Ilovepdf Merged 6

1
2 INSTITUTE AND FACULTY OF ACTUARIES

3
4
5
6
7 EXAMINATION
8
9
19 April 2016 (am)
10
11
Subject CT8 – Financial Economics
12
Core Technical
13
14 Time allowed: Three hours
15
INSTRUCTIONS TO THE CANDIDATE
16
1. Enter all the candidate and examination details as requested on the front of your answer
17 booklet.
18
2. You must not start writing your answers in the booklet until instructed to do so by the
19 supervisor.
20
3. Mark allocations are shown in brackets.
21
4. Attempt all 10 questions, beginning your answer to each question on a new page.
22
23 5. Candidates should show calculations where this is appropriate.
24
Graph paper is NOT required for this paper.
25
26
AT THE END OF THE EXAMINATION
27
Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
28 question paper.
29
In addition to this paper you should have available the 2002 edition of the Formulae
30 and Tables and your own electronic calculator from the approved list.
31
32
33
34
CT8 A2016  Institute and Faculty of Actuaries
35
1 An investor measures the utility of her wealth using the utility function U(w) = ln(w)
for w > 0.
(i) Derive the absolute and relative risk aversions for this investor’s utility
function, and the first derivative of each. [4]
(ii) Comment on what this tells us about the proportion of her assets that this
investor will invest in risky assets. [2]
The investor has £100 available to invest in two possible assets, Asset A and Asset B.
The future value of Asset A depends on an uncertain future event.
 Every £1 invested in Asset A will be worth £1.30 with probability 0.75 and £0.40
with probability 0.25.
 Asset B is risk-free, so every £1 invested in Asset B will always be worth £1.
The investor does not discount future asset values when making investment decisions.
She decides to invest a proportion a of her wealth in Asset A and the remaining
proportion 1 – a in Asset B.
(iii) Express her expected utility of wealth in terms of a. [2]
(iv) Determine the amount that she should invest in each of Asset A and B to
maximise her expected utility, using your result from part (iii). [5]
[Total 13]
2 Consider an asset whose return follows the probability density function f(x).
(i) Write down a formula for the variance of the return on the asset, defining any
additional notation you use. [1]
(ii) Write down a formula for the shortfall probability for the return on the asset
below a level L. [1]
The returns on an asset follow a Normal distribution with mean µ = 6% per annum
and variance σ2 = 23% per annum. An investor buys €500 of the asset.
(iii) Determine the shortfall probability for the value of the asset in one year’s time
below a value of €480. [2]
(iv) Explain what can be deduced about an investor’s utility function if the investor
makes decisions based on:
(a) the variance of returns.

(b) the shortfall probability of returns.
[2]
[Total 6]
CT8 A2016–2
3 Consider a market with N securities. Let xi denote the weight of security i in a
portfolio, Vi the variance of the return on security i and Cij the covariance between the
returns on security i and security j.
(i) Write down an expression for V, the variance of the return on the portfolio. [1]
(ii) Describe how an efficient portfolio can be found under mean-variance

portfolio theory. [You do not have to include details of the partial derivatives
and their solutions.] [5]
(iii) Show that investors can diversify away specific risk by investing equal
amounts in an increasing number of independent securities. [3]
(iv) Show that the result in part (iii) still holds true when the securities are
correlated. [3]
[Total 12]
4 In a market where the assumptions of the Capital Asset Pricing Model (CAPM) hold,
there are a risk-free asset and two risky assets with the following attributes:
Rate of return (per annum)

State Probability Asset 1 Asset 2 Asset 3
1 0.2 5.0% 15.0% 26.0%

2 0.3 5.0% 22.0% 15.0%
3 0.1 5.0% 10.0% 24.0%
4 0.4 5.0% 28.0% 7.0%

Market capitalisation 30,000 70,000
(i) Determine the composition of the market portfolio. [1]
(ii) Determine the market price of risk. [5]
(iii) Calculate the beta of each risky asset. [2]
(iv) State the limitations of the CAPM. [3]

[Total 11]
5 (i) Define the three forms of the Efficient Markets Hypothesis. [3]
(ii) State two reasons why it is hard to test whether any of the three forms hold in
practice. [2]
[Total 5]
CT8 A2016–3 PLEASE TURN OVER

6 Suppose that at time t a portfolio (t , t) is held, where t represents the number of
units of a stock, with price St , held at time t and t is the number of units of a cash
bond, with price Bt , held at time t. The processes  and  are previsible.
Let V(t) = t St + t Bt be the value of the portfolio at time t.
(i) Explain what it means for this portfolio to be self-financing. [2]
Consider a stock paying a continuous dividend at a rate δ and denote its price at any
time t by St.
Let Ct and Pt be the price at time t of a European call option and European put option
respectively, written on the stock S, each with strike price K and maturity T ≥ t.
The instantaneous risk-free rate is denoted by r.
(ii) Prove put-call parity in this context by constructing two self-financing

portfolios whose value must be equal by the principle of no arbitrage. [6]
[Total 8]
7 Consider a non-dividend-paying share with price St at time t (in years) in a market

with continuously compounded risk-free rate of interest r.
(i) Show that the fair price at t = 0 of a forward contract on the share maturing at
time T is K = S0erT. [5]
A share is currently worth S0 = €20. The continuously compounded risk-free rate of

interest is 1% per annum.
(ii) Calculate the fair price at t = 0 of a forward contract written on the share with
delivery at t = 2. [1]
(iii) Give an expression for the value to the investor of the forward contract in part
(ii) at time t  2, in terms of St , t and r. [2]
An investor enters into the above forward contract at time t = 0. At time t = 1 the
risk-free rate of interest has increased to 4% per annum. The share price has not
changed.
(iv) Calculate the value to the investor of the forward contract at t = 1. [1]
(v) Determine each of the following Greeks for the contract value at time t = 1:
 delta
 theta
 vega
[3]
[Total 12]
CT8 A2016–4
8 Consider a three-period binomial tree model for the stock price process St.
Let S0 = 100 and let the price rise by 10% or fall by 5% at each time step.
Assume also that the risk-free rate is 4% per time period, continuously compounded.
(i) (a) State the conditions under which the market is arbitrage free.
(b) Verify that there is no arbitrage in the given market.

[2]
(ii) Calculate the price of a European call option on this stock, with maturity at the
end of the third period and a strike price of 103. [4]
A special option, called a European “Paylater” call option, has the following payoff at
maturity T:
( ST  K  c) if ST  K
and zero otherwise. K is the strike price and c is the premium paid for the option.
The premium is paid at maturity, and is only paid if the option expires in-the-money.
Further, the option premium is set such that the value of the option at time t = 0 is
zero.
Assume that K = 103 and the maturity of the contract is at time t = 3.
(iii) Determine the premium c of this contract. [3]

[Total 9]

9 (i) Draw a diagram to illustrate the Jarrow-Lando-Turnbull model for credit
default, defining any notation used. [4]
Consider a three-state credit model for a company in discrete time. The states are
Healthy (H), Unhealthy (U) or Defaulted (D). Transition probabilities from state i to
state j, pij, are constant:
pHU = 0.1
pUH = 0.05
pHD = 0.02
pUD = 0.3
pDj = 0 for all j ≠ D
Denote the probability that the company is in state i at time t (years) as pi(t).
A company is in the Healthy state at time 0.
(ii) Calculate pD(2), i.e. the probability that the company is in the Default state at
time 2. [2]
The company issues a zero-coupon bond at time 0, with maturity at time 2 and
nominal value £100. The continuously compounded risk-free rate of interest is 4%
per annum.
Assume that the bond returns its nominal value at time 2 if the company is not in
default, or x% of its nominal value at time 2 if the company is in default.
The fair price of the bond at time 0 is £87.63.
(iii) Calculate the value of x, the assumed percentage recovery on default. [2]
(iv) Calculate the credit spread on the bond. [1]
(v) Comment on the impact on the current price of the bond if it returned x% of its
nominal value at the time of default rather than at time 2. [1]
[Total 10]
CT8 A2016–6
10 In the Vasicek model, the short rate of interest under the risk-neutral probability
measure is given by:
t
rt    e  kt (r0  )    e k (t u ) dWu
0
where k, ,  > 0 and W is a standard Brownian motion.
Consider the related process:
t
Rt  rs ds
0
where rt is the short rate defined above.
(i) Show that Rt has a Normal distribution with mean and variance given by:
1  e kt
E ( Rt )  t  (r0  ) and
k
 2  2(1  e  kt ) 1  e 2 kt 
Var(Rt )  t    . [6]
k 2  k 2k 
Let P(0,t) be the price at time 0 of a zero-coupon bond with redemption date t > 0.
(ii) Show that, under the Vasicek model:
Var  Rt 
 E  Rt 
P (0, t )  e 2 . [3]
(iii) Show, by using the results from parts (i) and (ii), that:
 B(t ) r0
P(0, t )  A(t )e
1  e kt
where B(t ) 
k
  2  2 
and A(t )  exp  ( B (t )  t )    2   B (t ) 2  . [4]
  2k  4k 

(iv) State the main drawback of the above model for the term structure of interest
rates. [1]
[Total 14]
END OF PAPER
CT8 A2016–7
INSTITUTE AND FACULTY OF ACTUARIES
EXAMINERS’ REPORT
April 2016 (with mark allocations)

Core Technical
Introduction
The Examiners’ Report is written by the Principal Examiner with the aim of helping candidates, both
those who are sitting the examination for the first time and using past papers as a revision aid and
also those who have previously failed the subject.
The Examiners are charged by Council with examining the published syllabus. The Examiners have
access to the Core Reading, which is designed to interpret the syllabus, and will generally base
questions around it but are not required to examine the content of Core Reading specifically or
exclusively.
For numerical questions the Examiners’ preferred approach to the solution is reproduced in this
report; other valid approaches are given appropriate credit. For essay-style questions, particularly the
open-ended questions in the later subjects, the report may contain more points than the Examiners
will expect from a solution that scores full marks.
The report is written based on the legislative and regulatory context pertaining to the date that the
examination was set. Candidates should take into account the possibility that circumstances may
have changed if using these reports for revision.
F Layton
Chairman of the Board of Examiners
June 2016
 Institute and Faculty of Actuaries

Subject CT8 (Financial Economics Core Technical) – April 2016 – Examiners’ Report
A. General comments on the aims of this subject and how it is marked
1. The aim of the Financial Economics subject is to develop the necessary skills to construct
asset liability models and to value financial derivatives. These skills are also required to
communicate with other financial professionals and to critically evaluate modern financial
theories.
2. The marking approach for CT8 is flexible in the sense that different answers to those
shown in the solution can earn marks if they are relevant and appropriate. Marks for the
methodology are also awarded.
B. General comments on student performance in this diet of the

examination
1. Students performed relatively well on bookwork questions, although many missed the
opportunity to be awarded full marks for these due to relatively superficial knowledge.
2. The majority of the students though seemed to struggle on the applications part of the
questions, through not being able to put together the pieces of information given and use
them. In a few instances this resulted in students re-calculating given data from basic
principles and therefore running out of time. Further, there is often a lack of knowledge of
how to use the distribution tables to compute probabilities (in the specific case of this
exam paper, the normal distribution), and relative sloppiness in getting the details right.
C. Pass Mark
The Pass Mark for this exam was 60%.
Solutions
Q1 (i) U’(w) = 1/w [½]
U’’(w) = –1/w2 [½]
Absolute risk aversion = A(w) = – U’’(w)/U’(w) [½]
= 1/w [½]
A’(w) = –1/w2 [½]
Relative risk aversion = R(w) = – wU’’(w)/U’(w) [½]
=1 [½]
R’(w) = 0 [½]
[Total 4]
Page 2
(ii) R’(w) =0 thus the log utility function exhibits constant relative risk aversion.
[1]
This is consistent with an investor who keeps a constant proportion of wealth

invested in risky assets as she gets richer. [2]
[Max 2]
(iii) Wealth after the uncertain event will be either:
100 × (1.3a + (1 – a)) = 100 + 30a with probability 0.75 [½]
or:
100 × (0.4a + (1 – a)) = 100 – 60a with probability 0.25. [½]
Thus expected utility of wealth is:
0.75 × ln(100 + 30a) + 0.25 × ln(100 – 60a). [2]

[Max 2]
(iv) Differentiate with respect to a:
30 × 0.75/(100 + 30a) – 60 × 0.25/(100 – 60a). [2]
Set equal to zero:
30 × 0.75 / (100 + 30a) – 60 × 0.25 / (100 – 60a) = 0

30 × 0.75 / (100 + 30a) = 60 × 0.25 / (100 – 60a)
30 × 0.75 × (100 – 60a) = 60 × 0.25 × (100 + 30a)
22.5 × (100 – 60a) = 15 × (100 + 30a)
2250 – 1350a = 1500 + 450a
750 = 1800a
a = 0.4167 [2]
Check for maximum:
Differentiate with respect to a again:
– 302 × 0.75/(100 + 30a)2 – 602 × 0.25/(100 – 60a)2.
This must be negative because of the square terms, hence this is a local
maximum. [1]
Page 3
So invest £41.67 in Asset Aand £58.33 in Asset B. [2]

[Max 5]
[TOTAL 13]
Early parts of this question were largely completed well, though some
students used the incorrect formulae despite them appearing in the tables
(sign problems mainly). The majority of the students were able to correctly
identify the nature of the utility function in terms of index of relative risk
aversion but failed to comment about the proportion of the assets that the
investor will invest in risky assets. The majority of students also failed to
express the expected utility of wealth, and calculated the utility of expected
wealth instead.
Q2 (i) Variance of return is defined as:
∞
 −∞ (μ − x)
2
f ( x)dx ,
where μ is the mean return at the end of the chosen period. [1]
L
(ii) Shortfall probability =  f ( x)dx . [1]
−∞
(iii) The shortfall probability required is the probability that the return is lower
than 480/500 – 1 = –4% i.e. P(N(6%, 23%) ≤ 4%) [1]
= P(Z≤ (–4% – 6%)/√(23%)) [½]
= P(Z ≤ –0.20851) [½]
= 0.417 [1]
[Max 2]
(iv) (a) This may imply that the investor has a quadratic utility function. [1]
(b) This corresponds to a utility function which has a discontinuity at the

minimum required return. [1]
[Total 2]
[TOTAL 6]
Well prepared students scored well on the bookwork parts of this question,
although some students failed to define in full the notation used in part (i).
Many students had problems in calculating the shortfall probability using the
distribution of the normal random variable, and in recognising that the
corresponding utility function has a discontinuity.
Page 4
Q3 (i) V = Σi xi2 Vi + Σi Σj, j≠i xi .xj Cij . [Total 1]
(ii) The aim is to choose xi to minimise V … [1]
… subject to the constraints
Σi xi = 1 [1]
and expectation of return E = EP, say, in order to plot the minimum variance
curve. [1]
One way of solving such a minimisation problem is the method of Lagrangian

multipliers. [1]
The Lagrangian function is:
W = V − λ(E − EP) − μ(Σi xi − 1). [1]
To find the minimum we set the partial derivatives of W with respect to all the
xi and λ and μ equal to zero. [1]
The result is a set of linear equations that can be solved. [1]
The usual way of representing the results of the above calculations is by

plotting the minimum standard deviation for each value of EP as a curve in
expected return – standard deviation (E – σ) space. [1]
In this space, with expected return on the vertical axis, the efficient frontier is
the part of the curve lying above the point of the global minimum of standard
deviation. [1]
Any portfolio on this efficient frontier is an efficient portfolio. [1]
[Max 5]
(iii) Where all assets are independent, the covariance between them is zero and the
formula for variance becomes:
V = Σi xi2 Vi . [1]
If we assume that equal amounts are invested in each asset, then with N assets
the proportion invested in each is 1/N. Thus:
V = Σi (1/N)2 Vi [1]
= 1/N[Σi Vi/N] = 1/N [1]
where represents the average variance of the stocks in the portfolio.
Page 5
As N gets larger and larger, the variance of the portfolio approaches zero. [1]
[Max 3]
(iv) With equal investment, the proportion invested in any one asset xi is 1/N and
the formula for the variance of the portfolio becomes:
V = Σi (1/N)2 Vi + Σi Σj (1/N)(1/N).Cij . [1]
Factoring out 1/N from the first summation and (N − 1)/N from the second
yields:
V = 1/N Σi Vi /N + (N − 1)/N Σi Σj Cij/N(N − 1). [1]
Replacing the summation by averages we have:
V = 1/N Vi + (N − 1)/N. ̅ . [1]
The contribution to the portfolio variance of the variances of the individual

securities goes to zero as N gets very large. [1]
This shows that the individual risk of securities can be diversified away. [1]
The contribution of the covariance terms approaches the average covariance as

N gets large. However, this does not represent specific risk i.e. risk relating to
individual securities. [1]
[Max 3]
[TOTAL 12]
Early parts of this question were largely completed well. The majority of the
students proceeded without problems although a few provided answers only
for the general case of dependent assets. Some students answered parts (iii)
and (iv) using the single index model despite the question being clear that
mean-variance portfolio theory was being examined.
Q4 (i) The composition of the market portfolio is as follows:

wi 0.3 0.7 [Total 1]
(ii) Mean returns: Asset 2: Asset 3:

21.8% 14.9% [1]
Page 6
Consequently:
3
ErM = wi Eri = 16.97% [1]
i =2
 3  
2
= E  wi ri   − ( ErM ) = 3.57% .
2
std. dev ( rM ) = σ M [2]

 i = 2 
  
( )
The market price of risk is given by ErM − r f / σ M [1]
And since the risk-free rate is 5.0%, this equates to:
(0.1697 – 0.05)/0.0357 = 3.35 [1]

[Max 5]
(iii) ( )(
From the Security Market Line it follows that βi = Eri − r f / ErM − r f . [1])
Hence β2 = 1.40 and β3 = 0.83 . [1 mark each]
[Max 2]
(iv) The assumptions made are unrealistic. [1]

Empirical studies do not provide strong support for the model. [1]
It does not account for taxes. [1]
Or inflation. [1]
Or situations in which there is no riskless asset. [1]
It does not consider multiple time periods. [1]
Or optimisation of consumption over time. [1]
Investors don’t always use the same “currency” [1]
Markets are not always perfect [1]
Investors don’t always have the same expectations [1]
Cannot lend/borrow unlimited amounts at the same risk-free rate [1]
Difficult to check as need to think about investment markets as well as capital
markets [1]
Unrealistic to invest in the market portfolio in practice as so many stocks [1]
[Max 3]
[TOTAL 11]
Many students answered all parts of this question correctly. A few either
made calculation mistakes, or did not cover a wide enough range of
limitations of the CAPM. Some students confused calculating the market
price of risk with the risk premium.
Page 7
Q5 (i) The three forms are: Strong – market prices reflect all current information
relevant to the stock, including information which is not public. [1]
Semi-strong – market prices reflect all current, publicly available information

relevant to the stock. [1]
Weak – market prices reflect all information available in the past history of the
stock price. [1]
[Total 3]
(ii) Tests need to make assumptions (which may be invalid) such as normality of
returns or stationarity. [1]
Transaction costs may prevent the exploitation of anomalies, so that the EMH
might hold net of transaction costs. [1]
Allowance for risk: the EMH does not preclude higher returns as a reward for
risk; however the EMH does not tell us how to price such risks. [1]
Testing the strong form EMH is problematic as it requires access to

information that is not in the public domain. [1]
It can be difficult to define “public information” or to determine exactly when

information becomes public. [1]
It is impossible to test all of the possible trading rules that might be used by
technical analysts. [1]
The assumptions made about how security prices should react to new
information may be invalid. [1]
[Max 2]
[TOTAL 5]
Standard bookwork question which was largely well answered. Some

students referred to the investor knowing the information rather than the
security price reflecting the information or that the security price reflected
“only” the relevant information rather than “all” relevant information.
Q6 (i) This portfolio is described as self-financing if dV(t) is equal to φt dSt + ψt dBi.

