om

i.c
CM2-05: Stochastic models of investment returns Page 1

ng
si
om
o m
as
.m
w
w
w
Stochastic models of
investment returns
Syllabus objectives

3.1 Show an understanding of simple stochastic models for investment returns.

3.1.1 Describe the concept of a stochastic investment return model and the
fundamental distinction between this and a deterministic model.
3.1.2 Derive algebraically, for the model in which the annual rates of return are
independently and identically distributed and for other simple models,
expressions for the mean value and the variance of the accumulated
amount of a single premium.
3.1.3 Derive algebraically, for the model in which the annual rates of return are
independently and identically distributed, recursive relationships which
permit the evaluation of the mean value and the variance of the
accumulated amount of an annual premium.
3.1.4 Derive analytically, for the model in which each year the random variable
(1  i) has an independent lognormal distribution, the distribution
functions for the accumulated amount of a single premium and for the
present value of a sum due at a given specified future time.
3.1.5 Apply the above results to the calculation of the probability that a simple
sequence of payments will accumulate to a given amount at a specific
future time.

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 2 CM2-05: Stochastic models of investment returns

ng
si
om
o m
0 Introduction

as
.m
w
Financial contracts are often of a long-term nature. Accordingly, at the outset of many

w
contracts there may be considerable uncertainty about the economic and investment

w
conditions which will prevail over the duration of the contract.

Therefore, we might choose to develop models for future investment returns to help us study
that inherent uncertainty.

A deterministic model is a model that provides an output based on one set of parameter and
input variables. However, deciding which set of input variables to use may be a challenge.
Thus, for example, if it is desired to determine premium rates on the basis of one fixed rate
of return, it is nearly always necessary to adopt a conservative basis for the rate to be used
in any calculations, subject to the premium being competitive.

The deterministic approach can provide only a single fixed answer to a problem. This answer will
be correct only if the assumptions made about the future turn out to be correct.

When setting premium rates for insurance contracts, we need to make some assumption about
future investment returns, as the premiums received will (at least, in part) be invested in order to
meet the cost of future claims. Adopting a conservative basis for the rate of return means that
the rate selected is lower than we actually expect to receive in the future. The difference
between the assumed rate in the model and the expected rate provides a margin for uncertainty.
However, if the assumed rate of return is too low, this will result in premiums that are too high to
be competitive.

An alternative approach to recognising the uncertainty that in reality exists is provided by

the use of stochastic models. In such models, no single rate is used and variations are
allowed for by the application of probability theory.

Since we cannot specify in advance precisely what investment returns will be, in a stochastic
model, we make an assumption about the statistical distribution of future investment returns.
This enables us to consider the expected accumulated value of an investment at a future date and
the variance of that accumulated value. As we have seen, considering the variance of the future
value of a fund is one way of measuring the risk associated with our choice of investments.

The stochastic approach can give unreliable results if the statistical distribution used is not
appropriate.

Possibly one of the simplest models is that in which each year the rate obtained is
independent of the rates of return in all previous years and takes one of a finite set of
values, each value having a constant probability of being the actual rate for the year.

For example, the effective annual rates of return that will apply during each of the next n years
might be i1 , i2 ,..., in , where ik , k  1,2,..., n are random variables with the following discrete
distribution:

0.06 with probability 0.2


i k  0.08 with probability 0.7
0.10 with probability 0.1


© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 3

ng
si
om
o m
Question

as
.m
w
Calculate the mean, j, and the standard deviation, s, of ik .

w
w
Solution

The mean is:

j  E  ik   0.2  0.06  0.7  0.08  0.1  0.10  0.078

We can calculate the variance using the formula Var (ik )  E ik2   E 2  ik  :
 

E ik2   0.2  0.062  0.7  0.082  0.1  0.102  0.0062

 

 s2  Var (ik )  E  ik2   E 2  ik   0.0062  0.0782  0.000116  0.01082

 

So, the mean is 7.8% and the standard deviation is 1.08%.

Alternatively, the rate may take any value within a specified range, the actual value for the
year being determined by some given probability density function.

For example we might assume that the annual rates of return are uniformly distributed between
5% and 10%, or that the annual growth factor 1  i follows a lognormal distribution with given
parameter values, as we do in Section 2.

