reserve equal to the face value of the unit fund, plus any
necessary non-unit reserve, at any time. [1]
It is not possible to hold reserves of less than the face value of units because there are no
surrender penalties. [½]
The unit reserve in the first two years is very likely to exceed the asset shares at those times,
assuming that the allocation charges in years 2 and 3 are needed at least partly to cover the initial
expenses. [½]
Hence an initial strain will arise in the first two years as a result of the unit reserve alone. [½]
It is also very likely that a prudent projection of future cashflows, along with the effect of the
guaranteed charges and the policyholder options, will lead to a need for positive non-unit
reserves to be held. [1]
Alternatively, the risks associated with guaranteed charges and policyholder options could result
in an increased solvency capital requirement rather than added prudence in the non-unit
reserves. [½]
This could lead to further increases in initial strain. [½]

[Maximum 3]
(ii) Improving the product design
Could reduce anti-selection risk by allowing increases to benefits every year or not at all; once
policyholder misses an increase then no further increases would be allowable. [½]
This has to be considered in the light of the reduced flexibility of the contract, and of the reduced
marketability that might result. [½]
All charges and premium levels could be made to be reviewable, and more often, for example
every five years. [1]
Whilst the loss of guarantees implies reduction in marketability, … [½]
… it could enable benefit increases to be permissible over a longer period without undue
increases in risk, for example for 15 or maybe even 20 years, thereby making the contract
somewhat more attractive. [½]

The ability to review premiums and charges also allows potentially significantly lower charging
rates (especially for the mortality charge) … [½]
… which could enhance the product’s marketability considerably, further helping to offset the
negative effect on marketability of reducing the guarantees. [½]
A (monetary related) surrender penalty could be introduced in years 1 and 2, equal to (at least)
the amount of unrecouped allocation charges. [½]
A monetary per-policy expense charge could be introduced, to replace some of the fund
management charge. This would reduce some of the risk due to smaller than expected average
policy size. [1]
This could reduce marketability to an extent, as it might give the policy the appearance of having
more charges in total. [½]
Reducing the guarantees will reduce the solvency requirement. [½]
Furthermore, if no positive non-unit reserve is needed in the first two policy years, then there
may be scope to pre-fund some or all of the allocation penalties by holding negative non-unit
reserves in those two years (providing sufficient surrender penalties are also in place). [1]
This could reduce or even eliminate the new business strain from the contract, depending on the
level of initial expenses actually incurred. [½]
[Maximum 6]





using the course.

Solution X4.1
Without-profits business
Examining all of the elements of the basis in turn:
Interest rate
The fixed 4% basis makes no reference to the current yield … [½]
… or to the expected reinvestment yields on the actual asset mix. [½]
The basis might therefore be anything from excessively prudent to severely imprudent, depending
upon what the current and future yields are. [½]
The fixed basis does not reflect the asset valuation method. [½]
Mortality
The population mortality stipulated will probably overstate mortality. [½]
This will lead to higher reserves than should be necessary (except for annuity business). [½]
For annuity business, reserves will be lower than they should be. So there would be significant
under-reserving for companies specialising in such business. [½]
Expenses
The industry expenses may be out of date, eg if surveys are performed only every five years or if
there are time delays in publishing the results. [½]
The expenses may then be imprudent if an adequate allowance for inflation has not been
added. [½]
The expenses of any particular insurer will differ from the industry average, eg due to differences
in economies of scale or the extent of automation. [1]
So the method may be prudent for some companies and imprudent for others. [½]
Assets
Assets are valued at book value. This is generally prudent compared with market value, … [½]
… but there is always the risk that market value might be below book value. [½]
Especially in times of economic stress. [½]
So there needs to be a test of book value against market value. [½]

Normally the book value will be unrepresentatively low for equities and property, so the real
solvency position is not going to be shown – a very healthy company might appear weak. [½]
However, the impact of this will be small as the without-profits business sold in this country is
most likely to be backed by investment in bonds and cash (although some of the free assets may
be invested in equities). [½]
No reference is made to the compatibility of assets and liabilities, in particular matching by nature
and term. [1]
The current basis will not signal when any mismatch reserve may be required. [½]
Surrender values
Policy reserves should not be less than any guaranteed surrender value. This is appropriately
prudent. [½]
Some negative reserves are allowed. On its own this may be imprudent … [½]
… as it implies treating a policy as an asset on the grounds of its future cashflows, but the policy
might lapse tomorrow. [½]
However, in practice the negative reserve allowance is never likely to be invoked, because the
minimum surrender value under any contract cannot be less than zero, and the reserve cannot be
less than the surrender value. [1]
The $50 million limit is also fairly meaningless … [½]
… because for a small company it could represent a significant portion of the reserves, for a large
company hardly anything. Cannot say without knowing how much a $ really is. [½]
Solvency capital
Specifying a percentage of written premiums is a strange approach for life business, because it is
not linked to the amount of risk. [½]
(A method that is proportionate to reserves might be more suitable.)
For example, it could be too big for a rapidly expanding company (especially if writing a lot of
regular premium business). [½]
It could be too small for a company closed to new business, or one that wrote a lot of single
premium business in the past (which is still in force). [1]
Other valid examples would earn credit.
A risk-based solvency capital requirement approach, such as the Value at Risk measure, would
ensure that the level of capital held was a more accurate reflection of the risks taken on. [1]
Unit-linked
The market is about to start seeing unit-linked business so such business needs to be catered
for. [½]

The gross premium formula method is not appropriate for unit-linked business, eg due to the
complex manner in which cashflows change from year to year. [1]
Some method based on unit values and non-unit liabilities is needed. [½]
Setting up non-unit reserves will ensure that policies are not going to need extra financing by the
company in the future. [½]
The gross premium cashflow approach will be needed to determine such non-unit reserves. [½]
And the supervisory authority will need to decide on the extent to which negative non-unit
reserves will be allowable. [½]
The comments about surrender values and solvency capital mentioned earlier would also be valid
in the context of unit-linked policies: appropriate credit should be given for this (but only give
credit for one set of points if they have been included under both headings).
[Maximum 15]
Solution X4.2
Project each future year’s cashflow from the portfolio of existing policies. Each year’s cashflow
should consist of expected claims less expected future premiums (if we ignore expenses). [1]
We would allow for the impact of future mortality and lapses, so as to produce expected future
cashflows for each year. [½]
In the absence of direct market valuations of mortality and lapse risks, the company would use
best estimate assumptions of the future experience for these parameters, plus a risk margin to
represent the price the market might expect to pay in order to compensate for the uncertainty. [1]
Each year’s expected cashflow would then be discounted at the appropriate risk-free rate of
return. [½]
This could be based on Government bond yields or on swap rates … [½]
… of suitable terms to match the liabilities. [½]
It may be appropriate to make a deduction to the bond yields / swap rates to allow for credit risk.
[½]
However, it may be possible (depending on the purpose of the valuation and the regulations that
apply to it) to take credit for the illiquidity premium in corporate bonds and thereby use a higher
discount rate than the risk-free rate. [½]
The value of all the cashflows discounted at their appropriate risk-free rates plus the risk margin
would be the market-consistent valuation of the liabilities. [½]
The risk margin could be included by adding a margin to the mortality and lapse assumptions. [½]

Alternatively, an overall reserving margin in respect of these risks could be determined using the
‘cost of capital’ approach as follows: [½]
 firstly project the required capital at each future time period (ie the amount required in
excess of the projected liabilities) [½]
 multiply the projected capital amounts by the cost of capital [½]
 discount using market-consistent discount rates to give the overall risk margin. [½]
[Maximum 5]
Solution X4.3
Supervisory reserves are covered in Chapter 20.
Some healthy companies might be shown to be insolvent. [½]
Strong companies might seem weak. [½]
The extra prudence might only serve to protect against extremely improbable catastrophes … [½]
… eg if a ‘normal’ strong reserving basis contains margins to ensure that reserves are adequate
with a probability of 99% then a strengthening of the basis to ensure adequacy with a probability
of 99.9% is not going to have any useful consequence – in either case the company is solvent
normally, insolvent under financial / demographic catastrophes. [½]
A reserving basis can never guarantee solvency, because this depends on eventual future
experience. A line needs to be drawn somewhere. [½]
Financial results would be misleading (profits artificially delayed). [½]
Tax income to the government from taxed profits would be correspondingly delayed. [½]
The extra reserving would impinge on product profitability, by increasing the cost of capital. [1]
So for with-profits business companies might compensate by delaying the distribution of profits
for as long as possible. [½]
For without-profits business companies might compensate with higher premiums. [½]
Stronger reserves imply lower free assets, reducing companies’ investment freedom. [½]
So companies invest in less risky assets, and so achieve lower investment returns. Again this is to
the detriment of the policyholders’ eventual benefits. [½]
The higher capital requirement may discourage new companies from starting, … [½]
… and might be unfair to the insurance sector compared with other sectors (eg banks). [½]
[Maximum 6]