That is, at t + dt, there is no inflow or outflow of money necessary to make the
value of the portfolio back up to V(t + dt). [Total 2]
(ii) Consider two self-financing portfolios:
• Portfolio A: holding the call (long position) and a sold put (short position)
at time t. [1]
Its value at time t is Ct – Pt [1]
Page 8
and at time T, it is ST – K. [1]
• Portfolio B: holding a fraction e−δ(T–t) of the underlying asset for St e−δ(T−t)

and shorting (borrowing) cash of Ke–r(T–t) at time t. [1]
Its value at time t is then St e−δ(T−t) – Ke–r(T–t). [1]
Its value at maturity is then ST – K by taking into account the dividends

which are paid continuously at rate δ. [1]
By the principle of no arbitrage… [1]

… both portfolios must have the same value at all time t, since they have the
same value at time T. [1]
Hence: Ct – Pt = St e−δ(T−t) – Ke–r(T–t) [1]

[Max 6]
[TOTAL 8]
Standard bookwork question. The majority of the students answered correctly

although quite a few did not justify their argument on the basis of the no
arbitrage principle. Alternative valid approaches (including different portfolio
combinations) were of course acceptable.
Q7 (i) Let K be the forward price. Now compare the setting up of the following
portfolios at time 0:
A: one long forward contract. [1]

B: borrow Ke–rT cash and buy one share at S0. [1]
If we hold both of these portfolios up to time T then both have a value of

ST – K at T. [1]
By the principle of no arbitrage… [1]

… these portfolios must have the same value at all times before T. [1]
In particular, at time 0 both portfolios must have value zero (since the value of
a forward contract at t = 0 is zero). [1]
Since portfolio B has value S0 – Ke–rT at t = 0, this can only be zero if

K = S0erT. [1]
[Max 5]
(ii) K = €20 × e2×0.01 [1]

= €20.40 [1]
[Max 1]
Page 9
(iii) Value = (Ster(2–t) – 20.40)e–r(2–t) = St – 20.40er(t–2). [Total 2]
(iv) Using (iii), we get value = €20 – €20.40e–0.04 [1]

= €0.40 at time 1 [1]
[Max 1]
(v) Using (iii):
delta = d/dSt(St – 20.40er(t–2)) [1]

[½ mark for the definition of the greek, ½ mark for the actual formula]
=1 [1]
theta = d/dt(St – 20.40er(t–2)) [1]

[½ mark for the definition of the greek, ½ mark for the actual formula]
= – 20.40 r t er(t–2) = –0.784 at t = 1 [1]

vega = 0 [1]
as the value does not depend directly on the volatility of the share [1]
[Max 3]
[TOTAL 12]
Early parts of this question were answered well. The majority of the students
though confused the forward price (i.e. the delivery price) with the value to the
investor of the forward contract in part (iii), and consequently struggled with
the remaining parts, even if the large majority knew the definition of each
Greek.
Q8 (i) (a) The market is arbitrage free if and only if there exists a probability
measure under which discounted asset prices are martingales. [1]
In this case, the probability exists if and only if d < er Δt < u. [1]
(b) d = 0.95 < e 0.04 < 1.1 = u hence the condition is verified. [1]
[Max 2]
(ii)
Stock tree
Time 0 1 2 3
100.00 110.00 121.00 133.10

95.00 104.50 114.95
90.25 99.28
85.74
[½ mark for each of the prices of the stock at time 3]
Page 10
The price C0 of the option is computed via risk-neutral valuation. [½]
Let p̂ denote the risk-neutral probability of an up movement, then:
pˆ = { exp ( 0.04 ) − 0.95} / {1.1 − 0.95}} = 0.6054 [1]
and
3
 3
( )
+
C0 = e − rT    pˆ k (1 − pˆ )
3− k
S0u k d 3− k − K
k
k =0  
( )
= e− rT pˆ 3 × 30.10 + 3 pˆ 2 (1 − pˆ ) ×11.95 = 10.52. [4]
[Max 4]
For information, the detailed tree-based workings are provided below:
CALL
Time 0 1 2 3
10.52 15.45 22.04 30.10

4.04 6.95 11.95
0.00 0.00
0.00
(iii) As the premium is set so that the option price is zero, by risk-neutral valuation
it follows that:
C0
c = erT [2]
ˆ
P( ST > K )
C0
= e rT [2]
pˆ + 3 pˆ 2 (1 − pˆ )
3
= 18.09 [1]
[Max 3]
Page 11
Alternatively for the above:
From the figures in part (ii), must have:
[(30.1 – c) × p̂ 3 + (11.95 – c) × 3 p̂ 2 (1 – p̂ )] × exp-(3 × 0.04) = 0 [2]
c = [30.1 × p̂ 3 + 11.95 × 3 p̂ 2 (1 – p̂ )] / [ p̂ 3 + 3 p̂ 2 (1 – p̂ )] [2]
= 18.09 [1]
[Max 3]
[TOTAL 9]
Generally well answered, although some made calculation mistakes in

obtaining the prices of the given option contracts. In particular, too many
students simply used 1.04 for exp(0.04), which is of course not correct – a
simple check with the calculator would have shown this. The majority of
students adopted the correct approach to solve part (iii).
Q9 (i)
λ2j(t) j
2 λj2(t)
λj,n-1(t)
λn-1,j(t)
λ12(t)
λ21(t)
1 n-1
λ2n(t)
λ1n(t) λn-1,n(t)
[3 marks for diagram]
The n states represent n – 1 credit ratings plus default. [1]
λij(t) are the deterministic transition intensities from state i to state j at time t
under the real world measure P. [1]
[Max 4]
[1½ marks for diagram applied to specific example – 3 states model]
Page 12
(ii) pD(2) = pHD + pHH × pHD + pHU × pUD [1]

= 0.02 + (1 – 0.1 – 0.02) × 0.02 + 0.1 × 0.3 [1]
= 0.0676 [1]
[Max 2]
(iii) £87.63 = e–2×0.04 (( 1 – pD(2)) × 100 + pD(2) × 100x) [1]

= e–2×0.04 ((1 – 0.0676) × 100 + 0.0676 × 100x) [1]
so x = 25% [1]
[Max 2]
(iv) 87.63 = 100e–2×(r+c) where c is the credit spread. [1]

So c = 2.6%. [1]
[Max 1]
(v) The impact on cashflows would be that the bond might return the x% of its
nominal value earlier than time 2, so the value of the bond would increase.
[Total 1]
[TOTAL 10]
Generally, students answered this question correctly, although in the first part
quite a few considered only the particular case of the three states given in the
rest of the question (but which had not yet been introduced for part (i)). A few
students used the Merton formula for default to solve part (iii) of the question,
which was not appropriate for this model.
Q10 (i) By substitution and direct integration between 0 and t:
Rt = θ t . [1]
+ ( r −θ )
(1 − e ) − kt
[2]
0
k
t
+
σ
k
0
 (
1 − e ( ) dWs
−k t −s
) [3]
In virtue of the properties of the stochastic integral, Rt follows a Normal

distribution [1]
with the given mean (as the integral has zero mean) [1]
and the given variance – see Result 3.2 in Core Reading (Ito isometry). [1]
[Max 6]
Page 13
(ii) From risk-neutral valuation, the price of a bond is given by:
 − t r ds 
 s 
P ( 0, t ) = E  e 0  [2]
 
 
= E e− Rt ( ) [1]
which is the first moment of exp( − Rt ). [1]
As − Rt is normally distributed, the moment generating function gives the first

moment of − Rt as required:
Var( Rt )
− E ( Rt )+
=e 2 . [2]
Equivalently, use the results regarding the mean of the lognormal random
variable as per the Formulae & Tables.
[Max 3]
(iii) Notice that the variance can be written as:
σ2  1 − e− kt 1 − e −2 kt 
Var ( Rt ) =  t − B ( t ) − +  [2]
k 2  k 2k 
σ2 σ 2  1 − e − kt 1 − e −2 kt 
= − 2 ( B (t ) − t ) +  −2 + 
k 2k  k2 k2 
σ2 σ 2
2 ( ( )
B t −t) −
2
=− B (t ) .
k 2k
The result follows by substitution. [2]

[Max 4]
Page 14
(iv) The main drawback of the Vasicek model is that the short rate can take
negative values with positive probability. [1]
[TOTAL 14]
There was a wide range of quality of answers for this question. Generally,
students answered correctly the first and last parts of this question. Many
students managed to get through a few steps for part (ii), though often with
algebra issues; whilst in part (iii) there were relatively few comprehensive
attempts.
END OF EXAMINERS’ REPORT
Page 15
1
2 INSTITUTE AND FACULTY OF ACTUARIES
3
4
5
6
7 EXAMINATION
8
9
6 October 2016 (pm)
10
11
12
Core Technical
13
14 Time allowed: Three hours
15
16
17 booklet.
18
19 supervisor.
20
3. You have
Mark 15 minutes
allocations are of planning
shown and reading time before the start of this examination.
in brackets.
21 You may make separate notes or write on the exam paper but not in your answer
4. booklet. all
Attempt Calculators are beginning
10 questions, not to be used
yourduring
answerthetoreading time. You
each question on will
a newthen have
page.
22 three hours to complete the paper.
24
5. Attempt all 10 questions, beginning
Graph paper your
is NOT answerfor
required to each question on a new page.
this paper.
25
27 Graph paper is NOT required for this paper.
28 question paper.
29 AT THE END OF THE EXAMINATION
30 Hand in BOTHand your answer
Tables and booklet,
your ownwith any additional
electronic calculatorsheets firmly
from the attached,
approved list. and this
question paper.
31
32 In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.
33
34
CT8 S2016  Institute and Faculty of Actuaries
35
1 Consider an asset whose return follows the probability density function f(x).
(i) Write down a formula for the Value at Risk for the asset, at confidence level p.
[1]
(ii) Write down a formula for the downside semi-variance of the return on the
asset, defining any additional notation you use. [1]
(iii) State the arguments for and against using semi-variance as a risk measure. [2]
A farmer has a small apple tree which produces one harvest of apples per year. The
number of apples the tree produces follows a Poisson distribution with a mean and
variance of 8.
(iv) Determine the 10% Value at Risk level for the number of apples produced. [3]
(v) Determine the expected shortfall below a harvest of 5 apples. [3]

[Total 10]
2 (i) State the main assumptions of mean-variance portfolio theory. [4]
Consider a mean-variance portfolio model with two securities, with respective returns
SA and SB, where the expected return E[SB] = 0.25E[SA] and the variance of return
V[SB] = 0.25V[SA].
Let the correlation between the returns on the two securities be ρ.
(ii) Determine, in terms of E[SA], the expected return on the minimum variance
portfolio if:
(a) ρ=0
(b) ρ=1
[4]
(iii) (a) Calculate the variance of the return on the minimum variance portfolio
for part (ii)(b).
(b) Comment on the risk in this portfolio.

[2]
[Total 10]
CT8 S2016–2
3 In a market where the assumptions of the Capital Asset Pricing Model hold, there are
two risky assets with the following attributes:
Security A B
Expected return (p.a.) 20% 16%
(i) Determine the composition of the market portfolio with expected return 18%
per annum. [2]
(ii) Calculate the beta of each security under the assumption that the risk-free rate
of interest is 10% per annum. [2]
[Total 4]
4 Let Ri denote the return on security i given by the following multifactor model:
Ri  ai  bi ,1 I1  bi ,2 I 2  bi , L I L  ci
where ai and ci are the constant and random parts respectively of the component of
the return unique to security i, I1 , , I L are the changes in a set of the L indices and
bi ,k is the sensitivity (factor beta) of security i to factor k.
(i) State the category of the above model where:
(a) index 1 is a price index, index 2 is the yield on government bonds and
index 3 is the annual rate of economic growth.
(b) index 1 is the level of Research and Development expenditure, index 2

is the price earnings ratio, index 3 is the level of gearing.
[2]
Consider the following two-factor model for the returns on three assets A, B and C:
Asset A B C
ai 0.03 0.05 0.1
bi,1 1 3 1.5
bi,2 4 2 1.5

(ii) Determine the equation for the expected return on a portfolio which:
(a) equally weights the three securities.

(b) has weights x A  0.5, xB  1.5, xC  0.
[4]
(iii) Construct a portfolio of securities A, B, C that has a factor beta of 2 on the first
factor and 1 on the second factor, i.e. the expected return on the portfolio is:
RP  aP  2 I1  I 2  cP . [3]
[Total 9]
CT8 S2016–3 PLEASE TURN OVER

5 (i) State the key arguments against modelling market returns using a Gaussian
random walk. [3]
(ii) Describe the difference in time series modelling between a cross-sectional

property and a longitudinal property, including their dependence on the initial
conditions imposed on the model. [4]
[Total 7]
6 Let pt denote the value at time t (measured in years) of a European put option on a
non-dividend-paying stock with price St . The option matures at time T and has a
strike price K. The continuously compounded risk-free rate of interest is r.
(i) Derive a lower bound for pt in terms of St and K. [4]
Consider a market with the following two non-dividend-paying stocks:
Stock Volatility Current price
S1 10% £10
S2 20% £10
The following options are available on those stocks:
Derivative Underlying Strike Time to Current

asset price expiry price
(a) European call option S1 £8 1 £2.50

(b) European call option S2 £8 1 £2.26
(c) European put option S1 £8 1 £1.55
(d) European put option S1 £12 1 £1.20
(e) American put option S1 £12 1 £1.13
The continuously compounded risk-free rate is 3% per annum.
(ii) Identify, with reasons, five discrepancies in these option prices. [5]
[Total 9]
CT8 S2016–4
7 Consider a binomial tree model for the stock price St. Let S0 = 50 and let the price
rise by 10% or fall by 5% each month for the next three months. Assume also that the
risk-free rate is 5% per annum continuously compounded.
(i) State the conditions under which the market is arbitrage free. [2]
(ii) Calculate the price at time t = 0 of a European call option on this stock, which
expires in three months and is struck at-the-money (i.e. strike price K = 50).
[4]
A special option, called a knock-out barrier option, goes out of existence (i.e. expires
without any payoff or value) if the underlying asset reaches a pre-specified barrier
b > 0 either from above (down-and-out) or from below (up-and-out).
The down-and-out call has the following payoff at time T:
max( ST  K , 0) if min St  b,
0t T

0 otherwise.
Assume this special option is written on the given stock, has the same strike price and
maturity as the European call option described in part (ii) and the barrier b is fixed
at 48.
(iii) Calculate the price of this contract using the binomial tree model and risk-
neutral valuation. [3]
(iv) Determine the price of the down-and-out contract when b = 40, without
performing any further calculations. [2]
[Total 11]

8 Consider a non-dividend-paying stock, with price St , and a European call option on
that stock, whose value can be modelled using the Black-Scholes model.
(i) Write down the formula for the delta of this option under this model. [1]
Suppose that the stock price at time 0 is S0 = $40 and the continuously compounded
risk-free rate is 2% per annum. The call option has strike price $45.91, term to
maturity 5 years and a delta of Δ = 0.6179.
(ii) Determine the implied volatility of the stock to the nearest 1%. [4]
A second stock with price Rt is currently priced at R0 = $30 and has volatility
σR = 15% per annum.
An exotic option pays an amount c at time T if S1/S0 < kS and R1/R0 < kR.
(iii) Give a formula for the value of the option at time 0 if the two stocks are
independent, defining any additional notation used. [2]
(iv) Explain how the structure of the option could be simplified if the assets were
perfectly correlated. [3]
Assume now that the stock prices are independent. The option has term T = 1 year,
payoff c = $50 and strike prices kS = 0.8 and kR = 0.6.
(v) Determine the value of the option at time 0. [5]

[Total 15]
CT8 S2016–6
9 (i) Write down the properties of the following two models for interest rates:
(a) the one-factor Vasicek model

(b) the Cox-Ingersoll-Ross model
[You are not required to give any formulae for the models.] [4]
The Vasicek term structure model is described by the following stochastic differential
equation:
drt  a (b  rt )dt  dWt ,
with initial value r0 and a, b, σ > 0.
(ii) Show, by solving the Vasicek stochastic differential equation, that:
t
 
rt  r0e  at  b 1  e  at   e   dWs .
a t s
[4]
0
(iii) Determine the expectation, the variance and the distribution of the short
rate rt . [3]
[Total 11]

10 A company has issued zero-coupon bonds payable in five years’ time for a nominal
amount of £100m. The company has also issued 1 million non-dividend-paying
shares. A Black-Scholes model for the value of the company is adopted.
(i) Derive an expression for the value of the debt at time 0 using the Merton
model, in terms of the total value of the company and the value of a call
option. [4]
The current total value of the company is £200m. The continuously compounded
risk-free interest rate is 1% per annum.
The current arbitrage-free prices of options on the company’s shares, with maturity in
five years’ time and a strike price of £100, are as follows:
 put option = £17.30

 call option = £27.55
(ii) Calculate, using put-call parity, the value of the zero-coupon bonds per £100
nominal. [3]
The volatility of the total value of the company is 17% per annum.
(iii) Determine the approximate change in the share price and the bond price that
would arise from a £1m increase in the total value of the company.
[Hint: consider the delta of an appropriate option.] [4]
(iv) Comment on the relative change in the share and bond prices in part (iii). [2]
(v) Comment, without carrying out any calculations, on how the relative change
in part (iii) would differ if the total value of the company was lower. [1]
[Total 14]
END OF PAPER
CT8 S2016–8
EXAMINERS’ REPORT
September 2016

Core Technical
Introduction
exclusively.
Luke Hatter
Chair of the Board of Examiners
December 2016

Subject CT8 (Financial Economics Core Technical) – September 2016 – Examiners’ Report
theories.

examination
opportunity to be awarded full marks due to relatively superficial knowledge.
2. The majority of the students seemed to struggle on the applications part of the questions,
because they were not able to use and combine the information given to them in the
question. In a few instances this resulted in students re-calculating given data from basic
principles and therefore running out of time. Further, there is often a lack of knowledge of
how to use the distribution tables to compute probabilities (in the specific case of this
exam paper, the normal distribution), and relative inaccuracy in getting the details right.
C. Pass Mark
The Pass Mark for this exam was 60.
Page 2
Solutions
Q1 (i) VaR(X) = −t where P(X < t) = p. [1]
μ
 −∞ (μ − x)
2
(ii) f ( x)dx where μ is the mean return at the end of the chosen
period.
μ
 −∞ ( x − μ)
2
[Or equivalently f ( x)dx ] [1]
(iii) For:
Most investors do not dislike uncertainty of returns as such; rather they dislike
the possibility of low returns. One measure that seeks to quantify this view is
downside semi-variance. [1]
Against:
Semi-variance is not easy to handle mathematically. [½]
Semi-variance takes no account of variability above the mean. [½]
Furthermore if returns on assets are symmetrically distributed semi-variance is

proportional to variance, so it gives no extra information. [1]
Semi-variance measures downside relative to the mean rather than another

benchmark that might be more relevant to the investor. [½]
[Max 2]
(iv) P(X = 0) = (80e–8)/0! = 0.00034

P(X = 1) = (81e–8)/1! = 0.00268
P(X = 2) = (82e–8)/2! = 0.01073
P(X = 3) = (83e–8)/3! = 0.02863
P(X = 4) = (84e–8)/4! = 0.05725
So P(X ≤ 4) = 0.00034 + 0.00268 + 0.01073 + 0.02863 + 0.05725 = 0.09963
Alternatively, directly from the Formulae & Tables: P(X ≤ 4) = 0.09963 [1]
P(X = 5) = (84e–8)/5! = 0.09160

So P(X ≤ 5) = 0.191236 (or directly from the Formulae & Tables) [1]
So the 10% VaR level is 5 (or –5) apples. [1]
(v) P(X = 0) × (5 – 0) = 0.002

P(X = 1) × (5 – 1) = 0.011
P(X = 2) × (5 – 2) = 0.032
P(X = 3) × (5 – 3) = 0.057
P(X = 4) × (5 – 4) = 0.057 [2]
Page 3
Summing the above, we get 0.159
So the expected shortfall below 5 apples is 0.159 apples [1]

[Total 10]
Well-prepared students scored well here, though some confused semi-

variance with expected shortfall. Many students calculated probabilities
though they are listed in the Formulae & Tables.
A common mistake was to calculate the VaR and Expected Shortfall using the
distribution function rather than the probability mass functions, i.e. P(X ≤ x)
rather than P(X = x).
Q2 (i) Investors select their portfolios on the basis of the expected return and the
variance of that return over a single time horizon. [1]
The expected returns, variance of returns and covariance of returns are known
for all assets and pairs of assets. [1]
Investors are never satiated. At a given level of risk, they will always prefer a
portfolio with a higher return to one with a lower return. [1]
Investors dislike risk. For a given level of return they will always prefer a
portfolio with lower variance to one with higher variance. [1]
(ii) We use the following notation for i=A,B:
E(Si)=Ei
V(Si)=Vi
and CAB is the covariance between the returns of Asset A and Asset B.
(a) From the Core Reading
VB − C AB
xA = .
VA − 2C AB + VB
[1]
So xA = (0.25VA – 0) / (VA – 0 + 0.25VA) = 0.2
and xB = 0.8. [½]
Hence expected return = 0.2 × EA + 0.8 × 0.25EA [½]
= 0.4EA . [½]
Page 4
(b) Now CAB = √(VA × 0.25 × VA) = 0.5VA . [1]
So xA = (0.25VA – 0.5VA) / (VA – 2 × 0.5VA + 0.25VA) = –1
and xB = 2. [½]
Hence expected return = –1 × EA + 2 × 0.25EA [½]
= –0.5EA . [½]
[Max 4]
(iii) (a) The variance of the return on the portfolio in (b) is:
(–1)2 × VA + 22 × VB + 2 × (–1) × 2 × 0.5VA [½]
= 0. [½]
(b) So we have created a risk-free portfolio. [1]

[Total 10]
In part (i) the majority of students stated all the assumptions of MVPT but
marks were only available for the main assumptions (as asked for in the
question).
Part (ii) was a straightforward calculation based on bookwork and most

students scored well. A proportion of students attempted to derive the
formula for the minimum variance portfolio from scratch rather than using the
formula given in the Core Reading – this was time-consuming and the number
of marks on offer should have been a good guide that this was not required.
Page 5
Q3 (i) The market portfolio is the weighted portfolio of the risky securities in the
market, [½]
consequently ErM = 18% = w1Er1 + w2 Er2 . [½]
As w1 + w2 = 1 , then w1 = w2 = 0.5 . [1]
(ii) From the Security Market Line:
Eri − r f
βi = [1]
ErM − r f
therefore β A = 1.25 and βB = 0.75 . [1]

[Total 4]
A straightforward question and many students scored full marks. Part (i) was
well answered by all students but most students didn’t define the market
portfolio when deriving its composition.
Q4 (i) (a) Macroeconomic. [1]

(b) Fundamental. [1]
(ii) The results follow from the fact that the factor beta of a portfolio on a given
factor is the portfolio-weighted average of the individual securities’ betas on
that factor. This also applies to the constant and the random part. [1]
Working as follows:
Asset A B C
ai 0.03 0.05 0.1

bi,1 1 3 1.5
bi,2 –4 2 1.5
Weights
P1 0.33 0.33 0.33
P2 –0.5 1.5 0
P1 P2
aP 0.06 0.06
bP,1 1.83 4
bP,2 –0.17 5
Page 6
Hence:
(a) = 0.06 + 1.83 − 0.17 + [1]

for cP = ( c A + cB + cC ) / 3 . [1]
(b) = 0.06 + 4 +5 + [1]

for cP = −0.5cA + 1.5cB . [1]
[Max 4]
(iii) The required portfolio has weights such that the portfolio-weighted averages
of the betas equal to the target beta. Hence, we need to solve the linear
system:
 xA 
 1 3 1.5    2 
 −4 2 1.5  xB  =  1  [1½]
  x   
 C
which returns solution x A = 1/ 8 , xB = 3 / 8 and xC = 1/ 2 (recall that the

weights sum up to 1). [½ each = 1½ total]
[Total 9]
Part (i) was well answered by almost all students. Part (ii) was also answered
well, but some students wrote the expected return on the portfolio in terms of
the return on the indices rather than the expected values of the indices.
Most students correctly derived the equations to solve in part (iii) with the
majority also solving the equations correctly.
Q5 (i)
• Market crashes appear more often than one would expect from a normal
distribution. (The real world distribution has “fat tails”.) [½]
• While the random walk produces continuous price paths, jumps or

discontinuities seem to be an important feature of real markets. [½]
• Days with no change, or very small change, also happen more often than
the normal distribution suggests. (The real world distribution is “more
peaked”.) [½]
• The assumption of independent increments is contradicted by empirical

evidence of mean reversion and momentum effects.
[½]
• The assumption of a constant volatility is contradicted by empirical
evidence. [½]
Page 7
• It can be argued that expected returns on shares are likely to vary with
bond yields, which contradicts the assumption of a constant mean. [½]
• Random walks have a fractal dimension of 1½, (whereas) empirical

investigations of market returns often reveal a fractal dimension around
1.4. [½]
• Market returns are often (negatively) skewed [½]

