Subject3031999 2004

Faculty of Actuaries
Institute of Actuaries
EXAMINATIONS
April 1999
Subject 303 General Insurance
Time allowed: Three hours

INSTRUCTIONS TO THE CANDIDATE
1.
You have 15 minutes at the start of the examination in which to read the
questions. You are strongly encouraged to use this time for reading only but
notes may be made. You then have three hours to complete the paper.
2.
You must not start writing your answers in the booklet until instructed to
do so by the supervisor.
3.
Write your surname in full, the initials of your other names and your
Candidates Number on the front of the answer booklet.
4.
Mark allocations are shown in brackets.
5.
Attempt all 11 questions, beginning your answer to each question on a

separate sheet.
AT THE END OF THE EXAMINATION
Hand in BOTH your answer booklet and this question paper.

In addition to this paper you should have available
Actuarial Tables and an electronic calculator.
303A99
A small general insurance company underwrites only a single class of

business, which is long tailed. Describe briefly the factors it should consider
when deciding upon its investment policy.
[7]
An insurance company intends to offer a product to the motion picture

industry. The product will provide cover should the cost of producing a
finalised, distributable product exceed a pre-agreed budget. List the major
perils that could be included in the cover.
[4]
(i)
Describe the assumptions behind the Bornhuetter-Ferguson method of

estimating the ultimate cost of claims of a book of business.
[2]
(ii)
You have been given the following information about a particular class
of business.
Underwriting
Year
Initial
Expected
Loss ratio
Premium
Reported claims
cost as at
31 December 1998
1995
1996
1997
1998
70%
75%
80%
85%
1,000,000
1,500,000
2,000,000
2,500,000
500,000
500,000
1,000,000
400,000
Expected reported claims development pattern:

Development year
0
Expected proportion developed 10%
1
50%
2
80%
3
90%
Calculate the expected ultimate claims as at 31 December 1998

according to the Bornhuetter-Ferguson method.
(iii)
Comment briefly on the reliability of the results in part (ii).
4
100%
[3]
[3]
[Total 8]
List, and state the purposes of, the major actuarial investigations undertaken
by general insurers in relation to investment.
[5]
(i)
List the reasons why a general insurer models claims.
(ii)
List the steps involved in applying a deterministic modelling method.

[3]
[Total 7]
3032
[4]
Two general insurance companies both have a solvency margin equal to 50% of
written premiums. Explain why the underlying financial strengths of the two
companies may differ. You may ignore the question of the adequacy of the
technical reserves.
[5]
List the main regulatory restrictions that may be encountered by a general

insurance company when transacting insurance business.
[4]
List the reasons why general insurance claims experience may vary from that
assumed in the premium basis.
[6]
(i)
Explain the terms reinstatements and reinstatement premiums in

the context of catastrophe excess of loss reinsurance.
[4]
(ii)
Define the term rate on line.
[1]
A property catastrophe reinsurance layer with unlimited reinstatements

would have a theoretical rate on line, given a detailed actuarial assessment of
the risk, of 5%. A prospective reinsured has asked for two quotations for cover
on this layer one with no reinstatements, and one with a single
reinstatement at 100% additional premium.
10
(iii)
State with reasons which of the two alternative structures you would
expect to require the higher rate on line.
[4]
[Total 9]
(i)
List the factors that a general insurance company writing many

different classes of business would consider in determining appropriate
reinsurance arrangements.
[5]
A general insurance company A writes only commercial property
business. One risk which it coinsures with three other insurers B, C
and D has a sum insured of $10 million, but an expected maximum loss
(EML) of $500,000. Company A accepts 40% of this risk, with B, C and
D accepting 20% each.
Company A reinsures with company X 5% of every risk under a quota
share treaty. It is agreed that A will not write business for which its
gross share of the EML exceeds $250,000.
Company A also has a three line surplus treaty with companies Y and Z,
each taking 50%, which operates after the quota share, and is based on
company X taking 5% of company As gross business. The surplus
treaty has a maximum EML retention of $50,000.
A single large claim gives rise to a loss of $750,000.
3033
PLEASE TURN OVER
11
(ii)
Calculate the amount of the claim which Company A will pay, net of all
reinsurance recoveries due. State any assumptions you make.
[4]
(iii)
Explain how your answer to (ii) would differ if, immediately prior to
this claim, companies B and Y were declared insolvent.
[3]
(iv)
State the information you would expect Company A to provide to

Company Z during the handling of this claim.
[3]
[Total 15]
You are the actuary of a general insurance company with two Private Motor
products, A and B. The monthly Gross Written Premium (GWP) of products A
and B has been 10m split equally between the two products for many years up
to and including December 1998. You have been asked by the Finance
Director to evaluate the adequacy of the Unearned Premium Reserve (UPR) of
60m in Private Motor at 30 June 1999. The UPR has been calculated gross of
Deferred Acquisition Costs (DAC) and on a 24ths basis.
In your testing of the adequacy of the Claim Reserves as at 31 December 1998,
you determined an expected ultimate loss ratio for Accident Year 1997 for
Private Motor of 90%, split 75% for product A and 105% for product B.
In the light of the emerging poor performance of Product B during 1998, the
company imposed two rate increases to Product Bs rates, each of 6%, which
were applied at 1 January and 1 April 1999. This had the effect of reducing
Product Bs share of the monthly GWP to 20% by June 1999. This reduction
occurred evenly over the business written in the first 6 months of 1999. Most
customers switched to Product A, and the total monthly GWP remained at
10M in each month. No other rate increases had been applied for a number of
years.
Claim cost inflation has been 0.5% per month for many years. DAC is 15% of
UPR. Claims handling expenses are 2.5% of the outstanding claim amount.
The present value of investment return at 30th June 1999 is 10% of the
outstanding claim amount. You may ignore reinsurance.
(i)
Define the following terms:

(a)
(b)
(c)
(d)
Unearned Premium Reserve

Deferred Acquisition Cost
Unexpired Risk Reserve
Additional Reserve for Unexpired Risks (AURR)
[4]
(ii)
Calculate the expected ultimate loss ratio for the total Private Motor
business written in July 1999. State clearly any assumptions you make
in your calculation.
[8]
(iii)
Calculate the expected ultimate claim cost of the unexpired risks as at

30 June 1999. State any additional assumptions made.
[10]
3034
(iv)
Determine whether or not an AURR is required in respect of Private

Motor business as at 30 June 1999.
[4]
(v)
Explain why the AURR calculated for Product A alone might differ from
that calculated for Product B alone.
[4]
[Total 30]
3035
EXAMINATIONS
April 1999
EXAMINERS REPORT
Subject 303 (General Insurance) April 1999 Examiners Report

Comments on each specific question appear in italics at the end of each solution.
Main objective is to maximise investment return whilst meeting all its

contractual obligations.
Match characteristics of the assets with those of the liabilities by term,
amount, nature and currency.
Premiums will be received in advance of claims, and these can be invested.
Need to allow for the split between annual, and monthly premium business,
and for the fact that some monies may be held by third parties.
Consider the effect of inflation on the liabilities. Can these be matched?
Matching considerations may run counter to solvency requirements - long
term assets have high volatility.
Marketability of assets - the insurer is small, and claims may fluctuate from
one month to the next.
Consider the free assets - you want to maximise the return, but in deciding on
the investment policy the insurer needs to consider the size in relation to:
annual premium income

expected claims each year
absolute size of the liabilities at any time
SMSM
The effect of the reinsurance arrangements should be taken into account.

The insurer will not want to stray too far from industry norms.
Page 2
Actor,director or producer illness, accident or death

Contract disputes
Weather perils disrupting timetable of filming
Faulty film
Damage to negative during editing
(i)
Whatever claims have developed in relation to an origin year, the

future development pattern will follow that experienced for other
origin years.
The past development for a given origin year does not necessarily
provide a better clue to future claims than the more general loss ratio.
[2]
(ii)
U/W
Year
Initial
Estimated
Ultimate
Claims
% not
reported
yet
Expected
Future
Reported
Claims
Actual
Reported
Claims
Total
Estimated
Ultimate
Claims
1995
1996
1997
1998
700,000
1,125,000
1,600,000
2,125,000
10%
20%
50%
90%
70,000
225,000
800,000
1,912,500
500,000
500,000
1,000,000
400,000
570,000
725,000
1,800,000
2,312,500
The total estimated ultimate claims are 5,407,500

(iii)
Reported claims are lower than original estimate by 142,500

Is the original LR appropriate?
Are the reported reserve estimates reliable?
Is the development table accurate?
Older years are below the initial expected losses, recent years above
has the initial expected loss ratio allowed fully for market
developments?
There is a big jump in loss ratio from 1996 to 1997. Is there any
particular reason for this?
4
Investment portfolio evaluation, to assess the performance of investment
managers
Cash flow / asset-liability modelling to set investment policy
Allocation of income and capital between classes for pricing and
profitability measurement
Risk-based capital allocation to enable solvency evaluation or profitability
Calculation of return on capital to allow shareholders to assess
management performance.
(i)
To arrive at a risk premium per policy.

To select rating factors
To determine premiums using experience rating procedures
To demonstrate the estimated effect of changing the level of cover
To determine the effect of excess of loss reinsurance
To estimate the variability of claims experience
To estimate the impact on the reserves of industrial diseases
Financial planning
Workload management/staff planning
Page 3

Statutory requirement to do so
(ii)
Specify the purpose of the investigation

Collect data
Group and modify data
Compare data with a suitable density function
Fit the data to the density function
Check that goodness of fit is acceptable
Fit different model if not
Valuation of the Assets

The asset valuation bases, and
method used for recognising capital gains will affect the stated surplus
Investments held
Weak security of investments undermines high asset value
Undue concentrations of investments do likewise.
Mix of business
Differences between the mixes of business such as short tail or long
tail will distort comparison.
Long tail classes require a larger solvency margin for same relative
financial strength
Unusual events
Catastrophes a company with little catastrophe protection needs
more capital
High expenses from internal reorganisations
Third Party Exposure
Financial security of reinsurer affects quality of reinsurance levels
Or any other debtor, especially brokers, affects quality of assets
Presence of equalisation reserves, or other hidden types of capital.
Differences in reinsurance arrangements.
Restrictions on the type of business an insurer may write

Limits on the premium rates that may be charged.
A requirement to maintain a minimum level of solvency,
measured in some prescribed manner.
Restrictions on the type
or amount of assets that may be taken into account to demonstrate
solvency
A requirement to use prescribed bases for valuing assets/liabilities
Legislation to protect policyholders should a general insurer fail
Licensing of agents to sell insurance
Requirements to provide information
Control of sales outlets
Approval of staff and directors
Restrictions on volumes written
Page 4
Lack of previous experience

Variability of experience
Changing types of business
Changing risk characteristics
Antiselection by policyholders
Changing attitudes to claiming
Climate effects
Catastrophes
Exchange rate movements
Latent claims
New types of claim
Claims inflation
Legislative changes
Judicial changes
Poor management
Poor underwriting
(i)
A catastrophe excess of loss policy normally allows the reinsured to

make only one claim on the policy. Most policies allow the cover to be
restored after a claim, so that the reinsured still has protection. This is
called reinstatement. Normally there will be a limit on the number of
times the cover may be reinstated.
A reinstatement may be free, or a premium may be payable. If there is
a premium, it will normally be a proportion of the original premium. It
will be payable whether or not the whole of the cover has been
exhausted, and will be proportional to the amount of the claim to the
layer.
10
(ii)
Original premium width of layer. (Does not include reinstatement

premium in calculation.)
(iii)
The option with no reinstatements would have the higher rate on line.
The low rate on line of the underlying risk suggests that the probability
of two losses is very low. The expected value of a reinstatement
premium is greater than the expected value of a second loss, so the
expected net payments by the reinsured are lower under the second
option.
(i)
Class of business
Size of individual risks
Likely accumulation of risks
Volatility of claims experience - numbers and amounts
Size of class relative to total written premiums
Size of free reserves
Premiums written relative to size of free reserves
Availability of reinsurance
Availability of coinsurance
Page 5

Need for technical assistance
Management attitude to risk
Security status of reinsurers
(ii)
(iii)
Page 6
(a)
Company A is directly responsible for 40% of $750,000 = $300,000

under the coinsurance agreement.
Company X takes 5% of $300,000 = $15,000 under the quota share
agreement.
There are two possible approaches (or indeed anything between
the two). However, candidates need to recognise that there is a
need for an assumption, to state the assumption and then to
follow it through in the calculation.
Assumption 1 Company A takes the full three lines of cover,
ceding 75% to the surplus reinsurers. Therefore, its net claim is
25% ($300,000 $15,000) = $71,250.
Assumption 2 Company A keeps the maximum $50,000
retention, so that its proportionate retention is $50,000 (95%
40% $500,000) = 1 3.8. Therefore its net claim is ($300,000
$15,000) 3.8 = $75,000.
(b)
As company B is a coinsurer alongside A the insolvency does not

affect the amount to be paid by A.
As company Y is a reinsurer of A the insolvency will affect the
recoveries A can make. In the most extreme case, A will lose the
amount it is owed by Y; in practice, there will almost certainly be
a partial recovery from the liquidators of Y. The actual answer
will lie between the answer in (a) and a revised amount,
assuming that there is no recovery at all from Y. The answer in
that case depends on the assumption made in (a) above.
Assumption 1 In this case A will have to pay an extra 75% 50%
($300,000 $15,000) = $106,875, so the total will be $178,125.
Assumption 2 In this case, A will have to pay an extra 50%
($300,000 $15,000 $75,000) = $105,000, for a total of $180,000.
Details of loss:
Description of property, processes and materials used
Date and time of loss
Estimated cost
Cause of loss
Explanation why loss amount exceeded the EML
11
(i)
(ii)
(a)
UPR
The amount set aside from premiums
written before the accounting date
to cover risks incurred after that date
(b)
DAC
A deduction from unearned premium
as they become earned
for acquisition
and commission costs
(c)
URR
The reserve required to cover the claims
and expenses
which are expected to emerge from an unexpired period of cover
(d)
AURR
The reserve held in excess of the UPR to allow for any
expectation
that the UPR will be insufficient to cover the costs of
outstanding risks
Assumptions
Claims in accident year 1998 occurred on average at 1 July 1998
The business resulting in AY 1998 claim was written on average on 1
January 1998
Period of cover is twelve months
June 1999 business was written on average in the middle of the month
Claim cost inflation period for June 1999 business is 17 months
Claim frequency inflation is 0%
The mix/riskiness of business within products A and B is unchanged
Average written premium (AWP) of Product A is unchanged
AWP of product B is increased by 1.062 1 = 12.36%
Loss ratio of Product A at June 1999 is 75 1.00517.5 = 81.84%
Loss ratio of Product B is 105 1. 00517.5 1.062 = 101.97%
The GWP at June 1999 is 8 million for Product A and 2 million for
Product B
Combined loss ratio = 81.84% 0.8 + 101.97% 0.2 = 85.90%
Page 7

(iii)
Assumptions
Risk spread evenly throughout the year
Unexpired Loss ratio %

Cover Jul
for Product A
98 to Jun 99
in 24ths
Loss ratio %
for Product B
Jul 98 1
Aug 98 3
Sep 98 5
Oct 98 7
Nov 98 9
Dec 98 11
Jan 99 13
Feb 99 15
Mar 99 17
Apr 99 19
May 99 21
Jun 99 23
105 1.005^
105 1.005^
105 1.005^
105 1.005^
105 1.005^
105 1.005^
105 1.06 1.005^
105 1.06 1.005^
105 1.06 1.005^
105 1.062 1.005^
105 1.062 1.005^
105 1.062 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
75 1.005^
6.5
7.5
8.5
9.5
10.5
11.5
12.5
13.5
14.5
15.5
16.5
17.5
6.5
7.5
8.5
9.5
10.5
11.5
12.5
13.5
14.5
15.5
16.5
17.5
GWP
A
M
GWP
B
M
Ultimate
Claim
Cost
M
5.0
5.0
5.0
5.0
5.0
5.0
5.5
6.0
6.5
7.0
7.5
8.0
5.0
5.0
5.0
5.0
5.0
5.0
4.5
4.0
3.5
3.0
2.5
2.0
Total
0.387
1.168
1.956
2.752
3.556
4.369
4.948
5.657
6.352
6.888
7.564
8.229
53.827
Marks are given for:

Correct formula for one month (say July 1998)
= [75% 5.0 + 105% 5.0] 1.0056.5 1/24 = 0.387
Combined GWP Loss ratio for A + B
Claims cost inflation
Unexpired cover
Correct answer for July 1998
Correct answer for August 1998 to December 1998
Correct adjustment for January 1999 GWP
and loss ratio
Correct answer for January 1999
= 4.948
Correct answer for February 1999 to March 1999
Correct answer for April 1999 loss ratio
Correct answer for April 1999
= 6.888
Correct answer for May 1999 to June 1999
Correct overall unexpired claim cost = 53.827
(iv)
UPR
less DAC of 15% 60.0
equals
= 60.0
= 9.0
51.0
Unexpired claim cost

53.8
less investment return of 10% 53.8
= 5.4
plus claim handling costs of 2% of 53.8 = 1.3
equals
49.8
Surplus of 1.2
Therefore no need for an AURR
Page 8

(v)
Difference in expected ultimate loss ratios

Difference in GWP distribution throughout the twelve months ending
30 June 1999
The acquisition costs might be different due to:
different distribution channels, for example
Product A may be sold direct, Product B through brokers
Commission rates may therefore be different
Value of investment return depends upon claim payments
Product A may have a greater Comprehensive mix (more short tail)
than Product B
Page 9
EXAMINATIONS
September 1999

1.
2.
3.
4.
5.

separate sheet.

303S99
A general insurance company is constructing an asset-liability model to

determine the likely future relationship between its assets and liabilities.
(i)
State the particular factors that will need to be incorporated into the
model if it is to be used to assess future solvency.
[4]
(ii)
List, with brief reasons for their interest, the parties who may be
interested in the level of the solvency of the company.
(iii)
Discuss briefly whether it would be more effective to use a

deterministic or a stochastic model.
[3]
[3]
[Total 10]
State the reasons why a general insurer would analyse its claims data.
You are the actuary to an insurance company that writes industrial property
business with large deductibles. It is subject to large losses, which are usually
reported quickly. Reserves are always set up in respect of these losses, but
these are often found to be excessive or inadequate. You are concerned about
the estimation of ultimate claims from this line of business, which has been
carried out using the chain ladder method and has often been unstable in the
past. It has been suggested that the use of Bornhuetter-Ferguson methods
would provide an element of stability in reserve estimates.
Discuss this suggestion.
[5]
[5]
List the guidance notes published by the Institute and Faculty of Actuaries
which relate to general insurance business, and state what each relates to. [6]
List the portfolio movements that are normally monitored by a general

insurance company, and state the reasons for monitoring them.
3032
[5]
Company A is a medium-sized general insurance company, writing a variety of

property and liability risks. It expects a loss ratio of about 75% on its
commercial property portfolio, which accounts for about 60% of the companys
total premium. At the end of the companys past year it had capital of 28% of
the premium written during the year. It has recently taken out a reinsurance
treaty, from which the following information has been extracted.
Name of reinsured:
Company A
Subject business:
All business written by the reinsured in the class

commercial property with an inception date during
1999.
Reinsurance premium: 41% of original premiums of the subject business, to be

paid on 31 December 1999.
Commission:
A reinsurance commission of 9% of original premiums

of subject business shall be paid to the reinsured. This
is part of the 41% reinsurance premium.
Limits:
Losses shall be paid to the reinsured in excess of a

total loss equal to 70% of original premiums of subject
business, up to a maximum total loss of 111% of
original premiums of subject business.
Termination:
This treaty shall be terminated with effect 31

December 2003, and all monies due as claims or profit
share shall be paid to the reinsured immediately.
Profit share:
On termination of the contract a profit share shall be

payable to the reinsured. This shall be calculated as
97% of the profit. The profit shall be calculated as
41% of the original premiums of the subject business
less any claim payments made to the reinsured under
this contract.
Discuss this reinsurance treaty. You should consider the nature of the
contract, the nature and amount of the protection and any other benefits it
provides, the reasons for taking it out, and how it might meet those criteria.
[10]
3033
PLEASE TURN OVER
Give examples of errors which might be made in entering data into a computer
claims system, and suggest automatic proceedings which would reduce the
probability of them occurring.
[5]
You are the pricing actuary for a reinsurance company. You are being asked to
quote for a quota share reinsurance of the cedants book of liability business.
The inception date of the policy is to be 1 January 2000. The broker has
supplied you with the following information.
Between 1990 and 1997 premium increases kept pace with increases in
losses and the underwriting loss ratio was fairly stable at around 85%.
Since 1997 the market has softened. Premium rates fell by 5% per annum in
1998 and 1999 while losses increased at 10% per annum over the same
period. The company has cut back on its new writings as a result.
Written premium in 1989 was 500.
The following statistics have also been provided, and you have made the
calculations shown.
Underwriting year:
Written premium:
3034
1990
500
1991
600
1992
700
1993
800
1994
900
1995
1,000
1996
1,100
1997
1,200
1998
600
1999
300
Paid losses
Data evaluated as at 31 December each year
Last diagonal as at 30 June 1999
Accident
Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
0
26
28
40
18
25
32
31
27
51
10
1
48
43
66
68
80
63
71
104
54
2
65
76
72
88
110
137
152
111
3
89
93
104
139
150
168
157
4
208
229
279
313
368
291
5
257
280
336
382
395
6
298
331
384
412
7
319
353
400
8
339
362
9
351
Year-on-year development
Accident
Year
01
1990
1991
1992
1993
1994
1995
1996
1997
1998
1.85
1.54
1.65
3.78
3.20
1.97
2.29
3.85
1.06
12
23
34
45
56
67
78
89
1.35
1.77
1.09
1.29
1.38
2.17
2.14
1.07
1.37
1.22
1.44
1.58
1.36
1.23
1.03
2.34
2.46
2.68
2.25
2.45
1.73
1.24
1.22
1.20
1.22
1.07
1.16
1.18
1.14
1.08
1.07
1.07
1.04
1.06
1.03
1.04
1.04
Average year-on-year development factors

All years:
2.15
1.49
1.29
2.27
1.18
1.14
1.06
1.04
All years excluding last diagonal:

2.39
1.59
1.36
2.43
1.22
1.16
1.07
1.06
1.68
1.19
2.13
1.16
1.13
1.06
1.04
1.22
1.16
1.07
1.06
Last three years:

2.10
1.04
Last three years excluding last diagonal:

2.64
3035
1.86
1.36
2.44
PLEASE TURN OVER
Incurred losses
Data evaluated as at 31 December each year
Last diagonal as at 30 June 1999
Accident
Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
0
123
135
169
201
220
242
267
295
372
202
Accident
Year
01
12
1990
1991
1992
1993
1994
1995
1996
1997
1998
4.15
4.16
3.91
3.80
3.95
4.01
4.02
3.98
2.50
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.87
511
561
661
763
869
970
1,073
1,174
930
3
404
445
524
606
685
766
824
4
416
458
542
625
708
779
5
421
463
547
631
712
6
425
468
553
634
7
425
468
553
8
425
468
23
34
45
56
67
78
89
1.05
1.06
1.06
1.06
1.05
1.06
1.02
1.03
1.03
1.03
1.03
1.03
1.02
1.01
1.01
1.01
1.01
1.01
1.01
1.01
1.01
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
384
420
496
572
651
726
805
1,026
9
425
Average year-on-year development factors

All years:
3.71
0.77
1.05
1.03
1.01
1.01
1.00
1.00
All years excluding last diagonal:

3.98
0.75
1.06
1.03
1.01
1.01
1.00
1.00
0.79
1.04
1.03
1.01
1.01
1.00
1.00
1.03
1.01
1.01
1.00
1.00
Last three years:

3.40
1.00
Last three years excluding last diagonal:

4.00
0.75
1.06
(i)
Describe in detail how you would use the Paid and Incurred
Bornheutter-Ferguson methods to estimate ultimate losses for this
business for each accident year. You should explain all choices and
assumptions you would use. You are not required actually to calculate
the estimated ultimate losses
[16]
(ii)
Estimate the ratio of discounted ultimate claims to undiscounted

ultimate claims as at the start of an accident year. Use a 10% p.a.
interest rate.
3036
[5]
(iii)
Assuming brokerage of 2%, commission of 5%, premiums paid evenly

throughout the year and a required profit load of 10% of undiscounted
premiums calculate the required ceding commission for this treaty.
State any assumptions made.
[5]
[Total 26]
In a particular country, all insurance business is accounted for on a one-year

basis, but may be accounted for on an accident year basis or an underwriting
year basis.
Company A and Company B operate in this country. Company A uses an
accident year basis and company B uses an underwriting year basis. Their
balance sheets as at 31 December 1998 are given below. Both companies have
had stable levels of premium income for a number of years.
Company A balance sheet as at 31 December 1998
Cash on hand
Other investments
Total assets
3,000
11,000
14,000
Unearned premium
less DACs
4,000
800
3,200
Claim reserves
6,000
Share capital
500
Share premium account 1,500
Retained profits
2,800
Total liabilities
14,000
Company B balance sheet as at 31 December 1998
(i)
Investments
14,000
Claim reserves
Share capital
Retained profits
12,000
500
1,500
Total assets
14,000
Total liabilities
14,000
Discuss how these two bases would differ in the context of one-year
accounting, and the consequences for the recognition of profits.
[5]
Company A has agreed to acquire all the shares of Company B as at 31

December 1998 for 2,500. At that date, Company Bs balance sheet will be
restated in accordance with Company As, and the two companies will be
merged. In accordance with the laws of this country, any goodwill arising from
the transaction will be written off immediately. Company As own loss ratio in
underwriting year 1998 is expected to be 60%. In the course of Company As
due diligence exercise, the following facts have emerged about Company B.
The premium written in 1998 was 3,200.
3037
PLEASE TURN OVER
No premium will be received in respect of the 1998 underwriting year after

the end of the year.
The expected loss ratio on underwriting year 1998 was 88%. This is the
conclusion of the due diligence exercise, rather than Company Bs
managements opinion.
Average commission on underwriting year 1998 was 25%.
Business is renewed predominantly on 1 January and 1 April, to the extent
that the average inception date of all policies was 1 April. (Which may be
considered to be exactly 25% through the year.)
All claim reserves need to be increased by 10% to conform to the standards
of Company A.
(ii)
Prepare the balance sheet of Company B at 31 December 1998 on an

accident year basis to Company As reserving standards.
[5]
(iii)
Prepare the balance sheet of Company A at 31 December 1998 after the

takeover has been completed.
[4]
(iv)
Comment on the differences between Company A and Company B that

may be inferred from the balance sheets and other information, and the
effect of the takeover on Company As financial position. You may
assume that all assets yield income of 7% per year.
[14]
[Total 28]
3038
EXAMINATIONS
September 1999
EXAMINERS REPORT
Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The examiners are mindful that a number of interpretations may
be drawn from the syllabus and Core Reading. The questions and comments are based
around Core Reading as the interpretation of the syllabus to which the examiners are
working. They have however given credit for any alternative approach or interpretation
which they consider to be reasonable.
The report does not attempt to offer a specimen solution for each question - that is, a
solution that a well prepared candidate might have produced in the time allowed. For
most questions substantially more detail is given than would normally be necessary to
obtain a clear pass. There can also be valid alternatives which would gain equal marks.
D S Brand
Chairman of the Board of Examiners
30 November 1999
Subject 303 (General Insurance) September 1999 Solutions

Overall, this was a difficult exam, and it was to be expected that candidates would find
difficulty in some areas. However, the examiners were concerned in a number of areas
where very few candidates appeared to have any understanding of the concepts being
examined.
(i)
The main problem with the answers to this question was that many
candidates failed to give any consideration to the question of models
intended to demonstrate solvency, rather than modelling in general.
Allow for any statutory requirements relating to the valuation of assets
and liabilities.
A comparison may then be made of the actual level of solvency with that
acceptable to the statutory authorities,
And any higher minima, such as those really acceptable to supervisors or
to rating agencies.
Assumptions:
Likelihood of catastrophes and accumulations.
Vulnerability of capital to major shocks such as these.
Effect of actual non-proportional reinsurance coverages on catastrophes.
Spread of different risk groups within the portfolio.
Insurers experience of writing different classes of business.
Expected level of profitability, both underwriting profit and level of
investment income generated (model tail length of classes).
(ii)
This section was generally well answered.

Statutory authorities to protect policyholders.
Company management duty to company / shareholders / policyholders.
Rating agencies provide accurate information.
Brokers to place business with solvent companies.
Shareholders to monitor their investments.
Policyholders to check the solvency of their insurers.
Competitors to try to discover competitively useful facts.
Potential purchasers of the company to assess its value.
(iii)
This section was generally well answered.

Might test variability using variations in assumptions with a
deterministic model.
Easier to test effects of particular scenarios using a deterministic model.
Stochastic more complex, and harder to get right.
Stochastic harder to parameterise.
Stochastic more effective in incorporating allowance for volatility is asset
values
And uncertainty in claims experience.
Stochastic may take into account correlations between and within assets
and liabilities.
Stochastic allows construction of probability distribution for outcomes.
Page 2
Most candidates knew the main reasons and scored reasonably well.
Estimating the cost of outstanding claims
to set reserves
Monitoring the actual runoff of outstanding claims
against estimated amounts
Monitoring the adequacy of reinsurance
Comparing the relative profitability of various parts of the account
Reviewing current premium rates
Pricing new products
Financial planning
Monitoring the asset-liability position
Almost no candidates considered the paid and incurred approaches separately,

despite the difference in the effect likely. Almost none realised that when the
claims incurred average 100% of ultimate in an early report then the weight
attached to the initial expected losses is minimal, and that therefore the incurred
Bornhuetter-Ferguson method has almost no effect on the incurred chain ladder
estimate, or that it is inappropriate to use the Bornhuetter-Ferguson approach
when a reduction in incurred claims is anticipated. Some candidates suggested
giving a credibility-inspired weight to the initial expected losses, which is a
sensible variation of the Burnhuetter-Ferguson. Some suggested that when claims
are reported quickly, as is normal in this class of business, it would be better to
reply on case estimates. This is a sensible suggestion, but the problem in this case
is that case estimates have proved unreliable, and to use this approach we would
need to find some way to improve the quality of case estimates.
It is not stated whether the paid or incurred Bornhuetter-Ferguson (BF) method
is intended, and whether it is likely to replace a paid or incurred chain ladder
method. If paid methods are intended, then the BF is likely to be an
improvement on the chain ladder, provided a sensible initial expected loss
estimate is available, as its development will be more stable, and less affected by
early swings in paid amounts. However, these methods ignore the amount of
reserves set up in respect of reported claims. This is normally thought to be a
valuable source of information, even though in this case it is noted that the
reserves are often badly inaccurate.
If it is intended that incurred methods be used, the BF suffers from a very
serious weakness. This is that if reductions in incurred claims are more normal
than increases, the formula does not make sense, predicting a reduction which is
a function of initial expected losses rather than the reported claims which are
subject to reduction. Also, even if there are only modest further increases in
incurred claims to be expected, then the initial expected losses are given little
influence in the projected result, and therefore little opportunity to exercise a
stabilising influence. From the information available, it appears likely that one
of these will be the case.
It may be found useful to consider the results of Bornhuetter-Ferguson methods
in addition to the current methods the company uses.
Page 3
This question was surprisingly badly answered, as it is straight bookwork. This

meant that while a reasonable number of candidates obtained full marks, other
gained little or no credit. Some candidates evidently felt that it was outside the
scope of the examination, as the specific application of the guidance notes is the
United Kingdom, but the answer is a direct quotation from the core reading. All
members of the Institute and Faculty should be aware of the nature of professional
guidance, which can be obtained only by familiarity with the guidance notes. Of
more importance, since the last amendment to the syllabus and core reading,
GN14 has been withdrawn and GN12 amended significantly. Under the normal
examination rules, an answer would have been accepted as fully correct if it
corresponded either to the syllabus/core reading or to the correct situation at the
time of the examination.
GN12:
An actuarial report on the reserves or financial soundness of an

insurance company or Lloyds syndicate.
GN14: Providing the actuarial reports on Lloyds run-off years of account
required under Lloyds byelaws.
GN18: Providing a Certificate of loss reserves to the International Insurers
Department in respect of a UK insurance company authorised in the
United States of America.
GN20: Providing a statement of actuarial opinion to the Council of Lloyds in
respect of solvency reserves under the Lloyds Valuation of Liabilities
Rules.
GN 32: The work of the Appropriate Actuary of a Friendly Society.
GN 33: Providing statements of actuarial opinion to the International Insurers
Department and the New York Insurance Department in respect of the
relevant US insurance or reinsurance business of a Lloyds syndicate.
This question was generally well answered.