Solution X4.4
Need an assumption for both base mortality and mortality trends. [½]
Many countries have exhibited a trend of decreasing mortality rates over time. [½]
Look at own experience data, which should be extensive. [½]
Compare observed with expected (using old premium basis). [½]
Look at a recent period (long enough to be credible). [½]
Analyse by major risk group, eg age, gender, smoker / non-smoker. [1]
Look at longer term trends in experience (this may be less reliable). [½]
Look at industry data. These data will be highly credible, but not as relevant to the company. [1]
Particularly useful for identifying longer term trends … [½]
… but watch out for heterogeneity in the data affecting the progression. [½]
Reinsurers may be able to provide further advice … [½]
… particularly with regard to future underwriting levels (and their effects on the standard
mortality experience). [½]
Population mortality can be particularly useful for projecting trends … [½]
… because large volumes of data are required to fit complex models such as Lee-Carter or
P-spline … [½]
… although population data is less relevant as it covers lives that don’t have insurance. [½]
Project the future mortality of each sub-group using the historical experience as a starting
point, … [½]
… allowing for expected future trends … [½]
… and considering the impact of probable future changes in:

 underwriting levels [½]
 sales distribution channel – still using brokers? [½]
 expected improvements in medical treatment and in healthcare [½]
 the effect of future selective withdrawals. [½]
Consider whether the data exhibits a cohort effect that would imply that the mortality
assumption should depend on date of birth as well as age. [1]

As well as extrapolating trends from the data, consider the expectations of experts and the results
of explanatory (process-based) models. [1]
However, be careful not to allow too generously for future mortality improvements, so as to
reduce the risk of loss due to there being less improvement than anticipated. [½]
If final result not competitive enough may need to reconsider underwriting strategy and/or sales
distribution and marketing strategies. [½]
[Maximum 11]
Solution X4.5
Supervisory reserves and capital requirements are covered in Chapter 20.
What is the impact of the changes on the solvency requirements?
Effect of interest rates down 2% pa
The effect of this depends much on what sort of business a company is writing. [½]
For unit-linked products, the unit reserve will be unaffected, while non-unit reserves will increase
slightly. [½]
Overall: probably a very small increase in reserves. [½]
For conventional business, reserves will increase significantly. [1]
The extent of the increase will depend on:

 Single or regular premium? – with regular premium business, the value of future
premiums will increase as the value of benefits increases (although not by the same
amount), diminishing the change; … [½]
 … the impact will therefore be greater for single premium business. [½]
 Term of the policies? – the longer the outstanding term of a policy, the greater the effect
of the decreased interest rate. [½]
A typical bad case could be a ten-year conventional with- or without-profits single premium
endowment policy, where reserves could increase by around 20%, from considering
10
 1.065 
  1
 1.045 
[1 mark available for some sensible attempt to quantify impact]
As the change in the valuation interest rate is to strengthen reserves, there is no reason to believe
that investment conditions have changed, and so there is no reason to believe that asset values
will have risen as a result of falling interest rates. [1]

Administration expenses up 25%
The impact of this depends on:

 The average policy size – the impact of the change will be relatively low for portfolios of
large policies (eg a company selling via specialised brokers to high net worth
individuals), … [½]
 … and high for portfolios of small policies (eg a company which does a lot of direct selling
to the masses). [½]
 The term of policies – the impact will be greater the longer is the term of the policy
considered. [½]
 The amount of administration expenses – relative to the other changes mentioned, the
impact will be greater for those companies that have high per-policy administration
expenses, ie for those companies which have a cost base which is large in relation to
portfolio size (eg new companies, badly managed companies). [½]
Mortality
The effect of the increase in assured lives’ mortality will depend on the mix of protection and
savings business. [½]
For term assurance business, solvency requirements will increase by almost 5%, … [½]
… whilst the increase for endowment assurance business will be in the region of, say, ½%,
depending on the age profile of the policyholders. [½]
The extent of the increase in solvency requirements for annuity business as a result of the
mortality improvements change will depend on the life expectancy of the annuitant as the effect
of mortality improvements is cumulative, eg mortality rates in 10 years’ time will have improved
by around 10%. [½]
Total impact
So depending on the mix of business, and the other factors mentioned above, solvency
requirements would increase for most companies by anything from 5–10% up to 30–35%. [½]
What is the impact of such a change on the financial management of life companies?
We shall consider here ‘average’ companies suffering, say, a 15–20% increase in solvency
requirements. However, the same comments will hold, but to a lesser or greater extent, in respect
of the ‘extreme’ companies.
Some companies will become insolvent, unless they manage to find large amounts of capital,
eg from shareholders. [1]
Or they may now be so weak that they close to new business. [½]
Other companies will see their free assets, and hence financial strength, reduced drastically. [½]

Such companies will have to reconsider their investment strategy, choosing a much lower level of
risk. [½]
This would be expected to lead to lower investment returns. [½]
Similarly, a lower risk tolerance will increase the need for risk premium reinsurance, … [½]
… and lower financial strength will increase the need for financial reinsurance. [½]
More reinsurance decreases expected future profits. [½]
The changes will greatly distort the emergence of surplus on with-profits business – it will appear
much later. [½]
And the distribution of what surplus does emerge on with-profits business is likely to be deferred
as much as possible, where the surplus distribution system permits this, … [½]
… in order to provide more working capital. [½]
Companies may stop writing the more capital intensive product lines or redesign in a more capital
efficient manner (eg replace conventional products by unit-linked). [1]
In general, companies will be more risk averse and so will be less likely to countenance any
project with big unknowns. [½]
Greater aversion to risk may translate into a higher risk discount rate in pricing and profitability
measurement work. [½]
Profits to shareholders will be hit badly in the year of the change. [½]
Companies may need to increase the proportion of earnings distributed as dividends, if

affordable. [½]
However, strengthening the solvency requirements has no direct impact on total future
cashflows, it merely changes the time at which future profits appear. [1]
[Maximum 15]

Solution X4.6
(i) Calculation of embedded value for unit-linked business
The embedded value can be calculated as the sum of:

 the shareholder-owned share of net assets, where net assets are defined as the excess of
assets held over those required to meet liabilities [½]
 the present value of future shareholder profits arising on existing business. [½]
The future shareholder profits can be calculated by projecting the following elements of the non-
unit fund:
 future charges (including surrender penalties) [½]
 less expenses [½]
 less benefits in excess of the unit fund [½]
 plus investment income earned on non-unit reserves [½]
 plus the release of non-unit reserves. [½]
[Maximum 3]
(ii) Financial and economic assumptions
These will be:

 unit growth rate
 interest on non-unit and negative non-unit reserves
 inflation of expenses and administration charges
 risk discount rate
 tax rates.
(iii) Principles in setting assumptions
This is a fairly new company so standard details might not apply. However, it is not totally new,
and so there will be some limited information.
The uncertainty that is likely to apply to many of the parameters here would be allowed for
through a suitable choice of risk discount rate, and so large margins in the individual assumptions
would not be appropriate. [½]
Alternatively, these margins could have been allowed for in the individual assumptions or by using
a stochastic model, and a lower risk discount rate used.
Mortality
The company will have very little data and these data will not be credible. [½]

The actuary would use their experience and judgement based on standard tables adjusted in line
with recent experience (from industry data) … [1]
… and also adjusted for the company’s target market and degree of underwriting. [1]
Also take account of any data available from reinsurers. [½]
This lack of data will not be so critical if the company can vary its charges, otherwise it might take
a margin for prudence in the basis. [1]
Need to allow for trends in mortality experience, especially improving longevity for the annuity
business. [½]
Persistency
This would be difficult to estimate since the company has only three years’ data. [½]
These data could be used as a starting point (depending on how much data there are), bearing in
mind that withdrawal rates tend to fall in the first few years after the sale. [½]
It should be borne in mind that the company has no experience of persistency rates at policy
durations of greater than three years. [½]
The actuary would use their experience and judgement. Advice might be sought from reinsurers,
consultants or employees with relevant experience. [1]
Care will be needed in using any information because persistency rates vary considerably over
time and for different target markets and sales channels. [½]
Expenses
It is a fairly new company but will have some expense analysis to refer to. This could be used to
estimate the items of expense (salaries, property etc). [½]
As it is a new company, the systems have been set up recently and so might be very good. [½]
There will be a few unusual problems, namely:

 company might base its expense loadings on a steady state situation, and will not yet be
at that stage [½]
 adjustments might be necessary for the exceptional costs of start-up. [½]
[Maximum 8]

Solution X4.7
(i) From the companies’ perspective:
For regular premium business, it will probably take many years before the 10% cut compensates
for initial expenses. [½]
But it could be reasonable, or even generous, for single premium business. [½]
There will probably be a problem with early withdrawals. It is likely that asset shares will be much
lower than the surrender value due to the impact of initial acquisition and administration costs,
causing losses on withdrawal. [1]
The mortality basis is inappropriate. It will probably overstate the mortality experience of the
insured lives. This will decrease surrender values, and will be to the company’s advantage. [1]
There will be a problem with surrenders close to maturity if the surrender value is a long way
below the imminent maturity value. [½]
This may create problems with disgruntled policyholders. [½]
The surrender value close to maturity might sometimes be significantly greater than the
guaranteed maturity benefit. [½]
And this will encourage surrenders, to the company’s cost. [½]
There will be a problem for companies in the event of any market crash, as asset shares might
then be much lower than the guaranteed surrender value. [½]
And in such a market crash situation there might be a large number of withdrawals, which would
cost companies a lot. [½]
And might cause solvency problems (if already weakened by a fall in asset values). [½]
It may be advantageous for policyholders to surrender their policy and take out a new one. [½]
There may be a trivially small surrender value for term insurance policies – companies will waste
time and resources managing the payment of these. [1]
There may be big inconsistencies with the treatment of paid-up policies, unless equivalently
curious regulations are introduced for these. [½]
The proposed basis is unacceptable for annuity business. There will be significant and costly
selection against life companies. [½]
The concept of surrender values being guaranteed is by itself a problem. Companies might
immediately need to increase reserves significantly. [½]
The extra reserving requirement will reduce solvency and product profitability. [½]