[Max 3]
(ii)
• A cross-sectional property fixes a time horizon and looks at the
distribution over all the simulations. [1]
• For example, we might consider the distribution of inflation next year. [½]
• Implicitly, this is a distribution conditional on the past information which

is built into the initial conditions, and is, of course, common to all
simulations. [½]
• If those initial conditions change, then the implied cross-sectional

distribution will also change. [½]
• As a result, cross-sectional properties are difficult to validate from past

data, since each year of past history typically started from a different set of
conditions. [½]
• However, the prices of derivatives today should reflect market views of a

cross-sectional distribution. [½]
• Cross-sectional information can therefore sometimes be deduced from the

market prices of options and other derivatives. [½]
• A longitudinal property picks one simulation and looks at a statistic

sampled repeatedly from that simulation over a long period of time. [1]
• For example, we might consider one simulation and fit a distribution to the
sampled rates of inflation projected for the next 1,000 years. [½]
• For some models, this longitudinal distribution will converge to some

limiting distribution (ergodic distribution) as the time horizon lengthens.
[½]
• Furthermore this limiting distribution is common to all simulations. [½]
Page 8
• Unlike cross-sectional properties, longitudinal properties do not reflect

market conditions at a particular date but, rather, an average over all likely
future economic conditions. [½]
[Max 4]
[Total 7]
Part (i) was standard bookwork which was answered well.
In part (ii), most students defined a cross-sectional and longitudinal property

correctly but failed to note their dependence on the initial condition and their
differences, as asked in the question. There were easy marks to be had by
giving an example of each, which few students did.
Q6 (i) Consider a portfolio, A, consisting of a European put on a non-dividend-

paying share and a share. [½]
At time T, portfolio A has a value of at least K, which is equal to that of the

cash alternative at time t of Ke–r(T–t). [½]
Thus by the principle of no arbitrage… [½]
pt + St ≥ Ke−r(T−t) . [1]
So pt ≥ Ke−r(T−t) − St . [1]
Moreover pt ≥0 as the payoff is always ≥0 so pt ≥ max(Ke−r(T−t) − St ,0) [1]

[Max 4]
(ii)
• (b) should be greater than (a) because the underlying asset is more volatile.
[1]
• (c) should be £0.26 by put/call parity (assuming that (a) is correct). [1]
[Alternatively, (a) should be £3.79 if (c) is correct, or both could be
incorrect.]
• (d) should be higher than (c) because the strike price is higher. [1]
• (d) is below the lower bound of £1.65. [1]
Page 9
• (e) should be higher than (d) because an American option is always worth
at least as much as a European option. [1]
[Total 9]
Part (i) was largely well answered, though common mistakes included using
incorrect portfolios or trivial arguments that did not really constitute a proof.
Part (ii) was well answered by the majority of students, though few identified
all five differences. Some students calculated theoretical prices using the
Black-Scholes formula, but this was time-consuming and not necessary to
identify the discrepancies. Many students failed to check the lower bound for
the put option despite having proved it in part (i).
Q7 (i) The market is arbitrage free if and only if there exists a probability measure
under which discounted asset prices are martingales. [1]
In this case, the probability exists if and only if d < erΔt < u. [1]
Using the figures in the question, 0.95 < e0.25 × 0.05 < 1.1 [1]
er − d
Alternatively we need 0 < q < 1 where q = [1]
u−d
[Max 2]
(ii)
Stock price tree
Time 0 1 2 3
50.00 55.00 60.50 66.55
47.50 52.25 57.48
45.13 49.64
42.87
[1 mark for the final prices; the bottom one is not necessary]
The price C0 of the option is computed via risk-neutral valuation; let p̂ denote
the risk-neutral probability of an up movement, then:
pˆ = { exp ( 0.05 / 12 ) − 0.95} / {1.1 − 0.95}} = 0.3612 [1]
and
 k =0  k  pˆ k (1 − pˆ )3−k ( S0u k d 3−k − K )+

3 3
C0 = exp(–rT)
= exp(–rT) ( pˆ 3 × 16.55 + 3 × pˆ 2 (1 − pˆ ) × 7.48) = 2.62. [2]

Detailed workings:
Page 10
CALL
Time 0 1 2 3
2.62 5.56 10.71 16.55
0.97 2.69 7.48
0.00 0.00
0.00
(iii) The only relevant trajectories, given the barrier set at 48, are the up-up-up, up-
up-down and up-down-up (i.e. the ones leading to a state in which the call
option is in the money). [1]
The price by risk-neutral valuation, therefore, is 2.0.

[1 for computation]
[1 for values of probabilities of the relevant trajectories – see table below]
Workings as follows:
K 50.0000
Split the stock tree DOC barrier b 48.0000
0 1 2 3 PATH min Payoff Probs Exp.
Value
50.0000 55.0000 60.5000 66.5500 uuu 55.0000 16.5500 0.0471 0.7797
50.0000 55.0000 60.5000 57.4750 uud 55.0000 7.4750 0.0833 0.6229
50.0000 55.0000 52.2500 57.4750 udu 52.2500 7.4750 0.0833 0.6229
50.0000 55.0000 52.2500 49.6375 udd 49.6375 0.0000 0.1474 0.0000
50.0000 47.5000 52.2500 57.4750 duu 47.5000 0.0000 0.0833 0.0000
50.0000 47.5000 52.2500 49.6375 dud 47.5000 0.0000 0.1474 0.0000
50.0000 47.5000 45.1250 49.6375 ddu 45.1250 0.0000 0.1474 0.0000
50.0000 47.5000 45.1250 42.8688 ddd 42.8688 0.0000 0.2607 0.0000
D0 2.0003
Alternative approach for the second and third marks:
Therefore option value = value of standard call option from part (ii) – the
value of the payoff under the duu path [1]
= 2.62 – 7.475 × .0833 × exp(–.05/4) = 2.00 [1]
Page 11
(iv) If the barrier is at 40, the barrier option is equivalent to the vanilla call option
in part (ii), so the price is 2.62 [1]
as it is never knocked-out. [1]
[Total 11]
This was well-answered with the majority of students scoring full marks.
Some students simply defined arbitrage in part (i) rather than stating the
conditions for the market to be arbitrage free.
Common mistakes included using an incorrect formula for the risk-neutral

probability or failing to correctly apply the annual risk free rate to the required
monthly time steps. Some students tried to price the option using the Black-
Scholes formula without appreciating that the underlying Black-Scholes
assumptions may not hold.
Q8 (i) Delta = Δ = Φ(d1)
using standard Black-Scholes notation. [1]
(ii) Δ = Φ(d1) = 0.6179 means that d1 = 0.3 [1]
So 0.3 = (log(40/45.91) + (0.02 + 0.5σ2) × 5) / σ√5 [1]
So –0.0378 – 0.6708σ + 2.5σ2 = 0 [½]
Solving the quadratic gives σ = 0.3161 or σ = –0.0478 [1]
Rejecting the negative root gives σ = 32% (or may quote variance = 10%) [½]
(iii) Under the risk-neutral probability measure Q, the fair price of the option is
ce–rT Q(S1/S0 < kS) Q(R1/R0 < kR) [2]
(iv) Under the Black-Scholes model , if the stocks are perfectly correlated then
S1/S0 = R1/R0. [1]
So if kS < kR then the option only depends on stock S and has value
ce–rT Q(S1/S0 < kS) [1]
Similarly if kS > kR then the option only depends on stock R and has value
ce–r TQ(R1/R0 < kR) [½]
If kS = kR then the option can be defined in terms of the price of either stock as
ce–rT Q(S1/S0 < kS) = ce–rT Q(R1/R0 < kS) [½]
Page 12
So overall the option can be defined in terms of the lower of kS and kR , and
either of the stock increases, i.e. has value
ce–rT Q(R1/R0 < min(kS,kR)) = ce–rT Q(S1/S0 < min(kS,kR)) [1]
[Max 3]
(v) ce–rT Q(ST/S0 < kS) Q(RT/R0 < kR)
= 50e–0.02 Q(ST/S0 < 0.8) Q(RT/R0 < 0.6)
= 50e–0.02 Q(S1 < 0.8 × 40) Q(R1 < 0.6 × 30) [1]
= 50e–0.02 (1 – Φ((log(S1/0.8S1) + (r – 0.5σS2))/σS)) (1– Φ((log(R1/0.6R1)

+ (r - 0.5σR2))/σR)) [1]
= 50e–0.02 (1 – Φ((log(1/0.8) + 0.02 – 0.5 × 0.322)/0.32) (1 – Φ((log(1/0.6)

+ 0.02 – 0.5 × 0.15)/√0.15) [½]
= 50e–0.02 (1 – (0.59982)) (1 – Φ(1.1769)) [½]
= 50e–0.02 (1 – 0.7257) (1 – 0.88039) [1]
= $1.61 (using σ = 0.32, or $1.59 using an exact σ = 0.3161) [1]

[Total 15]
Most students scored full marks in parts (i) and (ii). A number of students
used trial and error to find the volatility instead of simply solving the quadratic
equation.
Students struggled with parts (iv) and (v) with many only scoring low marks.
Most students calculated part (v) using the distribution of the share price
rather than Φ(d2) as an alternative solution. The best students showed their
workings so that some marks could be awarded even if the final answer was
not correct.
Q9 (i) (a) It incorporates mean reversion [½]
It is time homogenous, i.e. the future dynamics of r(t) only depend

upon the current value of r(t) rather than what the present time t
actually is. [1]
It is arbitrage free. [½]
It allows negative interest rates. [½]
Page 13
It is easy to implement since the characteristic functions of all related

quantities are available. [1]
It has constant volatility [½]
[Max 2]
(b) It incorporates mean reversion…. [½]

… is arbitrage free… [½]
… and time homogenous. [½]
Volatility depends on the level of the rates: it is high/low when rates

are high/low. [1]
It does not allow negative interest rates. [½]
However it is more involving to implement than Vasicek model [½]

as it is linked to the chi-squared distribution. [½]
It is a one factor model [½]

[Max 2]
(ii) Use Itô’s lemma on the auxiliary process Xt = e at rt : [1]
dX d2X dX
= eat , 2 = 0, = aeat rt . [1]
dr dr dt
And so Itô gives:
dX t = [e at a (b − rt ) dt + ae at rt ]dt + e at σdWt . [1]
And hence:
dX t = deat rt = abeat dt + σeat dWt . [½]
By direct integration from 0 to t, it follows that:
t
at
( at
)
e rt = r0 + b e − 1 + σ  e as dWs [1]
0
t
and hence, as required, rt = r0e − at
(
+ b 1− e − at
) + σe − a( t − s )
dWs . [½]
0
[Max 4]
[Alternatively using an integrating factor of eat gives the same result.]
Page 14
(iii) From Result 3.2 of the Core Reading, rt follows a Normal distribution [1]
with mean:
(
Ert = r0 e − at + b 1 − e − at ) [1]
and variance
 t 
2
= E   σ e ( ) dWs  
2 − −
Var ( rt ) = E ( rt − Ert )
a t s
[1]
  
  0  
t
−2 a( t − s )
= σ 2 e ds [1]
0
σ2
=
2a
(
1 − e −2 at . ) [½]
[Max 3]
[Total 11]
This was standard bookwork that was well-answered by the majority of

students, though some students gave formulae for the models despite being
told not to in the question.
Common mistakes in calculating the variance were to forget to square the

integrand or change the differential from dWs to ds.
Q10 (i) Under the Merton model the value at redemption is min(F(T), £100m), where
F(t) is the gross value of the company at time t. [1]
Thus the value at time 0 is:
e−5r E[min(F(5),100)] [1]
= e−5rE[F(5) − max(F(5) − 100,0)] [1]
(where the expectation is under the risk-neutral measure).
Thus the value at time 0 equals F(0) – C. [1]
where C is a call option on the total value of the company with strike £100m
and time to maturity five years. [1]
[Max 4]
Page 15
Alternatively:
In the Merton model, assuming no other claims on the company’s assets:
F(0) = E(0) + B(0) [1]
So: B(0) = F(0) – E(0)
At time T, if:
• F(T) ≥ L, the shareholders will repay the debt and E(T) = F(T) – L [½]
• F(T) < L, the shareholders will default and E(T) = 0 [½]
i.e. the shares are equivalent to a European call option on the assets of the
company with maturity T and a strike price equal to the par value of the debt,
L, and a current price of C. [1]
So: B(0) = F(0) – C [1]
where C is a call option with a term of five years and a strike price of £100m.
[1]
[Max 4]
(ii) Put-call parity gives a share price of £105.37 (= 27.55 + 100e–0.05 – 17.3) [1]
hence the total value of all shares in issue is £105.37m. [½]
The total bond value is therefore £200m – £105.37m = £94.63m. [1]
This is £94.63 per £100 nominal. [½]
(iii) The sensitivity of the share price to a change in the company’s gross value is
dSt/dVt . [1]
If we regard St as a call option on the asset value Vt (current value £200m,

strike £100m) then this is the Greek delta. [1]
From the Black-Scholes formula and the volatility above we find that
d1 = 2.145 [1]
so delta = Φ(d1) = 0.984, [1]
so a £1m increase in asset value will give a £0.984m increase in the total share
value [½]
and a £0.016m increase in total bond value. [1]
This is an increase of £0.984 in the share price and an increase of £0.016 in the
bond price per £100 nominal. [½]
[Max 4]
(iv) The current value of the company is well in excess of the nominal value of the
bonds… [½]
Page 16
… so bondholders are highly likely to receive the full nominal amount on

maturity. [½]
The bond price is therefore not very sensitive to small changes in the value of
the company… [½]
… and the share price moves almost in line with the value of the company.
[½]
(v) If the company value was lower then the value received by the bondholders at
maturity would be more likely to fall short of the nominal amount. [½]
So any change in company value would impact the bond price more (and
hence impact the share price less). [½]
[Total 14]
Part (i) was standard bookwork with most students scoring well.
Students struggled with parts (ii) and (iii), but many picked up some marks by
showing all their working.
Page 17
EXAMINATION
20 April 2017 (pm)

Core Technical
Time allowed: Three hours
booklet.
supervisor.
3. You have 15 minutes of planning and reading time before the start of this examination.
You may make separate notes or write on the exam paper but not in your answer
booklet. Calculators are not to be used during the reading time. You will then have
three hours to complete the paper.
6. Candidates should show calculations where this is appropriate.
question paper.
CT8 A2017  Institute and Faculty of Actuaries

1 (i) State the expected utility theorem. [2]
A risk averse investor makes decisions using a quadratic utility function:
U(w) = w + dw2.
(ii) Derive an upper bound for d for this investor. [2]
(iii) Explain why the investor can only use this utility function to make decisions
over a limited range of wealth, w. Your answer should include a statement of
this range. [2]
The investor states that the upper limit of wealth where she can use this utility
function is w = $1,000.
(iv) Determine the value of d in the investor’s utility function. [1]
The investor wins a prize of $250 in a gameshow. She is then offered the opportunity
to exchange this prize for a larger prize of $600 if she can answer one more question
correctly. However, she will receive no prize at all if she gets the question wrong.
She estimates her chances of answering the question correctly to be 50%.
(v) Determine whether the investor should take this opportunity to exchange. [3]
[Total 10]
2 Describe the empirical evidence relating to the continuous-time lognormal model for
security prices. [8]
CT8 A2017–2
3 Consider a non-dividend-paying security with price St at time t. The security price
follows the stochastic differential equation:
dSt = St(μdt + σdZt)
where:
• Zt is a standard Brownian motion

• μ = 16% per annum
• σ = 25% per annum
• t is the time from now measured in years
• S0 = 1
(i) Derive the distribution of St. [4]
An investor has taken out a house loan, with a repayment of £100,000 due in six
years’ time.
(ii) Determine the amount that the investor would need to invest in the security to
give a 75% probability of having an investment value of at least £100,000 in
six years’ time. [4]
The investor only has £50,000 available, which he invests in this security at time
t = 0.
(iii) Calculate the following risk measures applied to the difference between the
value of the security and £100,000 at time t = 6:
(a) 90% Value at Risk relative to £100,000

(b) expected shortfall or surplus relative to £100,000
[5]
(iv) Comment on the implications for the investor of your answers to part (iii). [2]
(v) Suggest two changes that the investor might therefore make to his portfolio.
[2]
[Total 17]

4 (i) Define the following Greeks algebraically:
(a) delta
(b) vega
(c) theta
(d) gamma
[4]
Consider a call option with price ct at time t (in years) written on an underlying non-
dividend-paying asset with price St at time t and volatility σ.
Using Taylor’s expansion, it can be shown that the change in value of the option is
approximately given by:
dct = delta × dSt + 0.5 × gamma × (dSt)2 + theta × dt + vega × dσ
At time t = 0, the underlying asset price is €23 and the volatility is 20% per annum.
The option is priced at €6.17 and has the following properties:
• delta = 0.822
• vega = 0.104
• theta = –0.855
• gamma = 0.033
At time t = 1, the security price has fallen to €20 and its volatility is now 15% per
annum.
(ii) Estimate the value of the call option at time t = 1. [2]
The delta of a call option is always positive, whilst the delta of a put option is always
negative.
(iii) Justify this result. [2]
The vega of both call and put options is always positive.
(iv) Justify this result. [1]

[Total 9]
CT8 A2017–4
5 Consider a three-period binomial tree model for a stock price process St , under which
the stock price either rises by 18% or falls by 15% each month. No dividends are
payable.
The continuously compounded risk-free rate is 0.25% per month.
Let S0 = $85.
Consider a European put option on this stock, with maturity in three months (i.e. at
time t = 3) and strike price $90.
(i) Calculate the price of this put option at time t = 0. [4]
(ii) Calculate the risk-neutral probability that the put option expires out-of-the-
money. [2]
(iii) Assess whether the probability calculated in part (ii) would be higher or lower
under the real-world probability measure. [No further calculation is required.]
[3]
[Total 9]
6 The market price St of a traded security satisfies the following stochastic differential
equation:
dSt = (μ − λσ) St dt + σSt dWt ,
where Wt is a standard Brownian motion under the probability measure P* .
(i) Determine the value of λ such that the discounted asset price process
St = e− rt St is a martingale under the given probability measure. [3]
(ii) Explain whether the probability measure P* is the real-world or risk-neutral

measure, for the value of λ obtained in part (i). [3]
[Total 6]

7 The current price of a non-dividend-paying share is £7 and its volatility is thought to
be 40% per annum. The continuously compounded risk-free interest rate is 5% per
annum.
A European call option on this share has a strike price of £6.50 and term to maturity
of one year.
(i) Calculate the price of this call option, assuming that the Black-Scholes model
applies. [4]
The market price for the option is actually £2.
(ii) Show that the volatility of the share implied by the true market price of the
option is 60% per annum, to the nearest 1% per annum. [6]
[Total 10]
8 (i) List the desirable characteristics of a term structure model. [3]
Let B (t , T ) be the price at time t > 0 of a zero-coupon bond which pays a value of 1
when it matures at time T.
Let F (t , S , T ) be the forward rate at time t for a deposit starting at time S > t and
expiring at time T > S.
Consider the following two investment strategies implemented at time t:
A At time t:
Purchase one zero-coupon bond maturing at time T.
Continue to hold the bond to time T.
B At time t:
Purchase α = e − F (t ,S ,T )(T − S ) zero-coupon bonds maturing at time S < T.
At time S:
Invest the redemption amount from the bond at the forward rate F (t , S , T ) and
continue to hold this deposit to time T.
(ii) Show that:
B(t ,T ) = e− F (t , S ,T )(T − S ) B(t , S ). [4]
(iii) Derive an expression for B (t , T ) in terms of the instantaneous forward rate,

using the result from part (ii). [3]
[Total 10]
CT8 A2017–6
9 Let Ri denote the return on security i in a two-factor model.
(i) Write down the return equation for this two-factor model, defining all
additional notation that you use. [2]
(ii) Describe the three main categories of multifactor models. [3]

[Total 5]
10 In a market in which the Capital Asset Pricing Model (CAPM) holds, there are two
securities with the following attributes (expressed per annum):
security A B
E(ri) 0.196 0.164
Cov(ri, rj) A 0.05 0.01
B 0.01 0.03
(i) Determine the composition of the market portfolio with expected return 18%
per annum. [2]
(ii) Calculate the beta of each security, under the assumption that the risk-free rate
of interest is 10% per annum. [2]
(iii) State the limitations of the CAPM. [3]

[Total 7]
11 (i) Describe the three main approaches to modelling credit risk. [5]
Consider the Merton model for credit risk. Let F(t) be the total value at time t of a
corporate entity which has issued zero-coupon debt with promised repayment amount
L due at time T.
(ii) State the condition under which the corporate entity is assumed to default in
this model. [1]
(iii) State the type of process that F(t) can be assumed to follow. [1]
(iv) Give an expression for the risk-neutral probability of default of the corporate
entity at time 0, defining any additional notation used. [2]
[Total 9]
END OF PAPER
CT8 A2017–7
EXAMINERS’ REPORT
April 2017

Core Technical
Introduction
exclusively.
Luke Hatter
July 2017

theories.