The movements which are normally monitored are:
New business
Lapses at renewal
Endorsements
Mid-term cancellations
By studying movements and their trends, an insurer can:
Measure the growth or contraction of the business and of its different parts.
Get an early indication of unusual losses or gains in business volume. These
may indicate if the insurers rates or conditions are out of line with the
market.
Assess the effects of a new set of rates or a marketing campaign on the
business, and thus the sensitivity of the portfolio to market forces.
Movements will give an early indication of adverse or favourable changes. These
may indicate the need for a review of premiums or terms and conditions. They
may also be used to help reconcile accounts data and in-force statistics.
Page 4
This question was generally answered very badly. Few candidates realised that
the policy in question was financial reinsurance, and most of those who did failed
to analyse the policy on its merits as a financial reinsurance considering the
insurers probable motives in effecting such cover. Many candidates analysed the
companys gross and net position, rather than looking at the effect on the company
of the reinsurance, which would have revealed the lack of risk transfer. Of
candidates who did identify the lack of risk transfer, many identified it as a flaw
in the policy design, rather than a fundamental part of it. Past examinations have
suggested that candidates are familiar with the concept of financial reinsurance,
but the response to this question suggests that they lack knowledge on how such
arrangements might be constructed. A good number of candidates did identify
correctly that the policy would have a beneficial effect on the companys solvency
purely in terms of the solvency margin:net premium ratio. Unfortunately, if
candidates failed to recognise the nature of the policy, it was not generally possible
to award significant marks. The examiners gave favourable consideration to
candidates who had analysed the reinsurance as if it was a conventional stop loss
product, and had got very low marks, but whose analysis of it as a stop loss
product was sensible and detailed. However, such analysis would generally
suggest that there were a number of major inconsistencies in the product, which
many candidates who took this approach spotted. Some candidates thought that
this was a proportional reinsurance, apparently on the basis that the reinsurance
premium is expressed as a proportion of the subject premium. It should be noted
that almost all reinsurance premiums, whether of proportional reinsurance or not,
are expressed in this way. The worst candidates simply listed the conventional
reasons for effecting reinsurance, or examined the policy in the light of what the
companys overall reinsurance strategy ought to be. Some candidates understood
the commission clause to indicate that 41% premium was net of the 9%
commission. This had not been intended, and would change the financial nature
of the policy profoundly, however, its purpose was understood.
The policy transfers almost no risk. The company expects a loss ratio of 75%.
Therefore, it is unlikely that a paid loss ratio of 70% will arise by 31 December
2002, so no money is likely to be paid to the reinsured before 31 December 2003,
when the treaty terminates. On the other hand, a profit share 97% of any
claim payments which might be made means that the total amount then paid to
Company A will vary very little as a result of the loss experience of the reinsured
business. This type of reinsurance is known as financial reinsurance, or finite
risk reinsurance.
The total amount that Company A receives back is likely to be very close to 40%
of the underlying premium, whatever the loss ratio may be.
Since it provides almost no protection against loss, Company A must have had
some other motive in taking it out. The policys nature is close to what is known
as a time and distance policy, a form of financial reinsurance, or finite risk
reinsurance. It may be viewed as placing money on deposit, or as a method of
improving the Companys apparent solvency position. As the companys solvency
ratio is only 28% of premiums, which would normally be considered low, the
latter is probably the companys motive.
The yield on the funds placed is (40/32) 1 = 5.7% per year. This may or may
not be a reasonable yield, depending on the current yields for a four year deposit
Page 5

in the currency in which Company As business is written. The tax treatment of
a transaction such as this in Company As domicile, and the security rating of the
reinsurer, are both matters that will affect whether or not it may be considered a
good and appropriate investment of the companys funds.
In taking out this reinsurance the company loses assets equivalent to 32% of the
premium for the class. However, it will be able to reduce reserves by 40% of the
premium in anticipation of reserve recoveries. This will release 8% of this classs
premium, and improve the solvency ratio. Since premiums have been ceded by
this transaction, measuring the ratio on net premium will give a greater rise.
However, it does not change the companys underlying position at all. If the
company had had problems with its capital adequacy this transaction will not
relieve them in the way that a quota share cession, for example, would have
done. Because it is artificial, some jurisdictions prohibit the treatment of this
type of policy in this way, but presumably this is not the case in Company A's
jurisdiction, or it would not have taken out the treaty.
It is possible that Company A might be permitted to use discounted reserves in
its balance sheet. This might achieve the same effect without the aid of artificial
transactions.
This question was reasonably well answered. Few candidates noted the point
about currency, which may be a major concern to an insurer operating in a
number of countries, or in the London Market, or a marine, aviation or travel
insurer, to name just three examples. A number of candidates were concerned with
dates of birth, rather than dates in general, which would not be a matter of
concern to most insurers.
Information could be entered onto the wrong claim record.
The claim and policy numbers should both be in series that mistakes are difficult
to make a single digit wrong or two numbers swapped will give an invalid
number, and it should not be possible for a policy number also to be a claim
number.
Incorrect amounts may be entered, or correct amounts in the wrong

currency.
There should be some check on amounts. Very large or small claims should be
queried if entered. This is especially important if working in a variety of
currencies. A query should be raised if an amount is entered in a different
currency from previous entries. A query should be raised if the claim is not in
the currency of the country of the address of the policyholder (this will not apply
for marine insurance, travel insurance and some other classes).
A claim may be entered for an incorrect date.
The system should automatically check that the policy was on-risk on the day
when the claim occurred. A query should be raised if there is a very long gap
between the date of loss (or reporting to the insured) and reporting to the
insurer, or if the date of loss was later than the date reported.
Page 6
Information may be entered against the wrong policy.
Other details, such as policyholder surname, deductible, and the fact that paid +
estimated outstanding < sum insured, should be checked against the information
on the policy record.
Information may be missed out.
A claim should not be accepted until all fields have been filled in, possibly with
null entries.
This question was answered moderately. The most important common fault was
not considering whether or not tail factors were required. A number of candidates
noted that the paid claims were not fully developed, but concluded that the method
could not be used, or used it without incorporating a tail factor. One strange
misapprehension of many candidates appeared to be that paid claims should
always be associated with written premium and incurred claims with earned
premium. This entirely depends on whether accident year or underwriting year
claim cohorts have been used to compile the paid and incurred claims, and then
both sets of data should be compared with the same premium.
(i)
The first decision is which set of development factors to use to project the
losses. The last diagonal of data represents six months development,
rather than a years, so it is not appropriate to use the sets of factors
which include the last diagonal. Therefore one of those calculated
excluding the last diagonal should be used. Some candidates noted that
the last diagonal was significantly different from the previous development
factors, but failed to connect it with the half years experience. Some
candidates suggested adjusting the last diagonal of data to incorporate
this, some noting that it was a pity to have to throw away data. This is a
reasonable alternative approach, but, especially at early durations, the
precise amount to adjust by is problematic
Neither triangle of development factors appears to suggest that the
development pattern is changing over time. The paid one is erratic, but
there is no evidence of systematic increase or decrease. If there were, it
would be appropriate to base our chosen factors on recent diagonals only,
or to extrapolate the trend into the future. As it is, it seems best to use
the average development factor for all years excluding the last diagonal.
In practice we might smooth the observed average factors. A significant
number of candidates thought that the paid triangle was extremely volatile,
some suggesting that it was so volatile that it was necessary to discard it.
In fact the triangle would not be considered to be particularly volatile in
practice.
The question of whether or not a tail factor is needed ought to be
addressed. With incurred claims the answer is almost certainly not. All
nine observations of claims after development year six show nil
development, so development may be assumed to be complete at that
stage. With paid claims, development continues in the tail, and may be
assumed to continue beyond the development horizon for which we have
Page 7

data. We would have to extrapolate the trend we observe in the tail. In
practice we would do so, but also refer to benchmarks, discussions with
the broker, the direct writer and any other available source of information
in forming a judgement. A possible value would be (continuing on from
the observed value of 1.06): 1.04 1.03 1.025 1.02 1.015 1.01
1.007 1.005 1.003 1.002 1.001 1.001 = 1.17. Most candidates
made sensible comments on which set of averages to choose, and it was
quite acceptable to prefer the last three years if sensible arguments were
stated. Few candidates mentioned the alternative sources of information.
The use of accident year implies the use of earned premium to estimate
initial expected losses. We should estimate this value using the average
of written premium in the accident year and in the prior year. However,
we should check with the broker whether or not this is appropriate the
class of business may have a particular renewal season.
A question arises as to whether or not it is appropriate to use the B-F
method to project incurred losses from development year 1, since these
generally fall substantially in development year 2. The reduction cannot
be a result of new claims being reported (as this would require them to be
negative claims), but must be the taking down of reserves, which have
already been reported. Therefore, it is inappropriate to project the
reduction on the basis of initial expected losses. It may also be
inappropriate to use the method without amendment after development
year 0, for the same reason.
As part of the procedure, proportionate future development of claims must
be estimated. This is done for each development year by taking the
selected development factors for later development years, and the tail
factor if any, and multiplying them together. However, in this context
they must be adjusted to mid-development year values. It may be enough
to do this by simple interpolation, but especially in the early stages, it
may be that development is slowing down and the second half of the year
should have less of the factor than the first. It may also be the case that
cedants tend to report losses shortly before the end of the year, in which
case the reverse may be the case. Particular care needs to be taken over
the value used for the 1999 accident year, which at the moment is at
development year . There are standard interpolation methods that
may be used, and the underlying situation should be discussed with the
cedant or the broker. Most candidates demonstrated by the construction of
future development factors appropriately, but few pointed out the need for
mid-year interpolation, or the difficulties surrounding its estimation.
Page 8

A similar problem exists in relation to the 12 factor on the incurred,
which is a substantial reduction. It may be that this actually happens
over a longer period or a shorter one for example a reserve review at
the end of each year of claims of a certain age, which tends to produce
reductions. The actual underlying half-year development factors could be
similar to either of these two examples:
or
0
5
2.5
1
0.79
1.59
11
.85
1.25
12
.88
.60
22
.9
1.04
23
1.18
1.02
Whichever is chosen will have a profound effect on the results and should
be based on a knowledge of the underlying situation.
The process of calculation is as follows:
1.
2.
3.
4.
5.
6.
7.
(ii)
Estimate earned premium for each accident year.

Calculate initial expected losses for each accident year, as the
product of the result in (1) and an initial expected loss ratio based on
the brokers pricing information.
Take the chosen development factors, and estimate development
factors to ultimate from each development year, as the product of
later year-on-year factors, and any tail factor.
Interpolate the results in (3) in an appropriate way to get mid-year
factors.
Estimate future development percentages as the complement of the
reciprocal of the results in (4).
Estimate future development as the product of the results in (2) and
(5).
Estimate ultimate losses as the sum of the results in (6) and the
reported or paid claims.
The worst common mistakes were to use the incurred claims pattern and to
use cumulative, rather than incremental, amounts. Some candidates used
a particular year to do the calculations. This was acceptable.
-
Use paid development pattern

Using the uninterpolated pattern, then assuming payment midway
through the period we should discount by year, 1 years, 2
years etc.
Must spread out paid tail or assume tail is paid on average two to
three years after payment period before the tail. (Answer shows 3
years.)
01
12
23
34
45
56
67
78
89
9Ult
Incremental uninterpolated pattern:

4.2%
5.9%
6.0%
5.8%
31.3%
11.7%
10.4%
5.3%
4.8%
14.5%
01
12
23
34
45
56
67
78
89
910
Page 9

Discount:
0.953
0.867
0.788
0.716
0.651
0.592
0.538
0.489
0.445
0.350
Multiply and add back up: 0.608

(iii)
Most candidates failed to attempt this question.

-
(i)
Assume brokerage and commission are accrued at the same rate as

premium.
Assume premium paid on average after six months, that is a
discount factor of 1.1 = 0.953.
Ignore internal expenses (assume they are included in profit
loadings).
Equate present value of income and present value of outgo.
Premium 0.953 = Losses 0.608 (or whatever was calculated in
part (ii) ) + profit + commission 0.953 + brokerage 0.953 + ceding
commission 0.953.
Assuming premium adequacy in 2000 is the same as in 1999 we
expect accident year 2000 loss ratio to be 114%.
Required ceding commission is 100% 114% 0.608 / .953 10% /
.953 2.5% 5% = 9.28%.
This question required candidates to think beyond the circumstances

outlined in core reading, which treat underwriting year accounting purely
in terms of three-year accounting. It should have helped that there were
example balance sheets later in the question. Many candidates did
understand the central point of the question, which was the treatment of
the exposure still to take place. Some students were confused by the
concepts of funded accounting and closed and open year in a context of
one-year accounting each year would be closed at 31 December, so that
there would be no open years. Others mentioned reinsurance to close,
which is only ever relevant in the context of Lloyds syndicates.
The underwriting year basis anticipates all losses from business already
written, whether or not the exposure period has expired by the accounting
date. Therefore, there would be a need for a reserve in respect of any
losses that will occur after the balance date on any unexpired cover at
that date, which would not be needed under accident year accounting.
Conversely, there will be no need for any reserve in respect of unearned
premium, which does need to be set up under accident year accounting,
not would there be any credit in respect of deferred acquisition costs.
The consequence of using the underwriting year basis rather than the
accident year basis is that any profit or loss which is expected on the
unexpired period of cover will be anticipated in the accounts, with possible
follow-on effects on the timing of tax and dividend payments. Under the
accident year basis, the recognition of profit is deferred until the cover
period has expired; the recognition of losses may be similarly deferred,
unless an additional amount for unexpired risks is set up as a reserve, in
which case the timing of the losses will be the same.
Page 10

(ii)
Most candidates correctly calculated the unearned premium and DACs.

Some candidates simply increased claims reserves by 10%, without
removing the unexpired portion, and double counts reserves for the
unexpired period of cover. Some candidates changed the reserve amounts,
but instead of taking the difference. It would be nice to be able to do this in
real life. Some candidates tried to reckon whether or not an additional
amount for unexpired risk would be appropriate. This was a sensible
approach, but as this part of the question is a step towards calculating a
merged balance sheet, it is at that point that this needs to be considered.
This was not clear from the question, and candidates who did this would
not have been penalised.
First, restate Company Bs balance sheet in accident year format, and
adjust to Company As standards, which may or may not be realistic. The
unearned premium should be 3,200 25% = 800. DACs on this should be
25%, or 200. Losses on the unearned premium should be 800 88% =
704, but given that reserves need to be increased by 10%, The Company B
will have reserved only 704 1.1 = 640. Consequently, the reserves
should be (12,000 640) 1.1 = 12,496. Therefore the accident year
balance sheet is
Investments
14,000
Total assets 14,000

(iii)
Unearned premium
Less DACs
Claims reserves
Share capital
Retained profits
Total liabilities
800
200
600
12,496
500
404
14,000
This part of the question was generally badly answered. Common mistakes
included not taking out of the assets the purchase price that Company A
has paid to the shareholders of Company B, or adding it to Company Bs
assets in the apparent belief that the purchase price is received by the target
company itself, and adding Bs share capital on to As. Few candidates
seemed to understand the nature of a share premium account, and believed
that it would be affected by the transaction. Many candidates did not
appreciate the comment that goodwill must be written off immediately.
The implication of this is simply that the assets and liabilities of Company
B can be taken into Company As balance sheet, any difference going
directly into retained profits. This simplifies, rather than complicates, the
calculations required.
Company As commission ratio is 20% (to judge by the relationship
between DACs and unearned premium). Add on the loss ratio and we get
80%. Therefore, there is a profit margin of at least 20% on unearned
premium. This amounts to 800. A similar calculation on B gives a loss of
13%, or 104. Overall, there seems to be no need for an additional amount
for unexpired risk, even if we were to take account of the expenses of
running off the business.
Page 11

Company A has paid 2,500 for Company B. Any goodwill is written off.
Therefore, adding the companies balance sheets together, but deducting
the purchase price, we get:
Investments
25,500
Total assets 25,500
Unearned premium
Less DACs
4,800
1,000
3,800
Reserves
18,496
Share capital
500
Share premium a/c 1,500
Retained profits
1,204
Total liabilities
25,500
The remaining cash of 500 could be shown separately from the rest of the
investments.
(iv)
This part was generally poorly answered. Candidates did not seem to
understand many of the implications of simple balance sheets. Some
candidates suggested that Company A was poorly invested as so much of its
assets were in cash. However, this ignores the fact that cash would always
be placed at overnight interest rates and would therefore yield investment
income to Company A, and that a company on the verge of buying another
for cash needs to have a cash hoard available to fund the purchase. There
were also some comments that revealed that candidates did not understand
the existence of the share premium account. Some candidates evidently
thought that it reflected the performance of Company As shares on the
stock market. It actually indicates that Company A has at some time
issued shares for a consideration in excess of their face value.
The ratio of reserves to annual premium suggests strongly that the two
companies write different types of business. As is short tail, with
reserves being only about fifteen months claims. Many candidates made
this point, although some looked at the ratio and simply concluded that
Company B was better reserved than Company A, despite the fact that we
know that in the restated balance sheet, the two companies are reserved on
the same basis. (Assuming that the unearned premium is half the years
written premium.) Bs reserves, as they are now restated, amount to 47
months claims. This suggests that As business is property or motor, and
may include domestic and smaller commercial business. Bs is more likely
to be industrial risks, with a large proportion of liability business.
The average tail of claim payment is much longer for Company B than
Company A roughly four years compared with nine months. A appears
to be rather more profitable than B. Taking account of investment income
alters this position to some extent. As profit before management
expenses per 100 of premium is 100 60 1.075 20 = 23.17. Bs, on
the other hand, is 100 88 1.0754 25 = 9.1. After management
expenses this is likely to be only a modest profit at best. The merger will
add little to As profit, unless either there are synergetic expense savings,
or else management action can be taken to improve Bs performance.
Before the merger, Company As solvency ratio was 4800/8000 = 60%,
which would probably be considered healthy for a company writing
Page 12

generally low-risk, short-tail lines of business. On its own assessment, Bs
was 2000/3200, or 62%. Some candidates included only retained profit,
or only share capital, in their solvency calculations. This suggests a lack of
understanding of the elements of a balance sheet. This was higher than As
but as it was involved in higher risk, longer tailed business, it was
probably less satisfactorily capitalised. A number of candidates got the
first half of this point, but failed to notice the second. On Company As
valuation its solvency ratio was 992/3200, or 31%, which is a much less
satisfactory position. This is an important point, which many candidates
failed to notice. It is possible that Company Bs valuation of liabilities is
correct, but if we know that the two companies valuation of liabilities are
on different bases, then a comparison without adjustment is liable to be
misleading.
The combined solvency ratio is 3204/11200 = 28.6%. This would generally
be considered low for a company with a significant long tailed liability
book of business. It may raise concern in the market, and is a lot lower
than Company As previous level, although Company A may well have
built up its cash hoard in anticipation of the takeover, so its previous
position may have been misleading. Many candidates realised that
Company As solvency position had worsened as a result of the merger, but
few seemed to realise that that is an automatic consequence of a major
acquisition for cash.
The overall expected profit if operations continue as before, including
investment income on shareholders funds, is 11,200 2,400 + 0.075
25,500 7,616 = 3,096 before management expenses are deducted. This
will perhaps allow retained profits to be built up to boost the solvency
margin, but will place significant constraints on growth, and possibly on
Company As dividend prospects.
It is difficult to see what benefit the takeover has brought to the
shareholders of Company A, unless it can bring about significant
improvements in the performance of what was Company B.
Page 13
EXAMINATIONS
17 April 2000 (am)

1.
2.
3.
4.
5.

separate sheet.

303A2000
Describe the consequences for a general insurance company of setting

inappropriate levels of technical reserves.
[5]
List the major areas of risk and uncertainty facing a general insurer.
[3]
You are the chief actuary for a general insurance company that is currently
assessing the possible takeover of another insurance company.
The target insurer has a large worldwide property and liability portfolio of
business.
In the last year severe ice storms and typhoons occurred in North America
and the Far East respectively, resulting in an unusually high level of
claims for the insurer, and claims being made on its reinsurers.
There have been a number of recent high-profile passive smoking claims on

which the courts have found against the insurer with very large awards for
damages paid.
One of the many key ratios you are considering is the claim ratio. Discuss the
factors you would take into account in analysing this ratio.
[5]
You are the chief actuary to a general insurance company. The companys
main lines of business are employers liability, marine, aviation and private
motor insurance.
(i)
Describe the features of the business that would influence the choice of
investments to be held in respect of these classes of business.
[4]
(ii)
Describe how the investment portfolio will affect the choice of discount
rate used in calculating the technical reserves.
[2]
[Total 6]
(i)
Describe briefly the main stages in the life cycle of a claim.
(ii)
Explain how the periods between those stages might differ between
claims for home buildings insurance and employers liability insurance.
[2]
[Total 4]
3032
[2]
You have been given the following summary information, extracted from the
accounts of a general insurance company. All figures are in (millions).
Profit and loss account for the year ended 31 December 1999
Gross premium earned
Adjustment for reinsurance
Net premium earned
Gross claims incurred
Adjustment for reinsurance
Net claims incurred
Investment income
Net commission paid
Adjustment for deferred acquisition costs
Net commission earned
Other expenses of management
Gross profit
Taxation on Profit
500
240
720
450
54
2
260
270
210
52
60
88
22
66
Balance sheet as at 31 December 1999

Technical provisions net of reinsurance
Unearned premium net of DACs
Outstanding claims
IBNR
Total
Issued and paid-up shares
Retained profits
Ordinary shares
Land and buildings
Fixed interest securities
Broker balances
108
475
1,517
1,500
200
800
100
2,100
100
400
2,600
2,600
Discuss the nature of this companys business and its financial condition,
insofar as they can be determined from these accounts.
[10]
(i)
List the reasons why a general insurance company models claims.
[4]
(ii)
Describe the main requirements of a financial model.
[4]
(iii)
List the items that would be included in a financial model that is to be

used in framing investment policy for a general insurer.
[6]
[Total 14]
3033
PLEASE TURN OVER
You are the actuary of a small general insurance company that writes only
home buildings and contents insurance. Legislation has recently been
introduced in the territory where your company writes business that allows
local governments to serve notices on the owners of properties that have been
contaminated by pollution, requiring them to restore the land to an
uncontaminated state. The Marketing Director has asked you to consider
adding to the standard buildings insurance policy a new option to cover the
potential liability of the homeowner for the cost of cleaning up land arising
from the service of such a notice.
(i)
Describe the particular risks to the company of writing business of this

type.
[3]
(ii)
Describe the exclusions that might be associated with the proposed new
area of cover and the problems that might be associated with these
exclusions.
[3]
(iii)
State with reasons the existing reinsurance arrangements you would

expect a company such as this to have in place before these new
proposals were made.
[3]
(iv)
Explain how those reinsurance arrangements might need to be changed

if the proposal is carried out.
[3]
[Total 12]
In a particular country, there are three general insurance companies,

Company A, Company B and Company C, writing a similar mix of property and
liability insurance. Each company analyses its claims experience in a different
way. Company A analyses its experience by year of accident, Company B by
year of reporting and Company C by year of underwriting. Explain briefly the
principles of each method, including the treatment of IBNR claims, and
discuss each methods advantages and disadvantages.
[15]
10
A large industrial corporation has recently acquired a general insurance

company that has traditionally written personal motor business. The new
management wishes to expand the business into homeowners insurance, small
commercial lines insurance, larger commercial risks insurance and personal
computer extended warranty insurance. The motor portfolio has traditionally
been written through independent brokers. The insurance company operates
in a region that is susceptible to windstorm.
(i)
List the various options that the insurer may have in distributing its
products.
[3]
(ii)
Describe briefly how the general insurance companys choice of

distribution channels may affect its business and the decisions it must
make, under the following headings.
3034
Volume/market
Expenses/setup
Risk premium
Experience in particular products and distribution channels
[10]
(iii)
(iv)
3035
List the types of traditional reinsurance protection available to the

company and indicate the suitability of each type of cover for the
portfolio it intends to build.
[6]
List the areas of cover that may be provided in the personal computer
warranty policy and the data from each that should be recorded in
order to enable future claims investigations.
[7]
[Total 26]
EXAMINATIONS
April 2000
EXAMINERS REPORT
This question was largely bookwork and was reasonably well answered.
Inadequate levels of technical reserves can lead to.
Inappropriate pricing decisions.
If they are overstated, this might:
Worsen the apparent results, leading to a loss of confidence by shareholders,

brokers and the stock market.
Reduce the apparent solvency margin, causing possible problems with the
regulators.
Tie up assets that could be applied more usefully to other projects, or have an
effect on the investment policy for example in relation to matching liabilities.
If they are understated, this might:
Lead to profits being prematurely distributed, leading to future problems

meeting liabilities.
Tax payments could be higher in the short term than would otherwise be the
case.
Might set inappropriate reinsurance arrangements.
This bookwork question requiring a list was very well answered by most
candidates.
Premium risk
Claims experience
Expenses
Commission
Investment income
Catastrophes
Latent claims
The acquisition and renewal of business
Poor policy wordings
Inflation
Judicial decisions
Legislation
Insurance Cycle
Failure of third parties
Fraud
Management risk
Page 2
This question involved considering the factors influencing the claim ratio. Many
candidates failed to gain marks by giving general answers and not giving proper
consideration to the specific points raised in the question.
The claims ratio is defined as Claims incurred/Premium earned.
It is usually calculated net of reinsurance.
Although will need to calculate gross to assess the performance of the target
companys underwriters bearing in mind the recent catastrophes.
Will need to look at overall ratio and separate ratios by class of business and
territory.
Will probably wish to analyse in local currency to avoid the impact of exchange
rate movements,
but also in home currency terms.
Will need to consider the strength of the reserving basis.
Claims incurred incorporates estimates of outstanding claims, including IBNR.
The former are likely to be high because of recent storms and the latter high
because of the recent court decision.
The trend in this ratio over the last five to ten years is likely to be severely
distorted because of the above.
So it is useful also to look at this ratio with the effects of each removed, if
possible.
There may be prior year movements that will affect the ratio, especially on the
tobacco claims.
Comparison with other companies and hence form a benchmark.
Several candidates did not structure their answer around the business categories
mentioned in the question and produced a generic answer. In addition some
candidates showed a lack of knowledge of the coverage given under different types
of insurance. In particular their were several cases of confusion regarding
aviation and travel insurance.
(i)
Many employers liability claims are reported and settled quickly.

However, some may take some time to determine / agree liability, some of
which may go to court and experience lengthy delays.
Also the potential for latent claims is now considerable.
Settlements are likely to be in real terms.
Marine and Aviation claims may be for damage and/or liability. Delays in
reporting are not significant.
However, settlement delays, especially for liability claims can often be
lengthy due to the number of parties involved.
Private motor claims are mostly reported and settled quickly.
A few liability claims may take a few years to settle.
Many claims are for repairs and will be affected by wage, inflation and the
level of prices of motor vehicle parts.
Country of business and hence currency of liabilities.
Page 3

(ii)
If used it is likely to reflect the total return on assets backing the

technical reserves.
May use a different discount rate for different parts of the business.
Where mismatched a notional portfolio reflecting a more matched position
may be used instead.
A lower discount rate than that calculated may be used to make a small
implicit allowance for contingencies
There may be statutory rules about whether or not you discount, and
what rate should be used.
No discounting of liabilities is likely

(i)
Period (for some types of liability claim) for the condition to develop.
Occurrence of insured event.
Claim reported to the insurer.
Claim processed by the insurer.
In some cases a period of time may be allowed for any medical conditions
to settle.
Claims accepted by insurer and any disagreements in respect of the
amount to be paid are sorted out.
In some cases disputes may result in lengthy court proceedings
Partial payments and outstanding estimates.
Claim settled and file closed.
Claim reopened if conditions reappear
(ii)
Some liability claims may have a long period of gestation, thereby leading
to long delays before the condition is reported,
whereas property insurance claims encounter few reporting delays. The
main exception to this is probably subsidence.
May be lengthy delays in determining liability in employers liability
claims long tailed.
Usually delays are short in property claims
Page 4
This very good question regarding the interpretation of accounts was very poorly
answered by most candidates. Many candidates failed to calculate simple ratios
proerly and to recognise when their calculations were wrong, for example when the
proportioned reinsured was greater than 100%. In addition, many candidates
failed to reconcile their calculations with the information given by for example
stating that the company was losing money as the combined ratio was 147%,
despite the P&L account showing a profit.
The companys outwards reinsurance premium is 48% of its inwards premium.
This would generally be considered to be a high proportion. It may indicate that
the company writes high-risk business that needs a lot of protection, or possibly
that it obtains a lot of business from other companies with a reciprocal
arrangement. It is not likely to need surplus relief, as it has a high level of free
capital.
The gross loss ratio is 144%, but the net is only 104%. This may indicate the
clever use of a soft reinsurance market, or possibly one or two major losses
occurred or came to notice during 1999 that meant that the reinsurance
programme was extensively used.
The adjustment for DAC is small, suggesting that premium levels were stable in
1999.
Technical reserves are more than 8 years premium. Unless the companys
business has reduced greatly, and we know that it has not in the past year at
least, this suggests that the companys business is very long tail. This is how the
company can be profitable: it generates enough investment income to offset the
effects of a loss ratio of 104%, commission of 20%, and expenses of management
of 23%, a combined ratio of 147%.
Capital is 500, about 200% of premium. This would normally be considered a
very high ratio, but with such a long tail account, and the possibility of needing
to strengthen reserves, it is probably merely prudent. There is no mention of
whether or not reserves are discounted. If they are not, then the effective
solvency position is even stronger than this.
The assets show a larger proportion of equity type assets than is usual for a
general insurance company. This is probably related to the long tail nature of
the account. Presumably this reflects a large portion of liabilities being to some
extent linked to inflation.
The return on capital is 15%. This is healthy, but not extremely high, and would
be lower still if reserves were discounted.
Broker balances are 20% of net premium, suggesting an average credit term of
about 2 months.
The companys assets yielded about 8% during 1999.
Page 5

(i)
To arrive at a risk premium per policy.

To select rating factors.
To determine premiums using experience rating procedures.
To demonstrate the estimated effect of changing the level of cover by
changing the level of deductibles or the effect of reinsurance.
To estimate the likely variability of claims experience.
To estimate the impact on reserves of latent claims/industrial diseases.
Financial planning.
To set the reserves.
For asset / liability modelling.
(ii)
The model must be valid, rigorous and adequately documented.

The model chosen must reflect adequately the distribution of the classes of
the business being modelled.
The parameter values being used should be appropriate to the classes of
business being modelled.
The outputs from the model should be capable of independent verification
for reasonableness and should be readily communicable to those to whom
advice will be given.
The model must be robust to internal or external changes.
Must be realistic and easy to use.
(iii)
Actual investments held at start of projection

Amount of liabilities at start of projection split by currency
The expected payment profile and variability associated with payments
Nature of liabilities and how strongly correlated to price and wage inflation
An asset / economic model that produces projections of expected wage and price
inflation and their variability, expected investment returns by type and currency of
investment and variability of investment return by type of investment
The amount of business to be written in the future, split by currency and nature
Tax rate
Reinsurance
Any restrictions on investment policy, either statutory or management specific
Definitions of solvency and return to be used in assessing the benefits of any one
investment strategy
The amounts of any dividends paid
A rule or target
(i) A question about a non-standard coverage. Most candidates failed to
generate many valid points.
(ii) Again, poorly answered. Few people cosidered excluding death, bodily
injury or illness, or even mentioned the policy limit and excess.
(iii) A standard question. Generally well answered.
(iv) Few people mentioned the need to re-negotiate the existing treaties, or
the possibility of setting-up stand-alone cover for the new risks.
Page 6

(i)
There is little or no claims data available from any source.

So the ability to accurately estimate claims costs likely to arise from this
source is very limited.
There is always uncertainty surrounding new legislation and how it will
be interpreted and operate in practice.
The legislation may change in the future in such a way that the effects
are retroactive.
As an option to the homeowner there is a risk of adverse selection.
Those homeowners in high risk locations may be more likely to take up
the cover.
There is a risk of moral hazard,
where the homeowner knows that the land is already contaminated at the
time the cover is taken out.
Large accumulations of risk may arise,
where many homes in the same road or estate are covered.
(ii)
Deliberate pollution caused by the homeowner. However this may be

difficult to prove.
Contamination which arises as a direct or indirect result of a homeowners
actions. For example, interfering with a precautionary measure, such as
a membrane, installed to prevent such contamination.
However, difficult to prove may be caused by a previous owner
Pre existing conditions
Exclude certain specified types of contamination
Death, bodily injury, disease, illness or injury to mental health.
Blight loss in value of the home as a result of the contamination
Criminal penalties in any way related to the contamination.
Costs below the designated excess level for such claims and above the
specified policy limit.
(iii)
Quota share treaty

to increase the spread of risks, as the insurer is small.
Risk XOL
for very expensive homes.
Aggregate XOL
to cover concentrations of risk by area or type of claim.
Cat XOL (or Stop Loss)
Page 7

to cover against the effects of a catastrophe and hence limit the risk of
insolvency.
(iv)
May desire a larger layer of cover on the Aggregate XOL policy, or

increase the limit on a risk XOL policy to cover the increased
concentration of risk and potential for large claims.
Treaties may have to be renegotiated
if the current reinsurers do not wish to insure the new option.
Alternatively, separate treaties may be arranged solely for this particular
option.
May use QS, Surplus, risk XOL or Aggregate XOL or a combination
thereof depending on what the reinsurers deem to be acceptable.
A bookwork question that was generally well answered.