The method has some advantages:

 Policyholders often expect surrender values to be comparable to the premiums they have
paid. [½]
 It is simple to do and to explain. [½]
 The proposal will automatically ensure that surrender values are in line with the
competition. [½]
 It will presumably be a stable basis in the future. [½]
 It is capable of being clearly documented. [½]
 To some extent it removes the problem of surrender values which are perceived to be too
low, since any lowness (whether real or perceived) can be justified by reference to
supervisory / legal obligations. [½]
The question does not specify whether the measure will be retrospective – if so there could be a
big problem with policyholders’ reasonable expectations for existing policies. [½]
For normal surrenders (not early in the policy term, or in market crash conditions) the surrender
profit to the company will probably be greater than before. This should result from assets earning
a real return above price inflation, and from the mortality basis. There should also be an expense
profit at longer durations. [1]
The proposal will tend to hurt young companies more than it hurts more mature companies
(younger companies lose more from early surrenders, and don’t make so much profit from later
surrenders). [½]
[Maximum 12]
(ii) From the policyholders’ perspective:
The basis will favour those who surrender early, or subsequent to a market fall. [1]
It will probably be worse at longer terms. [½]
Companies may increase premiums if they find an overall increase in surrender costs, and to
compensate for any reduction in product profitability caused by any significant extra reserving
requirement. [1]
The basis imposes a constraint on how companies are able to distinguish themselves from each
other, and so from the point of view of interfering with a free market, it ultimately harms
policyholders by reducing their ability to choose. [½]
[Maximum 2]

Solution X4.8
(i) Gross premium policy value
We want the reserve at the start of the tenth year, ie nine years after the start of the policy. So
the expression for the gross premium prospective policy value is:
(12)
9V  50,000 A  0.05P 
a  P 
a @6% [2]
39:16| 39:16| 39:16|
Evaluating the functions:
l
A  A39  v16 55 1  A55 
39:16| l39
[1]
1
9,557.8179
 0.11688  
16 9,864.8688
1  0.26092   0.39876
1.06
l 1 9,557.8179

a a39  v16 55 
  a55  15.602    13.057
39:16| l39 1.06 16 9,864.8688
 10.6221 [1]
11  16 l55  11  1 9,557.8179 
a (12) |  
 a  1  v   10.6221   1   
39:16 39:16| 24  l39  24  1.06 16 9,864.8688 
 10.339 [1]
So the reserve is:
9V  50,000  0.39876  0.05  12  82.20  10.6221  12  82.20  10.339
 10,263.49 [1]
[Total 6]
(ii) New monthly premium after alteration
The current reserve plus the value of future premiums must be sufficient to fund the future
benefits and expenses plus the alteration expense. So the new premium equation is:
a (12) |  150  80,000 A

10,263.49  P   0.05P 
a @6% [2]
39:21 39:21| 39:21|
Using tabulated values where possible, we have:
A6% |  0.30327 [½]

39:21
a 6% |  12.309
 [½]
39:21

11  21 l60  11  1 9,287.2164 
a (12) |  
 a  1  v   12.309   1   
39:21 39:21| 24  l39 24  1.06 21 9,864.8688 

 11.978 [1]
So the new premium satisfies the equation:
10,263.49  11.978P  150  80,000  0.30327  0.05P  12.309
Solving this, we find that P  1,245.15. So the new monthly premium is 103.76. [1]
[Total 5]
(iii) Comment
This is a higher premium than before. [½]
The higher sum assured will require the premium to increase. [½]
Extending the term will have the opposite effect, because more premiums are paid and are
accumulated for longer. [1]
However, the increase in sum assured is too great to be covered by an increase in term alone. [½]
[Maximum 2]
Solution X4.9
This question combines together the material on active and passive valuation approaches from
Chapter 20 and embedded values from Chapter 18.
(i) Definition of passive and active valuation approaches
A passive valuation approach is one which uses a valuation methodology which is relatively
insensitive to changes in market conditions … [½]
… and a valuation basis which is updated relatively infrequently. [½]
In contrast, an active approach would be based more closely on market conditions … [½]
… with the assumptions being updated on a frequent basis. [½]

[Total 2]
(ii) Items needed to calculate an embedded value
An embedded value is the sum of the shareholder-owned share of net assets (the excess of the
assets over the liabilities) and the present value of future shareholder profits arising on existing
business.
A methodology is required to:

 value the assets [½]
 value the liabilities. [½]

A basis is required for the following items in order to calculate the present value of the future
profits: [½]
 investment return earned on reserves [¼]
 mortality [¼]
 expenses [¼]
 expense inflation [¼]
 surrender rates [¼]
 rates of conversion to paid up [¼]
 risk discount rate. [½]
A basis is also required to value the liabilities. [½]

[Maximum 3]
(iii) Examples of a passive and active valuation approach
Passive valuation approach
The assets could be valued at historic cost or book value. [½]
The liabilities could be valued passively using a net premium valuation (as this method is fairly
insensitive to yield changes as the net premium is recalculated). [½]
The bases used to value the liabilities and project the profits could be changed infrequently. [½]
For example, the investment return assumption may be ‘locked in’ (ie remain unchanged) over
the life of the contract. [½]
Active valuation approach
A market-consistent approach could be used as follows:

 the market value could be used to value the assets [½]
 the risk-free rate could be used as the investment return to value the liabilities [½]
 the risk-free rate could be used to project the investment return earned on the reserves
when projecting the profits [½]
 the risk-free rate could be used as the discount rate to calculate the present value of the
profits. [½]
The non-economic assumptions could be updated regularly. [½]
The examples given above have been chosen because they are described in Chapter 20. However,
equally valid examples should also be given credit.
[Maximum 3]





using the course.

Solution X5.1
(i) Reserving for the cost of the guarantee
The cost of the guarantee will depend mainly on the difference between interest rates at maturity
and the 4.5% pa guaranteed … [½]
… but also on the relationship between the specified mortality basis and revised expectations of
future mortality rates. [½]
The cost can be quantified as
 Uy
ay@i  Uy
 @4.5%   if i  4.5%

a
 y [1]

0 otherwise
where y is the age at maturity, i is the then current rate of interest, Uy is the value of the unit
fund at maturity age y, and Uy ay@4.5% is the guaranteed minimum annuity payment. [1]
ay@i % will use

The annuity, ay@4.5% will be calculated on the specified mortality basis whereas 
prudent (low) expected future mortality rates. [½]
It has been assumed that everyone will exercise the guaranteed annuity option if it is in the
money (as this is the most prudent assumption to make). [½]
Similarly, it has been assumed that there are no surrenders (or transfers to other providers). [½]
Credit should also be given for other assumptions regarding option take-up rates and surrender as
long as they are reasonably prudent.
As interest rates are the most significant (and uncertain) factor, for the policy under consideration
the company should model interest rates stochastically until maturity age and calculate the cost
of the guarantee for each run. [½]
This would be built into the company’s calculation of its non-unit reserves, the cost being included
as a negative cashflow in the final policy year, for each simulation. [½]
The projected value of the unit fund at maturity should be consistent with the interest rates
projected in each stochastic run. [½]
Then perform, say, 5,000 to 10,000 simulations and take a suitably high percentile (say 97.5th
highest) value of the resulting distribution of non-unit reserves, so as to obtain a prudent estimate
of the required reserve. [1]
[Maximum 5]

(ii) Free choice of retirement age
In each run of projected interest rates, it would be assumed that the policyholder would take the
annuity option at whatever retirement age caused the greatest cost to the insurance company. [1]
(iii) Pricing
In reserving, it was assumed that everyone behaved in such a way as to produce the most adverse
financial effect on the company. This is the most prudent thing to do. In pricing, a more realistic
and less prudent view can be taken – eg that a proportion less than 100% take the option at the
most adverse retirement age. [1]
This proportion would vary, depending on interest rates, since the more valuable the option is,
the less likely it is that anyone would not take it. [½]
However, some policyholders might still prefer to take the cash now (if allowed by regulation)
even if the annuity guarantee was more valuable. [½]
ay@i % .
Less prudent (not so low) mortality rates might be used in the calculation of  [½]
In addition, a lower, less prudent, percentile might be used to value the guarantee, eg 75% rather
than the 97.5% used in reserving. [1]
[Maximum 2]
(iv) Derivative contracts that match the guarantee
A guaranteed annuity rate corresponds to a call option on the bonds that would be necessary to
ensure the guarantee was met. [½]
The exercise price would be chosen so that the bonds generated the required return of 4.5% pa.
[½]
Alternatively the guarantee can be mirrored by a swaption. [½]
This option would swap the floating rate returns at the option date for a fixed return of 4.5% pa.
[½]
[Total 2]
(v) Problems with holding the derivative contracts
Markers, please give credit for alternative examples of the problems described below.
The necessary derivatives may not be available in the market. [½]
For example, swaptions may not have a sufficiently long term to cover the life expectancy of the
policyholders. [½]
The term of the annuity is unknown as it depends on the longevity of the policyholder. [½]
So the company will not know what term of bonds it should buy options for. [½]
The company does not know how large the unit-linked fund will be at retirement. [½]