examination
opportunity to be awarded full marks due to relatively superficial knowledge.
2. The majority of the students seemed to struggle on the application parts of the questions,
because they were not able to use and combine the information given to them in the
question. In a few instances, students did not know how to go from the lognormal
distribution to the Normal and then to the standard Normal. Further, there is often a lack
of knowledge of how to use the distribution tables to compute probabilities (in the specific
case of this exam paper, the normal distribution), and relative inaccuracy in getting the
details right.
C. Pass Mark
Solutions
Q1 (i) The expected utility theorem states that a function, U(w), can be constructed
representing an investor’s utility of wealth, w, at some future date. [1]
Decisions are made on the basis of maximising the expected value of utility
under the investor’s particular beliefs about the probability of different
outcomes. [1]
(ii) U’(w) = 1 + 2dw, and [½]
U’’(w) = 2d. [½]
Because the investor is risk averse, we must have U’’(w) < 0 (alternatively to
satisfy the condition of diminishing marginal utility of wealth (risk aversion))
[½]
Page 2
So we must have d < 0. [½]
(iii) The condition of non-satiation requires U’(w) > 0. [½]
Hence 1 + 2dw > 0 and w < –1/(2d) [½]
So the quadratic utility function can only satisfy the condition of non-satiation
over a limited range of w:
Specifically –∞ < w < –1/(2d) [1]
(iv) 1,000 = –1/2d [½]
=> d = –1/2000 = –0.0005 [½]
(v) U(250) = 250 – 0.0005 × 2502 = 218.75 [1]
E[U(exchange)] = 0.5 × U(600) + 0.5 × U(0) [½]
= 0.5 × (600 – 0.0005 × 6002) = 210 [½]
So the investor should not accept the opportunity to exchange… [½]
… because the expected utility of the exchange opportunity is lower than that
of the prize. [½]
[Total 10]
Generally well answered. In part (i) many students covered the axioms on
which the theory is based rather than the theorem, as asked. In part (ii) some
candidates confused non satiation with risk aversion. A significant number of
candidates did not know how to reply to part (iv). In part (v) students
appeared to have difficulties distinguishing between utility of expected wealth
and expected utility of wealth.
Q2 Share prices are always positive, which is consistent with this model. [½]
The increments of share prices are proportional to the share price itself. [½]
However, estimates of σ vary widely according to what time period is considered. [1]
Examination of historic option prices suggests that volatility expectations fluctuate

markedly over time. [1]
One way of modelling this behaviour is to take volatility as a process in its own right.
This can explain why we have periods of high volatility and periods of low volatility.
[1]
Page 3
One class of models with this feature is known as ARCH: autoregressive conditional
heteroscedasticity. [½]
A more contentious area relates to whether the drift parameter μ is constant over time.
[½]
There are good theoretical reasons to suppose that μ should vary over time. [½]
For example, if interest rates are high, we might expect the equity drift, μ, to be high
as well. [½]
One unsettled empirical question is whether markets are mean reverting, or not. [½]
There appears to be some evidence for this… [½]
… but the evidence rests heavily on the aftermath of a small number of dramatic
crashes. [½]
Furthermore, there also appears to be some evidence of momentum effects. [½]
A further strand of empirical research questions the use of the normality assumptions
in market returns. [½]
Actual returns tend to have many more extreme events, both on the upside and
downside, than is consistent with such a model. [1]
While the random walk produces continuous price paths, jumps or discontinuities
seem to be an important feature of real markets. [1]
Furthermore, days with no change, or very small change, also happen more often than
the normal distribution suggests. [1]
However, whilst a non-normal distribution can provide an improved description of the

actual returns observed, the improved fit to empirical data comes at the cost of losing
the tractability of working with normal (and lognormal) distributions. [1]
Market jumps are consistent with the arrival of information in packets rather than
continuously. [½]
After a crash, many investors may have lost a significant proportion of their total
wealth; it is not irrational for them to be more averse to the risk of losing what
remains. [½]
Page 4
Many orthodox statistical tests are based around assumptions of normal distributions.
If we reject normality, we will also have to re-test various hypotheses. In particular,
the evidence for time-varying mean and volatility is greatly weakened. [1]
[Max 8]
Standard bookwork question. Overall most of the candidates described some key
points worth some marks, but not everyone covered all the necessary points to get
full marks. Most students focussed on the appropriateness of the normality
assumptions. A few students instead discussed Brownian motion rather than the
lognormal model.
Q3 (i) Using Ito’s Lemma:
dlogSt = 1/St dSt – 1/(2St2)(dSt)2 [½]
= (μ – σ2/2)dt + σdZt [1]
Integrating both sides gives
logSt = logS0 + (μ – σ2/2)t + σZt [½]
=> St = S0 exp((μ – σ2/2)t + σZt) [½]
As Zt is normal, [1]
then St is lognormal [½]
with parameters
(μ – σ2/2)t = 0.12875t [½]
and σ2t = 0.0625t [½]

[Max 4]
(ii) To find the initial investment we need the 25th percentile of logSt over 6 years,
i.e. with parameters 0.12875 × 6 = 0.7725 [½]
and 0.0625 × 6 = 0.375. [½]
25th percentile is calculated from:
P( (log S6– 0.7725) / √(0.375) < X) = 0.25 under a normal distribution [1]
 X = –0.6745 [½]
Page 5
 log S6= -0.6745×√(0.375) + 0.7725 = 0.35945 [½]
 S6= 1.4325 [½]
So:
Initial investment required = £100,000 / 1.4325 = £69,806 [½]
(iii) (a) We need the 10th percentile of log S6, which is
P((log S6– 0.7725) / √(0.375) < X) = 0.1 [1]
 X = –1.2816 [½]
 log S6= -1.2816×√(0.375) + 0.7725 = –0.01232 [½]
 S6= 0.98776 [½]
So the VaR is £100,000 – (£50,000×0.98776) = £50,612 [½]
(b) E[S6] = exp(μ + σ2/2) = exp(0.7725 + 0.1875) = 2.61170 [1]
So expected value = 50,000 × 2.61170 = £130,585 [½]
Hence expected surplus of £30,585.
(iv) The investor has an expected surplus, and therefore expects to repay the
loan… [1]
… but there is also a chance of a very large shortfall. [1]
He therefore may wish to change the components of his portfolio, to reduce

the risk of not being able to pay off the loan. [1]
[Max 2]
[Note to markers: please award marks for any reasonable point which is
consistent with answers in 3(ii) and 3(iii) – even if those results were wrong.]
(v) The investor might move his investments to an asset with a lower expected
return but also lower variance. [1]
The investor might decide to diversify his portfolio between a large number of
different securities. [1]
The investor might decide to pay off some of the loan early. [1]
Page 6
The investor might decide to buy an insurance product rather than using
securities. [1]
[Max 2]
[Total 17]
Large variety of answers on this question where well prepared candidates

scored full marks and less prepared candidates struggled. The main issues
seem to be the parts of questions requiring calculations. In part (iii) most
students struggled to calculate the VaR or calculated it at the 10% level rather
than 90%, as asked. For both part (iv) and (v) most of the candidates did not
manage to give valid points.
Q4 (i) Let f be the value of the derivative, S the price of the underlying, σ its
volatility, T maturity of the derivative and t current time.
f f
(a)   (t St ) [1]
s s
(b)   f [1]

f
(c) Either   or [1]
t
2 f
(d)  [1]
s 2
(ii) Change in value of option = 0.822×(–3) + 0.5 × 0.033 × (–3)2 – 0.855× 1 +

0.104 × (–0.05) = –3.178 [1]
So new value of option = 6.17 – 3.178 = €2.992 [1]
(iii) The delta for a call option is always positive because an increase in the share
price makes an option to buy the share for a set price more valuable. So as the
share price increases, the call option price also increases, hence the relative
change (the delta) is positive. [1]
Similarly, the delta for a put option is always negative because an increase in
the share price makes the option to sell the share for a set price less valuable.
So as the share price increases, the put option price reduces, hence the relative
change (the delta) is negative. [1]
Page 7
(iv) The more volatile an asset is, the more valuable the choice offered by an
option. [1]
[Total 9]
Generally well answered by most candidates (especially part (i)). However

some candidates failed to calculate the correct figures in part (ii). In part (iii)
the majority of candidates assumed that Black Scholes applied which is why
full marks were not awarded.
Q5 (i)
Stock tree
time 0 1 2 3
85.00 100.30 118.35 139.66
72.25 85.26 100.60
61.41 72.47
52.20
The price C0 of the option is computed via risk-neutral valuation; let ̂ denote
the risk-neutral probability of an up movement, then:
e0.0025  0.85
pˆ   0.4621 [1]
1.18  0.85
And
3
 3
 
3 k 
C0  e  rT    pˆ k 1  pˆ  K  S0u k d 3k [2]
k
k 0  

 e 0.0075 17.53  3 pˆ 1  pˆ   37.80  1  pˆ 
2 3
  12.82 [1]
[Max 4]
The detailed working are provided below – in case attempts to answer this
question go through the whole tree [and it carries same marks as above].
PUT
time 0 1 2 3
12.82 5.05 0.00 0.00
19.55 9.41 0.00
28.36 17.53
37.80
Page 8
(ii) Risk-neutral probability that the put option expires out-of-the-money =

P  S3  K  [1]
Hence P  S3  K   3 pˆ 2 1  pˆ   pˆ 3  0.4433 [1]
(iii) Under the risk-neutral probability the expected rate of return on the stock is
the risk-free rate of return, i.e. 0.25% per month. [1]
Under the real-world probability measure instead, the stock is expected to earn
a much higher rate of return… [1]
… in order to justify the higher risk it carries compared to the risk-free bond
[1]
Consequently we would expect the probability of the put option to expire out-
of-the-money to be higher than 0.4433. [1]
[Max 3]
[Total 9]
Generally well answered, although not as well answered as binomial tree

questions in past exams. A common mistake was failing to appreciate the
number of possible combinations comprising the probability in part (ii). There
were a lot of numerical slips; candidates struggled with the reasoning required
for part (iii).
Q6 (i) The SDE of St  e rt St is:
dSt     r    St dt  St dWˆt , [1]
For the martingale property to hold, set the drift to zero [1]
which implies  
  r  . [1]

(ii) By substitution of the value of λ in the given SDE we obtain
dSt  rSt dt  St dWt , [1]
In other words the expected rate of return on the stock is given by the risk-free
rate of interest [1]
Page 9
Hence the probability measure P* is risk-neutral [1]

[Total 6]
Large variety of answers on this question: only well prepared candidates

scored well – some candidates did not answer. In particular, the main
difficulty was recognising that for the martingale property to hold, the drift
needed to be zero.
Q7 (i) Data: S0  7, K  6.50,   40% p.a., T  1 year, r  5%.
The Black-Scholes formula returns:
d1 = 0.5103 [½]
d2 = 0.1103 [½]
N(d1) = 0.6951 [½]
N(d2) = 0.5439 [½]
So C0  7  0.6951  6.50e 0.05  0.5439 [1]
= 1.50 [1]
(ii) 60% returns a call price of 1.995, following the same calculations as in part (i)
d1= 0.5068, [½]
d2 = -0.0932, [½]
N(d1) = 0.6939, [1]
N(d2) = 0.4629 [1]
We need to check that the volatility is closer to 60% than 61%, so we calculate
the call price with a volatility of 60.5%. We need this price to be larger than 2.
[1]
60.5% returns a call price of 2.0073

d1= 0.5076, [½]
d2 = –0.0974, [½]
N(d1) = 0.6941, [1]
N(d2) = 0.4612 [1]
Page 10
Therefore, since the 60% result is smaller than 2 and the 60.5% result is larger
than 2, the implied volatility is 60% to the nearest 1%. [1]
[Max 6]
[Total 10]
Some struggled to get the d1 and d2 terms of the Black-Scholes formula

correctly. Many candidates just verified 60% as the volatility instead of
proving that this was the closest volatility to within 1%. Any meaningful values
used by candidates for verification purposed has been awarded accordingly
Q8 (i) The model should be arbitrage-free. [½]
Interest rates should ideally be positive. [½]
Interest rates should exhibit some element of mean reversion. [½]
The model should be computationally tractable / produce simple formulae for

bond and option prices. [½]
It should produce realistic dynamics. [½]
It should give a full range of possible yield curves. [½]
It should fit historical data. [½]
Can be calibrated easily to current market data. [½]
Flexible to cope with a range of derivatives. [½]

[Max 3]
(ii) Both strategies pay a value of 1 at time T [1]
By the no arbitrage principle… [1]
… if they have the same value at time T then they must have the same value at
time t [1]
Hence B(t,T) =  B(t,S) [1]
and so the requested relationship follows, with α = e–F(t,S,T)(T–S). [½]

[Max 4]
(iii) From the expression in part (ii), it follows that:
log B  t ,T   log B  t , S 
F t, S ,T    . [1]
T S
Page 11
The instantaneous forward rate is defined as f  t ,T   lim F  t , S , T  , [½]

S T
i.e.
log B  t ,T   log B  t , S   log B  t ,T 

f  t ,T    lim  . [1]
S T T S T
Solving the last equality with respect to the ZCB price, we obtain:
T
  f  t , s ds
B  t ,T   e t [½]
[Total 10]
Large variety of answers and only well prepared candidates scored well. In
part (i) many candidates failed to list all the key points; in part (ii) a number of
candidates failed to use the given portfolio to obtain the required result and,
instead, used the result they had to show as given. In part (iii) candidates
struggled to recognise that they needed to define the instantaneous forward
rate.
Q9 (i) Ri  ai  bi,1 I1  bi,2 I 2  ci [½]
where ai and ci are the constant and random parts respectively of the
component of the return unique to security i [½]
I1 , I 2 are the changes in a set of the two indices [½]
bi,k is the sensitivity (factor beta) of security i to factor k [½]
(ii) Macroeconomic factor models
These use observable economic time series as the factors, such as the annual
rates of inflation and economic growth, short term interest rates, the yields on
long term government bonds, and the yield margin on corporate bonds over
government bonds. [1]
Fundamental factor models
These use company specific variables as the factors, e.g. the level of gearing,
the price earnings ratio, the level of research and development spending, the
industry group to which the company belongs. [1]
Page 12
Statistical factor models
These do not rely on specifying the factors independently of the historical

returns data. Instead a technique called principal components analysis can be
used to determine a set of indices which explain as much as possible of the
observed variance. [1]
[Total 5]
Well answered by most candidates. Generally candidates answered part (i)

correctly with some candidates failing to define correctly ci and some failing to
write down the equation correctly. Most of the candidates obtained marks in
part (ii), with some failing to give a proper definition of statistical models and
some failing to name the three main categories of multifactor models.
Q10 (i) The market portfolio is the weighted portfolio of the risky securities in the
market, consequently
ErM  18%  wA ErA  wB ErB [1]
As wA  wB  0.5 , then w A  wB  0.5 . [1]
(ii) From the security market line
Eri  r f
i 
ErM  r f
Therefore  A  1.2 and  B  0.8 . [1 each]
(iii) Empirical studies do not provide strong support for the model. [½]
The underlying assumptions are not realistic. [½]
Investors cannot necessarily borrow or lend unlimited amounts at the same

risk-free rate. [½]
The markets for risk assets may not be perfect. [½]
Investors may not have the same estimates of expected returns, standard
deviations and covariances of securities. There are basic problems in testing
the model since, in theory, account has to be taken of the entire investment
universe open to investors, not just capital markets. [½]
It does not account for taxes. [½]
Page 13
It does not account for inflation. [Or: some investors may measure in real
terms and some in money terms.] [½]
It does not account for situations in which there is no riskless asset. [½]
The basic model does not allow for currency risk. [Or: investors may not
measure in the same currency.] [½]
It does not consider multiple time periods. [Or: investors do not all have the
same one-period time horizon.] [½]
It does not consider optimisation of consumption over time. [½]

[Max 3]
[Total 7]
Well answered by most candidates. However, in part (ii) there was some
evidence of students not being able to calculate the market portfolio correctly
instead calculating the minimum-variance portfolio under Mean-Variance
Portfolio Theory.
Q11 (i) Structural models [½]
Structural models aim to link default events explicitly to the fortunes of the
issuing corporate entity. [1]
An example of a structural model is the Merton model. [½]
Reduced form models [½]
Reduced form models use observed market statistics rather than specific data
relating to the issuing corporate entity. [1]
The market statistics most commonly used are the credit ratings… [½]
… issued by credit rating agencies such as Standard and Poor’s and Moody’s.
[½]
The output of such models is a distribution of the time to default. [½]
Intensity-based models [½]
An intensity-based model is a particular type of continuous-time reduced form

model. [1]
Page 14
It typically models the “jumps” between different states,… [½]

… which are usually credit ratings,… [½]
… using transition intensities. [½]
[Max 5]
(ii) Default occurs if the value of the assets is not enough to cover the face value
of the debt at maturity
Or alternatively: F(T) < L. [1]
(iii) F(t) follows a geometric Brownian motion (or continuous time lognormal
model). [1]
(iv) Hence, by risk-neutral valuation, the Merton model-based probability of

default is:
 F 0  2  
 ln r q   T 
 L  2 
P  F T   L   1  N   
T 
 
 
 
where P(.) is the probability under the risk-neutral measure, F (0) is the
current value of the firm, σ is its volatility, r is the risk-free rate and q
denotes any potential payout cashflow. [2]
[Total 9]
Candidates familiar with the study material scored well. Part (i) was often
answered correctly although some candidates failed to describe properly the
main approaches and some mixed the names of the approaches. In part (iii)
many candidates missed the “q”.
Page 15
EXAMINATION
28 September 2017 (pm)

Core Technical
booklet.
supervisor.
5. Attempt all nine questions, beginning your answer to each question on a new page.
question paper.
CT8 S2017  Institute and Faculty of Actuaries

1 (i) Define the following terms:
(a) absolute dominance

(b) first order stochastic dominance
(c) second order stochastic dominance
[4]
Consider four assets which will deliver a one-year return ri on asset i with
probabilities as set out below:
P(ri = –5%) P(ri = –3%) P(ri = 0%) P(ri = +3%) P(ri = +5%)
Asset 1 0.2 0.2 0.2 0.2 0.2

Asset 2 0.3 0.2 0.1 0.2 0.2
Asset 3 0.1 0.3 0.2 0.3 0.1
(ii) Determine which type of dominance, if any, is exerted by:
(a) asset 2 over asset 3.

(b) asset 3 over asset 1.
(c) asset 1 over asset 2. [6]
[Total 10]
2 (i) Define in the context of mean-variance portfolio theory:
(a) an inefficient portfolio

(b) an efficient portfolio [2]
(ii) State the two assumptions about investor behaviour that are needed for the
existence of efficient portfolios. [1]
An investment universe includes two assets, A and B, with expected return on asset i
of ri and variance vi as set out below:
Asset i Expected return ri Variance of return vi
A rA = 0.05 vA = 0.16
B rB = 0.07 vB = 0.25
The correlation of returns is cAB = –0.2.
In an efficient portfolio, let a be the proportion which is held in asset A.
(iii) Express the portfolio variance V in terms of a quadratic function in a, showing

your workings. [3]
CT8 S2017–2
Let R be the expected return on the portfolio.
(iv) Express the portfolio variance V in terms of a quadratic function in R, using

your result from part (iii) and showing your workings. [Your expression
should not include a.] [3]
The expression in part (iv) represents the efficient frontier.
An investor uses a utility function that gives rise to an indifference curve

V = 16R – 200R2.
(v) Determine the two portfolios on the efficient frontier that also lie on the
investor’s indifference curve. [4]
(vi) Comment on the implications for part (v) if short selling is not allowed in the
market. [2]
[Total 15]
3 Consider a European call option with price ct written on an underlying non-dividend-

paying security with price St at current time t.
(i) State whether each of the following changes in underlying factors would
increase or reduce the price of this option:
(a) a fall in the price of the underlying security

(b) an increase in the strike price of the option
(c) an increase in the volatility of the underlying security price
(d) a fall in the risk-free rate of interest
[You should assume that each change occurs on a standalone basis, i.e. all
other factors are unchanged.] [2]
(ii) Explain each of your statements in part (i). [4]
Consider a European put option with price pt written on the same underlying security,
with the same strike price K and the same maturity T as the call option described
above.
The continuously compounded risk-free rate of interest is r.
(iii) Write down a formula that relates the values of ct and pt. [1]
The call option has value £0.50 at time t = 0, and the put option has value £1.00. Both
options are written on a security with current value S0 = £5, and both options have
strike price £6.00 and maturity T = 3 years.
(iv) Determine the continuously compounded risk-free rate r. [2]
(v) Suggest, with justification, how the formula in part (iii) can be rewritten as an
inequality if both options are American options. [3]
[Total 12]

4 Consider a one-period binomial tree model for the stock price process St.
Let S0 = $100 and assume that in three months’ time the stock price is either $125 or
$105. No dividends are payable on this stock.
Assume also that the continuously compounded risk-free rate is 5% per annum.
(i) Verify that this market is not arbitrage-free by considering the relationship
between the risk-free rate and the stock price movements. [2]
(ii) (a) Identify a portfolio which would generate an arbitrage profit.

(b) Calculate this profit.
[4]
Now assume that the continuously compounded risk-free rate is 20% per annum.
Consider a European put option on this stock, expiring in three months’ time and with
strike price K = $120.
(iii) Calculate the current price of this put option. [3]

[Total 9]
5 (i) State the Cameron-Martin-Girsanov theorem. [3]
(ii) State an important property of the discounted value of a security price process
under the risk-neutral measure. [1]
The price process St of a traded security satisfies the following stochastic differential
equation:
dSt = μSt dt + σSt dWt ,
where Wt is a standard Brownian motion under the real-world probability measure,

and μ and σ are constants, with σ > 0.
Let r > 0 be the continuously compounded risk-free rate of interest.
(iii) Show, using parts (i) and (ii), that Wt + λt is a Brownian motion under the
risk-neutral probability measure, if λ = (μ − r ) . [3]
σ
(iv) Calculate the value of λ in the case in which μ = 0.04 + r and σ = 0.4. [1]
Another traded asset has a price process satisfying the stochastic differential equation
dAt = (0.06 + r ) At dt + γAt dWt .
(v) Determine the value of the volatility coefficient γ, using your result from
part (iv). [2]
[Total 10]
CT8 S2017–4
6 (i) Write down an expression for the price of a derivative in a Black-Scholes
market in terms of an expectation under the risk-neutral measure, defining any
additional notation that you use. [3]
Consider an option on a non-dividend-paying stock when the stock price is £50, the
exercise price is £49, the continuously compounded risk-free rate of interest is 5% per
annum, the volatility is 25% per annum, and the time to maturity is six months.
(ii) Calculate the price of the option using the Black-Scholes formula, if the option
is a European call. [4]
(iii) Determine the price of the option if it is an American call. [1]
(iv) Calculate the price of the option if it is a European put. [2]
(v) Determine how the prices of the contracts in parts (ii) to (iv) would change in
the case of a dividend-paying underlying stock. [Note that you do not have to
perform any further calculations.] [3]
[Total 13]
7 (i) State the main potential drawback of the Vasicek model. [1]
(ii) Discuss the extent to which this drawback may be a problem. [3]
(iii) Explain how the Cox-Ingersoll-Ross model avoids this drawback. [3]
The Vasicek term structure model is described by the following stochastic differential
equation:
drt = a(b − rt )dt + σdWt ,
and a, b, σ > 0.
Under this model, the short rate rt follows a Normal distribution with mean
E (rt ) = r0e− at + b (1 − e− at )
σ2
and variance Var (rt ) = (1 − e −2 at ).
2a
(iv) Assess, using the information provided above, whether the model generates
interest rates that are mean reverting and, if so, the value to which they revert.
[2]
(v) Assess, using the information provided above, the relevance of the
parameter a to any mean reversion. [2]
[Total 11]

8 In a market in which the Arbitrage Pricing Theory (APT) model holds, the expected
return is given by
E[ Ri ] = λ 0 + λ1bi ,1 +λ 2bi ,2 +…+ λ nbi ,n
(i) Define all the terms in this equation. [2]
Let rf denote the risk-free rate of interest.
(ii) Construct a risk-free portfolio to prove that λ0 = rf . [2]
Assume that rf = 0.075. Consider a two-factor model (i.e. n = 2) and two well-
diversified portfolios (P1 and P2) with the following features:
P1 P2
E[Ri] 0.18 0.15

bi,1 1.5 0.5
bi,2 0.5 1.5
(iii) Determine the values of λ1 and λ2. [3]
Suppose that in the market there is another portfolio with the following features:
E[ R3 ] = 0.16, b3,1 = 0.75, b3,2 = 0.7.
(iv) Comment on the feasibility of such a portfolio under the APT model
assumptions. [3]
[Total 10]
9 Consider the Merton model for credit risk.
Assume that a firm has issued a zero-coupon bond maturing in five years’ time with
maturity value €100m, and that the current value of the firm’s assets is €110m.
Further assume that the estimated volatility of the firm’s assets is 25% per annum and
the risk-free rate of interest is 2% per annum continuously compounded.
(i) Show that the current value of the debt of the firm is €76.88m. [5]
(ii) Calculate the yield to maturity of the debt. [3]
(iii) Calculate the credit spread on the debt. [2]