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 4 CM2-05: Stochastic models of investment returns

ng
si
om
o m
1 Simple models

as
.m
w
1.1 Fixed rate model

w
w
At this stage we consider briefly an elementary example, which – although necessarily
artificial – provides a simple introduction to the probabilistic ideas implicit in the use of
stochastic rate models.

Suppose that an investor wishes to invest a lump sum of P into a fund with compound
investment rate growth at a constant rate for n years. This constant investment return is
not known now, but will be determined immediately after the investment has been made.

This model, where the effective annual rate of return is a single unknown rate i, which will apply
throughout the next n years, is often known as the fixed rate model.

The accumulated value of the sum will, of course, be dependent on the investment rate. In
assessing this value before the investment rate is known, it could be assumed that the
mean rate will apply. However, the accumulated value using the mean rate will not equal the
mean accumulated value. In algebraic terms:

n
 k   k 
P  1
  (i j p j )   P 
 
 j 1
p j (1  i j )n 

 j 1   

where:

i j is the jth of k possible investment rates of return

p j is the probability of the investment rate of return i j

The above result is easily demonstrated with the following simple numerical question.

Question

The returns from an investment are assumed to conform to the fixed rate model with the
distribution of rates as specified below:

0.06 with probability 0.2


i k  0.08 with probability 0.7
0.10 with probability 0.1


(i) Calculate the expected accumulated value at the end of 5 years of an initial investment of
£5,000.

(ii) Calculate the accumulated value at the mean rate of return.

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 5

ng
si
om
o m
Solution

as
.m
w
(i) The expected accumulated value is:

w
w
5,000E[(1  ik )5 ]  5,000(0.2  1.065  0.7  1.085  0.1  1.105 )
 5,000  1.4572
 £7,286

(ii) The mean rate of return is:

E  ik   0.2  0.06  0.7  0.08  0.1  0.10  0.078

Therefore the accumulated value at the mean rate of return is:

5,000  1.0785  £7,279

As expected, we see that the expected accumulated value is not equal to the accumulated value
calculated at the expected rate of return.

Question

Calculate the variance of the accumulated value of the investment in the previous question.

Solution

The variance of the accumulated value, s2 , is:

s2  5,0002 (E[(1  ik )10 ]  (E[(1  ik )5 ])2 )

 5,0002 ((0.2  1.0610  0.7  1.0810  0.1  1.1010 )

(0.2  1.065  0.7  1.085  0.1  1.105 )2 )

 5,0002  (2.128791  1.4572262 )  (£363)2

It’s important to keep a few extra decimals in this last calculation to avoid losing accuracy, since
the calculation involves subtracting two numbers of similar magnitude.

For the fixed rate model, the mean and variance of the accumulated value of an investment must
be calculated from first principles.

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 6 CM2-05: Stochastic models of investment returns

ng
si
om
o m
1.2 Varying rate model

as
.m
In our previous example the effective annual investment rate of return was fixed throughout

w
w
the duration of the investment. A more flexible model is provided by assuming that over

w
each single year the annual yield on invested funds will be one of a specified set of values
or lie within some specified range of values, the yield in any particular year being
independent of the yields in all previous years and being determined by a given probability
distribution.

This model is often called the varying rate model. The main difference between this and the fixed
rate model is that, in the varying rate model, the rates of return can be different in each future
year, whereas, in the fixed rate model, the same (unknown) rate of return will apply in each
future year.

Measure time in years. Consider the time interval [0, n ] subdivided into successive periods
[0,1], [1,2], , [ n  1, n ] . For t  1,2, , n let it be the yield obtainable over the tth year, ie the
period [t  1, t ] . Assume that money is invested only at the beginning of each year. Let Ft
denote the accumulated amount at time t of all money invested before time t and let Pt be
the amount of money invested at time t. Then:

Ft  (1  it )(Ft 1  Pt 1) , for t  1,2,3, (1.1)

It follows from this equation that a single investment of 1 at time 0 will accumulate at time n
to:

Sn  (1  i1)(1  i2 ) (1  in ) (1.2)

Similarly, a series of annual investments, each of amount 1, at times 0,1,2, , n  1 will

accumulate at time n to:

An  (1  i1)(1  i2 )(1  i3 ) (1  in )

 (1  i2 )(1  i3 ) (1  in )

 (1.3)
 (1  in 1)(1  in )

 (1  in )

Note that An and Sn are random variables, each with its own probability distribution
function.

