Examinations: 18 April 2000 (Am)

Faculty of Actuaries Institute of Actuaries
EXAMINATIONS
18 April 2000 (am)
Subject 105 — Actuarial Mathematics 1
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1. Write your surname in full, the initials of your other names and your
Candidate’s Number on the front of the answer booklet.
2. Mark allocations are shown in brackets.
3. Attempt all 17 questions, beginning your answer to each question on a

separate sheet.
Graph paper is not required for this paper.
AT THE END OF THE EXAMINATION
Hand in BOTH your answer booklet and this question paper.
In addition to this paper you should have available

Actuarial Tables and an electronic calculator.
 Faculty of Actuaries
105—A2000  Institute of Actuaries
1 In the context of a pension scheme, explain the term “prospective service
benefit” and state one example. [2]
2 In a select mortality investigation, θx,r corresponds to the number of deaths

aged x next birthday at entry with duration r at the policy anniversary
following death. θx,r divided by the appropriate central exposed to risk gives
an estimate of µ[y]+t .
Derive the values of y and t to which this estimate applies. State clearly any
assumptions used. [2]
3 Mortality levels for a certain country have been studied at national and
regional level. Explain the circumstances under which a particular region
may have an Area Comparability Factor of 0.5. [2]
4 A 25 year annual premium endowment assurance policy was sold to a life

aged 40 exact at outset. Death benefits are payable at the end of the year of
death. Calculate the Zillmerised net premium reserve at the end of the tenth
year per unit sum assured.
Basis: Mortality: A1967–70 Select

Interest: 3% per annum
Initial expense: 2.5% of the sum assured [3]
5 A life insurance company sells an annual premium whole life assurance policy
with benefits payable at the end of the year of death. Expenses are incurred at
the start of each year, and claim expenses are nil.
(a) Write down a recursive relationship between the gross premium

reserves at successive durations, calculated on the premium basis.
Define all symbols used.
(b) Explain the meaning of this formula. [3]
6 1
Calculate A30: 30:30 using A1967–70 mortality and interest of 4% per annum. [3]
105—2
7 A pension scheme provides a pension of 1
45
of final pensionable salary for each
year of service, with a maximum of 2
3
of final pensionable salary, upon
retirement at age 65.
Final pensionable salary is defined as average annual salary over the 3 years
immediately preceding retirement.
A member is now aged exactly 47 and has 14 years of past service. He earned
£40,000 in the previous 12 months.
Calculate the expected present value now of this member’s total pension on
retirement, using the symbols defined in, and assumptions underlying, the
Formulae and Tables for Actuarial Examinations. [3]
8 The random variables Tx and Ty represent the exact future lifetimes of two
lives aged x and y respectively.
Let the random variable g(T) take the following values:
R|a if max{ Tx ,T y } ≤ n
g(T) = S|a
n
if max{ Tx ,T y } > n
T max{ Tx ,T y }
(i) Describe the benefit which has present value equal to g(T). [2]
(ii) Express E[g(T)] as concisely as possible in the form of an annuity

function. [1]
[Total 3]
9 Define the term “asset share” in the context of a with-profit policy. [3]
10 The number of people sick with a new disease is expected to increase

according to the logistic model. The initial number sick is 100,000 and it is
believed that the number sick with the disease will never exceed 250,000. At
the outset, sickness is assumed to grow at 5% per annum.
Calculate the number of people who are sick after exactly 10 years. [3]
11 A multiple decrement table is subject to two forces of decrement α and β.

Under the assumption of a uniform distribution of the independent
decrements over each year of age, ( aq )αx = 0.2 and ( aq )βx = 0.05.
Calculate q αx and q βx . [3]
105—3 PLEASE TURN OVER

12 An insurer sells combined death and sickness policies to healthy lives aged 35.
The policies, which are for a term of 30 years, pay a lump sum of £20,000
immediately on death, with an additional £10,000 if the deceased is sick at the
time of death. There is also a benefit of £3,000 per annum payable
continuously to sick policyholders. There is no waiting period before benefits
are payable. Annual premiums of £500 are payable continuously by healthy
policyholders.
The mortality and sickness of the policyholders are described by the following
multiple state model, in which the forces of transition depend on age.
σx
H = healthy S = sick
ρx
µx νx
D = dead
p xgh,t is defined as the probability that a life aged x who is in state g(g = H, S
or D) is in state h at age x + t (t ≥ 0 and h = H, S or D). The force of interest is
δ.
Express in integral form, using the probabilities and the various forces of
transition, the expected present value of one such policy at its commencement.
[4]
13 A pension scheme provides the following benefit to the spouse of a member,

following the death of the member in retirement:
A pension of £10,000 per annum payable during the lifetime of the spouse, but
ceasing 30 years after the death of the member if that is earlier. All payments
are made on the anniversary of the member’s retirement.
Calculate the expected present value of the spouse’s benefit in the case of a
female member retiring now on her 60th birthday, who has a husband aged
exactly 64.
Basis: a(55) Ultimate mortality at 8% per annum interest [8]
105—4
14 (i) Discuss the suitability of the crude death rate, the standardised
mortality rate and the standardised mortality ratio for comparing
(a) the mortality, at different times, of the population of a given

country
(b) the mortality, at a certain time, of two different occupational

groups in the same population [6]
(ii) The following table gives a summary of mortality for one of the
occupational groups and for the country as a whole.
Occupation A Whole Country

Exposed Exposed
Age group to risk Deaths to risk Deaths
20–34 15,000 52 960,000 3,100
35–49 12,000 74 1,400,000 7,500
50–64 10,000 109 740,000 7,100
37,000 235 3,100,000 17,700
Calculate the crude death rate, the standardised mortality rate and the
standardised mortality ratio for Occupation A. [4]
[Total 10]
15 An insurer issues 15 year term assurance policies to lives aged exactly 50 who
have provided satisfactory answers on a basic medical questionnaire. The sum
assured of £100,000 is payable at the end of the year of death during the policy
term. The policy includes an option at the end of the term which allows
policyholders to convert their policy to a whole life policy for the same sum
assured (payable at the end of the year of death). The premiums payable for
this whole life policy are the office’s standard premium rates, irrespective of
the health of the policyholder effecting the option.
The insurer calculates annual premiums for all products using A1967–70
Select mortality and 4% per annum interest, with an expense allowance of 5%
of all premiums.
(i) Describe:
(a) the North American method and
(b) the conventional method
for pricing mortality options. [5]
(ii) Using the conventional method calculate the extra annual premium the
insurer should charge above that for a term assurance policy with no
option. [5]
(iii) Without performing any further calculations, describe what other

considerations would arise if the option were such that the policy could
be converted on the 10th or the 15th anniversary. [3]
[Total 13]

16 A life insurance company issues a 4 year unit-linked policy with a level
premium of £1,000 payable annually in advance to a life aged exactly 61. The
death benefit at the end of the year of death is £4,000, or the bid value of the
units if greater. The maturity value is the bid value of the units.
95% of each premium is invested in units at the offer price. The bid price is
95% of the offer price. Premiums payable in the first two years are invested in
capital units which are subject to a management charge of 6% per annum.
Subsequent premiums are invested in accumulation units for which the
management charge is 1% per annum. Management charges are deducted at
the end of each year from the bid value of units before benefits are paid.
Capital units are actuarially funded using factors of A61+t :4 −t calculated using
A1967–70 Ultimate with 5% per annum interest for t = 0, 1, 2 and 3.
The company uses the following assumptions to profit test this contract:
Rate of interest on unit investments: 8% per annum

Rate of interest on sterling fund: 4% per annum
Mortality: A1967–70 Ultimate
Initial expenses: £100 plus 20% of the first premium
Renewal expenses: £20 on the first policy anniversary, and
increasing with inflation at 5% per
annum on each subsequent anniversary
(i) Using a risk discount rate of 12% per annum calculate the expected net
present value of the profit on this contract. [12]
(ii) Without performing any further calculations, state with reasons

whether your answer in (i) would be higher or lower for each of the
following, if
(a) the risk discount rate were 10% per annum

(b) the policyholder were aged 50 exactly
(c) capital units were actuarially funded at 4% per annum [5]
[Total 17]
105—6
17 A man aged exactly 30 effected a 35 year with profit endowment assurance for
a sum assured of £50,000. Level annual premiums are payable throughout the
policy term, ceasing on earlier death. The sum assured, with attaching
bonuses, is payable at the end of the year of death, or on maturity. Compound
reversionary bonuses vest at the end of each policy year.
(i) Show that the premium (to the nearest £1) is £990 per annum using the
following basis:

Expenses: Initial: £250 plus 60% of the annual premium
Renewal: 2.5% of second and subsequent premiums
Bonuses: 1.923% per annum [7]
(ii) The random variables Tx and Kx represent the exact future lifetime and
the curtate future lifetime of a life aged x, respectively. Using Tx , Kx or
both, express, in stochastic form, the gross future loss random variable
for this policy at duration t, where t is an integer and 0 < t < 35. Use
those elements of the basis set out in part (i) as needed. Assume bonus
declarations have been in line with the original bonus loadings. [3]
(iii) Immediately before the 11th premium is due, and just after the 10th
bonus has brought the sum assured plus accumulated bonuses to
£60,000, the policyholder wishes to convert the policy to a non-profit
whole life policy, with premiums of an unchanged amount payable until
death.
Using the mortality and interest elements of the premium basis set out
in part (i), and allowing for renewal expenses of 2.5% of all future
premiums as well as an alteration expense of £100, calculate the revised
sum assured. [6]
(iv) State one other consideration, if any, that the office should take into
account before completing the alteration in (iii), and explain why they
should do so. [2]
[Total 18]
105—7
EXAMINATIONS
April 2000
EXAMINERS’ REPORT
ã Faculty of Actuaries
ã Institute of Actuaries
Subject 105 (Actuarial Mathematics 1) — April 2000 — Examiners’ Report
1 “Prospective service benefit” means a benefit not dependent on either past or

future service explicitly, although it may depend on total expected service.
Examples include — lump sum death benefit of 4 ´ salary or spouse’s pension

n
death in service of ´ final salary where n is based on deceased member’s
120
total potential service to NPA, including any past service.
Many candidates confused “prospective” with “future”.
2 x next birthday at entry Þ x - ½ on average at entry assuming birthdays

uniformly distributed over policy year.
r at policy anniversary after death means exact duration r - 1 at the

anniversary before death (the start of the policy year rate interval for
duration) and hence r - ½ mid-year when the force of mortality is
estimated. No assumptions are necessary.
The force estimated is m[x-½]+r-½ , so y = x - ½, t = r - ½.
3 If its age/sex profile is such that if it experienced the same age/sex specific
mortality rates as the country, then its crude death rate would be twice that
of the country, i.e. the region has a much older age structure (and/or higher male
proportion) than the country.
[ x ]+t:n -t
a a[ x ]+t:n-t
4 tV
Zillmer
= 1- -I
[ x ]:n
a a[ x ]:n
a50:15 æ a ö
50:15
Here 10V = 1 - - (.025) ç ÷
[40]:25
a ça ÷
è [40]:25 ø
11.671 æ 11.671 ö
= 1- - (.025) ç ÷
17.180 è 17.180 ø
= 0.30368
Page 2
5 (a) ( tV ¢ + GP - et) (1 + i) = qx+t (S) + px+t (t+1V ¢)
where tV ¢ = gross premium reserve @ time t
GP = office premium
et = expenses incurred at time t
i = interest rate in premium/valuation basis

S = Sum Assured
px+t (qx+t) probability life aged x + t survives (dies within) one year on
premium/valuation mortality basis.
(b) Income (opening reserve plus interest on excess of premium over expense,
and reserve) equals outgo (death claims and closing reserve for survivors)
if assumptions are borne out.
æ D ö
6 1
A30:30:30 1
= ½A 30:30
}
: 30
= ½ ç A30:30 - 60:60 A60:60 ÷
è D30:30 ø
é æ .04 ö D ì æ .04 ö üù
= ½ ê1 - ç ÷ 30:30 - 60:60
a í1 - ç ÷ 60:60 ýú
a
ë è 1.04 ø D30:30 î è 1.04 ø þû
é æ .04 ö 2487.2117 ì æ .04 ö üù

= ½ ê1 - ç ÷ (19.701) - í1 - ç ÷ (9.943) ýú
ë è 1.04 ø 10236.789 î è 1.04 ø þû
= ½[1 - .75773 - (.24297)(.61758)]
= .0461
7 Future service = 18 + 14 past Þ total = 32 > max of 30.
æ2ö s zC ra
Þ Value of benefit = ç ÷ (40,000) 47 s 65
è3ø s46 D47
æ2ö æ 4.28 ö æ 35846 ö

= ç ÷ (40,000) ç ÷ç ÷ = £62,033
è3ø è 4.18 ø è 15778 ø
Most candidates allowed for retirement at any age, not just 65, and many failed to notice
that service exceeded 30 years so the maximum of 2/3rds applied.
Page 3
8 (i) A continuous annuity of £1 p.a. payable for a minimum of n years and

continuing thereafter until the death of the survivor of x and y.
(ii) E[g(T)] = a .
xy:n
Rather than defining asset share, some candidates discussed bonuses and policy payouts.
9 The asset share for a with-profit policy is the accumulated value of premiums
less deductions plus an allocation of profits from non-profit business. The
accumulation is at actual earned rates of return.
The deductions include expenses, cost of benefits, tax, transfers to shareholders,

cost of capital and contribution to free assets.
Rather than defining asset share, some candidates discussed bonuses and policy payouts.
-1
é Kù é r ù
10 In logistic model P(t) = êCe -rt + ú or ê -rt ú
ë rû ë C re + K û
H .05
As t ® ¥ P(t) ® ÞK=
K 250,000
-1
é 1 ù
P(0) = êC + = 100,000 Þ C = 0.000006
ë 250,000 úû
-1
é 1 ù
Þ P(10) = ê(.000006) e -(.05)(10) +
ë 250,000 úû
= 130,904
Only a minority of candidates seemed familiar with the logistic model.
11 Under UDD in single decrement table
( aq )=x = qx= (1 - ½qx> ) = qx= - ½qx= qx> = 0.2
( aq )>x = qx> (1 - ½qx= ) = qx> - ½qx= qx> = 0.05
Þ qx= - qx> = 0.15 Þ qx= = qx> + 0.15
(qx> + 0.15) - ½(qx> + 0.15) qx> = 0.2
Þ - ½(qx> )2 + .925qx> + 0.15 = 0.2
Page 4
OR ( qx> )2 - 1.85qx> + 0.1 = 0
1.85 m 1.852 - 0.4

Roots are Þ 0.05573 (and q > 1 is invalid)
2
qx> = 0.05573 and qx= = 0.20573

Alternatively, qx= = ( aq )=x ¸ 1 - ½qx> and qx> = ( aq )>x ¸ 1 - ½qx=
Using iteration, and taking starting values in denominators of qx= » ( aq )=x etc.
1st iteration qx= = 0.2 ÷ [1 - (.5)(.05)] = .205128
qx> = 0.05 ÷ [1 - (.5)(.2)] = .055556
Similarly, 2nd iteration qx= = .20571, qx> = .05571
3rd iteration qx= = .20573, qx> = .05573
4th iteration qx= = .20573, qx> = .05573
Hence qx= = .20573, qx> = .05573
A large number of candidates used formulae appropriate when decrements are uniform in
the multiple decrement table, but the question specified that independent decrements were
uniform in the single decrement tables.
12 EPV = 500 ò 30
0 e
-dt hh
p35,t dt (premiums)
-20,000 ò 30
0 e
-dt hh
p35,t m35 +t dt (death from healthy)
-30,000 ò 30
0 e
-dt hs
p35,t n35+t dt (death from sick)
-3,000 ò 30
0 e
-dt hs
p35,t dt (sickness income)
ì 30 ¥
m tü
13 EPV f
= 10,000 í S (1 - t p60 m t
) t p64 f
v + S ( t -30 p60 f
- t p60 ) t p64 v ý
ît =1 t =31 þ
ì 30 m t ¥
f m t
¥
f m tü
= 10,000 í S t p64 v + S t - 30 p60 t p64 v - S t p60 t p64 v ý
ît =1 t = 31 t =1 þ
m
{
= 10,000 a64:30 + 30
m 30
p64 f m
v a60:94 f m
- a60:64 }
Page 5
m D94 æ 16.4 ö
a64:30 = a64 - a94 = 7.616 - ç 5844.0 ÷ (1.707) = 7.611
D64 è ø
m D94 16.4
30 p64 v30 = = = 0.002806297
D64 5844.0
f :m f m
a60:94 = 1.666 a60:64 = 6.854
Þ EPV = 10,000{7.611 + (.002806297)(1.666) - 6.854} = £7,617
Very few candidates provided a satisfactory answer. Many did not attempt to deal with
the term aspect of the question, and most of those who did assumed the annuity ended 30
years after retirement rather than 30 years after the pensioner’s death.
14 (i) Crude death rate is heavily influenced by mortality at older ages
(a) OK if population structures by age and sex are reasonably

stable.
Therefore beware large scale emigration/immigration. Easy and
practical.
(b) Not suitable — age and sex distributions in occupational groups

likely to vary significantly.
Standardised Mortality Rate
Again influenced by mortality at older ages.
(a) OK to use but need age specific mortality rates at each time
point.
Changing population structure has no effect.
(b) Copes well with age/sex variations provided age specific rates are
available for occupational groups.
But use of a fixed age structure may be unrepresentative of given

occupation.
Standardised Mortality Ratio
Heavily influenced by relative mortality at older ages.
(a) Fine but ensure standard rates used are same each time.
Page 6
(b) Good except for possible problems gathering the data on age
distributions.
Use of occupational age structure maintains relevance.
(ii) Occupational A
Crude Rate
= 235 / 37,000 = 0.00635
Standardised Mortality Rate
52
= (960,000 ´
15000
74
+ 1,400,000 ´
12000
109
+ 740,000 ´ ) ¸ 3,100,000
10,000
= (3,328 + 8,633.33 + 8,066) ÷ 3,100,000 = 0.00646
Standardised Mortality Ratio
ì 3,100 ü
ï15,000 ´ 960,000 ï
ï ï
ï 7,500 ï ì48.44 + 64.29 ü
= 235 ¸ í+ 12,000 ´ ý = 235 ¸ í ý
ï 1,400,000 ï î+ 95.95 þ
ï 7,100 ï
ï+ 10,000 ´ ï
î 740,000 þ
= 235 ¸ 208.68
= 1.126
Answered quite well in general, although some students tended to describe the various
measures in general rather than relate them to the specific situations described.
15 (i) (a) North American Method
Relies on double decrement table with explicit proportions who

choose to exercise option and a special mortality table for those
people post option.
While theoretically accurate, it is often difficult to obtain sufficient
data to estimate experience.
Page 7
(b) Conventional Method
Assumes all eligible lives actually take up option, and that they
are subject to Ultimate mortality as opposed to Select if normal
underwriting carried out. If there are many option dates etc., then
the most costly from the insurers point of view is assumed.
æ 1 ö
(ii) Insurer charges ç ÷ (P[65]) (100,000) per annum for whole life policy
è .95 ø
i.e. (.05254)(100,000) ¸ .95 = £5,530.53 p.a.
At option date (age 65), the value of benefits provided is
100,000 A65 = (100,000)(.58705) = £58,705
The insurers net liability at option date present value of benefits – (present value of
premiums less expenses)
= 100,000 A65 – (.95)(5,530.53) a65
= 58,705 – (.95)(5,530.53)(10.737)
= 58,705 – 56,412.20 = £2,292.80
Extra premium, P’, spread over term assurance policy term, is from:-
D65
.95P ¢ a[50]:15 = 2,292.80
D[50]
æ 2,144.1713 ö
Þ P ¢ = (2,292.80) ç ÷ ¸ (.95) (11.028)
è 4,581.3224 ø
Þ P¢ = £102.43 per annum
(iii) The office needs to decide which option is costlier, not just in the value of
the option benefit, but its impact on the overall premium required over
the period to the option exercise date.
In this case, it needs to compare the above option cost in premium terms
plus the 15 term assurance premium to the similarly calculated extra
premium for the 10 year option combined with a 10 year term insurance
premium.
It should then charge the higher combined premium, thereby having
option cost at any date more than covered.
Part (i) was well answered, but (ii) and (iii) were very poorly answered. Many candidates
treated the contract as a whole life from the start making the option cost the difference
between a term assurance and a whole life policy for the life aged 50.
Page 8
16 (i) q61 = .016 013 56 p61 = .983 986 44 0 p61 = 1.0 5%

A61:4 = .82703
q62 = .017 749 72 p62 = .982 250 28 1 p61 = 0.983 986 A62:3 = .86624
q63 = .019 654 64 p63 = .980 345 36 2 p61 = 0.966 521 A63:2 = .90792
q64 = .021 743 10 p64 = .978 256 90 3 p61 = 0.947 524 A64:1 = .95238
Capital unit fund — fully funded

Y/e fund Management Fund
Year Cost of alloc. Fund b/f after 8% growth Charge 6% c/f
1 902.50 – 974.70 58.48 916.22

2 902.50 916.22 1,964.21 117.85 1,846.36
3 – 1,846.36 1,994.07 119.64 1,874.43
4 – 1,874.43 2,024.38 121.46 1,902.92
Capital unit fund — a-funded

Available Needed at Extra death Management
Year Cost of alloc. Fund b/f @ y/e after 8% year end cost charge
1 746.39 – 806.10 793.67 1.96 10.47

2 781.78 793.67 1,701.49 1,676.35 3.02 22.12
3 – 1,676.35 1,810.46 1,785.17 1.75 23.54
4 – 1,785.17 1,927.98 1,902.92 – 25.06
Premium unit fund
1%
Fund Management Fund
Year Cost of alloc. Fund b/f @ year end charge c/f
3 902.50 – 974.70 9.75 964.95

4 902.50 964.95 2,016.85 20.17 1,996.68
Death cost (using full Cap. Units) Yr 1 Þ q61 (4000 – 916.22) = 49.38
Yr 2 Þ q62 (4000 - 1846.36) = 38.23
Yr 3 Þ q63 (4000 - 1874.43 - 964.95) = 22.81
Yr 4 Þ q64 (4000 - 1902.92 - 1996.68) = 2.18
Page 9
Sterling fund
(4%)
Premium less Sterling Death Management Profit Profit
Year cost of alloc. Expense interest cost charge vector signature
1 253.61 300.00 (1.86) 49.38 10.47 -87.16 -87.16

2 218.22 20.00 7.93 38.23 22.12 190.04 187.00
3 97.50 21.00 3.06 22.81 33.29 90.04 87.03
4 97.50 22.05 3.02 2.18 45.23 121.52 115.14
NPV = -87.16v + 187v2 + 87.03v3 + 115.14v4 = 206.37
Alternative approach whereby entire death cost is charged to sterling fund is also valid,
providing a-funded capital unit management charge is correspondingly increased.
(ii) (a) Given the shape of the cash flows, with the positives after the
negatives, a discount rate of 10% would mean larger NPV.
(b) Death cost would reduce, probability of being in force and hence
premium income would increase, causing NPV to increase.
A-funding factors would also decrease, accelerating the cash flows.
Given risk discount rate (12%) > sterling fund rate this will
increase NPV.
(c) At 4%, factors will be bigger, unit reserves increase and profit is
deferred. Because risk discount rate exceeds sterling fund rate,
NPV decreases.
Generally well answered, although candidates often failed to give reasons for their correct
conclusions in (ii).
æ 1 4% 4%
ö
17 (i) 6%
Pa30:35 = 50,000 ç
ç 1.01923
1
A30:35 1
+ A30:35 ÷ + 250
÷
è ø
6%
30:35
+ .025Pa + .575P
Because bonuses vest at year end, maturities get an extra bonus

compared to deaths in last year, and so the death benefit function is
divided by (1 + bonus loading).
6%
P .975a30:35 ì
ï
- .575 = 250 + 50,000 í
1 æ
ç A30:35 -
î1.01923 è
ï
D65 ö D65 ü
÷+
ï
ý
D30 ø D30 ï
þ
6%
30:35
a = 15.019
4%
A30:35 = .27483
Page 10
4%
D65 2144.1713
4%
= = .20551
D30 10433.31
Þ P(14.0685) = 250 + 50,000{.06801 + .20551} Þ P = 989.87 = £990 p.a.
(ii) Gross future loss = PV future outgo - PV future income
= PV future benefit payment + PV future expenses

- PV future premiums
= G(K30+t) + (.025)(990) amin[ K

30 + t +1, 35 -t ]
- (990) amin[ K
30 + t +1, 35 -t ]
where G(K30+t) = ïì50,000 (1.01923)t + K30 +t v.06

K 30 + t +1
K 30 +t < 35 - t
í 35 35 -t
î50,000 (1.01923) v.06
ï K 30 +t ³ 35 - t
(iii) Reserve before alteration = reserve after alteration + cost of alteration
Before
ì
ï 1 æ 4% D65 ö
4%
D65 ü
ï 6%
10V = 60,000 í A
ç 40:25 - ÷ + 40:25
ý - (.975)(990)( a )
ï1.01923 è D40 ø D40 ï
î þ
ì 1 ü
= 60,000 í (.40005 - .30690) + .30690ý - (.975)(990)(13.081)
î1.01923 þ
= 23,897.55 - 12,626.44 = 11,271.11 say £11,271
After
6% 6%
10V = x A40 40
- (.975)(990) a
= x (.15807) - (.975)(990)(14.874) = (.15807)(x) - 14,357
Þ 11,271 = (.15807)(x) - 14,357 + 100 Þ x = 161,498 say £161,500
(iv) The amount at risk is immediately significantly increased (by £100,000)

and the term for which there is a death strain has been extended.
There is a grave risk of adverse selection against the office unless it
underwrites the alteration as effectively a new business case. A simple
declaration of health will not suffice in this case given the size of the
change of the immediate risk.
Parts (i), (iii) and if attempted (iv) were well answered although most students missed the
different bonus treatment needed for death benefits compared with the maturity benefit.
Few candidates seemed familiar with the concept of the gross future loss as a random
variable and answers to part (ii) were weak.
Page 11
EXAMINATIONS
19 September 2000 (am)

separate sheet.
Graph paper is not required for this paper
In addition to this paper you should have available, Actuarial

Tables and an electronic calculator.
105—S2000 ã Institute of Actuaries
1 Two lives, each aged x, are subject to the same mortality table. According to the
mortality table and a certain rate of interest, Ax = 0.5 and Ax x = 0.8.
Calculate Ax 2x , using the same mortality table and interest rate. [2]
2 The following data are available in relation to a particular country and one of its
regions:
Region A Country
Population at Deaths in Population at Deaths in
Age group 30 June 2000 2000 30 June 2000 2000
(000s) (000s)
0–39 645 350 13,580 8,347
40–59 450 2,295 8,100 45,360
60+ 385 27,500 6,290 489,860
Calculate the standardised mortality ratio for region A by reference to the

country as a whole. [2]
3 (i) A life insurance policy provides a benefit of £10,000 payable immediately

on the death of a life (x), if (x) dies after a life (y). Express in integral form
the expected present value of the benefit under this policy. [1]
(ii) Set out, giving a reason, the most appropriate annuity factor to value
annual premiums payable under the policy. [1]
[Total 2]
4 A healthy life aged exactly 35 has a policy providing an income benefit of £50 per
week payable during sickness. The benefit is not payable beyond age 60. There
is no deferred or waiting period.
Calculate the present value of this benefit.
Basis: Mortality: English Life Table No. 12-Males

Sickness: Manchester Unity Sickness Experience 1893/97
Occupation Group AHJ
Interest: 4% per annum [3]
5 An annuity of 1 is payable annually in arrears while at least one of two lives, (x)
and (y), is alive.
Derive an expression in terms of joint-life and single life functions for the
variance of the present value of the annuity. [3]
6 Describe three types of bonus that may be given to a with profits contract. [3]
105—2
7 In the context of a life insurance contract, explain how an asset share may be
built up using a recursive formula. [3]
8 (i) On 1 January 1990 a life insurance company issued a 20-year annual

premium without profits endowment assurance policy to a life then aged
exactly 40, which is still in force. The sum assured of £100,000 is payable
at the end of the year of death within the term of the policy, or on
survival. The company values the policy using a modified net premium
method, with a Zillmer adjustment.
Calculate the reserve for the policy on 31 December 1999.