Accident Year
Claims are grouped by the time period during which they occurred.
All claims in a cohort belong to the same period of exposure.
They will all therefore be subject to the same risk environment.
Even though they may have been written under different rating and policy
conditions.
IBNR claims will be included (eventually) in the time period in which they
occurred,
as will recoveries and re-opened claims.
Year of Reporting
Claims are grouped by the time period in which they are reported to the
company,
irrespective of when they occurred.
An apparent advantage is that no further claims will be added after the end of
the time period.
The method does not allow for the cost of IBNR and reopened claims.
Claims will arise from different exposure periods
which may have different volumes of business, cover applying and claims
settlement patterns.
Hence any claims patterns derived may not represent the current position.
Underwriting Year
Page 8
Claims are assigned to the calendar year in which the policy was written
Irrespective of whether the claim occurred in the first or second calendar year
of the policy.
Therefore claims occurring on two consecutive calendar years will be assigned
to the same claim year.
The method follows the way in which funded accounts are divided
and follows the total outcome of all policies written in each year.
It takes up to two calendar years before all claims have occurred
10
and an additional period before all are reported.

Each cohort of claims will have occurred over a wider risk period than an
accident year cohort.
IBNR claims included (eventually)
(i) Bookwork. Well answered.

(ii) Poorly answered. Candidates did not consider that different
distribution channels will attract different market segments, and hence
have
different risk premiums. Few mentioned the possibility of using market or
reinsurer data to counter the stated lack of experience.
(iii) Bookwork. Generally well answered.
(iv) Few people seemed to know what benefits were likely to be provided.
Most failed to mention labour or transportation/call-out costs. A number
seemed to think coverage would be given for electrocution, or for missing
components in the system.
(i)
Possible distribution channels are:
(ii)
personal lines/small commercial lines brokers

employed staff paid by salary and/or commission
direct telephone
direct mail
direct internet
affinity groups
banks and other financial institutions
specialist brokers for large commercial risks and inwards reinsurance
personal computer retailers and manufacturers.
The implications are:

Volume/market
A move to direct marketing may alienate existing source for motor
business.
Different channels may give access to different segments of the market.
Different channels may have different profitability owing to competitive
position.
Direct methods are suitable only for personal lines and small commercial.
Page 9

Expenses/set up
Direct operations would require significant investment in infrastructure
and training.
Persistency of business will determine the spreading of new business
costs.
Rates may need to vary by channel to reflect differing expense structure.
If distributing through banks or other financial institutions who does
all the processing and claims handling net rate agreement.
Risk Premium
Market segment may vary by distribution channel; therefore rates may
need to vary by distribution channel.
Lack of experience
The company appears to lack experience in both the new products and
also the distribution channel. Need to acquire this by using market data
and also possible assistance from reinsurers.
Even reinsurers may not have adequate experience for the PC extended
warranty product.
(iii)
Quota Share
These are new classes of business and distribution channels, so we may
want to share the experience with a suitable reinsurer for the
homeowners/motor/small commercial/PC extended warranty. Also there
will be overriding commission to assist new business strain.
Surplus
Large commercial risks will be very variable in size, so a surplus treaty
may be suitable. This will give the insurer the opportunity to write large
risks without taking on excessive risk on its own account.
Aggregate XL
The quota share will not protect the company against the accumulation of
claims from a single source (e.g. subsidence) so aggregate XL may be
needed.
Risk XL
There will be exposure to large single losses, so risk XL may be required.
Page 10

CAT XL
The area is known to be susceptible to windstorms, and other catastrophic
events are always possible. This could be protected against by cat XL.
(iv)
Policy benefits:
Replacement of faulty parts

Labour for replacement and repair
Transport costs/callout charge
Replacement of system if beyond repair
Factors:
Manufacturer
Manufacturers warranty period
Manufacture date
Components of system (VDU/CPU/printer)
Model numbers
Retailer
Extended warranty period
Cost of system (possibly broken down by component)
Purchaser type individual/business
Cover provided
Date of purchase
Location
Date of loss
Claim amount by component
Page 11
EXAMINATIONS
18 September 2000 (am)

1.
questions. You are strongly encouraged to use this time for reading only
but notes may be made. You then have three hours to complete the paper.
2.
3.
4.
5.

separate sheet.

303S2000
(i)
State the advantages and disadvantages of discounting technical reserves.

[4]
(ii)
Company A and Company B have always had exactly the same portfolio of
business, each taking 50% of any risk on a coinsurance basis. In their
published accounts, company A sets reserves that are intended exactly to
pay for outstanding claims, discounted at the rate of interest the company
expects to earn on its assets. Company Bs reserves always include a 20%
margin for prudence above the amount it believes necessary to pay these
claims, and it does not discount its reserves.
Discuss the differences between the accounts of Company A and Company
B, as they would appear to an analyst who was not aware of this
information.
[10]
[Total 14]
(i)
Define experience rating.
[1]
(ii)
Describe how prospective and retrospective rating methods work.
[4]
(iii)
Explain how number-based and cost-based experience rating systems are

used.
[6]
[Total 11]
(i)
List the main uses of policy and claims data in a general insurance
company.
(ii)
[3]
List the main aspects of an insurers experience that should be analysed if

the insurer is to maintain adequate control of its business.
[2]
[Total 5]
A general insurance company that writes only motor business exclusively

through brokers has experienced a gradual but increasing reduction in its market
share over the past five years. The company has decided to use the internet with
a view to preventing further reduction and achieving 50% of its sales through
this medium within the next five years. Discuss the possible effect this strategy
may have in the following areas of the companys business:
3032
The Cost Structure

Broker Arrangements
Calculation of Premiums
Business Mix and Volumes
Reinsurance Arrangements
[10]
You are the actuary to a general insurance company writing product liability
insurance. The Chairman of the company has noticed that the Government has
recently issued a new index linked government loan. This pays a nominal
coupon of 1.5% per year, and both the coupon and the capital value of the loan
are indexed to the local consumer price index. It matures in 2016. The
Chairman has suggested that this loan would be an appropriate investment for
the company. Explain, with reasons, whether or not the stock should be
purchased.
[5]
A large reinsurance company has recently received a request from a small

insurance company for technical assistance. The insurer, which has total
reserves of $60 million and currently writes $25 million of annual premiums, all
of which is private household insurance, has decided to expand into commercial
property insurance with a view to writing roughly the same volume of business
in that class in terms of premium income, thus doubling its total premium
income. The reinsurance company has been asked to provide assistance in
setting the premium rates to be charged, and to provide suitable reinsurance
cover for this new class and for the company as a whole. As actuary to the
reinsurance company you have available risk premiums for different types of
commercial property that you can provide to the insurance company.
(i)
Explain how the insurance company should formulate the total premium
to be charged for each risk.
[5]
(ii)
Describe:
(iii)
3033
(a)
the reinsurance arrangements you would recommend, including

any limits or restrictions to the cover, and
(b)
the information the reinsurance company would require from the

insurance company with regard to the risks it underwrites.
[5]
Describe how the reinsurance company should price each part of the
reinsurance policy.
[6]
[Total 16]
PLEASE TURN OVER
You are the actuary of a major insurer that for the last ten years has been
writing twelve year warranty business covering major damage caused by
building defects in new homes. The table below shows the policies issued and
resulting claims outgo for a particular builder whose premium rates are under
review at his own request.
U/W
Year
No of
policies
1
2
3
4
5
6
7
8
9
10
20,000
18,000
12,000
11,000
10,000
8,000
13,000
15,000
25,000
26,000
7
9
10
9
8
10
10
9
6
4
15
15
20
15
25
30
20
15
10
Claims paid in year after writing (thousands)

2
3
4
5
6
7
8
20
20
20
30
40
35
25
15
30
25
40
30
50
45
25
40
55
85
80
70
55
95
125
135
120
90
160
145
150
115
195
175
155
220
170
9
230
(i)
Describe the main claims features of this business that can be deduced
from the above table.
[3]
(ii)
You are aware that a competitor insurance company has quoted for this
business. You know that the rate the competitor quoted is lower than the
rate you are currently charging, but you are not sure precisely what that
rate is. Explain with reasons how you would calculate the new office
premium to be quoted, given that the necessary loadings for expenses,
commission and reinsurance are already known.
[9]
[Total 12]
A general insurance company wishes to review its premium rates. It has

maintained a full database of the policy and claims data required to analyse the
experience for a particular risk group.
Discuss why it may be necessary to make adjustments to the base data before it
can be used in a rating exercise.
[15]
3034
You are the actuary to a small general insurance company that writes personal
accident business.
One type of policy has a term of five years. Policyholders may buy it in units,
each of which costs 20, payable at the start of the term. Each unit pays a
monthly income of 100, payable at each monthly anniversary of the policy from
the policyholders accidental death until and including the end of the term. In
addition, a lump sum of 1000 per unit is payable at the end of the term, if the
policyholder has died as the result of an accident during the term of the policy.
The policy is sold through agents, each of whom receives a commission of 5 per
unit sold.
Describe in detail how you would set technical reserves for this policy. You may
assume that the business is profitable so that there is no need for an additional
amount for unexpired risks, and that the company does not discount its technical
reserves. You need not consider the effects of outwards reinsurance.
[12]
3035
EXAMINATIONS
September 2000
EXAMINERS REPORT
Subject 303 (General Insurance) Examiners Report 2000 Solutions

Candidates are reminded that under the current examination system the answers have to
be provided in a hand written form and therefore the examiners can only take into
account solutions to the extent that they are legible.
The examiners are required to test the candidates knowledge of the subject and also to a
limited extent, the ability of the candidate to apply the knowledge and principles in
practical situations. Most candidates do not succeed in obtaining a pass as a direct result
of not being able to demonstrate to the examiners that they are able to apply the
knowledge and principles of the subject in a practical situation. At first glance it may
appear that examiners go out of their way to contrive situations and contracts that are so
unusual that most candidates may never meet these situations in practice. However these
situations are the ones where the well-prepared candidate can demonstrate that he
understands the principles of the subject and is not just regurgitating bookwork
The bookwork component of this question was answered well by the majority of
candidates, however very few candidates were able to apply the knowledge in a
particular situation. The number of marks for a given question does provide an
indication of the number of points that are required for full marks, against this
criteria the majority of candidates provided a superficial solution for part (ii).
(i)
Advantages
Discounted reserves give a more realistic value of assets required to meet

expected liability
More useful for assessing solvency
More useful for pricing and comparing profitability of classes
Releases profit earlier (if no additional security margin is introduced)
Reserves smaller, appears more solvent
Disadvantages
Removes margins which would help if experience is worse than expected
Difficult to derive payment pattern for some classes
Complex and difficult to determine rate
Need to allow for amounts not invested (for short tail classes difficulties may
outweigh any small benefit of discounting)
Earlier release of profit also leads to earlier crystallisation of tax liability
(ii)
As Company A has set up lower reserves for the same premium income, it will
appear to be more profitable on an inception-to-date basis. However, the position
in any particular year will depend on whether the companies are growing or
shrinking. If they are growing then the reserves will increase, and Company Bs
will increase by more than company As. As the companies will have the same
earned premium, but Company B will have a higher claims cost, it will appear to
be less profitable. Conversely, if the amount of business being written is
reducing, Company B will appear to be more profitable.
Page 2
Subject 303 (General Insurance) September 2000 Examiners Report

Company B will have a higher amount of reserves for the same premium. The
analysts may deduce that its business is longer tailed than that of Company A.
However, with higher reserves higher investment income will be allocated to its
technical account, which will make its insurance business appear more profitable
than Company As. In a year when premiums are rising this will offset, partially
or totally, the premium effect; if premiums are falling, it will exacerbate
Company Bs apparent greater profitability.
One consequence of appearing to be more profitable, at least on an inception-todate basis, is that Company A is likely to pay more tax on its profits than
Company B, assuming that that is the taxation basis under which they operate.
It may also have released more in dividends. If it has, and its retained profit is
equal to Company Bs, the companies will appear to be of equal solvency strength
on the most widely used basis. However, in that case its assets will be lower than
Company Bs and it will generate less investment income.
On the other hand, if it has maintained its total assets at the same level as
Company B, then, with a higher tax bill it must have released lower dividends to
its shareholders despite its higher apparent profitability, since both companies
must have paid out the same amount in total to achieve this. In that case,
although its underlying strength is the same as Company B, it would appear to
be more solvent, as more of its assets would be accounted for as capital and less
as liabilities. The investment income of the two companies will be the same, but
Company A will appear to be generating more from the investment of
shareholders funds and Company B more from its insurance business.
Effect of tail of business
Effect on smoothness of profit
This question was answered well by the majority of candidates
(i)
Experience rating is a system in which the premium for an individual risk

depends, at least in part, on the claims experience for that risk.
(ii)
With prospective rating the premium at the renewal date depends on the
experience of the risk prior to that date.
The insurer takes on all of the underwriting risk in the next policy period.
With retrospective rating the premium for the current policy period is
adjusted, based on the experience of that period of risk.
A deposit premium is paid at the start of the period of cover.
And will be followed by an adjustment premium or refund at the end of
the period.
(iii)
Number based systems:

Systems based on claims frequency are normally used for small individual
risks with relatively low expected number of claims.
This is because the variability of the claim cost in any one policy year
would be too great to judge the relative severity of the underlying risk.
The accumulation of a sufficiently credible experience would take too
many years.
Page 3

Cost based systems:
These systems tend to be used for larger risks with numerous claims.
In these cases, the aggregate cost of claims experienced in any one year is
a more stable indicator or the relative level of that risk.
The large number of claims makes a number-based system impractical.
This question was answered well by the majority of candidates

(i)
Administration
Accounting
Statutory returns
Investment
Financial control
Management information Risk management
Reserving
Experience statistics
Premium rating
Product costing
Marketing
Reinsurance
(ii)
Claims data: frequency and average cost, by location and class of business
Exposure data
Portfolio movements
Expenses
Persistency and profitability by source
Investment performance
Solvency
Reinsurance
Run-off experience
This question, though requiring some thought and application to the scenario
given, was largely bookwork in nature. However, given this background, it was
answered well by very few candidates. Candidates on the whole did not provide
sufficient breadth or depth in their answers.
Cost structure changes will include:
IT development costs recruitment and equipment costs.
Overheads may fall as efficiency gains are achieved.
Advertising costs may increase in the short term to help achieve the desired
target level of business, though may fall in the longer term where less expensive
means are available on the internet.
Cheaper claims handling may become possible where smaller claims are handled
over the internet and by use of e-mail.
Brokers fees may fall overall as less business is put through them.
Page 4

However, the costs per policy may rise where less cost effective deals are struck
(i.e. the costs fall less than the business going through the brokers). Effect on
broker relationship. Capital efficiency should improve as the balance of
premiums with the brokers should fall. This will free up a portion of the insurers
assets which were otherwise tied up. Shift to fixed expenses from variable.
The cost structure will change as more of the expenses (maintaining an
infrastructure) are fixed and less (commission) are variable.
Premiums calculation:
Another complete set of premium rates will be required.
Different rating structure
It should be easier and more cost effective to update these periodically to take
account of new experience.
More frequent updates will become possible.
More sophisticated premium calculation system may be possible.
More extensive policyholder information can possible be captured via this
medium by developing easy to use system.
Also greater flexibility in data collection enables potential new rating factors to
be collected and tested.
Test the market with the rates
Take account of competitors rates
Business mix and volumes:
The mix of business will change as those purchasing insurance via the internet
will have different characteristics.
Likely to be more sophisticated.
Different geographical spread
Possibly increased proportion of policyholders in the 2040 age range.
Thus an increase in the concentration of risk is likely.
The projected volume of business is highly uncertain.
A lot will hinge on advertising
the design of the website and
social trends towards the increasing use of the internet at home.
Renewal rates may increase due to the ease of renewal via this medium.
This in turn may help to reduce costs per policy and thus premiums.
With this sales medium likely to become more popular in the future, early
entrance to this potential market may reap rewards in the longer term through
generating customer loyalty.
Reinsurance arrangements:
Page 5

Premiums may need to be negotiated where reinsurers perceive an increase in
risk perhaps due to potential increased concentrations.
New aggregate excess levels may be desired due to the increased concentrations.
Any quota share arrangements may need to be renegotiated as the reinsurer
may be less comfortable with the change in approach.
There may be increased risk of fraud early on as the system developed will still
be in its infancy.
Technical help from reinsurer
This question, though not completely bookwork, was fairly straightforward,

requiring a systematic approach to consideration of the liabilities and
corresponding appropriateness of the asset in question. As a result it was
answered reasonably well. However, many candidates did not describe the nature
of the liabilities for which the asset was being considered, and even when
described the description was often incomplete or inaccurate. This led to
candidates not con ncluding whether the asset was a good or bad match for
the liabilities in question.
Product liability covers faulty design, faulty manufacture, faulty packaging and
incorrect instructions.
Property damage / bodily injury
Usually on a claims incurred basis (rather than relating to year of sales).
There are some long reporting and settlement delays.
Where time is taken to attribute physical conditions to the correct cause and
where claims relate to rogue drugs causing settlements to continue for some
years.
Possible accumulation of claims.
The suggested asset purchase is appropriate by reference to the real nature of
some of the liabilities caused by the delays.
However, the term of the asset is considerably longer than the likely term of the
liabilities.
And the volatility of the suggested asset (long term, low, index-linked coupon) is
less suitable where the liabilities are uncertain rather than fixed in nature.
May be a suitable purchase for part of the free reserves.
However, expected return on this asset is likely to be lower than that which can
be earned on equity.
However, part of the free reserves must cover the SMSM, so may be suitable for
part as less volatile than equities.
Does this asset fit in with the portfolio currently held?
Other classes of business
Percentage of total business
Currency
Page 6

Type of inflation link
Marketability in respect of accumulation of claims
Regulatory needs
Value for money
On balance, argue whether or not worth bidding.
Given the bookwork nature of most of this question, it was for the most part poorly
answered. Candidates, on the whole, failed to cover the full range of items
requiring consideration. In particular, many candidates failed to provide
adequate explanation of the points in part (i) (as asked by the question).
Particularly poorly answered were parts (ii)(b) and (iii).
(i)
Office premium = risk premium + expenses + commission + reinsurance

investment returns + loading for profit and contingency
Risk premium = obtained from reinsurer.
Possibly adjust risk premium.
Calculate risk premium for home business from claims data.
Expenses = loading for variable and fixed expenses (i.e. per policy, per $
premium, per claim and per $ claim)
Must also allow for expense inflation over the period until all claims
arising are fully settled.
Anticipated changes in office efficiency.
Renewal / lapse rates to determine how many policies to average the costs
over.
Commission will depend on the proposed method of sale and average
anticipated future rate.
Reinsurance may vary by risk perhaps a variable and fixed
component depends on the reinsurance arrangements set-up.
Investment return will need to allow for current investment conditions,
inflation, tax and late receipt of premiums.
Unlikely to be significant as the business is relatively short-tailed.
Profit and contingency loading should allow for desired return on capital
and risk reward, risk free rate of return, inflation, degree of uncertainty,
market conditions and tax.
Variable expenses will automatically vary by size of risk through the
different premium paid and different expected claim amounts.
Fixed expenses may wish to vary this loading by size of risk.
Commission likely to vary by size of risk to some degree.
Reinsurance Quota share / Surplus will vary per premium:
Page 7
Risk XL premiums should be based on the characteristics of risk.

Aggregate XL / Cat XL should vary by degree of risk and uncertainty.
Profit and contingency should vary by perceived degree of risk and
uncertainty.
Risk free rate of return, inflation, degree of uncertainty, competitiveness
and tax.
(ii)
(a)
Quota share with maximum size of risk

Or surplus reinsurance treaty
Or Risk XL working layer with maximum size of risk
Or combination
There are likely to be limits to business volumes in the first few
years.
Aggregate XL to cover against accumulations of risk.
Cat XL to cover against widespread storm damage, etc.
Perhaps stop loss for commercial property or for whole business to
protect against insolvency.
May also include limits to the size of risks, number of each type of
risk, volumes of risk by location.
(b)
Owner, location, type of business.

Size (floor area), type of property, age of property, method of
construction, number of floors.
Security systems, fire prevention systems.
Policy number, value, premium, EML, SI, Period of cover.
Date incepted, underwriters name. Expiry date, if policy not
annual.
Previous insurer, claims in last five years for whole company
insured.
(iii)
Proportional reinsurance arrangements simple relevant proportion of

premiums + commisions / loadings, etc.
In theory:
Risk and Aggregate XL covers will need to be modelled by frequency and
average cost from ground up though the reinsurer will have to base this
on data available from its existing book of similar reinsurances.
Use different sets of assumptions to see how results may vary.
Page 8

Base on assumed business volume with adjustment clause for significant
deviations from expected.
Premium will reflect expected claims outgo + contingency loading to
reflect the expected variance in the results.
In practice:
The theoretical rate will need to be assessed.
In practice though the rates charged will be largely dictated by market
conditions.
However, in this case since the insurer has asked the reinsurer for
technical assistance the market rates are of less relevance.
Cat XL look at past experience on other similar Cat covers offered.
Likely to reflect market rates more than actual experience.
7
This question relating to a less common type of insurance was very poorly
answered. Even though the table was headed Claims paid in year, many
candidates proceeded to assume that the table related to cumulative costs.
Practically all candidates showed little understanding of the term long-tailed,
which was irrelevant in this question given that the earliest exposure periods were
still incomplete. Part (ii) was especially badly answered, with practically all
candidates suggesting that the usual chain-ladder methods could be applied to
this class of business.
(i)
It is a long-term policy and claims will be affected by inflation. At present, the
earliest years are not fully run off.
Costs are highly variable between calendar years, underwriting years and
duration.
Earlier years are more volatile than later ones in their development.
Policy is long term in terms of claims, but we cannot tell whether or not it is longtail.
Most claims are long delayed from inception towards end of policy term.
Relatively low cost per policy. Fixing houses is expensive, as claim frequency is
almost certainly very low.
No obvious catastrophe.
Significant step up from year 4 to year 5.
(ii)
Pure risk premium per unit exposure
= expected claim amount per unit of exposure, or
= (no. claims / no. pols) (no. pols / exposure) (total claim amount / no. claims)
Office premium p.u. = Risk premium p.u. + Expenses p.u. + Commission p.u. +
Reinsurance prem p.u. + Profit p.u. + Contingency loading p.u. investment
return
Page 9

What period are the rates to be applied for?
What is the mid-point? Ok for short term agreement
Or should they be indexed? Ok for long term agreement.
Need to allow for inflation from time point at which risk premium estimated to
mid point of base period for future rate to be charged and claim payment date
To calculate risk premium need to analyse claims data.
None of the insurance years in the current book of business are fully run-off or
even the risk period expired
Therefore need to find external sources of data if available,
Otherwise need to estimate run-off pattern by some other means
Data in respect of other builders may also be helpful for identifying trends
Chain ladder methods useful as a cross-check,
but fail to provide sensible answers because;
Triangle at later durations is highly uncertain,
The run-off pattern is not necessarily consistent over time,
Due to variation of weather conditions, economic conditions homeowners may
delay reporting claim if the policy has not yet expired, unless wishing to sell the
home and move on
Because the policy term is 12 years very little is known about the more recent
insurance years as most of the risk is as yet unexpired.
The quality of risk is likely to vary over time for a number of reasons. E.g.
minimum building standards, quality of workmanship, mix of business (change
in proportion of types of home)
Need to analyse the data at least making allowance for differences between
underwriting year, development year and calendar year Can use a multivariate
analysis,
Model the data to estimate parameters for each of the three sets of variables,
though information is limited with regard to most recent underwriting years.
These can then be used for estimating the bottom half of the triangle
To what extent is the result from this builder credible? Should we rely on its own
results or use industry-wide statistics?
This is a large builder (26,000 homes in one year). To what extent are there
statistics on the probability of such a builder going bankrupt.
Need to allow for any underlying calendar year trends believed to exist
apparent within the data
E.g. Social inflation and the effects of global warming
Expected changes in the quality of the building work
Assume average run-off pattern, unless there are clear reasons for a change to
occur in the future.
E.g. changes in the types of home constructed so that different types of claim
which appear at different durations become more likely to arise
Page 10

Adjust for known catastrophe or rare claims within the data not expected to
occur regularly
And add a contingency loading in respect of these items
The analysis of total claims paid per unit exposure (home) is unlikely to be very
informative on its own.
Would help to analyse frequency per home and average cost per claim separately
if possible to identify different trends therein
For example rising average costs may indicate rising inflation in rebuilding costs
or a change in the types of claims seen
What profit margin should be allowed
perhaps base it on the degree of perceived risk, allowing for the very long term of
the policy.
How do the calculated rates compare with those quoted by the competitor,
Or those previously charged.
Does the recent analysis in any way indicate that a lower premium can be
charged
Should the business be priced as a loss leader with a view to maintaining market
share or expanding
Contingency loading should allow for the risk of catastrophe
and the considerable random variation as demonstrated within the data shown
Due to the length of term of the policies written will need to allow for investment
returns on net balance of premium remaining throughout the policy duration.
In calculation of the risk premium the assumed rate of return should be based on
a matching portfolio of assets likely to be held as technical reserves.
Also the rate should not be chosen independent to the inflation assumptions
used. There should be consistency between them it is the difference between
the assumed investment return and inflation rate that is most important
This bookwork question provided the candidate with little guidance on the areas
to be covered and therefore required a systematic approach to consideration all the
potential areas. Probably as a consequence this bookwork type question was not
answered well by the majority of candidates.
Unusually heavy or light experience:
Claims experience tends to go in cycles

For some classes unusually heavy or light years may be experienced in
isolation
Especially if the risk is affected by climate
If experience is untypical then choose another base year
Or aggregate more years experience
Or apply an adjustment factor to the base year
Page 11
Which would obviously be subjective

Although industry data may be available.
Large or exceptional claims:
May be left in the data

Or truncated and spread
Or removed
Depending on the extent to which similar claims are likely to occur in the
future
Trends in claims experience
If trends are detected in the base data, it is important to attach more weight
to recent experience
Allowance for inflation
Trends should also be investigated to see whether or not they are likely to
continue into the future
Or of they are the results of a one-off change in company or market practice.
If they are expected to continue then an assumption will be needed to allow
for them.
It may be necessary to adjust past data.
Changes in risk:
Changes in risk can be difficult to deal with.

They may show up as trends and be dealt with as such.
Alternatively, major elements of the risk could be separated in the base data
And projected separately
And combined with an assumption about the future mix of risks.
Changes in cover:
Page 12
Changes in cover can be difficult to allow for.

Major changes are likely to involve the perils covered
Or the limits and excesses applied to each claim.
They may also arise from changes to underwriting
Or to claims settlement procedures.
If a peril is no longer to be insured
It may be possible to exclude these claims from the data.
If a new peril is to be insured
It may be necessary to use external data
Such as market statistics, consumer or manufacturer data, government
statistics.
Changes to limits or excesses are more complicated.
If there is a detailed database allowing all claims to be separately considered,
it may be possible to adjust each claim to the original gross amount
And project the gross data to the new rating period.
Otherwise it will be necessary to make more approximate adjustments
Based on any knowledge of the underlying claims cost distribution.
Either way the information will be incomplete
As many insureds will not notify claims below or near the excess points.
Future changes in the risk environment other than normal trends will need
to be identified.
Changes in reinsurance cost:
It will be necessary to allow for changes in reinsurance cost.

Maybe necessary to incorporate IBNR
Margins in reserves (positive and negative)
Errors in data
Changes in claim definition including treatment of NIL cost claims
Candidates are expected to be able to apply the knowledge and principles

underlying the subject in various situations. This question required the candidate
to apply his knowledge of technical reserves in the context of a particular type of
contract. In setting out the formula for calculating the Unearned Premium
Reserve most candidates made the appropriate assumption that the incidence of
claims is approximately uniform over the period of cover, however many
candidates failed to then allow for the reduction in the claim amount over the
period of the contract, and merely stated that the usual formula could be used.
Many candidates failed to observe that once a claim was made the claim amount
payable could be calculated exactly and therefore an exact method should be used
to calculate the outstanding claims reserve in preference to the usual triangulation
based methods used for other classes of business.
The technical reserves that might be required in respect of this product are:
unearned premium reserve

outstanding claims reserve
IBNR
Claims expense reserve
Catastrophe reserve
No need for UPR for policies already claiming
Unearned premium reserves cannot be calculated in accordance with the usual

formulae since the term is not one year and the risk is not constant throughout
the term. If a policyholder should die accidentally in the first month of the policy,
the total benefit would be 100 60 + 1000 = 7000, whereas if he should die
accidentally during the last month it would be 1100. In general, if he dies in
month n (n = 0,1,2, 59), the benefit will be 100 (60 n) + 1000 = 7000
100n. The amount of premium earned each month should reflect this.
The probability of a person dying accidentally each month should not vary unless
some extremes of age are included. If the most dangerous ages for young male
accidental death are included then assuming a level probability will be slightly
conservative, and might be acceptable. If older ages are included some
adjustment to this assumption may be needed.
The premium earned in each month should be proportional to the risk, or to
70 n. (n = 0, 1, 2, 59)
Page 13

The sum of these weights is 60 70 60 59 / 2 = 4200 1770 = 2430
After m complete months of the policy the proportion of the risk remaining will
be
(70 n) 2430
59
n =m
= ((60 m) 70 (59 + m) / 2 (60 m)) / 2430

= (4200 70m 1770 30m + 59m / 2 + m2 / 2) / 2430
= (2430 70.5m + m2 / 2) / 2430
= 1 47m / 1620 + m2 / 4860
Hence the unearned premium per unit of policy for a unit taken out exactly m
months ago is 20 (1 47m / 1620 + m2 / 4860). To hold this for a policy that
had a curtate elapsed duration of m months would be slightly conservative. The
correct value could be approximated by adding on half a month to get 20 (1
47m / 1620 + m2 / 4860 (70 m) / 4860). The exactly correct value for a policy
that had had d days since its last monthly anniversary is, in a 30-day month, 20
(1 47m / 1620 + m2 / 4860 (70 m) / 2430 d / 30).
From this, 25% may be deducted to allow for deferred acquisition costs.
Setting reserves for outstanding claims should be relatively simple. We know
exactly how many monthly payments remain on each claim. If this is t, then for
each unit of the policy we should reserve 100t + 1000. We may need to add
any unpaid payments between the date of death and the valuation date if
payment has not yet started, for example if it is still not proved who is entitled to
claim the benefit. If a claim is in dispute, for example if it is not clear whether or
not a death was a result of an accident, then a proportion of a full reserve might
be set up.
To estimate IBNR, the time taken to report claims should be investigated. The
value of an IBNR claim is straightforward, along the lines of outstanding claims.
We should calculate the claim amounts of all policies if they had become claims
in each previous month, in the same way as outstanding claims, and get an
average value across all policies for each individual month.
We may get the expected number from a delay table. This should be constructed
from the record of past claims by month of occurrence and by number of months
delay until reporting. This may be used to give the proportion of claims being
reported by the nth calendar month end after the accident (n = 1, 2, 3, ). Let
this proportion be pn , let the number of claims reported in the nth month prior to
the valuation be Nn , and let the average amount of claim on policies in force
during that month be Vn. Then the IBNR is equal to
(1 p ) N
n =1
Page 14
Vn / pn

Claims expense reserve would most likely be a percentage of the claims cost or
premium.
Page 15
EXAMINATIONS
2 April 2001 (am)

1.
2.
3.
4.
5.

separate sheet.

303A2001
State the reasons why a general insurance company may undertake

investigations with respect to its investments.
[3]
You are the actuary for a general insurance company that writes household
buildings and contents insurance. List distinct rating factors that are likely to be
used to rate this business.
[4]
(i)
Define the term reinstatement premium.
[2]
A $1m xs $1m excess of loss reinsurance treaty has the following terms and loss
history in the year it was written.
up front premium of $200,000

1 reinstatement at 120% additional premium
In chronological order the only losses (from ground up) to potentially impact this
treaty are:
1.
2.
3.
4.
$1.5m
$5m
$1.8m
$1.5m
(ii)
Calculate how much of each loss is recoverable.
[2]
(iii)
Calculate the reinstatement premiums generated by each loss.
[2]
(iv)
Calculate the rate on line of this contract.
(i)
Define three year accounting.
[1]
(ii)
Explain, with examples, why three year accounting is used.
[4]
(iii)
At the beginning of year one, 100 premium is received and 5% commission

is paid. 10 is paid out in claims halfway through year one, 30 halfway
through year 2 and 20 halfway through year 3, and claims handling
expenses are 50% of the claims paid.
[2]
[Total 8]
Calculate the first declared profit for this tranche of business using the
three year accounting basis, given investment return of 4% per annum.
State any assumptions that you make.
303 A20012
[5]
[Total 10]
Discuss the different approaches which could be used to protect policyholders

following the insolvency of a general insurance company.
[8]
Discuss briefly the validity of each of the following statements.