For example, if the yield each year is 0.02, 0.04, or 0.06 and each value is equally likely, the
value of Sn will be between 1.02n and 1.06n . Each of these extreme values will occur with
probability (1 3)n .

Question

Determine the probability that Sn will take the value 1.02  1.04n1 .

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 7

ng
si
om
o m
Solution

as
.m
This value of Sn will occur if the rate of return is 2% in one year and 4% in the remaining n 1

w
w
w
years. Since there are n different years in which the 2% could fall, the probability is:

n1
1 1 n
n   
3 3 3n

In general, a theoretical analysis of the distribution functions for An and Sn is somewhat

difficult. It is often more useful to use simulation techniques in the study of practical
problems. However, it is perhaps worth noting that the moments of the random variables
An and Sn can be found relatively simply in terms of the moments of the distribution for
the yield each year. This may be seen as follows.

Moments of Sn

Let’s consider the kth moment of Sn , where Sn denotes the accumulated value at time n of an
initial investment of 1 made at time 0.

From Equation (1.2) we obtain:

n
(Sn )k   (1  it )k
t 1

and hence:

 n 
E Snk   E   (1  it )k 
   t 1 

n
  E (1  it )k  (1.4)
t 1

since (by hypothesis) i1, i2 , , in are independent. Using this last expression and given the
moments of the annual yield distribution, we may easily find the moments of Sn .

For example, suppose that the yield each year has mean j and variance s 2 . Then, letting
k  1 in Equation (1.4), we have:

n
E [Sn ]   E [(1  it )]
t 1

n
  1  E [it ]
t 1

 (1  j )n (1.5)

since, for each value of t, E [ it ]  j .

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 8 CM2-05: Stochastic models of investment returns

ng
si
om
m
With k  2 in Equation (1.4) we obtain:

o
as
.m
n
E [Sn2 ]   E (1  2it  it2 )

w
w
t 1

w
n
  (1  2E [it ]  E [it2 ])
t 1

 (1  2 j  j 2  s 2 )n (1.6)

since, for each value of t:

E [ it2 ]  (E [ it ])2  Var (it )  j 2  s 2

The variance of Sn is:

Var (Sn )  E [Sn2 ]  (E [Sn ])2

 (1  2 j  j 2  s 2 )n  (1  j )2n (1.7)

from Equations (1.5) and (1.6).

Alternatively, we can write:

n n
E  Sn2    E[(1  it )2 ]     E 1  it   Var (1  it ) 
2
   
t 1 t 1

using a rearrangement of the variance formula: Var ( X )  E  X 2    E  X  .

2
 

Then, using the results that:

 E 1  it   1  E it 
2 2
 (1  j)2 and Var (1  it )  Var (it )  s2

for all values of t , we can write:

   
n n
E  Sn2    (1  j)2  s2  (1  j)2  s2
 
t 1

This means that:

 
n
Var (Sn )  (1  j)2  s2  (1  j)2n

These arguments are readily extended to the derivation of the higher moments of Sn in
terms of the higher moments of the distribution of the annual investment rates of return.

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 9

ng
si
om
o m
Question

as
.m
w
Calculate the mean and variance of the accumulated value at the end of 25 years of an initial

w
investment of £40,000, if the annual rate of return in year k is independent of that in any other

w
year and ik  Gamma (16,200) for all k .

Solution

Since ik  Gamma(16,200) , using the formulae for the mean and variance of the gamma
distribution from page 12 of the Tables:

 16
j  E  ik     0.08
 200

 16
s2  Var (ik )    (0.02)2
 2
200 2

So, the mean of the accumulated amount is:

40,000E  S25   40,000(1  j)25  40,000  1.0825  £273,939

and the variance is:

Var (40,000 S25 )  40,0002 Var (S25 )


 40,0002 [(1  j)2  s2 ]25  (1  j)50 
 40,0002 [(1.082  0.022 )25  1.0850 ]

 (£25,417)2

Moments of An

Recall that An is a random variable that represents the accumulated value at time n of a series of
annual investments, each of amount 1, at times 0, 1,2, , n  1 . i1 , i2 ,  , in are independent
random variables, each with a mean j and a variance s2 .