Zillmer adjustment: 2% of the sum assured [3]
(ii) Without carrying out any further calculations, explain how the value of
the policy would differ if the company used a Zillmer adjustment of 1% of
the sum assured, with the same mortality and interest assumptions. [2]
[Total 5]
9 A life insurance company issues a special reversionary annuity contract. Under

the contract an annuity of £10,000 per annum is payable monthly for life, to a
female life now aged exactly 60, on the death of a male life now aged exactly 65,
provided the male life dies within 10 years of the start date of the policy.
Payments commence on the first monthly policy anniversary after the date of
death.
Calculate the single premium required for the contract.
Basis: Mortality: a(55) Ultimate mortality, male or female as appropriate

Expenses: none [5]
10 A pension scheme provides an ill-health retirement pension of 1/60 of Final

Pensionable Salary for each year of company service, with fractions of a year to
count proportionately, subject to a maximum pension of 40/60 of Final
Pensionable Salary. Retirement due to ill-health may take place at any age
before age 65. Final Pensionable Salary is defined as the average annual salary
over the three-year period preceding retirement.
Derive commutation functions to value the ill-health retirement pension for a

member aged exactly 25, who has completed exactly 5 years company service to
date. Define carefully all the symbols that you use. [7]
11 Describe the component method of population projection used for British Official
Projections, stating carefully any assumptions that you make and defining all the
symbols that you use. [7]

12 A life insurance company issues only single premium without profit term
assurance policies.
The premium is to be calculated for a special 3-year term assurance for lives aged
exactly 60 where the basic sum assured is £100,000, payable at the end of the
year of death.
This special policy carries a “guaranteed insurability” option that may be selected
at the outset of the 3-year policy in return for the payment of an additional single
premium.
This option provides a guarantee to the policyholder that a further £100,000 of

sum assured may be purchased, at a subsequent policy anniversary, on normal
premium rates and without evidence of health.
The further sum assured purchased will not itself carry any further options, and
will expire at the end of the 3-year term of the original policy.
A policyholder who has paid the additional single premium can subsequently
decide whether or not to effect the increase in sum assured and then at which
policy anniversary — the first or second, but not both.
The company uses the “North American experience” method for pricing the
option.
Calculate the additional single premium payable at outset for a policyholder

choosing the option.
Basis: Mortality: A1967–70 Select, except in the case of policyholders who

decide to exercise their option to increase the sum assured.
For these policyholders, the mortality basis assumed to
apply, from the point of increase in sum assured, is 150% of
A1967–70 Ultimate.
Interest: 5.5% per annum
Proportion of policyholders at the first anniversary who decide to

increase their sum assured at that point: 20%
Proportion of policyholders at the second anniversary who decide to

increase their sum assured at that point: 20%
Expenses: none [7]
105—4
13 A life insurance company uses the following 3-state model, to estimate the profit
in respect of a 2-year combined death benefit and sickness policy issued to a
healthy policyholder aged exactly 55 at inception.
Healthy (H) Sick (S)
Dead (D)
In return for a single premium of £6,000 payable at the outset the company will
pay the following benefits:
£16,000 if the policyholder dies within 2 years, payable at the end of the year of
death;
£8,000 at the end of each of the 2 years if the policyholder is sick at those times.
Let St represent the state of the policyholder at age 55 + t, so that S0 = H and for
t = 1 and 2, St = H, S or D.
The company uses transition probabilities defined as follows:
ij
p55 +t = P(St+1 = jSt = i)
For t = 0 and 1 the transition probabilities are:
HD SD SH HS
p55 +t = 0.08 p55 +t = 0.15 p55 +t = 0.75 p55 +t = 0.12
The transitions in the multiple state model are the only sources of randomness.
(i) One possible outcome for this policy is that the policyholder is healthy at
times 0, 1 and 2. List all the possible outcomes and the associated cash
flows. [3]
(ii) Calculate the probability that each outcome occurs. [5]
(iii) Assuming a rate of interest of 8% per annum, calculate the net present
value at time 0 of the profit for each outcome. [2]
(iv) Calculate the mean and standard deviation of the net present value of the
profit at time 0 for the policy. [5]
[Total 15]

14 On 1 September 1992, a life insurance company issued a whole life with profits
policy to a life then aged exactly 45. The basic sum assured was £100,000. The
sum assured and attaching bonuses are payable immediately on death. Level
monthly premiums are payable in advance to age 85 or until earlier death. The
company calculated the premium on the following basis:
Mortality: A1967–70 Select

Bonus loading: 0.97087% per annum compound, vesting at the
beginning of each policy year
Expenses: initial: 50% of the first year’s premiums, incurred at the outset
renewal: 5% of the second and each subsequent year’s premiums,
incurred at the beginning of the respective policy years.
(i) Show that the monthly premium is £229, to the nearest £. [7]
(ii) Immediately before payment of the premium due on 1 September 2000, at

the request of the policyholder, the insurance company alters the policy to
a paid-up policy, with no future premiums payable. The sum assured
under the policy is reduced, with no further bonuses payable.
The insurance company calculates the reduced sum assured after

alteration by equating prospective gross premium policy reserves
immediately before and after alteration, allowing for an expense of
alteration of £100.
Bonuses have vested at the rate of 4% per annum compound at the

beginning of each policy year from the date of issue of the policy. The
company calculates prospective gross premium policy reserves for the
purpose of the alteration using the following assumptions:

Expenses: none
Allowance for future bonuses: none
Calculate the sum assured after alteration. [6]

[Total 13]
15 A life insurance company issues a 3-year unit-linked endowment assurance

contract to a male life aged exactly 62 under which level annual premiums of
£4,000 are payable in advance throughout the term of the policy or until earlier
death. 101% of each year’s premium is invested in units at the offer price.
The premium in the first year is used to buy capital units, with subsequent years’
premiums being used to buy accumulation units. There is a bid-offer spread in
unit values, with the bid price being 95% of the offer price.
The annual management charges are 5.25% on capital units and 1.25% on
accumulation units. Management charges are deducted at the end of each year,
before death, surrender or maturity benefits are paid.
105—6
On the death of the policyholder during the term of the policy, there is a benefit
payable at the end of the year of death of £10,000 or the bid value of the units
allocated to the policy, if greater. On maturity, the full bid value of the units is
payable.
A policyholder may surrender the policy only at the end of each year. On
surrender, the bid value of the accumulation units plus a proportion of the capital
units is payable. The proportion of the capital units payable on surrender is
determined by the year of surrender, as follows:
Year of surrender Proportion of capital

units paid out
1 0.85
2 0.90
3 1
The life insurance company uses the following assumptions in carrying out profit
tests of this contract:

Expenses: initial: £300
renewal: £60 at the start of each of the second and third policy
years
Unit fund growth rate: 9% per annum
Sterling fund interest rate: 4.5% per annum
Risk discount rate: 15% per annum
Surrender rates: 15% of all policies still in force at the end of each of
the first and second years
(i) The company holds unit reserves equal to the full bid value of the
accumulation units and a proportion A62+t:3−t (calculated at 4%), of the full
bid value of the capital units, calculated just after the payment of the
premium due at time t (t = 0, 1 and 2). The company holds no sterling
reserves.
Calculate the profit margin on the contract. [17]
(ii) Assume instead that the company holds unit reserves equal to the full bid
value of both the accumulation and capital units and that the company
also holds sterling reserves, at the start of each policy year, equal to 10%
of the annual premium. Calculate the revised profit margin on the
contract. [6]
[Total 23]
105—7
EXAMINATIONS
September 2000
EXAMINERS’ REPORT
Subject 105 (Actuarial Mathematics 1) — September 2000 — Examiners’ Report
1 Ax = 0.5
Axx = 0.8
2 A1xx = Axx
Axx2 = Ax − Axx1 = Ax − ½Axx = 0.5 − ½ × 0.8 = 0.1.
åE c
x ,tmx ,t
2
åE
x
The standardised mortality ratio (SMR) =
c s
x ,t mx ,t
x
350 + 2,295 + 27,500

=
æ 8,347 45,360 489,860 ö
ç 645 * + 450 * + 385 *
è 13,580 8,100 6,290 ÷ø
30,145
=
32,899.93
= 0.9163.
∞
3 (i) £ 10,000 ò vt (1 − t py ) t px µ x +t dt
0
where x = age of (x)
y = age of (y)
(ii) The premium should be payable as long as (x) is alive, while the benefit is
still payable. It does not matter whether (y) is alive. The most
appropriate annuity factor is, therefore:
(m )
x , where m denotes frequency of payment.
a
50( K 35 − K 60 )
4 Value =
D35
K35 13
= K 35 + K 35
13 / 13
+ K 35
26 / 26
+ K 35
52 / 52
+ K 35
104 / all
= 462592 + 143625 + 154161 + 179711 + 716291 = 1656380
K60 = 970852.7
D35 = 23986
Page 2
Value = £1429.02
Alternative
æ D60 ö
50 ç 69.056 − 129.405 ÷ , based on value of 1 p.w. all periods, whole of life
è D35 ø
= £1428.99
5 Required: Var( a K )
xy
Var( a K ) = Var( a

K +1
− 1)
xy xy
æ 1 − v K xy +1 ö
= Var ç ÷
ç d ÷
è ø
1 K +1
= 2
Var(v xy )
d
1 2
= ( Axy − Axy )2 )
d2
1 2
= ( Ax + 2 Ay − 2 Axy − ( Ax + A y − Axy )2 )
d2
where “2” denotes evaluation at rate of interest i2 + 2i. Other functions are
evaluated at rate of interest i.
6 The following are three types of guaranteed reversionary bonuses. The bonuses
are usually allocated annually in arrears, following a valuation.
Simple — the rate of bonus each year is a percentage of the initial basic sum
assured under a policy. The effect is that the sum assured increases linearly over
the term of the policy.
Compound — the rate of bonus each year is a percentage of the basic sum
assured and the bonuses previously added. The effect is that the sum assured
increases exponentially over the term of the policy.
Super compound — two compound bonus rates are declared each year. The first
rate (usually the lowest) is applied to the basic sum assured. The second rate is
applied to the bonuses previously added. The sum assured increases
exponentially over the term of the policy. The sum assured usually increases
more slowly than under a compound allocation in the earlier years and faster in
the later years.
Page 3
7 An asset share is evaluated for an individual policy or for a block of policies,

usually for non-unit linked policies.
The asset share is the accumulation of premiums less deductions associated with
the contract plus an allocation of profits on non-profit business, all accumulated
at the actual rate of return earned on investments. Deductions include all
expenditure associated with the contract or contracts.
The asset share may be built up recursively on a year-to-year basis. Initially, the
asset share is zero. Each year, the cash flows including premiums received,
deductions made to cover actual costs and provisions made to cover other
liabilities and provision for profits allocated to the policy or group of policies are
recorded. A suitable rate of return is used to accumulate the asset shares plus
premiums less deductions plus profit allocations to the year-end to determine the
asset share. The process is repeated for subsequent years.
æ a
ö a
8 (i) Reserve = 100,000 ç1 − 50:10
÷ − 2,000 50:10
ç a
÷ a

è [40]:20 ø [40]:20
a

50:10
= 8.207
a

[40]:20
= 13.772
Reserve = £39,216.24
(ii) Using a Zillmer adjustment has the effect of reducing the policy value.
Changing the Zillmer adjustment from 2% of the sum assured to 1% of the
sum assured has the effect of reducing the amount of the Zillmer
adjustment and hence increasing the policy value, as at 31 December
1999.
æ ö
9 Premium = 10,000 ç a65
(12)
60
(12) ÷
− 10 a65 60
ç ÷
è m f m f ø
a

(12)
65 60
= a60 − a60:65 = 10.996 − ½(7.753 + 7.335)
m f f f m
= 3.452
Page 4
m f æ ö
D75 l70
(12)
= ç a70 − a70:75 ÷
10 a65 60

D65 l60 ç f ÷
m f è f m ø
6809 780683
= × × (8.328 − ½(4.9 + 4.525))
17857 897001
= 0.381307 × 0.870326 × 3.616
= 1.200
∴ Premium = 10,000(3.452 − 1.2) = £22,520.
10 Define a service table:
lx+t = no. of members aged x + t last birthday
ix+t = no. of members who retire due to ill-health age x + t last birthday
sx+t / sx = ratio of earnings in the year of age x + t to x + t + 1 to the earnings in

the year of age x to x + 1
Define zx+t = 1
3
(sx−3 + sx−2 + sx−1); axi = value of annuity of 1 p.a. to an ill-health
retiree aged exactly x + t.
Let (AS) be the member’s expected salary earnings in the year of age 25 to 26.
Assume that ill-health retirements take place uniformly over the year of age.
Consider ill-health retirement between ages 25 + t and 25 + t + 1, t < 35.
The present value of the retirement benefits related to future service:
( t + ½) ( AS ) z25+t +½ v25+t +½ i25+t i

a =
(t + ½)( AS ) z C25
ia
+t
25 25+ t +½ s
60 s25 v l25 60 D25
z ia 25+t+½ i
where C25 + t = z25+t+½ v i25+t a25 +t +½
s
and D25 = s25 v25 l25
Similarly it may be shown that the present value of the benefits is, in total:
( AS )
é½z C25
ia
+ 1½z C26
ia
+ ... + 35z C60
ia
+ 35z C61
ia
+ ... + 35z C64
ia
ù
60 s D25 ë û
Page 5
( AS )
= é½z C25
ia
+ 1½z C26
ia
+ ... + 35½z C60
ia
+ 36½z C61
ia
+ ... + 39½z C64
ia
60 s D25 ë
(
− ½z C60
ia ia ù
+ ... 4½z C64 )û
=
60 s D25 ë
(
( AS ) é z ia z ia
M 25 + M 26 + ... z M 64
ia
− ) ( z ia
M 60 + z M 61
ia
+ ... + z M 64
ia
)ùû
64 − x
[where z M xia = Σ z
Cxia+t − ½z Cxia ]
t =0
( AS ) z ia z ia
= é R25 − R60 ù
60 s D25 ë û
64 − x
where z Rxia = Σ z
M xia+t
t =0
Similarly it may be shown that the present value of benefits related to past
service is:
5( AS ) z ia
M 25
60s D25
åC
30
z ia
where z M 25
ia
= 25+ t .
t =0
11 Px(n) = Survivors to n of Px−1(n −1) + migrants during (n − 1, n) who

survived to be age x at n (net migrants are considered, i.e.
migrants less emigrants).
P0(n) = Births during (n − 1, n) + migrants during (n − 1, n) who survived

to be age 0 at n
where Px(n) is the population age x last birthday at n, where n refers to

mid-year n.
Let
B(n) = births during (n − 1, n)
Mx(n) = net migrants during (n − 1, n) who survive to be age x last

birthday at n
qx−½(n − 1) = probability that a life aged last birthday x − 1 at n − 1 dies in

(n − 1, n), assuming those aged x − 1 last birthday at n − 1 have
birthdays uniformly distributed over the calendar year.
½q0(n − 1) = probability that a life born in (n − 1, n) dies in (n − 1, n).
Page 6
Then we have:
Px(n) = Px−1(n − 1) × (1 − qx−½(n − 1)) + Mx(n)
P0(n) = B(n) × (1 − ½q0(n − 1)) + M0(n)
Projections are carried out separately for each sex to give values P0(n),
P1(n), ..., Px(n), ....
B(n) and Mn(x) are determined using separate models. Total births in (n − 1, n),
B(n), are projected using
{ }
B(n) = Σ½ Pxf ( n − 1) + Pxf (n ) fx ( n)
where Pxf ( n) is the number of females aged x last birthday at n.

{ }
½ Pxf ( x − 1) + Pxf (n) gives the average female population aged x last birthday
over the year (n − 1, n).
fx(n) is the fertility rate over (n − 1, n) for women aged x last birthday at the date
of birth.
The summation is taken over all ages where fx(n) > 0.
The sex ratio at birth has been estimated empirically to be 1.06 : 1 (males :
females). This ratio is used to obtain male and female births, as follows:
1.06 B( n) B( n)
Bm(n) = Bf(n) =
2.06 2.06
Migration numbers are estimated directly from the International Passenger

Survey.
Page 7
12 Construct multiple decrement tables
For those not exercising the option:
Age No. alive No. of

deaths
60 100,000 669.90
61 79,464.08 770.94
62 62,954.51 1,117.42
For those exercising the option:
Age No. alive No. of

deaths
61 19,866.02 477.19
62 35,147.46 935.79
Premiums payable:
P0 = 100,000[0.0066990v + (1 − .006699) * 0.00970168v 2 +

(1 − .006699) * (1 − .00970168) * 0.01774972v 3 ] = 2,987.67
P1 = 100,000[0.00970168v + (1 − .00970168) * 0.01774972v2] = 2,498.85
P2 = 100,000[0.01774972v] = 1682.44
where P0 is the premium payable at the outset, P1 is the premium payable at the
first anniversary for additional cover and P2 is the premium payable at the
second anniversary for additional cover.
Cost of benefits =
100,000
100,000
[
669.90v + (770.94 + 2 * 477.19)v 2 + (1,117.42 + 2 * 935.79)v 3 ]
=4,730.567
Value of premiums=
1
[2987.67*100,000+2,498.85*19,866.02v+1,682.44*15,738.63v2]
100,000
=3,696.12.
Option premium=
4,730.57-3,696.12 = £1,034.45
Page 8
13 (i)
Outcome Cashflow
(1) HHH 6,000, 0, 0

(2) HHS 6,000, 0, −8,000
(3) HHD 6,000, 0, −16,000
(4) HSH 6,000, 8,000, 0
(5) HSS 6,000, −8,000, −8,000
(6) HSD 6,000, −8,000, −16,000
(7) HD 6,000, −16,000
“−” indicates cashflow to policyholder
(ii) Complete the set of transition probabilities:
HH SS
p55 +t = 0.8, p55 +t = 0.1
The probability that each outcome occurs is:
Outcome Probability
(1) 0.64
(2) 0.096
(3) 0.064
(4) 0.09
(5) 0.012
(6) 0.018
(7) 0.08
1.000
(iii) The net present value of each outcome is:
Outcome NPV of Profit
(1) 6,000
(2) −858.711
(3) −7,717.421
(4) −1,407.407
(5) −8,266.118
(6) −15,124.829
(7) −8,814.815
Page 9
(iv) Mean = Σ NPV × liability

= 6000 × 0.64 − 858.711 × 0.096 + ...
= £2,060.36
Variance = ΣNPV2 × Probability − (Mean)2

= 34,009,436.35
Standard deviation = £5,831.76
14 (i) 1.04 / 1.0097087 = 1.03
Þ death benefits evaluated at 3% p.a.

Value of death benefits = 100,000 A[45]
= 100,000 × 1.04½ × A[45]

3%
= 100,000 × 1.04½ × 0.42060
= 42,892.952
(12) 4%
Value of premiums = Pa[45]:40
N[45] − N 85 11 æ 1 − D85 ö
a

(12)
= − ç ÷
[45]:40
D[45] 24 çè D[45] ÷ø
99744.168 − 1095.1562 11 æ 241.28824 ö

= − ç1 − ÷
5680.3705 24 è 5680.3705 ø
= 17.366651 − 0.438864
= 16.92779
Value of premiums = 16.92779P
Value of expenses = 0.45P + 0.05P a

[45]:20
= 0.45P + 0.05P × 17.366651
= 1.31833P
∴ 16.92779P = 42892.952 + 1.31833P
P = 2747.882
Monthly premium = £228.99
Page 10
(ii) Sum assured = 1.048 × 100,000 = 136,856.91
Value of reserves before alteration =
136,856.91 A53 − 228.99 × 12a

(12)
53:32
A53 = 1.04½ A53

4%
= 1.04½ × 0.4226 = 0.430969
N 53 − N 85 11 æ D85 ö
a

(12)
53:32
= − ç1 − ÷
D53 24 è D53 ø
60363.851 − 1095.1562 11 æ 241.28824 ö

= − ç1 −
4020.9326 24 è 4020.9326 ÷ø
= 14.74004 − 0.43083
= 14.30921
Value of reserves = 19,661.094.
= SA53 = S × 0.430969
∴ S × 0.430969 + 100 = 19,661.094
∴ S = £45,388.63
15 (i) Multiple decrement table
age (x) qxd qxs (al)x ( ad )dx ( ad )sx
62 0.01774972 0.15 100,000 1774.972 14733.754

63 0.01965464 0.15 83491.274 1640.991 12277.542
64 0.02174310 0 69572.741 1512.727 0
65 68060.014
Page 11
Unit Funds (ignoring actuarial funding)
Year, t 1 2 3
Value of Capital Units at start 0 3963.790 4093.703

Premium to CUs 3838 0 0
Interest on CUs 345.42 356.741 368.433
Management charge on CUs 219.630 226.828 234.262
Value of CUs at end 3963.790 4093.703 4227.874
Value of Accumulation Units at start 0 0 4131.127

Premium to AUs 0 3838 3838
Interest on AUs 0 345.42 717.221
Management charge on AUs 0 52.293 108.579
Value of AUs at end 0 4131.127 8577.769
Surrender value of units 3369.222 7815.460 12805.643
Capital Unit Fund (allowing for actuarial funding)
Year, t 1 2 3
Actuarial funding factor 0.89097 0.92528 0.96154

Value of CUs at start 0 3667.616 3936.259
Premium to CUs 3419.543 0 0
Interest on CUs 307.759 330.085 354.263
Management charge on CUs 195.683 209.879 225.252
Value of CUs at end 3531.618 3787.822 4065.270
Sterling Fund
Year, t 1 2 3
Unallocated premium 580.457 162 162

Expenses 300 60 60
Interest 12.621 4.59 4.59
MC on Capital Units 195.683 209.879 225.252
MC on Accumulation Units 0 52.293 108.579
Surrender profit 23.927 15.218 0
Extra death benefit 114.813 40.902 0
Cost of extra allocation 113.546 123.692 162.604
End of year cashflow 284.329 219.297 277.817
Probability in force 1 0.834913 0.695727

Discount factor 0.869565 0.756144 0.657516
Expected present value 247.243 138.445 127.088
Expected p.v. of profit = 512.776
Expected p.v. of premiums = 4000 × 2.25208 = 9008.323
Profit margin = 5.69%
Page 12
(ii) Revised Sterling Fund (ignoring reserves)
Year, t 1 2 3
Unallocated premium 162 162 162

Expenses 300 60 60
Interest −6.21 4.59 4.59
MC on Capital Units 219.630 226.828 234.262
MC on Accumulation Units 0 52.283 108.579
Extra death benefit 107.141 34.89 0
End of year cash flow 55.881 411.01 449.449
Reserves at start of year 400 400 400

Interest on reserves 18 18 18
Change in reserves at year end −66.035 −66.683 −400
Revised cashflow −260.084 95.693 467.449
Expected present value −226.160 60.412 213.835
Expected present value of profit = 48.087
Page 13
EXAMINATIONS
11 April 2001 (am)

separate sheet.

105 A2001 ã Institute of Actuaries
1 In the context of a unit-linked contract, state a key reason for the use of actuarial
funding of capital units. [2]
2 In the context of a pension fund, state what is meant by a transfer value. [2]
3 Under the Manchester Unity model of sickness, you are given the following
values:
t px = 1 − .05t2 (0 ≤ t ≤ 1)
z x +t = 0.1 (0 ≤ t ≤ 1)
Calculate sx . [3]
4 Some time ago, a life office issued an assurance policy to a life now aged exactly
55. Premiums are payable annually in advance, and death benefits are paid at
the end of the year of death. The office calculates reserves using gross premium
policy values. The following information gives the reserve assumptions for the
policy year just completed. Expenses are assumed to be incurred at the start of
the policy year.
Reserve brought forward at the start of the policy year £12,500

Annual premium £1,150
Annual expenses £75
Death benefit £50,000
Mortality A1967–70 ultimate
Interest 5.5% per annum
Calculate the reserve at the end of the policy year. [3]
5 Life insurance company A calculates paid-up policy values for endowment

assurance policies by applying the net premium reserve as a single premium at
the time of the alteration. It holds net premium reserves based on A1967–70
ultimate mortality and 3% per annum interest.
Life insurance company B calculates its corresponding paid-up values by

reducing the sum assured to (t/n) times the original sum assured, where n is the
original term of the policy and t is the number of premiums which have been paid
at the time of the calculation of the paid-up sum assured.
The sum assured is paid at the end of the term or the end of the year of death, if
earlier. Premiums are payable annually in advance.
Identify, showing your calculations, which company pays the higher paid-up sum
assured after 15 years, immediately before payment of the 16th premium, for a
25-year endowment assurance policy originally taken out by a life then aged
exactly 40. [4]
105 A2001—2
6 A life office prices sickness insurance contracts using the following three state
model in which the forces of transition depend on age:
σx
healthy sick
ρx
µx υx
dead
Level premiums are payable continuously. Benefits are payable continuously

during periods of sickness. There is no death benefit, and the contracts have a
deferred period of three months and include a waiver of premiums during periods
of benefit payment. Reserves are always positive under the normal premium
basis.
State briefly, with reasons, what effect the following changes will have on the
premium (certain increase, certain decrease, not certain), if the same net present
value of profit is to be achieved:
(a) an increase in the death rate from the sick state together with an increase
in the rate of transition from the healthy state to the sick state
(b) a fall in the death rate from the sick state together with a fall in the rate
of transition from the sick state to the healthy state. [4]
7 A population is subject to two modes of decrement, α and β, between ages x and

x + 1. In the single decrement tables
2
α æ x ö
t px = ç ÷
èx +tø
and
3
β æ x ö
t px = ç ÷
èx +tø
where 0 ≤ t ≤ 1.
Write down an integral expression for ( aq )αx . Hence or otherwise obtain an

expression for this probability in terms of x only. [6]
105 A2001—3 PLEASE TURN OVER

8 The following data relate to a population projection being carried out using the
component method, and specifically give information about the female population
of the country.
P(x, t) = population at 1 January 2001+ t aged x last birthday, x, t = 0, 1, 2, ..
M(x, t) = estimated net number of emigrants from the population during the
year 2001 + t, aged x last birthday at 1 January 2001 + t
q(x, t) = independent probability that a life who attains exact age x during the
year 2001 + t, dies during that year
The following is a selection from the available data.
x P(x,0) q(x,0) q(x,1) M(x,0) M(x,1)
54 728,610 .0121 .0115 37,013 31,461

55 700,369 .0136 .0129 35,868 30,126
56 678,123 .0152 .0144 34,312 28,994
57 620,975 .0170 .0161 31,179 24,943
Calculate, from the information given, the projected number of females aged 57
last birthday at 1 January 2003. State any assumptions you make. [6]
9 Define each of the following terms and give one example of each:
(a) class selection

(b) selective decrement
(c) spurious selection [6]
10 A life office has just sold a single premium deferred pension policy to a lady aged
exactly 45. This policy guarantees to pay a cash sum of £200,000 on her 60th
birthday, which must be used to buy a whole of life annuity at that time. The
policy also carries an annuity option whereby the policyholder can elect to receive
a pension of £15,000 per annum payable monthly in advance from the same date,
until her death.
The office invests the single premium such that the value of related assets on the
policyholder’s 60th birthday will be normally distributed with a mean value of
£250,000 and a standard deviation of £50,000. It also believes that the annual
interest rate, i, which will be available on the policyholder’s 60th birthday is a
random variable where i =
.04 with probability .25

The distribution of the value of assets on the policyholder’s 60th birthday is

independent of this annual interest rate, i.
105 A2001—4
Calculate the probability that the value of assets on the policyholder’s 60th
birthday is less than the cost of providing the annuity benefit, assuming the
policyholder is alive at age 60.
Basis for annuity rates: Mortality: a(55) female ultimate

Interest: i per annum
Expenses: Nil [8]
11 An annuity of £40,000 per annum is payable annually in arrear in respect of two

lives both aged 40.
• The first payment is deferred until the end of the year in which the first of
the two lives dies, and
• Payments continue until 5 years after the death of the survivor.
Assume the two lives are independent with respect to mortality.
Calculate the expected present value of this annuity.
Basis: A1967–70 ultimate mortality

4% per annum interest [10]
12 Two life offices operating in the same economy have maintained the following
records in respect of their male assured lives data:
deaths, subdivided by age last birthday at the preceding policy

anniversary, and duration at the preceding policy anniversary
in force, each 1 January, subdivided by age next birthday at issue, and

calendar year of policy issue
(i) Describe how you would calculate select forces of mortality, defining
carefully all symbols you use and stating necessary assumptions. State
clearly the ages and durations to which the resultant rates would apply.
[8]
(ii) Discuss briefly the advantages and disadvantages of pooling the data of
the two companies to form one mortality rate estimate for each
combination of age and duration. [3]
[Total 11]

13 A life insurance company sells with-profit whole life policies, with the sum
assured payable immediately on the death of the life assured and with premiums
payable annually in advance ceasing with the policyholder’s death or on reaching
age 65 if earlier.
The company markets two versions of this policy, one with simple reversionary
bonuses and the other with compound reversionary bonuses. In both cases the
bonuses are added at the end of the policy year.
The company prices the products using the following basis:
Mortality A1967–70 select

Interest 4% per annum
Expenses initial £250
renewal 2% of second and subsequent premiums
claim £150 at termination of contract
Bonuses simple 6% of basic sum assured per annum
compound 4% of accumulated sum assured and bonuses per
annum
(i) Write down an expression for the gross future loss at the point of sale for
each of these policies, assuming they are sold to a life aged x exact (x < 65)
at outset. Write the expression in terms of functions of the random
variables Tx and Kx , which represent the exact future lifetime and the
curtate future lifetime of (x), respectively. [5]
(ii) Calculate the gross premium required for each of the two policies for a
sum assured of £200,000 and a life aged 40 exact at outset, using the
equivalence principle. [8]
(iii) After 10 years, bonuses totalling £90,000 have been declared for the
compound reversionary bonus contract. Calculate the net premium
reserve for that policy at that time, using A1967–70 ultimate mortality
and interest of 4% per annum. [4]
[Total 17]
105 A2001—6
14 A life office issues an endowment assurance with a term of five years to a life
aged exactly 55. The sum assured is £100,000, payable at the end of the five
years or at the end of the year of death if earlier. Premiums are payable
annually in advance throughout the term of the policy.
The office assumes that initial expenses will be £300, and renewal expenses,
which are incurred at the beginning of the second and subsequent years of the
policy, will be £30 plus 2.5% of the premium. The funds invested for the policy
are expected to earn 7.5% per annum, and mortality is expected to follow the
A1967–70 select life table. The office holds net premium reserves, using
A1967–70 ultimate mortality and interest of 4% per annum.
The office sets premiums so that the net present value of the profit on the
contract is 15% of the annual premium, using a risk discount rate of 12% per
annum.
(i) Calculate the premium. [12]
(ii) Without carrying out any further calculations, state with brief reasons
what the effect on the premium would be in each of the following cases:
(a) The reserves are calculated using a lower rate of interest.