So it does not know how many options / swaptions it should buy. [½]
Similarly, the company does not know how many policyholders will be eligible to exercise their
option … [½]
… as some policyholders may die or surrender before retirement. [½]
The counterparty to the derivative contracts may default. [½]

[Maximum 3]
Solution X5.2
Here the question asks how the company would prepare the data for the validation process. This
is different from the more usual question about how the company would validate the data.
Validation is still relevant, but only part of the answer.
The first step would be to request the data from the systems department. [½]
This must include all the information that will be needed to project future cashflows and profits
for the in-force business, so will include for each policy: policy type, number of units held, sum
assured, age, term, entry / expiry date, amount of premium, charging rates, etc. [1]
[Full mark for at least four examples, provided they are relevant to unit-linked]
Checks should be performed to ensure that the data are complete, that the data file includes
everything that is needed and it is in the right format. [½]
Then perform validity checks such as reconciling ons and offs, … [½]
... checking highest / lowest / average of age / sum assured / premium, … [½]
... reconciling with other sources … [½]
... and checking that all data are reasonable (eg dates are valid, allocation rate is not 1,000%). [½]
The data would then be split by class of business, contract type and possibly sales method. [½]
The company might value the contracts by model policies or on an individual policy basis. [½]
If model policies are being used then group the data together, maybe by premium size, age, term,
expired duration, different charging structures and so on. [½]
For each group of data decide upon a suitable ‘average’ model policy. [½]
[Maximum 5]

Solution X5.3
(i) Calculation of standard premiums
The premium for a one-year term assurance is:
10,000  0.01
P1   96.62 [1]
1.035
The premium for a two-year term assurance is:
10,000  0.01 10,000  0.01  0.99

P2    189.04 [1]
1.035 1.0352
[Total 2]
(ii) Underwriting the special contract
The terms of the contract are that the option can be exercised without further underwriting at
that time. [½]
However, term assurances would normally be underwritten at outset (for the reasons described
below). [½]
There is a very large sum at risk as the sum assured is considerably bigger than the premium. [½]
The sum at risk will increase further if the policyholder exercises the option. [½]
So the insurer can potentially make large losses if policyholders have heavier mortality than
expected. [½]
Anti-selection may occur if applicants in poor health know that they will be able to obtain cover
without underwriting. [½]
[Maximum 2]
(iii) Calculation of special premium
Let P be the single premium for the special contract.
The expected present value of the cashflows in the first year is:
40,000  0.01
P  P  386.47 [1]
1.035

Half the policyholders do not exercise the option at the end of the first year. The other half
exercise the option, pay the premium P2 and then experience heavier mortality from that point.
So, the expected present value of the cashflows in the second year is:
 40,000  0.01  0.99   P2  0.99 50,000  0.03  0.99 

0.5    0.5   
2
 1.035   1.035 1.0352 
 184.84  90.41  693.13 [1½]
 787.56
The 50% of policyholders that exercised the option at the end of the first year continue to
experience heavier mortality. Of the remaining policyholders, 70% do not exercise the option at
the end of the second year and experience standard mortality, and the other 30% exercise the
option, pay the premium P1 and then experience heavier mortality from that point. So, the
expected present value of the cashflows in the third year is:
 50,000  0.03  0.99  0.97   40,000  0.01  0.992 

0.5    0.5  0.7  
 1.0353   1.035 3 
 
 P  0.992 50,000  0.03  0.992 
0.5  0.3  1   [2]
 1.0352 1.0353 
 
 649.60  123.76  13.26  198.90
 959.00
So the premium is given by
P  386.47  787.56  959.00  2,133.03 [½]

[Total 5]
(iv) Factors to include in premium calculation
Cashflow projections should be performed so that a full profit test of the premium can be made.
A profit loading should be included … [½]
… this would increase the premium. [½]
An allowance for initial, renewal and claim expenses should be included … [½]
… and the additional expenses associated with the option (eg notifying policyholders and
processing their decisions) should also be included … [½]
… these would all increase the premium. [½]
An allowance for surrenders could be made … [½]
… this would reduce the premium if some lives in poor health surrendered … [½]
… but might increase the premium if a generous surrender value is paid. [½]

An allowance for reserves and solvency capital requirements should be included … [½]
… this would increase the premium … [½]
… as the providers of capital usually require a higher return than that earned on the reserves. [½]
[Maximum 4]
(v) Reasons for low sales
The contract may not be a good match for customers’ needs. [½]
There is no flexibility in the term of the contract and three years may be too short. [½]
There is no flexibility in the sum assured and customers may need more or less than the fixed
amount. [½]
Competitors may be offering a similar product for a lower premium. [½]
Competitors may offer contracts with improved features, … [½]
… eg greater flexibility in the timing of exercising the option and/or the amount of extra sum
assured that can be chosen … [½]
… or an option to convert the contract into an endowment assurance or whole life assurance. [½]
The public may be unaware of the product, eg due to a lack of advertising. [½]
The company may have a low reputation with the public, eg following a regulatory fine. [½]
The contract may be unpopular with intermediaries, eg if the company offers low commission. [½]
The economy may be in recession so that many potential customers cannot afford the contract.
[½]
There may be little need for the contract if the state or employers offer generous benefits. [½]
[Maximum 4]
Solution X5.4
The main form of reinsurance wanted will be risk premium reinsurance for contracts offering
significant protection benefits. [1]
The main reason the company wants this sort of reinsurance is to reduce the impact of mortality
fluctuations on free assets … [½]
... and on shareholders’ profits. [½]
With a relatively small pool of in-force business, mortality variations could be a very large
percentage of the expected amount paid in excess of the value of units. [½]
Retention levels might therefore need to be quite low. [½]

The risk premium reinsurance would be operated just like a mortality charge. [½]
The individual surplus method would be most suitable, putting a maximum ceiling on the retained
sum at risk. [½]
If the volume of past experience data is quite small, the company may seek reinsurance to protect
it against pricing errors. [½]
However, this would be less important if charges to policyholders were reviewable frequently, as
they probably would be under unit-linked business. [½]
Quota share reinsurance would be used to reduce the exposure to pricing risks, if needed. [½]
Reinsurers can also be used for their technical assistance, and advice with pricing can help reduce
the risk of the company setting incorrect charging rates. [½]
Other areas for help might be in underwriting and product design. [½]
The company may need financial reinsurance to reduce new business strain, … [½]
... depending on the extent to which the unit-linked products have been designed to reduce initial
capital requirements. [½]
Any such financing requirement could be effected on the back of the risk premium reinsurance by
a suitable combination of commission and premium levels, … [½]
… eg by receiving large initial reinsurance commissions in return for increased risk premiums. [½]
A quota share approach would be most effective here. [½]
However, this form of financial reinsurance may not be effective under accounting or supervisory
regimes where a realistic liability has to be held in respect of the increased risk premiums. [½]
The company might seek catastrophe reinsurance to cover any large effects of non-independent
claims. [½]
However, with a low normal retention limit on the individual surplus cover, the extra benefit
provided by catastrophe cover would probably not justify the cost. [½]
As the company is small, stop-loss reinsurance might be used to cover total losses on the whole
portfolio of policies (although, again, this is likely to be expensive). [½]
[Maximum 8]

Solution X5.5
(i) Retention limit
A simulation approach might be adopted. This would involve the following steps:
 Decide upon some criterion for claim volatility beyond which the company cannot go. The
aim may be to keep the probability of insolvency below a specified level. [½]
 Combine a stochastic model for expected claims rates together with a model of the
business to project forward claims, assets and liabilities. [½]
 The business may not be easy to model because it is group life and so the data available
may not be of sufficient detail. [½]
 Use a simulation method to determine a retention level such that the company stays
solvent for, say, 995 times out of 1,000 runs. The model will need to be run lots of times
at differing retention levels, in order to find the level that meets this criterion. [1]
 Sensitivity analysis may also be carried out, varying the mortality probability
distribution. [½]
 The company may also check to see if the protection can be done more cheaply using a
mortality fluctuation reserve. [½]
 This would involve increasing the insurer’s retention and using the reduction in cost to
finance a mortality fluctuation reserve. [½]
 Model the net profit assuming this new arrangement and compare the protection against
that offered by the previous arrangement. [½]
 Try this for various levels of reinsurance / mortality fluctuation reserve and decide on
which offers the required protection for lowest cost. [½]
 Another possible approach is based on the theory of efficient investment frontiers, and
looks at reinsurance as an asset class that allows the firm to optimise its risk and reward
trade off. [½]
 The result should be checked against the market norm for reasonableness. [½]
[Maximum 5]
(ii) Apportionment
(a) Claim of 90,000
Reinsurer A incurs one-third of the cost: 30,000.
The balance is less than the excess point and so is retained by the company: 60,000. [½]
(b) Claim of 290,000
Reinsurer A incurs 96,667.
The balance above the excess point is taken by Reinsurer B: 113,333.