[Total 10]
END OF PAPER
CT8 S2017–6
EXAMINERS’ REPORT
September 2017

Core Technical
Introduction
The Examiners’ Report is written by the Principal Examiner with the aim of helping
candidates, both those who are sitting the examination for the first time and using past papers
as a revision aid and also those who have previously failed the subject.
The Examiners are charged by Council with examining the published syllabus. The
Examiners have access to the Core Reading, which is designed to interpret the syllabus, and
will generally base questions around it but are not required to examine the content of Core
Reading specifically or exclusively.
For numerical questions the Examiners’ preferred approach to the solution is reproduced in
this report; other valid approaches are given appropriate credit. For essay-style questions,
particularly the open-ended questions in the later subjects, the report may contain more points
than the Examiners will expect from a solution that scores full marks.
The report is written based on the legislative and regulatory context pertaining to the date that
the examination was set. Candidates should take into account the possibility that
circumstances may have changed if using these reports for revision.
Luke Hatter
December 2017

1. The aim of the Financial Economics subject is to develop the necessary skills to
construct asset liability models and to value financial derivatives. These skills are
also required to communicate with other financial professionals and to critically
evaluate modern financial theories.
2. The marking approach for CT8 is flexible in the sense that different answers to
those shown in the solution can earn marks if they are relevant and appropriate.
Marks for the methodology are also awarded.
B. General comments on student performance in this diet of the examination
1. Students performed relatively well on bookwork questions, although many missed

the opportunity to be awarded full marks due to relatively superficial knowledge.
2. Some students seemed to struggle on the application parts of the questions,

because they were not able to combine and use the information given to them in
the question.
C. Pass Mark
Solutions
Q1 (i) (a) Absolute dominance exists when one investment portfolio [1]
provides a higher return than another in all possible circumstances.
(b) The first order stochastic dominance theorem states that, [½]
assuming an investor prefers more to less, A will dominate B
(i.e. the investor will prefer portfolio A to portfolio B) if:
FA ( x)  FB ( x), for all x, and [½]
FA ( x)  FB ( x) for some value of x. [½]
(c) The second order stochastic dominance theorem applies when the [½]
investor is risk averse, as well as preferring more to less.
In this case, the condition for A to dominate B is that
x x
 a FA ( y)dy   a FB ( y)dy, for all x, [½]
Page 2
with the strict inequality holding for some value of x, [½]
and where a is the lowest return that the portfolios can possibly [½]
provide.
[Max 4]
(ii)
PDF –5% –3% 0% 3% 5%
1 0.2 0.2 0.2 0.2 0.2
2 0.3 0.2 0.1 0.2 0.2
3 0.1 0.3 0.2 0.3 0.1
CDF –5% –3% 0% 3% 5%

1 0.2 0.4 0.6 0.8 1
2 0.3 0.5 0.6 0.8 1
3 0.1 0.4 0.6 0.9 1
[1 mark for CDF table]
∫CDF –5% –3% 0% 3% 5%

1 0 0.004 0.016 0.034 0.05
2 0 0.006 0.021 0.039 0.055
3 0 0.002 0.014 0.032 0.05
[2 marks for ∫CDF table]
(a) None [1]
(b) Second order [1]
(c) First order [1]
Most students knew the definitions of the different types of

dominance and stated them clearly, either in words or
formulae. Fewer students were able to determine the types of
dominance exhibited by the assets in part (ii). In particular,
very few students integrated the CDF correctly to check for
second order dominance. There was one anomaly in the
question whereby students were asked to “consider four assets”
when only three were given in the table.
Page 3
Q2 (i) (a) A portfolio is inefficient if the investor can find another [½]
portfolio with the same expected return and lower variance,
or the same variance and higher expected return. [½]
(b) A portfolio is efficient if the investor cannot find a better one [½]
in the sense that it has both the same or higher expected return
and the same or lower variance. [½]
(ii) The assumptions are:
(a) Investors are never satiated. [At a given level of risk, they [½]
will always prefer a portfolio with a higher expected return to
one with a lower return.]
(b) Investors dislike risk. [For a given level of return, they will [½]
always prefer a portfolio with lower expected variance to one
with higher variance.]
(iii) V = a2VA + (1 – a)2VB + 2a(1 – a)(VAVB)0.5CAB [1]
= 0.16a2 + 0.25 (1 – a)2 – 2a(1 – a) (0.16  0.25)0.5  0.2 [1]
= 0.49a2 – 0.58a + 0.25 [1]
(iv) R = aRA + (1 – a)RB [1]

= –0.02a + 0.07
So R2 = 0.0004a2 – 0.0028a + 0.0049 [1]
So V = 1225R2 – 142.5R + 4.2225 [1]
(v) 1225R2 – 142.5R + 4.2225 = 16R – 200R2 [1]
So R = 0.0670 or 0.0442 [2]
Hence a = 0.1497 or 1.2889 [1]
(vi) The second solution implies a proportion of –0.2889 invested in asset B [1]
so would not be allowed, hence only the first solution would remain. [1]
Page 4
This question was largely well-answered. Many students made

mistakes in the algebra but these were penalised only for the
mistake itself if the remaining workings were correct. Some
students found the point where the efficient frontier and the
indifference curve were tangential but did not check that they
touched.
Q3 (i) (a) reduce [½]
(b) reduce [½]
(c) increase [½]
(d) reduce [½]
(ii) (a) This is because there is a lower intrinsic value (or, where the [1]
intrinsic value is currently zero, a smaller chance that the option
is in-the-money at maturity).
(c) This is because there is again a lower intrinsic value, or a smaller [1]
chance that the option is in-the-money at maturity.
(d) This is because the higher the volatility of the underlying share, [1]
the greater the chance that the share price can move significantly
in favour of the holder of the option before expiry.
(e) This is because the money saved by purchasing the option rather [1]
than the underlying share has to be invested at this lower rate of
interest, thus decreasing the value of the option.
(iii) ct + Ke–r(T–t) = pt + St [1]
(iv) 0.5 + 6e–3r = 1 + 5 [1]
=> r = 2.9% p.a. [1]

(v) As no dividend is paid, American call options will never be exercised [1]
before maturity. Therefore, their price should be the same as for
European call options with the same characteristics.
It is sometimes optimal to exercise an American put option early, so the [1]

value of an American put option can be higher than a European put option.
and the formula becomes an inequality:

[1]
ct + Ke–r(T–t) ≤ pt + St
Page 5
Almost all students stated the correct option price changes in

part (i), and most gave good reasons for the changes in part
(ii). The question did not state that the assumptions underlying
the Black-Scholes formula applied so explanations based on
this were not valid. Fewer students were able to explain clearly
what the impact was of changing the options to be American ,
though well-prepared students built an answer around the
impact of exercising each option early.
Q4 (i) The market is arbitrage free if and only if there exists a probability [1]
measure under which discounted asset prices are martingales
In this case, the probability exists iff d  ert  u [1]
The given market does not satisfy this property as [1]

er t  1.0126  d  1.05  u  1.25
Alternatively, it can be seen that investment in the stock will gain [1]
more than the risk-free rate…
… under any possible outcome / with no downside risk. [1]

[Max 2]
(ii) (a) The investor could buy the stock at 100 by borrowing money [1]
at the risk-free rate of interest.
In three months, the investor then could sell the stock and [1]
repay the loan + interest
(b) This would result in a profit of either 23.74 – in the case in [1]
which the stock is worth 125
Or 3.74 – in the case in which the stock is worth 105 [1]
(iii) The price C0 of the option is computed via risk-neutral valuation; let 𝑝̂ denote
the risk-neutral probability of an up movement, then
e0.20.25  1.05
pˆ   0.0064 [1]
1.25  1.05
C0  e0.20.25 (15  (1  pˆ ))  14.18. [2]
Page 6
This question was broadly well-answered. Part (ii) caused the

most difficulty, with some students struggling to construct valid
portfolios or trying to explain in general terms without any
specific portfolios. There were also some solutions to part (iii)
involving probabilities either less than zero or greater than one,
which were clearly not valid.
Q5 (i) Suppose that Z t is a standard Brownian motion under P. [1]
Furthermore, suppose that  t is a previsible process. [½]
Then there exists a measure Q equivalent to P [½]
t
and where Zt  Zt  0  s ds is a standard Brownian motion under Q. [1]
Conversely, if Z t is a standard Brownian motion under P and if Q is

equivalent to P then there exists a previsible process  t such that
t
Zt  Zt   0  s ds is a Brownian motion under Q. [1]
[Max 3]
(ii) Under the risk-neutral probability measure, the discounted value of asset
prices are martingales. [1]
(iii) First determine the SDE of St  e rt St :
dSt  (  r ) St dt  St dWt , [1]

Then change the Brownian motion and the probability measure (using the
CMG theorem) so that the above reads
dSt  (  r  )St dt  St dWˆt , [1]
(  r )
If    then the drift term is zero, as required (for a martingale). [1]
0.04  r  r
(iv)   0.1 [1]
0.4
Page 7
(v) The risk driver is the same, therefore the market price of risk is identical. [1]
0.06 0.06
Hence     0.6 . [1]
 0.1
This question was answered poorly on the whole. Few students

knew the Cameron-Martin-Girsanov theorem well enough to
score full marks. The later parts were answered better, with
many students picking up marks in parts (iv) and (v).
Q6 (i) Ct  E (e r (T t )CT Ft ) [1]
where Ft denotes the filtration at time t > 0, [½]
CT is the payoff under the derivative [½]
at maturity time T, [½]
Ct is the derivative value at time t, [½]
and the expectation is taken under the risk-neutral martingale measure. [½]
[Max 3]
Data: S  50; K  49; r  5%;   25%;T  0.5

(ii) The Black-Scholes formula returns:
d1 = 0.3441 [½]
d2 = 0.1673 [½]
N(d1) = 0.6346 [½]
N(d2) = 0.5664 [½]
So Call  50  0.6346  49e0.050.50  0.5664  4.66 [2]
(iii) Same as European call (as the stock is non-dividend-paying), i.e. 4.66 [1]
(iv) Using put-call parity (or otherwise):
pt = ct + Ke-r(T-t) - St [1]
Hence pt = 2.45. [1]
Page 8
(v) If the stock is dividend-paying, the payment of the dividends would cause the
value of the underlying asset to fall – which follows from the no arbitrage
principle [1]
Alternatively: in valuing the option we must take account of the fact that
dividends are payable on the underlying asset which do not feed through to the
holder of the option. [1]
Therefore the price of the European call would decrease… [½]
… since by buying the option instead of the underlying share the investor
forgoes the income [½]
Similarly, the price of the European put would increase [½]
The American call would now be more expensive than the European call due
to potential early exercise opportunity [1]
[Max 3]
This question was answered well by most students. There were

a number of numerical mistakes in the Black-Scholes
calculations in part (ii) despite this being a very common skill
examined in CT8. Knowledge of how dividends affect the
option pricing was weak. There were also many cases of
students rounding d1 and d2 too aggressively in the Black-
Scholes calculations resulting in a materially incorrect answer.
Page 9
Q7 (i) It allows negative interest rates. [1]
(ii) The extent of the problem depends on the probability of negative interest
rates… [½]
… within the timescale of the problem in hand (or, for example, less of an
issue if the time horizon is short)… [1]
… and their likely magnitude if they can go negative. [1]
It also depends on the economy being modelled, as negative interest rates have
been seen in some countries. [1]
[Max 3]
(iii) The CIR model does not allow interest rates to go negative. [1]
This is because the volatility under the CIR increases in line with the square
root of r(t). [1]
Since this reduces to zero as r(t) approaches zero… [½]
… and provided the volatility parameter is not too large… [½]
… r(t) will never actually reach zero. [½]
… provided σ2 ≤ 2αµ
[Max 3]
(iv) Letting t   , we note that the mean converges to b. [1]
Hence interest rates under the model are mean reverting [½]
To the long-run mean b [½]
(v) Letting t   , we note that the variance converges to 2 / 2a [1]
Hence, the variance of the short rate is inversely proportional to a [½]
This implies that the convergence of the rate to the long run mean b is faster
the bigger a [1]
So a controls the speed of the mean convergence [1]
[Max 2]
Page 10
Most students here knew that the Vasicek model allows negative
interest rates, and were able to explain why the Cox-Ingersoll-
Ross model does not. Part (ii) required students to think about
why negative interest rates might (or might not) be a problem in
the real world –many students just repeated bookwork about the
Vasicek model and scored no marks. Parts (iv) and (v) were
answered fairly well, though not all students worked through
the algebra correctly.
Q8 (i) E[ Ri ] is the expected return on security i [½]
bi ,k is the response of (or sensitivity of) the rates of return on security i to

factor k [1]
 k is the risk premium per unit of exposure corresponding to factor k [1]

[Max 2]
(ii) The risk-free portfolio has zero exposure to all risk factors [1]
i.e. bi,k  0 for all i, k [½]
And the expected return on the risk-free portfolio is r f [½]
From which the result follows: λ0 = r f . [½]

[Max 2]
(iii) We need to solve the linear system: [1]
0.18  0.075  1.51  0.5 2

0.15  0.075  0.51  1.5 2
which returns 1  0.06,  2  0.03 [1 mark each]
(iv) The given portfolio represents an arbitrage opportunity [1]
as the given expected return does not satisfy the given APT equation [1]
Indeed E[ R3 ]  0.075  0.06  0.75 0.03  0.7  0.141 [1]
It is therefore not a feasible portfolio… [1]
… under an assumption of no arbitrage [½]

[Max 3]
Page 11
This question was answered well by most students. Parts (ii)

and (iii) caused some difficulty, though most students spotted
what they needed to do.
Q9 The Merton model for credit risk is based on the Black-Scholes formula. Hence.
(i) The current value of the debt, say D0 , of the firm is the value of a risk-free
zero coupon bond with the same face value and maturity of the firm debt
corrected by the cost of default. That is, where F0 is the value of the firm’s
assets (=110) and L is the face value of the debt (=100):
D0  L e rT  ( L e rT N (d 2 )  F0 N (d1 )) [1½]
Alternative approach: Or equivalently it is the current value of the firm’s

assets less the value of equity, where the latter is the value of a call option on
the assets of the company with strike price equal to L, i.e.:
D0  F0  ( F0 N (d1 )  L e rT N (d 2 )) [Alternative 1½]
d1 0.6289
d2 0.0699
[½ for each]
N(d1) 0.7353
N(d2) 0.5279
N(-d1) 0.2647
N(-d2) 0.4721
[½ for each, but only give for either + or –, i.e. Max 1]
Giving overall value of either (in €m):
100 e0.1  (100 e0.1  0.4721  110  0.2647) or

110  (110  0.7353 100 e0.1  0.5279) [1]
= €76.88m as required [½]
(ii) The yield to maturity solves D0  Le yT [1]
i.e. 76.88 = 100 e5y [1]
Consequently y  5.26% per annum [1]
Note: may quote 5.25% if didn’t round to 76.88.
Page 12
Note to markers: give full marks for correct answer, even if no working
shown.
(iii) Credit spread: y  r [1]
= 3.26% per annum. [1]
Most students applied the Merton model correctly, though some

just explained the model. Parts (ii) and (iii) were also
answered well by most students.
Page 13
EXAMINATION
24 April 2018 (am)

Core Technical
1. Enter all the candidate and examination details as requested on the front of your
answer booklet.
supervisor.
question paper.
CT8 A2018 © Institute and Faculty of Actuaries

1 A horse racing fan assesses her utility of wealth using the utility function
U(w) = 2(w 0.5–1).
(i) Prove algebraically that the horse racing fan is:

(a) non-satiated
(b) risk averse. [2]
(ii) Prove that the horse racing fan exhibits constant relative risk aversion. [2]
The horse racing fan is attending a race and she intends to place bets on two horses.
The table below shows the pay-out per £1 bet on each of these horses if it wins the
race, and the investor’s estimated probabilities of each horse winning the race. The
pay-out is the total paid and is not in addition to the bet being returned.
Horse Winning pay-out per £1 bet Probability of winning

A £1.69 60%
B £6.25 10%
The horse racing fan has total wealth of £1,000 and she will bet all of her wealth on
this race. Negative bets are not allowed.
(iii) Calculate the amount she should bet on each horse to maximise her expected
utility of wealth. [7]
(iv) Calculate the expected wealth resulting from the bets in part (iii). [1]
(v) Explain how and why this differs from the utility of the horse racing fan’s
initial wealth. [2]
[Total 14]
2 Describe the difficulties in estimating parameters for asset pricing models. [5]
3 The value of an investment asset follows the equation A(t) = exp(Bt ), where Bt follows
standard Brownian motion.
(i) State the five defining properties that apply to Bt as a standard Brownian
motion.[5]
An actuarial student invests $1,000 in the asset at time 0.
(ii) Calculate the expected value of this investment at time 5. [2]
(iii) Calculate the probability that the value of the student’s investment is less than
$10,000 at time 5. [2]
[Total 9]
CT8 A2018–2
4 Mr and Mrs Jones both wish to buy stocks in Widgets Inc. They don’t have enough
money right now, so they are considering buying either forwards or options on the
stocks, both with a term of 4 years.
The stock price at time 0 is £10 with standard deviation of 12% per annum. The stock
does not pay any dividend. The continuously compounded risk-free rate of interest is
5% per annum.
(i) Calculate the 4 year forward price on one stock. [1]
(ii) Calculate the price at time 0 of a 4 year call option on one stock with a strike
price of £12.21. [3]
Mrs Jones enters into one forward contract, while Mr Jones buys one call option. At
time 4 the stock is worth £12.
(iii) Calculate the accumulated profit or loss at time 4 for Mrs Jones. [1]
(iv) Calculate the accumulated profit or loss at time 4 for Mr Jones. [2]
(v) Explain why Mr Jones makes a loss despite having an option that does not
force him to buy the stock. [2]
(vi) Calculate the range of stock prices at time 4 which would leave Mr Jones
better off than Mrs Jones. [3]
[Total 12]
5 Consider a zero-coupon bond Bt with three years to maturity. The bond pays $100
at maturity if it has not defaulted, or $30 if it has defaulted. The continuously
compounded risk-free rate is r. In a two-state model the default intensity is l under
the probability measure P, and the bond price is:
Bt = 30e–r(3 – t) if the bond has already defaulted by time t, or

Bt = e–r(3 – t)(30(1 – e–l(3 – t)) + 100e–l(3 – t)) if the bond has not yet defaulted by time t
(i) Show that P is an equivalent martingale measure. [4]
A derivative pays $35 at time 3 if the bond has defaulted and $0 otherwise.
(ii) (a) Determine a constant portfolio containing the bond and cash which
replicates this derivative.
(b) Derive an expression for the arbitrage-free price for the derivative at
time 0 in terms of r and l.[5]
(iii) Explain how your answers to parts (i) and (ii) are related through the value of
the portfolio in part (ii) also being a martingale. [3]
[Total 12]

6 Consider a three-period binomial tree model for the non-dividend paying stock price
process St , in which the stock price either rises by u% or falls by d % each period till
maturity. Let r denote the continuously compounded risk-free rate of interest.
(i) State the conditions under which this market is arbitrage free. [1]
Let S0 = £95 and assume this price either rises or falls by 20% each year for the
next three years. Assume also that the risk-free rate is 5% per annum continuously
compounded.
(ii) Calculate the price of a vanilla European put option with maturity in three
years and strike price 110. [4]
Assume a change in market conditions such that the same share price now either rises
or falls by 5% each year for the next three years.
(iii) Determine how this change would impact on the option price. [2]
[Total 7]
7 (i) Define a complete market. [1]
The price process of a traded security satisfies the following stochastic differential
equation
dSt = μSt dt + σSt dWt ,
where Wt is a Brownian motion under the real-world probability measure P. Let r > 0
be the continuously compounded risk-free rate of interest, with r ≠ μ.
(ii) Show that the discounted stock price e–rtSt is not a martingale under the real-
world probability measure P.[3]
(iii) Demonstrate how the discounted asset price e–rtSt can be a martingale under
an equivalent martingale measure Q. [3]
[Total 7]
CT8 A2018–4
8 The current price of a non-dividend paying stock is £65 and its volatility is 25% per
annum. The continuously compounded risk-free interest rate is 2% per annum.
Consider a European call option on this share with strike price £55 and expiry date in
six months’ time. Assume that the Black-Scholes model applies.
(i) Calculate the price of the call option. [4]
(ii) Define algebraically the delta of the call option. [1]
(iii) Calculate the value of the delta of the call option. [2]
(iv) Calculate the value of the delta of a European put option written on the same
underlying, with the same strike and maturity as above. [2]
[Total 9]
9 Consider a market with the following bonds in issue.
Principal Expire (years) Coupon Price Zero rate Forward rate

value T (annual*) R(0, T) F(0, S, T)
100 0.25 0 97.5 (a)
100 0.5 0 94.9 (b) F(0, 0.25, 0.5) = 10.81%
100 1 0 90.0 10.54% F(0, 0.5, 1) = (d)
100 1.5 8% (c) 10.68% F(0, 1, 1.5) = (e)
(* half the stated coupon is paid every 6 months)
(i) Calculate the values of (a), (b), (c), (d), (e) in the table above. [5]
(ii) Write down the stochastic differential equations of two standard models for the
short rate of interest. [2]
[Total 7]

10 Consider a call option on a non-dividend paying stock S when the stock price is £15,
the exercise price, K, is £12, the continuously compounded risk-free rate of interest
is 2% per annum, the volatility is 20% per annum and the time to maturity is three
months.
(i) Calculate the price of the option using the Black-Scholes model. [4]
(ii) Determine the (risk neutral) probability of the option expiring in the money.
[1]
A special option called a “digital cash-or-nothing” option has a payoff in three

months’ time of:
1 if ST > K
0 otherwise
(iii) Calculate the price of the digital option. [2]
(iv) Describe the limitations of the Black-Scholes model. [2]

[Total 9]
CT8 A2018–6
11 Consider a market in which the Capital Asset Pricing Model (CAPM) holds.
(i) List the assumptions, additional to those used in modern portfolio theory, of
the CAPM. [2]
(ii) Prove that the market portfolio has unit beta. [2]
In the same market as above, there are two assets with the following attributes.
Rate of return (per annum) Variance/Covariance Matrix