(b) The office uses a risk discount rate of 15%.
(c) Mortality is assumed to be A1967–70 ultimate.
[6]
[Total 18]
105 A2001—7
EXAMINATIONS
April 2001
EXAMINERS’ REPORT
Examiners’ Comments
The overall standard of scripts was somewhat disappointing, especially with regard to
answers offered for Questions 10, 11 and 12 where candidates were required to apply
principles from the syllabus to a problem that they may not have seen before. It is also
obvious that many candidates did not read the question asked, or at least did not
address the specifics of the question in their solution. Comments on the individual
questions follow.
Question 3
A surprisingly high number of candidates used the half-year approximation to work out
the integral without any justification, and a number omitted the factor of 52.18
Question 4
Many students used q55 and not q54 to calculate the death cost
Question 7
The majority of candidates could not quote any of the acceptable integrals and a further
number did not know how to obtain an expression for the force of the decrement from
the available data.
Question 8
Common errors included adding rather than subtracting the net number of emigrants
and not applying a survival factor to these emigrants. Clearly students did not read the
question correctly.
Question 9
Generally answered well, although a number of candidates seemed unaware of selective

decrements.
Question 10
This question was answered very poorly. Far too many candidates averaged the interest
rate as a first step and calculated the cost of annuity benefits at the expected interest
rate (6%). A further number then overlooked the option when finalising their solutions.
Page 2
Question 11
This question was answered poorly by many candidates. Most candidates who made
progress favoured the alternative answer shown below. While a number of students
identified the reversionary element correctly, few were able to correctly handle the 5-
year element.
Question 12
While many candidates had a general understanding of the construction of select

mortality rates, most had trouble applying it with the data given. Only a minority of
candidates appreciated that the in force data would have to be manipulated in order to
get a central exposed to risk which corresponded with the data for deaths.
Question 13
Part (ii) was generally well done. The commonest errors were to limit the benefit term to
25 years and to ignore that it was paid immediately on death. Part (i) was not answered
well as students seemed to confuse random variables and their expected values. Part(iii)
proved the most difficult. Many students calculated a gross premium reserve, others
correctly omitted allowance for expenses and future bonuses but used the gross
premium from (ii) and finally most of those who could correctly compute a net premium
did so using select mortality.
Question 14
This was generally answered quite well. A significant minority failed to recognise it as a
profit test question, and others constructed the reserves using the gross premium. In
part (ii), most students identified the correct effect on the NPV of the profit ( a
reduction) but wrongly assumed that the premium must also reduce.
Page 3
1 The office can more closely match income and outgo. The initial strain caused by
high initial expenses is reduced by capitalising the higher management charges
from the capital unit fund.
2 When a member leaves a pension scheme with an entitlement to deferred

benefits, they may elect in lieu to have a cash payment made by the scheme to
either a new scheme or an individual pension policy.
1
3 sx = 52.18 ò 0
t px zx +t dt
1
1 é .05t3 ù
ò
2
= 52.18 (1 − .05t ) (0.1) dt = 5.218 êt − ú
0
ë 3 û0
é .05 ù
= (5.218) ê1 − = 5.131
ë 3 úû
4 In general: (tV + P − E) (1 + i) = (qx+t) (S) + (px+t) (t+1V)
Here (12,500 + 1,150 − 75) (1.055) = (q54) (50,000) + (p54) (t+1V)
q54 = .00755572
Þ p54 = .99244428
1
Þ t+1V = {(13,575)(1.055) − (.00755572)(50000)}
.99244428
1
= {14,321.62 − 377.79} = 14,050
.99244428
Page 4
a55:10 8.371
5 Company A: 15 V40:25 = 1− =1− = 0.5124

a40:25 17.169
15 V40:25 0.5124
PUP SA = = = 0.6777
A55:10 0.75619
t 15 3
Company B: t = 15 n = 25 Þ PUP SA = = = = 0.6
n 25 5
Therefore Company A provides the higher PUP SA.
6 (a) Uncertain → healthy to sick transition rate increase causes

reduced premium income and higher claims but
increased sickness death rate leads to reduced
claims.
Depends on interaction of both effects.
(b) Definite increase → Both effects mean people are sick for longer.
Therefore higher premiums will be needed to
meet higher claims.
1 1
7 ( aq )αx = ò 0
t ( ap )x ( aµ )αx +t dt = ò 0
(t pxα )(t pxβ )(µ αx +t ) dt
1
= ò 0
(t pxα µαx +t ) t pxβ dt
d
But t pxα µαx +t = − (t pxα )
dt
2
α α d æ x ö 2 −3 2x 2
Þ t px µ x +t = − ç ÷ = (−1) (−2) (x ) (x + t) =
dt è x + t ø ( x + t )3
1 2x 2 x3 1 1
∴ ( aq )αx = ò .
0 ( x + t )3 ( x + t )3
dt = 2x5 ò 0 ( x + t )6
dt
1
é 1 1 ù 2é x5 ù
5ê ú = ê1 −
= 2x − ú
ê 5 ( x + t )5 ú 5ë ( x + 1)5 û
ë 0û
Page 5
8 We need P(57, 2) so we need to project P(55, 0) for 2 years, allowing appropriately

for emigrants.
2001 P(55, 0) = 700369 all have their 56th birthday during 2001 Þ use q(56, 0)
P(56, 1) = (700,369) (1 − .0152) − (35,868) [1 − (½) (.0152)] assuming net

emigration is spread uniformly across year
= 689723.4 − 35,595.4 = 654,128.0
2002 P(57, 2) = (654,128) (1 − .0161) − (28,994) [1 − (½) (.0161)] using (q57, 1)
= 643,596.5 −28,760.6 = 614,835.9 = 614,836
9 (a) Class selection: Refers to a factor affecting relative mortality which

is a permanent feature, e.g. age, sex, smoking
status etc.
(b) Selective decrement: When lives grouped by one decrement experience

different levels of another decrement, e.g. ill health
retirers usually experience heavier mortality than
other scheme members or retired members of
similar age/sex
or
marriage/mortality.
(c) Spurious selection: An investigation wrongly suggests that a certain

selection is present when it is not.
It usually results from unrecognised heterogeneity

in the data, with perhaps changing proportions of
lives subject to different underlying mortality
rates, e.g. occupational differences being the
underlying cause of “regional” mortality effects
or
changing sex mix leading to wrongly attributing

mortality rate progressions to temporary initial
selection.
Page 6
13
10 a (12)
60 = a60 + Þ @ 4% = .542 + 13.294 = 13.836
24
@ 6% = .542 + 10.996 = 11.538
@ 8% = .542 + 9.294 = 9.836
Þ annuity per annum available is @ 4%: (200,000 ÷ 13.836) = 14,455 p.a.

@ 6%: (200,000 ÷ 11.538) = 17,334 p.a.
@ 8%: (200,000 ÷ 9.836) = 20,333 p.a.
so the guaranteed minimum annuity option will only be chosen if i = .04
Þ if i = .06 or .08 (occurs with probability 0.75), the office will only make a loss if
assets at age 60 < £200,000.
(12)
60
The guarantee has value at vesting of (15,000) a
= (15,000) (13.836) = 207,540
so the insurer makes a loss if i = .04 (with prob = .25) if assets at age
60 < 207,540
Þ Probability of loss = (.75) [Prob (Assets < 200,000)] + (.25) [Prob(Assets < 207,540)]
é æ A − 250,000 200,000 − 250,000 ö ù 2

= (.75) ê P ç < ÷ ú where A ~ N(250,000, 50,000 )
ë è 50,000 50,000 øû
é æ A − 250,000 207,540 − 250,000 ö ù

+ (.25) ê P ç < ÷ú
ë è 50,000 50,000 øû
= (.75) [P(z < − 1)] + (.25) [P(z < − .8492)]
= (.75) [1 − .84134] + (.25) [1 − .80234] = (.75) (.15866) + (.25) (.19766)
= .118995 + .049415 = .16841 = 0.17
Page 7
11 From considering only the second condition, the first five payments are certain
to occur:
5
Þ EPV = 40,000 åv
t =1
t
= 40,000a5
Payments thereafter will be made if either life was alive five years earlier, with
probability t −5 p40:40
∞
Þ EPV = 40,000 åv
t =6
t
t −5 p40:40
Setting n = t − 5 (∴ t = n + 5) we get
∞ ∞
EPV = 40,000 åv
n =1
n +5
n p40:40 = 40,000v5 åv
n =1
n
n p40:40
= 40,000 v5 a40:40
However, the first condition means that no payment occurs on any date when
both lives are still alive, which occurs with probability t p40:40 .
∞
The previous values are too high by 40,000 åv
t =1
t
n p40:40 = 40,000 a40:40 .
Therefore the total expected present value is given by:
40,000[ a5 + v5 ( a40:40 ) − a40:40 ]
= 40,000[ a5 + v5 ( a40 + a40 − a40:40 ) − a40:40 ]
é ì 2N 41 N 41:41 ü N 41:41 ù
= 40,000 ê a5 + v5 í − ý− ú
êë î D40 D40:40 þ D40:40 úû
é ì (2)(125,015.43) 109,071.05 ü 109,071.05 ù

= 40,000 ê4.45182 + (.82193) í − ý− ú
ë î 6,986.4959 6,794.7238 þ 6,794.7238 û
= 40,000[4.45182 + (.82193) {(2)(17.89387) − (16.05232)} − 16.05232]
= 40,000[4.45182 + (.82193) (19.73542) − 16.05232]
= (40,000)(4.62063) = 184,825
Page 8
Alternative solution:
The first condition is the sum of two reversionary annuities:
(40,000)(2)( a40 40 ) = (80,000)( a40 − a40:40 )

or
is the difference between a last survivor annuity and joint life annuity:
(40,000)( a40:40 − a40:40 ) = (80,000)( a40 − a40:40 )
The second condition pays a 5 year annuity certain, with the first payment at the
end of the year of death of the survivor:
(40,000)( a5 ) A40:40
Using A40:40 = 1 − da40:40 = 1− d(2 a40 − a40:40 )
evaluating and summing the 2 elements leads to the same solution as above.
12 (i) Force of mortality Þ Deaths ÷ Corresponding central exposed to risk
In this case let θy,r = number of deaths where
y = age last birthday at previous policy anniversary
r = duration at previous policy anniversary
Þ both are policy year rate intervals
θ y ,r
estimates µ[y+½−r]+r+½
E yc ,r
At the start of the policy year rate interval lives are aged y last birthday,
with duration in force of exactly r years. Assuming birthdays are spread
uniformly over the policy year, this gives an average exact age at the start
of the rate interval of y + ½, or y + ½ − r at the start of the policy.
The duration in force half-way through the rate interval (appropriate for
force of mortality) is clearly r + ½.
For the central exposed to risk, initially we define
n Px,t = number of lives in force on 1.1.n where
x = age next birthday at issue
t = calendar year of issue
Page 9
Þ x + n − t = age next birthday at following policy anniversary
Þ x + n − t − 1 = age last birthday at following policy anniversary
Þ x + n − t − 2 = age last birthday at previous policy anniversary
and also n − t − 1 = duration at previous policy anniversary
For correspondence with deaths we need
n Py+2+t−n, n−r−1 and n+1 Py+1+t−n, n−r
and the appropriate central exposed to risk for calendar year n is:
c
n E y ,r = ½(n Py+2+t−n, n−r−1 + n+1 Py+1+t−n, n−r)
assuming all movements (new business, deaths, lapses etc.) are spread
evenly throughout the calendar year.
θ y ,r
Then µ[y+½−r]+r+½ =
å
n
c
n E y ,r
summing the central exposed to risk over the years of the study.
Alternative method for (i)
It is probably easier to actually restructure in force as follows:
At each census date, calculate r = census year − calendar year of issue
Let y = x + r
Þ age y − 2 last birthday at preceding policy anniversary
Þ duration r − 1 at preceding policy anniversary
Redefine in force as n Py,r = no. of lives in force on 1.1.n where
y = age next birthday at following policy

anniversary
r = duration at following policy anniversary
Page 10
Þ Appropriate central exposed to risk for calendar year n is
c
n E y ,r = ½(n Py+2, r+1 + n+1Py+2, r+1)
assuming n Py,r varies linearly over the calendar year
(rest of part (i) solution is as above)
(ii) Pooling the date will give rise to more credible estimates of true
underlying mortality rates, since greater exposure means lower variance.
However, one must be wary of heterogeneity in the data from the two
offices:
e.g. differing geographical coverage

differing underwriting standards
different distribution or target market etc.
13 (i) Simple bonus version:
{
L = 250 + (S[1 + (.06)Kx] + 150) vTx − P (.98)amin[1+ K
x ,65 − x
+ .02P }
Compound bonus version:
{
L = 250 + (S[(1.04)K x + 150) vTx − P (.98) amin[1+ K
x ,65 − x
+ .02P }
(ii) Equivalence principle Þ E[L] = 0
Also we shall assume E[T] * E[K] + ½

Simple bonus:
D[ x ]+1
Þ 250 + (S + 150) A[ x ] + (.06S ) ( IA )[ x ]+1 = P é(.98) a[ x ]:65− x + .02ù
D[ x ] ë û
"""""""" " """"""""" !
more easily valued as
(.94S + 150) A[ x ] + .06S ( IA )[ x ]
Page 11
In this case:
é R[40] ù
250 + (1.04)½ ê{(.94) (200,000) + 150} A[40] + (.06) (200,000) ú
êë D[40] úû
= P [(.98) a[40]:25 + .02]
é 57,705.359 ù
Þ 250 + (1.04)½ ê188,150 (.27284) + (12,000)
ë 6,981.5977 úû
= P[(.98) (15.609) + .02]
Þ 250 + (1.04)½ [51,334.85 + 99,184.22] = P[15.31682]
Þ P = 153,749.94 ÷ 15.31682 = £10,038 p.a.
Compound bonus:
é A* ù
250 + (1.04) ê 200,000 + 150 A[ x ] ú = 15.31682P
½ [40]
ê 1.04 ú
ë û
i −b
* at i.e. 0%
1+b
é 200,000 ù
Þ (250) + (1.04)½ ê + (150) (.27284)ú = 15.31682P
ë 1.04 û
Þ 250 + (1.04)½ [192,307.69 + 40.926] = 15.31682P
Þ P = 196,407.87/15.31682 = £12,823p.a.
(iii) Net Premium Reserve for WP policies
(i) allows for accrued bonuses only

(ii) net premium ignoring any bonuses
Þ 10V = 290,000 A50 − (NP) a50:15
A40 (200,00) (1.04)½ (.27331)

where NP = 200,000 =
a40:25 15.599
= 3,573.60 p.a.
Page 12
Þ 10V = (290,000) (1.04)½ (.38450) − (3,573.60) (10.995)
= 113,713.23 − 39,291.73 = 74,421.50
a56:4 3.720
14 (i) Reserves 1 V55:5 =1− =1− = 0.18188
a55:5 4.547
Similarly 2V = .37189
3V = .57115
4V = .78007
5V = 0 assuming all claims paid in cash flow outgo
Also q[55] = .00447362 p[55] = .99552638 Þ 0 p[55] =1

q[55]+1 = .00625190 p[55]+1 = .99374810 1 p[55] = .995526
q57 = .01049742 p57 = .98950358 2 p[55] = .989302
q58 = .01168566 p58 = .98831434 3 p[55] = .978917
q59 = .01299373 p59 = .98700627 4 p[55] = .967478
Opening Closing Profit

Year Premium Expense Reserve Interest Claim Reserve Vector
1 P 300 0 .075P−22.5 447.36 18106.63 1.075P−18876.49

2 P .025P+30 18188 .073125P+1361.8 625.19 36956.50 1.048125P−18061.89
3 P .025P+30 37189 .073125P+2786.9 1049.74 56515.44 1.048125P−17619.28
4 P .025P+30 57115 .073125P+4281.4 1168.57 77095.44 1.048125P−16897.61
5 P .025P+30 78007 .073125p+5848.3 100000 0 1.048125P−16174.70
Þ Profit Signature NPV of Profit Signature
1 1.075P − 18,876.49 ×v = .95982P − 16,854.01

2 1.043446P − 17,981.08 ×v2 = .83183P − 14,334.41
3 1.036912P − 17,430.79 ×v3 = .73805P − 12,406.89
4 1.026027P − 16,541.36 ×v4 = .65206P − 10,512.33
5 1.014038P − 15,648.67 ×v5 = .57539P − 8,879.48
3.75715P − 62,987.12
NPV = .15P = 3.75715P − 62,987.12
Þ P = 62,987.12 ÷ (3.75715 − .15) = £17,462
Page 13
(ii) (a) Lower interest rate Þ larger reserves

Larger reserves Þ profit deferred
Profit deferred Þ NPV reduced
(since risk rate (12%) > earned (7.5%))
Þ profit falls below 15% premium
Þ need to increase premium to re-establish 15% margin
(b) Higher risk rate Þ lower NPV etc.

Þ need higher premium to meet profit requirement
(c) Ultimate mortality Þ more death claims

Þ earlier claims (NPV claims increases) and reduced
premium income
Þ NPV falls
Þ need to increase premium
Page 14
EXAMINATIONS

separate sheet.

105 S2001 ã Institute of Actuaries
1 Under the Manchester Unity model of sickness, you are given the following
values:
sx = 5
1
ò 0
t px dt = 0.9
Calculate the value of zx . [2]
2 Give a formula for P21 (2003) in terms of P20 (2002) , based on the component
method of population projection. Px (n ) denotes the population aged x last
birthday at mid-year n.
State all the assumptions that you make and define carefully all the symbols that
you use. [3]
3 A life insurance company issues a policy under which sickness benefit of £100 per
week is payable during all periods of sickness. There is a waiting period of 1 year
under the policy.
You have been asked to calculate the premium for a life aged exactly 30, who is
in good health, using the Manchester Unity model of sickness.
Describe how you would allow for the waiting period in your calculation, giving a
reason for your choice of method. [3]
105 S2001—2
4 An employer recruits lives aged exactly 20, all of whom are healthy when
recruited. On entry, the lives join a scheme that pays a lump sum of £50,000
immediately on death, with an additional £25,000 if the deceased was sick at the
time of death.
The mortality and sickness of the scheme members are described by the following
multiple-state model, in which the forces of transition depend on age only.
σx
ρx
µx νx
Dead (D)
All surviving members retire at age 65 and leave the scheme regardless of their
state of health.
pxab,t is defined as the probability that a life who is in state a at age x (a = H, S, D)

is in state b at age x + t (t ≥ 0 and b = H , S , D ) .
Write down an integral expression for the expected present value, at force of
interest δ , of the death benefit in respect of a single new recruit. [3]
5 A pension scheme provides a pension of 1/60 of career average salary in respect of

each full year of service, on age retirement between the ages of 60 and 65. A
proportionate amount is provided in respect of an incomplete year of service.
At the valuation date of the scheme, a new member aged exactly 40 has an
annual rate of salary of £40,000.
Calculate the expected present value of the future service pension on age
retirement in respect of this member, using the Pension Fund Tables in the
Formulae and Tables for Actuarial Examinations. [3]
105 S2001—3 PLEASE TURN OVER

6 A life insurance company issues a special annuity contract to a male life aged
exactly 70 and a female life aged exactly 60.
Under the contract, an annuity of £10,000 per annum is payable monthly to the
female life, provided that she survives at least 10 years longer than the male life.
The annuity commences on the monthly policy anniversary next following the
tenth anniversary of the death of the male life and is payable for the balance of
the female’s lifetime.
Calculate the single premium required for the contract.
Basis: Mortality: a(55) Ultimate, males or females as appropriate

Expenses: none [4]
7 The staff of a company are subject to two modes of decrement, death and
withdrawal from employment.
Decrements due to death take place uniformly over the year of age in the
associated single-decrement table: 50% of the decrements due to withdrawal
occur uniformly over the year of age and the balance occurs at the end of the year
of age, in the associated single-decrement table.
You are given that the independent rate of mortality is 0.001 per year of age and
the independent rate of withdrawal is 0.1 per year of age.
Calculate the probability that a new employee aged exactly 20 will die as an
employee at age 21 last birthday. [4]
8 The following data are available from a life insurance company relating to the
mortality experience of its temporary assurance policyholders.
θx ,d The number of deaths over the period 1 January 1998 to 30 June 2001,
aged x nearest birthday at entry and having duration d at the policy
anniversary next following the date of death.
Py ,e (n ) The number of policyholders with policies in force at time n, aged y

nearest birthday at entry and having curtate duration e at time n, where
n = 1.1.1998, 30.6.1998, 30.6.2000 and 30.6.2001.
Develop formulae for the calculation of the crude central select rates of mortality
corresponding to the θx ,d deaths and derive the age and duration to which these
rates apply. State all the assumptions that you make. [6]
105 S2001—4
9 (i) State the conditions necessary for gross premium retrospective and
prospective reserves to be equal. [3]
(ii) Demonstrate the equality of gross premium retrospective and prospective

reserves for a whole life policy, given the conditions necessary for equality.
[4]
[Total 7]
10 A life insurance company issues a special term assurance policy to two lives aged
exactly 50 at the issue date, in return for the payment of a single premium. The
following benefits are payable under the contract:
(i) In the event of either of the lives dying within 10 years, a sum assured of
£100,000 is payable immediately on this death.
(ii) In the event of the second death within 10 years, a further sum assured of
£200,000 is payable immediately on the second death.
Calculate the single premium.
Basis: Mortality: A1967–70 Ultimate

Expenses: None [8]

11 A life insurance company sells term assurance policies with terms of either 10 or
20 years.
As an actuary in the life office, you have been asked to carry out the first review
of the mortality experience of these policies. The following table shows the
statistical summary of the mortality investigation. In all cases, the central rates
of mortality are expressed as rates per 1,000 lives.
All policies 10-year policies 20-year policies
Age Number Central Number Central Number Central

in force mortality in force mortality in force mortality
rate rate rate
–24 6,991 1.08 6,013 0.86 978 2.12

25–44 6,462 2.05 5,438 1.74 1,024 3.68
45–64 5,815 13.26 4,942 11.55 873 22.94
65– 3,051 75.70 2,570 71.53 481 97.70
Total 22,319 18,963 3,356
(i) Calculate the directly standardised mortality rate and the standardised
mortality ratio separately in respect of the 10-year and 20-year policies.
In each case, use the “all policies” population as the standard population.
[6]
(ii) You have been asked to recommend which of these two summary
mortality measures should be monitored on a regular basis.
Give your recommendation, explaining the reasons for your choice. [3]
[Total 9]
105 S2001—6
12 A life insurance company offers an option on its 10-year without profit term
assurance policies to effect a whole life without profits policy, at the expiry of the
10-year term, for the then existing sum assured, without evidence of health.
Premiums under the whole life policy are payable annually in advance for the
whole of life, or until earlier death.
(i) Describe the conventional method of pricing the mortality option, stating
clearly the data and assumptions required. Formulae are not required.
[3]
(ii) A policyholder aged exactly 30 wishes to effect a 10-year without profits

term assurance policy, for a sum assured of £100,000.
Calculate the additional single premium, payable at the outset, for the
option, using the conventional method.
The following basis is used to calculate the basic premiums for the term
assurance policies.

Expenses: none [4]
(iii) Describe how you would calculate the option single premium for the policy
described in part (ii) above using the North American method, stating
clearly what additional data you would require and what assumptions you
would make. [4]
(iv) State, with reasons, whether it would be preferable to use the

conventional method or the North American method for pricing the
mortality option under the policy described in part (ii) above. [3]
[Total 14]

13 (i) On 1 September 1996, a life aged exactly 50 purchased a deferred annuity
policy, under which yearly benefit payments are to be made. The first
payment, being £10,000, is to be made at age 60 exact if he is then alive.
The payments will continue yearly during his lifetime, increasing by
1.923% per annum compound.
Premiums under the policy are payable annually in advance for 10 years
or until earlier death.
If death occurs before age 60, the total premiums paid under the policy,
accumulated to the end of the year of death at a rate of interest of 1.923%
per annum compound, are payable at the end of the year of death.
Calculate the annual premium.
Basis: Mortality: before age 60: A1967–70 Ultimate
after age 60: a(55) Males Ultimate
Expenses: initial: 10% of the initial premium, incurred

at the outset
renewal: 5% of each of the second and

subsequent premiums, payable at the
time of premium payment
claim: £100, incurred at the time of payment

of the death benefit [9]
(ii) On 1 September 2001, immediately before payment of the premium then

due, the policyholder requests that the policy be altered so that there is no
benefit payable on death and the rate of increase of the annuity in
payment is to be altered. The premium under the policy is to remain
unaltered as is the amount of the initial annuity payment.
The life insurance company calculates the revised terms of the policy by
equating gross premium prospective reserves immediately before and
after the alteration, calculated on the original pricing basis, allowing for
an expense of alteration of £100.
Calculate the revised rate of increase in payment of the annuity. [7]

[Total 16]
105 S2001—8
14 A life insurance company issues a 3-year unit-linked endowment assurance
contract to a male life aged exactly 60 under which level annual premiums of
£5,000 are payable in advance throughout the term of the policy or until earlier
death. 102% of each year’s premium is invested in units at the offer price.
The premium in the first year is used to buy capital units, with subsequent years’
premiums being used to buy accumulation units. There is a bid-offer spread in
unit values, with the bid price being 95% of the offer price.
The annual management charges are 5% on capital units and 1% on

accumulation units. Management charges are deducted at the end of each year,
before death, surrender or maturity benefits are paid.
payable at the end of the year of death of £12,000 or the bid value of the units
payable.
The policy may be surrendered only at the end of the first or the second policy
year. On surrender, the life insurance company pays the full bid value of the
accumulation units and 80% of the nominal bid value of the capital units,
calculated at the time of surrender.
The company holds unit reserves equal to the full bid value of the accumulation
units and a proportion, A60+t :3−t (calculated at 4% interest and A1967-70 Ultimate
mortality), of the full bid value of the capital units, calculated just after the
payment of the premium due at time t (t = 0,1 and 2). The company holds no
sterling reserves.
The life insurance company uses the following assumptions in carrying out profit
tests of this contract:
Expenses: initial: £400
renewal: £80 at the start of each of the second and third

policy years
Sterling fund interest rate: 5% per annum
Surrender rates: 20% of all policies still in force at the end of each of
the first and second years
Calculate the profit margin on the contract. [18]
105 S2001—9
EXAMINATIONS
September 2001
EXAMINERS’ REPORT
Examiners’ Comments
Overall the standard of attempts was lower than the examiners would have expected. There was
evidence that many candidates spent too much time on the earlier questions, with consequent time
problems later on.
Questions 1,2,3,4,9 and 11 were well answered. In question 5, career average salary was not dealt
with well. Candidates had difficulty with year-end decrements in question 7 and with the duration in
question 8. Question 9(i) was poorly answered, although it was a standard question.
Question 10 was the most poorly attempted, with few candidates scoring more than half marks. There
was an ambiguity in this question: the benefit payable on the second death could have been
interpreted as £200,000 or £300,000. Candidates were given credit for either approach. Many
candidates did not give sufficient detail in their answers to question 12(i). Question 13(ii) was poorly
attempted and the answers to question 14 were not as strong as one would have expected for a fairly
standard question.
Page 2
1
1 sx = 52.18 ò0
t px zx +t dt
zx =
ò
52.18 lx +t zx +t dt
0
=
sx * lx
=
sx
=
5
= 5.56
1 1 1
0.9
ò 0
lx +t dt ò
0
lx +t dt ò
0
t px dt
2 The required formula is:
P21 (2003) = P20 (2002)(1 - q20 1 2 (2002)) + M 21 (2003)
q20 1 2 (2002) is the probability that a life aged 20 last birthday at mid-year 2002
dies between mid-year 2002 and mid-year 2003, assuming those aged 20 last
birthday at mid-year 2002 have birthdays uniformly distributed over the
calendar year.
M 21 (2003) denotes the number of migrants entering the population during mid-
year 2002 and mid-year 2003 who survive to be aged 21 last birthday at mid-year
2003.
The formula is applied separately to males and females.
3 To allow for the fact that benefit cannot be paid for at least one year, the sickness
benefit could be valued using the factor
K 31 - K x
100 , where x is the ceasing age for benefits.
D30
However, the factor K 31 is not accurate as it takes into account sickness of all
durations, whereas a new policyholder aged 30 cannot experience sickness of all
durations from age 31. For this reason and because the numerical effect is not
significant, I would use the factor K 30 rather than K 31 in the above formula.
4 The required expression is
45
ò e -dt {2 p20,
HH HS
25,000 t .m 20 +t + 3 p20,t .n 20 +t }dt
0
Page 3
5 The value of the benefits is
40,000 s R40 ra
40,000 1,758,471
= = £63,816.35
60 s39 1 2 D40 60 1
2 (3.48 + 3.58).5,204
6 The required single premium is given by
f
D70
f 10
10000 10 p60 (12)
v a&&70m
|70 f
= ( ) (
a f - a70m :70 f
D60 70
)
3571.2
= 10000 (7.308 - 5.106 )
8858.7
= £8,876.90
7 Construct a multiple decrement table.
Age No. alive No. deaths No. withdrawals No. withdrawals

over year at year end
20 100000 97.50 4997.5 4745.25
21 90159.75 87.9058
At age 20, no. of deaths = 100000*0.001(1-0.5*0.05) = 97.50
no. of withdrawals over year = 100000*0.05*(1-0.5*.001) = 4997.5
no. of withdrawals at year end = 100000*(1-0.05)*(1-0.001)*

0.05 = 4745.25
Required probability = 87.9058/100000 = 0.00087906.
8 Define a census taken at time t after the start of the period of investigation
(1.1.98), Px' ,d (t ) , of those lives having a policy in force at time t, who were x
nearest birthday at entry and will be duration d on the policy anniversary next
following time t.
The central exposed to risk is then given by
t =3.5
E xc ,d = ò
t =0
Px' ,d (t )dt
Page 4
Assuming that Px' ,d (t ) varies linearly between the census dates the integral can
be approximated by
2 { }
2 { } 2 { }
1 * 1 Px' ,d ( 0 ) + Px' ,d ( 1 2 ) + 1 * 2 Px' ,d ( 1 2 ) + Px' ,d ( 2 1 2 ) + 1 *1 Px' ,d ( 2 1 2 ) + Px' ,d ( 3 1 2 )
2
However, the census data have been recorded according to age x nearest birthday
at entry and curtate duration d at time t. The following formula may be written:
Px' ,d (t ) = Px ,d -1 (t ) .
Substituting this into the equation above gives
E xc ,d = 1 * 1 {Px ,d -1 ( 0 ) + Px ,d -1 ( 1 2 )} + 1 * 2 {Px ,d -1 ( 1 2 ) + Px ,d -1 ( 2 1 2 )}
2 2 2
+ 1 *1 {Px ,d -1 ( 2 1 2 ) + Px ,d -1 ( 3 1 2 )}
2
o qx ,d
mx ,d = estimates m[ x ]+ d -1 because the average age at entry is x assuming
E xc ,d
birthdays are uniformly distributed over the policy year and the exact duration
at the start of the rate year of death is d - 1 for all lives (no assumptions are
necessary).
9 (i) Gross premium retrospective and prospective reserves will be equal if:
· The mortality and interest rate basis is the same for the retrospective
and prospective reserves and is the same as that used to determine the
gross premium at the date of issue of the policy.
· The same expenses (excluding the initial expenses) are valued in the
retrospective and prospective reserves and also the expenses valued in
the retrospective reserves are the same as those used to determine the
original gross premium.
· The gross premium valued in the retrospective and prospective

reserves is that determined on the original basis using the equivalence
principle.
(ii) The prospective reserves at time t are given by
SAx +t + ea&&x( m+t) + fAx +t - Ga&&x(m+t) ………… (a)
where S is the sum assured

e is the annual rate of renewal expenses
f is the claim expense
G is the annual rate of gross premium
Page 5
The retrospective reserve at time t is given by
Dx
Dx +t
{ }
Ga&&x :t| - SA1x :t| - I - ea&&x(m:t|) - fAx1:t| ……………(b)
where I is the additional initial expense.
The original gross premium is given by
Ga&&x( m ) - SAx - I - ea&&x( m ) - fAx = 0 ………….. (c)
Dx
Add
Dx +t
{ }
Ga&&x( m ) - SAx - I - ea&&x( m ) - fAx , which is identically 0, to (a).
Dx D
Combining terms, e.g. Ga&&x( m ) - Ga&&x(m+t) = x Ga&&x( m:t|) gives (b), the
Dx +t Dx +t
expression for the retrospective reserve.
10 The value of benefit (a) and (b) is
æ ö
100000ç 2 A501 :10| + A50 ÷
67 18
ç : 50 :10| ÷
è ø
1 æ D D60:60 ö
» (1.04 ) 2 ç 1 - da&&50:50 :10| - 2 60 + ÷
6718
A50 : 50:10| ç D50 D50:50 ÷
è ø
a&&50:50:10| = 2a&&50:10| - a&&50:50:10|
N 50:50 - N 60:60 59513.103 - 24729.51

a&&50:50:10| = = = 7.98781
D50:50 4354.5857
a&&50:10| = 8.207
D60 2855.5942
= = 0.621178
D50 4597.0607
D60:60 2487.2117
= = 0.57117
D50:50 4354.5857
1 M 50 - M 60 1 1767.5555 - 1477.0842
A501 :10| = 1.04 2
= 1.04 2 = 0.064438
D50 4597.0607
\ the required premium is
Page 6
( )
100000 2 * 0.064438 + (1.04 ) 2 (1 - 0.038462(2 * 8.207 - 7.98781) - 2 * 0.621178 + 0.57117 )
1
= £13,369.55
Alternative solution
The benefits payable may be regarded as a sum of £200,000 payable on either

death, less a sum of £100,000 payable on the first death.
\ the value of the benefits is:
200000 * 2 A501 :10| - 100000 A50

6718
:50:10|
A501 :10| = 0.064438
1 æ D60:60 ö
A50
6718
= 1.04 2
ç
ç 1 - da&&50:50:10|
- ÷÷
:50:10|
è D50:50 ø
= 1.04 2 (1 - 0.038462 * 7.98781 - 0.57117 ) = 0.124011

1
\ the required premium is
200000*2*0.064438-100000*0.124011
=£13,374.10
(The difference in the two answers is due to rounding)
Page 7
11 (i) The directly standardised rate (DSR) is given by
åE m
x
cs
x ,t x ,t
DSR =
åE x
cs
x ,t
The standardised mortality ratio (SMR) is given by
åE x
c
x ,t mx ,t
SMR =
åE x
c s
x ,t mx ,t
For 10-year policies:
6.991 * 0.86 + ... 6.013 * 0.86 + ...