The company is left with the excess amount: 80,000. [½]
(c) Claim of 1,000,000
The balance above the excess point is taken by Reinsurer B: 586,667.
The company is left with the excess amount: 80,000. [½]
(d) Claim of 1,200,000
The balance above the excess point is greater than the maximum reinsured and so Reinsurer B
incurs just the maximum: 600,000.
The company is left with the remainder: 200,000. [½]

[Total 2]
(iii) Protection against very large claims
‘Very large claims’ means claims exceeding 1.02m, as under the existing arrangements claims up
to this amount will leave the company with a retention of not greater than 80,000. [1]
To protect against even larger individual claims, additional surplus cover could be bought. [½]
This may be purchased through a treaty but since the number of individuals that could lead to
very large claims is small, facultative cover may be found for these cases. [½]
To protect against a number of claims resulting from one incident, catastrophe excess of loss
reinsurance could be used. [½]
The contract will specify how much the reinsurer will pay if a catastrophe occurs and also the
definition of a ‘catastrophe’. [½]
The amount might be the excess over a retention limit although there is likely to be a maximum
claim amount. [½]
Claims above this maximum could be reinsured by a second line of catastrophe cover, protecting
the company from much rarer and larger claim events and paid for by a separate premium. [½]
More lines of catastrophe cover could be added as required. [½]
Stop-loss reinsurance might also be used. This targets the company’s claim experience over its
whole portfolio over a year, paying total claims in excess of a certain level up to a maximum. [1]
[Maximum 4]

Solution X5.6
Status of the legislation
There may be some possibility that the law is changed or rejected, so exact implications are
difficult to predict. [½]
However, need to act in the meantime on the assumption that the legislation is enacted as
proposed. [½]
Effect of legislation on existing business
In-force business was written with guaranteed premiums, and making certain assumptions about
contract terms. [1]
The implications of these contract terms being changed retrospectively are potentially severe. [½]
AIDS is a significant problem in this country, and so the potential losses could be very large. [½]
This is compounded by the fact that the company has written significant volumes of business that
is potentially affected. [½]
The losses would damage the financial strength of the company … [½]
... and lead to reduced dividends to shareholders. [½]
The extent to which it is adversely affected by the new law depends also on: [½]
 how successful its underwriting has been in keeping actual and potential HIV sufferers out
of its insured population [½]
 how bad the AIDS epidemic turns out to be [½]
 how effective the exclusion clause would have been in identifying the AIDS-related
deaths. [½]
The exclusion would very likely have been much less than 100% effective … [½]
... as the cause of death might not be stated as AIDS, and proving AIDS may be difficult or even
impossible in some cases. [½]
The company could investigate past levels of applied exclusions to try and quantify the effect of
not being able to apply it. [½]
This may understate the likely cost because:

 some potential claimants may have been put off from claiming because of the exclusion
clause [½]
 the proportion of persons with HIV in the population was probably lower in the past
(although approximate adjustment may be made for this). [½]

The lapse rate may decrease. Previously, policyholders with HIV / AIDS may have lapsed their
policies in the belief that any claim was unlikely to be paid due to the exclusion clause. [1]
The target markets for the business will also be important. [½]
It may be, for example, that the subsets of the population targeted are relatively free of AIDS. [½]
Connected with the issue of the future losses is the question of solvency requirements. [½]
The expected increase in claims means that the supervisory reserves would have to be increased.
[½]
And because we are dealing with a very uncertain future cost, … [½]
… the associated reserves should be calculated assuming a big margin on top of ‘best estimate’
… [½]
… and/or the solvency capital held against mortality risks should increase sharply. [½]
This effect could be quite substantial and might make a noticeable difference to the company’s
free asset ratio. [½]
Knock on consequences of this would include:

 reduced investment freedom [½]
 lower confidence from potential customers and intermediaries [½]
 reduced ability to write new business [½]
 in the extreme, the company’s solvency position may be threatened. [½]
Could try to reinsure the existing business against the future risk, … [½]
... but suitable reinsurance should cost more than the expected future cost (unless a very
optimistic reinsurer). [½]
However, such reinsurance could at least reduce any capital problems. [½]
Effect on future business
The company could increase premiums to allow for the increase in likely AIDS related claims. [½]
However, if there is considerable uncertainty about future levels of AIDS-related claims, … [½]
... the necessary increase could be very large and would make the contracts much less attractive
to most policyholders than they are now. [½]
Without a large increase in premiums, the company might find that the business still required
substantially higher reserves. [½]
This would decrease the capital efficiency of the product and decrease free asset ratios. [½]
The company could try to underwrite more severely to exclude all people who are HIV positive or
likely to become so (assuming the legislation allows this). [½]

An undesirable consequence is that the level of underwriting required would be unpopular (the
company might have to ask all applicants to have HIV blood tests and to answer highly personal
lifestyle questions) … [½]
... and expensive. [½]
The company might even introduce periodic re-underwriting for AIDS, although again this would
be expensive. [½]
Introducing reviewable premiums, either within the current non-unitised contract or in unitised
form, … [½]
... would reduce the risk from future uncertainty. [½]
This might reduce the increase in premiums needed. [½]
If there was a strong correlation between incidence of AIDS and some other factor such as
occupation, then could try to rate by that. [½]
Although it may be difficult to get any useful and reliable data. [½]
The company might seek to increase its reinsurance, … [½]
... but reinsurers will be aware of the problem and so the cost of reinsurance may be high. [½]
In the extreme, the company could consider withdrawal from the market. [½]
Even though this is a significant class of business for the company, there is clearly no point in
writing new business if the company does not believe it can do so profitably. [½]
[Maximum 21]
Solution X5.7
Sources of information
The proposal form can ask for health related information about the applicant eg height, weight,
smoking and drinking habits, personal and family medical history. [1]
It can ask lifestyle questions eg travel and dangerous pastimes. [1]
The proposal form can also assist financial underwriting by asking the applicant’s salary and for
information about the applicant’s dependants. [1]
Reports can be requested from doctors that the applicant has consulted. [½]
A medical examination can be carried out on the applicant. This will include simple checks
eg height, weight, blood pressure, and possibly blood or urine tests. [1]
More specialised tests may be required eg electrocardiograms and chest X-rays. [1]

Underwriting procedures
The proposal form will be checked by a member of the clerical staff or by specialised underwriting
software. Further information may be required if the proposal form indicates any potential
problems eg serious illnesses at a young age in family medical history. [1½]
More complex cases will be passed to a professional underwriter who will decide the information
required to make a decision. [½]
They may be assisted by underwriting manuals provided by reinsurers. [½]
Underwriting will be stricter for higher sums assured. [½]
For sums assured above a certain level, further information may be required automatically
regardless of the information provided on the proposal form. [½]
[Maximum 8]
Solution X5.8
(a) Continues not to offer discount
It may be uncompetitive for non-smokers and attract only smokers. [½]
May cause a loss in non-smoker business written, as they take their business elsewhere. [½]
This may cause a deterioration in experience. [½]
This would lead to worse premium rates that would attract only the worst risks, engendering a
downward spiral. [½]
(b) All contracts, same basis
It may be possible to get useful data (ie based on reinsurers’ / industry data) to calculate the
discount. [½]
Following the market should prevent anti-selection that would be a risk under (a), … [½]
... but does nothing innovative to attract business. [½]
There will be little effect on contracts that are fundamentally for saving. [½]
For short-term savings endowments and other savings contracts, a smoking question may be
resented as irrelevant. [½]
(c) Strict definition
May discourage non-disclosure. [½]
But, medical examination is costly – it may outweigh the reduction in risk. [½]
Medical examination will also discourage business. [½]

Reinsurers’ / industry data will not be useful. [½]
Relatively few people will be entitled to the discount, so terms for the majority of potential
business may be unattractive, reducing volumes. [½]
May be unnecessarily over-cautious (eg in requiring people never to have smoked). [½]
(d) Non-smoker discount lapses if they smoke
How will the company know? – impractical. [½]
Unlike for general insurance, life insurance policyholders do not generally tell insurer of changes
in the risk. [½]
Unfair to repay – eg if someone starts smoking nine years into a ten-year contract. [½]
People may lapse rather than repay the discount. Such lapses may also be selective. [½]
(e) Same basis, some contracts
Probably the best option. Protects the company against anti-selection that was a risk in (a), … [½]
... but any underwriting and admin cost is concentrated on contracts where it is most relevant and
needed. [½]
Useful data (ie reinsurers’ / industry statistics) are available. [½]

[Plus up to 1½ marks for a reasonable discussion]
[Maximum 10]
Solution X5.9
(i) Why use reinsurance to manage the mortality risk?
Without reinsurance, some cases may be too large for the company to take on board even if the
lives are acceptable at standard rates. This could seriously impair the marketability of the
product. [½]
Using reinsurance will pass some of the mortality risk to the reinsurer. Therefore large cases may
be accepted and the excess above a certain level reinsured. Reinsuring a higher proportion of
larger risks (by using individual surplus, say) will reduce claims volatility and help produce more
stable profits. If the company is a proprietary, this will help meet its shareholders’
expectations. [1]
Catastrophe excess of loss reinsurance will be needed where there is a significant concentration
of risk. This is often the case with group business, where for example, an explosion in a head
office building may lead to many non-independent claims. [½]
Since reinsurance reduces risk, it can be used to help a company expand more quickly, without
increasing the risk beyond acceptable levels. This may be useful for a small company or a
company writing a new class of business where future volumes are hard to predict. [1]