From Equation (1.3):

An1  (1  i1 )(1  i2 )  (1  in1 )

 (1  i2 )  (1  in1 )


 (1  in2 )(1  in1 )

 (1  in1 )

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 10 CM2-05: Stochastic models of investment returns

ng
si
om
m
Also:

o
as
.m
An  (1  i1 )(1  i2 )  (1  in1 )(1  in )

w
w
 (1  i2 )  (1  in1 )(1  in )

w


 (1  in1 )(1  in )

 (1  in )

It follows from Equation (1.3) (or from Equation (1.1)) that, for n  2 :

An  (1  in )(1  An 1) (1.8)

Equation (1.8) can also be deduced by general reasoning. An1 is the accumulated value at time
n 1 of a series of annual payments, each of amount 1, at times 0, 1, 2,  , n  2 . The value, at
time n 1 , of the same series of payments together with an extra payment at time n 1 is
1  An1 . Accumulating this value forward to time n gives (1  in )(1  An1 ) and this is equivalent
to An .

The usefulness of Equation (1.8) lies in the fact that, since An 1 depends only on the values
i1, i2 , , in 1 , the random variables in and An 1 are independent. (By assumption the yields
each year are independent of one another.) Accordingly, Equation (1.8) permits the
development of a recurrence relation from which may be found the moments of An . We
illustrate this approach by obtaining the mean and variance of An .

Let:

n  E [ An ]

and let:

mn  E [ An2 ]

Since:

A1  1  i1

it follows that:

E[ A1 ]  E[1  i1 ]  1  E[i1 ]  1  j  1  1  j

and:

m1  E[ A12 ]  E[(1  i1 )2 ]  1  2E[i1 ]  E[i12 ]  m1  1  2 j  j 2  s 2

where, as before, j and s 2 are the mean and variance of the yield each year.

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 11

ng
si
om
m
Taking expectations of Equation (1.8), we obtain (since in and An 1 are independent):

o
as
.m
n  (1  j )(1  n 1) n2

w
w
w
This equation, combined with initial value 1 , implies that, for all values of n:

n  sn at rate j (1.9)

We obtain this result by applying the recursive formula for n repeatedly for each year:

n  (1  j)  (1  j)n1

 (1  j)  (1  j)2 [1  n2 ]

 (1  j)  (1  j)2  (1  j)2 n2

 ...

 (1  j)  (1  j)2  (1  j)3    (1  j)n

 , calculated at the mean rate of return.

Thus the expected value of An is simply s n

Recall that the symbol sn denotes the accumulated value at time n of payments of 1 at times
0, 1, 2,…, n 1 . At the interest rate j , it is equal to:

(1  j)n  1
sn  (1  j)  (1  j)2  (1  j)3    (1  j)n 

d

j
where d  .
1 j

Next, we consider the variance of An .

Since:

An2  (1  2in  in2 )(1  2 An 1  An2 1 )

by taking expectations we obtain, for n  2 :

mn  (1  2 j  j 2  s 2 )(1  2 n 1  mn 1) (1.10)

As the value of n 1 is known (by Equation (1.9)), Equation (1.10) provides a recurrence
relation for the calculation successively of m2 , m3 , m4 , . The variance of An may be
obtained as:

Var ( An )  E [ An2 ]  (E [ An ])2  mn  n2 (1.11)

In principle the above arguments are fairly readily extended to provide recurrence relations
for the higher moments of An .

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 12 CM2-05: Stochastic models of investment returns

ng
si
om
o m
Question

as
.m
A company considers that on average it will earn interest on its funds at the rate of 4% pa.

w
w
However, the investment policy is such that in any one year the yield on the company’s

w
funds is equally likely to take any value between 2% and 6%.

For both single and annual premium accumulations with terms of 5, 10, 15, 20, and 25 years
and single (or annual) investment of £1, find the mean accumulation and the standard
deviation of the accumulation at the maturity date. (Ignore expenses.)

Solution

The annual rate of return is uniformly distributed on the interval [0.02,0.06] . The
corresponding probability density function is constant and equal to 25 (ie 1 (0.06  0.02) ).
The mean annual rate of interest is clearly:

j  0.04

and the variance of the annual rate of return is:

1
s2  (0.06  0.02)2  34  104
12

The formulae for the PDF, mean and variance of a uniform random variable are given on page 13
of the Tables.
1 1
We are required to find E [ An ] , Var ( An )  2 , E [Sn ] , and Var (Sn )  2 for n  5, 10, 15, 20 and
25.

Substituting the above values of j and s 2 in Equations (1.5) and (1.7), we immediately
obtain the results for the single premiums.