DSR = = 13.56053 SMR = = 0.920149
6.991 + 6.013 *1.08 + ...
For 20-year policies:
6.991 * 2.12 + ... 0.978 * 2.12 + ...

DSR = = 21.06187 SMR = = 1.424669
6.991 + 0.978 *1.08 + ...
Note: In each of the above the DSR is expressed as the number of deaths
per 1,000.
(ii) I would favour the standardised mortality ratio. The directly standardised
mortality rate requires mx ,t to be recorded for each age group, for the 10-
year and 20-year policies separately. The data may not be readily
available. The SMR requires the number of deaths in each age and policy
group only to be recorded: these data should be easily recorded.
12 (i) In pricing the mortality option using the conventional method, the actuary
pricing the option assumes:
· that all lives eligible to take up the option will do so, and
· that the mortality experience of those who take up the option will be
the Ultimate experience which corresponds to the Select experience
that would have been used as a basis if underwriting had been
completed as normal when the option was exercised
The mortality basis used is not usually assumed to change over time, so
that the only data required are the Select and Ultimate mortality tables
used in the original pricing basis.
In pricing the mortality option, the actuary values the premium income
assuming that the premium payable at the end of the ten years is
calculated using Select rates according to the original premium basis and
Page 8
values the premiums assuming Select rates apply only from the date of
issue of the original policy. The actuary values the liabilities similarly.
The difference in the present value of the premium income and benefit
liability per policy originally issued gives the additional option single
premium, per policy issued.
(ii) Whole life premium payable = 100000P[40]
Whole life premium which should be paid according to the actuary’s basis
= 100000P40
Option premium = present value of the difference in premiums =
ìï ü
( ) l ï
100000 í P40 - P[40] v10 40 a&&40 ý
l[30]
îï ï
þ
ì 1 33542.311 ü
= 100000 í( 0.01063 - 0.01058 ) 10
*14.874 ý = £41.18
î 1.06 33828.764 þ
(iii) I would require the following data:
· an estimate of the probability of those reaching age 40 as

policyholders, who exercise the option
· a multiple decrement table to describe the mortality and other

relevant decrements (such as surrender) of those who exercise the
option, commencing at age 40
· the basis on which the whole life premium payable is to be calculated:

this would normally be assumed to based on the 1967-70 Select
mortality, similar to the premium basis set out
I would calculate the present value of the additional liability, using the
multiple decrement table from age 40 and allowing for the probability of
exercise of the option and A1967-70 Select mortality before age 40.
I would then calculate the whole life premium payable and also the
present value of the whole life premiums payable, similarly to the method
used to calculate the additional liability.
The difference between the two values, per term assurance policy issued,
would be the option premium.
Page 9
(iv) The more accurate method is the North American method. However, I
would favour the conventional method for the following reasons.
· There may not be sufficient data available to apply the North

American method.
· If the policy were being sold in a market where the conventional

method was generally used for pricing, then there would be adequate
experience of the use of the method in the market.
· Even if sufficient data were available in respect of the North American

method, they might not be appropriate for pricing the portfolio
concerned, particularly if the pricing were being done when the
business was first issued.
13 (i) Present value of annuity payments:
1.01923 1.06 = 1.04 Þ annuity payments are valued at 4% p.a.
6%
D60 1 l60
Value = 100000 a&&60
4%
= 100000 (11.625 + 1)
D50 1.0610 l50
1 30039.787
= 10000 12.625 = 64,821.99
1.0610 32669.855
Present value of death benefits:
Present value of death benefit at age 50 + t
= P&s&t1|.923% v6t % , where P is the annual premium.
=P
1.01923
0.01923
(1.01923t - 1)
1
1.06 t
=
1.01923
0.01923
P (v 4t % - v6t % )
\ the present value of the death benefits is
1.01923
0.01923
4%
( 6%
P A501 :10| - A501 :10| )
4% D60 2855.5942
A50:10|
1 = A50:10| - = 0.68436 - = 0.063182
D50 4597.0607
6% 1 l60 30039.787
A50:10|
1 = A50 - 10
A60 = 0.25736 - 0.55839 * * 0.39136
1.06 l50 32669.855
=0.056421
Page 10
\ the present value of the death benefits is
1.01923
(0.063182 - 0.056421)P = 0.358347 P
0.01923
Present value of premiums less expenses:
= 0.95 Pa&&50:10| - 0.05 P - 100 A50:10|

1
= 0.95 P * 7.599 - 0.05 P - 100 * 0.056421
= 7.16905 P - 5.6421
\ 7.16905P - 5.6421 = 64821.99 + 0.358347 P
P = £9,518.49
(ii) At the date of alteration:
Present value of annuity payments before alteration
6%
D60 4% 1 l60
= 10000 a&&60 = 10000 (11.625 + 1) = 10000 * 0.708453 *12.625
D55 1.065 l55
= 89442.19
Present value of death benefit before alteration
4%
Ps&&5|1.923% A55:5|
1 +
1.01923
0.01923
(
4%
P A55:5|
1
6%
- A55:5|
1 )
1.01923
= P *5.295953* 0.045886 + P ( 0.045886 - 0.043249 )
0.01923
= 0.382777 P
= 3643.459
Present value of annuity payments after alteration
=89442.19+3643.459-100 = 92985.649
Present value of annuity payments before alteration, based on a rate of

interest of 6% after age 60
10.813
= * 89442.19 = 76605.02
12.625
Page 11
Estimated interest rate underlying annuity payments after alteration
92985.649 - 89442.19
0.04 - * 0.02 = 0.034479
89442.19 - 76605.02
Revised rate of escalation
1.06
= - 1 = 2.467%
1.034479
14 Multiple decrement table
Age qxd ( al )dx ( ad )dx ( ad )sx

60 0.014432 100000 1443.246 19711.35
61 0.016014 78845.4 1262.596 15516.56
62 0.01775 62066.25 1101.658 12192.92
Probabilities of survival
Age t px
60 1
61 0.788454
62 0.620662
Unit fund (ignoring actuarial funding)
Year 1 2 3
Fund brought forward 0 4970.97 5100.215
Premiums allocated to CU 4845 0 0
Interest 387.6 397.6776 408.0172
Management charge 261.63 268.4324 275.4116
Fund carried forward 4970.97 5100.215 5232.821
Fund brought forward 0 0 5180.274

Premiums allocated to AU 0 4845 4845
Interest 0 387.6 802.0219
Management charge 0 52.326 108.273
Fund carried forward 0 5180.274 10719.02
Surrender values 3976.776 9260.446 15951.84
Unit fund (with actuarial funding)
Year 1 2 3
Actuarial funding factor 0.890605 0.925148 0.961538
Fund brought forward 0 4598.885 4904.053
Premiums allocated to CU 4314.979 0 0
Interest 345.1983 367.9108 392.3242
Page 12
Fund carried forward 4427.169 4718.456 5031.558
Sterling fund
Year 1 2 3
Unallocated premium 685.021 155 155
Expenses 400 80 80
Interest 14.251 3.75 3.75
Mortality charge 109.2946 33.64881 0
Additional allocation 135.3906 146.0999 201.2624
Fund at year end 376.3742 325.2797 250.5794
Present value of profit = 623.4689
Present value of premiums = 10774.61
Page 13
EXAMINATIONS
9 April 2002 (am)
1. Enter all the candidate and examination details as requested on the front of your answer
booklet.
2. You must not start writing your answers in the booklet until instructed to do so by the
supervisor.
4. Attempt all 14 questions, beginning your answer to each question on a separate sheet.
Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available Actuarial Tables and
your own electronic calculator.
105 A2002 ã Institute of Actuaries
1 (i) Explain what is meant by the following expression:
3|4 q [40]+1 .
(ii) Calculate its value using A1967–70 mortality. [3]
2 Two lives, each aged exactly 35, are independent with respect to mortality and are
each subject to a constant force of mortality of 0.02.
Calculate the value of the following expression for these lives:
1
10|20 q35:35 [4]
3 (i) Define the term Total Fertility Rate and explain the difference between rates
calculated on a cohort basis and a period basis. [3]
(ii) In the context of population projection, state, with a reason, which basis is
preferable. [2]
[Total 5]
4 Due to a downturn in the economy, the numbers unemployed in a certain country are
expected to increase. The current number unemployed is 100,000, and this is
expected to rise towards but not exceed 300,000 following the logistic growth model.
The initial rate of increase in unemployment will be 25% per annum. Calculate how
long it will take for the unemployed population to reach 200,000. [5]
5 A life insurance company sold a number of 4-year single premium policies with a
guaranteed amount payable at maturity.
The closest matching investment available was a 5-year zero-coupon bond. Interest
rates at the time of the insurance company selling the policies and investing the
money in the bonds were 5.25% effective per annum.
The office invested all the premiums received in these assets. The insurance company
guaranteed a return of 5.0% per annum at maturity. On death, the return was not
guaranteed but the company promised to pay out the full market value of the related
asset immediately at the date of death.
If the distribution of 1 + i is log-normal with parameters m = 0.05 and s = 0.01, and

mortality follows English Life Table No. 12 – Males, calculate the probability that the
office makes a loss on a policy sold to someone aged exactly 56. You should assume
that the company sells the matching asset at the time of any claim.
(Remember that if X~lognormal with parameters m and s, then loge(X) is normally

distributed with mean m and standard deviation s.) [5]
105 A2002—2
6 A retiring employee aged exactly 60 is given a choice between the following two
pensions:
Pension A is payable annually in arrear throughout the pensioner’s lifetime, with at

least 4 payments guaranteed to be made. The first payment is £20,000 and payments
increase by 0.9709% per annum compound thereafter.
Pension B is payable annually in arrear, with an initial payment of £13,000. Each

subsequent payment increases by £1,000 and payments cease immediately on death.
Calculate the expected present value of each pension using the following basis:
Mortality: A1967–70 Select

7 Describe how nutrition and education affect mortality. [6]
8 A mortality investigation of pensioners who retired due to ill health is being

undertaken to investigate if there are any initial temporary select effects. Data in
respect of deceased pensioners are categorised as follows:
x: age last birthday at death

r: curtate number of years between retirement and death
(i) Estimates of m[y]+t are made by dividing the death data by its corresponding
central exposed to risk. Derive the values of y and t in terms of x and r,
stating clearly any assumptions you need to make. [3]
(ii) The following data are also available in respect of one pensioner:
Date of birth 1 August 1936

Date of retirement 1 November 1998
Date of death 1 July 2001
Calculate the contribution of this individual to each of the appropriate central

exposed to risk figures corresponding to the available summary data. [3]
[Total 6]

9 A life insurance company sells 4-year unit-linked endowment assurance contracts to
males aged 61 exact. Premiums of £1,000 are payable annually in advance.
Capital units are bought by the premium in the first year, and accumulation units are
bought thereafter. 102% of each year’s premium is invested in units at the offer price.
There is a bid-offer spread in unit values, with the bid price being 95% of the offer
price.
Capital units bear a management charge of 5% per annum of the bid value of the fund,
and this charge is deducted at the end of every year.
The death benefit under the policy is paid at the end of the year of death, and is the
full bid value of units under the policy, after deduction of relevant management
charges.
The pricing actuary assumes that fund growth will be 7.5% per annum and that
mortality experience will follow A1967–70 Select. He is contemplating using part of
the management charge for actuarial funding of capital units. The actuarial factor at
duration t would be A[61]+t:4-t at a rate of interest of 4% per annum, with mortality as
above.
Assuming non-unit fund growth is 5% per annum, and ignoring expenses, calculate
the non-unit fund cash flow for the first year of the policy if:
(a) the full value of capital units is held

(b) only the actuarially funded value of capital units is held [6]
10 The following 3 state model is used to price various sickness policies. The forces of
transition s, r, m and u depend only on age.
sx
H: healthy S: sick
rx
mx ux
D: dead
The following probabilities are defined:
p ijx,t is the probability that a life aged x in state i will be in state j at age x + t;
piix,t is the probability that a life aged x in state i will remain in state i until age x + t;
t p ijx, z is the probability that a life aged x in state i will be in state j at age x + t, having
been in state j for period z.
105 A2002—4
Using these probabilities and / or forces of transition, write down an expression for
the expected present value of each of the following sickness benefits for a life
currently aged 35 and healthy. The constant force of interest is d.
(a) £1,000 per annum payable continuously while sick, but all benefits cease at
age 65
(b) £1,000 per annum payable continuously while in the sick state for any
continuous period in excess of a year. However, any benefit period is limited
to 5 years payments, but the number of possible benefit periods is unlimited
(c) £1,000 per annum payable continuously throughout the first period of sickness
only [6]
11 (i) A life insurance company prices endowment assurance policies allowing for
mortality, expenses and interest. For surrenders, it wants to base the values it
is prepared to pay on gross premium reserves, using the premium basis
unchanged except for the interest rate. If it is to make a profit on surrenders,
state in what direction it should change the interest rate element of the basis, if
the reserves are:
(a) retrospective
(b) prospective
Give reasons for your answers. You should assume that experience is the
same as the premium basis. [4]
(ii) A policyholder aged exactly 60 has 5 years remaining on his endowment

policy which has a sum assured of £100,000 payable immediately on death, or
maturity, whichever occurs first. He can no longer afford to pay any further
premiums. He is offered a choice of:
(a) a surrender value of £41,000
(b) a paid up sum assured of £54,000
(c) a whole life policy, without future premiums, with a death benefit,
payable immediately on death, of £100,000
Show which he should choose, assuming he wants the one with the highest
expected present value of benefits.

[Total 8]

12 A small employer decides to set up a pension scheme for his 2 employees, who are
described by the following details:
Age Past service Expected salary

(exact) (years) in next year (£)
30 5 25,000
35 6 20,000
The scheme will provide a pension of 1/60th of pensionable salary for each year of
service (fractions of a year counting proportionally) on retirement for any reason.
Pensionable salary is the average annual salary earned in the final 36 months of
employment.
The employer meets the full cost of the scheme. The contribution rate is determined
by equating the expected present value of the total scheme liabilities to the expected
present value of contributions. Contributions are calculated to be a constant
percentage of the total salaries of the members at any time.
(i) Using the symbols defined in, and assumptions underlying, the Formulae and
Tables for Actuarial Examinations, calculate the contribution rate required for
the scheme. Ignore the possibility of new members joining the scheme. [8]
(ii) Immediately after the scheme is set up, a new employee joins the company
and pension scheme. She is aged exactly 40, and will earn £30,000 in the next
year. The employer decides to maintain the contribution rate determined in
part (i) and to apply it to the new total salaries. Determine whether the funding
rate is sufficient to meet the liabilities of the extended membership. [3]
[Total 11]
13 100 people aged exactly 50 are each sold a 15-year endowment assurance policy with
sum assured £100,000. The premiums are paid annually in advance, and the sum
assured is paid on maturity or at the end of the year of earlier death.
The life insurance company’s assumptions are:
Mortality: A1967–70 Ultimate, and the lives are independent with respect to
mortality
Expenses: Initial: £300

Renewal: 2.5% of each premium, including the first
Let P be the gross annual premium.
(i) State the gross future loss random variable for one policy at the outset. [3]
105 A2002—6
(ii) Using your answer to part (i) or otherwise, evaluate, in terms of P,
(a) the mean and variance of the loss (in present value terms) for a single
policy at outset
(b) the mean and variance of the loss (in present value terms) for the entire
portfolio at outset. [7]
Note: A50:15 at 12.36% per annum = 0.20426
(iii) Show what values the gross annual premium P can take if the company
requires that the probability it incurs a loss (in present value terms) on the
entire portfolio has to be less than 2.5%. Use the Normal approximation. [4]
[Total 14]
14 A life insurance company issues a number of 3-year term assurance contracts to lives
aged exactly 60. The sum assured under each contract is £200,000, payable at the end
of the year of death. Premiums are payable annually in advance for the term of the
policy, ceasing on earlier death.
The company carries out profit tests for these contracts using the following
assumptions:
Initial expenses: £200 plus 35% of the first year’s premium
Renewal expenses: £25 plus 3% of the annual premium, incurred at the beginning of
the second and subsequent years
Investment return: 7% per annum
Reserves: One year’s office premium
(i) Show that the office premium, to the nearest pound, is £4,396, if the net
present value of the profit is 25% of the office premium. [10]
(ii) Calculate the cash flows if the company held zero reserves throughout the
contract, using the premium calculated in part (i). [2]
(iii) Explain why the company might not hold reserves for this contract and the
impact on profit if they did not hold any reserves. [3]
[Total 15]
105 A2002—7
REPORT OF THE BOARD OF EXAMINERS ON

THE EXAMINATIONS HELD IN
April 2002
Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however
given credit for any alternative approach or interpretation which they consider to be
reasonable.
K Forman
Chairman of the Board of Examiners
11 June 2002
Page 1
EXAMNINER’S COMMENTS
The overall standard of scripts was better than in recent sittings. However answers were
very disappointing for questions 5, 10 and 13(iii) in particular, where the question posed a
problem not seen in recent examinations. It is also clear that many candidates’ statistical
knowledge or understanding is not up to the standard required. Finally candidates are urged
to read the questions carefully. In many cases the answers for questions 5, 8, 9, 10 and 12
omitted elements asked for or added details not required for the question.
Comments on individual questions follow after the solution to each question.
Page 2
1 3|4 q [40]+1is the probability that a life now aged 41, who entered the population of
interest a year ago subject to select mortality at that time, will survive for 3 more
years, and die during the following 4, when aged between 44 and 48.
= ( 3 p [40]+1 )( 4 q 44 ) for a 2 year select table
l44 l44 - l48 l44 - l48

= . = = (33,309.271 – 32,934.221) / 33,484.739
l[40]+1 l44 l[40]+1
= 0.0112006
Comment on Question 1
Well answered, with only a small minority of candidates mixing up the survival and death
periods.
1
2 10|20 q35:35 = (.5)( 10|20 q 35:35 ) = .5[ 10 p 35:35 {1 - 20 p 45:45 }]
= .5[ 10 p 35 2 - 30 p 35 2] = .5[(e-.2)2 – (e-.6)2] = .5(.67032 - .30119) = .1846
Answers were generally of a reasonable standard. The commonest errors related to the
factor of .5, and a range of errors in evaluating the required integrals.
3 Total fertility rates summarise the age specific fertility rates fx (i.e. the ratio of births
to population of women aged x generating them). The summation is over all ages for
which fx > 0, often taken as 15-49.
Cohort: fertility rates are summed (over a period of time) for women born in a
specified period e.g. all those born in the same calendar year
Period: fertility rates are summed at a point of time (e.g. the rates experienced in one
calendar year) for women of different ages
Cohort rates are generally preferred for their greater stability and their smooth rate of
change over time
or
Period rates are quicker and easier to obtain, and therefore suitable for immediate use
Other sensible reasons also gained credit.
Good standard, although some candidates mixed up cohort and period rates while others
provided formulae that dealt with numbers of births rather than fertility rates.
Page 3
4 Logistic model
{1/P(t)}{dP(t)/dt} = r - kP(t) or also P(t) = [Ce-rt + (k/r)]-1 or = r/[Cre-rt + k]
Current rate: r - k100,000 = .25

Limiting population: r - k300,000 = 0
leading to k = (.25/200,000) = 1/800,000 and r = .375 (=3/8)
We want t such that P(t) = 200,000
From t = 0 (now) 100,000 = [C + (1/300,000)] -1 so C = 2/300,000 or 0.00000666667
200,000 = [(2/300,000) e-.375t + (1/300,000)] –1
(1/200,000) - (1/300,000) = (2/300,000) e-.375t
e-.375t = (1/4) so using logs t = 3.70 years
Overall standard was quite good, although a surprising number of candidates did not seem to
know the logistic model at all. The commonest error was to use r = 0.25.
5 Insurance company received P so guaranteed maturity benefit

= [(1.05)4 ] * P = 1.21550625P
The company invests P @5.25% so is due to receive 1.2915479P in 5 years.
On death, the office breaks even because it pays out exactly the value of asset
available. This occurs with probability 4q56 = (1 - [l60/l56]) = 0.0690
At maturity (t = 4) office loses money only if yields at the time are j such that
{1.2915479P / (1 + j)} < 1.21550625P i.e (1 + j) > 1.06256
Prob (1 + j > 1.06256) for lognormal (1 + j)

= Prob (z > [Ln 1.06256 – 0.05] / 0.01) from standard normal
= Prob (z >1.07) = 1 - .85769 = 0.14231
Maturity occurs with probability 1 - .0690 = .9310 so the overall probability of a loss
is 0.9310 * 0.14231 = 0.1325 = 0.13
Very poor standard of answers. Many made no reasonable attempt. Of those who did, some
tried to calculate a surrender profit or loss, even though this was clearly zero. Many tried to
calculate the value of the zero coupon bond at the end of 4 years (one year short of
redemption) by considering the distribution of (1+i)4 and accumulating rather than using the
distribution of 1+i directly and discounting.
Page 4
6 Pension A
EPV = (20,000/1.009709) (a4 + v44 p[60] a64) @ (1.04/1.009709) = 3%
= (20,000/1.009709)[3.7171 + (.88849)(.947214)(11.962 -1)] = 256,363
Pension B
EPV = 12,000 a [60] + 1,000(Ia) [60] = 12,000 a [60] + (1,000S[60]+1/D[60])
= 12,000[12.710 - 1] +1,000 [307,254.58/2,815.3028] = 249,657
Well answered. Common mistakes in A were not getting initial level correct (missing divisor
of 1.009709) and using 4% interest for the deferred period until life annuity commences. In
B, many evaluated (Ia)[60] using S[60].
7 Nutrition influences morbidity and (in longer term) mortality.

Lack of nutrition leads to general weakening of body
Poor quality increases risk of disease / hinders recoveries.
Excessive / inappropriate can lead to obesity and associated diseases (e.g. hypertension,
heart disease). This can arise from social factors e.g. ready processed food / fast food etc.
Poor / lack of nutrition can arise from adverse economic circumstances.
Education (covering formal and also general awareness from public health
campaigns).
It influences awareness of elements of healthy lifestyle. This can affect behaviour in
many areas e.g. nutrition / diet; personal health and hygiene; awareness of effects of
tobacco, alcohol, drugs;
Education level will also have a bearing on income level, occupation , standard of
housing and general lifestyle, all of which are themselves regarded as influencers of
mortality..
Other reasonable points also received credit.
Well answered overall. Some candidates were inclined to repeat the same point rather than
identifying distinct influences on mortality.
8 (i) Age last birthday = x at start of rate interval in which dies

Curtate duration = r at start of rate interval in which dies
No assumptions needed
x + .5 at mid-point of interval, with duration r +.5 so we are estimating

m[x+.5-(r+.5)]+r+.5 = m[x-r]+r+.5
This does require an assumption of an even spread of retirements over the year
of age, because we have no other information about ages at entry (we can only
deduce that they can range from x - r + 1 to x - r -1).
Page 5
(ii) Retired aged 62 years and 3 months. Dies aged 64 years and 11 months so
total exposure is 2 years 8 months.
(62,0) 9 months
(63,0) 3 months
(63,1) 9 months
(64,1) 3 months
(64,2) 8 months
Very poorly answered, especially part (ii). Some otherwise correct answers omitted
assumptions completely while others gave “standard” assumptions e.g. policy anniversaries
spread evenly over the year of age when there are no policies (only retirements). Overall, the
understanding of the different rate intervals and the associated assumptions seems confused.
In part(ii), many candidates calculated the total exposure incorrectly, including in some
cases not even calculating the age at death correctly.
9 (a)
Capital units no actuarial funding
Year Cost of Fund at end Management Fund at end

investment before m.c. Charge
1 969 1,041.67 52.08 989.59
Non-unit fund
Year Premium Interest Death cost Management Cashflow

less cost of charge
allocation
1 31 1.55 0 52.08 84.63
(b)
A funding factors
A[61]:4 0.85697
A[61]+1:3 0.89045 from (M[61]+1 - M65 + D65) / D[61]+1
= (1,337.8829 - 1,258.7316 + 2,144.1713) / 2,541.7641
Page 6
A funded capital unit fund
Year Cost of Fund available Fund needed Management

investment at end at end Charge
1 830.40 892.68 881.18 11.50
Non-unit fund
Year Premium Interest Death cost Management Cashflow

less cost of charge
allocation
1 169.60 8.48 0.78 11.50 188.80
The death cost is q[61]*(full capital unit fund – A funded capital unit fund @ t = 1)
i.e. 0.00723057*(989.59-881.21)
Handled very well. Errors, where they occurred, were to include a death cost in (a), use of
the wrong funding factor at t=1 and incorrect calculation of the death cost in (b). Some
candidates completed a full profit test for each year of the contract, wasting valuable time.
30 -dt HS
10 (a) EPV = 1, 000ò e p35,t dt
0
¥ 6 -dt HS
(b) EPV = 1,000 ò ò e t p dzdt
0 1 35, z
or
¥ HH 6
1,000ò e -dt p s 35+t ò e -dr p35
SS
+t ,r drdt
0 35,t 1
¥ HH ¥
(c) EPV = 1,000 ò e -dt p s35+t ò e-dr p35
SS
+t ,r drdt
0 35,t 0
or
¥ HH ¥
1,000ò e -dt p SS
s 35+t ò a r| p35+t ,r ( r 35+t + r + n 35+t + r )drdt
0 35,t 0
This was a testing question that was not answered well at all. Most attempted (a) but often
got it wrong, while very few candidates made any real attempt at (b) and (c), even though (b)
in particular just required direct use of a formula given in the appropriate Core Reading.
Where an attempt was made at (b) or (c), candidates often used the benefit ceasing age for
(a), although none applied in (b) or (c).
Page 7
11 (i) The retrospective and prospective reserves equal each other on the premium
basis. We want the SV calculation to result in a lower reserve.
Retrospective reserve needs to be done at a smaller interest rate, as it is

accumulating past excess premiums over claims/ expenses.
Prospectively, the interest rate needs to be higher than the premium basis, so
that the discounting of the excess of future outgo (claims / expenses) over
premium income results in a lower answer.
(ii) SV = 41,000
1 1
PUPSA = 54,000 A60:5 = 54,000((1.06).5 { A60:5 - A }+ A )
60:5 60:5
1
where A = v5 5 p60 = (.747258)(27,442.681/30,039.787) = .68265
60:5
EPV of PUPSA = 54,000(1.02956{.75477 - .68265} + .68265) = 40,873
Whole Life option 100,000 A60 = (1.06).5 (.39136) = 40,293
So SV is best
Well answered. Those who got (i) wrong often wrestled with reserving formulae rather than
considering the underlying concept needed. In part (ii), a surprisingly large number of
candidates overvalued the paid up option by multiplying the entire endowment factor by
1.060.5 rather than just the death element.
Page 8
12 EPV of past pensions: (n/60)(Sal)( z M xia + z M xra )/sDx
z ia z ra
EPV future pensions: (1/60)(Sal)( R x + R x )/sDx
s
EPV of contributions @ 1% of salary: (.01)(Sal)( N x )/sDx
s
age salary past service Dx z
M xia z
M xra z ia z ra s
Rx Rx Nx
30 25,000 5 28,043 8,636 88,345 231,941 2,915,486 540,020
35 20,000 6 22,276 8,513 88,345 188,977 2,473,760 417,224
EPV past pension EPV future pension EPV cont. 1%sal
7,204.78 46,764.89 4,814.21