(a)
When calculating loss ratios to maintain correspondence
paid claims should be compared with written premium
accident year claims should be compared with earned premium
incurred claims should be compared with earned premium
underwriting year claims should be compared with written

premium
(b)
Surplus Reinsurance is a form of non-proportional reinsurance.
(c)
Deferred acquisition costs are shown as a liability on the balance sheet.
(d)
The Basic Chain Ladder method assumes that the oldest year is fully runoff.
(e)
The paid chain ladder method produces lower IBNR estimates than the
incurred chain ladder estimate due to outstanding claims being included
in the latter.
[9]
You are an actuary for a general insurance company which writes private motor
insurance.
Describe briefly the reasons why you may undertake an actuarial investigation
of premium rates for this business.
[6]
303 A20013
PLEASE TURN OVER
You are the actuary for a general insurance company and are about to perform a
premium rating exercise for a class of personal lines business.
(i)
List the claims data needed to perform the premium rating exercise.
[4]
(ii)
Explain briefly why the introduction of new premium rates is likely to

take place some time after the period of the base claims.
[3]
(iii)
(a)
Describe briefly the problems that may arise as a result of

introducing a new premium rating basis some time after the
period of the base data used in deriving the new premiums.
(b)
Explain how these problems may be mitigated.

[9]
(iv)
State the non-claims related factors to be taken into account before

deriving the final premium to be charged to the customer.
[6]
[Total 22]
You are the actuary for a medium sized general insurance company that writes
household insurance business. The cumulative paid and notified (paid plus
outstanding) claims for this class of business for the last five years of
underwriting are shown below:
Paid Claims ($000)
Notified Claims ($000)
Underwriting
Year
Underwriting
Year
1996
1997
1998
1999
2000
10
9
10
10
11
20
21
23
27
30
32
35
40
41
50
1996
1997
1998
1999
2000
40
40
41
46
46
80
86
77
79
78
80
87
77
77
77
18%
41%
61%
80%
100%
53%
102%
103%
100%
100%
Assumed
Development
Pattern
Assumed
Development
Pattern
(i)
Discuss the reasonableness of the above assumed development patterns of

paid and notified claims.
[8]
(ii)
Discuss how the ultimate claim cost estimated by the Basic Chain Ladder
method would be affected if during 2000 there had been a major
hurricane that had caused the paid losses in underwriting year 2000 to be
50 instead of 11 and the notified losses to be 100 instead of 46.
[6]
(iii)
Discuss how the accuracy of the ultimate claim cost estimated by the
Basic Chain Ladder method in part (ii) could be improved.
303 A20014
[3]
(iv)
Calculate, in respect of the 2000 underwriting year, the discount factor to

be applied to the undiscounted outstanding claims reserve to give the
discounted reserve as at 31/12/2000. Assume a 10% interest rate and that
the assumed development patterns given in the question are correct.
State any further assumptions you make.
(v)
[6]
Discuss the effect on the technical reserves the company may hold if the
information given was for Accident Year instead of Underwriting Year
but all the numbers within the triangles were unchanged and you were
preparing Accident Year accounts instead of Underwriting Year accounts.
(You are not required to explicitly calculate any reserves.)
State what additional information you would need to be able to quantify
this.
[7]
[Total 30]
303 A20015
EXAMINATIONS
April 2001
EXAMINERS REPORT
There were many points that the candidate could make in response to this
question, as a consequence most candidates scored well.
Evaluation of the existing portfolio
ALM to assist in setting investment policy

Allocating investment income between classes of business
Allocating capital between classes of business
Allocating risk based capital
Determining the return on capital
Determine discount rate to apply to liabilities
Determine the return made on the investments compared to appropriate
benchmarks
To evaluate the performance of the investment managers
Statutory assessment
Determine level of non-investible assets
Check liquidity
Check security
Put value on assets
Check matching by currency / term / nature
This question tested the candidates understanding of rating factors in relation to

household product. Many candidates missed the obvious factors of sum insured
and number of bedrooms. Some candidates showed their lack of understanding of
rating factors by including risk factors which could not be used as rating factors
i.e. difficult to measure or verify. The better candidates managed to list enough
factors to score reasonably well though.
Sum insured
Number of rooms / bedrooms
Location / postcode
Excess
Business use
Owner occupied or rented
Type of property
Type and standard of construction
Age of building
Age of policyholder
Locks / alarm / neighbourhood watch security discounts
Claims history / NCD
Occupied day / night
Heating method
Occupation
Sex
Periods property is vacant
High value items
Page 2
The answer given below is the one the examiners had expected the candidates to
come up with. It was however clear to the examiners that the course material must
indicate an alternative way in which such contracts work. As a consequence, the
examiners gave full credit for the alternative approach providing the candidate
was consistent in parts (ii) and (iii). Most candidates scored very high on this
question.
(i)
(ii)
(iii)
(iv)
The premium payable for the restoration of full cover following a claim.
For higher levels of XOL reinsurance a claim my lead to the amount of
cover for the remaining period of insurance to be reduced or terminated
unless a further premium is paid for reinstatement. Such a premium may
also be required for lower layers of cover if there is a limited number of
free reinstatements.
$0.5m, $1.0m, $0.5m, $0m
$0.12m, $0.12m, $0m, $0m
20% as the reinstatement premium is not included
This question was answered very well by the majority of candidates

(i)
This is the usual form of funded accounting
underwriting profits are first recognised at the end of the third
accounting year from the start of the underwriting year.
(ii)
Used because underwriting year is of fundamental importance
e.g. at Lloyds (determines who is on risk)
e.g. for reinsurance contracts that operate on a policies incepting basis
because for some classes considerably more information is available
after three years compared with after only the first year regarding
premium payment, claims settlement and making reinsurance
recoveries
e.g. marine and aviation insurance and non-proportional reinsurance
(iii)
Assumptions
no other expenses
claims expenses incurred at same time as claims paid
no further claims incurred
valid assumption about taxation
Brought forward
Premium
Commission
Expenses
Claims
Investment return
Carried forward
Year 1
Year 2
Year 3
0
100
5
5
10
3.5
83.5
83.5
0
0
15
30
2.45
40.95
40.95
0
0
10
20
1.05
12.00
i.e. profit at the end of the third year is 12.00
Page 3
Most candidates did not score very well on this question. Candidates showed a
lack of wider thinking other than the basic understanding of the Policyholders
Protection Board
In the event of insolvency there will be two broad categories of policyholder
liability outstanding, outstanding claims not yet settled and unexpired periods of
risk
Appoint insolvency practitioners, with any excess outstanding liabilities to be
met by the Government from taxes.
This offers the maximum protection.
But is unfair as the cost is met by all taxpayers
Meet outstanding liabilities via levies on the insurance industry.
As above, but unfair on those policyholders who are more astute and companies
which are better run.
Require deposits to be held in an insolvency fund which can then be used in the
event of insolvency.
As above, but to a lesser degree since the insolvent insurer will have contributed
at least in part to the outstanding liability.
Could apply the above systems to just private policyholders, or those who
purchased compulsory insurance on the basis that corporate policyholders are
more able to assess the likelihood of future insolvency and take steps to avoid or
withstand the effects.
Could apply the above systems to just specific types of insurance or outstanding
liability which are deemed of greater importance.
E.g outstanding claims, rather than unexpired risk
Or liability claims rather than property damage
Apply the above approaches to only outstanding claims.
Insureds lose out to the extent of cover not then provided
May not be able to get such advantageous rates on new cover
May need to cover a specified period after failure as insureds would otherwise
without realising be without insurance cover
Could give refunds in respect of unexpired periods of risk
By covering all remaining periods of risk, ensures little or no risk periods of noninsurance
Additional marks were given to those candidates who mentioned ways of
reducing the risk of insolvency in the first place. Marks were also awarded for
additional valid points such as requesting capital from parent (if one exists) or
finding a purchaser for the company.
Page 4
This question proved to be one of the main questions which showed those
candidates who understood some basic principles of general insurance and could
demonstrate to the examiners their understanding. Some candidates thought
that everything must be false in such questions and therefore proceeded to argue
against all such statements, whilst others limited their answer to each part by
one word statement of true or false.
(a)
Only if the paid claims are underwriting year paid claims.

True. As claims incidence reflects exposure.
Only if the incurred claims are accident year incurred claims.
True. As claims incidence reflects exposure.
(b)
False. It is a form of proportional reinsurance where the cedant is free to

choose(within limits) the proportion of each risk ceded.
(c)
False. They appear as an asset held to reflect the spreading of

acquisition costs over the exposure period.
(d)
False. The method assumes that past development pattern is stable but
nothing else.
(e)
False. Both methods estimate the same thing, the ultimate claim. IBNR
is simply the ultimate claim less a known figure, the incurred claim. If all
the assumptions behind the two methods are valid they should produce
very similar answers. The paid development method is as likely to
produce higher IBNR estimates as it is lower ones.
Many candidates could not demonstrate to the examiners sufficient reasons for
carrying out such an exercise, although most mentioned the more common
reasons.
To ascertain the overall profitability of the current premium rates
To ascertain the overall profitability of proposed new premium rates
To analyse segment level profitability of the current premium rates
To analyse segment level profitability of proposed new premium rates
Performance of current premium rates not in line with expectations
Comparison of current rates with competitors
Review the suitability of current rating structure in light of the current risk
environment, allowing for changes in the political areas, legislation, traffic, new
technology etc.
Assess effect of lapse rates
Assess new potential rating factors
Assess impact of cover changes, new perils in, old perils out, excess etc.
Assess extent of X-subsidies
Assess change in cost of reinsurance
Assess need for APUR
Change marketing strategy
Page 5
This was a fairly straightforward bookwork question which the examiners thought
the well prepared candidate would score highly on without too much difficulty.
However, it proved that many candidates failed to score sufficient marks on this
question owing to lack of bookwork detail and the skills needed to answer
part (iii b).
(i)
Date claim occurred

Date claim notified
Dates of payments
Amounts of payments
Date(s) of settlement
Date(s) of reopened
Estimates of the outstanding amounts
Rating factor details (at the time the claim occurred)
Cause of claim
Type of peril
Claim number
The length of time that can elapse before sufficient claims have been
notified on which to base a rating exercise
Delays in processing and analysing the claims experience
The time taken in assessing and receiving agreement that the
premium rates can be changed
The administrative time taken to implement a rate change
Time taken to receive approval from a regulatory body (necessary in
some countries)
(ii)
(iii)
(a)
Experience in the intermediate period may turn out to be different

to that expected because
(b)
Page 6
Claims trends not as expected

Claims inflation not as expected
Commission rates have changed
Expense levels have changed
Volume assumptions have proved inaccurate
Competition and market has changed
Investment returns have changed
Changes in legislation and court awards
Cover changes in product
Changes in risk levels
scenario test the profitability so the range of possible outcomes is

understood
use the most up to date data that is available and developed
enough, use different periods for different perils
consider competitors likely reaction to any rating changes that are
in the pipeline (i.e. where the decision has been taken but the rates
are not in the market)
analyse the results of previous rating changes to try and assess the
likely changes in volumes
(iv)
Expenses fixed
variable
Loading per policy
proportional to premium
per claim
proportional to claims
Commission if applicable
Investment return both income and capital growth
Reinsurance costs i.e. the net cost to the insurer of buying
reinsurance
Profit margin required by the company
Discounts available e.g. loyalty discounts
Payment method (admin fee for monthly payers)
Competitive analysis
Required growth of business volumes by number of policies and
premium and hence standing in the market in respect of market share
The examiners were very disappointed in the majority of answers given to this
question. There did not appear to be a time problem for the exam as a whole and
in fact a significant number of candidates attempted question 9 earlier than
others. In recent examinations for 303 there has been a trend of candidates not
being able to demonstrate their interpretation of a set of data. The class of
business in the question was one that most candidates should be fairly familiar
with, but many candidates failed to mention more than a couple of the obvious
points and most did not demonstrate their understanding of this particular
situation. Part (ii) was not very well answered in that there was very little
discussion given in most candidates answers most candidates merely made a
couple of calculations. Despite the question not referring to pricing the business at
all, several candidates indicated curtating the effect of the hurricane and
spreading the cost over all years which is not an approach used in reserving..
Although part (iv) was a straightforward discounting question many candidates
managed to make it more complicated by not reading the question carefully and
noting the instructions given. In addition many candidates attempted to calculate
the discount based on the notified triangle as well as the paid. The examiners were
generally pleased though with the answers given by those candidates who
managed to get as far as the last part in showing their understanding of the
different bases of accounting year definitions.
(i)
The notified pattern shows slight redundancy after 24 months.

This is not untypical for this type of business that is typically reported
quickly with fairly accurate case reserves being posted soon after
notification. The main exception to this would be in the case of
subsidence claims on buildings policies.
Page 7
As this is an underwriting year triangle we would expect around 50%

of the exposure and hence 50% of all losses to have been reported by
month 12.
The paid pattern seems unusually slow for such a normally short
tailed class of business. Actual pattern will depend upon the mix of
contents and buildings.
It appears to be developing linearly which is unusual.
The paid pattern is clearly not fully run-off after 60 months. The
notified pattern probably is and it shows that the paid claims are only
about 50/77=65% paid at this stage.
The selected paid pattern looks wrong because it is not fully run-off
and should therefore be queried.
The paid and notified data both seem very stable. There is no evidence
of unusual losses in the data that could distort the development
patterns. The paid claims would be expected to lag behind the notified
data.
The selected patterns are consistent with the data (apart from the lack
of a paid tail).
(ii)
The true ultimate claims should be increased by the cost attributable

to the hurricane.
Whether the ultimate claims calculated using an unadjusted chain
ladder method accurately reflects this depends on the relative
reporting and payment patterns of hurricane compared to nonhurricane losses.
If hurricane related losses are reported faster than non-hurricane
losses then the unadjusted notified chain ladder estimate will
overstate the overall claims and vice-versa.
Similarly, if hurricane related losses are paid faster than nonhurricane losses then the unadjusted paid chain ladder estimate will
overstate the overall claims and vice-versa.
The relative reporting speeds of hurricane and non-hurricane losses
will be influence by when during the year the hurricane happened. If
the hurricane happened early in the year it is likely to be more
reported than non-hurricane losses.
If most policies are 12 months in duration then there will be some
hurricane related losses in the 1999 underwriting year.
The ratio of hurricane paid/notified losses is (5011)/(10046)=72.2%.
This is much higher than the non-hurricane ratio of around 25%. This
would imply that Hurricane related losses are shorter tailed than
average and/or that the hurricane took place towards the beginning of
the year.
(iii)
Page 8
The simplest way to improve the accuracy would be to remove all the
hurricane related losses and project those separately.
We are still left the problem of how to accurately project the hurricane
loss but at least is does not distort the non-hurricane triangle.
If the hurricane occurred in the early part of the year we could assume
all loss notification have now been received and assume zero (or even
slightly negative) hurricane related claims.
If the hurricane happened towards the end of the year we could use an
exposure based approach (i.e. look at all the properties insured by X in
the path of the hurricane and estimate the likely frequency and
severity of future claims.
(iv)
Assume claim payments made mid-year (or any reasonable

assumption).
Assume ultimate claims are selected based on selected payment
pattern (or adjust calculations accordingly).
Use incremental paid pattern 18%, 23%, 20%, 19%, 20%.
Discount to end of 31/12/2000 for the 2000 underwriting year
i.e. pattern is 12 months old already.
Formula is
23% *1.1 ^ 0.5 + 20% *1.1 ^ 1.5 + 19% *1.1 ^ 2.5 + 20% *1.1 ^ 3.5
(100% 18%)
Answer is 0.84
(v)
For accident year accounts we need a UPR reserve in addition to the

outstanding claims reserves.
Depending on the profitability of the book we may also require a URR
in addition to the UPR.
You may be allowed to discount before deciding whether a URR is
required
As data in triangles in unaltered the numerical value of the IBNR
calculated will be identical.
The difference is that for the underwriting year triangle the IBNR will
be represent future losses in respect of all business written up to
31/12/2000 and not just the amount earned up to 31/12/2000.
The overall level of reserves will be lower for the company with the
underwriting year data by the amount of the UPR (plus any URR if
applicable).
Additional data required:
Premium volume with inception and expiry dates to enable
UPR to be calculated
Knowledge of the underlying business to decide if even
incidence of risk or not
Information required to estimate (discounted) loss ratio for
recent years to see if URR required
Page 9
EXAMINATIONS

1.
2.
3.
4.
5.

separate sheet.

303S2001
A general insurance company is considering writing property damage insurance

cover.
State the main types of property likely to be insured.
Describe briefly the areas of risk and uncertainty for a general insurance
company writing a small heterogeneous book of business.
[3]
[10]
A general insurance company is wishing to review its reinsurance arrangements.

State the factors that may influence its decision.
[7]
Explain briefly what is meant by:
(a)
(b)
(c)
(d)
discovery period
facultative reinsurance
IBNER
moral hazard
[6]
(i)
Explain what is meant by EML.
[2]
(ii)
Write down the maximum EML that can be insured, and not go beyond
the maximum reinsurance cover, if the only reinsurance the company has
available is a surplus treaty with N lines of cover and a maximum
retention of R.
[1]
(iii)
An insurer (C) takes on a risk with EML of $10m. C has a risk excess of
loss contract $5m xs $1m with reinsurer A which applies before a surplus
treaty with reinsurer B. The surplus treaty provides a maximum
retention of $3m and 4 lines with the minimum always being ceded.
Determine how much A, B and C each pay if a claim for $9m occurs.
[4]
[Total 7]
(i)
Describe two distinct methods used to account for general insurance

business.
(ii)
[3]
Describe the four accounting concepts which are widely used in preparing
the financial accounts of a general insurance company.
[5]
[Total 8]
303 S20012
(i)
State the prime objective regarding the investment of assets of a general

insurance company.
[2]
(ii)
List the factors which influence a general insurance companys

investment policy.
(iii)
Describe briefly the ways a supervisory authority could influence the

investment policy of a general insurer.
[4]
[Total 9]
You are the actuary for a general insurance company that writes only ten-year
new home warranty business, in a country where all homes must be built to a set
of minimum standards. The warranty is purchased by the builder before
construction commences, on behalf of the purchaser for a single up-front
premium the purchaser of the home then becomes the policyholder on legal
completion. The policy covers new homes from the date first occupied against
the cost of major damage resulting from building defects where the cost of
repairs exceeds $2,000 (at 1/4/2001 prices linked to the house re-building cost
index), and the builder is no longer in business. New IT systems are being
developed for all parts of the business. As part of the design process you have
been asked to specify your data requirements.
State the items of data you are likely to require for reserving and pricing
purposes.
10
[3]
[12]
You are the actuary for a general insurance company which was established five
years ago. Since then it has written only third party motor insurance.
(i)
Describe, with reasons, the extent to which the basic chain ladder method
may be relied upon for assessing the required level of outstanding claims
reserves.
[10]
(ii)
Describe the additional steps you might take to assess the required level
of reserves.
[3]
[Total 13]
You are the actuary for a general insurance company writing only commercial
lines business. You have recently been asked by one of the underwriters to
assist with a request from the owner of a small Christmas tree farm for
insurance cover against all major insurable risks in respect of that business.
(i)
Describe the main insurance products the farm is likely to need and the
associated perils covered by each product.
[8]
(ii)
List the rating factors you would use in premium rating each part of the
cover in (i).
[7]
(iii)
Describe the main characteristics of claims likely to arise under each

element of cover.
[10]
[Total 25]
303 S20013
EXAMINATIONS
September 2001
EXAMINERS REPORT
Introduction
The report does not attempt to offer a specimen solution for each question - that is, a
K Forman
20 November 2001
Most candidates scored full marks on this question although some failed to mention
anything more than Buildings Insurance.
Residential buildings
Moveable property / contents / possessions
Commercial buildings
Land vehicles
Marine craft
Crop
Aircraft
There were many items that could have been covered in this question which was
largely bookwork. Candidates who clearly knew their bookwork scored very high
marks. There was a good range of marks on this question which demonstrated that
some candidates could not recall or did not understand the wide range of areas of
risk and uncertainty in general and how in particular such a book of business would
be affected.
Premium rating structure does not accurately reflect the cost of the insured risks.
This may result in adverse selection.
The overall level of office premiums may be too low
Allowance for expenses is too low due to business volumes lower than expected
Business volumes / profit may be volatile due to competition.
Poor policy wording may result in additional unanticipated claims
The insurer will have to take appropriate measures to avoid moral hazard.
The insured risk profile may not be even over the policy duration
Claims costs or frequency may be greater than expected adverse fluctuations.
Claims costs or frequency distributions may be different from those expected.
There may be concentrations of risk, for example for fleet business where more than
one driver from the same fleet is travelling to the same destination
Geographical spread of risks should be considered, to avoid unnecessary
accumulations of risk
Possible under reserving by the insurer poses a serious risk of insolvency.
Incorrect recording of data poses considerable risk as accurate data is crucial for the
correct assessment of premiums and reserves.
Higher inflation than anticipated represents a minor risk
if liability cover is included then most of the liability is short-tailed, though there may
be some longer tailed third-party liability / injury elements which have greater impact.
Poor performing, illiquid or mismatching investments pose considerable risk.
Where the market is competitive, underwriting margins will be tight and investment
returns a more significant element.
If investments are not sufficiently liquid then there is risk of exposure to random
fluctuations in the level of claims
Risk of investment default
Catastrophes represent a considerable area of uncertainty.
Inappropriate or insufficient reinsurance levels of reinsurance
Third party default
Exposure to risk from political and legal changes / precedents
Page 2
Although this was a fairly straightforward bookwork question many candidates failed
to mention more than a few distinct points. Very few candidates scored more than half
marks on this question.
Classes of business written
Change in company strategy
Reduce volatility of claims experience / smooth profits
Size of free reserves
Total premium written
Geographic area covered (e.g. flood is only a risk in some places)
Accumulations of risk, too much in one geographic area
Accumulations of risk, too much of one class of business
Accumulations of risk, risks where claims can occur in more than one class of
business
Accumulations of risk, inwards reinsurance (but may not even know what you are
reinsuring)
Availability of reinsurance / capacity
Value for money of reinsurance return commission
Security status of available reinsurers
Any regulations on the amount and types of reinsurance that must be purchased
Existing relationship with broker / reinsurance
Risk aversity of management
Most candidates could define the four terms. The main one causing difficulty was
IBNER. Clearly candidates were guessing in some cases what it meant and in others
clearly had no idea such as stating that it stood for Incurred But Not Ever Reported.
(a)
A time limit, usually defined in the policy wording or through legislative

precedent, placed on the period within which claims must be reported.
Generally applies to classes of business where a long delay may occur from
origin of claim to awareness of condition giving rise to a claim and subsequent
reporting e.g. employers liability.
(b)
A form of reinsurance covering a single risk, commonly used for very large
risks or portions of risk written by a single insurer, that are shared among
several reinsurers. More time consuming than a Treaty.
(c)
A reserve reflecting expected changes (increases and decreases) in estimates

for reported claims.
(d)
The risk that an insured may attempt to take an unfair advantage of the insurer,
for example by suppressing information relevant to the assessment of risk or
by submitting a false claim.
Page 3
This question provided a wide range of answers. Although an unusual reinsurance

arrangement, not only did candidates not understand the order in which the two
contracts worked but many had difficulty in understanding that for the surplus treaty
that the minimum was always ceded. Some demonstrated that they did not understand
the significance of the EML. A few candidates thought that the EML was the largest
claim that could arise.
(i)
EML is the expected or estimated maximum loss

The largest loss expected to arise from a single event in respect of an insured
property which may well be less than its market or replacement value.
Used as an exposure measure in rating certain classes, including reinsurance.
(ii)
(1+N) * R
(iii)
$10m EML is split as follows

A:B:C 5:2:3
Claim is $9m, A will pay $5m, B will pay 2/5 of (95)= $1.6m, C will pay
$2.4m
An alternative answer of 5:0.75:3.25, was accepted if explained as it was
apparent that from the tuition material that candidates could have interpreted
this question in that way although the examiners consider the first solution to
be more valid.
Page 4
Very few candidates had problems with this straightforward bookwork question.
(i)
One year accounts, which consider all income earned and outgo incurred in a
year and permit the release of any profits at the end of the year
Funded accounts, which consider the business written in each year (and
income and outgo pertaining to that business) and do not permit the release of
profits until the end of a subsequent year (usually the third year)
(ii)
Going concern, that the business will continue to operate for the foreseeable
future
Accruals, revenue and costs are recognised as they are earned and incurred not
as they are received and paid
Consistency, like items are treated in a similar manner within each period and
from one accounting period to the next
Prudence, revenue and profits are not anticipated and provisions are made for
all known liabilities.
Most candidates gained good marks on this question. The main area where marks was
lost was in part (iii) in describing ways a supervisory authority could influence the
investment policy.
(i)
Maximise return
subject to meeting all contractual obligations and
recognising the uncertainty involved
(ii)
Term of liabilities
Nature of liabilities
Amount of liabilities
Currency of liabilities
Absolute size of free reserves
Size of free reserves relative to written premium
Size of non-investible funds
Culture / risk averseness
Regulatory requirements
Moral / marketing aims
Asset deemed over / under priced
Expected return
Investment managers advice
Availability of assets
Liquidity / marketability
Tax
Expenses
(iii)
Restriction on which asset types are admissible for establishing solvency and
meeting a minimum level of solvency
Restriction in amount for which each asset is admissible
Custodianship of assets
Prevent some assets being held
Prescribe that other assets must be held and amount thereof
Require mis matching reserves to be held if certain assets are held
Requirement of investment manager being experienced / take professional
advice
Specify asset valuation methodology for statutory purposes
There were many items of data which could have been mentioned for this class of
business. Scores were generally low because candidates could not extend the normal
data items needed to the particular product in the question. Nearly all candidates
could list the basic items but only the better candidates could state many items
relating to the data required about the builders.
Page 5
Policy Data:
Cover / Identifiers
Policy number
Policy status
Address
Geographical location (postcode, etc.)
House purchase price / sum insured
House-rebuilding cost index at purchase
Fees paid / premium
Endorsements / Exclusions
Policy Limits
Site number
Commission
Source of business
Excess
Dates
Date policy purchased
Date home started
Dates of various stages of construction
Date home completed
Date first occupied
Date of legal completion
Associated dates on which information is recorded
Term of policy / end date
Technical / Risk information
General risk information
Site size
House type (detached, terraced, flat, etc.)
Construction type (brick & block, timber frame, etc.)
Number of floors
Number of bedrooms / size of property
Floor space area
Site manager
Special features
Buildings standards in force
Inspection certificate
Page 6
Construction information
Ground conditions / treatment
Trees / subsoil
Foundations type
Ground floor type
Wall type
Window / door types
Roof type
Basement?
Builder data
Builder current trading status
Trade association
Date commenced trading
Date ceased trading
Reason for ceasing to trade (bankruptcy, retired, etc.)
Builder size (number of homes completed in last n years)
Turnover
Profit
Builder Identifier
Any subcontracting information
Other accounts information
Credit Rating
Company Directors
How long has the builder been in business
Premium rating category / measures
Past claims history
Claims Data:
Identifiers
Claim number (link to policy)
Name of current owner
Claim status
Claim accepted as valid?
Claim investigated or cash settled
Remedial works contractor identifier
Dates
Date damage first noticed
Date reported
Page 7
Date investigated
Date accepted as valid
Date initial damage case estimate made
Dates of payments
Dates closed
Dates reopened
Technical / Damage information
Type of damage / defect (code)
Damage / defect description
Location of damage
Causation of damage
Financial information
Initial estimate of repair cost
Current estimated cost of repairs / outstanding amounts
Paid so far
Type of cost repairs / alternative accommodation / technical investigation
Recoveries made from third parties
Other data
Competitors premiums
Investment returns for discount rate
Relevant inflation indices: RPI, NAE, House re-building
Salaries of staff employed to handle claims
Legal and professional costs
Equipment costs (computers, stationary, etc.)
Economic cycle
Building cycle
Profit loadings
Contingency loading
Reinsurance arrangements
Earning pattern of premiums
Page 8
Many candidates failed to demonstrate to the examiners the reasons why the Basic
Chain Ladder method was not the best to adopt in this case. Some candidates failed to
distinguish between the damage and injury type claims. In part (ii) almost all
candidates mentioned Bornhuetter- Ferguson and other methods but failed to expand
on further additional steps.
(i)
Outstanding Claims Reserves = Outstanding Reported Claims + IBNR + Outstanding
Claims Handling Expenses + Reopened Claims Reserves
Should always use more than one method to establish technical reserves.
BCL often more useful as just a broad brush check on other methods.
BCL method is very simple and has numerous implicit assumption.
Effect of large claims in data
Claims triangle unlikely to be fully run-off, so BCL cannot be used on the existing
data to estimate the tail-end liabilities.
BCL assumes a stable run-off pattern from year to year. This is unlikely to be the
case with TP cover which has low claim frequency and a skewed cost distribution.
Earlier years less credible as growing account from nothing
Might be reasonable for more frequent damage claims which are reported and settled
quickly for relatively small amounts.
However, particularly poor for injury claims which can often take considerably longer
to settle and are less frequent.
Future inflation is implicitly assumed to follow that experienced in the past data,
which is quite likely not to be the case.
Internal factors such as changes to the claims handling process may have occurred to
invalidate the same pattern as the past assumption.
As too can External factors, such as changes in the legal process.
Changes may also have occurred to the underlying risk due to things like;
Increased traffic conditions, changes to court award levels, changes to the risk profile
due to competition or rate changes.
The most recent years are very uncertain, especially in respect of the injury claim
liabilities, due to the longer delays.
Extent of reliability for IBNR will depend on basis of data.
Paid claims triangle little is known of latest year.
Incurred claims triangle greater information for latest year.
Extent of usefulness for estimating claims handling expense reserves will depend on
the extent to which the claims data includes or excludes associated claims handling
costs
(ii)
External data and curve fitting might be available in order to assess the likely extent
of the tail
May use additional methods of assessment, including amongst others;
Inflation adjusted chain ladder
Page 9
Bornhuetter-Fergusson
Average Cost per claim methods
Analysing the various delays in the lifecycle of a claim may provide insight into how
the tail may look
May approach reinsurers for technical assistance
May adjust the data for known internal and external factors
Ask the underwriters or other experts
Other reserves such as UPR, APUR etc.
10
This question tested the candidates ability to determine which types of insurance
products would be required for such a commercial enterprise. Most candidates
managed to mention most of the products but some demonstrated that they did not
understand the difference between perils, risk factors and rating factors. Some
candidates clearly ran out of time and failed to complete part (iii)
(i)
Employers Liability:
This insurance indemnifies the insured against legal liability to compensate an
employee or their estate for bodily injury, disease or death suffered, owing to
negligence of the employer, in the course of employment.
Perils are largely grouped as;
Accidents caused by negligence of the employer or by other employees
Exposure to harmful substances
Exposure to harmful working conditions
Public Liability:
The insured is indemnified against the legal liability for the death or bodily
injury to a third party or for damage to property belonging to a third party,
other than those liabilities covered by other liability insurance.
Perils will include compensation for injury from falling objects, pollution, etc.
Motor Fleet Third Party Liability:
Third Party Liability indemnifies the owner of a motor vehicle against
compensation payable to third parties for personal injury or damage to their
property.
Motor Property Damage
Indemnifies insured against loss or damage arising to their vehicle from
specified perils such as theft, subject to any limits or excesses.
Commercial Property Damage:
As per motor property damage..
For Commercial Fire perils include mainly fire, but can also include
Explosion, lightning, theft, storm, flood and vandalism
Pecuniary Loss:
Protects the insured against bad debts or other failure of third parties or effects
of recession, as specified in the policy
Page 10
Fidelity Guarantee:
Covers the insured against financial losses caused by dishonest actions by its
employees, including loss of money or goods owned by the insured or for
which the insured is responsible, and reasonable fees incurred in establishing
the size of the loss.
Business Interruption:
Indemnifies the insured against losses made as a result of not being able to
conduct business
Perils will include items such as fire at insureds property, and fire at
neighbouring properties.
Product Liability
Indemnifies insured against legal liability for death or injury to third party, or
damage to property belonging to a third party arising from product fault. Perils
include faulty design, packaging, misleading instructions etc.
Group Medical and Personal Accident Insurance
Indemnifies the insured against some or all sorts of the costs for medical
treatment and fixed amounts for loss of limbs etc.
Crop insurance:
Indemnifies the incurred against losses made to the crop
(ii)
disease
fire
storm
drought
Employers and Public Liability:

Payroll / number of employees
Type of business
Claims experience
Location of workforce
Equipment used / Processes involved
Materials handled
Health & Safety measures / procedures
Training
Turnover
Deductible
Motor Fleet:
Use of vehicle
Age of vehicle
Number of vehicles
Page 11
Value of vehicles
Type of cover
Excess
Occupation of drivers
Sex of driver
Age of driver
Type of vehicle
Location of vehicle overnight
Weight / capacity
Area of use (local / national)
Maintenance procedures
Level of use
Claims experience
Commercial Property
Size of farm / number of buildings / sum insured
Use of buildings
Construction design
Location
Value of stock
Age / condition of buildings / machinery
Excess
Claims experience / training provided
Fire precautions
Security measures
Pecuniary Loss, Fidelity Guarantee, Business Interruption:
Type of business
Turnover / earnings / sum insured / profits / projected sales
Value of work in progress
Materials handled
Equipment Used
Indemnity period (3-5 years)
Years trading to date
Previous bankruptcies
Financial Controls / Security(Cash kept on site)
Product Liability
Size of trees
Associated components
Packaging
Chemicals used in farming process
Location of sales (country)
Medical / Personal Accident
Age
Sex
Cover
Page 12
Number of people covered

Region
Medical history
Sum insured for PA
Activities in workplace
Crop Insurance:
Location
Past claims
Variety of tree
Use of pesticides
Sum insured / crop value
Security measures
Exclusions
Tree density
Fire prevention measures
(iii)
Employers Liability:
Many claims likely to be reported and settled quickly. However, some may be
complicated due to origination over long periods of time.
So long reporting delays (due to appearance and notification)
And long settlement delays (due to determination and agreement of extent of
liability and settlement of conditions).
E.g. Asbestosis, or here perhaps long-term effects of use of chainsaw.
Therefore long-tailed.
Claim frequency quite low, but seasonal effects.
Average cost of claim distribution quite skewed.
Potential for quite large individual claims and accumulations of risk arising
from the same cause.
Public Liability:
Similar to Employers liability, though reporting delays unlikely to be long
Motor Liability Claims:
Greater frequency than Employers and public liability claims.
Cost distribution not dissimilar
Reporting and settlement quicker, though not as fast as property damage
claims.
Accumulation of risk not quite so great class action case shouldnt arise.
Motor and Commercial Property:
Claim event sudden and easily determinable, so few reporting and settlement
delays.
However, some settlement delays can occur on larger claims to verify value of
stocks, etc.
Frequency is greater than liability claims.
Average cost distribution is less skewed.
Page 13
Accumulations are generally possible where there is a risk of fire on a

particular business park site. Though this is unlikely to be the case unless
there are neighbouring farms also covered
Pecuniary Loss & Fidelity Guarantee:
Generally few reporting and settlement delays short-tailed.
Frequency and cost of claims usually low.
Business Interruption:
Generally few reporting and settlement delays short-tailed.
Frequency and cost of claims usually low.
Though settlement delays may be longer than for Pecuniary Loss owing to
greater need for verification.
Product Liability
Generally few reporting delays
May be long settlement delays in determining liability
Low claim frequency
Very skewed cost distribution
Medical / Personal Accident
Generally few reporting and settlement delays short tailed
Sometimes subject to dispute
Claims often quite small in size
Crop Insurance:
Short tailed, except for claims arising from disease
Catastrophic
Frequency effects
Low frequency
High average cost
Page 14
EXAMINATIONS
17 April 2002 (am)

1.
Enter all the candidate and examination details as requested on the front of your answer
booklet.
2.
You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.
3.
You must not start writing your answers in the booklet until instructed to do so by the
supervisor.
4.
5.
Attempt all 7 questions, beginning your answer to each question on a separate sheet.
Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available Actuarial Tables and
your own electronic calculator.
303A2002
Describe the areas of risk and uncertainty in respect of expenses and investments for a
general insurance company writing only a small volume of business.
[9]
(i)
(a)
Describe the benefits provided by Employers Liability insurance.
(b)
Describe briefly, with examples, the distinct groups of insured perils

for such business.
[5]
(ii)
List the main rating factors that are used to set premium rates for Employers
Liability insurance.
[2]
[Total 7]
(i)
You are the actuary for a recently formed but rapidly growing general
insurance company writing only motor insurance business. The assets of the
company consist solely of index-linked government securities and cash.
Discuss the appropriateness of the investments currently held.
[7]
(ii)
Describe briefly the controls that a supervisory authority might impose on the
assets of a general insurance company.
[4]
[Total 11]
(i)
List the uses of a statistical claims model.
[3]
(ii)
Explain the difference between deterministic and stochastic modelling.
[3]
(iii)
Discuss the factors that would influence your choice between a deterministic
and a stochastic model.
[4]
[Total 10]
A general insurance company writes only motor insurance business. Five years ago it
replaced one of its existing rating factors with a new one, with a view to improving its
underwriting result and overall profitability. Pre-tax profits since then, however, have
fallen steadily each year. The Board of the company is of the (currently unfounded)
opinion that these poor results are due to the replacement of the rating factor made
five years earlier. As Chief Actuary, you have been asked to investigate the cause of
the recent reduction in profits.
Describe briefly the main actuarial investigations you would carry out to do this and
their respective purposes.
[12]
303 A20022
A large industrial company is seeking to expand its operations on a global scale. The
company wishes to ensure that it has appropriate insurance arrangements to
adequately manage the risks involved in the existing business, and make any
appropriate changes in respect of the expansion.
You are the Chief Actuary for a large multi-national general insurance company that
provides most forms of insurance cover, with the exception of fixed benefits related
insurance. Your company already insures some of the above companys risks and
those of some other industrial companies with global operations.
You have been asked to assist the company with the insurance aspects of their existing
business and possible expansion of business.
(i)
Describe in general terms how you would determine which risks are insurable.
[2]
(ii)
Describe briefly the various types of insurance product that might be

appropriate in helping to manage the industrial companys risks.
[6]
(iii)
Describe briefly the factors that you would consider in assessing the extent to
which your company can insure the industrial companys various insurable
risks.
[6]
(iv)
Describe in detail the various reinsurance arrangements that your company

will need to have in place, assuming that your company decides to provide
insurance coverage for the industrial company for all the insurance products in
(ii).
[6]
(v)
What factors do you expect your lead reinsurer to consider in deciding to what
extent they will accept the placement of the active risks.
[5]
[Total 25]
303 A20023
PLEASE TURN OVER
(i)
State why insurance reserves may be calculated for a general insurance

company.
[4]
You are the actuary for a general insurance company that writes only MIG (Mortgage
Indemnity Guarantee) business for a number of small regional building societies.
MIG covers mortgage lenders for the difference between the amount owed by a
borrower who has defaulted on their mortgage and the resale value of the repossessed
house. The amount owed by the borrower consists of the outstanding mortgage,
interest payment in arrears and any costs incurred in selling the house. MIG cover is
taken out at the same time as a mortgage and covers the lender for a specified period,
usually the term of the mortgage. A claim occurs when a repossessed property is sold.
(ii)
Discuss the characteristics of this business that affect which reserves might be
needed and describe the methods that may be used to estimate such reserves
for MIG. You should indicate which method(s) are likely to be used for each
of the reserves.
[18]
(iii)
Discuss the relative sizes of the reserves that are likely to be estimated for
MIG.
[4]
[Total 26]
303 A20024
REPORT OF THE BOARD OF EXAMINERS ON

THE EXAMINATIONS HELD IN
April 2002
Introduction
The report does not attempt to offer a specimen solution for each question that is, a
K Forman
25 June 2002
On the whole, this question was poorly answered by many candidates. Although
substantially bookwork, this question required candidates to pull together a broad
range of points. Some candidates focused too much on one of either Expenses or
Investment, rather than covering both in detail. The examiners were additionally
looking for some application of the details given in the question i.e. referring to the
small company; some candidates did this well.
Expenses
Commission is normally expressed as a percentage of the premium paid and so
Only possible uncertainty is if different types of intermediaries are paid different
levels of commission.
In this case a change in the mix of the source of business could result in actual
commission payable being different to that projected even though business volumes
are in line with expectations.
Change in staff and accommodation costs as the business increases/decreases.
This could be particularly relevant for a small company as the decision about when to
change premises etc. can significantly increase costs.
Changes in legal and professional costs. No control, affected by supply and demand
Changes in the rates of inflation that affect expenses, e.g. medical costs, salaries
Using an expense allocation that is not appropriate leading to anti-selection
Volumes of business different to expected and hence different contribution to fixed
expenses
Mix of risks different to expected
More relevant for small company as fixed expenses higher proportion of total
Claims expenses higher than expected
Business risks which affect expenses
Investments
Premiums take credit for expected investment performance of investments held to
meet liabilities in respect of the business.
There is uncertainty regarding the performance of these assets income and gains.
Page 2
Market conditions worsen

Proportion of investible assets less than expected
Claims paid faster than expected or any other mismatching comment
Forced to realise assets at an unfavourable time
Limited diversification of investments for small company
Investment expense higher than expected
Poor investment management

Tax
Default risk
Investment legislation
This was a bookwork question for which a well prepared student could score
maximum marks.
For part (i)(a), the examiners were surprised by the number of candidates who could
not describe Employers Liability benefits, which is contained within core reading.
Some candidates mistook rating factors for perils in part (i)(b). Part (ii) was
reasonably well answered.
(i)
(a)
Benefits
Employers Liability insurance indemnifies the insured against legal
liability to compensate an employee or their estate
for bodily injury, disease or death suffered
owing to the negligence of the employer
in the course of employment.
(b)
Perils
accidents caused by the negligence of the employer or other employees
e.g. safety guards
exposure to harmful substances
e.g. chemicals, coal dust, asbestos
exposure to harmful working conditions
e.g. loud noises, repetitive strain, stress.
(ii)
Rating factors
Payroll, no of employees
Type of industry or occupation
Previous claims experience
Location of the workforce
The materials handled
The processes involved
Turnover
Size of deductible
Level of staff training / score risk assessment
Provision of first aid facilities
Page 3

In part (i), some candidates described motor business as short tail and did not discuss
the suitability of the investments in terms of matching bodily injury claims. The
weaker candidates talked about the need for liquidity whereas the stronger candidates
recognised that premium income would be sufficient to pay most claims in the short
term owing to the growing account of the business. Part (ii) was well answered.
(i)
Would want to maximise return subject to meeting liabilities as they fall due
Rapidly growing => size of free reserves relative to written premium and
statutory solvency may be under pressure
Different matching considerations for vehicle damage and bodily injury
Balance of these liabilities depends on type of policies written (e.g.
comprehensive vs third party liability)
Vehicle damage claims reported and settled quickly
Bodily injury claims longer tail
Cash is liquid
.and has stable capital value (good for solvency)
. but does not provide inflation protection for either the property damage
claims or the bodily injury claims
Index linked government securities provide some inflation protection
..but inflation affecting bodily injury claims in particular will not
necessarily correspond with the type of inflation underlying the index
Can get ILGS with different terms
so can try to match tail of liabilities
ILGS capital values can be volatile and may not be suitable if solvency under
pressure
However marketability of assets not an issue as can pay claims out of
premiums currently being received but may depend upon currency of assets
and liability
Risk of default with both cash and ILGS is low
Value for money: would expect other asset classes to produce higher returns
over the longer term e.g. equities
So would consider using other asset classes for matching the free reserves
Page 4
However other assets may not be appropriate if there is then a currency

mismatch
Different taxation rules for different assets may affect decision
May depend upon any statutory solvency test
May be compulsory type of investment
(ii)
Restriction on the amount and type of certain types of assets that can be taken
into account when assessing solvency
Custodianship of assets
Prevention from holding certain assets
Prescription to hold certain assets
Requirements to hold mismatching reserves
Specify asset valuation methodology
Requirement of investment manager to be experienced/take professional
advice
Forced disclosure of investment policy
Specification of minimum solvency margin
The examiners were pleased by the standard of the answers to this question.
Candidates seemed to apply their knowledge well in answering part (iii).
(i)
Calculate a risk premium per policy

Select rating factors
Determine premiums using experience rating procedures
Estimate the effect of changing the level of cover by changing the levels of
deductibles
Demonstrate the effect of reinsurance
Estimate the likely variability of claims experience
Examine effect of actual versus expected claims experience
Estimate the possible effect of industrial diseases on the reserves
Produce reserve estimates
A check against case estimates
Assess the degree of solvency
Allocate capital to different classes or categories of business
Financial planning
Workload management / staff planning
Assess profitability of different classes
As part of Asset Liability Model to set investment policy
Page 5
(ii)
Each of the parameters in a deterministic model have fixed values

The model produces results in the form of a point estimate
Possible to sensitivity test the results of a deterministic model by running the
model with different parameters values
Some of the parameters in a stochastic model (e.g. number of claims or claim
amounts) are allowed to vary and have their own distribution function
A stochastic model must be run many times using a random sample from the
distribution functions
The model produces results in the form of a probability distribution
(iii)
Time and manpower available: stochastic models are more complex and timeconsuming
The nature of the parameters within the model.
e.g. stochastic models are more effective in allowing for volatility in asset
values
The availability of data: there may not be sufficient data to permit the fitting
of distribution functions with any level of certainty
The need to be able to explain the model and communicate findings e.g. to
those with less modelling experience; stochastic models can be more difficult
to explain
Purpose of the investigation
Deterministic automatically done to get stochastic
Current procedures
More informative: any additional information obtained by using a stochastic
model may not be useful
e.g. reserving: requiring only a best estimate reserve for the accounts
or a reserve that is likely to be adequate in 95% of possible outcomes
The answers to this question were on average disappointing. Many candidates wrote
out a lot of bookwork without relating their answer to the details given in the
question.
Some candidates assumed that the Board of the company was correct in its
assumption about the cause of the poor profitability. These candidates therefore
focused on the impact of the change in rating factor without considering other
drivers. Some candidates assumed the Board was incorrect and did not discuss the
impact of the change in rating factor at all. Those candidates who considered a
range of factors scored highest.
Need to investigate the following:
Underwriting profitability
- to check if poor results are due to quality of underwriting
- look at underwriting procedures and guidelines and adherence to them
Claims analyses. E.g. multiple regression, etc. using both the old and new factor.
- to investigate the effect of the change in rating factor
The effects of competition on persistency e.g. lapse rate investigation
to see if the change in approach 5 years ago mirrored those made by competitors or
has there been anti-selection.
Page 6
Has the insurance cycle affected the results

Compare with market experience in respect of ULR
Compare premiums required with premiums charged for deliberate undercharging
Has there been an increase in competitive pressures, through either more players or
lower premiums (soft market)
Underwriting performance on homogeneous cohorts of business
- to check for any adverse selection
Investment performance against peer groups, benchmarks and past performance
- to check if in line with targets / expectations
Effects of any internal changes
E.g. Change in sales medium, commission rates, changes to internal processing,
changes to claims handling, etc.
- To ensure various expenses are in line with expectations.
Effects of any external changes on profitability.
E.g. Tax, legislation, solvency requirements, road traffic factors, etc.
- To identify if poor results due to unexpected high claims
General claims trends.
E.g. Unusually large claims, high frequency, catastrophe events, etc.
- To see if due to natural random variation
Effectiveness of reinsurance arrangements
- to see if due to inappropriate cover
third party / reinsurance bad debt
Mix of business
- to investigate whether poor performance is due to poor coverage of overhead and
other fixed expenses. Different cohorts will have different contribution rates
Coverage and policy wording
- to see if any changes in premium were in line with the resulting changes in claims
experience.
Adequacy of reserving / changes to reserving practices
- The level of reserves may have been set at increasingly prudent levels.
Any changes in methodology of reserving
Any strengthening of prior years reserves
Part (i) was well answered.

Despite a similar question in the September 2001 paper, some candidates struggled to
describe the common insurance products. Most candidates wrongly assumed that
because fixed benefit products were not currently offered by the insurance company,
they would not be appropriate for the industrial company. Several candidates
confused the names of different products (e.g. pecuniary loss and fidelity guarantee).
In part (iii), some candidates wrote at some length about rating factors rather than
discussing more general factors that would influence the extent to which the company
can insure the risks.
Those candidates who applied the details in the question, rather than describing every
type of reinsurance contract indiscriminately, scored better on part (iv).
Page 7
The stronger candidates were able to distinguish between the considerations of the
insurance company in (iii) and those of the reinsurance company in (v).
(i)
The company must have an insurable interest in the risk being considered.
The risk must be of a financial and reasonably quantifiable nature.
Also, ideally the risks should be independent of each other
Should be an ultimate limit
Should minimise moral hazard
And the probability of incidence should be relatively small
(ii)
Employers liability
- Indemnifies the insured against legal liability to compensate an employee or
their estate for bodily injury, disease or death suffered, owing to negligence of
the insured, in the course of employment.
Public liability
- Indemnifies the insured against legal liability for the death of or bodily injury
to a third party or for damage to property belonging to a third party, other than
where covered by other liability insurances.
Fleet motor 3rd party liability
- Indemnifies the insured against compensation payable to third parties for
personal injury or damage to their properties.
Product liability
- Indemnifies the insured against legal liability for the death of or bodily injury
to a third party or for damage to property belonging to a third party, that
results from a product fault.
Property (General)
- Indemnifies the insured against value of loss or damage to the property or its
contents, subject to any limits or excesses.
Commercial Property
- resulting from pre-specified perils E.g. Fire, storm, lightning, flood, theft,
explosion, etc.
Fleet motor property
- resulting from accidental or malicious damage, fire, theft, etc.
Marine & Aviation property (if oil industry then own tankers etc.) and Goods
InTransit
- resulting from fire, explosion, jettison, piracy, etc.
Professional Indemnity
- if professionals in the company are negligent in the provision of their
services
Directors and Officers
- for protection against company being sued for acts D&O performed
Fixed Benefits
- for medical benefits / sickness scheme
Pecuniary Loss
- Protects the insured against bad debts or failures of a third party
Fidelity Guarantee
- covers the insured against financial losses caused by dishonest actions by its
employees
Page 8
Business Interruption
- indemnifies the insured against losses made as a result of not being able to
conduct business
Other valid types e.g. Project Insurance in case project to expand costs more
than expected
(iii)
The extent to which risks are already covered for this company
The extent to which similar risks are covered in respect of other companies.
Relationship with insured and past profitability
Likely profitability of additional business
How will the cover be structured? Will the company be looking for a multiyear contract?
Any other potential concentrations of risk
- by class of business
- geographically
Current level of free reserves. What scope is there for new business.
Reinsurance / co-insurance arrangements in place
- Do these risks fall within existing treaties
- If not, how easy will it be to arrange additional cover, facultative or
additional treaties
Any legislative requirements / restrictions
The Boards attitude to risk
The potential for long-term involvement/desire to maintain existing
involvement
Current classes of business authorised
Willingness to extend classes authorised to write
Business strategy
Staff expertise in areas of potential insurable risks
Competition clearly this would bring in a considerable volume of business /
premium income
What data is available to assess the risks to be insured
Large company, so quota share treaties unlikely to be used
Surplus may be needed for large commercial property risks if insurer does not
write much of this business
Need to determine retention and number of lines for each risk
However, likely to use the full range of non-proportional reinsurance products
available.
XOL policies cover the insured for losses arising above a pre-specified lower
limit up to a pre-specified upper limit
Risk XOL relates to single risks
Aggregate (clash) relates to accumulations on multiple risks, due to a single
event, or from a single cause through time
Cat XOL relates to losses arising within a pre-determined time span from prespecified events
Stop Loss relates to cohorts or portfolios of risks
These policies will often have a Stability Clause (particularly for liability
business) i.e. indexed limits
Page 9
Risk XOL is likely to be arranged to cover risks such as marine & aviation
property damage.
Aggregate XOL may be arranged and include several layers;
- for each class of business separately
- aggregated over several-classes
- aggregated by insured
Place business with different insurers to spread risk of reinsurer default
Cat XOL may be arranged to cover against specific pre-defined events, such
as Hurricane, Earthquake, etc.
Stop Loss may be arranged, though for a large multi-national it may not be
available
(v)
If the risks fall under existing treaties then they will be automatically covered.
However, if not
The current relationship they have built up with your company
Their confidence in the ability of the multi-nationals underwriters to
accurately assess the risks
Confidence in insurer to deal with claims in acceptable manner
Influence of business written by insurer
Availability of reinsurance for business it accepts
Availability of profit sharing arrangements
The quality of data provided
The cover already provided in respect of:
- the insured in question from all cedants
- other risks with your company
- each class of business for all cedants
- within the companys geographical regions of operation
Available capacity
Claims experience in respect of each of the classes risks
Whether it is authorised to cover all of the classes required
This question was a reasonably good indicator of those who understood the subject of
claims reserving and could apply their bookwork knowledge to a non standard class
of business.
Part (i) was mostly well answered. A few candidates discussed why different types of
claims reserve were required rather than why reserves would be calculated.
There were a number of errors in part (ii):
Page 10
The most common error arose from candidates using a different definition of
a claim occurring to that clearly given in the question. This often revealed a
lack of understanding of an IBNR claim.
Consequently, some candidates talked about the potentially long delays
between a mortgage default and the sale of the property and incorrectly
concluded that IBNR reserves would need to be large.
Many candidates incorrectly described MIG as long tail. Whilst the

exposure period is very long (theoretically up to the term of the mortgage)
the reporting and settlement delays are short compared with many other
classes of business.
Candidates generally discussed a number of methods but did not make an
appropriate selection for each type of reserve.
Weak candidates suggested using the 365ths or 24ths method for the UPR.
Very weak candidates suggested using case estimates for UPR or IBNR!
Part (iii) was reasonably well answered.

(i)
Reasons reserves are calculated include:

Determine the liabilities for the insurers published accounts
Determine the liabilities for the accounts for the supervision of
solvency
Determine the liabilities for the internal management accounts
Estimate the cost of claims in recent periods to get a base for
estimating future premiums required for a given level of profit
Value the company for sale/purchase
Value the company for setting the terms of a merger
Value the company for setting performance related pay for directors
To calculate tax liabilities
To determine investment strategy
Determine liabilities for any profit sharing arrangement
(ii)
MIG is characterised by uncertainty and as such presents challenges for

reserving.
The features of MIG are that:
Exposure period is up to 25 years.
However most insurers consider the exposure nil after 10 years as risk
is not uniform throughout term of cover and generally reduces over
time with rising house prices.
Average claims amounts are surprisingly consistent over time for a
particular lender but mar vary by lender owing to different lending and
repossession practices.
Hence the loan acceptance terms of the building societies are an
important factor and the reserving should be done at this level as well
as top down to check that performance is comparable to what would be
expected given the underwriting standards.
Most of the variability of claims experience is based on the number of
claims.
In particular circumstances, e.g. growing unemployment, increasing
mortgage rates and falling house prices, catastrophes can occur over a
long period.
As such any UPR may not be enough to cover future claims on
business already written
Page 11
The likely reserves for this class of business are:

Additional Provision for Unexpired Risks
Outstanding Claims reserve
Claims expense reserve
IBNR
IBNER
Maybe Statutory Claims Equalisation Reserve
Catastrophe reserve
Case Estimates and various triangulation methods can be used to estimate such
reserves both by accident and underwriting year.
Do average costs and numbers of claims separately.
Split data by lender, region, type of mortgage, loan to value ratio
UPR
Use underwriting triangles of earlier years claims experience inflation
adjusted chain ladder to see pattern of claims and earned premium
accordingly.
Patterns may change from one underwiting year to another owing to changes
in economic conditions.
Outstanding Claims Reserves
Claims are normally settled very quickly and hence likely that only a few will
be outstanding. Hence the most likely approach is case estimates
Claims Expenses
Likely to be a percentage loading of any claims costs.
IBNR
Information may be available from lenders as to which properties have been
sold but not yet notified to the insurer and thus likely to lead to claims. In
which case a count of number of claims can be taken and an assumed average
cost adopted based on recent experience. Alternatively calculate as a
percentage of outstanding claims or use delay table method
IBNER
As stated, average costs are fairly consistent and as such the need for an
IBNER is unlikely. In fact a negative IBNER may arise if mortgagees are
pursued and recoveries for bad debts are made
Statutory CER
Rules for calculation likely to be set by Regulatory authority e.g. a percentage
of premium is transferred to CER providing loss ratio is below a certain
amount otherwise a transfer from CER
Catastrophe reserve
Company may set up a catastrophe reserve which may be in line with its
premium rating basis such that for years when no catastrophe occurs whatever
the loading in the premiums for catastrophes is transferred to the reserve.
Page 12
Re-opened Claims Reserve

Likely to be zero (or in fact negative) so no method required
APUR
For business written to date need to assess if UPR is sufficient to meet cost of
claims and expenses for the unexpired period of risk.
Need to project likely number of claims and average costs. Could use the B-F
method as this takes the belief about the claims development for each year and
modifies it only slightly unless there are large deviations in the future from the
projections.
Easy to allow for some years having different development patterns to others.
Alternatively, if available, could model claims arising from arrears cases in the
past and based on current arrears cases project future claims resulting from
business already written.
(iii)

This is likely to be the largest reserve given that a claim only occurs once the
house has been sold, and exposure can last for up to say 25 years.
Additional Provision for Unexpired Risks
The size of this reserve depends on whether business written in the past has
been written at adequate rates.
In particular the stage of the economic cycle when business was written and
now will have a big impact upon the size of this reserve
If past rates have been inadequate then the APUR could be the largest reserve.
Outstanding Claims reserve/Claims expense reserve
Will be small as the time between a claim occurring and paying it will be
small.
IBNR/IBNER
IBNR will comprise of houses that have been sold but the insurer not
informed. Very small. IBNER will be smaller than IBNR and most probably
nil.
CER/Cat reserve
Both likely to be small compared to the UPR. Although the Cat reserve may
be large if built up over many years because when things go bad they go bad
in a big way.
Page 13
EXAMINATIONS

1.
booklet.
2.
3.
supervisor.
4.
5.
question paper.
303S2002
Explain why it is important to understand the nature of any flaws, inconsistencies,

errors and omissions within a set of data that is to be used for a review of pricing of a
general insurance policy.
[3]
(i)
Explain what is meant by moral hazard.
(ii)
Describe the various moral hazards associated with household buildings and
contents insurance.
[7]
[Total 9]
[2]
Describe briefly the factors that should be taken into account when determining the
allowance that should be made for future investment return when pricing a general
insurance product.
[9]
A general insurance company writes only Public Liability business.

(i)
(a)
Describe the characteristics of the liabilities that you would expect for
this class of business.
(b)
Suggest, with reasons, suitable assets that could be used to match these
liabilities.
[6]
(ii)
State the factors that will influence whether the company decides to match
liabilities and assets.
[3]
[Total 9]
(i)
Explain what is meant by the terms case estimation and statistical estimation.
[3]
(ii)
Compare and contrast these two methods of estimation.
[6]
A general insurance company writes a large book of household buildings insurance

business. It is considering switching from a case estimated approach to a statistical
approach for assessing the cost of outstanding claims.
(iii)
303 S20022
Explain, giving reasons, which of the two methods in (i) is likely to be the
most appropriate for this book of business.
[3]
[Total 12]
You are the Actuary to a general insurance company which started transacting
household buildings and contents insurance on 1 January 1998. All business written in
a year is based on the same premium rating basis. All reinsurance is on a quota share
basis. The only accounting information available to you is:
Year
1998
1999
2000
2001
Net Written Premium*

Net Earned Premium*
Net Loss Ratio*
25
10
150%
50
35
130%
100
70
100%
200
140
75%
* the loss ratio is calculated on an occurrence basis and premiums are net of
reinsurance ceded
(i)
Calculate the gross loss ratio for each of the four underwriting years, stating
any assumptions that you make.
[7]
(ii)
Explain why the figures in (i) are unlikely to occur in practice.
(iii)
Discuss the effects that such rapid premium growth may have on an insurers
balance sheet and future business plans.
[5]
[Total 16]
303 S20023
[4]
PLEASE TURN OVER
You are the Actuary for a medium sized general insurance company that writes mainly
commercial property business. Much of the property insured is in the North Eastern
area of the country in which it operates.
The company has the following two reinsurance treaties, one with Company A, the
other with B and C, which have been in place since 1 January 1997:
A
B
C
4 line surplus treaty with maximum retention of 500,000

200,000 xs 500,000 individual risk excess of loss
300,000 xs 700,000 individual risk excess of loss
The treaties operate in the order above. A stability clause provides for the indexation
of the limits for the treaty with Company C only. This index started at a value of 100
on 1 January 1997. To date six claims have arisen from risks written in 1997
(amounts shown in 000s).
Claim
Number
Expected
Maximum Loss
Total Paid
Index
1
2
3
4
5
6
1,000
2,300
1,600
2,800
1,200
1,500
500
3,500
300
1,600
800
4,000
102
104
101
105
108
101
(i)
Describe briefly the main features of surplus reinsurance.
[4]
(ii)
Discuss briefly the features of an individual risk that your company should
consider when deciding how much of that risk it should cede to company A.
[3]
(iii)
For risks written in 1997, the company retained the maximum proportion
allowable under the terms of the surplus reinsurance treaty.
Calculate the amounts payable by your company and by each of companies A,
B and C in respect of claim 2, stating any assumptions you make.
[4]
(iv)
Comment on problems that the insurer may face with the existing reinsurance
programme for the above six claims, suggesting alternative reinsurance
arrangements that would address these problems for future business.
[6]
[Total 17]
You are the Actuary for a small general insurance company writing only personal lines
motor insurance business.
(i)
Discuss the areas of risk and uncertainty inherent in the claims experience of
this insurer.
[18]
(ii)
Describe the business risks to which this insurer may be exposed.
303 S20024
[7]
[Total 25]
EXAMINATIONS
September 2002
EXAMINERS REPORT
Introduction
K Forman
26 November 2002
Many candidates mentioned points relating to only if the premiums were too high or
too low and did not go into the greater detail which the examiners were looking for.
As a result of this most candidates did not score many marks on
this question.
(i)
Need to know the potential variability in the results produced from analysing
the data.
Different data problems will affect the results in different ways and to a
varying extent. Understanding the nature of the problems will help in
determining what allowance must be made.
This will help to ensure that premiums are set at an appropriate level to ensure
desired profitability is achieved. If the premiums are set too high then likely
loss in business. If the premiums are set too low then could get selected
against.
At the high level, this will help to ensure the right decisions are made.
Helps in determining the cause of the data problems which can in turn help to
decide what actions to take to prevent such problems arising again in the
future.
Helps in finding improvements to the data capture methods.
The answer given below in part (i) below relates to the Core reading definition. The
examiners accepted also an alternative definition which they thought was also
appropriate policyholder acts in a way that makes the insured event more likely just
because they have insurance. In this case marks were given in part (ii) where the
candidate gave reasonable and relevant examples. Some candidates did not read the
question carefully and gave examples not relating to household buildings and
contents business.
(i)
The risk that an insured may attempt to take unfair advantage of the insurer,
for example by suppressing information relevant to the assessment of risk or
by submitting a false claim.
(ii)
The homeowner may under-declare the rebuilding cost of the home at

proposal.
This may simply be through failing to provide accurate information at
purchase.
Or may be due to failing to advise the insurer where modifications have taken
place.
E.g. Extentions, loft conversions, etc.
Page 2
With regard to contents insurance, the homeowner may deliberately underdeclare the value of their contents at proposal
Or may fail to revise the figure through time where affected by inflation
The insured may fail to notify of changes to personal circumstances relevant
for assessing the risk.
E.g. The home may now be empty during the day
The homeowner may now work at home.
Or may have provided misleading information regarding the security measures
at the property.
The homeowner may submit a false claim
Or inflate valid claim, perhaps including items that were never there or were
not damaged
The homeowner may have failed to disclose risk related information on the
proposal form
E.g. Past claims, etc.
Fraud, arson and deliberate damage
Household security, e.g leaving window open when gone out
Most candidates made a reasonable attempt at this question but only the better
candidates managed to get enough points to earn well in excess of half marks. Some
candidates concentrated upon the choice of assets and thus failed to mention many of
the points which the examiners were looking for.
Mix of assets held to back the required level of reserves (including free reserves)
- This will be important in order to assess the likely future returns, investibility of
assets (including premium payment pattern), risk of default.
Expected level of investment return
- when the expected rate of return is low, this component is less important.
Term of policy / Exposure profile
- more relevant where policy term > 1 year and significant portion of exposure is later
in the policy term
Length of tail of the business being underwritten
- long-tailed means more relevant
The capital allocated to back this particular part of the business
Page 3
- how is the capital allocation determined

- just technical reserves, with free reserves elsewhere, or does it include a portion of
the free reserves
Extent to which assets held are mismatched to the liabilities
- mismatching increases the level of risk so a downward adjustment to the assumed
return may be appropriate.
The degree of uncertainty in the expected level and timing of claim payments
- increased certainty means less of a reduction for mismatching
Any legislative factors that affect the extent to which such allowance can be made.
Consistency with other assumptions, especially inflation
Investment expenses.
Tax treatment of investment returns
- any allowance should be made net of tax
4
The examiners were expecting the candidates to discuss liabilities as detailed below.
The examiners considered the cases where candidates had interpreted the meaning of
liabilities as UPR etc. Marks were awarded for sensible remarks in these cases.
Most candidates did not find this question difficult and hence scored reasonably well.
The main reason for candidates not gaining marks on this question was owing to not
mentioning the difference between property and bodily injury type liabilities.
(i)
(a)
(b)
Page 4
Liabilities
Property part is short to medium term
Linked to price inflation
Less variable amounts than bodily injury part
Generally smaller amounts than bodily injury part
Bodily injury part has more variable term than property, some very long
term
Delays occurrence-notification and notification-settled
Earnings linked
Court award inflation
Prone to court award accumulation
Occasionally very large bodily injury claim
Susceptible to moral hazard (e.g. slip/ trip)
Assets
Consider property and bodily injury separately
(ii)
decide to match / not by:
Property is short tailed so, same currency, short-dated government

securities, other fixed interest, cash
Bodily injury is longer-tailed, real, for which equities may offer the best
match.
However, will need to match short term claims in the same way as
property unless there is enough cashflow to pay out of premiums.
size of company
absolute size of free reserves
size of free reserves cf GWP
size of free reserves cf ultimate liabilities in a year
size of free reserves cf absolute liabilities at a point in time
existence of any required statutory minimum
reinsurance arrangements
liquidity of free reserves
attitude to risk of the company
statutory regulations e.g. admissability
availability
expected return on assets
value for money
desire to diversify / security considerations
extent of positive cashflow
Most candidates answered this question well. The main reason that most marks were
not gained was owing to the lack of comment in part (iii) for the possible need of
some case estimating for e.g. subsidence claims.
(i)
These terms relate to the determination of reserves for outstanding claims.