In the latter case, the reinsurer will be able to help set the mortality assumption to be used in the
pricing basis, so reducing the risk of parameter error. [½]
Reinsurance also offers the company some protection against where it has its pricing basis or
underwriting strategy wrong. [½]
If a reinsurance treaty is in place, large risks may be accepted quickly, without the delay of trying
to find a third party willing to take on part of the risk. [½]
[Maximum 4]
(ii) Why still underwrite?
The most important reason is that the terms of the treaty are likely to require it. The likely claims
experience is heavily dependent on the level of underwriting in place, so this will have a material
impact on the terms agreed. [1]
The reinsurer will expect the same underwriting process to remain in place throughout the
duration of the treaty. If the process changes, the treaty will need to be renegotiated. [½]
The better the initial level of underwriting, the better the terms available from reinsurers. [½]
Since the direct writer will be keeping some of the risk, it will still want to reduce its expected
claims experience and volatility as much as it is cost effective to do so. [½]
If the treaty contains a profit sharing clause, this becomes even more important. [½]
Underwriting will reduce the opportunity for anti-selection. Reinsurance is not an alternative
strategy for achieving this. [½]
[Maximum 3]





using the course.

Assignment X6 Solutions
Solution X6.1
Markers should bear in mind that there is not a unique right answer, so credit should be given for
anything that is sensible.
(a) Overhead costs
These may be broken down into smaller groups, eg Director of Pensions allocated to ‘pension
business’. They would then be allocated to the policies on a per-policy basis by some reasonable
method. [1]
For example, there might be more detail from the manager’s timesheet, such as time spent on
new / existing business, between policy classes, etc. [½]
Another example would be fixed ‘investment’ overheads, which might be allocated in proportion
to funds under management. [½]
(b) Branch office costs
The branch costs will probably be new business costs. [½]
However, they could do some on-going admin and so this part could be included as servicing
costs. [½]
Salary costs would be allocated between these, and between different classes of business, in
proportion to time spent (from analysing timesheets). [½]
Costs such as staff involved in new business could be allocated on the basis of per new business
policy or as a percentage of commission. [½]
The latter implies that there is a bigger cost involved in large cases. This is probably true to a
limited extent. [½]
If the properties are rented then these costs would be included. [½]
If the properties are owned by the policyholders’ fund then a notional rent should be charged as the
cost. [½]
This would be subdivided between departments in proportion to floor space occupied, … [½]
… then allocated between new / renewal and product type in the same proportion as the relevant
salaries. [½]

(c) New computer
This is an overhead cost and so would be treated the same as (a). There could be some initial
allocation to products if only part of the system is being replaced. [½]
An annual cost would be derived from spreading the cost of the system over its estimated
working lifetime. [½]
The cost could then be allocated between products in proportion to computer usage, if known. [½]
(d) Head office buildings
Again this is an overhead cost. The office will probably be an asset of the company, in which case
once the office is built a notional rent would be charged to each department, probably based on
the floor space occupied. [1]
This would be split between products, and between new / renewal / claim costs, in the same
proportion as for salaries (similar method to branch property costs). [½]
[Maximum 9]
Solution X6.2
Expenses are likely to be different for term assurance and pension policies. [½]
In particular, heavy underwriting for term assurances means that expenses for this class are likely
to be heavier. [½]
Term assurances and pensions policies may differ in other ways affecting expenses, … [½]
… for example the mix between single and regular premium business. [½]
Not all expenses are ‘per policy’. [½]
For example, underwriting expenses may be more closely linked to size of sum assured. [½]
The period of three months may not be typical, for example new business sales often vary by time
of year. [½]
The period may not reflect what is expected in the future, especially with regard to business
volumes. [½]
The historical expenses would also have to be adjusted for inflation since the middle of the period
… [½]
… and up to the middle of the period for which the premium rates are expected to be in force. [½]
New business processing is by no means the only aspect of initial expenses. [½]
For example there will be product development expenses to cover. [½]
Just using processing expenses could seriously understate total initial expenses. [½]

On the other hand, the department may not be exclusively concerned with new business. [½]
[Maximum 6]
Solution X6.3
The company would start by analysing recent persistency experience on its term assurance
portfolio. [½]
It will need to choose a long enough period to give credible results, … [½]
… but will not wish to look back so far that the data are not relevant. [½]
It will wish to classify the data by the relevant factors that affect persistency experience. [½]
The most important factors are likely to be:

 duration in force [½]
 sales method and target market. [½]
Other variables that will affect persistency rates and which may be used to classify the data are:
 frequency and size of premium [½]
 premium payment method [½]
 original term of contract [½]
 gender and age. [½]
The extent to which this classification is possible will depend on the volume of data available. [½]
When classifying by variables such as the above, groupings will need to be quite large (eg age by
ten-year bands) to ensure reasonable quantities of data per cell. [1]
Once the data are classified as required, persistency rates over each year could be determined as
the number of policies surviving the year divided by the number of policies in force at the
beginning of the year. [1]
Withdrawal rates by year in force could then be derived as one minus the persistency rates over
each year. [½]
Alternatively divide the numbers of withdrawals by the appropriate exposed-to-risk at each

duration.
Persistency rates can also be calculated cumulatively as the number surviving the year divided by
the number in force at the outset of the contract. [½]
Might also compare actual persistency with that expected according to the current persistency
assumption. [½]

Need to bear in mind how past economic conditions may have affected the past persistency
experience, … [½]
… and it may be necessary to adjust the estimates if they have been significantly affected by any
one-off or unusual past events. [½]
The company might also look at industry persistency data for this class of business … [½]
… but would be used with caution due to the heterogeneity in the data. [½]
It may provide some information on trends or indicate more clearly the effects of economic
recession, etc. [½]
In forming pricing assumptions the company will need to adjust these figures for anything that
has changed since the period which has been analysed. [½]
Examples might be:

 changes in the tax regime and other legislation affecting the contracts [½]
 the economic situation [½]
 the competitive situation [½]
 change in sales channel. [½]
The company will be particularly concerned about the effect of early lapses on profitability, and so
its assumptions may be on the prudent side of realistic. [1]
It may also sensitivity test on higher rates of early lapse. [½]

[Maximum 12]
Solution X6.4
Accumulating with-profits contracts are discussed in Chapter 6. The topic of risk is covered in
Chapters 10, 11, 12 and 29.
(i) How pricing allows for profit distribution method
The ‘price’ charged for AWP products is usually in the form of charges, that may be deducted from
the premium and/or from the benefit fund. [1]
These charges will be designed to cover the insurer’s expenses. [½]
The insurer will usually use profit-testing pricing methods to determine the appropriate level of
charges to meet its profit objectives. [½]
In deriving the charges, the company may have to make projections of future bonus rates so that
such things as future fund-management charges can be projected. [1]
The projected future bonus rates will be based on the assumed return on the underlying
investments and the intended split between regular and terminal bonus payments. [1]

The prices charged will reflect the extent of guarantees provided. [½]
For example:
 a charge may be levied to cover the (minimum) guarantee of zero future fund growth [½]
 a charge for the smoothing of bonuses may be levied, to cover the cost of the capital
needed [½]
 the expense charges may be guaranteed for the full term of the policy. [½]
The first two of these would probably be paid for via reductions in the asset share calculations,
which will result in lower bonus rates being declared. [½]
The third example would be met by explicit margins in the expense charges. [½]
An alternative is for there to be no explicit charges but the insurer to net off the expense costs
through reductions in the regular bonus rates. [½]
In this case the product does not need to be priced, as such, as all the costing is retrospective –
ie the experience impacts on the benefits after it is happened, from year to year. [½]
[Maximum 5]
(ii) Risks from adopting the bonus system
There is the risk that the product and the bonus distribution system will not be easily understood by
those involved in the sales process (be they brokers, a direct salesforce or telesales people), and by
the customers. [½]
This could impinge on sales volumes, and if the product is not sold in sufficiently large volumes,
development costs may not be recouped, … [1]
... and there may be an insufficient contribution to other fixed overheads. [½]
If the new with-profits design sells well, other companies will begin to enter the market and reduce
the company’s market share. [½]
To ensure high sales volumes of the product the company will probably need to quote high final
benefits at maturity; … [½]
... there is then the risk of not achieving the investment performance implied by the projection, and
disappointing PRE. [½]
There is a risk of high withdrawal rates after a few years if policyholders perceive their regular bonus
interest payments to be low, perhaps not realising the possible impact of terminal bonus at
maturity. [½]