8,696.18 39,844.63 3,745.95
Total 15,900.96 86,609.52 8,560.16
Total Liability = 15,900.96+86,609.52 = 102,510.48
Contribution rate needed = 102,510.48/8,560.16 = 11.98% of salary
New employee
s
Age salary past Dx z ia z ra s
Rx Rx Nx
service
40 30,000 0 18,629 147,045 2,032,033 317,121
EPV EPV EPV

past future cont.
pension pension 1%sal
0.00 58,486.18 5,106.89
Contribution needed = 58,486.18 / 5,106.89 = 11.45%.
Therefore the contribution rate of 11.98% established for the original 2 members is
more than that required to meet the costs of the new entrant, and the scheme is in
surplus.
Answered very well, but a disappointing number of candidates overlooked the ill-health
retirement benefits.
Page 9
13 All values at t = 0
(i) Future loss random variable =

100, 000v min( K50 +1,15) + 300 - .975 Pa&&min( K
50 +1,15)
1 - v min( K50 +1,15)

= 100, 000v min( K50 +1,15) + 300 - .975 P
d
(ii) (a) Mean for single policy just take expected value of random variable
X = EV one policy = 100,000 A50:15 + 300 - .975 P a&&50:15
= (100,000)(.44395) + 300 -(.975)P(9.823)
= 44,695 – 9.577425P
Y = Variance one policy
Variance = (using 2nd form of loss r.v.)
= [100,000 + (.975P/d)]2Var ( v min( K50 +1,15) )

2 2
= [100,000 + (.975P/d)]2 ( A50:15 - [ A50:15 ] )
where the 2 superscript denotes at i2 + 2i
= [100,000 + (17.225P)]2(.007168397)
or Standard Deviation = (100,000 + 17.225P)(.084666)
(b) 100 policies
Mean = 100X
Variance = 100Y assuming the lives are independent or
Standard Deviation = 10 Std Dev above
Page 10
(iii) Using Central Limit Theorem (n = 100) we can assume normality of portfolio
loss.
We want Prob (loss > 0) < .025
Prob ([loss – mean]/ Std Dev > [0 - mean]/std Dev) < .025
Prob (z > - mean/Std Dev) < .025
This means that (- mean / Std Dev) > 1.96 or (mean/ std Dev) < -1.96
(100)(44,695 - 9.577425 P) < (-1.96) (84,666 + 14.5837P)
P ³ 463,5445/929.158448 = 4,988.86 say 4,989
A very mixed standard. It is clear some candidates do not have a good understanding of the
difference between a random variable and its expectation, at least in this context. Common
errors in (i) were to use assurance or life annuity functions, to miss the +1 in the Kx+1 terms
or to give a profit (rather than loss) random variable.
In (ii), very few got the variance correct for a single policy, usually not making the
conversion of the annuity into (1-vn)/d format used in the model solution, and then missing
the cross-product or covariance term between the benefit and premium random variables. A
surprising number of candidates missed the independence of lives within the portfolio and
therefore concluded that the variance of the portfolio was 1002 times the variance of one
policy.
In (iii), of the few candidates who attempted this part, many started with considering a
loss < 0, when the loss has to be > 0 to be a loss.
Page 11
14 (i)
age qx px t-1px
60 0.01443246 0.98556754 1
61 0.01601356 0.98398644 0.98556754
62 0.01774972 0.98225028 0.96978510
Year Prem. Expense Opening Interest Death Closing Profit vector Profit NPV
reserve claim reserve signature
1 P 0.35P 0 0.0455P 2,886.49 0.985568P -0.290068P -0.290068P -0.252233P

+200 -14 -3,100.49 -3,100.49 -2,696.08
2 P 0.03P P 0.1379P 3,202.71 0.983986P 1.123914P 1.107693P 0.837575P

+25 -1.75 -3,229.46 -3,182.85 -2,406.69
3 P 0.03P P 0.1379P 3,549.94 0 2.107900P 2.044210P 1.344101P

+25 -1.75 -3,576.69 -3,468.62 -2,280.67
1.929443P
-7,383.44
Therefore 1.929443P - 7383.44 = .25P Þ Premium = 4,396.36 = 4,396.
(ii) If we use this premium, and ignore reserves, the cash-flows per policy in force
at the start of each year are (-43, 1,334, 986).
(iii) As the cash flows in years 2 and 3 are all positive, there is no need to establish
reserves at the end of any year.
In such a scenario, the profits emerge earlier and because the discount rate
exceeds the earned rate of interest, the NPV increases.
Answered well overall. The most common error was mishandling of reserves. A
disappointing number of students started from a commutation function approach when a
cashflow model was needed. In (iii), a number of candidates made the general statement that
it was not necessary to hold reserves for term assurance contracts because the probability of
death was low, without any reference to the specifics of the cashflows in this case.
Page 12
EXAMINATIONS
booklet.
supervisor.
question paper.
1 Explain in words what is meant by the logistic model for projecting the size of a
population.
Write down a differential equation whose solution gives a formula for the size of a
population based on the logistic model.
Define carefully all symbols that you use. [3]
2 Describe how option pricing techniques may be used to determine the value of the
guarantee under a deferred annuity policy with a guaranteed minimum annuity. [3]
3 Define each of the following terms and give an example of each in life assurance
business:
(a) class selection

(b) spurious selection
(c) adverse selection [3]
4 Explain what is meant by the following, in the context of with profit life insurance
contracts:
(a) earned asset share

(b) retrospective valuation reserve [4]
5 A life insurance company issues a whole life assurance policy to a life aged exactly
60, paying a sum assured S, together with attaching bonuses, immediately on death.
Compound bonuses are added annually in advance. Premiums under the policy are
payable annually in advance, ceasing at exact age 85 (the last premium is payable at
age 84 exact) or on earlier death.
Write down an expression for the net future loss random variable at outset for this
policy. Define carefully all the symbols that you use. [4]
6 A life insurance company issues a temporary annuity policy to two independent lives,
each aged exactly 60. The annuity of £10,000 per annum is payable quarterly in
arrears, while at least one of the lives is alive. The annuity is payable for a maximum
of 10 years.
Basis: mortality: A1967–70 Ultimate

interest: 4% per annum
expenses: ignore [4]
105 S2002—2
7 Two lives, (x) and (y), are assumed to be independent with respect to mortality and
are each assumed to be subject to a constant force of mortality of 0.01. Calculate the
probability that (x) dies more than 10 years after (y). [4]
8 Calculate
2
A[20]:[20]
using A1967–70 mortality and interest of 4% per annum. [4]
9 Members of a pension scheme are subject to two modes of decrement namely death
(d) and withdrawal (w). The following assumptions are made in respect of the two
decrements:
Independent rate qxd is A1967–70 Ultimate;
Independent rate qxw is 0.05 per annum at age 20 last birthday and increases by
5% at each successive age attained. (For example, the annual rate of
t
withdrawal at age 20 + t last birthday is ( 0.05) ´ (1.05 ) );
the decrements are statistically independent;
each decrement is uniformly distributed in its single decrement table.
Calculate the probability that a new entrant aged exactly 20 will withdraw from the
scheme at age 22 last birthday. [5]
10 (i) Define, giving a formula, the term “Standardised Mortality Ratio”. Define all
the symbols that you use. [2]
(ii) Show how the Standardised Mortality Ratio may be expressed as a weighted
average, setting out clearly what function is averaged and what the weights
are. [3]
[Total 5]

11 A pension scheme provides the following deferred benefits for a member aged 55
exact, who leaves service before Normal Pension Age (NPA), which is age 65 exact.
(a) A deferred pension of £10,000 per annum, payable monthly in advance from
NPA. The pension is payable for a minimum of 60 monthly payments. The
pension increases monthly in deferment and payment at the effective rate of
3.846% per annum compound.
(b) On the death of the member after NPA, a dependant’s pension of 50% of the
member’s pension entitlement at the date of death. The pension is payable
monthly in advance beginning on the first day of the month following the date
of the member’s death, or the fifth anniversary of the member’s NPA, if later
and increases monthly in payment at the effective rate of 3.846% per annum
compound.
Calculate the expected present value of the deferred benefits.

Proportion with
dependants: 90% of members have dependants at the date of
retirement
Age difference: members are the same age as their dependants
(assume that females are treated exactly the same as
males)
Expenses: none [11]
12 A life insurance company issues a disability insurance contract to a healthy life aged
exactly 30. Under the contract, a benefit of £20,000 per annum is payable weekly in
the event of disability. The benefit continues to be payable during disability, until the
policyholder recovers or reaches age 65. The benefit increases continuously in
payment at the rate of 3% per annum compound.
There is no waiting period or deferred period. Premiums continue to be payable

during periods of disability.
Disability benefit payments are valued using rates of claim inception and termination.
(i) Describe the method of valuing disability benefit payments under this
contract, setting out the data required. [6]
(ii) Derive commutation functions for valuing the benefits payable under the
contract, stating clearly any assumptions that you make and defining carefully
all the symbols that you use. [7]
[Total 13]
105 S2002—4
13 A life insurance company issues a long-term care contract to a healthy life aged 50
exact. Under the contract, the life insurance company will pay the costs of long-term
care while the policyholder satisfies the conditions for payment.
The conditions for payment are assessed each year on the policy anniversary, just
before payment of the premium then due. If the policyholder satisfies the conditions,
the full annual amount of the benefit payable is paid immediately.
Regular premiums are payable annually in advance under the policy until death and
are waived during periods of benefit payment.
For those lives needing care at 100% of maximum, the current payment on the policy
anniversary is £50,000. The company uses the following data in respect of the
expected proportions of lives at each age needing care at different expected cost
levels, for pricing the long-term care contract.
Exact age Proportion needing Proportion needing

care at 50% of care at 100% of
maximum maximum
51–70 0.01 0.01

71–85 0.04 0.06
86+ 0.08 0.10

Benefit inflation: Maximum payment at 100% care level at policy
anniversary t(t = 1, 2,..) = £50,000 ´ (1.019231)t
Expenses: 10% of each premium
(i) Write down an expression for the expected present value of benefits (including
the waiver of premium benefit) at outset for the contract. Define carefully all
(ii) Calculate the annual premium payable under the contract. [10]
[Total 14]

14 A life insurance company issues a 2-year unit-linked endowment assurance contract
to a male life aged exactly 63, under which level annual premiums of £6,000 are
payable in advance throughout the term of the policy, or until earlier death.
102% of each year’s premium is invested in units at the offer price.
The premium in the first year is used to buy capital units, with the second year’s
premium being used to buy accumulation units. There is a bid-offer spread in unit
values, with the bid price being 95% of the offer price.
The annual management charges are 5% of the bid value of capital units and 1% of
the bid value of accumulation units. Management charges are deducted at the end of
each year, before death, surrender or maturity benefits are paid.
payable at the end of the year of death of £10,000, or the full bid value of the units
payable.
The policyholder may surrender the policy only at the end of the first policy year.
The surrender value is equal to 87% of the bid value of the capital units.
The life insurance company uses the following assumptions in carrying out profit tests
of this contract:

Surrender: 10% of policies then in force, occurring at the end of the
first policy year
Renewal: £100 at the start of the second policy year
Non unit fund interest rate: 4% per annum
(i) Calculate the net present value on this contract, assuming that the company
holds unit reserves equal to the full bid value of the accumulation units and
capital units. [12]
(ii) Assume that the company holds unit reserves equal to the full bid value of the
accumulation units and a proportion, A63+t:2-t (calculated at 4% and
A1967–70 Ultimate mortality), of the full bid value of the capital units
(t = 0, 1).
Calculate the net present value on the contract. [9]
(iii) Explain what the effect would be on the answers in parts (i) and (ii) if the
mortality assumption were changed to mortality of A1967–70 Select. [2]
[Total 23]
105 S2002—6
EXAMINATIONS
September 2002
EXAMINERS’ REPORT
Introduction
reasonable.
K G Forman
12 November 2002
EXAMINERS’ COMMENTS
The overall standard of attempts by candidates was high. A number of questions were
answered very well. The more challenging questions were less well answered, such as
Questions 2,6,7, 12 and 13, with evidence of lack of proper preparation. A common mistake,
which was also a feature of previous examinations, was to misread some of the questions.
Detailed comments are given after the solution to each question.
Page 2
1 A logistic model for projecting the size of a population is a model under which an
initial rate of growth for the population is assumed to decrease over time in proportion
to the size of the population.
The model may be expressed in the form
1 dP ( t )
= r - kP ( t ) , where
P ( t ) dt
P(t) is the size of the population at time t

r is the rate of growth, a constant, >0
k is a constant, >0.
Full credit was given for the solution set out. However, a fuller treatment would be to assume
that the initial rate of growth, r , may also be negative, in which case r may be assumed to
either increase or decrease over time in proportion to the size of the population. Few
candidates gave this fuller treatment.
The question was answered well in general. A number of candidates were vague in the
definition, omitting the point that the initial rate of growth decreased over time in proportion
to the size of the population.
2 A guaranteed annuity rate corresponds to a call option on the bonds that would be
necessary to ensure that the guarantee was met, i.e. at an exercise price that generated
the required fixed rate of return. Alternatively, it can be modelled by an option to
swap floating rate returns at the option date for fixed rate returns sufficient to meet the
guaranteed option.
It is difficult to ensure that the whole investment fund corresponds to a single option
traded in the market. An approximation is possible by using options written on
indices.
At the date of policy issue, all guarantees will be out of the money, i.e. they will have
no intrinsic value because current market rates are more than sufficient to meet the
guarantees, but will have a time value that is the result of the views of many investors
(“the market”) of the present value of the likely future costs of the option.
Thus the market price of a suitable option produces a way of pricing the guaranteed
annuity rates.
This question was poorly answered. Many candidates described other techniques rather than
option pricing techniques.
3 (a) Class selection is the process whereby lives are divided into separate groups,
within which mortality or morbidity is homogenous, where each group is
Page 3
specified by a category or class of a particular characteristic of the population.

An example in life assurance business is the use of individual rating factors
which produce mortality differences, e.g. smoking status.
(b) Spurious selection is the process whereby lives are divided into separate
groups, within which mortality or morbidity is homogenous, where the
differences in mortality or morbidity are due to factors other than those used to
form the groups. An example is a change in underwriting over time leading to
mortality improvements, where such improvements are ascribed to the passage
of time.
(c) Adverse selection is the process whereby lives are divided into groups that
tend to act against a controlled selection process imposed on the groups, in
respect of mortality or morbidity. An example is where smokers will tend to
select policies from a life office that does not use smoking status as a rating
factor.
This question was generally well answered. Credit was given for all reasonable examples.
Some candidates confused self-selection with adverse selection in part (c).
4 (a) The asset share for a with profit contract is the accumulation of premiums less
deductions associated with the contract plus an allocation of profits on non-
profits business, all accumulated at the actual rate of return earned on
investments. The deductions include expenses, claims, cost of capital and
transfer to shareholder funds, if relevant and are based on actual experience.
(b) The retrospective valuation reserve is the expected accumulation of past

premiums received, less expected expenses and benefits including any
reversionary or interim and terminal bonuses included in past claims.
This question was well answered. Some candidates omitted the allocation from non-profit
business in part (a).
5 The net future loss random variable is given by
K60 +1 T60
S (1 + b ) v - Pa&&min( K .
60 +1,25)
b is the annual rate of future bonus
K 60 , T60 are the curtate and complete future lifetimes of a life aged 60
P is the annual premium
This question was well answered. A common error was the inclusion of K 60 rather than
K 60 + 1 .
Page 4
6 The required single premium is given by
( 4) ( 4)
( ( 4)
10000a60:60:10| = 10000 2a60:10| - a60:60:10| )
( 4) D70
a
60:10|
= a60 + 83 -
D60
(
a70 + 83 )
1516.9972
= 11.551 + 83 -
2855.5942
(
7.957 + 83 )
= 7.500
( 4) æD ö
a60:60:10| = a60:60 + 83 - ç 70:70 ÷ a70:70 + 83 ( )
è D60:60 ø
æ 1039.0172 ö
= 8.943 + 83 - ç
è 2487.2117
÷ 5.498 + 8
ø
(
3
)
= 6.865
\ the single premium is 10000 * (2 * 7.500 - 6.865) = £81,350
This question was relatively poorly answered. Many candidates struggled with the correct
evaluation of the annuity factors.
7 The required probability is
¥
p y m y +t t +10 p x dt , where t p x = exp æç - ò 0.01ds ö÷
t
ò0 t
è 0 ø
¥
= ò e-0.01t *0.01* e-0.1-0.01t dt
0
¥
-0.1 é e -0.02t ù
= 0.01* e * ê- ú
ëê 0.02 ûú 0
= 0.5* e-0.1
= 0.45242
This question was answered less well than the examiners had expected, with many candidates
setting out the initial integral expression incorrectly.
Page 5
8 2
A[20]:[20] ( 1
= A[20] - A[20]:[20] )
(
= A[20] - 12 A[20]:[20] )
A[20 ] = 0.13312
A[20 ]:[20 ] = 1 - da&&[20 ]:[20 ]
= 1 – 0.038462*21.509
= 0.17272
2
\ A[20]:[20] = 0.13312 - 0.5*0.17272
= 0.04676
This question was well answered by most candidates.
9 Independent rates of withdrawal:
Age attained Rate

20 0.05
21 0.0525
22 0.055125
Probability of survival to age 22 =
l22 34029.283
0.95* 0.9475* = 0.95* 0.9475* = 0.898568
l20 34088.257
(aq )22w = q22w (1 - 12 q22d ) = 0.055125 * (1 - 12 * 0.00079739) = 0.055103
Required probability = 0.898568*0.055103 = 0.049514
in place of (aq )22 .

w w
This question was well answered. Some candidates used q 22
10 (i) The Standardised Mortality Ratio is the ratio of the actual deaths in a
population compared with the expected deaths, based on standard mortality
rates.
The formula is
å Exc,t mx,t
x
, where
å Exc,t s mx,t
x
E xc,t is the central exposed to risk in the population between ages x and x + t
Page 6
mx,t is the central rate of mortality for the population between ages x and x +
t
s
mx,t is the central rate of mortality for a standard population between ages x
and x + t
(ii) The Ratio may be written in the form
m
å Exc,t s mx,t s mx,t
x x ,t
å Exc,t s mx,t
x
which is the weighted average of the age-specific mortality differentials

between the population being studied and the standard population.
mx,t
i.e. s
mx,t
weighted by the expected deaths in the population being studied based on

standard mortality.
i.e. Exc,t s mx,t
Part (i) was well answered in general. A common error was not basing expected deaths on
the mortality of a standard population.
Part (ii) was very poorly answered, with few candidates obtaining full marks.
11 (a) The net rate of interest is 4% per annum in deferment and payment.
The value of the deferred pension in payment is
D65 æ (12 ) D70 (12 ) ö

10000* * ç a&& + * a&& ÷
D55 è 5| D65 70 ø
Page 7
D65 2144.1713
= = 0.585109
D55 3664.5684
(12 ) i
a&&5| = * a5 = 1.021537 * 4.4518 = 4.5477
(12)
d
D70 1516.9972
= = 0.707498
D65 2144.1713
(12 )
a&&70 = 8.957 - 0.458 = 8.499
Required value = 10000*0.585109*(4.5477+(0.707498*8.499)) = 61792
(b) The value of the pension on death in retirement is
D65 æç D70 æl æ l70 ö (12 ) ö ö÷

0.9 * 0.5 *10000 * * * çç 70 * a&&70
(12 )
ç
|70 + ç1 -
÷÷ * a&&70 ÷
÷÷
D55 çè D65 l
è 65 è l 65 ø øø
a&&70
(12 )
|70 = a 70 | 70 = a 70 - a 70 : 70 = 7.957 - 5.498 = 2.459
l70 23622.102
= = 0.8607797
l65 27442.681
Required value =
(
0.9*0.5*10000*0.585109* 0.707498* ( 0.8607797*2.459 + (1 - 0.8607797 ) *8.499 ) )
= 6147
Total value = £67,939
Attempts at valuing the benefits in part (a) were reasonable in general. Attempts at valuing
the benefits in part (b) were very poor.
12 (i) The approach uses two double decrement tables. One table relates to healthy
policyholders and decrements of falling sick and dying. Recovery and
subsequent rates of sickness are allowed for in the table. The table is used to
calculate probabilities of surviving to be a healthy policyholder at age 30+t,
(al )30+t
, 0 £ t £ 35 . The table is also used to calculate the dependent initial
(al )30
rate of falling sick at age 30+t, (aq )30+t , 0 £ t £ 35 . The rate (aq )30+t is called
the inception rate for disability.
The second table relates to policyholders receiving disability benefits and has
decrements of recovery from disability and dying (while disabled). The
survival probabilities from this double decrement table are used, together with
an appropriate interest rate, to determine the present value at the date of
becoming disabled of a disability annuity of £20,000 per annum, increasing in
payment continuously at the rate of 3% per annum compound and payable
Page 8
according to the policy conditions until the policyholder dies while disabled,
recovers or reaches age 65.
The probabilities of surviving as a healthy policyholder to age 30+t, the

inception rates for disability at age 30+t and the disability annuity payable
from age 30+t are calculated for each value of t and are integrated or summed
over the range 0 £ t £ 35 .
The data required are those set out in the two multiple decrement tables above,
for ages from 30 to 65.
(ii) The value of a disability benefit of £1 p.a. payable weekly to a healthy life
now aged 30 exact is
t =35
( al )30+t
ò ( al ) ( am )30+t v ( 20000* a30+t ) dt , where
i t i¢
t =0 30
i
( al )30+t , ( am )30 +t
are based on a double decrement table for healthy
policyholders described in part (i) above and gives the number surviving to
age 30 + t while healthy and the force of inception of disability at age 30 + t.
i¢
a30 +t is a continuous annuity based on the second double decrement table
described above, evaluated at rate of interest i¢ , where
1+ i
i¢ = -1
1.03
Assume that lives becoming disabled in ( 30 + t ,30 + t + 1) do so on average at

age 30 + t + 12 and the integral is approximated by
t =35
å
( ad )i30+t v
30+t + 12
( 20000* a
i¢
30+t + 12 ) , where
t =0 ( al )30 v 30
( ad )i30+t is the number of lives becoming disabled at age 30 + t last birthday

in the first decrement table described in part (i) above.
Page 9
Define commutation functions as follows
i 30 12 i¢
¢ia+t = ( ad )30+t v
C30 a30+t + 1
2
D30 = ( al )30 v30

t =¥
¢ia = å C30
M 30 ¢ia+t
t =0
So the value is approximated by
¢ia
M 30
20000*
D30
Poorly answered in general. The evidence was that many candidates were not well prepared
for this topic, with some candidates not attempting the question and others giving very poor
answers. For a well-prepared candidate, the question should have been relatively
straightforward and a small number of candidates did achieve high marks.
13 (i) The expected present value of benefits at outset is given by
¥ ìï D 4% é 2 c ùü ¥ ì 6% é 2 c ù üï
ï
50+t
( I Lc ï
)
å í D50 êêå 50+t úú ý å í D5050+t
+ P
ïD
ê å I 50+t ú ý ,
êë c =1 úû þï
t =1 î ë c =1 û þï ï
t =1 î
using A1967-70 Ultimate
c
I 50 +t is the proportion of policyholders needing care at exact age 50+t, at
benefit level c = 1, 2
Benefit levels: c = 1 is the benefit level at 50% of maximum

c = 2 is the benefit level at 100% of maximum
L1 = £25, 000, L2 = £50, 000
P is the annual premium
Page 10
(ii)
Exact age Proportion Proportion Total Expected
needing care needing care proportion payment
at 50% of at 100% of per current life
maximum maximum
51-70 0.01 0.01 0.02 750
71-85 0.04 0.06 0.10 4000
86+ 0.08 0.10 0.18 7000
Present Value of Long Term Care Benefits (all calculated at 4% interest):
D50 = 4597.0607
Exact age Expected ( N x - N x +t +1 ) / D50 Value

(x to x+ t) payment
per current life
A B A*B
51-70 750 12.37730 9282.98
71-85 4000 2.44004 9760.16
86+ 7000 0.18574 1300.18
Total Present Value 20343.32
Present value of Waiver Payments (all calculated at 6% interest)
Present value is = P*{(0.02*(D51+D52+….+D70)+0.1*(D71+D72+….+D85)

+0.18*(D86+D87+……)}/D50
This can be rewritten as:
PV = P*{0.02*a50+0.08*v20*(l70/l50)*a70+0.08*v35*(l85/l50)*a85}
= P*(0.02*12.120+0.08*0.31180*0.723055*7.018
+0.08*0.13011*0.20712*3.297)
= 0.376084*P
So the final premium equation allowing for expenses is:
0.9*P* a&&50 (at 6%) = 20343.32+0.376084*P
i.e. P = 20343.32/(0.9*13.120-0.376084)
i.e. P = £1,779.52
The examiners expected this to be a challenging question. It required the application of basic
actuarial techniques to pricing a product that was probably relatively unfamiliar to most
candidates. A number of candidates performed very well, achieving full, or near to full,
Page 11
marks. However, the majority of candidates performed poorly, with part (i) being better
answered than part (ii).
14 (i)
Multiple decrement table
x ( al ) x
qxd qxs ( ad )dx ( ad )sx
63 0.01965464 0.1 100000 1965.464 9803.454
64 0.0217431 0 88231.08 1918.417 0
65 86312.67
Unit Fund
Year, t 1 2
Value of Capital units at start 0 5965.164

Premium to Capital units 5814 0
Interest on Capital units 465.120 477.213
Management charge on CUs 313.956 322.119
Value of Capital units at end 5965.164 6120.258
Value of Accumulation units at start 0 0

Premium to Accumulation units 0 5814
Interest on Accumulation units 0 465.120
Management charge on Aus 0 62.791
Value of Accumulation Units at end 0 6216.329
Total value of units 5965.164 12336.587

Surrender value 5189.693 0
Non-unit Fund
Unallocated premium 186 186

Expenses 500 100
Interest -12.56 3.44
MC on Capital units 313.956 322.119
MC on Accumulation units 0 62.791
Surrender profit 76.023 0
Extra death benefit 79.303 0
End of year cash flow -15.884 474.350
Probability in force 1 0.882311

Profit signature -15.884 418.52
Discount factor 0.869565 0.756144
Expected present value -13.812 316.465
Net present value 302.65
Page 12
(ii)
Unit Fund
Year, t 1 2
Actuarial funding factor 0.92528 0.96154

Value of Capital units at start 0 5735.744
Premium to Capital units 5379.578 0
Interest on Capital units 430.366 458.860
Management charge on CUs 290.497 309.730
Value of Capital units at end 5519.447 5884.873
Total value of units 5519.447 12101.202*

Surrender value 5189.693 0
*including accumulation units
Non-unit Fund
Unallocated premium 620.422 186
Expenses 500 100
Interest 4.817 3.440
Surrender profit 32.327 0
Extra death benefit 88.064 5.118
MC on Capital units less cost of additional
99.656 79.463
allocation
MC on Accumulation units 0 62.791
End of year cash flow 169.158 226.576

Profit signature 169.158 199.91
Discount factor 0.869565 0.756144
Expected present value 147.094 151.161
Net present value £298.25
(iii) If A1967-70 Select mortality were used in the profit tests instead of A1967-70
Ultimate mortality, the cost of the extra death benefit would decrease and, separately,
the profit signature would increase. The effect of these two factors would be to
increase the net present value of profit in part (i) and part (ii).
Part (i) was generally well answered. A surprising number of candidates calculated the
probability of being in force for year 2 incorrectly.
A number of candidates achieved full marks for part (ii). However, in general this part was
less well answered than part (i). Common errors were the incorrect calculation of the
surrender profit, extra death benefit and management charge on capital units less the charge
of additional allocation.
Page 13
Part (iii) was well answered. A number of candidates described the effect of changing the
mortality basis from Ultimate to Select in the actuarial funding factors, in addition to the two
factors given in the solution. Full credit was also given for this approach.
Page 14
EXAMINATIONS
9 April 2003 (am)
booklet.
supervisor.
question paper.
105—A2003 ã Institute of Actuaries
1 In the context of Manchester Unity sickness functions, state the relationship between
26|26 26|26
s26 and z26 . [2]
2 (i) In the context of with profit policies, describe the super compound method of
adding bonuses. [2]
(ii) Give a reason why a life insurance company might use the super compound
method of adding bonuses as opposed to the compound method. [1]
[Total 3]
3 Under a policy issued by a life insurance company, the benefit payable on death, at
the end of the year of death, is a return of premiums paid without interest. A level
premium of £1,500 is payable annually in advance, throughout the term of the policy.
For a policy in force at the start of the tenth year, you are given the following
information:
Reserve at the start of the year, 9V: £11,300