For example:

E[ S5 ]  1.045  1.21665

Var (S5 )  (1  0.08  0.042  43 104 )5  1.0410  0.000913

 standard deviation [ S5 ]  0.000913  0.03021


For the annual premiums we must use the recurrence relation (1.10) (with n 1  s at
n 1
4%) together with Equation (1.11).

Equation (1.9) is used to calculate E[ An ] . For example:

1.045  1
E[ A5 ]  
s5 @4%   5.63298
0.04 1.04

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 13

ng
si
om
o m
To calculate the standard deviation of A5 , we first need to calculate m5 from the recursive

as
formula:

.m
w
w
mn  (1  2 j  j2  s2 )(1  2n1  mn1 )

w
starting with 1  1  j and m1  1  2 j  j2  s2 .

The values required are tabulated below:

n mn n

1 1.08173 1.04
2 4.50189 2.1216
3 10.54158 3.24646
4 19.50853 4.41632
5 31.73933018 5.6329755

Therefore, using Equation (1.11):

Var (A5 )  m5  (5 )2  31.73933018  5.63297552  0.0089172

So the standard deviation is:

0.0089172½  0.09443

This answer is very sensitive to rounding.

The results are summarised in Table 1. It should be noted that, for both annual and single
premiums, the standard deviation of the accumulation increases rapidly with the term.

Single investment £1 Annual investment £1

Term
(years) Mean Standard Mean Standard
accumulation (£) deviation (£) accumulation (£) deviation (£)
5 1.21665 0.03021 5.63298 0.09443
10 1.48024 0.05198 12.48635 0.28353
15 1.80094 0.07748 20.82453 0.57899
20 2.19112 0.10886 30.96920 1.00476
25 2.66584 0.14810 43.31174 1.59392

Table 1

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 14 CM2-05: Stochastic models of investment returns

ng
si
om
m
This approach is also easily extended to provide recurrence relations for other series of

o
as
investments. Let Ft again represent the accumulated amount at time t of all money invested

.m
w
before time t and let Pt be the amount of money invested at time t. We stated in Equation (1.1)

w
w
that:

Ft  (1  it )(Ft 1  Pt 1 )

Therefore the mean of Ft can be found from the recursive relationship:

E[F0 ]  0

E[Ft ]  (1  j)(E[Ft 1 ]  Pt 1 )

Question

An investor invests 1 unit at time t  0 and a further 2 units at time t  2 . The expected rate of
return in each year is 10%. Calculate the accumulated value of the fund at time t  5 :

(i) using recursive formulae, and assuming the varying rate model applies

(ii) using the corresponding deterministic model.

Comment on your answers.

Solution

(i) Using a recursive approach, where Ft represents the accumulated amount at time t of all
money invested before time t, we obtain:

E[F0 ]  0

E[F1 ]  1.1  [E[F0 ]  P0 ]  1.1  (0  1)  1.1

E[F2 ]  1.1  [E[F1 ]  P1 ]  1.1  (1.1  0)  1.21

E[F3 ]  1.1  [E[F2 ]  P2 ]  1.1  (1.21  2)  3.531

E[F4 ]  1.1  [E[F3 ]  P3 ]  1.1  (3.531  0)  3.8841

E[F5 ]  1.1  [E[F4 ]  P4 ]  1.1  (3.8841  0)  4.27251

(ii) Using the corresponding deterministic model, with an annual rate of return of 10%, gives:

1.15  2  1.13  4.27251

So the corresponding deterministic model gives the same answer.

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 15

ng
si
om
o m
2 The lognormal distribution

as
.m
w
In general a theoretical analysis of the distribution functions for An and Sn is somewhat

w
w
difficult, even in the relatively simple situation when the yields each year are independent
and identically distributed. There is, however, one special case for which an exact analysis
of the distribution function for Sn is particularly simple.

Due to the compounding effect of investment returns, the accumulated value of an investment
bond grows multiplicatively. This makes the lognormal distribution a natural choice for modelling
the annual growth factors 1  i , since a lognormal random variable can take any positive value
and has the following multiplicative property:

If X1  log N( 1, 12 ) and X2  log N( 2, 22 ) are independent random variables, then:

X1 X2  log N( 1   2, 12   22 )

The graph below illustrates the shape of the PDF of a typical lognormal distribution used to model
annual growth factors, 1  i .

log N(0.075,0.12 )

Question

Given that 1  i  log N(0.075,0.12 ) , calculate the mean and standard deviation of the annual
growth rate, i .