Withdrawals will be a risk:

 at early policy durations, depending on the extent to which there may be a penalty to
recover unrecouped initial expenses on surrender [½]
 where fund benefit values exceed asset shares, eg through smoothing, depending on the
extent to which a market value reduction is applied [½]
 through loss of future profits from any foregone future charges, on surrender at any
time. [½]
Any deduction from surrender values made for any or all of the above will also create the risk that
policyholders will be disappointed, leading to bad publicity and lower future sales. [½]
There will be a similar risk if the surrender values (even without penalties) do not match up with any
illustrations given to policyholders earlier, eg at time of sale. [½]
If there are significant guarantees, and lack of early surrender penalties, it is possible for the policy
to have a significant capital requirement. [½]
There is then a risk to the insurer’s solvency of selling too much of the new product too quickly. [½]
The company will need reasonable levels of free assets to be able to smooth out the investment
return volatility from year to year. [½]
If the company is not sufficiently strong, such smoothing will be minimal – the resulting fluctuations
in regular bonus (and terminal bonus, if applicable) may lead to bad publicity and low sales. [½]
The company will ideally want to back at least some of its future (discretionary) bonus liability,
including any terminal bonus, by equities and property. [½]
The extent of this will again be limited if the company does not have enough free assets, because of
the need to smooth bonus payments from year to year. [½]
The extent of profit deferral is likely to be low, because of marketing pressures to keep regular
bonuses high, so little additional investment flexibility is likely to arise from this source. [½]
Therefore, if this is the company’s only product using this bonus system, it will be difficult to build
up a strongly diversified equity and property portfolio, further exposing the company to the specific
risk of the individual investments. [½]
The company may also have little experience in the equity / property sectors (although this may not
be an issue if the company, which is small, has decided not to invest in house.) [½]
The guaranteed benefits should be backed by government bonds of similar term, but many of the
policies will have a very long term and it could be difficult finding the appropriate stocks. [½]
Thus there could be a mismatch risk. [½]
There may be a risk from high future expenses and inflation if the expense charges on the policy are
guaranteed. [½]

Even if charges are reviewable, increasing the charges in the future can lead to marketing risks,
similar to those caused by poor investment performance. [½]
The admin systems and the company’s actuarial expertise may be inadequate to manage the new
system well. [½]
The directors may be very tempted to overpay bonuses for marketing reasons, to the significant
detriment of any free assets and overall strength of this small company. [1]
[Maximum 11]
(iii) How bonus system will affect financial management of product
‘Financial management’ is a not very well defined term. It is like saying ‘how can the company act
to improve its profits and solvency capital?’. Answers will therefore call for very broad idea
generation.
To reduce capital strain, the company will try to reduce its supervisory solvency requirements. [½]
It can help achieve this by:

 avoiding onerous guarantees (such as a non-zero guaranteed fund growth rate or
non-reviewable charges) [½]
 having a matching charging structure, eg a high initial expense charge followed by small
regular charges [1]
 having an appropriate surrender penalty. [½]
This will make it easier for the company to sell higher volumes of new business, and so generate
more profit more quickly. [½]
The company could also try to improve its capital position by trying to defer as much surplus as
possible to be distributed in terminal bonus form. [1]
This is a way of creating working capital to help finance further new business … [½]
... and to allow more risky assets to be chosen for the policyholders’ fund. [½]
This will be constrained by competitive pressures, because policyholders like to see their guaranteed
benefits growing well. [½]
This is especially true for a pensions product, where a person’s financial planning will be easier if
they are more sure about their eventual benefit. [½]
There is a big change in the management of the bonus distribution, as all the bonuses are now fully
discretionary as opposed to the automatic system of the revalorisation approach. [½]
Asset share calculations and model office projections will be more important tools in helping to
determine appropriate bonuses to declare. [½]
Model office projections will require much more careful modelling of the future bonus distributions
than hitherto, as a non-automatic bonus distribution strategy will need to be modelled. [½]

Considerable judgement is needed in determining, for example, appropriate dynamic links to use in
the model. [½]
Great care will be needed not to overstate the solvency position of the company (by using
over-optimistic assumptions, for example). [½]
The company will need to use asset share methodology in other areas, for example when
determining how much terminal bonus to include in surrender values and for setting the Market
Value Reduction factor to be used. [1]
In the event of expense experience on the product being disappointing, the company may be able to
reduce bonuses or increase charges accordingly, … [1]
... although this would be subject to competitive and PRE considerations, and on any guarantees in
place. [½]
[Maximum 8]
Solution X6.5
Free assets are covered in Section 3.3 of Chapter 28. The term ‘financial strength’ is defined in the
glossary.
(i) Measuring financial strength
The financial strength of a life company refers to its ability:

 to withstand adverse changes in experience, [½]
 to fulfil its new business plans, and [½]
 to meet policyholders’ reasonable expectations. [½]
This is often measured by the free assets of a company, because free assets will assist the
company in continuing to meet its business objectives. [½]
Adverse changes in experience will reduce the company’s free assets, … [½]
… either due to actual losses on the revenue account or due to the changes being reflected in
higher solvency requirements (or both). [1]
Free assets will allow this to happen without jeopardising cover of any required solvency
capital. [½]
Free assets could also be used to explicitly reserve for fluctuations – eg a mortality fluctuations
reserve. [½]
Writing new business will normally involve new business strain. [½]
Substantial free assets will allow the company to suffer this strain without endangering solvency,
… [½]
… and without having to resort to other options which it might not want to do (eg financial
reinsurance, or redesign of products). [½]

Policyholders’ reasonable expectations: to satisfy these the company might need to make
financial decisions which it would not otherwise make … [½]
… such as high bonus declaration to honour a marketing ‘promise’ despite disappointing

investment performance. [½]
Free assets will allow the company to do this. [½]
In addition free assets will allow the company to pursue a riskier investment strategy than it could
otherwise do, … [½]
… hence achieving the high returns that the policyholders might be expecting. [½]
[Maximum 7]
(ii) Size of free assets
A company will normally aim to achieve some target rate of return on its embedded value. [½]
The total return will result from the future returns achieved on the life insurance operation (ie the
present value of future shareholder profits from the life portfolio), … [½]
… and the return achieved on the free assets. [½]
The desired rate of return will normally be several percentage points higher than market yields on
low-risk investments (eg fixed interest government stock). [½]
Free assets will almost certainly earn less than that desired rate of return, otherwise the company
would rationally dispense with the life operation and the associated risk. [½]
So the company seeks to achieve a high return on the insurance business, and this return is then
reduced by a lower return on free assets: ie holding free assets reduces the return earned by the
company. [1]
So the company should want the minimum level of free assets which gives it sufficient financial
strength to pursue its life insurance operations as it wants to. [½]
[Maximum 3]
Solution X6.6
(i) Reasons for discrepancy
This part may initially feel like it can be answered using the persistency experience analysis
bookwork in the Core Reading – but it goes beyond that. It is all about consistency in comparing
investigation results, so it is necessary to think about what consistency can mean in this context.
However, a number of relevant points can be generated by thinking about the sort of factors that
influence persistency rates, such as economic conditions and duration in force.
There are three investigations under consideration: the ALAC set, the company’s own experience,
and investigations by the company into the experience of individual competitors.

The investigations may involve different time periods, … [½]
… in particular the latest industry wide investigation might reflect nothing more recent than two
years ago, while the company’s latest internal investigation should be in respect of last year. [1]
Such a time difference could be significant if economic conditions have changed much over the
last one or two years, because persistency experience can be very sensitive to this. [½]
The investigations may focus on different products, and the company’s major products may be
minor ones by industry standards. [½]
If the company’s main competitors are not big players in the market (competing in a small niche
market?) then their experience will not necessarily be expected to be in line with industry
experience. [½]
The methodology of the investigations might be different. [½]
In particular, one set could be expressed in terms of numbers of policies, while another could be
weighted by the premiums or benefit levels involved. [½]
Similarly, one investigation might have considered only lapses and surrenders, while another
consider partial withdrawals and paid-up contracts too. [½]
Categorisation might be inconsistent between the three investigations, eg if ‘early’ withdrawals

means year 1 to this company but years 1–3 for ALAC. [½]
There might be other inconsistencies, eg do the studies include all sales channels or does the
ALAC study involve broker business only? [½]
How does the company know its competitors’ rates? If done by working from their published
financial information and trying to deduce movements, the estimate may contain a large margin
of error. [½]
[Maximum 4]
(ii) Checks
Check the persistency data with reference to the formula:
in-force at start + new policies – withdrawals = in-force at end [½]
by product class and duration in force (moving from, for example, duration t at start to duration
t  1 at end of year). [½]
Check that the data underlying the investigation reconcile with published accounts / any other
published information (eg supervisory returns). [½]
Check that the results are consistent with previous years’ results. [½]
Look at the analysis of valuation surplus: is the withdrawal surplus consistent with the experience
results? [½]
Likewise look at the analysis of the change in embedded value. [½]