Reserve at the end of the year per survivor, 10V: £13,200
Probability of death during the year: 0.04
Rate of interest earned: 5% p.a.
Calculate the profit expected to emerge at the end of the tenth year per policy in force
at the start of that year. Ignore expenses and all decrements other than death. [3]
4 Compare the use of the component method and the logistic mathematical modelling
method for projecting the size of the population in a certain country. [4]
5 A researcher into international mortality experience is interested in comparing death

rates in different countries by cause of death (cancer, heart disease, accidents etc.).
An initial study compares crude death rates by cause of death for each country, and
indicates a wide range of experience among the different countries.
(i) Comment on the approach of using crude rates for this comparison, indicating
any advantages and disadvantages of this method. [2]
(ii) Suggest an alternative approach which addresses any shortcomings identified

in (i). You should assume any data required are available. [2]
[Total 4]
105 A2003—2
6 A life insurance company sells disability insurance contracts, under which the benefit
is £100 per week, payable while a life insured is alive, disabled and aged not more
than 65. It calculates premiums and reserves using the inception rate / disability
annuity methodology.
Calculate the expected present value of future benefit payments for the following two
policyholders:
(a) A 45 year old who is healthy at the valuation date, and whose policy has a
deferred period of one year. The value should take into account all possible
future periods of sickness claims.
(b) A 55 year old who has been receiving benefit payments for the last two years.
The value should allow only for the remaining payments under the current
sickness claim.
Basis: Interest: 6% per annum

Morbidity & Mortality: S(ID) in the Actuarial Tables [4]
7 A life insurance company issues a reversionary annuity policy to a male and a female,
both of whom are aged exactly 60.
The annuity commences immediately on the death of the first of the lives to die and is
payable subsequently while the second life is alive, for a maximum period of 20 years
after the commencement date of the policy.
The annual amount of the annuity is £10,000 and is payable continuously.
Calculate the single premium for the policy.
Basis: Mortality: PMA92C20 for the male life and PFA92C20 for the female life.
The lives are independent with respect to mortality.
Expenses: Initial: £300 incurred at the outset
Annuity: 2% per annum of the annuity payment, incurred

continuously while the annuity is being paid
[7]

8 An insurer sells a special 3-year single premium non-profit term assurance policy for
an initial sum assured of £250,000. This policy includes an option such that the
policyholder can double the sum assured at the end of the second year of the policy by
paying an additional premium at that time, based on normal mortality rates, without
evidence of health.
All death benefits are payable immediately on death.
The company uses the North American method for pricing this policy.
Calculate the premiums payable by a female life aged exactly 55 at the outset who
does take up the option.
Basis: Normal mortality: ELT15 (Females)
Mortality of those who

exercise the option: 300% of ELT15 (Females)
Expenses: None
Proportion of policyholders 40% of those alive on the second

who exercise the option: policy anniversary
[8]
9 (i) Explain why a life insurance company might need to set up non-unit reserves
in relation to a unit-linked assurance contract. [3]
(ii) A ten-year contract has the following profit signature before non-unit reserves
are set up:
(-1, 0, +1, -2, +1, +1, 0, -1, 0, +1)
If positive non-unit reserves are set up to zeroise negative cash flows, write
down the revised profit signature. You should ignore interest. [2]
(iii) State the advantages of cash flow techniques for product pricing compared
with traditional commutation functions. [3]
[Total 8]
105 A2003—4
10 A member of a pension scheme is aged exactly 40, having joined the scheme at age
exactly 22. He earned £30,000 in the immediately preceding 12 months. Final
pensionable salary is defined as the annual average earnings over the three years
immediately prior to retirement. Normal Retirement Age is a member’s 65th
birthday.
Using the functions and symbols defined in, and assumptions underlying, the
Example Pension Scheme Table in the Actuarial Tables, calculate the expected
present value of each of the following:
(i) A pension on ill-health retirement of two-thirds of final pensionable salary. [3]
(ii) A pension on retirement at any stage on grounds other than ill-health of one-
eightieth of final pensionable salary for each year of service (fractions of a
year counting proportionately), subject to a maximum of 40 years. [3]
(iii) A lump sum on retirement at any age for any reason of £50,000. [3]
[Total 9]
11 In a select mortality investigation, tabulations of in force populations are available for

a certain class of business, in the following 2 ways:
On each of 1 January 2000, 2001 and 2002, Px,t is available where x and t are defined
as:
Method x t
A Age last birthday Curtate duration
B Age next birthday at Duration at policy
issue plus calendar year anniversary during year of
of census minus census
calendar year of issue
Two different tabulations of deaths in each of the years 2000–2002 are also available,
qy,r where y and r are defined as:
Method y r
1 Age last birthday at Duration at policy
policy anniversary prior anniversary following
to death death
2 Age last birthday at Curtate duration at death
death
These data are to be used to estimate select forces of mortality. For each tabulation of
deaths:
(i) Determine the ages and durations to which these estimates apply, stating all
assumptions you make. [6]
(ii) Indicate which of the tabulations of census data gives the best match to each of
the tabulations of deaths and write down an appropriate approximation to the
required exposed to risk. State all assumptions you make. [4]
[Total 10]

12 You are the actuary of a life insurance company which issued 5,000 with-profit
endowment assurance policies to lives then aged exactly 40 on 1 January 2002. Each
policy had an original sum assured of £100,000 and a term of 20 years, with annual
premiums of £4,300 payable in advance throughout the term, ceasing on earlier death
or discontinuance.
You are given the following information, most but not all of which is needed to
calculate asset shares:
· The office holds net premium prospective reserves for in force policies based
on AM92 Ultimate mortality and 4% per annum interest.
· On death, policies receive the original sum assured plus previously declared
reversionary bonuses and any applicable terminal bonuses. The claim
payment is made at the end of the calendar year of death.
· On discontinuance within the first two years, policies receive a surrender value
equal to 25% of premiums paid. The surrender value is payable at the end of
the calendar year of discontinuance.
· On 31 December 2002, the office declared a reversionary bonus of 2% of the

original sum assured for all policies fully in force on that date (i.e. not
including any policies terminating during 2002 for reason of death or
surrender).
· On 31 December 2002, the office also declared a terminal bonus for death
claims which arose in the previous 12 months whereby the total death benefit
payable is 125% of the original sum assured plus 125% of any attaching
reversionary bonuses.
· Expenses incurred were £15.0 million on 1 January 2002.
· During 2002, 4 policyholders died and 200 discontinued.
· The office earned interest of 6.5% on its assets during 2002.
The company uses actual death claims when calculating asset shares and ignores all
other factors affecting profit or expenses not given above.
(i) Calculate the asset share per in force policy on 31 December 2002. [7]
(ii) State with reasons which information given is not required for your calculation
in (i). [3]
[Total 10]
105 A2003—6
13 (i) Describe the benefit whose present value is shown below. Tx and Ty are the
complete future lifetimes of two lives aged x and y respectively:
ì g (T ) = £100, 000vTy if T < T

ï x y
í [2]
ïî g (T ) = 0 otherwise
(ii) The policy in (i) was originally paid for by a single premium at outset. The
policyholders, who are both still alive, now request that the benefit be
modified immediately to be paid on the earlier death of either life.
Calculate the level premium payable annually in advance from now until the
first death of either life if the policy is amended in the manner requested.
Basis: Mortality: (x) subject to force of mortality of 0.02

(y) subject to force of mortality of 0.03
(x), (y) independent with respect to mortality
Interest: force of interest of 0.04
Renewal expenses: 2.5% of all premiums payable from the

alteration date
Alteration expenses: £100 [8]
(iii) State, with reasons, any actions the life insurance company should undertake
before proceeding with the alteration described in (ii). [2]
[Total 12]

14 A life insurance company sells 4-year decreasing term assurance policies, with level
premiums payable annually in advance for the term of the policy, but ceasing on
earlier death. The initial sum assured is £200,000 decreasing by £50,000 at each
policy anniversary and the death benefit is payable at the end of the year of death.
The company allows for the following when calculating premiums:
Initial expenses: £300 plus 25% of the annual premium
Renewal expenses: £30 per annum plus 2.5% of annual premium, incurred at the
time of payment of the second and subsequent premiums
Mortality: AM92 Select
Interest: 4% per annum (for all rates needed)
For a male aged exactly 60 at outset:
(i) Write down the gross future loss random variable at the outset of the policy.
[3]
(ii) Calculate the office premium using commutation functions, setting the
expected value of the gross future loss random variable to zero. [4]
(iii) Calculate the office premium using a discounted cash flow projection,
assuming no withdrawals, ignoring reserves and using the same profit criterion
as in (ii). [6]
(iv) Without further calculation explain the effect of:
(a) allowing for the setting up of reserves in the calculation in part (iii)
(b) having set up the reserves in (iv)(a), increasing the discount rate to
10% per annum
[3]
[Total 16]
105 A2003—8
REPORT OF THE BOARD OF EXAMINERS
April 2003
EXAMINERS’ REPORT
Introduction
reasonable.
J Curtis
3 June 2003
Overall Comments
The standard this year was generally good, slightly improved from last year. Candidates
seemed to cope well with the new areas that were examined this time (mainly questions 6 and
12). However, the following areas continue to prove the most difficult for candidates:-
estimation of select forces of mortality (question 11), mortality options (question 8), and
contingent assurances / reversionary annuities (question 7 and 13(ii)), despite the questions
asked being very standard for these topics.
Comments for individual questions follow after each question which we hope will assist
students.
26|26 26|26 1 26|26

1 s26 = z26 ò p26 dt (» z26 0.5 p26 )
0t
Well answered. The main error, if one was present, was to confuse the exact and
approximate relationships.
2 (i) The super compound bonus method is a method of allocating annual bonuses
under which two bonus rates are declared each year. The first rate, usually the
lower, is applied to the basic sum assured and the second rate is applied to the
bonuses added in the past.
(ii) The sum assured and bonuses increase more slowly than under other methods
for the same ultimate benefit, enabling the office to retain surplus for longer
and thereby providing greater investment freedom.
This method also rewards longer standing policyholders and discourages

surrenders, relative to other methods.
Very well answered overall. In part (ii), other reasons, where valid, were accepted.
3 The death benefit in year 10 is £15,000
Profit emerging per policy in force at the start of the year is:
([9V + P]*1.05) - (15,000*0.04) - ([1 - 0.04]*10V) =

([11,300 + 1,500]*1.05) - (15,000*0.04) - (0.96*13,200) = £168
Well answered. Two common errors recurred, using a wrong death benefit (usually nine
times the premium) and omitting the survival probability of 0.96 for closing reserves.
4 The component method builds up recursively year on year, allowing explicitly for
each of the 3 key elements: births, deaths and net emigration. Each of these can be
modelled separately to incorporate changing trends, although to do so relies on
detailed data and / or assumptions, usually split by age and sex.
Page 2
The logistic model is easy to apply, but is restricted in the variation it can allow for a
population, relying on 2 parameters which give a limiting population and an initial
growth rate, which reduces as population increases. The model does not lend itself to
understanding mechanisms of population changes. In reality, growth varies over time
in a different manner and most recent projections using the logistic and similar
models have tended to overestimate the population.
Also well answered. Occasionally candidates gave extremely lengthy and detailed
descriptions of the two methods, too much for the marks available, while at the same time
overlooked the comparison of the two approaches, which was the main thrust of the question
asked.
5 (i) Crude rates are easily calculated, relying only on total population at risk and
total deaths for each cause of death in this case.
However, the relative results for different countries can vary widely if the
death rate for a certain cause of death (a) varies by age — as most do — and
(b) population structures vary by age between countries. Differences in the
crude rates for a cause of death would then be confounded with differences in
population structures.
(ii) The rates could be standardised. Direct standardisation is best, whereby each
countries actual age-specific death rates are applied to a common population.
Any reasonable standard population could be chosen, but where possible it

should have some relevance to the study e.g. a European study could
standardise according to the population in Europe sorted by age as follows:
å s Exc mxA
x
Directly standardised death rate for cause A for a given country =
å s Exc
x
s c
Where Ex is the central exposed to risk at age x in the standard population
and mxA is the central mortality rate from cause A at age x in the country in
question.
Generally very well answered, especially part (i). While many candidates did not relate their
answers to the specific question which concerned a cause of death study and wrote about
mortality rates generally, this was accepted by the examiners. In part (ii), alternative
suggestions were also accepted, where justified.
HS (1/ all )
6 (a) (0.242488)(100)(52.18) using a 45 = £1,265.30
SS
(b) (5.4952)(100)(52.18) using a55,2 = £28,673.95
Very well answered. The only common error was the omission of the 52.18 factor.
Candidates seemed clearly familiar with the new examination tables.
Page 3
7 Value = 10, 000*1.02 I + 300 where
m
I = a60:20 +a f - 2* a m f
60:20 60:60:20
l80 6,953.536
m
a60:20 = a&&60 - 12 - v 20
l60
( )
a&&80 - 12 = 15.132 - (0.456387)
9,826.131
(7.006) = 12.869
l80 7, 724.737
af
60:20
= a&&60 - 12 - v 20
l60
( )
a&&80 - 12 = 16.152 - (0.456387)
9,848.431
(8.489) = 13.113
l80f l80
m
am f
60:60:20
= a&&60:60 - 12 - v 20 ( )
a&&m f - 1 = 13.590 - (0.456387)
l60 l60 80:80 2
6,953.536 7, 724.737
9,826.131 9,848.431
(5.357) = 12.233
I = 12.869+13.113-2*12.233=1.516
Value =10000*1.02*1.516+300
Þ Premium = £15,763
This question caused considerable problems to candidates . Common errors were to only
allow for one reversion (usually on death of male), omit the factor of 2 for joint life annuity,
use a factor of 0.98 instead of 1.02 for expenses, or assuming that the annuity ran for 20
years from the first death. A surprisingly high proportion of candidates used erroneous
formulae to convert annuities from annually in advance to continuous, often dividing by
1.040.5. This is a basic actuarial function which is given in the examination tables.
8 Let the full single premium at commencement = P
The premium (based on normal mortality) payable at the time of exercising the option
on the 2nd anniversary =
d 57 554
(1.05) 0.5 250,000q57 v = 250,000v 0.5 = 250,000v 0.5 = 1,444.30
l57 93,583
Therefore the premium required at duration 2, if the option is exercised, is £1,444.30
Thus equating the expected present value of all premium income with the expected present value
of all claims, we get:
Page 4
P + (0.4) 2 p55 v 2 (1,444.30) = (1.05) 0.5 250,000( q55 v + 1| q55 v 2 + 2 p55 [(.6) q57 v 3 + (.4)( 2)( 3q57 v 3 )])
l 57 2 250,000
Þ P + ( 0 .4 ) v (1,444.30) = (1.05) 0.5 ( d 55 v + d 56 v 2 + 3d 57 v 3 )
l 55 l55
93,583 2 250,000
Þ P + ( 0 .4 ) v (1,444.30) = (1.05) 0.5 ( 450v + 499v 2 + 3(554)v 3 )
94,532 94,532
Þ P + 518.75 = 6,278.54 leading to P = 5,759.79
Alternative approach based on non-option policy
If the policy were a simple 3-year term assurance without any options, the single
premium at commencement would be:
1
250,000A = (1.05) 0.5 250,000[ q55 v + p55 q56 v 2 + 2 p55 q57 v 3 ]
55:3|
250,000
= (1.05) 0.5 ( d 55 v + d 56 v 2 + d 57 v 3 )
l55
250,000
= (1.05) 0.5 ( 450v + 499v 2 + 554v 3 ) = 3,684.80
94,532
To allow for the option, the initial single premium needs to be increased by:
0.4 2 p55v 2 {1.050.5 [250,000( q57
* *
- q57 )v + 250,000( q57 - q57 )v ]}
*
q57 represents the mortality of optioners post-option = 3q57
(The 1st term in square brackets represents the extra mortality of optioners on the original SA,
and the 2nd term represents the extra mortality on the additional SA over and above that paid
for by the normal rates premium paid at the time of exercising option, t=2)
4d 57 3
= (0.4)( 2 p55 )v 3 (1.05) 0.5 250,000( 4q57 ) = (0.4)(1.05) 0.5 250,000 v
l55
4 * 554 3
= (0.4)(1.05) 0.5 250,000 v = 2,075.00
94,532
The total single premium at outset = 3,684.80+2,075.00 = 5,759.80 (same as above, allowing
for rounding)
The premium payable by policyholders at t=2 when exercising their option is (unchanged
from original solution):
d 554
(1.05) 0.5 250,000q57 v = 250,000v 0.5 57 = 250,000v 0.5 = 1,444.30
l57 93,583
This proved the most difficult question for students, with few fully correct answers. A number
of candidates seemed to misread the question and tried to calculate the cost of the option
(instead of the premiums) while others treated the policy as annual premium. Many students
calculated the basic premium for a policy with no option and tried to calculate the additional
premium required for the option so the examiners have provided an alternative solution
along these lines.
Page 5
9 (i) To zeroise future negative cash flows.
The office must meet all future outgo (additional to unit liabilities) e.g. death
claims in excess of units, expenses, maturity guarantees. It can take credit for
future income to the non-unit fund but cannot assume recourse to future
capital.
If there are negative cash flows, we cannot assume that they will be met from
subsequent positive cash flows or future capital (lapse risk, regulations). They
are future losses which we need to reserve for now. With adequate non-unit
reserves established, the minimum expected cash flow in future years,
allowing for release of reserves, is zero, hence the “zeroisation” of cash flows.
(ii) (-2, 0, 0, 0, +1, 0, 0, 0, 0, +1)
(iii) Cash flow approach is more flexible in general and allows for clarity of
thought and ease of presentation of results
Allows for complex policies (varying benefits, options)
Permits variable or stochastic premium basis e.g. interest basis
Best (often only) approach for multiple state model situations
Allows amount and timing of cash flows to be observed
Provides net cash flows useful for investment strategy
Allows for explicit amount of profit to be calculated.
Makes explicit allowance for cost of capital
Only way to calculate non-unit reserves
Facilitates repeating with altered basis for sensitivity testing (once spreadsheet
or program set up)
Generally well answered, especially part (ii). Some candidates only gave examples of outgo
in part (i), without considering offsetting income while in part (iii) some candidates tended to
concentrate on only one reason.
2 s z M ia 2 7.814 58, 094

10 (i) (30, 000) 40 s 40 = (30, 000) = £47,527.51
3 s39 D40 3 7.623 25,059
30, 000 s40 æç 18 z M 40 + R 40 - R 62 ö

ra z ra z ra
(ii) ÷
80 s39 ç s ÷
è D40 ø
30, 000 7.814 (18)(128, 026) + 2,884, 260 - 159, 030

=
80 7.623 25, 059
= £77,153.73
Page 6
æ M i + M 40
r ö
æ 369 + 782 ö
(iii) 50, 000 ç 40 ÷ = 50, 000 ç ÷ = £17, 945.12
ç D ÷ è 3, 207 ø
è 40 ø
Well answered throughout. The commonest mistakes related to omitting the salary
adjustment, treating (i) as service-related, omitting the factor for age 62 in the future service
part of (ii). In (iii), some candidates used annuity functions and / or omitted one of the types
of retirement.
11 (i) We are estimating m[x]+t
From 1qy,r y is policy year rate interval and lives are aged between y and y + 1
at the start of the interval in which death occurs, giving an average age at the
policy anniversary before death of y + .5, assuming an even spread of
birthdays over the policy year.
r is also a policy year rate interval, and is the same as a duration of r - 1 years
exact at the policy anniversary before death, without assumption.
The age at entry is y + .5 - (r -1) = y - r + 1.5 and the duration midway

through the rate interval (needed for the duration when estimating forces of
mortality) is r - 1 + .5 = r - .5 so we are estimating m[y-r+1.5]+r-.5. No further
assumptions are required.
From 2qy,r y is age last birthday at death giving a life year rate interval, with
lives y exact at the start of the interval without assumptions needed.
r is again a policy year rate interval, giving duration r years exact at the policy
anniversary before death, without assumption.
The average age at entry is y - r, but we must assume an even spread of

birthdays over the policy year because the 2 rate intervals are not the same (the
age at entry could range from y - r -1 to y - r + 1 based on the information we
have) and the duration midway through the rate interval is r + .5 so we are
estimating m[y-r]+r+.5.
(ii) Census A gives a life year for age, with y last birthday, and a policy year for
duration with r curtate.
Census B gives y next birthday at next policy anniversary, which is also y - 2

last birthday at previous policy anniversary. It also gives duration r at policy
anniversary following census or r - 1 curtate at census.
For the 1qy,r deaths, census B fits perfectly but we just need to be careful with
age labels. To get y last birthday at previous policy anniversary, and r - 1
curtate, we need Py+2,r.
Page 7
The approximate exposed to risk is é 12 B Py2000 B 2001 1 B 2002 ù

+ 2,r + Py + 2,r + 2 Py + 2,r û for
ë
estimating m[y-r+1.5]+r-.5
For the 2qy,r deaths, census A fits perfectly. To get y last birthday, and
r curtate, we need Py,r.
The approximate exposed to risk is é 12 A Py2000 A 2001 1 A 2002 ù

,r + Py ,r + 2 Py ,r û for
ë
estimating m[y-r]+r+.5
We assume that Px,t.varies linearly between census dates.
Generally not well answered , especially part (ii). In (i), some candidates based their answer
on the census data rather than on the death data, listed standard assumptions regardless of if
they applied here. Others, having defined the age and duration labels correctly did not
define the force of mortality at all or incorrectly.
In part (ii), while many students correctly matched the censuses to the death tabulations,
almost none got the correct age / duration labels for census B matched with deaths method 1.
There was a slight discrepancy in the question between the number of years of death data (3)
and the time period spanned by the censuses (2). This was not central to any of the answers
required, but the examiners accepted all valid interpretations / assumptions made by students
in this regard.
12 (i) Death claims in 2002 get SA, no reversionary bonus, and terminal bonus =
125,000
Discontinuances in 2002 get 0.25*4,300 = 1,075
2002 money flows:
Premium income: 5000*4,300 = 21,500,000

Expenses: 15,000,000
Balance: 6,500,000
Interest during 2002 @ 6.5%: 422,500
Balance @ 31/12/2002 before claims: 6,922,500
Death claims 2002: 4*125,000 500,000

Surrender claims: 200*1,075 215,000
Total funds 31/12/2002: 6,207,500
No. of policies in force 31/12/2002: 5000 - 4 - 200 4,796
Page 8
6, 207,500
Asset share per policy in force at 31/12/2002 = = £1,294
4, 796
(ii) The basis for net premium reserves and the 2002 reversionary bonus
declaration were the unnecessary items.
Neither affected the cash flows during 2002 nor therefore the year end asset
share.
Well answered, especially as this was the first time an asset share calculation had appeared.
The main error was to allow for reserves in some way. Some students tried to do the
calculation per policy sold but this usually led to errors.
13 (i) A (contingent) whole life assurance with benefit of £100,000 paid

immediately on the death of (y) providing it occurs after (x)’s death
(ii) Reserve before alteration
¥
V = 100, 000 ò e-dt (1 - t px ) t p y m y +t dt
0
¥ ìï¥ ¥ üï
-.04t -.02t -.03t -.07t
= 100, 000 ò e (1 - e ).03e dt = 3, 000 í ò e dt - ò e-.09t dt ý
0 îï 0 0 þï
ìæ 1 ö æ 1 ö ü
= 3, 000 íç ÷-ç ÷ ý = 9,523.81
îè .07 ø è .09 ø þ
Reserve post alteration:
100,000 A xy - (0.975P )a&&xy
¥
A xy = ò e-dt t px t p y (m x +t + m y +t )dt
0
¥ ¥
-.04t .02t -.03t
= òe e e (0.05)dt = 0.05 ò e-.09t dt
0 0
æ 1 ö
= .05 ç ÷ = 0.555556
è .09 ø
¥ ¥ ¥
e0.09
a&&xy = å e -dt t px t p y = å e-.04t e -.02t e -.03t = å e-.09t = = 11.6186
0 0 0 e0.09 - 1
Page 9
or alternatively
¥
1 1+ i
å e-.09t = a&&¥ (at i = e.09 - 1 = 0.094174) =
d
=
i
= 11.6186
0
Reserve before = Reserve after + alteration expense
9,523.81 = (100,000)(.555556) - P(.975)(11.6186) + 100
so P = £4,072.32 p.a.
(iii) Both lives should be underwritten at this time.

The proposed change increases the probability of claim payout by the insurer
substantially with regard to life y. Previously if y was worse than assumed
mortality, it was a margin for the office, but now the office is at immediate
risk in relation to y. The risk with regard to x is similar to that before the
alteration as regards the likelihood of a claim arising, but because the claim
would now be paid immediately on x’s death, the present value could increase
significantly.
Part (i) was well answered. In part (ii), many candidates made a good effort but many
omitted or could not calculate the pre-alteration reserve. In part (iii), many candidates made
general comments about underwriting without explaining why in the context of this particular
alteration.
14 (i) Gross future loss random variable (GFL r.v.) =
K[60] +1
(v ){(200, 000 - (50, 000)( K[60] )} + 300 + 30aK - P(.975a&&K - .225)
[60] [60] +1
for K[60] < 4
or 300 + 30a3 - P(.975a&&4 - .225) for K[60] ³ 4
(ii) E(GFL r.v.) = 0
1 1
Þ 250, 000 A[60]:4 - 50, 000( IA)[60]:4 + 300 + 30(a&&[60]:4 - 1) = P(.975a&&[60]:4 - .225)
1 M [60] - M 64 400.74 - 372.69

A[60]:4 = = = 0.0318547
D[60] 880.56
1 R[60] - R64 - 4M 64 7380.21 - 5813.76 - 4(372.69)

( IA)[60]:4 = = = 0.08595666
D[60] 880.56
Page 10
N[60] - N 64 12475.24 - 9186.74

a&&[60]:4 = = = 3.734555
D[60] 880.56
leading to 7,963.68 - 4,297.83 + 300 + 82.04 = 3.41619P P = £1,184.91
(iii)
q[60] 0.005774 p[60] 0.994226 0p[60] 1

q[60]+1 0.00868 p[60]+1 0.99132 1p[60] 0.994226
q62 0.010112 p62 0.989888 2p[60] 0.985596
q63 0.011344 p63 0.988656 3p[60] 0.97563
Year Prem Expense Interest Claim Cash flow Profit Signature NPV
1 P 0.25P+300 0.03P-12 1154.8 0.78P-1466.8 0.78P-1466.8 0.75P-1410.38

2 P 0.025P+30 0.039P-1.2 1302 1.014P-1333.2 1.008145P-1325.5 0.932087P-1225.5
3 P 0.025P+30 0.039P-1.2 1011.2 1.014P-1042.4 0.999394P-1027.39 0.888458P-913.342
4 P 0.025P+30 0.039P-1.2 567.2 1.014P-598.4 0.989289P-583.817 0.845648P-499.049
Total NPV = 3.416193P - 4,048.28
So P = £1,185.03
(same as above except for rounding due to use of commutation functions)
(iv) (a) Profit is deferred but as earned interest and risk discount rate are equal,
there is no impact on NPV or premium.
(b) Profit is deferred but because the discount rate exceeds earned rate,
NPV falls and premium would have to increase to satisfy the same
profit criterion.
Parts (ii) and (iii) were handled well throughout, with only the death benefit element of part
(ii) causing any difficulty. In part (i), a number of students gave the expectation of the
random variable, and among those who did give a random variable many omitted the select
notation and / or struggled with the benefit element. In part (iv), many gave correct answers
for (b), but in (a) very few students recognised that there would be no impact on the premium
because the earned interest rate equalled the discount rate.
Page 11
EXAMINATIONS
booklet.
supervisor.
question paper.
1 A life insurance company issues a number of 3-year unit-linked policies to lives each
aged 40 exact. The year-end non-unit fund cash flows ( NUCF )t , per policy in force at
the start of policy year t, are as follows (in £’s):
Year (t) 1 2 3
( NUCF )t 100 100 -150
Non-unit fund reserves are to be set up at each year end for each policy then in force
to zeroise future negative cash flows. Calculate the adjusted value of ( NUCF )t at the
end of year 1, assuming that interest is earned on reserves at the rate of 5% per annum
and that the mortality basis is AM92 Select. [3]
2 Describe four benefit options that may be available to an individual member of a

pension scheme who leaves the scheme before normal pension age. [4]
3 A life insurance company issues a disability insurance policy to a healthy life aged
exactly 45.
The benefits under the policy are as follows. There is no waiting period, but there is a
deferred period of one year. A benefit of £10,000 per annum is payable continuously
while the policyholder is sick, after the completion of the deferred period. The benefit
is payable until the policyholder reaches age 65, dies or recovers. Premiums are
waived while the policyholder is in receipt of benefit payment.
Level annual premiums are payable continuously under the policy until age 65 or the
policyholder’s earlier death.
Basis: Sickness: S(ID) Tables
Mortality: ELT(15) Males
Expenses: Initial: 60% of the annual rate of premium
Regular: £50 per annum, assumed incurred continuously in all

years of the policy, including periods of sickness
Claim: 1.5% of sickness benefit payments made to the

policyholder
[4]
105 S2003—2
4 Describe the calculation of a surrender value for a without-profit endowment
assurance policy, under which level annual premiums are payable monthly in advance
and cease on earlier death or surrender and the sum assured is payable immediately on
death. Give formulae, defining carefully all the symbols that you use. [5]
5 A life insurance company issues a term assurance policy to a life aged 55 exact for a
term of 10 years. The sum assured is payable immediately on death. The sum assured
is given by
£100, 000 ´ (1 + 0.05t ) t = 0,1, 2....,9.
where t denotes the curtate duration in years since the inception of the policy.
Level premiums are payable monthly in advance for a period of 10 years or until
earlier death. The life insurance company calculates the premium using the
equivalence principle.
Basis: Mortality: AM92 Select

Expenses: None [5]
6 A pension scheme provides a pension of 1/60 of Final Pensionable Salary for each
year of scheme service upon retirement for any reason. Fractional years of service
count proportionately. Final Pensionable Salary is defined as the average annual
salary in the three years immediately prior to retirement. Members are required to
contribute continuously at the rate of 5% of salary.
You are given the following data in respect of Member A as at 1 January 2003:
Age: 50 exact
Annual rate of salary: £50,000
Using the data in the Actuarial Tables, calculate, in respect of Member A:
(i) The expected present value of future contributions payable. [3]
(ii) The expected present value of the pension benefits on retirement for any
reason based on future service. [2]
[Total 5]