Solution

Using the formulae on page 14 of the Tables:

E[1  i ]  e  ½  e 0.075  ½  0.1  1.0833

2 2

Var (1  i)  e2   (e  1)  e2  0.075  0.1 (e 0.1  1)  0.10862

2 2 2 2
and:

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 16 CM2-05: Stochastic models of investment returns

ng
si
om
m
So, the annual growth rate will have a mean value of:

o
as
.m
E[i ]  8.33%

w
w
w
and a standard deviation of:

sd (i)  sd (1  i)  10.86%

Suppose that the random variable log(1  it ) is normally distributed with mean  and
variance  2 . In this case, the variable (1  it ) is said to have a lognormal distribution with
parameters  and  2 .

So, if log(1  it )  N( , 2 ) , then 1  it  log N( , 2 ) .

Equation (1.2) is equivalent to:

n
log Sn   log(1  it )
t 1

The sum of a set of independent normal random variables is itself a normal random
variable. Hence, when the random variables (1  it ) (t  1) are independent and each has a
lognormal distribution with parameters and  and  2 , the random variable Sn has a
lognormal distribution with parameters n  and n 2 .

Since the distribution function of a lognormal variable is readily written down in terms of its
two parameters, in the particular case when the distribution function for the yield each year
is lognormal we have a simple expression for the distribution function of Sn .

So Sn  log N(n , n 2 ) or log Sn  N(n , n 2 ) , and the distribution function of Sn can therefore
be written:

 log s  n 
P(Sn  s)    
  n 

Question

A man now aged exactly 50 has built up a savings fund of £400,000. In order to retire at age 60,
he will require a fund of at least £600,000 at that time. The annual returns on the fund, i , are
independent and identically distributed, with 1  i  log N(0.075,0.12 ) .

Calculate the probability that, if the man makes no further contributions to the fund, he will be
able to retire at age 60.

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 17

ng
si
om
o m
Solution

as
.m
w
If the man makes no further contributions, his accumulated fund at age 60 will be 400,000 S10 .

w
w
So, the probability that the fund will be sufficient for him to retire is:

 600,000 
P(400,000 S10  600,000)  1  P  S10 
 400,000 

 log1.5  10  
1 
  10 
 1  (1.0895)  (1.0895)  0.862

Similarly for the present value of a sum of 1 due at the end of n years:

Vn  (1  i1)1  (1  in )1

 logVn   log(1  i1)    log(1  in )

Since, for each value of t, log(1  it ) is normally distributed with mean  and variance  2 ,
each term on the right-hand side of the above equation is normally distributed with mean
  and variance  2 . Also the terms are independently distributed. So, logVn is normally
distributed with mean  n and variance n 2 . That is, Vn has lognormal distribution with
parameters  n and n 2 .

So Vn  log N(n , n 2 ) or log Vn  N(n , n 2 ) , and the distribution function of Vn can

therefore be written:

 log s  (n )   log s  n 

P(Vn  s)       
  n    n 

By statistically modelling Sn and Vn , it is possible to answer questions such as:

 to a given point in time, for a specified confidence interval, what is the range of values
for an accumulated investment

 what is the maximum loss which will be incurred with a given level of probability.

It can also be noted that these techniques may be extended to calculate the risk metrics
such as VaR, as introduced in a previous chapter, of a series of investments.

Question

The annual returns on a fund, i , are independent and identically distributed. Each year, the
distribution of 1  i is lognormal with parameters   0.075 and  2  0.0252 .

Calculate the upper and lower quartiles for the accumulated value at the end of 5 years of an
initial investment of £1,000.

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 18 CM2-05: Stochastic models of investment returns

ng
si
om
o m
Solution

as
.m
w
By definition, the accumulated amount 1,000 S5 will exceed the upper quartile u with probability

w
w
25%, ie:

0.75  P(1,000 S5  u)  P(S5  u 1,000)

So, using the formula for the distribution function of S5 :

 log(u 1,000)  5 
0.75  P(S5  u 1,000)    
  5 

From page 162 of the Tables, we find that (0.6745)  0.75 . So, we must have:

log(u 1,000)  5
 0.6745 ie u  1,000e5 0.6745 5  £1,511
 5

Similarly, the lower quartile is:

l  1,000e5 0.6745 5  £1,401

So far in this section, we have assumed that the annual growth factors in each year are
independent, ie we have assumed that the varying rate model applies. We can also use a
lognormal distribution for annual growth factors in conjunction with the fixed rate model.