Check actual withdrawals vs expected withdrawals from recent model office projection work. [½]
[Maximum 3]
(iii) Advantages and disadvantages
If the company’s results are incorrect, and these lower persistency rates better reflect reality,
then the suggestion is sound for all of the areas of work concerned. [½]
However, assuming that the internal results are accurate and that the MD is suggesting that the
company uses persistency rates that are lower than reality, it needs to consider what a lower
persistency assumption means in terms of prudence, … [½]
… and also to what extent the company wants prudence in the areas of work mentioned. [½]
First, what does a lower persistency assumption imply for prudence?
Lower persistency will be more prudent where discontinuance terms are such that the company
loses out. [½]
This is probable at early durations (especially with negative asset share), … [½]
… but should not be the case later on unless the company is deliberately paying out more than
asset share for PRE / marketing reasons. [½]
Lower persistency rates will also be more prudent if there are potentially onerous guaranteed
discontinuance terms. [½]
Lower persistency rates will be less prudent if the discontinuance terms are such that the
company, overall, makes a profit on discontinuance (ie a profit in excess of that which it would
have accrued were the policies concerned to have continued). [½]
This is unlikely, unless there are no PRE / marketing pressures on the company. [½]
Next, is extra prudence wanted?
Reserving
For reserving for demonstration of solvency to supervisors, possibly yes – if there is doubt about
the validity of the persistency assumptions, it is better to use a more prudent set. Also, using
industry data may be more easy to justify to supervisors. [1]
But the extent to which this is an issue will depend on the local statutory reserving regulations. [½]
For instance, if these prescribe ‘choose a 0% or 100% withdrawal rate for each product class (or
policy) depending on which is more prudent’ (which has been a common approach), then the
whole question is immaterial. [½]
The extent to which extra prudence is wanted will depend on how prudent the basis already is in
aggregate, … [½]
… the financial impact of any strengthening of reserves, … [½]
… and how much extra security any required solvency capital may already add. [½]

Embedded value
Normally (eg internal company work) a best estimate is required. [½]
It does not really matter too much if a parameter is wrong for one year, as after one year, the
company should have a better idea of the situation. (Unless, that is, the EV is being used to set a
sale value.) So it is better not to have extra prudence. [1]
Pricing
A best estimate is likely to be wanted, but it is very important to get assumptions right, especially
for conventional without-profits business. So normally go for the prudent side of best estimate.
[1]
If certain that the company’s figures are correct, stick with them (ie use the company’s figures
with a small margin, depending on the guarantees involved). [½]
If there is some doubt about the validity of the company’s own assumptions, it is better to move a
bit more in the direction of prudence, especially if rates are guaranteed. So the MD’s suggestion
could be reasonable here (if higher w x = more prudent for the products concerned), assuming
that this does not put the company out of the market. [1]
Greater prudence in the assumptions could justify a decrease in the risk discount rate. [½]
[Maximum 10]
Solution X6.7
(i) Principles of investment
A company should select investments that are appropriate to the nature, term and currency of
the liabilities. [1]
The investments should also be selected so as to maximise the overall return on the assets, where
overall return includes both income and capital gains. [1]
The extent to which the first principle may be departed from in order to meet the second will
depend, all other things being equal, on the extent of the company’s free assets and the
company’s appetite for risk. [1]
[Total 3]
(ii) Suitability of investments
(a) Long-dated domestic government bonds
These assets would be appropriate to match long-term liabilities fixed in monetary terms, … [½]
… such as the guaranteed sum assured and any declared bonus on policies some way from
maturity. [1]

This is true for the UK-style ‘additions to benefits’ surplus distribution method (ATB) and the
revalorisation method. [½]
Index-linked securities (if available) are a reasonable match for the company’s expenses. [½]
Government bonds are less suitable for discretionary benefits under ATB, … [½]
… because bonds are expected to under-perform equities in the long term … [½]
… and do not offer protection against unexpected inflation. [½]
Under the revalorisation method there is no deferral of distribution of profit. [½]
This makes equity investment prohibitively risky. [½]
Government bonds may therefore be the predominant asset class backing with-profits business
using revalorisation. [½]
The returns are denominated in the local currency and so domestic bonds are a good currency
match. [½]
(b) Policy loans
These would offer fixed returns above cash yields, … [½]
… are appropriate for liabilities fixed in monetary terms and are reasonably secure. [½]
Policy loans may not offer returns as high as those on equities, but this depends, among other
things, on the rate of interest used to calculate the loan. [½]
Also, the term of the investment would be chosen by the policyholder and so the company would
not be in control of this. [½]
The returns are denominated in the local currency and so are a good currency match. [½]
(c) Ordinary shares
Equities offer real long-term returns, which is arguably what policyholders are seeking from
with-profits savings policies. [½]
Equities are therefore a suitable ‘match’ for future bonuses. [½]
This is particularly so for the discretionary bonuses, eg terminal bonus under ATB. [½]
Also, the highest long-term nominal return would be expected from this class. [½]
The returns are denominated in the local currency and so are a good currency match. [½]
The market value is volatile, but with a high level of free assets the company can absorb
movements in the capital value. [½]
(Equity investment would be prohibitively risky for the revalorisation method – see earlier.)

(d) Property
Property investment also gives an expected real long-term return … [½]
… and so would be suitable for non-guaranteed liabilities. [½]
Its disadvantages (volatile capital values in medium term, large unit size, expensive to administer,
illiquid) … [½]
… would not greatly trouble a company with significant free assets investing for the long term. [½]
Property is therefore a useful diversification from equities in pursuit of best returns. [½]
(e) Overseas shares
For with-profits liabilities these provide a good match in terms of the nature of any discretionary
liabilities, … [½]
… but not in terms of currency. [½]
The uncertainty of capital values and the currency risk both make these poorly suited to with-
profits business under the revalorisation method. [½]
The company might want the currency mismatch because it gives more diversification by
economy; … [½]
… alternatively, the currency risk could be hedged (at a cost). [½]
In the long-term overseas equities might provide better returns than domestic equities and so
might be more appropriate from the point of view of maximising returns. [½]
It is possible that the company has some overseas business originating from an overseas branch,
in which case overseas equities could give a suitable currency match. [½]
[Maximum 15]
(iii) Small company with limited free assets
Closer matching [½]
The company would need to ensure that the nature, term and currency of the assets are closer to
those of the liabilities. [½]
This is because there is less of a cushion against adverse experience. [½]
This might mean more bond investment and less in equities and property as the volatility of
equities and property poses too great a solvency risk for this level of free assets. [1]
Liability valuation
It is possible that the valuation of liabilities is performed at an interest rate that is derived, or in
some way related to, the yield on the asset portfolio. [½]

Pressure on the free assets is likely to favour fixed interest in order to reduce the value of the
liabilities. [½]
This would also lead to increased investment in fixed-interest bonds, less in equities and
property. [½]
Similarly, any solvency capital requirement is likely to be greatest for equities and property
leading to increased investment in fixed-interest bonds. [½]
[Total 4]





using the course.

SP2 Study Mat

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SP2 Study Mat

Uploaded by

Copyright:

Available Formats

Life Insurance

The Actuarial Education Company

You must take care of your study material to ensure that it is

Legal action will be taken if these terms are infringed. In

These conditions remain in force after you have finished using

© IFE: 2019 Examinations The Actuarial Education Company

The guide is split into two parts:

Part 1 Section 1 Before you start Page 2

The Actuarial Education Company © IFE: 2019 Examinations

1.1 Before you start

© IFE: 2019 Examinations The Actuarial Education Company

1.2 Core study material

The Actuarial Education Company © IFE: 2019 Examinations

B (t ) is the base value, or divisor, at time t.

© IFE: 2019 Examinations The Actuarial Education Company

1.3 ActEd study support

Successful students tend to undertake three main study activities:

LEARNING LEARNING & REVISION REVISION & REHEARSAL

Combined ASET Mock Exam

The products and services are described in more detail below.

The Actuarial Education Company © IFE: 2019 Examinations

Each chapter includes:

‘Learning & revision’ products

Combined Materials Pack (CMP)

Don’t underestimate the benefits of doing and submitting assignments:

© IFE: 2019 Examinations The Actuarial Education Company

Regular and Block Tutorials

The Actuarial Education Company © IFE: 2019 Examinations

You can choose one of the following types of tutorial:

 Do you have a regular train or bus journey?

 Do you want to fit more study into your routine?

© IFE: 2019 Examinations The Actuarial Education Company

‘Revision & rehearsal’ products

ActEd Solutions with Exam Technique (ASET)

The Actuarial Education Company © IFE: 2019 Examinations

© IFE: 2019 Examinations The Actuarial Education Company

The Subject SP exams

The Actuarial Education Company © IFE: 2019 Examinations

© IFE: 2019 Examinations The Actuarial Education Company

The Actuarial Education Company © IFE: 2019 Examinations

Revision and exam skills

© IFE: 2019 Examinations The Actuarial Education Company

Exam question skill levels

Questions may be set at any skill level:

It also gives the following advice:

The Actuarial Education Company © IFE: 2019 Examinations

1.5 The examination

What to take to the exam

Remember to take with you:

Past exam papers

© IFE: 2019 Examinations The Actuarial Education Company

1.6 Queries and feedback

Questions and queries

The Actuarial Education Company © IFE: 2019 Examinations

2.1 Subject SP2 – background

Subject SP2 is Life Insurance.

Links to other subjects

© IFE: 2019 Examinations The Actuarial Education Company

2.2 Subject SP2 – Syllabus and Core Reading

On successful completion of this subject, a student will be able to:

1. Life insurance products and general business environment (15%)

The Actuarial Education Company © IFE: 2019 Examinations

Detailed syllabus objectives

1. Life insurance products and general business environment (15%)