7 A life insurance company uses the following 3-state model, to calculate premiums for
a 3-year sickness policy issued to healthy policyholders age 50 exact at inception.
Dead (D)
In return for a single premium of P payable at the outset the company will pay a
benefit of £10,000 at the end of each of the 3 years if the policyholder is sick at that
time.
Let St represent the state of the policyholder at age 50 + t , so that S0 = H and for
t = 1, 2 and 3, St = H , S or D.
The life insurance company uses transition probabilities defined as follows:
ij
p50 +t = P ( St +1 = j | St = i )
For t = 0, 1 and 2 the transition probabilities are:
HD SD SH HS
p50 +t = 0.05 p50 +t = 0.15 p50 +t = 0.80 p50 +t = 0.1
The life insurance company calculates P as the expected present value of the benefit
payments, assuming interest at 6% per annum and expenses of 5% of P.
Calculate P. [5]
8 A life insurance company issues 10-year unit linked policies to lives aged exactly 50.
Premiums paid in the first two years of the policies are applied to purchase capital
units, with premiums in subsequent years being applied to purchase accumulation
units.
The management charge on the capital unit fund is 5% of the bid value of the units,
deducted at the end of each policy year. The management charge on the accumulation
unit fund is 1% of the bid value of the units, deducted at the end of each policy year.
The life insurance company wishes to use actuarial funding assuming a rate of interest
of 3% per annum. In calculating the actuarial funding factors, the life insurance
company assumes that mortality is constant, with
qx = 0.001 for 50 £ x £ 60 .
The life insurance company ignores surrenders.
105 S2003—4
(i) Calculate the actuarial funding factor to be applied at the end of the third year
of a policy. [4]
(ii) The life insurance company is considering using a higher rate of interest for
actuarial funding factors. It wishes to assume the same mortality basis and to
ignore surrenders in calculating the revised actuarial funding factors.
Describe how you would determine the maximum rate of interest it would be
prudent to use in calculating the actuarial funding factor to be applied at the
end of the third year of the policy. Set out the considerations you would take
into account. [5]
[Total 9]
9 You are a consulting actuary to a client who wishes to invest £1m now to provide an
immediate income for his partner and himself in retirement. Both the client and his
partner are aged 60 exact.
The client wishes to provide a payment annually in advance each year while either he
or his partner is alive. He wishes the amount of the payment to be
£ I ´ (1.05)t t = 0,1, 2.......
where I denotes the amount of the initial payment and t denotes the curtate duration in
years since the inception of the policy.
The client further requests that he wishes the amount of the initial payment I to be
such that the capital of £1m is at least 95% likely to be sufficient to provide the
required payments and he asks you to advise what the maximum value of the initial
payment I should be.
In carrying out the calculations, you assume that the only source of random variation
is the future mortality of the client and his partner.
Calculate the required value of I based on the following assumptions.
Mortality: The client and his partner are independent with

respect to mortality and are each subject to the
mortality of PMA92C20.
Rate of future investment returns: 6% per annum
Expenses: none
[9]

10 A life insurance company offers an option on its 10-year level term assurance policies
to effect a whole life without profits policy, for the sum assured, without evidence of
health. The option may be exercised once only, either on the fifth anniversary of the
policy or at the expiry of the 10-year term. If the option is exercised on the fifth
policy anniversary, the term assurance policy ceases immediately.
The sums assured under the 10-year term assurance policy and under the whole life
policy are both payable immediately on death. A single premium, inclusive of the
option premium, is payable at the outset under the term assurance policy and level
premiums under the whole life policy are payable annually in advance until death.
The premiums under the whole life policy are calculated using the company’s normal
annual premium basis.
(i) Describe the conventional method of pricing the mortality option, stating
clearly the data and assumptions required. [4]
(ii) A policyholder aged exactly 45 wishes to effect a 10-year without profits term
assurance policy, for a sum assured of £200,000.
Calculate the total single premium payable under the term assurance policy,
using the conventional method to calculate the option premium.
The following basis is used to calculate the basic term assurance premium:
Basis: Mortality: AM92 Select

Expenses: none [5]
[Total 9]
11 On 1 January 2000, a life insurance company issued an endowment assurance policy

to a life aged exactly 50 for a term of 10 years.
Under the policy, a sum assured of £100,000 is payable on survival to age 60 exact or
at the end of the year of death on earlier death. Level premiums are payable annually
in advance for 10 years or until earlier death.
On 1 January 2003, the policy is still in force and the life insurance company
calculates on a prospective basis both the gross premium reserve and the net premium
reserve for the policy at this date, using the assumptions shown below. The same
assumptions were used to calculate the gross premium at inception as follows:
Mortality: AM92 Ultimate

Expenses: Initial: £300 incurred at the outset
Renewal: 5% of each premium
(i) Calculate the gross premium reserve as at 1 January 2003. [3]
(ii) Calculate the net premium reserve, with Zillmer adjustment, as at 1 January
2003. Identify clearly the Zillmer adjustment. [2]
105 S2003—6
(iii) Explain why the net premium reserve with Zillmer adjustment calculated in
part (ii) might be used in preference to the net premium reserve with no
Zillmer adjustment, calculated as at 1 January 2003, using the same
assumptions. [2]
(iv) Assume instead that the life insurance company calculated the gross premium
reserve as at 1 January 2003 using a rate of interest of 3.5% per annum
following a general fall in market interest rates, with all other assumptions
unchanged. Assume also that the net premium reserve with a Zillmer
adjustment, calculated in part (ii), is unchanged.
State, giving a reason, whether you consider it appropriate to use this

unchanged net premium reserve with a Zillmer adjustment for reserving
purposes. [2]
[Total 9]
12 A life insurance company issues a two-year without-profit policy to a member, aged

exactly 50, of a certain club. The policy provides the following benefits:
(a) on death as a member during two years, a sum of £10,000
(b) on withdrawal from the club within two years, a return of 75% of premiums
paid without interest
(c) on survival as a member to the end of two years, the sum of £5,000
Death and withdrawal benefits are payable at the end of the year of death or
withdrawal respectively and the survival benefit is payable on the maturity date of the
policy. There are no decrements from membership of the club other than death or
withdrawal.
A premium of £3,000 is payable annually in advance under the policy for 2 years or
until earlier death or withdrawal.
Calculate the net present value of the profit under the policy to the life insurance
company.
Basis: Mortality: the independent rate of mortality is that of

AM92 Select
Withdrawal: the independent rate of withdrawal is 5% per annum
Rate of decrements: Mortality and withdrawal occur uniformly throughout
each policy year in the respective associated single
decrement tables.
Expenses: £150 incurred at outset
Rate of interest: 5% per annum
Reserves: Ignore
[9]

13 A life insurance company issues a special annuity policy to a male and a female life,
both aged exactly 60.
Under the policy, an annuity is payable annually in arrear for a maximum of 4 years,
ceasing on the first death of the two lives. The first payment under the policy is
£10,000 and subsequent payments increase by 1.9231% per annum compound.
(i) Calculate the standard deviation of the present value of benefits under the
annuity policy.
Basis: Mortality: The male and the female lives are independent with
respect to mortality and are subject to the mortality of
PMA92C20 and PFA92C20 respectively.

[8]
(ii) State, with reasons, whether the standard deviation would be higher, lower or
the same if the annuity were to cease on the second death of the two lives,
other conditions remaining unchanged. [2]
[Total 10]
14 A life insurance company uses the following multiple-state model for pricing and
valuing annual premium long-term care contracts, which are sold to lives that are
healthy at outset.
0: Healthy 1: Claim level 1 2: Claim level 2
3: Dead
Under each contract, the life company will pay the costs of long-term care while the
policyholder satisfies the conditions for payment. These conditions are assessed every
year on the policy anniversary, just before payment of the premium then due. If the
policyholder satisfies the conditions, the annual amount of the benefit payable is paid
immediately. A maximum of four benefit payments may be made under the policy,
after which time the policy expires. The policy also expires on earlier death.
Premiums are payable annually in advance under the policy until expiry, and are
waived if a benefit is being paid at a policy anniversary.
105 S2003—8
For lives at claim level 1, benefits of 60% of the maximum level are paid, while lives
at claim level 2 receive 100% of the maximum level. The current maximum level is
£50,000 per annum and is expected to increase by 6% per annum compound in the
future.
pijx is the probability that a life aged x in state i will be in state j at age x+1 and the
insurer uses the following probabilities for all values of x:
px00 = 0.87 px01 = 0.1 px02 = 0.0

p11
x = 0.6 p12
x = 0.3 px22 = 0.6
(i) Calculate the annual premium under the contract.

Expenses: 7.5% of each premium
[9]
(ii) A policyholder has already received two benefit payments at level 1, and is
about to receive a third benefit instalment. Calculate the reserves the office
should hold for this policy immediately after the benefit payment is made, if
the policyholder is assessed as entitled to either:
(a) benefit at level 1 = £42,000 per annum

(b) benefit at level 2 = £70,000 per annum
Reserve basis: Transition probabilities: as given
Benefit inflation: Inflation of the maximum benefit

level of 7% per annum.
[5]
[Total 14]
105 S2003—9
REPORT OF THE BOARD OF EXAMINERS
September 2003
EXAMINERS’ REPORT
Introduction
reasonable.
J Curtis
11 November 2003
© Faculty of Actuaries
© Institute of Actuaries
EXAMINATIONS
September 2003
EXAMINERS’ REPORT
© Faculty of Actuaries
© Institute of Actuaries
Overall Comments
The standard of answering overall was at a lower level than the examiners expected.
Candidates found particular difficulty with questions 4, 8, 9, 13 and 14. Attempts at questions
9 and 13 in particular were generally unsatisfactory. In relation to the other questions many
candidates performed well.
Individual comments follow after each question and we hope that these will be of assistance
to students.
1 p[ 40] = 0.999212
p[ 40]+1 = 0.999038
150
The reserve required per policy in force at the end of year 2, 2V = = 142.857
1.05
The cost of this, at the end of year 2, per policy in force at the start of
year 2 = p[40 ]+1 *2 V = 142.720
The adjusted value of ( NUCF )2 = 100 − 142.720 = −42.720
42.720
The reserve required per policy in force at the end of year 1, 1V = = 40.686
1.05
The cost of this, at the end of year 1, per policy in force at the start of
year 1 = p[40 ] *1V = 40.654
The adjusted value of (NUCF )1 = 100 − 40.654 = 59.35
The question was well answered in general. A number of candidates used incorrect mortality
rates.
2 Return of the member’s contributions
Under this option, the total of the member’s contributions are returned, with or
without interest. This option is available normally only after a short period of service.
There is likely to be a tax charge on the sum paid to the member.
A deferred pension payable from normal pension age
This option provides for the member to receive, from the scheme the member is
leaving, a pension payable from normal pension age. The pension is normally based
on the number of years’ service to the date of leaving and final pensionable salary at
Page 3
the date of leaving. The basic amount of the deferred pension is increased each year,
from the date of leaving to normal pension age, by a revaluation rate.
An immediate pension from the date of leaving
This option provides an immediate pension payable from the scheme, from the date of
leaving. This option is normally restricted to members close to normal pension age.
The pension can be calculated in a number of ways: a common method is to
determine the pension amount as that which is actuarially equivalent to the deferred
pension the member would otherwise have received.
A transfer cash equivalent
The transfer cash equivalent is an amount determined by the scheme actuary as a fair
assessment of the present value of the deferred pension and other benefits given up by
the member leaving the scheme. The transfer cash equivalent may be paid to a new
scheme that the member is joining, or to a special individual policy that a member can
effect for this purpose with a life insurance company.
This question was well answered in general. Some candidates just listed the benefit options,
whereas use of the word “Describe” required a fuller treatment.
3 Let P be the annual premium.
P is given by
(
P a45:20 − a
HS (1/ all )
45:20 ) = 10000*1.015* a
HS (1/ all )
45:20
+ 0.6 P + 50a45:20
∴ P (11.299 − 0.242488 ) = 10150*0.242488 + 0.6 P + 50*11.299
∴ P = £289.41
Overall this question was answered well. Some candidate had difficulty with valuing the
waiver benefit.
4 The retrospective policy value is determined, using a basis that reflects the experience
of the policy and takes account of the cost of surrender. The formula for the policy
value is as follows:
Dx
Dx +t { (12 ) (12 )
}
Ga − SA1x:t − I − ea − fA1x:t − C , where
x:t x:t
x is the age of policyholder at inception
Page 4
t is the policy duration at which the surrender value is being calculated

G is the annual office premium
S is the sum assured
I are the initial expenses, in excess of the regular expenses occurring each year
e are the regular annual expenses
f are the additional expenses that occur when the contract terminates
C are the surrender expenses
The prospective policy value is calculated using a basis that reflects the future
expected investment earnings, future expected expenses and future expected mortality
experience of the surrendering policyholders, less the cost of surrender. The formula
is as follows:
(12 ) (12 )
SAx +t:n−t + ea + fAx +t:n−t − Ga −C
x +t:n −t x +t:n −t
Additional definition: n is the original term of the policy.
A table of surrender values by policy duration is produced. The surrender value at a

particular duration is usually a blend of the retrospective and prospective policy
values, subject to a minimum of zero. Generally, the retrospective policy value is
given a greater weighting at earlier durations and the prospective value is given a
greater weighting at later durations. Other considerations, such as the asset share and
marketing influences, are also generally taken into account. Where possible, the
surrender value should be less than the asset share. Marketing considerations may
mean adjusting surrender values upwards.
Most candidates did not answer this question well. The examiners’ view was that this was a
standard theoretical question and well-prepared candidates should have scored reasonably.
Very few candidates mentioned both prospective and retrospective reserves; most formulae
given were not fully correct; and very few candidates dealt with the considerations set out in
the final part of the solution.
Page 5
5 Let P be the annual premium.
P is given by
Pa(12) = 95000 A 1 + 5000( IA) 1

[55]:10 [55]:10 [55]:10
⎛ D ⎞ ⎛ 689.23 ⎞
a(12) = a − 0.458 ⎜1 − 65 ⎟ = 8.228 − 0.458 ⎜ 1 − ⎟ = 8.056
[55]:10 [55]:10 ⎜ D[55] ⎟ ⎝ 1104.05 ⎠
⎝ ⎠
⎛ D ⎞
A1 = 1.040.5 ⎜ A − 65 ⎟
[55]10 ⎜ [55]10 D[55] ⎟
⎝ ⎠
= 1.040.5 ( 0.68354 − 0.624274 ) = 0.060439
R[55] − R65 − 10* M 65

( IA) 1 = 1.040.5 *
[55]:10 D[55]
9482.75 − 5441.07 − 10*363.82

= 1.040.5 * = 0.372692
1104.05
95000*0.060439 + 5000*0.372692
P= = £944.04
8.056
Candidates attempted this question well in general. There were some minor errors in the
formulae and numerical calculations.
Page 6
6 (i) Salary at age 50 exact ⇒ salary earned between age 49.5 and 50.5, assuming
that the salary increase was given at age 49.5.
s49.5 = 0.5*(9.031 + 9.165) = 9.098
Value of future contributions
s
N50 163638
= 0.05*50000* = 2500* = £25, 036.40.
s49.5 * D50 9.098*1796
(ii) Value of future retirement benefits
50000 z R50
ra
+ z R50
ia
50000 1604000 + 363963
= * = * = £100,365.26.
60 s49.5 D50 60 9.098*1796
The solution given is based on the assumption that Member A’s salary was increased 6
months before the valuation date. The examiners gave full credit for any other sensible
assumption so long as the assumption was stated. For example, assuming that the salary had
just been increased, s 50 would be used in place of s 49.5 . Candidates answered the question
well, in general.
7 The remaining transition probabilities are:
+t = 0.85 p50+t = 0.05

HH SS
p50
Probability of being sick at t = 1
= 0.1
p51 + p50
HH HS
= p50 p51 = 0.85* 0.1 + 0.1* 0.05 = 0.09
HS SS
= p50HH p51HH p52HS + p50HH p51HS p52

SS
+ p50HS p51 p52 + p50HS p51
SS SS SH
p52HS
= 0.85*0.85*0.1+0.85*0.1*0.05+0.1*0.05*0.05+0.1*0.8*0.1 = 0.08475
Page 7
P is given by
(
0.95P = 10000 * 0.1v + 0.09v 2 + 0.08475v 3 )
P = £2,585.23
Most candidates scored well on this question, with many getting full marks.
8 (i) The actuarial funding factor is given by
A53:7
at a rate of interest of 3% and mortality given by qx = 0.001 53 ≤ x ≤ 60 .
A53:7 = 1 − da53:7
a53:7 = 1 + 0.999v + (0.999v)2 + ... + (0.999v)6
1 − (0.999v)7
= = 6.39873
1 − 0.999v
0.03
A53:7 = 1 − *6.39873 = 0.81363
1.03
(ii) In assessing the maximum rate of interest, I would make a prudent

estimate of the level of the company’s future renewal expenses (including
renewal commissions) and express this as a regular percentage of the projected
bid values of the funded capital and accumulation units, say i%.
I would use discounted cash flow techniques to calculate i.
Conventionally, the rate i% tends to be the management charge used for

accumulation units, 1% in this case. In practice, we might tend to increase the
3% interest rate to 4% (5%-1%).
Mathematically, however, the maximum rate of interest is

(5% − i% ) (100% − 5% ) . In this case, assuming i = 1% , this would give a
maximum theoretical rate of 4.21%.
In assessing whether this would be prudent to use, I would compare the funded
value of capital units at the end of the third year using the revised actuarial
funding factor with the surrender value of capital units at that time. The
funded value should not be less than the surrender value. A further check
should be made to ensure that this remains the case at all subsequent policy
durations.
Page 8
I would also consider whether the mortality assumption was appropriate for
calculating the actuarial funding factor. The assumed level of mortality should
not be lighter than that prudently expected for the group of policyholders.
Otherwise the company would be anticipating future management charges it
might not receive.
Part (i) was not well answered. Many candidates did not show that the actuarial funding
factor as the present value of an endowment benefit.
Credit was given for variations from the solution set out: if a candidate assumed that the
amount of the management charge being pre-funded was 3% per annum and used a rate of
0.03
interest of for the present value of the endowment benefit, credit was given; if a
0.95
candidate assumed that the death benefit was payable immediately on death rather than at
the end of the year of death in the calculation of the present value of the endowment benefit,
credit was also given.
Part (ii) caused particular difficulties. Few candidates mentioned the use of discounted cash
flow techniques or the considerations set out in the final two paragraphs of the solution.
9 Let t be the future lifetime of the joint status. For the payments to be exactly 95%
likely to be sufficient, since the lives are independent with respect to mortality, the
value of t is given by
t p60:60 = 0.05
⇒ 2 t p60 − t p60:60 = 0.05
⇒ ( t p60 ) 2 − 2 t p60 + 0.05 = 0
2 ± 4 − 4*0.05
⇒ t p60 = = 0.02532 or 1.9747
2
⇒ t p60 = 0.02532
l60+t
⇒ = 0.02532 ⇒ 39 < t < 40
l60
Therefore, for the payments to be at least 95% likely to be sufficient, there must be at
least 40 payments.
Alternative derivation that there must be at least 40 payments
Page 9
For payments to be at least 95% likely to be sufficient, t is given by
t q 60:60 ≥ 0.95
⇒ ( t q 60 ) ≥ 0.95
2
⇒ t q 60 ≥ 0.97468
⇒ t p 60 ≤ 0.02532
l 60 = 9826.131 ⇒ l 60+t ≤ 248.798
⇒ t ≥ 40
I is given by
1.06
1000000 = Ia40 i = − 1 = 0.9524%
1.05
a40 = 33.44892
I = £29,896
This was the most poorly answered of all the questions, with few candidates gaining many
marks. The question was based on a practical application of standard joint life mortality and
the examiners would have expected candidates to have performed much better.
10 (i) Under the conventional method, the premiums that should be charged and the
premiums that will be charged for the new policy or policies that the
policyholder can opt to take are determined. The present value of the
differences between the premiums is then calculated and this is the present
value of the cost of the option. Where there is more than one option, the
present value of one option only is taken into account: the option chosen is the
one that gives the highest present value of the differences in premiums.
In carrying out the calculations, the following assumptions are made:
all lives eligible to take up the option will do so;
the mortality experience of those who take up the option will be the
Ultimate experience which corresponds to the Select experience that
would have been used as a basis if underwriting had been completed as
normal when the option had been exercised.
Page 10
The mortality basis used is not usually assumed to change over time, so the
only data required are the Select and Ultimate mortality rates used in the
original pricing basis.
(ii) The present value of the differences in premiums are as follows:
Option exercised at the fifth anniversary
D50 ⎛ A50 A[50] ⎞

Present value = 200000 ⎜ − ⎟ a50
D[ 45] ⎜ a50 a[50] ⎟
⎝ ⎠
1366.61 ⎛ 0.32907 0.32868 ⎞

= 200000* *1.040.5 ⎜ − ⎟ *17.444
1677.42 ⎝ 17.444 17.454 ⎠
= 96.10
Option exercised at the tenth anniversary
D55 ⎛ A55 A[55] ⎞

Present value = 200000 ⎜ − ⎟ a55
D[ 45] ⎜ a55 a[55] ⎟
⎝ ⎠
1105.41 ⎛ 0.38950 0.38879 ⎞

= 200000* *1.040.5 ⎜ − ⎟ *15.873
1677.42 ⎝ 15.873 15.891 ⎠
= 154.62
The cost of the option is the greater value, i.e., £154.62
The basic single premium is given by
M [ 45] − M 55
P = 200000*1.040.5 A 1 = 200000*1.040.5 *
[45]:10 D[ 45]
462.68 − 430.55
= 200000*1.040.5 = £3,906.75
1677.42
∴ The total single premium = £3,906.75 + £154.62 = £4,061.37.
Candidates performed well on this question in general. In part (ii) there is a subtle point that
if the 5 year option is taken then a release of the Term Assurance reserve would take place.
The Examiners did not expect students to cover this and the solution is based on this
assumption. A few candidates did point this out and due credit was allowed within the total
marks in these cases.
Page 11
11 (i) The original gross premium is given by
0.95 Pa50:10 = 300 + 100000 A50:10
a50:10 = 8.314
A50:10 = 0.68024
P = £8, 650.47
The gross premium reserve
= 100000 A53:7 − 0.95*8650.47 * a53:7
a53:7 = 6.166
A53:7 = 0.76286
Gross premium reserve = £25,614.14
(ii) Net premium reserve with Zillmer adjustment
⎛ a ⎞ a
= 100000 ⎜1 − 53:7 ⎟ − 300* 53:7
⎜ a ⎟ a50:10
⎝ 50:10 ⎠
= 25,835.94 − 222.49 = £25,613.45
£222.49 is the Zillmer adjustment.
(iii) The net premium reserve with Zillmer adjustment equals the gross premium
reserve calculated in part (i) (subject to rounding errors). If the insurance
company actuary is satisfied that there are sufficient margins in the gross
premium reserve then the net premium reserve with Zillmer adjustment would
be adequate. In addition, the use of the net premium reserve with Zillmer
adjustment compared with the use of the reserve without adjustment would
reduce the company’s funding requirements.
(iv) If the life insurance company’s actuary decided that the gross premium reserve
using 4% interest was no longer adequate given the fall in market interest rates
and that 3.5% interest should be used, this would give a higher value for the
gross premium reserve. The net premium reserve calculated in part (ii) was
equal to the gross premium reserve using 4% interest and this net premium
reserve would not be adequate.
Page 12
Many of the well prepared performed well on this question. A surprising number of
candidates showed a lack of understanding of a Zillmer adjustment.
12 The multiple decrement table is as follows.
Age ( x) (al ) x (ad ) dx (ad ) wx
50 100000 192.17 4995.07

51 94812.76 252.55 4734.16
52 89826.04
Values for the multiple decrement table are calculated from formulas of the following
type:
(
(aq ) dx = qxd 1 − 12 qxw )
(ad ) dx = (al ) x *(ad ) dx
(al ) x +1 = (al ) x − (ad ) dx − (ad ) wx
The profit test is set out as follows.
Year 1 2
Premium 3000 3000

Expenses 150
Interest 142.5 150
Death benefit 19.217 26.637
Withdrawal benefit 112.389 224.692
Survival benefit 4737.02
Cash flow 2860.894 −1838.349

Discounted cash flow 2487.734 −1317.954
Net present value £1,169.78
Candidates performed well on this question in general. Where errors occurred, they were
mostly in respect of the multiple decrement table. A number of candidates did not use a cash
flow approach which is what the Examiners were expecting.
Page 13
13 (i) With i = 0.06 and payments increasing at the rate of 1.9231% per annum, we
can value at 4%, but we must make the initial payment = 10000/1.019231.
age l x (male) l x (female) k kpxy Pr(Kxy = k)
60 9826.131 9848.431 0 1 0.004504

61 9802.048 9828.163 1 0.995496 0.005364
62 9773.083 9804.173 2 0.990132 0.006361
63 9738.388 9775.888 3 0.98377 0.007514
64 9696.99 9742.64 ≥4 0.976257 0.976257
k amin( k ,4)| amin( k ,4) 2 E[x] E[x 2]

0 0 0 0 0
1 0.961538 0.924556 0.005158 0.00496
2 1.886095 3.557353 0.011998 0.02263
3 2.775091 7.70113 0.020851 0.057863
≥4 3.629895 13.17614 3.543709 12.86329
3.581717 12.94875
Variance = 12.94875 − (3.581717)2 = 0.120052
Std Dev: (0.120052)0.5 = 0.346485
∴ Std Dev for this annuity is (10000/1.019231)*0.346485=3399.48
Alternative solution
With i = 0.06 and payments increasing at the rate of 1.9231% per annum, we
can value at 4%, but we must make the initial payment = 10000/1.019231.
We require
⎛ 1 − v min (K xy +1,5 ) ⎞
Var (amin (K , 4 ) | ) = Var (amin (K +1,5 )| − 1) = Var ⎜⎜ ⎟
⎟
xy xy
⎝ d ⎠
=
1
d 2
(A
2
60:60:5|
(
− A60:60:5| ))
2
age l x (male) l x (female) k Pr(Kxy = k)
60 9826.131 9848.431 0 0.004504

61 9802.048 9828.163 1 0.005364
62 9773.083 9804.173 2 0.006361
63 9738.388 9775.888 3 0.007514
Page 14
64 9696.99 9742.64 ≥4 0.976257
k v 4k%+1 * Pr (K xy = k ) (
v8k.+161% * Pr K xy = k )
0 0.0043308 0.0041642
1 0.0049593 0.0045852
2 0.0056549 0.0050272
3 0.0064230 0.0054904
≥4 0.8024121 0.6595242
0.8237801 0.6787912
(A60:60:5|
)
2
= 0.82378 2 = 0.6786136
Variance =
1
(0.6787912 − 0.6786136) = 0.12005
d 42%
Std Dev: (0.12005)0.5 = 0.34648
∴ Std Dev for this annuity is (10000/1.019231)*0.34648=3399.43
(ii) If the annuity were a last survivor annuity, the standard deviation would be
smaller. The chances of both lives dying during the 4 years would be much
lower, so more annuities would be payable for 4 years, with a consequent
reduction in the deviation from the average present value of the annuity
payments.
This question was very poorly answered in general. Many candidates were unable to make
any reasonable attempt. The examiners had expected the question to be challenging, but not
to the extent experienced. 2 alternative solutions are given which the Examiners hope will
assist.
Page 15
14 (i) With no recovery to the healthy state, premiums are payable only until the first
claim.
t px00 = t px00 = (0.87)t
∴ EPV premiums P{1 + 0.87v+(0.87v)2 + (0.87v)3 +...} = 5.578947P
Valuing the benefit from the point when the first claim arises, we get the
following probabilities:
the first claim payment will be at level 1;
the second claim payment will be at level 1 with probability 0.6 and
level 2 with probability 0.3;
the third claim payment will be at level 1 with probability

0.62 = 0.36 and at level 2 with probability 0.6*0.3 + 0.3*0.6 = 0.36 ;
the fourth claim payment will be at level 1 with probability

0.6 3 = 0.216 and at level 2 with probability
0.6 * 0.3 * 0.6 + 0.6 * 0.6 * 0.3 + 0.3 * 0.6 * 0.6 = 0.324 .
If the first claim is in n years time, the expected present value will be
50000*0.6*1.06n * v n . With v at 6%, this is 30,000 for all n. Similarly the
present value of any level 2 claim will be 50,000, so we can ignore interest in
valuing claims.
The EPV of all claims at the point of the first claim payment arising is
therefore:
30, 000*(1 + 0.6 + 0.36 + 0.216) + 50, 000*(0 + 0.3 + 0.36 + 0.324) = 114, 480
Finally the probability that the first claim occurs at the end of year 1 is 0.1, at
the end of year 2 is (0.87)*(0.1), at the end year 3 is (0.87)2*(0.1) and in
general at the end of year n is (0.87)n−1*(0.1).
The probability of a claim is therefore
0.1
0.1*(1 + 0.87 + 0.87 2 + ...) = = 0.76923
0.13
The EPV of all claims = (0.76923)*(114,480) = 88,061.45
Page 16
The equation of value is:
(1 − 0.075)*5.578947*P = 88,061.45 ⇒ P = £17,064.43
(a) If the third instalment is at level 1, then the fourth claim will be at level
1 with probability 0.6, or at level 2 with probability 0.3.
However, interest and claim inflation no longer cancel, so the reserve

immediately after paying the third claim is:
⎛ 1.07 ⎞ ⎛ 1.07 ⎞
V = 42, 000* ⎜ ⎟ *(0.6) + 70, 000* ⎜ ⎟ *(0.3) = £47, 080
⎝ 1.05 ⎠ ⎝ 1.05 ⎠
(b) If the third instalment is at level 2, then the fourth can only be at level
2, and will occur with probability 0.6.
This gives the following reserve value
⎛ 1.07 ⎞
V = 70, 000* ⎜ ⎟ *(0.6) = £42,800
⎝ 1.05 ⎠
This question was also not answered well. Many candidates valued the policies as four-year
policies only and many also failed to appreciate that interest could be ignored in valuing
claims in part (i) after which the question became much easier to complete. Few candidates
made reasonable attempts at part (ii).
Page 17
EXAMINATIONS
21 April 2004 (am)
Subject 105 Actuarial Mathematics 1
booklet.
supervisor.
question paper.
Faculty of Actuaries
105 A2004 Institute of Actuaries
1 (a) Give a formula for the Area Comparability Factor, defining all terms you use.
(b) Explain the role of this Factor in calculating standardised mortality rates,
indicating any advantages it has over other available methods.
[3]
2 A life insurance company uses the following model for pricing and valuing sickness
and other contracts.
1: Healthy 2: Sick
x x
3: Dead
pxab,t is the probability that a life now aged x and in state a will be in state b when aged
x+t
p xaa,t is the probability that a life now aged x and in state a will remain continuously in
state a until age x + t
Explain what is represented by each of the following integrals:
65 x t 12
(a) 12, 000e p x,t dt
0
30 30 t (t r ) 11 22
(b) 10, 000e p35,t x t p35 t ,r drdt
0 0
[3]
3 Explain the main differences in approach between the conventional and North
American methods for pricing mortality options in life assurance contracts. [4]
105 A2004 2
4 In a certain country, the population has reached a stationary size, and there is no
immigration or emigration. Women between the ages of 20 39 inclusive are regarded
as being of childbearing age and mortality in this age range is zero. In the past every
woman had a new baby on each of her 21st, 26th, 31st and 36th birthdays. From
1 January 2004, a change in birth patterns means that every woman is expected to
have a new baby on each of her 23rd, 28th, 33rd and 38th birthdays. During the
transition from one pattern to the other, it is expected that every woman will still have
4 babies, with a gap of at least 5 years between consecutive births.
Calculate the Total Fertility Rate for:
(a) the calendar year 2003

(b) the calendar year 2004
(c) women born in 1962
(d) women born in 1982
[4]
5 (a) Explain what is meant by
2
n q[ x ][ y ]
2
(b) Evaluate 25 q[40][40] assuming both lives are subject to AM92 mortality.
[4]
6 In a select mortality investigation, x,r corresponds to the number of deaths aged x

nearest birthday at death with duration r at the policy anniversary preceding death. x,r
divided by the appropriate central exposed to risk gives an estimate of [y]+t.
Derive the values of y and t to which this estimate applies, stating any assumptions
used.
[5]
7 The future lifetimes of two individuals aged x and y are independent, and subject to
constant forces of mortality of 0.02 and 0.03 respectively.
(i) Calculate the probability that their first death occurs after 3 years and before 8
years from now. [3]
(ii) Calculate the probability that their second death occurs after 3 years and
before 8 years from now. [3]
[Total 6]
105 A2004 3 PLEASE TURN OVER

8 A company issues a block of 5-year single premium investment policies to lives each
aged 60 exact at commencement of a policy. It guarantees simple annual reversionary
bonuses of 8% per annum of the single premium, with the possibility of a terminal
bonus at maturity. The death benefit is 5 times the single premium.
All premiums received under this policy are invested in an asset class where 5-year
returns have a normal distribution with a mean of 50% and standard deviation of 25%.
The company intends to declare terminal bonuses on maturity such that the proceeds
of the policy are the greater of the guaranteed amount and 90% of the underlying asset
value.
Calculate the probability that:
(a) the insurer makes a loss on a particular policy.
(b) a policyholder receives a terminal bonus.
Basis: Mortality: ELT15 (Females)

Expenses: Ignore
[6]
9 A retirement benefits scheme provides a lump sum retirement benefit equal to 3/80ths
of the salary rate at retirement for each completed year of service in the scheme.
Fractions of a year do not get credit. Retirement can occur at any age after attaining
age 60 but not later than a member s 65th birthday.
Calculate the total service liability for the lump sum benefit in respect of a member
aged 63 exact on the valuation date who has exactly 30 years of past service and is
earning £40,000 per annum.