Suppose that the rate of return on an investment, i , is currently unknown, but once determined,
it will be the same in all future years. In this case:

Sn  (1  i)n

If 1  i  log N( , 2 ) , then:

log(1  i)  N( , 2 )

and: log(1  i)n  n log(1  i)  N(n , n2 2 )

So: Sn  (1  i)n  log N(n , n2 2 )

Note that the distribution of Sn obtained here is different to that derived for the varying rate
model, where Sn  log N(n , n  2 ) , which applies in the case where the returns in each year are
independent.

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 19

ng
si
om
o m
Question

as
.m
w
A lump sum of $14,000 will be invested at time 0 for 4 years at an annual rate of return, i . The

w
rate of return, once determined, will be the same in each of the four years. 1  i has a lognormal

w
distribution with mean 1.05 and variance 0.007.

Calculate the probability that the investment will accumulate to more than $20,000 in 4 years’
time.

Solution

We first need to find the values of the parameters for the lognormal distribution. Using the
formulae for the mean and variance from the Tables:

  12  2
e  1.05 (Equation 1)

e2    e  1   0.007
2 2
and: (Equation 2)
 

Squaring Equation 1, we have:

2
   12  2  2   2
e  e  1.052
 

Substituting this into Equation 2 gives:

 0.007 
1.052  e  1   0.007   2  ln 
2
 1   0.006329
   1.052 

So, using Equation 1:

  ln(1.05)  12  2  0.04563

We know that 1  i  log N( , 2 ) . Since we have a constant rate of return over the 4 years, the
fixed rate model applies. Letting Sn denote the accumulated value at time n of an investment of
1 made at time 0, we have Sn  log N(n , n2 2 ) .

We need the probability P 14,000 S4  20,000  where:

S4  log N(4  ,16 2 )  log N(0.1825,0.1013)

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 20 CM2-05: Stochastic models of investment returns

ng
si
om
m
So:

o
as
.m
P 14,000 S4  20,000   P  S4  1.429 

w
w
 P ln S4  0.3567 

w
 0.3567  0.1825 
 P Z  
 0.1013 

 P  Z  0.547 

 1    0.547   0.2922

© IFE: 2019 Examinations The Actuarial Education Company

 om
i.c
CM2-05: Stochastic models of investment returns Page 21

ng
si
om
o m
as
Chapter 5 Summary

.m
w
w
A stochastic model of investment returns provides information about the distribution of

w
financial outcomes. This distribution can be used to find best estimates and probabilities.

The varying rate model and the fixed rate model provide formulae for the mean and variance
of the accumulated amount of a fund or the present value of a future payment.

Varying rate model (single premium)

E  Sn   (1  j)n Var (Sn )  [(1  j)2  s2 ]n  (1  j)2n

Fixed rate model (single premium)

E  Sn   E (1  i)n  Var (Sn )  E (1  i)2n   (E[(1  i)n ])2

   

For the varying rate model, the variance and higher moments of the accumulated amount of
a series of payments can be calculated using recursive formulae.

Varying rate model (annual premium)

E  An   
sn| at rate j

Recursive formulae for E  An  :

E  A0   0

and:

E  Ak   (1  j) 1  E  Ak 1  ( k  1,2,..., n )

Recursive formulae for Var (An ) :

E  A02   0
 

and:

   
E  Ak2   (1  j)2  s2 1  2E  Ak 1   E  Ak21 
   ( k  1,2,..., n )

Then:

Var (An )  E  An2    E  An 

2
 

The Actuarial Education Company © IFE: 2019 Examinations

 om
i.c
Page 22 CM2-05: Stochastic models of investment returns

ng
si
om
o m
For the fixed rate model with annual premiums, calculate the mean and variance directly

as
.m
from the definitions.

w
w
The lognormal distribution can be used to model the annual growth factor, 1  i . This allows

w
probabilities to be determined in terms of the distribution function of the normal distribution.

The lognormal model formulae for the varying rate model are:

 log s  n 
Sn  log N(n , n 2 ) P(Sn  s)    
  n 

 log s  (n )   log s  n 

Vn  log N(n , n 2 ) P(Vn  s)       
  n    n 

© IFE: 2019 Examinations The Actuarial Education Company