State Probability Asset 1 Asset 2 Asset 1 Asset 2
1 0.2 5.00% 11.00% Asset 1 0.00068 0.00102
2 0.3 10.00% 15.00% Asset 2 0.00102 0.00181
3 0.1 8.00% 12.00%
4 0.4 4.00% 5.00%
(iii) Calculate the beta of each security. [3]
(iv) Determine the value of the risk-free rate of interest which is consistent with
the results obtained in part (iii), under the assumption that the CAPM holds.[2]
[Total 9]
END OF PAPER

EXAMINERS’ REPORT
April 2018

Core Technical
Introduction
Luke Hatter
June 2018
 Institute and Faculty of Actuaries

1. Performance by candidates on this paper was, on the whole, considerably worse

than in recent sittings.
2. In general, the real differentiators in those who scored well were attention to detail
in their algebraic steps, and the breadth of knowledge in being able to score the
bookwork marks and even attempt most questions. The majority of candidates
seemed unable to gather from the text of the question the relevant information, and
translate it in the appropriate equivalent statistical concepts. For example,
candidates struggled with formulating the probability that an event occurs in
appropriate mathematical terms, and determining from the information in the
question the direct way to recover required variances and covariances. This showed
a lack of sufficient confidence with the fundamental statistical concepts which
Financial Economics so heavily relies on.
3. Students performed relatively well on bookwork questions, although many missed

the opportunity to be awarded full marks due to relatively superficial knowledge.
4. The majority of candidates seemed to struggle on the application parts of the

questions, because they were not able to use and combine the information given to
them in the question.
C. Pass Mark
Page 2
Solutions
Q1
i)
a. U’(w) = w-0.5 > 0 for w > 0 [1]
b. U’’(w) = -0.5w-1.5 < 0 for w > 0 [1]
ii) R(w) = w*(-U‘’(w)/U’(w)) = 0.5 [2]
iii) E[U] = 0.6U(1.69a) + 0.1U(6.25b) +0.3U(0) [½]

= 0.6*2*((1.69a)^0.5-1) + 0.1*2*((6.25b)^0.5-1) – 0.6 [½]
= 1.2*(1.3a^0.5-1) + 0.2*(2.5b^0.5-1) – 0.6
= 1.56a^0.5 + 0.5b^0.5 – 1.4 – 0.6 [½]
= 1.56a^0.5 + 0.5(1000-a)^0.5 –2 [½]
dE[U]/da = 0.78a^-0.5 – 0.25(1000-a)^-0.5 [1]
Setting dE[U]/da = 0 gives
0.78a^-0.5 = 0.25(1000-a)^-0.5 [½]
=> 0.78 = 0.25a^0.5(1000-a)^-0.5
=> 3.12 = (a/(1000-a))^0.5 [½]
Squaring both sides
=> 9.7344 = a/(1000-a) [½]
=> 9,734.4 = 10.7344a or 8.7344a
=> a = £906.8 or £1,114.50 [½]
Rejecting the figure >£1,000 gives
a = £906.80 and b=£93.20 [1]
Checking the second derivative
d^2E[U]/da^2 = -0.39a^-1.5 – 0.25(1000-a)^-1.5 < 0 [½]
hence this is a maximum [½]
[Note to markers: rounding accepted]
iv) E[U] = 49.801 [1]

[Note to markers: assign [1] mark to answers containing the expected wealth]
v) U(1000) = 61.2456 [1]

So the maximum expected utility of wealth is less than the current utility of
wealth. [½]
This is because the odds offered pay out less than would be required based on the
investor’s estimated probabilities of each horse winning. [½]
Based on expected utility, the investor would be better off not betting at all. [½]
There may be other horses in the race where this position is reversed. [½]
[Note to markers: assign [1] mark to any valid comment]
[Max 2]
Page 3
Part i) and ii) were very well answered with most students scoring full
marks.
Many students found iii) challenging with only a few able to formulate
the expected utility correctly allowing for the possibility of neither horse
winning. Some students understood the method required of setting the
first derivative to zero and so were able to score method marks even if
they were unable to obtain the correct formula for the expected utility.
Very few candidates checked for conditions on the second order
derivative.
Quite a few candidates were able to derive the expected wealth correctly
for iv) based on their answer in iii) and so were able to score this mark.
Some candidates were able to score a mark for calculating the utility of
the initial wealth although few were able to make sensible comments to
score additional marks.
Q2
The estimation of parameters is one of the most time-consuming aspects of [½]
stochastic asset modelling.
The simplest case is the purely statistical model, where parameters are [½]
calibrated entirely to past time series. Provided the data is available, and
reasonably accurate, the calibration [½]
can be a straightforward and mechanical process.
Of course, there may not always be as much data as we would like, and [½]
the statistical error in estimating parameters may be substantial.
Furthermore, there is a difficulty in interpreting data which appears to [½]

invalidate the model being fitted.
For example, what should be done when fitting a Gaussian model in the [½]
presence of large outliers in the data?
Perhaps the obvious course of action is to reject the hypothesis of [½]

normality, and to continue building the model under some alternative
hypothesis. After all, in many applications, the major financial risks lie in
the outliers, so it seems foolish to ignore them.
In practice, a more common approach to outliers is to exclude them from [½]

the statistical analysis, and focus attention instead on the remaining residuals
which appear more normal.
The model standard deviation may be subjectively nudged upwards after [½]
the fitting process, in order to give some recognition to the outliers which
have been excluded.
Page 4
It has often been the practice in actuarial modelling to use the same data [½]
set to specify the model structure, to fit the parameters, and to validate the
model choice.
A large number of possible model structures are tested, and testing stops [½]
when a model which is found which passes a suitable array of tests.
Unfortunately, in this framework, we may not be justified in accepting a [½]

model simply because it passes the tests.
Many of these tests (for example, tests of stationarity) have notoriously [½]
low power, and therefore may not reject incorrect models.
Indeed, even if the “true” model was not in the class of models being [½]
fitted, we would still end up with an apparently acceptable fit, because
the rules say we keep generalising until we find one.
This process of generalisation tends to lead to models which wrap [½]

themselves around the data, resulting in an understatement of future risk,
and optimism regarding the accuracy of out-of-sample forecasts.
In the context of economic models, the calibration becomes more [½]
complex. The objective of such models is to simplify reality by imposing

certain stylised facts about how markets would behave in an ideal world.
This theory may impose constraints, for example on the relative [½]
volatilities of bonds and currencies. Observed data may not fit these
constraints perfectly.
In these cases, it is important to prioritise the features of the economy [½]

that are most important to calibrate accurately for a particular application.
[Note to markers: please award ½ mark for any valid idea presented by the
candidates. We only need the concept to earn a half mark, not all the detail
above.]
[Max 5]
This was generally poorly answered with most candidates picking up one
or two marks at most. Many candidates focused on why the CAPM and
APT models were unrealistic rather than focusing specifically on the
challenges of estimating parameters and hence were unable to score
well.
Page 5
Q3
i) EITHER Bt has independent increments, OR Bt - Bs is independent of {Br , r <=
s} whenever s < t, OR BOTH
EITHER Bt has stationary increments, OR the distribution of Bt - Bs depends
only on t – s, OR BOTH
EITHER Bt has Gaussian increments, OR the distribution of Bt - Bs is N(0, t - s),
OR BOTH
Bt has continuous sample paths t -> Bt.
B0 = 0.
[1 Mark each]
[Total of 5]
ii) A(0) = exp(0) = 1, so the students buys 1,000 units of the asset. [½]
E[A(5)] = exp(0.5*12*5) = 12.182 [1]
So the expected value of the investment is $12,182. [½]
iii) P(Investment<$10,000) = P(A(5)<10) [½]

= P(Z<(ln(10)/sqrt(5)) [½]
= P(Z<1.03) [½]
= 0.8484 (0.85) [½]
Part i) was bookwork and was well answered with the majority of
candidates scoring full or close to full marks.
Parts ii) and iii) proved challenging for many candidates although the
techniques required were quite standard. Students in general struggled
with identifying the correct parameters and which expectation and
probability to calculate.
Q4
i) F(4) = 10exp(4*0.05) = £12.21 [1]
ii) We can obtain d1 and d2 from the Black Scholes formula as:
d1 = 0.1214 [1]
d2 = -0.1186 [1]
Call price = £0.96 [1]
iii) Paid £12.21 for a stock worth £12 = loss of £0.21 [1]
iv) Paid £0.96 for call option = 0.96*exp(4*0.05) = £1.17 at time 4 [1]
Call expires worthless, hence loss of £1.17 [1]
Page 6
[Note to marker: please award ½ mark for using £0.96 as the loss on the position]
v) Mr Jones must pay for the optionality. [1]

He makes a loss if the option expires worthless, but that loss is never [1]
larger than £1.17.
This capped loss has a cost. [1]
[Note to Markers: please accept valid answers carrying forward an erroneous
£0.96 as loss]
[Max 2]
vi) If the stock is worth more than £12.21 at time 4 then Mrs Jones will make [½]
a profit
and this profit will always be larger than Mr Jones’ profit because he had [1]
to buy the option.
If the stock is worth less than £12.21 at time 4 then Mrs Jones will make [½]
a loss
of £12.21 minus stock price [1]
Mr Jones will always make a loss of £1.17 [½]

So the crossover is at a stock price of 12.21 – 1.17 = £11.04 [1]
[Max 3]
ALTERNATIVE ANSWER: Profit from Call > Profit from forward [1]
Max{S(T) – 12.21, 0} – 1.17 > S(T) – 12.21 [½]
=> Max{– 12.21, – ST} > – 11.04 [½]
=> S(T) < 11.04 [1]
[Note to Markers: please accept valid answers carrying forward an erroneuos £0.96
as loss]
In general, parts i), ii) and iii) were well answered although there were
some calculation errors. Some candidates though did not reflect on the
magnitude of their answer to determine if it was realistic.
For iv) many candidates did not accumulate the premium paid to expiry
as required by the question and so were not able to score full marks.
In part v) most candidates did not explain properly why a premium is
required to enter the option and hence lost a mark.
Part vi) was quite poorly answered and quite a few candidates struggled
as they were unable to consider all the different possible outcomes.
Page 7
Q5
i) We need E[e-rtB t | F s ] = e-rsB s [½]
If the bond has already defaulted by time s then e-rsB s = e-rs30e-r(3-s) = e-rtB t [1]
Otherwise e-rtB t = 30e-3r if the bond defaults before time t, or [½]
e-rtB t = e-3r(30(1 - e-λ(3-t)) + 100e-λ(3-t)) otherwise [½]
Then E[e-rtB t | F s ] = 30e-3r(1 - e-λ(t-s)) + e-3r(30(1 - e-λ(3-t)) + 100e-λ(3-t)) e-λ(t-s) [½]

= e-3r(30(1 - e-λ(3-s)) + 100e-λ(3-s)) [½]
= e-rsB s hence this is a martingale [½]
ii) (a) Let the portfolio contain x cash and y bonds.

We need the value at time 3 to be $35 if the bond has defaulted, [1]
so xe3r +30y = 35
We also need the value at time 3 to be zero if the bond has not defaulted, [1]
so xe3r +100y = 0
Hence x = 50 e-3r and y = -0.5. [1]
(b) The price at time zero must be the cost of buying this portfolio [½]
= x+ yB 0 = 50e-3r – 0.5e-3r(30(1 - e-3λ)) + 100e-3λ) [1]
= 35e-3r(1-e-3λ) [½]
iii) Let the portfolio value be V. [1]

We need the value of the portfolio at time 0 to be E[e-3rV 3 ] under
probability measure P.
E[e-3rV 3 ] = 35e-3r * [Probability that bond defaults] [1]
= 35e-3r(1-e-3λ) hence the requirement holds. [1]
This fits with the fact that being able to hedge a derivative price without [1]
arbitrage means we can price it under the Equivalent Martingale Measure.
[Max 3]
Page 8
This was a difficult question that was very poorly answered with
candidates either not attempting or only superficially attempting this
question. Overall the lowest scoring question from the paper.
For i), the majority of candidates appeared unfamiliar with the term
“equivalent martingale measure”, and did not realise they needed to
focus on the discounted price process.
There were more attempts for ii) although the majority were unable to
formulate the required equations for the replicating portfolio.
Part iii) was very poorly attempted and any reasonable comment was
given credit.
Q6
i) EITHER: The market is arbitrage free if and only if there exists a probability
measure under which discounted asset prices are martingales. [1]
OR
The probability exists if 1 − 𝑑𝑑 < 𝑒𝑒 𝑟𝑟𝑟𝑟𝑟𝑟 < 1 + 𝑢𝑢. [1]
OR BOTH [max 1]
[Note to markers: please accept solutions with 𝑑𝑑 < 𝑒𝑒𝑒𝑒𝑒𝑒(𝑟𝑟𝑟𝑟𝑟𝑟) < 𝑢𝑢, and 𝛿𝛿𝛿𝛿 = 1.]
(Note to markers: both answers are acceptable)

ii)
Stock tree
time 0 1 2 3
95.00 114.00 136.80 164.16
76.00 91.20 109.44
60.80 72.96
48.64
The price P 0 of the option is computed via Risk Neutral Valuation; let 𝑝𝑝̂ denote the risk
neutral probability of an up movement, then
𝑒𝑒 0.05 − 0.80
𝑝𝑝̂ = = 0.6282
1.20 − 0.80
[1]
and
3
3
𝑃𝑃0 = 𝑒𝑒 −𝑟𝑟𝑟𝑟 � � � 𝑝𝑝̂ 𝑘𝑘 (1 − 𝑝𝑝̂ )3−𝑘𝑘 (𝐾𝐾 − 𝑆𝑆0 𝑢𝑢𝑘𝑘 𝑑𝑑3−𝑘𝑘 )+
𝑘𝑘
𝑘𝑘=0
[2]
Page 9
= 𝑒𝑒 −0.05∗3 �0.56 × 3𝑝𝑝̂ 2 (1 − 𝑝𝑝̂ ) + 37.04 × 3𝑝𝑝̂ (1 − 𝑝𝑝̂ )2 + 61.36 × (1 − 𝑝𝑝̂ )3 �

= 11.23
[1]
The detailed working are provided below – in case attempts to answer this
question go through the whole tree.
PUT
time 0 1 2 3
11.23 4.87 0.20 0.00
23.53 13.44 0.56
43.84 37.04
61.36
iii) The given market conditions imply that 1 − 𝑑𝑑 = 0.95, 1 + 𝑢𝑢 = 1.05; [1]
however the discount factor is 𝑒𝑒 0.05 = 1.0513
Hence the condition of no arbitrage is violated and no pricing is [1]

possible
[Alternatively, candidates can recalculate the risk neutral probability 𝑝𝑝̂ , which in this
case would give 1.0127, hence there is arbitrage in the market
Alternatively, candidates can obtain a negative option price ].
Parts i) and ii) were well answered although there were a few
calculation errors in ii). Lost marks were mostly for getting an incorrect
probability value, forgetting the combination factor in the final
calculation or slipping up with the numbers. A few students also got
confused and tried to price a call option.
There were quite a few good answers to iii) although about half did not
spot that the no arbitrage condition would not hold under the new
scenario.
Q7
i) The market is complete if for any contingent claim X there is a [1]
replicating strategy (Φ𝑡𝑡 , Ψ𝑡𝑡 )
i.e. is a self-financing strategy, defined for 0 ≤ t < U , capable of [1]

reproducing the derivative terminal payment at 𝑈𝑈 without risk, for an
initial investment of 𝑉𝑉(0) at time 0.
[Max 1]
ii) The SDE of 𝑆𝑆̃𝑡𝑡 = 𝑒𝑒 −𝑟𝑟𝑟𝑟 𝑆𝑆𝑡𝑡 :
𝑑𝑑𝑆𝑆̃𝑡𝑡 = (𝜇𝜇 − 𝑟𝑟)𝑆𝑆̃𝑡𝑡 𝑑𝑑𝑑𝑑 + 𝜎𝜎𝑆𝑆̃𝑡𝑡 𝑑𝑑𝑊𝑊𝑡𝑡 , [1]
Page 10
For this process to be a martingale, the drift should be zero [1]
but 𝑟𝑟 ≠ 𝜇𝜇 in general. Hence it is not a martingale. [1]
[Alternative solutions based on solving the SDE and checking the martingale condition
are equally acceptable. This is equivalent to check that the identity E[𝑆𝑆̃𝑇𝑇 |𝐹𝐹𝑡𝑡 ]= 𝑆𝑆̃𝑡𝑡 for
𝑆𝑆̃𝑡𝑡 = 𝑆𝑆0 exp((𝜇𝜇 − 𝑟𝑟 − 0.5𝜎𝜎 2 )𝑡𝑡 + 𝜎𝜎𝑊𝑊𝑡𝑡 ) does not hold.]
iii) We need to change the Brownian motion by means of the Girsanov [1]
�𝑡𝑡 = 𝑊𝑊𝑡𝑡 + 𝜆𝜆𝜆𝜆 be a Brownian motion under a new probability
Theorem. Let 𝑊𝑊
measure 𝑄𝑄
then the above SDE becomes: [1]
𝑑𝑑𝑆𝑆̃𝑡𝑡 = (𝜇𝜇 − 𝑟𝑟 − 𝜆𝜆𝜆𝜆)𝑆𝑆̃𝑡𝑡 𝑑𝑑𝑑𝑑 + 𝜎𝜎𝑆𝑆̃𝑡𝑡 𝑑𝑑𝑊𝑊

�𝑡𝑡 ,
For the martingale property to hold set the drift to zero, which implies
(𝜇𝜇 − 𝑟𝑟)�
𝜆𝜆 = 𝜎𝜎 . [1]
[Alternative solution carrying equal marks: change the Brownian motion as above
and take the conditional expectation under the new measure; this returns 𝐸𝐸 𝑄𝑄 �𝑆𝑆̃𝑇𝑇 �𝐹𝐹𝑡𝑡 � =
𝑆𝑆̃𝑡𝑡 𝑒𝑒𝑒𝑒𝑒𝑒(𝜇𝜇 − 𝑟𝑟 − 𝜆𝜆𝜆𝜆)(𝑇𝑇 − 𝑡𝑡). The martingale condition requires 𝜆𝜆 = 𝜎𝜎 .]
In general, candidates who scored marks seemed to have a good

understanding of the underlying theory.
For i) many did not know the required bookwork definition although this
did not affect the rest of the question.
Part ii) was reasonably answered although many calculated an
expression for S(t) and struggled to show that the expectation condition
was satisfied rather than directly use the SDE.
Part iii) was not well answered and generally only the stronger
candidates scored well on this part.
Q8
i) Data: 𝑆𝑆0 = 65, 𝐾𝐾 = 55, 𝜎𝜎 = 25% 𝑝𝑝. 𝑎𝑎. , 𝑇𝑇 = 0.5 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦, 𝑟𝑟 = 2%
Let Ct be the price of the European call.
The Black-Scholes formula returns [½ Mark each]
𝑑𝑑1 = 1.09
𝑑𝑑2 = 0.9132
𝑁𝑁(𝑑𝑑1 ) = 0.8621
𝑁𝑁(𝑑𝑑2 ) = 0.8194
Therefore 𝐶𝐶0 = 65 × 0.8621 − 55𝑒𝑒 −0.02×0.5 × 0.8194 [1]
= 11.42 [1]
Page 11
𝜕𝜕𝜕𝜕
ii) 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = 𝜕𝜕𝜕𝜕
[1]
[Note to markers: please award ½ mark for stating 𝑁𝑁(𝑑𝑑1 ) ]
iii) In the Black-Scholes model 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = 𝑁𝑁(𝑑𝑑1 ) [1]
Using the results from above 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = 0.8621 [1]
iv) 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑎𝑎𝑝𝑝𝑝𝑝𝑝𝑝 = 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑎𝑎𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 − 1 [1]

Therefore, 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑎𝑎𝑝𝑝𝑝𝑝𝑝𝑝 = −0.1379 [1]
[Note to markers: if signs are incorrect in the formula and the actual value, award
½ mark for 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑎𝑎𝑝𝑝𝑝𝑝𝑝𝑝 = 0.1379 only]
This question was generally well answered with quite a few scoring full
marks or close to full marks. Many got (iv) correct, but a significant
number of candidates got the sign wrong.
Q9
i) Using continuous compounding.
[Note to markers: please accept any correct attempt using different compounding
convention]
1 97.5
a. − 0.25 𝑙𝑙𝑙𝑙 100 = 10.13% [1]
1 94.9
b. − 0.5 𝑙𝑙𝑙𝑙 100 = 10.47% [1]
c. 0.04 × (94.9 + 90) + 104 × 𝑒𝑒 −0.1068×1.5 = 96 [1]
(0.1054×1−0.1047×0.5)
d. 1−0.5
= 10.60% [1]
(0.1068×1.5−0.1054×1)
e. 1.5−1
= 10.97% [1]
ii) Two standard models for the short rate of interest are the Vasicek model
and the CIR model.
The corresponding SDEs are respectively
𝑑𝑑𝑟𝑟𝑡𝑡 = 𝑘𝑘(𝜃𝜃 − 𝑟𝑟𝑡𝑡 )𝑑𝑑𝑑𝑑 + 𝜎𝜎𝜎𝜎𝑊𝑊𝑡𝑡
𝑑𝑑𝑟𝑟𝑡𝑡 = 𝑘𝑘(𝜃𝜃 − 𝑟𝑟𝑡𝑡 )𝑑𝑑𝑑𝑑 + 𝜎𝜎�𝑟𝑟𝑡𝑡 𝑑𝑑𝑊𝑊𝑡𝑡
[1 mark each]
Alternatively: another standard model is the Hull and White model which extends
the Vasicek model to allow for time-inhomogeneity, therefore the parameters in
the SDE are time dependent.
Parts i) a) and b) were very well answered and d) was also well answered
by many. Parts c) and e) proved difficult with only a few getting the
marks here.
Page 12
Part ii) was very well answered.
Q10
i) Data:
𝑆𝑆0 = 15, 𝐾𝐾 = 12, 𝜎𝜎 = 20% 𝑝𝑝. 𝑎𝑎. , 𝑇𝑇 = 0.25 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦, 𝑟𝑟 = 2%.
Let Ct be the price of the European call.
The Black-Scholes formula returns [½ mark each]
𝑑𝑑1 = 2.3314
𝑑𝑑2 = 2.2314
𝑁𝑁(𝑑𝑑1 ) = 0.9901
𝑁𝑁(𝑑𝑑2 ) = 0.9872
Therefore 𝐶𝐶0 = 15 × 0.9901 − 12𝑒𝑒 −0.02×0.25 × 0.9872 [1]