Case estimation each outstanding claim is individually assessed to arrive at
an estimate of the total payments to be made.
Statistical estimation outstanding claims are assessed en masse in
relatively homogeneous cohorts, based on historical trends and patterns,
adjusting for known or anticipated future changes.
(ii)
Case estimation is:

- labour / time intensive
- Relies on skilled staff
- subjective
- suitable where claims are low frequency / high severity
- suitable where the book of business is small
- suitable if the company is relatively young, as there may be insufficient
history for statistical estimation, especially for long tailed business
Page 5
- are suitable where there are significant volumes of qualitative information

relevant for setting an accurate reserve, which require skilled interpretation by
an expert.
- this method can use all the available data including estimated inflation
- more likely to spot errors on individual claims
- it is difficult to ensure consistency through time and between assessors
- Separate estimates for IBNR are needed
- different statistical estimate methods allow for IBNR in different ways
- The reverse may be true for statistical estimation
- Statistical estimation may give inaccurate results if the past experience
contains unusual features unlikely to be repeated e.g. catastrophes and large
claims.
(iii)
Book of business is large, which means that case estimation may be

impractical and costly.
Household buildings insurance is not high frequency in the same way as
contents insurance, so statistical estimation may not be very reliable.
However, for a large book of business, the volume of claims may still provide
a sufficiently sound basis for statistical estimation.
Average Cost distribution is quite skewed. So for the lower frequency / high
severity claims statistical estimation will almost certainly be less accurate.
Given the information above, statistical estimation for the majority of claims is
likely to provide the most accurate estimate.
However, for the low frequency / high severity claims (e.g. subsidence), some
case estimation may be retained
Depends upon how long the company has been in operation.
The solution to part (i) is the approach that the examiners hoped most candidates
would use. In fact very few candidates used this approach or any other approach
which the examiners considered reasonable with a reasonable set of
assumptions. Most candidates failed to recognise that of the business written in each
year that 40% was earned in that year and that 60% was earned in the following year
thus making their assumptions invalid. Of those candidates who attempted this part of
the question many stated that the gross and net loss ratios were the same and
thus merely repeated the information in the question, although a few candidates
realised that this must be wrong for 7 marks! The main problem appeared to be able
to distinguish between the occurrence year and underwriting year concepts.
Even though some candidates failed in producing a reasonable solution to part (i)
they did go on to produce a reasonable answer to part (ii) for which there are many
reasons why the results in part (i) are unlikely in practice.
Part (iii) was answered well by the better candidates but some were put off by the first
two parts even though the third part could easily be answered without attempting the
other two parts.
Page 6
(i)
Assumptions
All policies are earned over a one year period

The terms of the quota share have remained constant over the period
1.1.98 to 31.12.2001
As the company began trading 1.1.98 all premium earned in 1998 was
written in 1998
All business written in a year has same loss ratio
Calculation
Note NIA is net incurred claims on accident year basis, and NIU is net
incurred claims on underwriting year basis.
Note that premium is earned 2/5 in the year it is written and 3/5 in the next
year
For 1998 NIA are 10 1.5 = 15, and NIU are 25 1.5 = 37.5. Hence net loss
ratio on an underwriting year basis for 1998 is 150%.
For 1999 NIA are 35 1.3 = 45.5, as 37.5 - 15 = 22.5 of these claims were on
business written in 1998, then 23 claims occurring in 1999 on 20 earned
premium, i.e. claim ratio of 115%. Hence net loss ratio on an underwriting
year basis for 1999 is 115% and thus NIU are 50 1.15 = 57.5.
Similarly for 2000 NIA = 70, 57.5 - 23 = 34.5 in respect of business written in
1999, thus 35.5 for business written in 2000 on 40 premium, hence claim ratio
of 88.75%.
For 2001, NIA = 105, 88.75 - 35.5 = 53.25 i.r.o. 2000 and 51.75 i.r.o. 2001 on
80 premium, i.e. claim ratio of 64.69 and NIU of 129.375.
NIA
NIU
NWLR
1998
15
37.5
150%
1999
45.5
57.5
115%
2000
70
88.75
88.75%
2001
105
129.375
64.69%
As QS reinsurance is assumed to remain constant the GWLR = NWLR

Notes
NWP = net written premium
NEP = net earned premium
NWLR = net written loss ratio
(ii)
Rapidly decreasing loss ratio and year on year growth are unlikely to
happen at the same time
Household insurance is a commodity market and so usually sells on price
So to build a book quickly you have to be cheaper than the competition
Page 7
Override or ceding commissions may mean that net ULR does not equal
gross ULR
Unrealistic that change premiums only once a year
Percentage reinsured likely to change over the years
Household profits tend to fluctuate with the occurrence/non-occurrence of
weather events
Any other comments about specific assumptions in the question not being
realised in practice
So a smooth(ish) increase in profits is unlikely

(iii)
Effects are:
Most candidates had few problems with most of this question. Parts (i)and (ii) were
generally well answered. Several candidates could not perform the calculation
required in part (iii). In particular, although the question stated that the company
retained the maximum proportion some candidates assumed that the maximum was
ceded i.e. all lines were being used. This error has occurred in the past which would
indicate that candidates do not read the question carefully or do not understand
Surplus Reinsurance. Part (iv) caused the most problems though with some
candidates assuming that all the 6 claims applied to the same risk even though the
question stated that the EML was different in each case. There were generally very
few comments regarding the problems faced by the insurer in respect of this given
arrangement even though detailed information was available from the question.
(i)
Proportional
Page 8
Fast growth only available by writing business at premiums less than the
market (unprofitably?) or having a unique customer proposition.
Reserve for claims / premiums would increase especially if rate are
unprofitable
New business strain could be considerable
Any imposed SMSM would mean capital considerations could restrict the
future rate of growth. If near the SMSM brokers may stop placing
business, even policyholder may hear that the company is not safe
Unless there is enough capital to comfortably support growth then plans
will need to be scaled back
May change investment strategy
May change reinsurance strategy
May increase staff accommodation, IT systems etc.
So claims shared in same proportion as premiums

Treaty
Proportion retained by direct writer may vary from risk to risk
Usually decided by direct writer
Subject to limits in treaty (numbers of lines / EML)
Which may be based on sum insured or EML
Enables insurer to write larger risks/ fine tune its exposure

Premium paid to reinsurer will be reduced for commission (return and
over-riding)
Administration more complicated than for quota share
Size of risk; bigger the risk, more company will cede to A

Uncertainty of risk, frequency and severity
Concentration of risk for this one policy
Extent to which it fits in with existing portfolio
.potential for accumulations through similar location of risks
.& by type of property
prior facultative cover
Own experience of this risk
EML of 2,300. Retain maximum of 500 => cede 1,800 to A i.e. 1,800/500
= 3.6 lines. Okay as <4.
Therefore A pays 3,500 x 3.6/4.6 = 2,739.
Net of surplus, company has claim of 761
Of which 200 paid by B
Lower limit for treaty C becomes 700 x 104/100 = 728
Therefore C pays 761 728 = 33
And your company pays the first 500 and the gap between B and C i.e. 28
so a total of 528
Check total payments equal 3500 and sensible comment if not
Assuming none of the companies default
There is only one payment made and that is 3500 at the time index is 104
(ii)
(iii)
(iv)
Where EML exceeds (4+1) 500,000, as it does for risk 4 above, terms
of treaty with A are not met
In such situations, could arrange for facultative cover (either XL or
proportional)
Going forward, may wish to increase maximum number of lines in treaty
with A or buy more lines elsewhere
Need to consider whether XL excess points provide sufficient cover.
Risk 6 has blown through the top of the programme
C being indexed but B not being indexed creates a gap in the programme
This could be a problem when inflation is high and should be eliminated
Greater use of proportional reinsurance could be used to spread risk of
accumulations
Eg reciprocity to avoid geographical concentrations
E.g. aggregate excess of loss, catastrophe excess of loss
Reinsurers complaining about estimation of EML
Risk of reinsurance default use more reinsurers
No problem with some of the claims
Page 9
Some candidates answered this question first rather than rushing their solution at the
end. There was a wide range in marks on this question with the better candidates
scoring over three quarters of the marks available for what was largely a bookwork
type question.
(i)
The uncertainty arises as the outcome of business already written and the
premiums to charge in future periods.
Claims
Motor insurance claims are subject to wide variability.
Especially as a small insurer there is uncertainty as to whether changes in
claims costs year on year are due to changes in the underlying risk or merely
random variation.
Variability will exist in terms of frequency, amounts, incidence and cost of
handling claims
Claims Delays
Delays from occurrence to notification result in uncertainty regarding the
number and cost of IBNR claims.
Delays from reporting to settlement result in uncertainty regarding the ultimate
cost of claims.
This uncertainty is greatest for the largest, bodily injury, claims especially if
they involve legal proceedings.
Changes in cover
If cover is added / deleted from the motor policies there probably wont be
sufficient data to make a reliable estimate of the impact of the change.
Characteristics of policyholders
If the company is aiming to attract different risks to those it has historically
held the claims experience may differ from the past.
It is difficult to determine how the claims will change.
There may be opportunities for anti-selection if the premium rates do not
correctly reflect the risk.
Unless there is superior accuracy in assessing the premiums which are less
than the market norm (to ensure they are set at a profitable level) then through
adverse selection any inadequacy will result in severe loss making.
Moral Hazard
Usage changes (e.g. fuel shortage)
Attitude to claims
Experience from the USA suggest that policyholder are starting to claim for
events they would not have done so previously.
i.e. increasingly litigious society.
Page 10
Crime / fraud rate

As this increases AD and theft claims may increase.
The timing of increasing crime rate is uncertain.
This will be correlated to other economic indicators.
Judicial decisions
This is often referred to as court award inflation. New precedents will be set
involving which types of claims are eligible for compensation and how much
the settlements for them are.
Occasionally new level of awards are set for existing types of claims.
This will immediately increase the average amount at which all future claims
are settled.
Sometimes these decisions will be retrospective meaning that the uplift applies
to all outstanding claims as well as future ones.
Legislation
Fiscal, changes in tax, cost of cars, cost of repairs
Change in cover, i.e. removing the upper limit on compensation or introducing
a requirement, e.g. to pay hospital charges.
Change in law, e.g. restricting the factors that can be used in underwriting, not
using driver age
Accumulation of risk
Could be exposed by business acquisition e.g. under 30s or as a small
company geographically exposed by writing a lot of business near the head
office, this could lead to an aggregation of claims from a weather event in the
area.
Catastrophe
A flood or hurricane could lead to many claims.
Currency risks
Paying claims in other territories exposes the company to the risks of
fluctuating currencies and currency mismatching if they do not hold reserves
in other currencies.
Reinsurance
This is subject to uncertainty as:
Inadequately appreciate the scale of the risks and purchase inadequate

reinsurance
Doubts about the value for money and availability of reinsurance
Page 11
For Cat whether the retention, reinstatement premium, upper limits etc. are
OK
Ability to make a recovery, solvency position of reinsurers
Policy wording
Must be precise so the only claims paid are those that the company intended to
provide cover for.
Also with regard to reinsurance contracts so that the company can recover
what it expects to.
Inflation
Uncertainty about future inflation especially for bodily injury claims will
affect the actual profit made and hence the assessment of premium required to
provide cover in future periods.
(ii)
Failure of third parties for example
Non payment from brokers

Staff dishonesty
Default by supplier
Timing, premiums late, recoveries late, claims early

Competition
Products are inferior
Prices uncompetitive, high expenses, inadequate analysis
Fall behind in technology, less business
Prices too cheap, unsatisfactory financial performance
Business affected by changes is legislation which other have allowed for
If lose business costs rise as fixed expenses have to be recouped
Data may be inaccurate / incomplete
Insurance cycle
As the insurer only writes motor business there is no opportunity to cross
subsidise with classes at different stages of the cycle.
As profits increase insurers enter the market, buy market share with low
premiums, market premiums decrease until losses are made. Insurers leave the
market, prices increase and the market edges back into profitability.
As only write one class at the bottom of the cycle will either lose business
(puts pressure on fixed expenses) or lower premiums and decrease the
solvency position.
Other business risks include expenses, investment and management
incompetence
Page 12
EXAMINATIONS
1 April 2003 (am)

1.
booklet.
2.
3.
supervisor.
4.
5.
question paper.
303A2003
A general insurance company writes only commercial property business. It writes 2

policies, both of which have a sum insured of 3,000,000. Policy A has an excess of
500,000 and policy B has a deductible of 500,000. Each of these policies
experiences a claim for 4,000,000.
(i)
(ii)
Explain the terms deductible and excess highlighting the primary

difference between them.
[4]
Calculate the amount of the 4,000,000 loss borne by the insured and the
insurer for policies A and B.
[2]
[Total 6]
You have been asked to determine the expected cost and variance of claims arising
from a homogeneous portfolio of insured risks by building a model.
(i)
State the type of model likely to be used, giving reasons.
(ii)
Outline the steps you would take in developing the solution.
[2]
[5]
[Total 7]
You are the actuary for a medium-sized general insurance company writing personal
motor business. The existing reinsurance programme has for the last 5 years consisted
of a small amount of quota share reinsurance and individual excess of loss cover of
10,000,000 xs 500,000. The Finance Director has suggested removing the excess
of loss cover entirely.
Explain how such a course of action would be expected to influence the
characteristics of the net liabilities and therefore the assets held by the company.
[6]
(i)
A global industrial company wishes to have a large commercial risk insured.

List the potential providers of this cover, outlining their key features.
[4]
(ii)
State, with examples, the main ways in which an insurance company may
acquire business.
[3]
[Total 7]
303 A20032
A small general insurance company (A) writing only property business cedes a quota
share reinsurance arrangement to a reinsurance company (B). The treaty cedes 40% of
the protected portfolio. A pays 25% commission to acquire the business and has
internal expenses of 7.5%. A expects to write 10m of business at a loss ratio of 65%.
B bears the ceded proportion of As acquisition costs and expenses. Bs own
expenses are 2% of gross premium received by B and B pays the broker 2.5% for the
business.
(i)
Calculate Bs expected profit on this treaty, stating any assumptions that you
make.
[4]
(ii)
As a pricing actuary for B you have been shown this treaty by an underwriter.
Describe what changes to the treaty you might suggest to the underwriter
before the terms are agreed.
[4]
[Total 8]
(i)
Describe the different ways expenses could be categorised for a general

insurance company when undertaking an expense analysis for pricing
purposes.
[4]
(ii)
Explain how the results of such an expense analysis would be included in

premium rates and why this is important.
[5]
[Total 9]
(i)
State a formula for calculating ultimate claims using the BornhuetterFerguson (BF) method of reserving.
[1]
(ii)
Explain why the paid BF ultimate is generally similar to the paid chain ladder
ultimate for older accident years.
[2]
(iii)
Discuss the advantages and disadvantages of the following approaches for

deriving an estimated loss ratio for use within the paid BF method:
303 A20033
(a)
Use market loss ratios derived from industry benchmark information.
(b)
Set the estimated loss ratio for year (x + 1) to the final selected ultimate
loss ratio for year x.
(c)
Use the ultimate loss ratios implied by the paid chain ladder method.
[5]
[Total 8]
PLEASE TURN OVER
(i)
(ii)
List 12 main rating factors which may be used in pricing personal motor
insurance.
[3]
(a)
State four risk factors that are not normally used in pricing this type of
business, describing the link between these factors and the rating
factors that are used.
(b)
Explain why the rating factors are used in preference to the

corresponding risk factors.
[5]
The marketing department of a general insurance company which does not currently
write motor insurance business has proposed issuing a special annual policy to low
mileage drivers for private use only. Only the policyholder is insured and at the start
of the policy year he pays a policy administration charge. During that year he then
pays for the insurance cover as needed on a daily basis by giving 24 hours notice and
having the appropriate cost deducted from his bank account. You have been asked to
identify the differences between this policy compared to a standard annual policy.
(iii)
Describe the advantages and disadvantages for the policyholder and the
insurance company.
[7]
(iv)
Describe the issues you would face initially in rating such a policy.
[3]
(v)
(a)
State four data items that are likely to be collected by the insurer which
would not normally be collected for a standard annual policy.
(b)
Explain why these are likely to be required.

[4]
[Total 22]
You are the actuary for a small general insurance company that writes only personal
lines motor business. A major motor manufacturer has approached your company
with an unusual offer. The manufacturer wishes to sell all its new cars with insurance
included in the purchase price i.e. their cars will have free insurance for a period of
3 years. Their proposal is that your company underwrites and administers the motor
insurance and the manufacturer will pay a one off fixed premium per car sold for the
3 years free insurance. This will not vary by any of the usual rating factors e.g. car
group, age of driver or past claims experience.
(i)
Describe the specific risks associated with the motor manufacturers proposal
from the insurers point of view.
[13]
(ii)
Outline possible ways of mitigating the risks in part (i).
303 A20034
[14]
[Total 27]
EXAMINATIONS
April 2003
EXAMINERS REPORT
Introduction
Mrs J Curtis
17 June 2003
The examiners were pleased with a slightly higher standard of solutions given by candidates
in this exam compared with recent diets. Comments on individual questions are given
below.
The examiners would like to remind candidates of the importance of reading the front cover
of the answer booklet and complying with the instructions. The examiners were pleased
with a significant overall improvement in handwriting. Examiners marking scripts are
under severe time pressure in marking all scripts by the deadlines given. Good
handwriting helps. An additional factor which seems to be creeping into some
candidates scripts is the failure to indicate the question number of the question that is
being answered, or in some cases indicate the wrong question number. Clearly examiners
take this into account when marking but such additional identification of which question is
being answered does delay the marking process.
The examiners would also like to remind candidates not to write their solutions in pencil,
only use the right hand side of the page and do not return scrap paper with the
answer booklet.
Most candidates were able to describe the terms deductible and excess, but only the
better candidates highlighted the differences. Being able to apply the definitions to a
particular case seemed to cause a lot of confusion. In particular some candidates
managed to get the calculations the wrong way around even though they got the
definitions correct. The examiners would like to emphasise to candidates that
application of knowledge is as important as the knowledge itself.
(i)
Deductible is the amount deducted from a claim which would otherwise have
been payable, i.e. it is borne by the policyholder. An excess is the sum that the
insured bears before any liability falls on the insurer. The primary difference is
that the deductible eats into the sum insured whereas the excess sits below the
sum insured. Hence on a policy with a deductible the maximum the insurer
will be liable to pay is the sum insured less the deductible.
(ii)
Policy A
Policy B
Page 2
Insurer
3,000,000
2,500,000
Insured
1,000,000
1,500,000
This question was generally very well answered, with many candidates scoring more
than 5 marks out of a maximum of 7.
(i)
A model with distributions is required as the mean and variance are required.
Stochastic is the most likely.
(ii)
Clarify purpose of investigation / specify the objective

Collect data
Modify data
Not necessary to group data as homogeneous
Choose a suitable frequency and severity density functions
(correlations should not be an issue as one homogenous portfolio)
Specify / estimate parameters

check goodness of fit of distributions
Allowance for change in new business volumes
construct model
run model multiple times and collect each outcome
sensitivity test the model
summarise the results mean, variance and distribution of outcomes
The examiners were disappointed with many of the answers to this question which
resulted in most candidates scoring less than half marks. Many candidates mentioned
the change in the net liabilities but did not go on to comment about the effect upon the
assets held.
Need to meet liabilities as they fall due so tend to match assets to liabilities by nature,
term, amount, currency
Changing reinsurance programme may impact the profile of the net liabilities and
therefore the assets that would match them
.although the company may not be following a closely matched policy, particularly
if free reserves are large
Removing the excess of loss cover will cause the impact of large losses on net claims
to increase
Large claims tend to be bodily injury claims
Therefore average term of net liabilities may increase if remove XL cover.
and payments for the bodily claims tend to be longer tailed e.g. court cases
So more inflation protection may be required.
.bodily injury claims inflation may be higher than damage claims inflation
So might wish to move to longer term assets that give good inflation protection e.g.
equities
Greater variation in retained claims cost
Removal of reinsurance may increase required statutory solvency so less investment
freedom
However with more exposure to random large loss movements, may need to choose
assets that are liquid
Page 3
Most candidates scored reasonably well on this bookwork question.

(i)
Companies
Public or private insurance limited liability, capital, distribute profits by
dividends.
Captives usually wholy owned Co. used by the owner to manage its risks.
Larger reinsurance contracts are then placed by the captive.
Self retention groups/Mutuals eg P&I clubs Companies owned by the
participants not for profit groups whose aim is to give value for money to
their members.
Government as insurer of last resort / high uncertainty
Lloyds syndicates capital provided by individual or corporate names to
one year ventures who underwrite on their behalf. Unlimited liability for
individual names, limited liability for corporates. All change here!
(ii)
Intermediaries brokers, banks

Retail linked to other sales travel insurance, extended warranty, mortgage
payment protection
Purchase another insurer
If captive, direct from parent
Employed sales staff
Mass marketing mailshots, newspaper TV and Radio ads.
Direct to customer internet, telesales
Although the solution given in part (I) was the one the examiners were looking for
alternative solutions with a slightly different interpretation of the information given in
the question were accepted. The main problem encountered by candidates was the
correct understanding of who pays what commission and expenses in such an
arrangement.
(i)
Premium
Claims
Commission
Ins Exp
Brokerage
Own expenses
40%
65%
25%
7.50%
2.50%
2%
Profit
-2%
4,000,000 (40% of 10m)

2,600,000 (65% of 4m)
1,000,000
300,000
100,000
80,000
(80,000)
Assume no investment income or other sensible alowance.

Expense allocation is accurate
Loss ratio will be as expected.
Sales targets are met
Page 4
No tax
No other prior reinsurance
No profit commission terms
(ii)
2% deficit. Realistically insurers expenses, brokerage, own expenses too

small to recoup deficit
suggest reducing commission to A.
Insurer may have fixed commission then difficult.
Insurer may have variable commissions with scope for reduction
Suggest sliding scale commission, or lower initial commissions with a profit
commission.
Suggest Insurer writes to a lower loss ratio
insist on tighter wordings or min rate rises.
difficult to guarantee, rate rises may lead to anti selection
Candidates showed that they generally understand the issue of expenses relating to
general insurance business. The examiners were looking for the distinctions between
fixed/ variable and direct / indirect. Many candidates also wrote at length regarding
claims expenses, investment expenses etc and credit was given for this approach.
(i)
Split fixed/variable.
Fixed expenses are those that do not vary with business volumes e.g. CEOs
basic salary.
Variable expenses are those that vary directly according to the level of
insurance business that is being handled at that time and may be linked to the
number of policies or claims or the amount of premiums or claims.
Split direct/indirect.
Direct expenses can be identified directly as belonging to a particular class of
business.
Indirect expenses are those that do not have a direct relationship to any one
class of business
All variable expenses are direct
but fixed expenses can be direct or indirect.
Split according to functionality
e.g. investment, acquisition, renewal etc.
(ii)
How use
Allocate as you wish the premium rates to be split,
so for personal business will split at least by product and possibly by cover
Link to business written or to claims..
Separate fixed and variable
Apportion indirect expenses across classes
Separate new business and renewal in theory if not practice
Why to get accurate costs / to cover expense costs overall
to get accurate rates
understand the level of cross subsidy in the rates
Page 5
e.g. renewals subsidise new business if there arent different premium rates for
each with the expenses split
understand the cost of writing business even if its not sold at the theoretically
correct rate
Credit was given in part (I) where candidates used a credibility formula or incurred
BF approach. Even though the question asked for a formula for Ultimate claims,
some candidates failed to give a formula and in several cases omitted to include the
current paid in their solution. Part (iii) turned out to be a good question to separate
those candidates who knew the topic and those who understood it.
(i)
Let PBF=ultimate from paid Bornhuetter-Ferguson method

cdf = selected factor to ultimate
EULR = expected ultimate loss ratio
Prem = ultimate earned premium
Then PBF = current paid + EULR * Prem * (1-1/cdf)
(ii)
(iii)
Page 6
The older the accident year, generally the more developed an account is
and hence the smaller the cdf.
For small cdfs, 1/cdf is large and therefore (1-1/cdf) is small. Therefore
BF ultimate become close to current paid. Meanwhile paid chain ladder
becomes closer to paid
(a)
+ Provides independent estimates

+ May reflect market rate changes and inflation effects as well as
trends in claims frequency and average cost
+ less need of own data which may be lacking or not credible
- Mix of business likely to be different for every company
- Need to check treatment of premiums is consistent (i.e. gross or net
of commission)
- Need to check consistency of reinsurance cover
- Market info not available
- Market info out of date
(b)
+ Gives some credit to the accounts own experience with its unique
portfolio of risks
+ Can be sure that premiums and claims are calculated on the same
basis
+ simple to apply
- May want to adjust for inflation and premium rate changes
- Year x may be an untypical year (e.g. exceptional large loss)
- Dependant on own estimate of previous year
- May be unreliable on long tail business
(c)
- Expected loss ratio would not be independent from claims on that

year.
- If use paid chain ladder loss ratio as expected loss ratio, PBF
becomes equal to paid chain ladder.
+ Could possibly use for reported BF projections
With so many rating factors to choose from which the examiners accepted as possible
main factors as those factors adopted by insurers do vary, most candidates scored full
marks on part (I). The examiners would like candidates to note that the question asked
for 12 and therefore were not expecting candidates to produce a list of more than 12.
Part (iii) onwards of this question sorted out the better candidates from those not so
well prepared in demonstrating their understanding. The poorer candidates gave
answers which were not much more than repeating the details in the question.
(i)
Excess
Cover
vehicle use
Vehicle age
Stated miles
driver address
driver occupation
claims record / NCD
years since passed test
gender of main driver
marital status
additional drivers
age of main driver
driving restricted to named drivers
make & model of vehicle
not in use location garaged
parked on or off street
security features
modifications
convictions / endorsements to licence
(ii)
(a)
traffic density where car driven To kept location

driver ability To claims record/age/sex/occupation
theft risk To driver address/not in use location/make & model
actual mileage To stated mileage
(b)
Rating factors must be capable of being objectively measurable. In the

cases above the risk factors can not be rating factors as they are not
objectively measurable and hence alternative rating factors are used.
(iii)
ph ads
only pay for cover as needed
potentially lower annual cost
Page 7
good if drive infrequent long distances

ph disads
may not want to or be able to give 24 hours notice / inconvenient
illegally driving without insurance cover if forget to arrange
may not include theft / fire damage when not insured
no other drivers allowed
daily cover may seem expensive
above a certain number of days the cost of cover will go above that of a
standard policy
may lose NCD within year of insurance
ins co ads
potential to penetrate niche market
could make affinity group deal
gain additional rating data re low mileage drivers
possibility of applying rate changes during the policy year.
ins co disads
additional costs of administration high if use call centre
high startup costs if use e technology like email or text message and lower
ongoing cost.
computer system may not be able to hold all data required
possible disputes over cover if crash occurs when not covered
cover runs midnight to midnight? Or a more specific 24 hrs, if so need to
record time and date of cover
more fraudulent claims / increased moral hazard
tendency to increase mileage on days insured
Page 8
(iv)
none of own data therefore need to buy in data or run a pilot

no data on the number of days low mileage drivers use their cars
so daily rate is not known with certainty
people identifying themselves as low mileage drivers may be high/medium
mileage drivers so current data may not be reliable.
The expenses and administration costs will not be known with certainty
(v)
collect individual dates of cover to collect true exposure period

if phone data collection then could ask about planned driving times &
locations and collect additional data e.g. mileage
bank details for deducting the daily charges could be different from details
first used, or previously used.
Passwords etc to verify caller ID more important than annual policy as
more frequently used.
This question proved to be the most difficult for most candidates. The best solutions
numbered the specific risks in part (I) and referred to each in turn in part (ii).
Many candidates considered that product liability would be a major risk which would
more likely be only of a 2nd order risk. In part (ii) many candidates wrote at length
about reinsurance and the different types. Hence they did not pick up on the many
different ways of mitigating the risks.
Time did not seem to be a problem for the majority of candidates.
(i)
Anti-selection by:
age
past driving experience
location
vehicle types
Catastrophe risk
Propensity to claim increases as will not worry about next year premium
Increase in moral hazard
Delay in receipt of premiums
Potential for business to be loss making, as small could threaten solvency.
Contact lasts for 3 years so could be tied into loss making rates, also the need
to predict inflation for 3 years.
Car sales are seasonaloperational problems as small company doing the
admin.
If volumes are small will not cover start up costs
Currency risk as manufacturer may sell in many countries
Difficulty in obtaining appropriate reinsurance cover.
Solvency capital considerations as small company and this is a major motor
manufacturer, it will substantially increase new business.
Difficulty in setting the fixed price long discussion as dont know the mix
of business are pricing for.
Small insurance company will not have much useful historic data especially of
writing brand new cars from this manufacturer.
Billing mechanism will need the facility for verification so both parties agree
how many cars have been sold
Risk that the manufacturer goes bankrupt and the insurance company does not
get paid although the cover is given.
Again there is the risk that the insurer itself becomes insolvent.
UPR variable over the 3 years as the risk varies over the term
If the cover is only for the original purchaser this will exacerbate the
variability of UPR/earnings.
Need to consider the investment strategy in respect of investing receipts for
longer than for an annual policy and risks associated with expected returns and
security
Introduction of, or increases in rate of, tax or levies imposed by the
government based on earned premiums.
(ii)
Ensure the motor manufacturer pays up front to mitigate the risk of the motor
manufacturer defaulting after cover has been provided.
Individually rate the business retrospectively when customer and vehicle
information is available (make assumptions for information gaps), run rather
like a fleet so collect deposit premium based on estimated car sales and
Page 9
customer mix and then an adjustment premium at the end of the month/quarter
based on actual numbers and mix.
Huge caution in the rates, probably means not cost effective for the
manufacturer.
Possibility of profit share but this would mean the manufacturer effectively
carried a significant proportion of the underwriting risk.
See manufacturers customer profile, e.g. mix by age, by postcode to assess the
people who buy their cars.
Say no to protect the company solvency. As the company is small this may be
the only response available.
Ensure the insurer has the right to frequent rate changes so the mix can be
analysed and changes implemented quickly. Need to consider the expense
implications of this frequent analysis and whether the resulting price will be
attractive/ acceptable to the motor manufacturer.
Consider the offers that competitors are likely to put forward, if any others are
prepared to take this business.
100% reinsure the business, cost? Will it be possible?
Write as admin only with the manufacturer keeping the risk perhaps via a
captive reinsurer.
Look at experience of similar schemes, if there are any currently operating in
the market.
If the scheme is already operating, get the current claims information and
adjust in the usual way adding margins as appropriate.
Use consulting actuaries
Have a short term notice in the contractual agreement to get out
Have a clause in the agreement that if taxes / levies are increased then these
are paid by the manufacturer to the insurer
Page 10
EXAMINATIONS

1.
booklet.
2.
3.
supervisor.
4.
5.
question paper.
303S2003
(i)
Define the following terms:

(a)
(b)
long-tailed business
latent claims
[2]
(ii)
Outline the difference between these two terms.
[1]
[Total 3]
You are the senior actuarial student at a large general insurance company writing
mortgage indemnity guarantee business. You have been asked to undertake a review
of the current premium rates for this business.
Outline the factors you would take into account in determining an appropriate
allowance for future investment return when pricing this product.
[6]
Your company has been underwriting commercial property insurance for several
years.
(i)
List the various delays that can occur in relation to the claims your company
receives in respect of this product.
[2]
(ii)
You have observed a recent unanticipated trend of increasing delays at various

stages in the life cycle of a claim. Explain the possible causes of this and the
likely effect upon the ultimate claims cost.
[3]
[Total 5]
(i)
List the 4 main types of insurance cover provided by general insurance

products.
[1]
A shipping company buys an annual hull property damage policy. The policy requires
that a description of the claim is first notified to the insurer at the time of the incident
or when the ship reaches a major port.
If the damage is minor the ship continues with its scheduled voyages until a
convenient time when a full inspection can take place. At this point a more detailed
assessment of the damage is made, the insurer is notified and the repair is planned. If
the damage is minor, the ship may continue its voyages until the repair is finally
carried out and the costs are finalised. If the damage is major then the ship is taken
out of service immediately and if repair is possible it is carried out as soon as possible.
(ii)
303 S20032
Describe the specific claim characteristics of this class of business.
[8]
[Total 9]
You are the actuary for a medium-sized general insurance company writing only
personal motor business. Your Finance Director has asked you to develop a model
that can be used to test the impact on profitability and solvency of changing the
companys reinsurance cover. The existing reinsurance programme has for the last 5
years consisted of a small amount of quota share reinsurance and individual excess of
loss cover of $10,000,000 xs $500,000.
(i)
(ii)
Discuss the factors that should be considered when deciding upon an

appropriate reinsurance programme for this company.
Describe how you would construct your model.
[8]
[10]
[Total 18]
A general insurance company writes only personal lines motor business via brokers.
(i)
Describe the perils covered by personal motor insurance.
[2]
(ii)
List the data required to investigate the appropriateness of the current risk
premium relativities.
[9]
(iii)
(a)
Define burning cost premium.
(b)
Describe how this may be adjusted, and state all the other information
that will be needed, in order to calculate the final office premium. [7]
(iv)
The proposed final office premiums are significantly different from those
currently charged. Explain the risks of implementing the new rating structure.
[7]
(v)
Discuss methods which may be used to mitigate the risks identified in (iv). [7]
[Total 32]
303 S20033
PLEASE TURN OVER
You are a consulting actuary for a large general insurance company that writes
commercial insurance for large multinational companies. You have been asked by
the company to review the methodology and assumptions used by the companys
reserving team for the motor fleet business as at 31 December 2002.
The companys reserving team has analysed non-large and large claims
separately. A claim is described as large if it has exceeded $500,000 at some point in
its history. The non-large claims have been projected using both paid and notified
claim chain ladder methods at a net of reinsurance level. The selected non-large
reserve figures have been based on an average of the results of the two chain ladder
methods. Reserves for large claims have been set to be 110% of the total large claim
case estimates, net of reinsurance, as at 31 December 2002.
The extract below shows historical development factors for the non-large paid and
notified claims triangles as at 31 December 2002, together with the companys
selected development patterns.
"Non-large" Net Paid development factors for cumulative amounts

Accident
Year
1994
1995
1996
1997
1998
1999
2000
2001
Development year
01
12
5.076
1.673
4.408
1.775
5.396
1.768
5.201
1.694
5.409
1.621
5.354
1.772
5.899
1.689
5.238
Selected paid idf

Selected paid cdf
5.500
23.933
1.700
4.351
23
1.357
1.495
1.456
1.411
1.430
1.426
34
1.286
1.294
1.416
1.217
1.369
45
1.189
1.176
1.082
1.223
56
1.161
1.065
1.110
67
0.997
1.057
78
1.031
1.430
2.560
1.300
1.790
1.180
1.377
1.100
1.167
1.040
1.061
1.020
1.020
1.000
1.000
"Non-large" Net Notified development factors for cumulative amounts

Accident
Year
1994
1995
1996
1997
1998
1999
2000
2001
Selected notified idf
Selected notified cdf
Development year
01
12
1.139
1.009
1.169
1.010
1.357
1.076
1.471
1.053
1.418
1.015
1.309
0.973
1.132
1.022
1.318
1.300
1.433
1.025
1.102
idf = incremental development factor

cdf = cumulative development factor
303 S20034
23
1.085
1.089
1.116
1.083
1.023
1.045
34
1.031
1.004
0.979
1.066
0.991
45
0.937
0.980
0.990
0.991
56
0.992
0.992
1.008
67
1.001
0.999
78
0.998
1.070
1.075
1.010
1.005
0.995
0.995
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
(i)
Suggest possible reasons why there are claims development factors less than
1.000 within the above triangles.
[9]
(ii)
(a)
Comment on the suitability of the companys methodology and

assumptions used to establish undiscounted reserves for its motor fleet
account.
(b)
Suggest alternative approaches.
[9]
As part of your work for the company, you have also been asked to assist with the
companys discounting adjustments. The companys accounting policy is to use a
discount rate of 5% per annum.
(iii)
(iv)
303 S20035
Using the companys chain ladder selections and stating any assumptions
made:
(a)
Derive a payment pattern to apply to the 2000 accident year.
(b)
Estimate a discount factor to convert the companys undiscounted

reserves to discounted reserves for the same accident year.
(c)
Derive the discounted mean term to settlement for this accident year.
[7]
List the factors you would consider in selecting a discount rate for reserving
purposes.
[2]
[Total 27]
EXAMINATIONS
September 2003
EXAMINERS REPORT
Introduction
J Curtis
25 November 2003
EXAMINATIONS
September 2003
EXAMINERS REPORT
This question was generally well answered. Each of the two terms examined in this
question can be found in the Core Reading glossary. The distinction between longtailed (long delays from occurrence to notification/settlement) and long term (long
exposure period i.e. long period of cover) is important but is not always appreciated
by students. Most candidates seemed unaware that latent claims could be shorttailed.
(i)
(ii)
(a)
Types of insurance in which a substantial weight of claims take several

years to be notified and/or settled.
.from the date of exposure and/or occurrence.
(b)
Claims resulting from perils or causes of which the insurer is unaware

at the time of writing a policy.
.and for which the potential for claims to be made many years later
has not been appreciated.
Long-tailed relates to known / anticipated features of claims at the time of

writing whereas latent claims are by definition not anticipated at time of
writing.
Latent claims can be short-tailed.
The question focuses on outlining considerations when allowing for investment

returns in pricing mortgage indemnity guarantee business. It does not require a
discussion about which assets would be suitable, although many candidates took this
approach. Points in relation to an appropriate choice of assets are only relevant in
this question to the extent that assets actually held are mismatched to liabilities.
Those candidates who discussed the nature of the liabilities and applied their
bookwork points scored better.
Need to consider likely return on assets backing technical reserves:
consider past investment returns on those assets
and investibility of those assets, including premium payment pattern of lender.
Technical reserves
Term of policy is the term of the mortgage, so premium may be received 25 years
or so before claim is paid.
Most of the risk though is in the early years of the mortgage term, and premium is
usually earned within the first 10 years, although even in those 10 years the risk is
not uniform.and premium earning patterns are not constant.
Delay between claims occurrence and settlement is relatively short so more
concerned with assets backing unearned premium reserve (UPR).
The UPR is very large.
and allowance for investment return is very significant.
Need also to allow for investment returns on assets backing capital:
These will be particularly significant.
.as MIG has relatively large capital allocation due to uncertainty of writing this
business.
Page 3
Allow for mismatching of assets and liabilities:

This is particularly important.
.as will be very difficult to determine future economic/property cycles and
hence claims costs.
Examples of mismatch are due to currency and legislative/solvency changes.
Downwards adjustment to investment return assumption may be high.
Other points to consider
Consistency with other assumptions eg inflation.
Regulatory / statutory restrictions on pricing methods/assumptions.
Investment return should be net of expenses.
Investment return should be net of tax.
Take into account likely future tax changes given long term nature of MIG.
Candidates scored reasonably well on this largely bookwork question, though there
were many more points available than required for full marks. Candidates did not
demonstrate a full understanding in part (ii) on how delays might impact the ultimate
claims cost.
(i)
Identification of damage / loss, manifestation

Reporting of damage / loss
Delays in processing claim
Assessment of extent of damage / loss
Monitoring / delays due to business interruption
Verification / agreement on extent
Processing payment / settlement delay
Recovery of reinsurance
Recovery of salvage
Closing of claim file
Reopening / resettlement
(ii)
Possible causes:
Change of brokers.
Change of claims handlers affects efficiency levels and delays.
Change in types of claim as different types of claim will have different
features affecting handling time.
Change in mix of business or cover eg increased excesses so change in
size of claim.
Surveyors assessing the cost of claims take longer to assess.
External factors such as postal strikes, legislation, judiciary, economic.
Internal factors eg staff shortage, system breakdown.
Likely effect on ultimate claims cost:
Claims may be turned down if reporting delay period extends dramatically
(depends on policy wording) so reducing overall cost.
Delays to repairs will result in increases in average settlements due to
inflation so increasing overall ultimate cost.
Page 4
Changes in delay will affect amount of investment return achieved, which

will affect discounted value of ultimate claims cost.
Although not required to gain full marks, some credit was given for the following
additional points:
If reinsurers default (infinite delay) then the net ultimate cost will incr
ease.
Claims handling costs may increase if more time is spent dealing with
claims.
For part (i), the majority of candidates know the four insurance cover types, which
appear in the Core Reading. However, candidates performance on part (ii) were
disappointing given the importance in general insurance of understanding the
characteristics of different claim types. Most noticeable was the lack of breadth and
detail in candidates answers, with many candidates confining themselves mainly to a
discussion of the delays in settlement and notification. It should have been clear to
candidates producing short solutions that they were too brief to score the full 8 marks
on offer.
(i)
Liability
Property Damage
Financial Loss
Fixed Benefits
(ii)
Claims origination:
sudden cause of loss
cause easily determinable
Claims notification:
usually notified quickly
biggest claims notified very quickly
although minor knocks may be delayed
Claims settlement / payment:
settled over time as repair scheduled
small claims can take a long time to settle
most claims settled within 3 years.
.longer than most property damage type claims
.but still relatively short-tailed
settlement delays due to disputed liability
Claims amount:
repairs will be subject to claims inflation
reported amount can rise when inspection reveals more damage
actual settlement cost can be quite different from estimate
policy excess may result in nil claims
salvage recoveries will lead to reductions in paid claims
Page 5
currency fluctuations may cause amounts to differ from expectations

damage exacerbated by continuation of journey
moral hazard due to false claims
lack of independence of claims assessor
Potential for accumulations / catastrophes

geographic concentration and exposure to storm, harbour fire
Claims distribution
low frequency (in ship years)
occasional total loss (actual/constructive)
skewed claims distribution
Well prepared candidates scored well on part (i),which was essentially bookwork.
Some candidates appeared to ignore the details given in the question (eg one class,
motor, medium-sized company) and lost marks for application. It was clear from
scripts that some students were confused about how risk XL and aggregate XL
contracts operate. In part (ii) the quality of candidates answers was variable. The
stronger candidates were able to build upon the Core Readings list of steps required
to build a model and thought about the specific requirements in this situation.
(i)
Page 6
Stability of profits.
.More excess of loss protection (i.e. lower excess point) may result in more
stable results
....stability of profits will affect ability to pay stable dividends (which
shareholders may prefer)
.particularly relevant as company writes only one class
Nature of inwards claims with respect to frequency and amount.
.several small claims but a few individual large losses on bodily injury
claims
Consider management/shareholder attitude to risk
Size of free reserves: to what extent can the free reserves withstand adverse
large loss experience? A couple of large bodily injury claims could cause
problems.
Company strategy: is the company expecting to expand its business? How
much of a strain will this place on the free reserves?
Potential for accumulations
.is there too much exposure in one geographical area?
.individual excess of loss will not address this
.may decide need more quota share in order to write a wider range of risks
but maintain similar levels of net exposure, or aggregate XL
Statutory solvency: how will changing the reinsurance protection impact the
statutory solvency position?
Technical assistance: does the company benefit from technical help from
existing reinsurers? How will a change in the reinsurance programme affect
this arrangement?
Relationship with existing reinsurers
Regulatory requirements to hold reinsurance
Market reputation: how will investors, analysts, brokers, customers react to

any significant change in reinsurance programme?
Security status: reinsurers with better security may charge more for the cover
Value for money
(ii)
Type of model/high level points:

Prefer to use a stochastic model but could use deterministic with lots of
sensitivity testing.
Construct a model that projects cashflows over a 5 to 10 year period.
Projections need to be realistic so assumptions must be on best estimate
basis.
Include new business
Define objectives:
By varying the retention level, determine the effect of the excess of loss
reinsurance on profitability and statutory solvency.
Need to define profitability and solvency eg statutory definition?
Projections need to be realistic so assumptions must be on best estimate basis.
Collect claims data:
Determine future gross claims expected therefore need gross and
reinsurance data separately.
Use individual claims data.
Collect and use individual claims data for last 5 10 years, gross of
reinsurance, grouping by risk categories
Group data by risk categories eg bodily injury vs property damage
Makes adjustments for:
Claims inflation
Cover provided
Deductibles
External factors
Changes to claims handling or settlement
IBNR by standard methods such as chain ladder or average cost per claim
Choose variables to be modelled stochastically
Fit separate claims frequency and amount distributions
So can calculate individual recovery amounts
Check goodness of fit (can apply this to frequency/amount distributions
and/or combined distributions)
Derive a combined claims distribution function
by simulating separate claims frequency and amount distributions.
Also model:
Gross premiums
Page 7
Reinsurance premiums investigate likely reinsurance market premiums

for XL cover at various excess points
Reinsurance recoveries - compare these reinsurance premiums against
modelled recoveries expected to be paid at different lower excess limits
Compare these reinsurance premiums against modelled recoveries
expected to be paid at different lower excess limits.
Expense cashflows
Investment cashflows
Specify correlation between variables

e.g. allow investment strategy to react to claims experience and level of
free reserves
For each lower excess limit, run large number e.g. 1000 (say) simulations,
projecting the cashflows of the model over the period.
Construct model so as to produce output relating to solvency and profitability.
Part (i) was straightforward with the perils listed in Core Reading. Candidates
knowledge of the bookwork in part (ii) was disappointing, with many candidates
failing to provide sufficient items. Partial credit was given to solutions that
mentioned rating factors without listing them, although the number of marks
available should have prompted more detail. A surprising number of candidates did
not provide a clear definition of burning cost premium, although most were able to
talk about adjustments to it. Noticeably, many candidates omitted to mention that the
burning cost premium would need to be adjusted for IBNR claims. Most candidates
appeared uncomfortable in dealing with the non-standard questions in part (iv) and
(v) and did not generate sufficient ideas.
(i)
Perils covered:
accidental damage to the insured vehicle
malicious damage to the insured vehicle
fire of insured vehicle
theft of insured vehicle
If the policy includes motor third party liability then the
insured is indemnified against compensation payable to a third party for
personal injury
or damage to their property.
Although not part of Core Reading, credit was given for the following point:
Breakdown where roadside assistance included as part of cover
(ii)
Page 8
Policy data as it was at the date the claim occurred:

policy number
start date
end date
endorsement date
policy term
type of cover
date of birth / age of drivers
gender of drivers
additional drivers
licence type
years held licence
risk address / postcode
broker code
details of previous accidents and convictions
claims history
security devices / safety devices
profession
marital status
transmission type
smoker / non-smoker (proxy for lifestyle)
vehicle group rating or other classification
make of car
model of car
engine size
fuel type
body type
modifications to vehicle
annual mileage
policy limits
excess compulsory and voluntary
use of vehicle e.g. private / business
where vehicle kept overnight
age of vehicle
value of vehicle
NCD
payment method
where insured to drive
policy number / link to claims data
Claims data:
claim number / link to policy data
accident date
notification date
claim amounts paid
dates claims amounts paid
claim recoveries made
dates recoveries made
estimated outstanding
dates the estimates were made
type of claim, TPBI, theft etc
claim description code e.g. collision with stationary vehicle
claims expenses e.g. loss adjusters fees, NHS charges
Page 9
claim event number, to link claims from the same event for reinsurance
recoveries
settled indicator, settled, outstanding, re-opened
driver at time of accident
currency
Current rating structure

Changes in policy conditions
Results of portfolio movement analysis
Risk premium relativities implied by relativities of competitors' office
premiums
Reinsurers opinions
Underwriters opinions
Changes in legislation on allowable rating factors
Future inflation assumption
(iii)
(a)
The actual cost of claims incurred during a past period as an annual

rate per unit of exposure
(b)
Include allowance for IBNR

Adjust for unusually heavy/light experience, build in an average
amount of bodily injury claims
Allow for trends in claims, e.g. greater recovery of health costs from
insurers by the government
Adjust for legislative / judicial changes
Allow for changes in risk and cover
Allow for change in mix of business eg new/renewal
Allow for estimated future claims inflation
Different expense / commission assumptions:

Underwriting expenses
Claims handling expenses split fixed variable
By new business and renewals
Allow for expense inflation
Allow for any discounts given
Commission if payable
Profit loading
Contingency margins
Tax
Reinsurance cost allocated to a policy
Levies
Investment return
Competitive issues, insurance cycle
(iv)
Page 10
New rates are higher:

Customers complain, expense impact in dealing with this.
Customers dont renew, expenses impact with the operation of fixed expenses.
New business drops, fixed expenses not spread.

Bad PR which further reduces new business and retention rates.
New rates are lower:
Lots of new business, financial impact of new business strain, cant get
enough staff, premises etc to cope with the influx of new business.
Existing customers feel that they have been overcharged in the past and move
their business elsewhere.
Brokers are unhappy with the resultant reduced commission.
Competitors do not react as expected.
New rates are differently structured.
Cost of implementation may exceed expectations.
Reinsurers not happy with the new rates.
Risk of anti-selection if the rates are different to those offered by competitors
in the market.
May write volumes of business where little previous experience and rates
uncetain.
(v)
Easiest method is not to implement the new rating structure. Could then move
the old one to look like the new one but gradually.
Renewal
Manage price offered to existing customers, for example no premium
increases or decreases by more than 10%. May mean that prices on renewal
will be different to those for new business.
Run the renewals as a closed book retaining the current pricing structure and
use the new rates for new business only. Again this will mean that the rates
will be different for new business and renewals which could lead to customer
complaints.
Give the customers an incentive to renew to soften the blow of price increases
e.g. free mobile phone / road map / holiday vouchers.
Rather than reduce premiums significantly, offer loyalty discounts.
New Business
Test the prices for a short period, measure the effect on volumes, conversion
rates and decide whether to fully implement.
Move gradually to the new prices.
Communicate to brokers, customers and staff the expected changes as a result
of the implementation.
Check if the levels of increase /decrease seen will be different to those being
applied by competitors.
Offer brokers profit commission.
Offer brokers volume commission.
Increase advertising / marketing campaign.
Spread distribution channels (not just broker).
Arrange more reinsurance / get technical assistance.
If all prices have increased then another option is to reduce the cover instead
of increasing the prices. The effectiveness of this depends on what is already
included in the cover that can be removed without making the product stand
Page 11
out in the market.
Page 12
Some candidates gave very strong answers in parts (i) and (ii) but in general the
examiners were disappointed with the lack of understanding that was shown in many
scripts. The examiners gave credit for alternative solutions in part (iii) where the
question had not specified the as at date for the discount factor. Although 31
December 2002 was the most logical choice, calculations as at 31 December 2000
were also accepted. The ability to produce payment patterns has been examined a
few times in recent years and yet candidates seemed to be very poorly prepared for
these calculations. Part (iv) was bookwork and well answered.
(i)
Paid triangles:
Factors less than one may appear in the paid triangles when the
incremental net paid for a particular accident year and development
year is negative.
This in turn may occur for the following reasons:
Data error some of the underlying records may have been recorded
incorrectly or the triangle may have been constructed incorrectly.
Extraction of large claims only if the history of a large claim isnt
extracted fully from the non-large triangle.
Currency if the underlying currencies have not been converted at
current exchange rates, some of the movements in the triangle may be due
to currency fluctuations.
Third party recoveries the insurance company may have been able to
recover monies from another insurer or via salvage from the insured.
Reinsurance lag there may be a time delay between payment of a gross
claim and a subsequent reinsurance recovery.
Policyholder repayment of amounts subsequently decided not covered eg
due to fraud.
There may be some premium adjustments due to experience rating, and it
is possible, although unlikely, that these may included incorrectly within
affect the paid triangle.
Incurred triangles:
Factors less than one may appear in the incurred triangles when the
incremental net incurred for a particular accident year and development year is
negative.
This is more common than with paid developments.
This in turn may occur for any of the above reasons plus the following:
Large claim settled at less than case reserve, outweighing positive incurred
claim movement on other claims.
Case reserving strength changes so that there is a one-off decrease in case
reserves, and this outweighs paid movements in that year.
Generally conservative case reserving philosophy so that negative incurred
but not enough reported (IBNER) claims development outweighs paid
claims.
(ii)
Comments on methodology and assumptions

Selected factors look reasonable.
.except no paid tail factor and paid development unlikely to stop dead.
Page 13
Incurred pattern reflects run-off savings => case reserving is prudent.

Using paid and incurred information good to use both but paid factors
to ultimate very high on latest accident years so paid chain ladder less
reliable.
Would expect more weight to incurred methods on later accident years as
incurred has more information and factors to ultimate are lower.
Large losses have been treated separately - good if distort factors in
triangle ie if large claims develop faster or slower than non-large.
Large losses in danger of being over/under projected if remain in triangle.
Large losses likely to be complex bodily injury claims therefore probably
longer-tailed on average than non-large claims.
If remove large losses from triangle won't project emergence of new large
losses.
Is $500,000 an appropriate level? Large loss definition of $500,000 is
being used on all accident/devt years. Inflation means that company will
effectively be removing more losses from the non-large triangle on the
2002 row than the 1995 row.
Large losses will be bodily injury claims case estimates use skill of
claims assessors but may not be consistent from year to year
Company has included an additional 10% of large loss case reserves. Is
this amount appropriate. This may be held to bolster case reserves if
believed to be insufficient or it may be held as an allowance for new large
losses to emerge.
Suggest use Bornhuetter-Ferguson or expected loss ratio method on 2002
accident year as a minimum
Suggested alternative approaches

Establish paid tailobtain further historical data, benchmark or do curve
fitting / fit tail by eye.
Apply more weight to incurred methods on later accident years.
Use Bornhuetter-Ferguson or expected loss ratio methods on the 2002
accident year where the factor to ultimate is high.
Should project gross then allow for reinsurance separately to avoid
distortions due to changes in reinsurance programme.
Project by country to avoid currency distortions.
Split into homogeneous groups (eg BI vs PD, Large fleets vs small fleets).
Test 10% allowance for large IBNER/IBNR by looking at historical claims
developments for individual large claims or compile triangle of large
claims.
Adjust large claim definition by year to allow for inflation.
Perform ACPC or inflation-adjusted ACPC for large &/or non-large losses.
(iii)
Page 14
(a)
Use paid pattern
[some candidates incorrectly used the incurred pattern or the average of the
paid and incurred]
Assume no tail factor required or make suitable adjustment
Development year
Use paid cdf (=ult/pd) 23.933
4.351
2.560
1.790
1.377
1.167
1.061
1.020
1.000
Calculate 1/cdf
(=pd/ult)
Cum % developed
4.18% 22.98% 39.07% 55.87% 72.63% 85.70% 94.27% 98.04% 100.00%
Subtract adjacent %s
Incr % developed
4.18% 18.80% 16.09% 16.80% 16.76% 13.07% 8.57% 3.77%
1.96%
27.57% 27.51% 21.45% 14.06% 6.19%
3.22%
For 2000 accident year*
27.57% in the above row = (100/(100 4.1818.816.09))*16.8%

[As at 31 December 2002, the 2000 accident year has reached the end of
development year 2. Need to normalise pattern so that percentages add up to
100% and can therefore apply pattern to reserves.]
(b)
Assume that payments occur on average mid-way through the year

Calculate xtv(t ) for each future time period where v = (1 + 0.05)(1)
and xt = % payable in year t:
Development year
0
1
2
3
4
5
6
7
8
t
1
2
3
4
5
6
xt
27.57% 27.51% 21.45% 14.06% 6.19% 3.22%
(t
)
0.9759 0.9294 0.8852 0.8430 0.8029 0.7646
v
(t
)
xtv
0.26905 0.25564 0.18991 0.11857 0.04969 0.02461
Discount factor = Discounted reserves / Undiscounted reserves

= Sum of xtv(t ) / Sum of xt
= 0.90746
(c)
Discounted mean term defined by:

Discount factor = (1 + i)(DMT)
Take logs of both sides:
Log(0.907458) = DMT*log(1.05) => DMT = 1.99 years
[Also possible to use accurate formula: Sum{(t-1/2)xtv(t-1/2)}/Sum{xtv(t-/2)}

but this would be more time consuming. Answer in that case would be 1.99
years]
(iv)
Legislation/supervisory regulations
Investment return earned on assets covering technical reserves.
.allowing for non-investible assets (e.g. broker/reinsurer balances)
Purpose of reserving / degree of prudence required
Extent to which assets are mismatched to liabilities / uncertainty
Rate used previously for consistency
Page 15
Allowance for tax

Allowance for investments expenses.
Page 16
EXAMINATIONS
26 April 2004 (am)
Subject 303
General Insurance

1.
booklet.
2.
3.
supervisor.
4.
5.

question paper.
303
A2004
State the reasons why a general insurance company might calculate IBNR reserves
separately from other claims reserves.
[2]
A general insurance company that writes commercial property insurance business

wants to purchase excess of loss reinsurance protection in order to reduce the overall
variability of its claims experience. You have been asked to recommend an
appropriate reinsurance structure (limit, deductible, number of reinstatements etc).
You decide to build a model of the likely claims experience and reinsurance
recoveries.
(i)
State the functionality that you would require of this model.
[3]
(ii)
State the other factors that the insurance company needs to consider before
deciding to purchase the reinsurance.
[5]
[Total 8]
You are the actuary for a general insurance company that writes motor fleet business.
One of the options available to the insured is for it to take a voluntary aggregate
deductible on its policy.
(i)
(a)
(b)
Define what is meant by a deductible.

Define what is meant by an aggregate deductible.
[2]
A particular insured currently pays 100,000 a year for insurance cover with no
deductibles of any kind, and would like to consider the impact on its premium of
adopting a 10,000 annual aggregate deductible. It has been suggested that each 1 of
aggregate deductible should reduce the premium by 1.
(ii)
Discuss, with reasons, whether the suggestion is correct, or whether the true
reduction would be higher or lower than this (assuming that any loadings to
the risk premium remain unchanged in monetary terms).
[6]
[Total 8]
(a)
State the ways in which a general insurance company can increase its gross
written premium.
(b)
State the constraints that it faces in doing so.
(i)
State what is meant by a soft market and explain how it occurs and persists.
[10]
[4]
(ii)
303 A2004
Describe the risks to a general insurance company of under-estimating the

softness of a market.
[6]
[Total 10]
A reinsurance company wants to introduce a new type of aggregate excess of loss

treaty covering subsidence claims on household insurance business. The treaty will
cover all subsidence claims notified, during a calendar year, to an insurer. The treaty
will be subject to an aggregate deductible and an aggregate limit.
(i)
State the characteristics of subsidence claims likely to be experienced by the

insurer.
[5]
(ii)
State the advantages and disadvantages to the reinsurer of the proposal.
(iii)
(a)
List the information that the reinsurance company would request from
the insurer if it were trying to price this treaty.
(b)
List the information likely to be required from other than the insurer in
order to price this treaty.
[8]
[Total 19]
[6]
You have been provided with the following financial information in respect of a
general insurance company.
Gross Earned Premium

Reinsurance Earned Premium
Net Earned Premium
Gross Claims Incurred
Reinsurance Claims Recoverable
Net Claims Incurred
Net (of reinsurance) Commission
Net Assets
Management Expenses
(i)
(ii)
(a)
Define the terms gross loss ratio, net loss ratio, expense ratio,
combined ratio and solvency ratio.
(b)
Calculate these ratios for each of the five years of data.
[7]
Describe the changes in the company finances from year to year in the
following areas calculating any additional statistics which support your
analysis.
(a)
(b)
(c)
(d)
(iii)
303 A2004
1999 2000 2001 2002 2003 Total

66
71 187 165 220 709
12
19 103
72
76 282
54
52
84
93 144 427
29
51 210 249 148 687
1
31 155 138
14 339
28
20
55 111 134 348
10
11
32
29
35 117
53
32
91 112 121 409
5
4
12
14
11
46
The efficiency of the management at controlling expenses.

Growth.
Reinsurance purchasing.
Underwriting profitability.
Suggest where the management focus should be in the coming year.
[10]
[2]
[Total 19]
PLEASE TURN OVER
You are advising a consortium of insurance professionals who wish to set up a new
reinsurance company. The business plan is to write international liability reinsurance
on a non-proportional basis. It is intended to reinsure professional indemnity covers
and employers liability covers. The company will be able to accept business from any
territory.
You have been asked to assist in deriving appropriate rating structures for the
company to meet its targets. The financial backers require a return on capital of 40%
per annum on average. The company wants to write 100m of premium income for
each of the two types of cover.
You have already calculated risk premiums for the two different lines of business.
(i)
List the adjustments that should be made to the risk premium to generate the
office premium.
[2]
(ii)
Describe the ideal requirements of an actuarial model.
[4]
The company wishes to use a stochastic actuarial model to allocate capital between
the two lines of business such that the probability of ruin is the same for each line of
business.
(iii)
Describe the steps involved in using a stochastic model to derive this common
probability of ruin.
[5]
(iv)
State the advantages and disadvantages of a stochastic model over a

deterministic model in this situation.
[4]
Using a stochastic model, you determine that professional indemnity reinsurance is a

more volatile class of business, and requires 50% more capital to support every 1 of
premium underwritten than employers liability reinsurance. The business plan shows
that the profit available (after allowing for investment returns) on employers liability
reinsurance is 12% of the premiums written, and it is intended that all lines of
business produce the same return on capital.
(v)
Calculate the total capital required for the business based on the business plan
if the financial backers target is to be achieved.
[4]
(vi)
Describe the other factors that the consortium should consider before deciding
upon a suitable amount of capital for this new company.
[5]
[Total 24]
END OF PAPER
303 A2004
EXAMINATIONS
April 2004
Subject 303
General Insurance
EXAMINERS REPORT
Subject 303 (General Insurance)
April 2004
Examiners Report
This question was generally well answered with many candidates scoring the majority
of the marks
The calculation method used for other claims reserves may not produce a result
including IBNR
The combined reported and non-reported claims in the reserve analysis process does
not enable a true assessment of the cost of claims.
Feedback into pricing enabling the true assessment of the cost of claims.
Feedback into the business control cycle.
To assist in asset / liability matching
Statutory reporting requirements may insist on the separate quantification and
reporting of IBNR claims reserves.
To allow accurate comparison of the profitability of different lines of business.
This question was well answered by most candidates.

(i)
The model must be capable producing claim distributions both gross and net
of a specified reinsurance structure.
Enable the user to test various reinsurance structures.
Must accurately describe the insurer s CPI book, present and prospective.
The model should run quickly and reliably with communicable answers,
encompasing key deterministic variables.
(ii)
Cost and availability of the various reinsurance structures vs their likely

benefit.
Security Status of the available reinsurers.
Size of free reserves.
Opportunities to find co-insurers.
Any constraints placed by regulators and rating agencies.
Impact on solvency requirements from regulators or rating agencies.
Company s attitude to risk.
Return on capital constraints.
How does the chosen reinsurance structure compare with competitors.
Size of book versus rest of business.
Long term relationship with insurers.
Alternatives to XOL reinsurance
Extent of in-house experience in this line.
Services available from reinsurers.
Most candidates could define what was meant by a deductible, however many
candidates struggled with the second part. Several candidates did not state whether
the suggestion was correct or not but talked around the possible reasons. Many
candidates mentioned fixed expenses in their solutions and sensible comments were
awarded marks accordingly.
(i)
Page 2
A deductible is the amount which, in accordance with the terms of the policy,
is deducted from the claim amount that would otherwise have been payable
and will therefore be borne by the policyholder.
April 2004
Examiners Report
The aggregate deductible applies across all claims in total. It is the total
amount of all claims borne by the policyholder before the insurance company
becomes liable.
(ii)
This is incorrect.
If true, in an extreme example
100,000 deductible implies 0 premium,
without removing all risk
The mean of the aggregate distribution is the risk premium
One pound of deductible reduces the premium by the expected value of claims
less than or equal to that pound
Assuming probability of claims exceeding the deductible is not 1
The reduction in risk premium would be less than the deductible
There were many possible points to make in this question, but some of the answers
given by candidates did not demonstrate to the examiners their understanding,
especially when it came to stating the constraints. Most candidates nevertheless
managed to score reasonable marks on this question.
Charge higher premiums for the same exposure.
Take on more exposure:
By reducing premiums where elasticity of demand is greater than 1
By expand into new classes of business (including inwards reinsurance and
coinsurance)
By expanding into new areas of risk within existing classes
By writing larger risks within existing classes
By development in new countries
By acquire another Insurance company
Expand into new distribution channels e.g. Direct marketing, internet
Increase advertising / marketing / customer services
Incentives for customers e.g. alarm clock or pen
Increase sales force
Increase commission levels offered to agents / brokers
Increase capital in order to obtain a higher credit rating which is more attractive to
potential clients.
Constrains:
Level of free reserves. If it grows too much it may not have enough capital to support
its business.
Competition from other insurers and self retention if it tries to increase its premiums
too much.
There may be regulatory constraints on:
premiums that it can charge.
authorisation to write lines of business / countries
distribution channels that can be used
minimum solvency ratio
It may not have the expertise to write new classes of business
It may not have the staff / infrastructure to expand quickly.
Page 3
April 2004
Examiners Report
Acquisition of another company may be too costly.