= 3.0650 [1]
ii) Probability of expiring in the money: 𝑃𝑃(𝑆𝑆𝑇𝑇 > 𝐾𝐾) = 𝑁𝑁(𝑑𝑑2 ) [½]
hence from above 𝑃𝑃(𝑆𝑆𝑇𝑇 > 𝐾𝐾) = 0.9872. [½]
iii) Risk neutral valuation applied to the given digital option returns
𝐸𝐸�𝑒𝑒 −𝑟𝑟𝑟𝑟 1𝑆𝑆𝑇𝑇>𝐾𝐾 � = 𝑒𝑒 −𝑟𝑟𝑟𝑟 𝑃𝑃(𝑆𝑆𝑇𝑇 > 𝐾𝐾) [1]
From above it follows that the price is 𝑒𝑒 −0.02×0.25 × 0.9872 = 0.98225 [1]
iv) Limitations [½ mark each]
a. Share prices can jump. This invalidates assumption that the stock price evolves
as geometric Brownian motion, as this process has continuous sample paths.
However, hedging strategies can still be constructed which substantially reduce
the level of risk.
b. The risk-free rate of interest does vary and in an unpredictable way. However,
over the short term of a typical derivative, the assumption of a constant risk-free
rate of interest is not far from reality. (More specifically the model can be
adapted in a simple way to allow for a stochastic risk-free rate, provided this is a
predictable process.)
c. Unlimited short selling may not be allowed, except perhaps at penal rates of
interest. These problems can be mitigated by holding mixtures of derivatives
which reduce the need for short selling. This is part of a suitable risk
management strategy.
d. Shares can normally only be dealt in integer multiples of one unit, not
continuously, and dealings attract transaction costs. Again we are still able to
construct suitable hedging strategies which substantially reduce risk.
Page 13
e. Distributions of share returns tend to have fatter tails than suggested by the
lognormal model.
[Note to Markers: please award ½ mark for any valid idea/comment. Just the concept
is enough for a half mark, no need for all the detail above.]
[Max 2]
Part i) was well answered although there were some calculation errors.
Parts ii) and iii) proved challenging for some although quite a few
candidates were able to score full marks.
Part iv) was generally well answered although quite a number of
candidates struggled to generate enough points to score full marks for a
standard bookwork question.
Q11
i) CAPM assumptions [½ mark each]
a. All investors have the same one-period horizon.

b. All investors can borrow or lend unlimited amounts at the same risk-free rate.
c. The markets for risky assets are perfect. Information is freely and
instantly available to all investors and no investor believes that they can affect
the price of a security by their own actions.
d. Investors have the same estimates of the expected returns, standard deviations
and covariances of securities over the one-period horizon.
e. All investors measure in the same “currency” e.g. pounds or dollars or in
“real” or “money” terms.
[Max 2]
ii) By definition the beta of each security is 𝛽𝛽𝑖𝑖 = 𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅𝑖𝑖 , 𝑅𝑅𝑀𝑀 )/𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅𝑀𝑀 ) [½]
where 𝑅𝑅𝑖𝑖 is the rate of return on security 𝑖𝑖, 𝑅𝑅𝑀𝑀 , 𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅𝑀𝑀 ) are respectively
the rate of return on the market portfolio and its variance [½]
Hence
𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅𝑀𝑀 , 𝑅𝑅𝑀𝑀 )
𝛽𝛽𝑀𝑀 = =1
𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅𝑀𝑀 )
as required [1]
(Note to markers: the same conclusion can be reached from the Security Market
Line, and is equally acceptable)
iii) As the market portfolio is the weighted portfolio of the risky securities
in the market, and the given weights are 0.4 and 0.6, then
𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅𝑖𝑖 , 𝑅𝑅𝑀𝑀 ) = 𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅𝑖𝑖 , 0.4𝑅𝑅1 + 0.6𝑅𝑅2 )
[1]
from which it follows that [½ mark each]
Page 14
𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅1 , 𝑅𝑅𝑀𝑀 ) = 0.4𝑉𝑉𝑎𝑎𝑟𝑟(𝑅𝑅1 ) + 0.6𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅1 , 𝑅𝑅2 ) = 0.00089

𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅2 , 𝑅𝑅𝑀𝑀 ) = 0.4𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅1 , 𝑅𝑅2 ) + 0.6𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅2 ) = 0.00150
Also:
𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅𝑀𝑀 ) = 0.42 ∗ 𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅1 ) + 0.62 ∗ 𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅2 ) + 2 ∗ 0.4 ∗ 0.6 ∗ 𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅1 , 𝑅𝑅2 ) =
0.00125
Consequently 𝛽𝛽1 = 0.70915, 𝛽𝛽2 = 1.1939 [½ each]
[Note to Markers: please accept any correct attempt with rounded figures. For
𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅1 , 𝑅𝑅𝑀𝑀 ) = 0.4𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅1 ) + 0.6𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅1 , 𝑅𝑅2 ) = 0.0009
we obtain 𝛽𝛽1 = 0.72, 𝛽𝛽2 = 1.2]
iv) From the Security Market Line it follows that 𝑅𝑅𝑓𝑓 = (𝐸𝐸𝑅𝑅𝑖𝑖 − 𝛽𝛽𝑖𝑖 𝐸𝐸𝑅𝑅𝑀𝑀 )⁄(1 − 𝛽𝛽𝑖𝑖 )
From the data 𝐸𝐸𝑅𝑅1 = 6.40%, 𝐸𝐸𝑅𝑅2 = 9.90% . [½ marks each]
Consequently
𝐸𝐸𝑅𝑅𝑀𝑀 = ∑2𝑖𝑖=1 𝑤𝑤𝑖𝑖 𝐸𝐸𝑅𝑅𝑖𝑖 = 8.5% [½]
and 𝑅𝑅𝑓𝑓 = 0.012797 [½]

[Note to Markers: if the rounding above and the corresponding betas are used, then
from the equation for asset 1 we obtain 𝑅𝑅𝑓𝑓 = 0.01, whilst from the equation for asset
2 we obtain 𝑅𝑅𝑓𝑓 = 0.015. Please accept any valid attempt.]
Part i) was answered reasonably although many candidates were unable

to identify the additional assumption of CAPM compared to modern
portfolio theory.
Part ii) was very well answered.
Part iii) was poorly answered with most candidates unable to calculate
the required covariances correctly. Many candidates indeed attempted
this task by only looking at the information regarding the rate of return
in each state (and corresponding probability), rather than actually using
the provided variance/covariance matrix
Part iv) was answered well and credit was given for calculating the
expected returns even if this was done in iii).
Page 15
EXAMINATION
26 September 2018 (pm)

Core Technical
1. Enter all the candidate and examination details as requested on the front of your
answer booklet.
supervisor.
question paper.
CT8 S2018 © Institute and Faculty of Actuaries

1 Describe the key findings in behavioural finance. [10]
2 An investor has taken out a $100,000 loan at a 10% per annum rate of interest,
annually compounded. The investor uses the loan to buy a portfolio of stocks whose
value follows a lognormal distribution, with parameters μ = 12% per annum and
σ2 = 25% per annum.
The investor plans to sell the stocks and repay the loan after five years.
(i) Calculate the mean and the variance of the lognormal distribution. [1]

(ii) Calculate the probability that the investor will have enough money to repay the
loan plus interest. [2]
After five years the stocks are only worth $120,000 so the investor cannot afford to
repay the loan plus interest.
(iii) Calculate the one-year 95% value at risk of the investor’s stock portfolio from
time t = 5 to time t = 6. [2]
The annual standard deviation of the investor’s stock portfolio at time t = 5 is $81,708.
The bank agrees to continue the loan for another five years, as long as the investor can
prove that the annual standard deviation of his portfolio is no higher than $40,000 at
time t = 5.
(iv) Calculate the proportion of the investor’s stock portfolio he would have to sell
in order to bring the value at risk down to a level acceptable to the bank. [1]
The bank also offers a cash deposit account returning a 6% per annum rate of interest,
annually compounded.
The investor sells the proportion of the stock portfolio in part (iv) and invests the
funds in the cash deposit account. The bank therefore continues the loan for another
five years.
(v) Calculate the probability that the investor’s stocks and cash deposit combined
are sufficient to repay the loan plus interest at time 10. [3]
(vi) (a) Comment on this result.

(b) Propose an alternative course of action for the investor. [2]
[Total 11]
CT8 S2018–2
3 In a market in which the Arbitrage Pricing Theory (APT) model holds, the expected
return is given by:
E[Ri] = l0 + l1bi,1 + l2bi, 2 + …
(i) Define all the terms in the equation. [2]
Assume the risk-free rate rf = 0.04. Consider two well diversified portfolios Pi with
the following features in a two factor model:
P1 P2
E Ri 15.50% 11.95%
bi,1 (a) (b)
bi,2 1.5 0.7
(ii) Determine the values (a) and (b) for l1 = 0.05 and l2 = 0.06. [3]
[Total 5]
4 An investor has £100 and is considering investing in two different stocks. The prices
of both stocks are assumed to follow the lognormal model with the parameters below.
Stock Current price Drift µ Volatility σ

A £5 5% 20%
B £5 8% 30%
(i) Calculate the expected value at time 3 of £100 invested in:

(a) stock A
(b) stock B. [2]
(ii) Calculate the standard deviation at time 3 of £100 invested in:

(a) stock A
(b) stock B. [4]
The investor decides to invest £50 in each stock.
(iii) Calculate the expected value of the investor’s portfolio at time 3. [1]
The correlation of the two stocks is 0.3.
(iv) Calculate the standard deviation of the value of the investor’s portfolio at
time 3. [3]
(v) Comment on the expected return and standard deviation of the portfolio
compared to investing the whole £100 in one stock. [4]
[Total 14]

5 The Ornstein-Uhlenbeck process is the solution to the equation dXt = – γXt dt + σdBt
where γ and σ are positive parameters.
Derive the solution for Xt.[8]
6 Consider a call option ct and a put option pt written on a non-dividend paying stock St.
(i) Prove the put-call parity relationship by constructing two portfolios that
produce the same value at maturity. [4]
A stock market includes four options set out below. All the options are for a term of
10 years and relate to a single non-dividend paying stock, currently priced at $5. The
continuously compounded risk-free rate is 3% per annum.
Type Strike price Option price

Option A European Call $8 $0.32
Option B European Put $8 ?
Option C European Put $10 ?
Option D American Put $10 ?
(ii) Calculate the price of Option B. [2]
(iii) Determine lower and upper bounds for the price of option C. [2]
(iv) Determine lower and upper bounds for the price of option D. [2]
[Total 10]
CT8 S2018–4
7 A company is currently financed entirely by equity with 100,000 shares in issue and
no debt. The current share price is $1. The company has total assets of $100,000 with
volatility of 15% per annum.
The company is considering raising $250,000 by issuing zero-coupon debt with a five-
year maturity date. The continuously compounded risk-free rate of interest is 3% per
annum.
The company intends to set the redemption value of the debt such that the share price
will remain unchanged under the Merton model.
(i) Give the value of the company’s assets immediately after issuing the debt. [1]
(ii) Calculate the redemption value of the debt using the Merton model. [5]
(iii) Calculate the credit spread on the debt. [2]
One year later, the company is struggling. The share price has fallen to $0.50 and the
current value of the debt has fallen to $50 per $100 of redemption value.
(iv) Calculate the proportionate fall in the value of:

(a) the equity.
(b) the debt.[2]
(v) Suggest why the value of the equity has fallen by proportionately more than
the fall in the value of the debt. [3]
[Total 13]
8 Consider a binomial tree model for the non-dividend paying stock with price St .
Assume this price either rises by 30% or falls by 20% each quarter (3 months) for the
next three quarters. Assume also that the risk-free rate is 2% per annum continuously
compounded. Let S0 = £60.
(i) Calculate the price of a vanilla European call option with maturity in nine
months’ time and a strike price of £55. [3]
(ii) Calculate the price of a vanilla European put option with the same maturity
and strike price as the contract in part (i). [1]
Assume the investor has a portfolio formed by a short position in the call option given
in part (i) and a long position in the put option given in part (ii).
(iii) Determine how the value of the portfolio would differ if the possible change in
the stock price was a fall of 30% instead of 20%. [3]
[Total 7]

9 The price process of a non-dividend paying stock St satisfies the following stochastic
differential equation
dSt = μ St dt + σSt dWt ,
where Wt is a Brownian motion under the real-world probability measure P. Let V(t)
be the value at t of a self-financing portfolio, consisting of Ft stocks and Yt cash
bond.
(i) Show that d(e–rtV (t)) = Ft d(e–rtSt).[3]
(ii) Determine the conditions under which the discounted value e–rtV (t) is a
martingale.[3]
[Total 6]
10 (i) State the main assumptions underpinning the Black-Scholes model. [3]
Consider a put option on a non-dividend paying stock when the stock price is £8, the
exercise price is £9, the continuously compounded risk-free rate of interest is 2% per
annum, the volatility is 20% per annum. and the time to maturity is three months.
(ii) Calculate the price of the option using the Black-Scholes model. [4]
(iii) Discuss how the price of the contract in part (ii) would change if the rate of
interest increases. (There is no need to carry out further calculations.) [2]
[Total 9]
CT8 S2018–6
11 Consider a market in which the Capital Asset Pricing Model (CAPM) holds.
(i) Write down the equation of the Security Market Line, defining all the notation
you use. [2]
In this market, the risk-free rate of interest is 9.44% per annum. There are only two
risky assets in the market with the following attributes.
Rate of return (per annum) Variance/Covariance Matrix

State Probability Asset 1 Asset 2 Asset 1 Asset 2
1 0.2 10.00 % 11.00 % Asset 1 0.00142 0.00379
2 0.3 15.00 % 30.00 % Asset 2 0.00379 0.01146
3 0.1 18.00 % 25.00 %
4 0.4 20.00 % 40.00 %
(ii) Determine the weight of each asset in the market portfolio to be consistent
with β1 = 0.46, β 2 = 1.36. [3]
(iii) Calculate the Market Price of Risk. [2]

[Total 7]
END OF PAPER

EXAMINERS’ REPORT
September 2018

Core Technical
Introduction
Mike Hammer
December 2018
 Institute and Faculty of Actuaries

1. In general, the real differentiators in those who scored well were attention to detail
in their algebraic steps, and the breadth of knowledge in being able to score the
knowledge marks and even attempt most questions. A number of candidates did not
gather relevant information from the text of the question, and translate it in the
appropriate equivalent statistical concepts. For example, candidates struggled with
formulating the probability that an event occurs in appropriate mathematical terms,
and determining from the information in the question the direct way to recover
required variances and covariances.
2. Students performed relatively well on knowledge based questions, although many

missed the opportunity to be awarded full marks.
C. Pass Mark
Page 2
Solutions
Q1
Anchoring and adjustment Anchoring is a term used to explain how people will produce estimates.
They then adjust away from this initial anchor to arrive at their final judgement. [1]
Prospect theory A theory of how people make decisions when faced with risk and uncertainty. It
replaces the conventional risk averse / risk seeking decreasing marginal utility theory. [1]
Framing (and question wording) The way a choice is presented (“framed”) and, particularly, the
wording of a question in terms of gains and losses, can have an enormous impact on the answer given
or the decision made. [1]
Myopic loss aversion This is similar to prospect theory, but considers repeated choices rather than a
single “gamble”. [1]
Estimating probabilities Issues (other than anchoring) which might affect probability estimates
include: [0.5]
• Dislike of “negative” events – the “valence” of an outcome (the degree to which it is considered as
negative or positive) has an enormous influence on the probability estimates of its likely occurrence.
[0.5]
• Representative Heuristics – people find more probable that which they find easier to imagine. As the
amount of detail increases, its apparent likelihood may increase (although the true probability can
only decrease steadily). [0.5]
• Availability – people are influenced by the ease with which something can be brought to mind. This
can lead to biased judgements when examples of one event are inherently more difficult to imagine
than examples of another. [0.5]
Overconfidence People tend to overestimate their own abilities, knowledge and skills. This may be a
result of: [0.5]
• Hindsight bias – events that happen will be thought of as having been predictable prior to the event,
events that do not happen will be thought of as having been unlikely prior to the event. [0.5]
• Confirmation bias – people will tend to look for evidence that confirms their point of view (and will
tend to dismiss evidence that does not justify it). [0.5]
Mental accounting People show a tendency to separate related events and decisions and find it
difficult to aggregate events. [1]
Effect of options Other issues include:
• Primary effect – people are more likely to choose the first option presented, but [0.5]
• Recency effect – in some instances, the final option that is discussed may be preferred! (The gap in
time between the presentation of the options and the decision may influence this dichotomy.) [0.5]
• Other research suggests that people are more likely to choose an intermediate option than one at
either end! [0.5]
• A greater range of options tends to discourage decision-making. On the other hand, a higher
probability is attributed to options explicitly stated than when included in a broader category. [0.5]
• Status Quo bias – people have a marked preference for keeping things as they are. [0.5]
• Regret aversion – by retaining the existing arrangements, people minimise the possibility of regret
(the pain associated with feeling responsible for a loss). [0.5]
• Ambiguity aversion – people are prepared to pay a premium for rules. [0.5]
[Max 10]
Page 3
The majority of students scored either full marks or nearly full

marks on this knowledge based question.
Some students confused Behavioural Finance with Expected Utility

Theory or the Efficient Market Hypothesis, for which there were no
marks on offer.
Q2
i) Mean = exp(µ+0.5*σ2) = exp(0.12+0.5*0.25) = 1.2776 [0.5]
Variance = exp(2µ+σ )*(exp(σ )-1) = exp(2*0.12+0.25)*(exp(0.25)-1) = 0.4636
2 2
[0.5]
ii) L(5) = 100,000 * 1.1^5 = 161,051 [0.5]

P(S(5) > 161,051) = P(S(5)/S(0) > 1.61051) = P(ln(S(5)/S(0)) > ln(1.61051))
= P(Z > ((ln(1.61051) – 5*0.12) / (0.25*5)^0.5) [0.5]
= P(Z > -0.110416) [0.5]
= 54.4% [0.5]
0.1*5
[If students have used continuous compounding (e ) deduct one mark]
iii) P(S(6)/S(5) < t) = 0.05 => P(Z < (ln(t)-0.12) / 0.25^0.5) = 0.05 [0.5]
=> (ln(t)-0.12) / 0.25^0.5 = -1.645 [0.5]
=> t = exp( -1.645 * 0.25^0.5 + 0.12) = 0.4954 [0.5]
=> VaR = £120,000 * 0.4954 = $59,446 [0.5]
[Or for £120,000 * (1 – 0.4954) = £60,552 lose one mark]
iv) The investor can retain 40,000 / 81,708 = 48.95% of his stocks, so he would need
to sell $61,254 of stocks. [1]
v) The loan at time 10 will be 161,051 * 1.1^5 = $259,374. [0.5]

The cash deposit account holds $61,254 at time 5, hence 61254 * 1.06^5 =
$81,972 at time 10. [0.5]
So we need the stocks to be worth at least 259,374-81,972 = $177,402 at time 10.
[0.5]
This needs a return of 177,402 / 58,746 = 3.0198 [0.5]
P(S(10)/S(5) > 3.0198) = P(Z > (ln(3.0198) – 5*0.12) / (0.25*5)^0.5) = P(Z >
0.4519) [0.5]
= 32.6% [0.5]
vi) There is a slightly less than 50:50 chance that the investor would be able to repay
the loan at time t=5. [1]
The cash account pays a lower rate of interest than the loan charges, so the
investor would be better off repaying $61,254 of the loan at time 5 if this is
possible. [1]
The investor could also seek other assets that deliver a higher potential return. [1]
Page 4
Or the investor could try to find other funds to cut his losses and repay the loan at
time 5. [1]
[Max 2]
Overall, students did not score well on the application of the

lognormal model.
A proportion of students confused the lognormal model and the
solution to Geometric Brownian Motion so had the incorrect drift
term. Another common mistake was not including the time factor
in the drift and volatility components.
Most students made suggestions in (vi) for how the investor could
reduce the risk in their portfolio.
Q3
i. 𝐸𝐸[𝑅𝑅𝑖𝑖 ] is the expected return of security i; [0.5]
𝑏𝑏𝑖𝑖,𝑘𝑘 is the responses of the rates of return on security i to factor k (alternatively the
sensitivity of security i to index k). [1]
𝜆𝜆𝑘𝑘 is the risk premium corresponding to factors k. [0.5]
ii. The risk free portfolio has zero exposure to any risk factor, i.e. 𝑏𝑏𝑖𝑖,𝑘𝑘 = 0 for all 𝑘𝑘, which
implies 𝜆𝜆0 = 𝑟𝑟𝑓𝑓 . [1]
Then, we look for the solution to

0.155 = 0.04 + 0.05𝑏𝑏1,1 + 0.06 × 1.5
[0.5 each]
0.1195 = 0.04 + 0.05𝑏𝑏2,1 + 0.06 × 0.7
which returns 𝑏𝑏1,1 = 0.5, 𝑏𝑏2,1 = 0.75 [0.5 each]
The majority of students scored either full marks or nearly full marks on
this knowledge question.
For part (i), common mistakes were confusing the different parameters
such as describing lambas as sensitivities and vice versa.
For part (ii), common mistakes were not including the risk-free rate in the
equations or minor calculation errors.
Page 5
Q4
i)
a. E[A 3 ] = A 0 exp(µt+0.5σ2t) = 100exp(0.05*3+0.5*0.22*3) = £123.37 [1]
b. E[B 3 ] = B 0 exp(µt+0.5σ2t) = 100exp(0.08*3+0.5*0.32*3) = £145.50 [1]
ii)
a. SD[A 3 ] = √(A 0 2exp(2µt+σ2t)(exp(σ2t)-1)) [0.5]
= √(1002exp(2*0.05*3+0.22*3)(exp(0.22*3)-1)) [1]
= £44.05 [0.5]
b. SD[B 3 ] = √(B 0 2exp(2µt+σ2t)(exp(σ2t)-1)) [0.5]
= √(1002exp(2*0.08*3+0.32*3)(exp(0.32*3)-1)) [1]
= £81.01 [0.5]
iii) E[P 3 ] = 0.5E[A 3 ] + 0.5E[B 3 ] = £134.44 [1]
iv) V[P 3 ] = 0.52V[A 3 ] + 0.52V[B 3 ] + 2*Correlation*0.5*0.5*SD[A 3 ]*SD[B 3 ] [1]

= 0.25*44.052 + 0.25*81.012 + 2*0.3*0.5*0.5*44.05*81.01
= 2,661.03 [1]
=> SD[P 3 ] = £51.59 [1]
v) The expected return of the portfolio falls halfway between the expected return on
each of the one-stock investment strategies. [1]
But the standard deviation is well below halfway between the two one-stock
strategies.
[1]
The price of risk for stock A is 23.37/44.05 = 0.53 [1]
The price of risk for stock B is 45.5/81.01 = 0.56 [1]
But the price of risk for the portfolio is 34.44/51.59 = 0.67 [1]
So the portfolio delivers a better expected return per unit of risk [1]
This is because the assets are not fully correlated… [1]
Which shows the benefit of diversification. [1]
[Max 4]
In general, students struggled with this question. The most common

difficulty was making the link between the price and the number of
shares held.
For part (iv), some students did not calculate a standard deviation
for the portfolio that was consistent with their answers in part (ii).
Their portfolio standard deviation was either much higher or much
lower.
Page 6
Some students calculated the proportion invested in the minimum

variance portfolio under Mean Variance Portfolio Theory despite this
not being asked for in the question.
Q5
[1]
[1]
[1]
[1]
[1]
[1]
[1]
[1]
[Or using a Taylor expansion:
𝑑𝑑𝑑𝑑(𝑋𝑋𝑡𝑡 ,𝑡𝑡) 1 𝑑𝑑 2 𝑓𝑓(𝑋𝑋𝑡𝑡 ,𝑡𝑡) 𝑑𝑑𝑑𝑑(𝑋𝑋𝑡𝑡 ,𝑡𝑡)

𝑑𝑑𝑑𝑑(𝑋𝑋𝑡𝑡 , 𝑡𝑡) = 𝑑𝑑𝑋𝑋𝑡𝑡 + 𝑑𝑑𝑋𝑋𝑡𝑡 2 + 𝑑𝑑𝑑𝑑 [1]
𝑑𝑑𝑋𝑋𝑡𝑡 2 𝑑𝑑𝑋𝑋𝑡𝑡 2 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑(𝑋𝑋𝑡𝑡 ,𝑡𝑡) 𝑑𝑑 2 𝑓𝑓(𝑋𝑋𝑡𝑡 ,𝑡𝑡) 𝑑𝑑𝑑𝑑(𝑋𝑋𝑡𝑡 ,𝑡𝑡)

Where 𝑑𝑑𝑋𝑋𝑡𝑡
= 𝑒𝑒 𝛾𝛾𝛾𝛾 , 𝑑𝑑𝑋𝑋𝑡𝑡 2
= 0 and 𝑑𝑑𝑑𝑑
= 𝛾𝛾𝑋𝑋𝑡𝑡 𝑒𝑒 𝛾𝛾𝛾𝛾 [1]
So 𝑑𝑑𝑑𝑑(𝑋𝑋𝑡𝑡 , 𝑡𝑡) = 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝑋𝑋𝑡𝑡 + 𝛾𝛾𝑋𝑋𝑡𝑡 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝑑𝑑 [1]
But 𝑑𝑑𝑋𝑋𝑡𝑡 = −𝛾𝛾𝑋𝑋𝑡𝑡 𝑑𝑑𝑑𝑑 + 𝜎𝜎𝜎𝜎𝐵𝐵𝑡𝑡 so 𝑑𝑑𝑑𝑑(𝑋𝑋𝑡𝑡 , 𝑡𝑡) = 𝑒𝑒 𝛾𝛾𝛾𝛾 𝜎𝜎𝜎𝜎𝐵𝐵𝑡𝑡 [1]
Page 7
𝑡𝑡 𝑡𝑡
Integrating between 0 and t gives ∫0 𝑑𝑑𝑑𝑑(𝑋𝑋𝑠𝑠 , 𝑠𝑠) = 𝜎𝜎 ∫0 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝐵𝐵𝑠𝑠 [1]
𝑡𝑡
So 𝑓𝑓(𝑋𝑋𝑡𝑡 , 𝑡𝑡) − 𝑓𝑓(𝑋𝑋0 , 0) = 𝜎𝜎 ∫0 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝐵𝐵𝑠𝑠 [1]
𝑡𝑡
So 𝑋𝑋𝑡𝑡 𝑒𝑒 𝛾𝛾𝛾𝛾 − 𝑋𝑋0 = 𝜎𝜎 ∫0 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝐵𝐵𝑠𝑠 [1]
𝑡𝑡
So 𝑋𝑋𝑡𝑡 = 𝑋𝑋0 𝑒𝑒 −𝛾𝛾𝛾𝛾 + 𝜎𝜎 ∫0 𝑒𝑒 𝛾𝛾(𝑠𝑠−𝑡𝑡) 𝑑𝑑𝐵𝐵𝑠𝑠 ] [1]
[Or using an integrating factor:
Use the integrating factor 𝑒𝑒 𝛾𝛾𝛾𝛾 . [1]
Then 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝑋𝑋𝑠𝑠 = −𝛾𝛾𝑋𝑋𝑠𝑠 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝑑𝑑 + 𝑒𝑒 𝛾𝛾𝛾𝛾 𝜎𝜎𝜎𝜎𝐵𝐵𝑠𝑠 [1]
So 𝛾𝛾𝑋𝑋𝑠𝑠 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝑑𝑑 + 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝑋𝑋𝑠𝑠 = 𝑒𝑒 𝛾𝛾𝛾𝛾 𝜎𝜎𝜎𝜎𝐵𝐵𝑠𝑠 [1]

𝑑𝑑
So (𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝑋𝑋𝑠𝑠 ) = 𝑒𝑒 𝛾𝛾𝛾𝛾 𝜎𝜎𝜎𝜎𝐵𝐵𝑠𝑠 [1]
𝑑𝑑𝑑𝑑
𝑡𝑡 𝑑𝑑 𝑡𝑡
Then ∫0 (𝑒𝑒 𝛾𝛾𝛾𝛾 𝑑𝑑𝑋𝑋𝑠𝑠 ) = ∫0 𝑒𝑒 𝛾𝛾𝛾𝛾 𝜎𝜎𝜎𝜎𝐵𝐵𝑠𝑠 [1]
𝑑𝑑𝑑𝑑
𝑡𝑡
So 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑋𝑋𝑡𝑡 − 𝑒𝑒 𝛾𝛾0 𝑋𝑋0 = 𝜎𝜎 ∫0 𝑒𝑒 𝛾𝛾𝛾𝛾 𝜎𝜎𝜎𝜎𝐵𝐵𝑠𝑠 [1]
𝑡𝑡
So 𝑒𝑒 𝛾𝛾𝛾𝛾 𝑋𝑋𝑡𝑡 = 𝑋𝑋0 + 𝜎𝜎 ∫0 𝑒𝑒 𝛾𝛾𝛾𝛾 𝜎𝜎𝜎𝜎𝐵𝐵𝑠𝑠 [1]
𝑡𝑡
So 𝑋𝑋𝑡𝑡 = 𝑋𝑋0 𝑒𝑒 −𝛾𝛾𝛾𝛾 + 𝜎𝜎 ∫0 𝑒𝑒 𝛾𝛾(𝑠𝑠−𝑡𝑡) 𝑑𝑑𝐵𝐵𝑠𝑠 ] [1]
The majority of students scored more than half marks on this

knowledge based question.
Common approaches to solve the SDE included using an

integrating factor or Ito’s Lemma.
Some students went on to calculate the distribution of the solution

and its long-term mean and variance despite not being asked for
this in the question.
6 Q6
i) Portfolio A = one call plus cash of Kexp(-r(T-t)) [1]
Portfolio B = one put plus one share [1]
Both portfolios have value max{K,S T } at expiry, hence by the principle of no
arbitrage they must have the same value at all earlier times. [1]
Page 8
Hence c t + Ke-r(T-t) = p t + S t [1]

[Note to markers: any alternative valid solution is acceptable]
ii) By put-call parity [0.5]

B = 0.32 + 8exp(-0.03*10) – 5 [1]
= $1.25 [0.5]
iii) C >= max{ 0 , 10exp(-0.03*10) – 5 } [0.5]

= $2.41 [0.5]
C<= 10exp(-0.03*10) [0.5]
= $7.41 [0.5]
Alternatively, option C is the same as B but with a strike price $2 higher. It can never be in
the money by more than $2 more than B, so it can never be worth more than $2 more than B.
[1]
Hence it can’t be worth more than $3.25. [1]
[Max 2]
iv) D >= 10 – 5 = $5 [1]
D <= $10 [1]

(The American option can be exercised early, but its value will never be more than $10.)
[Or the American option is worth at least as much as the European option for 0.5 marks.]
Part (i) was a knowledge based question with well-prepared

students having little difficulty answering. Common mistakes
included using incorrect portfolios or not explicitly applying the
assumption of no arbitrage to prove the portfolios had the same
value at the beginning.
In parts (iii) and (iv) most students identified the correct bounds but
some struggled with the American option.
Q7
i) 100,000 + 250,000 = $350,000 [1]
ii) Under the Merton model, we consider that the shareholders have a call option on the
company’s assets, with a strike price equal to the nominal value of the debt. [1]
We want the share price to remain unchanged, so the value of the ‘call option’ after
the debt has been issued must be $1 per share = $100,000 in total. [1]
Page 9
We need to find the nominal value of the debt at maturity, which will be the strike
price of this option. [1]
[Max 2 so far]
Trying a strike price of $250,000 gives a call value of $137,811 [1]
Trying a strike price of $350,000 gives a call value of $71,960 [1]
Sample answers to help with marking:

Strike price d1 d2 Call option value
$200,000 2.2834 1.948 $178,363
$250,000 1.6181 1.2827 $137,811
$300,000 1.0745 0.7391 $101,703
$350,000 0.6149 0.2795 $71,960
$400,000 0.2168 -0.1186 $49,148
Interpolating gives a strike price (i.e. nominal debt value at maturity) of $307,419
(actual value is $302,582) [1]
iii) The required yield on the debt is i where 250,000e5i = 302,582 => i = 3.82% [1]
=> credit spread = 3.82% - 3% = 0.82% [1]
iv) The equities fell to 50% of their original value [1]

The debt fell to (0.5 * 302,582) / 250,000 = 60.5% of its original value [1]
[Alternatively the debt fell by 39.5%.]
v) The debt ranks above the equities on company default. [1]

Hence the debt holders have a more secure investment. [1]
They will almost always receive something at maturity, and may receive the whole
value. [1]
The equity holders will receive nothing on default. [1]
And might receive nearly nothing even if the company does not default. [1]
[Max 3]
[Max 3]
Most students struggled with this question with most failing to score
more than a few marks.
In general, students did not understand they had been given the
value of equity and debt and had to solve for the redemption value
that was consistent with the values given.
Common mistakes included calculating the value of the equity using

the current value of the debt as a redemption value and then
proceeding through the question. This led to students using a
Page 10
present value of debt that was higher than the redemption value,
leading to a negative credit spread.
Q8
i.
Stock tree
time 0 3 month 6 month 9 month
60.00 78.00 101.40 131.82
48.00 62.40 81.12
38.40 49.92
30.72
The price C 0 of the option is computed via Risk Neutral Valuation; let 𝑝𝑝̂ denote the risk neutral
probability of an up movement, then
𝑒𝑒 0.02∗0.25 −0.80
𝑝𝑝̂ = = 0.41 [1]
1.30−0.80
and
3
𝐶𝐶0 = 𝑒𝑒 −𝑟𝑟𝑟𝑟 � �𝑘𝑘3 �𝑝𝑝̂ 𝑘𝑘 (1 − 𝑝𝑝̂ )3−𝑘𝑘 𝑚𝑚𝑚𝑚𝑚𝑚�0, 𝑆𝑆0 𝑢𝑢𝑘𝑘 𝑑𝑑3−𝑘𝑘 − 𝐾𝐾� [1]
𝑘𝑘=0
= 𝑒𝑒 −0.02∗0.75 (76.82 × 𝑝𝑝̂ 3 + 26.12 × 3𝑝𝑝̂ 2 (1 − 𝑝𝑝̂ ) ) = 12.87 [1]
The detailed workings are provided below – in case attempts to answer this
question go through the whole tree.
CALL
time 0 3 month 6 month 9 month
12.87 25.30 46.67 76.82
4.35 10.66 26.12
0.00 0.00
0.00
ii. Either from the put-call parity or by repeating calculation: 𝑃𝑃0 = 7.05 [1]
iii. The value of the position is 𝐶𝐶0 − 𝑃𝑃0 = 𝑆𝑆0 − 𝐾𝐾𝐾𝐾 −𝑟𝑟𝑟𝑟 [1].
This value would not change as it is independent of the expected movements of the
stock (i.e. its volatility) [2]
[Or students can recalculate the value for full marks.]
Parts (i) and (ii) were well-answered by the majority of candidates.
Page 11
For part (ii), some students calculated the put price from first
principles rather than simply using the put-call parity relationship.
This was a valid approach but took up more time in the exam.
For part (iii), the majority of students re-calculated the prices of

the put and call directly, though many made mistakes and hence
failed to conclude that the portfolio value remains unchanged.
Q9
i. From the definition, 𝑉𝑉 (𝑡𝑡) = Φ𝑡𝑡 𝑆𝑆𝑡𝑡 + Ψ𝑡𝑡 𝐵𝐵𝑡𝑡 [1]

−𝑟𝑟𝑟𝑟 −𝑟𝑟𝑟𝑟
therefore 𝑒𝑒 𝑉𝑉 (𝑡𝑡) = Φ𝑡𝑡 𝑒𝑒 𝑆𝑆𝑡𝑡 + Ψ𝑡𝑡 [1]
because the portfolio is self-financing it follows that 𝑑𝑑(𝑒𝑒 −𝑟𝑟𝑟𝑟 𝑉𝑉 (𝑡𝑡)) = Φ𝑡𝑡 𝑑𝑑(𝑒𝑒 −𝑟𝑟𝑟𝑟 𝑆𝑆𝑡𝑡 )
[1]
as required.
ii. Using the result from the previous part, the martingale property for the discounted
value of the portfolio is the same as for the discounted stock price. [1]
This requires a change of measure to adjust for the market price of risk 𝜆𝜆 = 𝜎𝜎
[2]
[Or we can apply Taylor’s theorem to d(S t e-rT) and check that the drift is zero.]
[Or we could use Ito’s Lemma.]
[Or just 𝜆𝜆 = 𝑟𝑟 as a possible solution for one mark.]
Several approaches were used to prove that this is a martingale. A

common approach included Ito's Lemma, while other students used
the five-step method and applied the Martingale Representation
Theorem.
In part (ii), some students simply repeated the definition or

properties of a martingale rather than considering the conditions
for the discounted share price process that would make it a
martingale.
Q10
i. The assumptions underlying the Black-Scholes model are as follows:

1. The price of the underlying share follows a geometric Brownian motion. [1/2]
2. There are no risk-free arbitrage opportunities. [1/2]
Page 12
3. The risk-free rate of interest is constant, the same for all maturities and the same
for borrowing or lending. [1/2]
4. Unlimited short selling (that is, negative holdings) is allowed. [1/2]
5. There are no taxes or transaction costs. [1/2]
6. The underlying asset can be traded continuously and in infinitesimally small
numbers of units. [1/2]
ii.
Data: 𝑆𝑆 = 8; 𝐾𝐾 = 9; 𝑟𝑟 = 2%; 𝜎𝜎 = 20%; 𝑇𝑇 = 0.25
By the Black-Scholes formula:
−𝑑𝑑1 = 1.0778 [0.5]
−𝑑𝑑2 = 1.1778 [0.5]
𝑁𝑁(−𝑑𝑑1 ) = 0.8594 [0.5]
𝑁𝑁(−𝑑𝑑2 ) = 0.8806 [0.5]
Therefore 𝑃𝑃0 = 9𝑒𝑒 −0.02×0.25 × 0.8806 −8 × 0.8594 [1]
= 1.01 [1]
iii. As interest rates increase in the market, the expected return required by investors in stock
tends to increase [0.5]
However, the present value of any future cash flow generated by option contracts decreases
[0.5]
The combined impact of these two effects is to decrease the value of the put option [1]
Rho is negative for a put option [0.5]
put options become less valuable in times of increasing interest rates because they
effectively defer the selling of a share and so delay access to the cash required to obtain the
risk-free rate
[0.5]
[Or students could explain how the terms in the formula change.]
[Max 2]
This was well-answered overall by the majority of students.
For part (i), some students included assumptions from Expected

Utility Theory, the Efficient Market Hypothesis or CAPM which
scored no marks.
For part (ii), simple calculation errors were the most common
mistake.
Q11
i. SML: 𝐸𝐸𝑅𝑅𝑖𝑖 = 𝑅𝑅𝑓𝑓 + 𝛽𝛽𝑖𝑖 (𝐸𝐸𝑅𝑅𝑀𝑀 − 𝑅𝑅𝑓𝑓 ) [1]
for
• 𝐸𝐸𝑅𝑅𝑖𝑖 : expected return on Asset i. [1/4]
• 𝑅𝑅𝑓𝑓 : risk-free rate. [1/4]
• 𝛽𝛽𝑖𝑖 : beta factor of security i defined as 𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅𝑖𝑖 , 𝑅𝑅𝑀𝑀 )/𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅𝑀𝑀 ). [1/4]
Page 13
• 𝐸𝐸𝑅𝑅𝑀𝑀 : expected return on the market portfolio. [1/4]

[Round up to nearest half mark.]
ii. Note that 𝐸𝐸𝑅𝑅𝑀𝑀 = 𝑥𝑥1 𝐸𝐸𝑅𝑅1 + 𝑥𝑥2 𝐸𝐸𝑅𝑅2 [1/4], and 𝑥𝑥1 + 𝑥𝑥2 = 1 [1/4]
Substitute into the SML and solve for 𝑥𝑥1 , so that
𝐸𝐸𝑅𝑅𝑖𝑖 −𝛽𝛽𝑖𝑖 𝐸𝐸𝑅𝑅2 −𝑅𝑅𝑓𝑓 (1−𝛽𝛽𝑖𝑖 )
𝑥𝑥1 = (𝐸𝐸𝑅𝑅1 −𝐸𝐸𝑅𝑅2 )𝛽𝛽𝑖𝑖
. [1]
From the data: 𝐸𝐸𝑅𝑅1 = 16.30% [1/4]
𝐸𝐸𝑅𝑅2 = 29.70% [1/4]
Substituting either for Asset 1 or Asset 2, 𝑥𝑥1 = 0.4 [1/2]
and therefore 𝑥𝑥2 = 0.6 [1/2]
[Alternatively, the beta of the market portfolio is 1, so x_1*0.46 + x_2*1.36 = x_1*0.46
+ (1-x_1)*1.36 = 1 => x_1=0.4]
[Round up to nearest half mark.]
iii. 𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅𝑀𝑀 ) = 0.42 ∗ 𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅1 ) + 0.62 ∗ 𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅2 ) + 2 ∗ 0.4 ∗ 0.6 ∗ 𝐶𝐶𝐶𝐶𝐶𝐶(𝑅𝑅1 , 𝑅𝑅2 ) = 0.00617.
[1]
Consequently �𝐸𝐸𝑅𝑅𝑀𝑀 − 𝑅𝑅𝑓𝑓 �/𝜎𝜎𝑀𝑀 = 1.897. [1]
The majority of students scored well in this question.
For part (i), some students confused the Security Market Line and
the Capital Market Line despite these being given in the Actuarial
Tables.
Page 14

Ilovepdf Merged 6

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ilovepdf Merged 6

Uploaded by

Copyright:

Available Formats

1

2 INSTITUTE AND FACULTY OF ACTUARIES

 Asset B is risk-free, so every £1 invested in Asset B will always be worth £1.

(iii) Express her expected utility of wealth in terms of a. [2]

(a) the variance of returns.

(ii) Describe how an efficient portfolio can be found under mean-variance

Rate of return (per annum)

1 0.2 5.0% 15.0% 26.0%

(i) Determine the composition of the market portfolio. [1]

(ii) Determine the market price of risk. [5]

(iii) Calculate the beta of each risky asset. [2]

(iv) State the limitations of the CAPM. [3]

CT8 A2016–3 PLEASE TURN OVER

Let V(t) = t St + t Bt be the value of the portfolio at time t.

(i) Explain what it means for this portfolio to be self-financing. [2]

The instantaneous risk-free rate is denoted by r.

(ii) Prove put-call parity in this context by constructing two self-financing

7 Consider a non-dividend-paying share with price St at time t (in years) in a market

A share is currently worth S0 = €20. The continuously compounded risk-free rate of

(b) Verify that there is no arbitrage in the given market.

Assume that K = 103 and the maturity of the contract is at time t = 3.

(iii) Determine the premium c of this contract. [3]

CT8 A2016–5 PLEASE TURN OVER

A company is in the Healthy state at time 0.

The fair price of the bond at time 0 is £87.63.

(iv) Calculate the credit spread on the bond. [1]

where k, ,  > 0 and W is a standard Brownian motion.

Consider the related process:

where rt is the short rate defined above.

(ii) Show that, under the Vasicek model:

Subject CT8 – Financial Economics

 Institute and Faculty of Actuaries

A. General comments on the aims of this subject and how it is marked

B. General comments on student performance in this diet of the

The Pass Mark for this exam was 60%.

Q1 (i) U’(w) = 1/w [½]

U’’(w) = –1/w2 [½]

Absolute risk aversion = A(w) = – U’’(w)/U’(w) [½]

A’(w) = –1/w2 [½]

Relative risk aversion = R(w) = – wU’’(w)/U’(w) [½]

This is consistent with an investor who keeps a constant proportion of wealth

(iii) Wealth after the uncertain event will be either:

100 × (1.3a + (1 – a)) = 100 + 30a with probability 0.75 [½]

100 × (0.4a + (1 – a)) = 100 – 60a with probability 0.25. [½]

Thus expected utility of wealth is:

0.75 × ln(100 + 30a) + 0.25 × ln(100 – 60a). [2]

(iv) Differentiate with respect to a:

30 × 0.75/(100 + 30a) – 60 × 0.25/(100 – 60a). [2]

Set equal to zero:

30 × 0.75 / (100 + 30a) – 60 × 0.25 / (100 – 60a) = 0

Check for maximum:

Differentiate with respect to a again:

– 302 × 0.75/(100 + 30a)2 – 602 × 0.25/(100 – 60a)2.

So invest £41.67 in Asset Aand £58.33 in Asset B. [2]

Q2 (i) Variance of return is defined as:

(b) This corresponds to a utility function which has a discontinuity at the

Q3 (i) V = Σi xi2 Vi + Σi Σj, j≠i xi .xj Cij . [Total 1]

(ii) The aim is to choose xi to minimise V … [1]

… subject to the constraints

One way of solving such a minimisation problem is the method of Lagrangian

The Lagrangian function is:

W = V − λ(E − EP) − μ(Σi xi − 1). [1]

The result is a set of linear equations that can be solved. [1]

The usual way of representing the results of the above calculations is by