Expanding into new distribution channels may alienate existing distribution channels.
Brokers may resist reduced premium as this implies reduced brokerage
Company s reputation may inhibit growth
Cashflow
New salesforce may be costly / difficult to recruit and train
Whilst the examiners had expected the solution as given in part (ii), it was clear to the
examiners in their post exam meeting that the term under-estimating could
have been interpreted as meaning either that premiums were higher or lower than the
market. Hence marks were also awarded for the alternative explanation
The main problem encountered seemed to be that candidates did not know the
difference between a soft market and a hard market.
(i)
It is caused by over-supply of insurance capacity in certain products.

Soft market is a term describing that part of the insurance cycle when business
is least profitable.
i.e. inadequate premiums being paid for the amount of risk assumed.
Its longevity and depth result from the true effects of
claims inflation,
broadening wordings cover
and increases in exposure
being under-estimated at the time of writing business.
Persists because premium increases lag behind claims notifications
(ii)
Risks
Write a larger volume of unprofitable business than planned leading to
Over statement of profits and under statement of reserves
managers not aware of true financial condition
leading to bad business decisions.
Risk of writing larger volumes of business than planned and not being able to
service it.
There is a risk that the true profitability of business written is not known for
some time
and the company writes business at less profitable rates than it expects.
Leading to a significant weakening of the financial condition of the company.
When the extent of this position is realised reserves must be increased
(sometimes dramatically)
leading to a reduced solvency position,
requiring additional capital support and/or other remedial action.
This reduced solvency may directly impact the quantity of profitable new
business that the company can write
and in the extreme may lead to the insolvency of the company.
Page 4
April 2004
Examiners Report
Many candidates did not demonstrate sufficient knowledge to perform very well in
this question. This was the most difficult question on the paper and certainly
sorted out those that could think around GI issues and those who struggled. In the
majority of cases there were just insufficient valid points made.
(i)
The frequency and average cost can be related to the economy (Housing
market) i.e. there is a strong dependency on property selling as many
subsidence claims arise following surveys for buyers.
Such claims are also related to several other events, e.g. the weather
(unusually dry spells), closeness of trees, pipes and soil conditions.
Long notification and settlement delays.

Low frequency and high severity
Can have high levels of reopened claims.
Claim size distribution can be highly skewed.
There can be significant accumulations of claim occurrences and notifications
during particular periods.
Claim costs subject to inflation.
Date of occurrence is usually uncertain
This may lead to disputes about coverage especially when there have been
changes in the insurance company providing the cover.
Need for individual loss assessment by a claims adjuster
Active claim management can significantly reduce the cost of a claim.
(ii)
Reinsurer Advantages:
New product type and hence it may attract significant new business.
Possible diversification away from its other products.
May be highly profitable if there is very little competition.
May enable it to attract more traditional business from its cedants. (Cross
selling).
Aggregate limit and deductible cap liability.
Claims notified wording removes problem of late notification
Such a product would be very attractive to insurers as it would enable them to
price their own policies with more certainty
Reinsurer Disadvantages:
Lack of data to price this business and possibly requirement for more capital.
Risk of anti selection. If there are long lags between adverse weather and
subsidence claim notifications, Insurers may be able to select against the
reinsurer.
Risk of anti selection. Exhaustion of aggregate deductible removes insurers
financial incentive to monitor claims.
How do you deal with reopened claims? They could just as easily be opened
as new claims.
Long settlement delays mean that the ultimate losses from the policy will not
be known for a long time. If mispriced initially, then this could be very costly.
Claim inflation is likely to be significant due to long settlement delays.
Page 5
April 2004
Examiners Report
Claim experience is potentially linked to the economic / housing cycle. This

can be very difficult to price for.
(iii)
(a)
Data on past years experience.

Notifications and ultimate losses by calendar year.
Listing of all subsidence claims showing details of:
Notification date.
Payments (dates and amounts)
Changes in outstanding case estimates (dates and amounts)
Additional details on the claims e.g. geographical area, likely
cause etc.
In force policy counts, premiums and sum insured by geographic
region. In order to check exposure to high risk areas.
Expected policy count, premiums and sum insured for the exposure
period by geographical region.
Clarification from the insurer on how it handles reopened claims.
Information about changes in excess levels and any changes planned
for the future.
Insurer s policy conditions
Insurer s underwriting philosophy
Expertise in handling claims
(b)
Industry data on claim notifications and claim costs.

Externally available models
Own claims experience of subsidence
Judicial findings and market practice
Economic data on Housing Market.
Building cost Index as a proxy for claim cost inflation.
Climate and weather data e.g. Rainfall data, Soil moisture and
composition data etc.
Applicable Tax rates
Expense loadings (including brokerage, commission and management
expenses).
Required Profit loadings
Investment yields
Cost of capital
Contingencies
Retrocession costs (if applicable)
Competitors prices (if available)
This question was generally well answered, however there were some points on which
candidates demonstrated that they did not understand accounts for a general
insurance company and important statistics.
In the calculation of the solvency ratio in part (i) candidates were expected to make a
reasonable assumption on how to obtain written premium from the data. Whilst not
particularly accurate the assumption of net written being the same as net earned was
accepted as in the solution below.
Page 6
(i)
April 2004
Examiners Report
Gross Loss Ratio = The ratio of the cost of claims to premiums, both gross of
reinsurance.
Net Loss Ratio = The ratio of the cost of claims to premiums, both net of
reinsurance.
Can be on an earned basis or a written basis.
Expense Ratio = The ratio of management expenses plus commission to
premium
Combined Ratio = The sum of the loss ratio (claim ratio) and the expense ratio
Solvency Ratio = The free reserves divided by the net (of reinsurance) written
premiums.
Gross Loss Ratio

Net Loss Ratio
Expense Ratio
Combined Ratio
Solvency Ratio
(ii)
1999
43.9%
51.9%
27.8%
79.6%
98.1%
2000
2001
2002
2003
Total
71.8% 112.3% 150.9% 67.3% 96.9%
38.5% 65.5% 119.4% 93.1% 81.5%
28.8% 52.4% 46.2% 31.9% 38.2%
67.3% 117.9% 165.6% 125.0% 119.7%
61.5% 108.3% 120.4% 84.0% 95.8%
(a)
Management Expense Ratio
Expense Ratio as calculated
above
9.3%
7.7% 14.3% 15.1%
7.6%
27.8% 28.8% 52.4% 46.2% 31.9%
Management expenses rise dramatically as the company expands in

2001 and 2002, almost doubling from the 2000 expense level.
Management expenses are brought under control in 2003
Similarly Expense Ratio increases dramatically in 2001.
(b)
Growth
GEP Growth
NEP Growth
7.6%
3.7%
163.4%
61.5%
11.8%
10.7%
33.3%
54.8%
Growth in Gross premiums of 163% in 2001 is dramatic, similarly the

33% growth in 2003 is large.
Large changes like this may be a result of acquiring another company.
Net Premium change is less volatile than the gross premium change
The large increase into 2001 follows the increase in gross premiums.
The large increase in 2003 is due to reduced reinsurance purchasing as
a percentage of gross income
(c)
Reinsurance purchasing
Page 7
Ceded Premium change

RI Cost = RI Prem
RI Recoveries
Ri Loss Ratio
April 2004
Examiners Report
18.2% 26.8% 55.1% 43.6% 34.5%

11
8%
(12)
163%
(52)
150%
(66)
192%
62
18%
Reinsurance loss ratio, and hence recoveries were low when the
company was small.
Ceded premiums have increased significantly with the growth in gross
premiums
Reinsurance spending has been good over the period with recoveries
exceeding premiums.
(d)
Underwriting profitability
Net UW Profit
11
17 (15)
(61)
(36)
Net profitability in the last three years has been negative

Although negative, the underwriting profit has improved in 2003.
(iii)
Many candidates scored very high marks on this question. Most of the solution was
straightforward bookwork. Several candidates did however seem to struggle with the
calculation part (v).
(i)
Page 8
Comment: After a period of rapid growth it would be sensible for the

company to consolidate.
Expenses have been reduced dramatically in the last year and these need to be
kept low
reducing them if possible.
Reinsurance purchasing has reduced probably as prices have risen due to the
high level of recoveries in the middle years.
A cost benefit analysis of reinsurance spending should be performed.
Effort needs to put into gross and net underwriting profits.
Expected investment return credited

Commission
Fixed expenses
Other variable expenses
Cost of retrocession
Profit
Adjustment for competition
Contingencies
Tax
Cost of capital
April 2004
Examiners Report
(ii)
the model being used should be valid, complete and adequately
documented
the model chosen should reflect adequately the risk profile of the classes of
business being modelled
the parameter values used should be accurate for the classes of business
being modelled
the outputs from the model and the degree of uncertainty surrounding them
should be capable of independent verification for reasonableness
The model, however, must not be overly complex so that either
The results become difficult to interpret and communicate
The model becomes too long or expensive to run.
(iii)
Definition of ruin
collect data
group and modify data
choose a suitable density function for each of the variables to be modelled
stochastically
specify correlations between variables
estimate the required parameters for the chosen density function(s)
check the goodness of fit is/are acceptable and attempt a fit with different
density function(s) if it is not
construct a model based on the chosen density function(s)
run the model many times, each time using a random sample from the
chosen density function(s)
count the number of scenarios in which each of the lines of business
produces ruin for the company
produce a summary of the results that shows the distribution of the
modelled results after many simulations have been run
perform sensitivity tests on results
Calculate the capital allocation that equalises the ruin probability for each
class
(iv)
Advantages
Captures interdependency of variables and general volatility better
Capital required for infrequent events
need concept of probability
Provides additional output to management
e.g. reinsurance purchasing
Best practice/competition
Deterministic model cannot do probability of ruin
Disadvantages
More complex and time consuming
Danger of spurious accuracy
(v)
Capital for EL = y
Page 9
April 2004
Examiners Report
Capital for PI = 3y/2

Total capital = 5y/2
Required return on capital = 40%
Required profit = 40% 5y/2 = y
Profit for EL = 12m
Same return on capital means profit for PI = 12m*3/2 = 18m
Total profit = 30m
Therefore, y = 30m
Total capital = 75m
(vi)
There may be a regulatory minimum level of solvency which this may not
reach
Need to consider how this compares to competitors, and the implication on
market confidence
Opportunity cost of capital may be better utilised elsewhere
No track record, therefore
The credit rating of the company will depend on the capital available
a lower credit rating will reduce the ability of the company to sell the
reinsurance
The availability of capital may impact on the company s ability to raise any
more
Market may not be able to sustain sufficient premium to produce desired
returns on a higher capital base
Quality of people affects ability to raise capital
END OF EXAMINERS REPORT
Page 10
EXAMINATIONS
Subject 303
General Insurance

1.
booklet.
2.
3.
supervisor.
4.
5.

question paper.
303
S2004
(i)
List the types of generic policy and claims data that would be required for a
premium rating exercise.
[5]
(ii)
Give examples of the specific rating factors that should be considered when
pricing:
(a)
(b)
personal medical expenses

employers liability insurance
[4]
[Total 9]
A general insurance company expects to write business at a loss ratio of 70% and
incurs expenses of 10% commission and 10% other management expenses. It pays
taxes of 30% of profits. It also pays out 10% of its post tax profits to shareholders in
the form of a dividend. The insurance regulator requires it to demonstrate a minimum
solvency ratio of 50% at the end of the year. Its current solvency ratio is 50%.
(i)
Calculate the maximum growth rate that it can sustain without recourse to
additional capital, stating any assumptions that you make.
[4]
(ii)
Suggest ways in which this insurer might be able to grow more quickly than
the rate calculated in part (i).
[5]
[Total 9]
State the advantages and disadvantages to a large multinational company of setting up

a captive insurer.
[6]
303 S2004
A general insurance company writes a wide range of general insurance products. The
reserve estimates for each homogenous portfolio are calculated using the BornheutterFergusson method. Company procedures dictate that the initial estimate of the
ultimate loss ratio is not changed once selected.
You are an actuary responsible for setting the reserve estimates for a stable portfolio
of short tailed business. You are reviewing the historical data for a fully developed
year. In this year $100m of premiums were written in a soft market. The portfolio is
reserved using the Bornheutter-Fergusson method based on reported claims, and at the
time of writing the expected cost of claims was under estimated. The initially
selected estimate of the ultimate loss ratio was 100% and the difference between the
initial and actual ultimate loss ratio is 20%.
The expected run off pattern for this portfolio is given below.
End of year
1
2
3
Paid claims
30% 90% 100%
Reported claims 50% 105% 100%
(i)
Explain why reported claims at the end of year 2 might exceed the ultimate
claims.
[4]
(ii)
Calculate the impact that the difference between the initial and actual ultimate
loss ratio has on the declared profits of the company at the end of each of the
three years, stating any assumptions you make.
[4]
(iii)
State the advantages and disadvantages to the company of its policy of not
changing the initial estimate of the Bornheutter-Fergusson ultimate loss ratio.
[4]
[Total 12]
(i)
List the third parties whose financial failure could have an impact upon a
general insurance company, giving in each case an example of the financial
effect upon the insurer.
[6]
(ii)
Describe the measures that an insurance regulator might impose in order to

reduce the risk of third party failure to the insurers that it regulates.
[4]
[Total 10]
303 S2004
PLEASE TURN OVER
A general insurance company specialises in writing commercial property insurance

business.
(i)
List the rating factors that it is likely to use in order to determine appropriate
premium rates.
[3]
(ii)
List the risks that the insurance company faces in relation to claims experience
from this class of business.
[4]
(iii)
Give examples of the measures that the insurance company could put in place
to control / reduce the variability of its claim experience.
[7]
[Total 14]
The government of a small developing nation wants to encourage exports from that
country. One of the measures that it has taken is to set up an export credit insurer. An
exporter can buy credit insurance from the insurer for each consignment of goods that
it exports. If the customer in the foreign country fails to pay for the goods, then the
insurer indemnifies the exporter.
(i)
Describe the characteristics of the claims that the export credit insurer is likely
to receive.
[5]
(ii)
List the policy data items that the export credit insurer will need to capture
when setting up a policy onto its system.
[2]
(iii)
Suggest appropriate ways to group claim development statistics for this insurer
when carrying out a reserving exercise to set outstanding claim reserves. [2]
(iv)
Describe how the unearned premium reserve would be calculated for this
insurer.
(v)
303 S2004
[2]
Suggest with reasons, suitable reinsurance structures for the export credit
insurer.
[4]
[Total 15]
You are the actuary for a large general insurance company selling household
buildings and contents insurance, through a broker network. The company has been
writing this business for 12 years. In common with its competitors in the market, the
company uses only geographical location and sum insured as rating factors.
The finance director of your company is considering the implications of introducing
some form of experience rating in order to set the company apart from the rest of the
market, and has asked for your assistance in analysing the impact of such a move.
(i)
Describe the different types of experience rating that could be considered, and
their consequences to the insurer.
[4]
(ii)
Compare the relative merits of the different options in this circumstance, and
make a recommendation between the alternatives.
[7]
(iii)
Discuss the impact that the introduction of experience rating may have on the
insurance company and its policyholders.
[9]
It has been decided to introduce the experience rating from 1 January 2005.
(iv)
Describe the portfolio movement analyses that could be put in place to

monitor the impact of this introduction.
[5]
[Total 25]
END OF PAPER
303 S2004
EXAMINATIONS
September 2004
Subject 303
General Insurance
EXAMINERS REPORT
Introduction
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however
given credit for any alternative approach or interpretation which they consider to be
reasonable.
M Flaherty
7 December 2004
September 2004
Examiners Report
This question was answered well by most candidates.

(i)
Policy data
Start and end dates of cover and dates of changes in cover
all rating factor details
all exposure measure details including limits, excesses etc.
source of business
details of premiums charged, unless they can be calculated by reference to
the details on rating factors and exposure
premiums net of commission etc.
Claims data
date of claim event
date reported
date settled
date reopened
dates and amounts of payments (including recoveries & expenses)
case estimates, if they exist, of amounts outstanding (and date)
other types of estimates (e.g. factor estimates)
rating factor details i.e. link between claims and policy details
type of claim
details of payee
type of peril
(ii)
Personal medical expenses

current age of policyholder
smoker/non-smoker
type of cover (hospital group)
extent of cover (excess, limits)
the extent of the family group covered
occupation
NCD as a proxy to medical history
post-code (address) as indicator of likely regional hospital costs
gender
Employers liability
payroll OR number of employees
type of industry or occupation
exposure and claims experience
location of the workforce
the materials handled
health and safety standards etc.
the processes involved
turnover
size of deductible
Page 2
September 2004
Examiners Report
The numerical solution in part (i) is the answer which the examiners had expected to
be given. However credit was also given where candidates had assumed an Accident
Year accounting basis.
(i)
Let growth in business = k

Premium in previous year = P
Premium next year = P * (1 + k)
Post tax profit = P * (1 + k) * (1 0.7 0.1 0.1) * 0.7 * 0.9
= P * (1 + k) * 0.063
Solvency margin required = P * (1 + k) * 0.5
Profit generate = additional solvency margin
P * (1 + k) * 0.063 = P * (1 + k) * 0.5 P * 0.5
Solving for k gives 14.4%
Assuming no investment return
Commission, management expenses and profit distribution continues at same
rate
Loss ratio is as assumed
Calculation is done annually
No reinsurance
(ii)
Grow more quickly by:

Charging higher premiums for the same exposure
Reducing cover for the same premiums
Stricter claims control.
Improved UW / risk selection
Reducing expenses and policy acquisition costs.
Reduce premiums with increase loss ratio more than offset by the reduction in
per policy expenses
Reducing the dividend that it pays out / not pay out a dividend at all.
Raising additional capital.
Buy additional quota share reinsurance.
A suitably structured financial reinsurance deal
Change investment strategy to increase investment return.
Incorporate in a tax haven.
Incorporate in a country with lighter insurance regulation.
Most candidates were able to make a reasonable attempt at this question, but there
was a wide variety of answers with the better candidates scoring almost full marks.
Advantages
Tax efficient
May be legislative or regulatory advantages
Fill gaps in insurance cover available in the market
Focus efforts on risk management
Manage overall Insurance spend of the company
Page 3
September 2004
Examiners Report
Improved financial planning due to more predictable insurance premiums, especially

the effects of the insurance cycle
Access reinsurance capacity directly.
Retain the expected profit that might otherwise be ceded to its insurers
Disadvantages
Administrative expense of setting up the captive
Capital cost required to set up the captive
The captive may need to get regulatory approval from all the countries that the
multinational operates in order to accept risk from all its operating subsidiaries.
Insurers may be able to provide assistance with risk mitigation / claims control. The
captive may not have the same expertise in all the different countries that the
company operates.
By setting up the captive, the multinational is retaining risk.
Setting up the captive may divert management focus from its core activities.
Compliance with insurance regulatory requirements
Many candidates used the paid claims information in part (ii) whereas the question
stated that the reserving is based on reported claims, hence a number of marks were
lost. Also in part (ii) some candidates derived the figures of ultimate claims and did
not go onto say what effect this had on the declared profits.
(i)
Claims made which end up as nil claims

E.g. Claims are made which fall outside the coverage of the policies
Claim assessments cautious which leads to an overstatement of outstanding
reserves
Reported claims do not allow for salvage or subrogation which reduce the
ultimate claims
(ii)
Premiums
Initial estimate lr
Actual LR
End of year
Paid
Reported
Dev ratio
Cum f
= 1 1/cum f
Reported claims
Claim reserve
Est Ult Claims
Actual ultimate claims
Overstatement of Profit
Page 4
100
100%
120%
1
30%
50%
2
90%
105%
3
100%
100%
2.100
2.000
0.5
60
50
110
120
10
0.952
0.952
0.05
126
(5)
121
120
1
1
0
120
0
120
120
September 2004
Examiners Report
End year 1 profit calculated 10 too big

Year 2
ultimate claims estimate deteriorates by 11, ultimate claims
overstated by 1
Year 3 zero difference
Assumptions: Written premiums all become earned.
Claims runoff according to the expected pattern.
(iii)
+ Simple method
avoids any subsequent investigation
+ For a short tailed line the errors generated will be relatively small and short
lived.
Does not allow for corrections once a different view of the likely ultimate
claims is known
May not satisfy accounting standards which require best estimate reserves
may give managers a false sense of security as to the true profitability of a
portfolio
either too high or too low
For long tailed lines the errors generated will potentially be large and long
lasting
- Pay more tax than necessary as profit overstated
The errors may overstate the true solvency position of the insurer, leading it
to take actions that may jeopardise the real solvency. In other cases the
reserves may be too prudent and so reserves that are too high are held
potentially restricting opportunities available.
The company may price using these reserve estimates, leading to incorrect
premiums being charged.
Several candidates gave examples of third parties who not in fact third parties e.g.
policyholders and staff. Apart from that most candidates answered this question
reasonably well.
(i)
Failure of a reinsurer could mean that it does not get reinsurance recoveries
that it is due. It may also have to purchase additional coverage.
Failure of an intermediary / Broker would mean that it may not get some
premiums owed to it.
Failure of a supplier of goods or services. The insurer would have to spend
additional money securing those goods or service.
Failure of a Financial Organisation / Bond issuer in which the Insurer has
assets. The insurer could lose those assets.
Failure of another insurer may mean that this insurer is asked to make
payments into a compensation fund. However the company could benefit from
a gain in extra business.
Page 5
(ii)
September 2004
Examiners Report
Limit the amount of reinsurance that Insurance companies can place with a
single Reinsurer.
Limit the amount of reinsurance that Insurance companies can take into
account with a single Reinsurer when determining solvency.
Impose or increase a Statutory Minimum Solvency Margin
Limit the amount of assets the insurer can invest in a single company / bond
issuer.
Require the licensing of intermediaries / brokers.
Require insurers to hold a bad debt reserve
Require insurers to hold certain asset classes or restrict asset classes
Require reinsurers to deposit assets / collateral to back their obligations to
Insurers.
Require that reinsurers and other financial organisations have sufficient credit
rating and are supervised more stringently.
Page 6
In answering this question some candidates failed to realise that commercial property
business is usually made up of buildings, contents and business interruption. In part
(iii) the answers were generally very short and many candidates did not mention
anything other than reinsurance.
(i)
Sum insured
Type of business (including hazardous chemicals)
Location of premises
Estimated maximum loss
Age of building
Safety equipment, including fire protection equipment
Construction type
Excess
Risks covered
Past claims experience
Risk management procedures
(ii)
Variability in claim frequency at any one time and from one period to another
Variability in claim severity at any one time and from one period to another
Individual large claim
Notification delays
Settlement delays
Moral Hazard. Fraudulent Claims.
Anti-selection.
Crime Rates (theft and arson claims)
September 2004
Examiners Report
Economic conditions, affecting attitude to claiming

Judicial decisions
Changes in legislation
Accumulations of risk
Catastrophes
Currency risks
Reinsurer security
Reinsurance mismatch
Policy wording
Inflation
Latent Claims
Increases in claims expenses
(iii)
Excess of Loss protection and Catastrophe XOL for large single events
Surplus reinsurance for ceding out larger risks.
Stop Loss / Aggregate XL reinsurance.
Exclusions (Terrorism / NCB)
Limits on size of risks that can be accepted.
Reduce exposure to certain industries / types of businesses that exhibit greater
variability of claim experience (e.g. Oil industry or large multinationals)
Track and control aggregates by geographic region.
Diversify by region, industry, type of business.
Match liabilities and assets by currency in order to reduce currency risk
Restrict the territories / currencies that it writes business in to reduce currency
risk.
Place reinsurance with financially strong reinsurers (AAA rated)
Spread the reinsurance placement amongst several reinsurers.
Employ specialist wordings staff to reduce risk from ambiguous policy
wording.
Impose excess / deductibles
Impose catastrophe / event limits or other sub-limits
Stricter claims underwriting to prevent fraudulent claims
Review rating structure and rates more frequently to reduce anti-selection
Diversify into other areas of insurance
The examiners were pleased with the general level of answers given to this question.
The main area of difficulty was in considering what was actually insured. Several
candidates included Goods In Transit within their answer. Examiners have also
noticed recently that question asking for suitable reinsurance structures tend to get
the standard answer of all types of insurance and not answers to the specific question.
(i)
Claims will generally be short tailed as typical trade credits are 3 6 months.
Claims will be in a number of currencies
Individual claims will generally be small and limited to the value of the export
consignment.
There will be nil claims as late payments are made to the exporter resulting in
the outstanding claim amount being reduced to zero
Page 7
September 2004
Examiners Report
There will be claim recoveries because some original invoices will be paid
after the insurer has paid the exporter s claim
There is likely to be accumulations of claims if a large number of exporters
supply the same customer.
Accumulations are also likely if a large number of foreign companies
experience financial difficulty due to economic conditions within those
countries.
There is potential for fraudulent claims if the exporter and the customer are
colluding.
(ii)
Unique policy identifier

Name & Address of policyholder
Premium for the policy.
Policy excess or deductible
Value to consignment
Currency of payment.
Name & Address of customer
Country of customer
Business type of customer (Retail / Manufacturing / Services etc)
Terms of trade for paying the invoice e.g. 30 days
Date policy issued
Date of despatch
Date when Invoice is payable.
(iii)
Split by Country
Split by Currency.
Group claims by the month in which the invoice is payable.
(iv)
The Premium would all be earned on the date the invoice is payable as that is
when all the exposure is.
Therefore UPR should simply be all the premiums on policies where the
invoice date falls after the accounting date netted down for commission and
other initial expenses.
(v)
Aggregate XL reinsurance to protect against accumulations from failure of a

single customer / multiple failures in foreign countries.
Stop loss reinsurance (if available) to protect poor experience across the book
of business.
Risk excess of loss to protect against individual large claims
Quota Share reinsurance to lay off some of this risk and enable it to write a
diverse portfolio.
Financial reinsurance to smooth results over economic cycles.
Page 8
September 2004
Examiners Report
Some candidates got the definitions the wrong way around, which meant that in part
(ii) the points made were the wrong way around. However if candidates were
consistently wrong then credit was given in part (ii) as appropriate. In
addition credit was given if points were made in answer to one part of the question
but fell into a different part in the examiners solution.
The question asked for a recommendation in part (ii) but some candidates failed to
produce a recommendation and therefore no marks could be awarded. Any well
reasoned recommendation would have scored marks.
(i)
Experience rating can be prospective or retrospective, numbers or amounts

based.
With prospective rating,
the premium at the renewal date depends to some extent on the experience of
the risk prior to that renewal.
The insurer takes on all the underwriting risk in such an arrangement.
With retrospective rating,
the premium for the current policy period is adjusted, based on the experience
of that period of risk.
A deposit premium, paid at the inception of the policy, will usually be
followed by an adjustment premium, or refund, at the end of the period
The insurer takes on less of the underwriting risk.
Numbers based systems depend upon the number of claims made.
Amounts based systems depend upon the total amount of claims made.
(ii)
Prospective vs retrospective
Broker network with many customers (personal lines) means prospective may
be easier to apply
Retrospective less risky to insurer
Retrospective more difficult to apply as there is uncertainty as to the amount
of some claims (e.g. subsidence)
Prospective means quotes change at the time of underwriting more likely to
have desired impact
Prospective requires more data up-front
12 years should be okay
Retrospective deals better with new customers
Prospective therefore rewards loyalty
Numbers vs Amounts
If claims frequency is small then neither is ideal
Numbers based more likely to discourage costly small claims
Numbers more appropriate when claims size is very variable
Amounts could be capped to make this more usable
Points for picking (my money is on prospective numbers (NCD))
(iii)
Potential large movements in individual premiums at point of renewal

Customers may not like the volatility introduced by such a system
May cause an exodus of policyholders; good and bad
In general, poorer risks see an increase in premiums
Leave and go to cheaper competitors
Company doesn t lose money on poor risks that leave
Page 9
September 2004
Examiners Report
Better risks see a reduction in premiums

more likely to stay, even at high
premiums relative to actual risk cost
Likely to attract better risks from other insurers too
Increased likelihood of attracting better risks if system can be applied to new
customers too
Important to get rates correct, or may be selected against
Need to monitor profile of any new business closely
Rapid growth from new business could cause new business strain
impacting on insurer s solvency
Rapid growth from new business could cause poor service to customers
resulting in loss of business and ability to write business in future
Take care to ensure underwriting and claims systems are capable of capturing
necessary information
Costs of changing UW and claims systems capability
Education of brokers is required
Business may be more complex to administrate may lead to restructured
commission
Difficult to apply for new business as may need to rely on honesty of
policyholder especially as no other insurer uses such a system
(iv)
new business
analyse the proportion of new quotes taken up, and/or the volume of business
growth
lapses or renewal
analyse the proportion of policies renewed each calendar year
endorsements
analyse the rate at which policies are changed during the year
mid-term cancellations
analyse the rate at which policies are cancelled during the year
conversion rate
analyse the proportion of quotes taken up
All of the above need to be analysed by rating factor to monitor the changing
nature of the portfolio.
Compare actual with expected
Page 10

Subject3031999 2004

Uploaded by

Document Information

Original Title

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

Subject3031999 2004

Uploaded by

Copyright:

Available Formats

Faculty of Actuaries

Time allowed: Three hours

Mark allocations are shown in brackets.

Attempt all 11 questions, beginning your answer to each question on a

Hand in BOTH your answer booklet and this question paper.

A small general insurance company underwrites only a single class of

An insurance company intends to offer a product to the motion picture

Describe the assumptions behind the Bornhuetter-Ferguson method of

Expected reported claims development pattern:

Calculate the expected ultimate claims as at 31 December 1998

Comment briefly on the reliability of the results in part (ii).

List the reasons why a general insurer models claims.

List the steps involved in applying a deterministic modelling method.

List the main regulatory restrictions that may be encountered by a general

Explain the terms reinstatements and reinstatement premiums in

Define the term rate on line.

A property catastrophe reinsurance layer with unlimited reinstatements

List the factors that a general insurance company writing many

PLEASE TURN OVER

State the information you would expect Company A to provide to

Define the following terms:

Unearned Premium Reserve

Calculate the expected ultimate claim cost of the unexpired risks as at

Determine whether or not an AURR is required in respect of Private

Subject 303 General Insurance

Subject 303 (General Insurance) April 1999 Examiners Report

Main objective is to maximise investment return whilst meeting all its

annual premium income

The effect of the reinsurance arrangements should be taken into account.

Actor,director or producer illness, accident or death

Subject 303 (General Insurance) April 1999 Examiners Report

Whatever claims have developed in relation to an origin year, the

The total estimated ultimate claims are 5,407,500

Reported claims are lower than original estimate by 142,500

To arrive at a risk premium per policy.

Subject 303 (General Insurance) April 1999 Examiners Report

Specify the purpose of the investigation

Valuation of the Assets

Restrictions on the type of business an insurer may write

Subject 303 (General Insurance) April 1999 Examiners Report

Lack of previous experience

A catastrophe excess of loss policy normally allows the reinsured to

Original premium width of layer. (Does not include reinstatement

Subject 303 (General Insurance) April 1999 Examiners Report

Company A is directly responsible for 40% of $750,000 = $300,000

As company B is a coinsurer alongside A the insolvency does not

Subject 303 (General Insurance) April 1999 Examiners Report

Subject 303 (General Insurance) April 1999 Examiners Report

Unexpired Loss ratio %

Marks are given for:

Unexpired claim cost

Subject 303 (General Insurance) April 1999 Examiners Report

Difference in expected ultimate loss ratios

Time allowed: Three hours

Mark allocations are shown in brackets.

Attempt all 9 questions, beginning your answer to each question on a

Hand in BOTH your answer booklet and this question paper.

A general insurance company is constructing an asset-liability model to

Discuss briefly whether it would be more effective to use a

List the portfolio movements that are normally monitored by a general

Company A is a medium-sized general insurance company, writing a variety of

All business written by the reinsured in the class

Reinsurance premium: 41% of original premiums of the subject business, to be