Salary increases: Nil
Independent mortality rates: PMA92Base
Independent retirement rates: Age 63 last birthday 10%
Age 64 last birthday 6%
State any other assumptions you rely on. [6]
10 List the main categories of costs incurred by life insurance companies, giving an
example of each, and indicating the manner in which they are usually allowed for in
calculating premiums. [8]
105 A2004 4
11 (i) In the context of Manchester Unity Sickness Tables, state the meaning of:
(a) the force of sickness z x

(b) the annual rate of sickness s x
[2]
(ii) An insurance sickness policy provides combined endowment and sickness

benefits. The sickness benefit is £200 per week for the first 26 weeks of
sickness, £150 per week for the next 26 weeks and £100 per week thereafter
while sickness lasts. All sickness payments cease on a policyholder s 65th
birthday. There are no waiting or deferred periods.
The endowment part of the policy pays £10,000 immediately on the death of
the policyholder or on survival to age 65.
Premiums are waived during periods of sickness.
Calculate the level premium per annum payable continuously by a new

policyholder aged 35. Premiums are payable to age 65 but cease on earlier
death.
Basis: Sickness: S(MU)

Mortality: ELT 15 (Males)
Expenses: Nil
[7]
[Total 9]

12 On 1 January 1993, a life insurance company issued a number of 25-year without
profit endowment assurance policies to lives then aged 35 exact. Level premiums
were payable annually in advance throughout the term of the policy, ceasing on the
earlier death of the life assured. The sum assured was payable on survival to the end
of the term, or at the end of the year of death, if earlier.
Premiums and reserves were calculated on the following basis:

Expenses: 60% of the first premium
5% of each premium excluding the first
Calculate, as at 31 December 2003, the profit or loss for the calendar year 2003 in
respect of these policies, given the following information:
The total sums assured in force on 1 January 2003 were £50,000,000.
The total death claims occurring during 2003 and paid on 31 December 2003 were
£200,000.
During 2003, policies with sums assured of £2,500,000 were surrendered.

Surrender values, paid on 31 December 2003, were calculated as the retrospective
reserve using the above basis, but with interest at 4% per annum.
During 2003, policies with sums assured of £1,000,000 (before alteration) were
made paid up with effect from 31 December 2003. Paid-up sums assured were
calculated on a proportionate basis, namely the original sum assured * t/25 where t
is the number of premiums actually paid.
The company incurred expenses of £100,000 on 1 January 2003.
The company earned a total return of 7% on its assets during 2003.
Ignore tax, and assume that reserves for paid-up policies ignore future expenses.
[10]
105 A2004 6
13 A life insurance company issues a policy to male lives aged 45 exact, providing the
following benefits:
A decreasing term assurance with a death benefit, which is payable immediately

on death, of £200,000 in the first year, £190,000 in the second year thereafter
reducing by £10,000 each year until the benefit is £10,000 in the 20th year, with
cover ceasing at age 65.
An annuity of £25,000 per annum, increasing by £2,000 each year, where the first
payment is made on the policyholder s 65th birthday, and continues annually for
life thereafter.
The policy is paid for by level quarterly premiums payable in advance for 20 years,
ceasing on earlier death.
Calculate the premium, using the equivalence principle.
Basis:
Expenses: Initial: £200 plus 35% of the premiums paid in the first year
Renewal: 5% of all subsequent premiums and £40 per annum,

increasing by 4% per annum compound, on each policy
anniversary
Claim: Death: £250*(1.04)t where t is the exact duration of

the policy at death, measured in years with
fractions counting
Annuity: 2% of annuity payments

[14]

14 (i) Under a 4-year unit-linked policy issued to a male aged 60 exact, the
following non-unit cash flows, NUCFt, (t = 1,2,3,4) are obtained at the end of
year t per policy in force at the start of year t.
Year t 1 2 3 4
NUCFt 400 210 190 450
Mortality follows AM92 Select.
(a) Show that the annual internal rate of return lies between 5% and 6%.
(b) If the rate of interest earned on non-unit reserves is 7.5% per annum,
calculate the reserves required at times t = 1, 2 and 3 in order to zeroise
future negative cash flows.
(c) Without doing any further calculations, explain what effect the
zeroisation of future negative cash flows in part (b) above will have on
the internal rate of return relative to that in (a) above. [7]
(ii) A unit-linked endowment policy with an annual premium of £5,000 and a term
of 2 years is to be issued to a male life aged 60 exact. 97.5% of each premium
will be allocated to units at the offer price. The units will be subject to a bid-
offer spread of 4%.
At the end of each year a management charge of 1% of the bid value of the
units will be deducted from the unit fund.
If the policyholder dies during the term of the contract the office will pay out
the greater of £40,000 and the bid value of the units at the end of the year of
death (after the deduction of the management charge).
The company carries out all profit test calculations on the contract using the
following basis:

Rate of growth on assets in the unit fund: 9% per annum
Rate of interest on non-unit fund cash flows: 6% per annum
Expenses: £250 at time 0; £50 at time 1
(a) If the policyholder dies in the second year of the contract, calculate the
amounts of the non-unit fund cash flows in both of the years of the
contract.
(b) Hence calculate the net present value of the profit assuming that the
policyholder dies during the second year of the contract.
(c) The policyholder could also die in the first year, or survive to the end
of the term of the contract. Calculate the net present value of the profit
for each of these two events.
(d) Hence or otherwise, calculate the expected net present value of the
profit under this contract.
[11]
[Total 18]
END OF PAPER
105 A2004 8
EXAMINATIONS
April 2004
EXAMINERS REPORT
Institute of Actuaries
Subject 105 (Actuarial Mathematics) April 2004 Examiners Report
In general terms this was a relatively straightforward paper of standard questions with the
possible exceptions of Questions 4 and 8. It was well done by the well prepared students.
The Examiners noted, however, that many students appeared unprepared for this
examination and often their marks were well short of the required pass mark resulting in
overall a disappointing pass ratio.
s
Exc,t s mx,t E xc,t s mx,t
1 (a) ACF x x
s
Exc,t Exc,t
x x
where
E xc,t : Central exposed to risk in population being studied between ages x and x + t
s
Exc,t : Central exposed to risk in standard population between ages x and x +t
s
mx,t : central rate of mortality either observed or from a life table in standard
population for ages x to x + t
(b) When multiplied by the crude death rate for the population or area under
consideration, the ACF provides a standardised mortality rate ( the indirectly
standardised rate ). This approach is often favoured when data required by
other methods, usually local age-specific mortality rates, are unavailable.
Question 1 was generally well done although clearly many students could not remember the
standard formula.
2 (a) Expected present value of a benefit of 12,000 p.a. payable continuously to a

life now aged x and healthy whenever x is sick, with the benefit ceasing at
age 65
(b) EPV of a benefit of 10,000 p.a. payable continuously to a life now aged 35 and
healthy throughout their first period of sickness, ceasing at age 65 in any event
This question was done reasonably well. In part (b) of the question there was an erroneous
symbol x in the formula which should have been 35. The examiners gave full credit for using
either x or 35 in the answer above.
Page 2
3 Conventional assumes all eligible lives exercise the option, experience ultimate
mortality according to some table and pay premiums on the option policy based on
select mortality from the same table as if underwriting took place at the time of
commencement of the option policy.
The North American approach assumes that only a certain proportion of eligible lives
exercise the option. Opters and non-opters are subject to different mortality levels.
This is normally achieved by having a double decrement table of mortality / exercise
of option for original policyholders and also a mortality table for post-option mortality
for those who exercise the option.
Question done well. Credit was given for other appropriate comments.
4 Total Fertility Rate = f x where fx is age specific fertility rate at age x.

x
For calendar years, we use the period rate approach where we sum the fx s observed in
that year.
For women born in a calendar year, we sum across the fx s observed over their
lifetime, each x coming from the rate observed in the calendar year in which they
were aged x. (None of this is required from the student, it is just explanation for the
following results).
Up to the end of 2003, fx = 1 for x = 21, 26, 31, 36 and fx = 0 otherwise.
Therefore, the answer to (a) and (c) = 4, seeing as all relevant births occur before the
change at the end of 2003.
From 1 January 2004, fx = 1 for x = 23, 28, 33, 38 and fx = 0 otherwise.
The answer to (d) is also 4. They will have babies when they are aged 21 (in 2003),
28, (in 2010), 33 (in 2015) and 38 (in 2020).
The answer to (b) is zero. There are no women who will have babies in 2004. Those
aged 23, 28, 33, and 38 all had babies during 2002 when aged 21, 26, 31 and 36, and
therefore will not have their next baby until 2009 when they are aged 28, 33 and 38
respectively.
This question was not done well and many students failed to understand the concept of a
Total Fertility Rate attempting often to construct probabilities.
The solution above is a full one. One mark was awarded for each part if the student just
wrote down the correct numerical answer.
Page 3
2
5 (a) n q[ x ][ y ] represents the probability that a select life, now aged y, will die within
n years, having been predeceased by a select life now aged x
2 1* q 1 *{( q 2 1 *{(1 l65 2

(b) 25 q[40][40] 2 25 [40][40] 2 25 [40] ) } 2
) }
l[40]
1 *{(1 8,821.2612 2
= 2 9,854.3036
) } 0.005495
Question was done reasonably well.
6 x nearest birthday at death x ½ at start of rate interval (life year from x ½ to x

+ ½) during which life dies or x at mid-point when the force of mortality is estimated.
No assumption necessary.
r at policy anniversary preceding death means exact duration r at the anniversary

before death (the start of the policy year rate interval for duration) and hence r + ½
mid-year. No assumption necessary
The average age at entry [y] is therefore [(x ½) r], but we must assume an even
spread of birthdays over the policy year because the two rate intervals are not the
same type and therefore not coincident. (Based on the information we have the age at
entry could range from (x ½) (r + 1) to (x + ½) (r) i.e. x r 1½ to x r + ½,
on average x r ½.)
Therefore we get an estimate of [x r ½]+r+½
Well prepared students scored well on this question. For full marks all comments regarding
assumptions needed to be stated.
7 (i)
3|5 q xy 3 p xy 8 p xy 3 px * 3 p y 8 px * 8 p y
3 3 8 8
exp[ 0.02dt ]*exp[ 0.03dt ] exp[ 0.02dt ]*exp[ 0.03dt ]
0 0 0 0
3 8 0.15 0.4
exp[ 0.05dt ] exp[ 0.05dt ] e e .8607 .6703 0.1904
0 0
Alternatively, the joint life status has constant hazard rate 0.02+0.03 = 0.05 giving a
probability of the first death occurring between time 3 and 8:
8 8 0.05t 0.15 0.4

3|5 q xy 3 t
pxy x t: y t dt 3
0.05e dt e e 0.1904
Page 4
(ii)
3|5 q xy 3 p xy 8 p xy ( 3 px 3 py 3 px * 3 p y ) ( 8 px 8 py 8 px * 8 p y )
.06 .09 0.15 .16 .24 0.40
(e e e ) (e e e )
(.9418 .9139 .8607) (.8521 .7866 .6703) .9950 .9685 0.0265
Alternatively,
8 0.02t 0.03t 0.05t

3|5 q xy 3|5 q x 3|5 q y 3|5 q xy 3
(0.02e 0.03e 0.05e )dt 0.0265
Although this question was a simple application of probabilities it was surprisingly not done
well overall.
8 (a) Insurer makes a loss if either the policyholder dies or the asset value s 5-year
return is less than 40% for survivors
Probability of loss = 5 q60 5 p60 * [ 402550 ] = 5 q60 5 p60 * [ 0.4]
l65 87, 093

5 p60 0.94942877
l60 91, 732
[ 0.4] 1 [0.4] 1 0.65542 0.34458
Probability of loss = 0.05057123 (0.94942877) *(0.34458) 0.3777
(b) Terminal bonus is received if both the policyholder is alive and the asset value
exceeds 155.556% of single premium
Probability of terminal bonus = 5 p60 *(1 [ 55.556

25
50 ]) =
5 p60 *(1 [.2222])
[0.2222] 0.58792 interpolating linearly between values for 0.22 and 0.23
Probability of terminal bonus = 0.94942877*(1-.58792)=0.3912
This question was done very poorly overall. Even though the question defined a Normal
Distribution very few students appreciated how to apply this in this case.
9 Dependent decrement rates from independent using:
(aq) rx qxr (1 1 qd )
2 x
etc. assuming a uniform distribution of decrements in the single
decrement tables.
Page 5
Age qr qd (aq)r (aq)d (ap)
63 0.1 0.009189 0.099541 0.00873 0.89173

64 0.06 0.010604 0.059682 0.010286 0.930032
Probability of retiring at 64 last birthday = (.89173)*(.059682) = 0.05322
Probability of retiring at 65 = (.89173)*(.930032) = 0.829338
Assuming that those retiring age 63 or 64 last birthday can be represented as retiring
on average half-way through the year then we get
Age Benefit Discount Probability EPV
63 45,000 0.971286 0.099541 4,350.73

64 46,500 0.916307 0.05322 2,267.61
65 48,000 0.889996 0.829338 35,429.16
Total 42,047.50
Despite the definitions given in the question, many students failed to appreciate that they
needed to use dependent decrements and produced an answer based merely on independent
decrements. Limited credit was given for a solution based on independent decrements and to
score well the dependent approach was necessary.
10
Category Example Pricing
Initial Commission Allow for directly, usually

premium related
Marketing, promotional Per policy on estimated volumes
Underwriting / Processing Usually per policy, although some

proposal / Issue of policy elements might be tied to other
documentation driver (e.g. medical expenses
might be sum assured related)
Renewal commission Allow for directly, usually

premium related
administration Per policy per annum, allow for

inflation
Claim Calculation and payment of Per policy, allow for inflation

benefit
Page 6
Overhead Central services e.g. IT, legal Per policy per annum
The solution above are the main items the Examiners were seeking. The question was open
to wide interpretation as the word Costs was used as opposed to Expenses. Thus the
Examiners gave full credit within the total marks for other valid references. Allowing for this
the question was well done.
11 (i) (a) the force of sickness z x is the probability that a person aged exactly x
is sick at that moment
(b) the annual rate of sickness s x is the expected number of weeks

sickness that a life aged exactly x will experience in the year of age x
to x + 1
(ii)
HS (0 / 26) HS (26 / 26) HS (52 / all ) P HS (0 / all )
Pa35:30| 10, 000 A35:30| 200a35 150a35 100a35 a35
52.18
using S(ID) notation and where all sickness benefit functions are understood
to terminate at age 65.
Using values from S(MU) tables, and noting that A35:30| 1 a35:30| we get
P *16.979 10, 000{(1 (0.039221)(16.979)}

P
200(10.813 1.931) 150(2.203) 100(2.972 11.859) (29.778)
52.18
P(16.979-0.571) = 3,340.72+(2,548.80+330.45+1,483.10) = 7,703.07
P = 469.47 per annum
(Theoretically, some adjustment to the age should be made to reflect the fact
that a healthy 35 year-old cannot receive the 2nd / 3rd levels of benefits
immediately, but the usual adjustments are approximate and have only minor
influence on the result, so are ignored here.)
Question done well by well prepared students.
Page 7
12 Premium per 100,000 given by:
100, 000 A[35]:25| 0.95 Pa[35]:25| .55 P

100, 000*(.24198) P *[(0.95)(13.392) 0.55] P 1,987.94
Reserve per 100,000 for fully in force policy at 31 12 2002 given by:
10 V 100, 000 A45:15| 0.95Pa45:15|
10V (100, 000)(.42556) (0.95)(1,987.94)(10.149) 23,389.18
Reserve per 100,000 for fully in force policy at 31 12 2003 given by:
11V 100, 000 A46:14| 0.95 Pa46:14|
11V (100, 000)(.45028) (0.95)(1,987.94)(9.712) 26, 686.47
SV at 31 12 2003 per 100,000 SA given by
D[35] 1
11 SV (0.95Pa[35]:11| .55P 100, 000 A[35]:11| ) @4%
D46
N[35] N 46 52, 662.65 29,905.96

a[35]:11| 9.0772
D[35] 2,507.02
1 M [35] M 46 481.53 460.84

A[35]:11| 0.0082528
D[35] 2,507.02
2,507.02
11 SV [{1,987.94}*{(.95*9.0772) 0.55} {100, 000*0.0082528}] 23, 690.44
1, 611.07
Cost of PUPs at 31 12 2003 per 100,000 SA given by:
11 PUPV ( 11
25
)100, 000 A46:14| 44, 000*0.45028 19,812.32
Total funds available at 31 12 2003 before paying any claims or setting up reserves:
[{500 * (P+10V)} 100,000] * (1.07) = 13,469,759.20 = A
Total claims paid on 31 12 2003 = (200,000 + 25 * 11SV) = 792,261 = B
Total closing reserves required:
(500 2 25 10)11V + 10 * 11PUPV = 12,553,958.81 = C
Profit for 2003 = A B C = 123,539.39
Page 8
Question not done well and very few complete answers were presented
13 Equivalence principle EPV premiums = EPV benefits + EPV expenses
Let P = quarterly premium
D65
EPV premiums: 4 Pa (4) 4 P[a[45]:20| 3 (1
8
)]
[45]:20| D[45]
4 P[13.785 3 (1 689.23 )] 54.2563P
8 1,677.42
EPV death benefit:
1 1
210, 000 A[45]:20| 10, 000( I A)[45]:20|
M [45] M 65 R[45] R65 20 M 65
210, 000*(1.04)0.5[ ] 10, 000*(1.04)0.5[ ]
D[45] D[45]
462.68 363.82 13,987.39 5, 441.07 20*363.82
210, 000*(1.04)0.5[ ] 10, 000*(1.04)0.5[ ]
1, 677.42 1, 677.42
12, 621, 61 7, 720.60 4,901.01
D65
EPV annuity: (23, 000a65 2, 000( Ia )65 )
D[45]
(0.410887)({23, 000*12.276} {2, 000*113.911})
209, 622.20
EPV expenses: Death claim: 250 * 20q[45] = 250 * (1 0.90030) = 24.92

Annuity .02 * EPV annuity = 4,192.44
Premium related:
(4) (4)
(0.05)(4 Pa[45]:20| ) (0.30)(4 Pa[45]:1| )
3 [1 D[45] 1
(0.05)(54.2563P ) (0.3)(4 P)(1 8 D[45]
])
2.712815P 1.182171P 3.8950 P
Other:
l65
160 40a[45]:20| @ 0% 160 40({1 e[45]} {1 e65 })
l[45]
160 40[35.282 (0.90030)(17.645)] 935.85
Page 9
54.2563P = 4,901.01 + 209,622.20 + 24.92 + 4,192.44 + 3.8950P + 935.85
50.3613P = 219,676.42
Hence P = 4,362.01
Quarterly Premium is 4,362 to nearer whole unit.
Very few complete answers were presented but many well prepared students did successfully
complete a number of parts.
For the item of Renewal Expenses, credit was also given if the student took the alternative
approaches of:
1. Assuming that this particular expense applied throughout life.
2. If the inflation escalator of 4% applied from year 3 rather than year 2
14 (i) (a)
Year t q[60]+t 1 p[60]+t 1 t 1p[60] NUCFt Profit NPV @ 5% NPV @ 6%

signature
1 0.005774 0.994226 1 400 400.00 380.95 377.36

2 0.008680 0.991320 0.99423 210 208.79 189.38 185.82
3 0.010112 0.989888 0.98560 190 187.26 161.77 157.23
4 0.011344 0.988656 0.97563 450 439.03 361.19 347.76
Total 7.85 1.01
Because there is a change in sign in NPV between 5% and 6%, there

must be a solution to NPV = 0 for an interest rate between 5% and 6%.
(b) 3V = 0 since policy has positive cash flow in year 4.

2V = 190 / 1.075 = 176.74
Clearly 1V = 0 since 210 (NUCF2) > p[60]+1*2V
(i) (c) It will increase it. The rate on non-unit reserves exceeds the IRR so in
this case the deferral of profits, by introducing reserves, will increase
NPV and IRR.
(Usually the discount rate exceeds the non-unit rate of return and
allowing for reserves would then reduce NPV and IRR)
Page 10
(ii) These preliminary calculations, while also an alternative way to get the answer
required in (ii)(d), are presented here as background calculations for (ii)(a), (b)
and (c).
Unit fund
Fund
Year Cost of allocation brought forward Interest Mgmt. charge Fund at end
t at bt ct dt et
5000*.975*.96 et-1 .09*(at+bt) .01*(at+bt+ct) at+bt-ct-dt
1 4,680.00 0.00 421.20 51.01 5,050.19

2 4,680.00 5050.19 875.72 106.06 10,499.85
Non-unit fund
Unallocated Mgmt.
Year premium Expense Interest Death cost charge Cash flow Prof Sig. NPV
t ft gt ht it jt kt lt mt
5,000 at .06*(ft gt) q[60]+t 1*(40,000 et) dt ft gt+ht it+ jt t 1p[60]*kt 1.12 t* lt
leading to
1 320.00 250.00 4.20 201.80 51.01 76.59 76.59 68.38
2 320.00 50.00 16.20 256.06 106.06 136.20 135.41 107.95
Total 39.57
(a) Yr 1: 320 250 +4.20 + 51.01 = 125.21

Yr 2: 320 50 +16.20 + 106.06 (40,000 10,499.85) = 29,107.89
(b) NPV = 125.21v 29,107.89v2 = 23,092.84 (@12%)

(c) Die Yr 1 Cash flow: 320 250+4.20+51.01 (40,000 5,050.19) = 34,824.60
NPV = 34,824.6v = 31,093.39 (@12%)
Survive: Yr 1 = 125.21
Yr 2 = 320 50 + 16.20 + 106.06 = 392.26
NPV = 125.21v + 392.26v2 = 424.50 (@12%)
(d) NPV for contract = 31,093.39q[60] 23,092.84p[60]q[60]+1 + 424.502p[60]

= 39.57
(same NPV as in preliminary calculations above in non-unit cash flows)
A very straightforward question done very well by well prepared students many of whom
scored virtually full marks.
END OF EXAMINERS REPORT
Page 11
EXAMINATIONS
booklet.
supervisor.
question paper.
105 S2004 Institute of Actuaries
1 A life insurance company issues an annuity policy to two lives aged 65 and 62 exact
in return for a single premium. Under the policy an annuity of £10,000 per annum is
payable monthly in advance while at least one of the lives is alive.
Basis:
Mortality: PMA92C20 in respect of the life aged 65 exact
PFA92C20 in respect of the life aged 62 exact
Expenses: none
[3]
2 A member of a pension scheme is aged 50 exact, having joined the scheme at age 30
exact. His current salary is £50,000 per annum. Final pensionable salary is defined as
the annual average earnings over the three years immediately prior to retirement.
Normal Retirement Age is a member s 65th birthday. Salary increases take place six
months before the member s birthday.
Using the functions and symbols defined in, and the assumptions underlying, the
Example Pension Scheme Table in the Actuarial Tables, calculate the expected
present value of the following:
A pension on retirement at any stage on grounds of ill health of one-sixtieth of final

pensionable salary for each year of service, with fractions of a year counting
proportionately. [4]
3 A life insurance company issues a policy to a life aged 50 exact. The policy
provides the following sickness benefit:
£100 per week for the first two years of sickness, reducing to £50 per week thereafter
during sickness. Sickness benefit ceases at age 65, or on earlier recovery or death.
There is no waiting or deferred period.
Level premiums under the policy are payable weekly in advance until age 65 or until
earlier death. Any premiums falling due during periods of sickness are waived.
Calculate the weekly premium.
Basis:
Mortality: ELT 15 (Males)
Sickness Table: S(ID) in the Actuarial Tables
Expenses: 5% of each premium
(Expenses continue even when premiums are waived)
[5]
105 S2004 2
4 (i) Describe the use of risk classification by life insurance companies in
underwriting life assurance policies. [2]
(ii) State two limitations to the use of risk classification and explain how life
insurance companies deal with these limitations. [3]
[Total 5]
5 A life insurance company issued a non-profit term assurance policy to a life aged x
exact at the outset, with a term of 20 years. Under the policy, the sum assured of
£100,000 is payable at the end of the year of death. Premiums under the policy are
level and payable monthly in advance for 20 years, or until earlier death.
The company values the policy at duration t years using a gross premium prospective
policy value, tV .
Derive algebraically the relationship between tV and t 1V . Define all the symbols
that you use, where necessary. [6]
6 On 1 January 2001, a life insurance company issued a 10-year joint life non-profit
term assurance policy to two lives aged 50 exact. Under the policy, the sum assured
of £500,000 is payable immediately on the death of the first of the lives to die.
Premiums of £1,000 per annum are payable annually in advance for 10 years, or until
the first death of the lives assured.
On 31 December 2003 the policy is still in force. Calculate the gross premium
prospective policy value at this date, using the following valuation assumptions:
Mortality: PMA92C20 for the first life and PFA92C20 for the second life
Expenses: Renewal: 3% of each premium
Claim: £200 on payment of a claim
[6]
105 S2004 3 PLEASE TURN OVER

7 A double decrement table is to be constructed from two single decrement tables. The
modes of decrement are and . For each of the single decrement tables you are
given
lx t = lx t 2 .d x and l x t = lx t 3 .d x for 0 t 1
where lxi = the number of lives in the single decrement table i at age x exact
(i = , )
d xi = the number of decrements over [x, x + 1] in the single decrement

table i (i = , )
(i) Show that
t px . x t = 2tq x for 0 t 1
where
t p ix = the probability that a life aged x exact survives t years
i
x t = the force of decrement by cause i at age x + t
qix = the probability that a life aged x exact becomes a decrement

by cause i over [x, x + 1]
in the single decrement table for cause i (i = , ).

[3]
(ii) Hence or otherwise show that the dependent initial rate of decrement at age x
exact due to cause is:
2
aq x
= qx 1 qx
5
[3]
[Total 6]
105 S2004 4
8 A life insurance company issues 10-year non-profit term assurance policies, for a sum
assured S, to lives aged x exact. It offers an option on the policies to effect, either on
the fifth policy anniversary or at the expiry of the 10-year term, a whole life non-
profit policy for the same sum assured, without evidence of health. Premiums under
the term assurance policies are payable annually in advance for 10 years, or until
earlier death, or until the fifth policy anniversary, if the option is then exercised.
Premiums under the whole life policy are payable annually in advance for the whole
of life. The sums assured under the term assurance and whole life policies are
payable at the end of the year of death.
An additional single premium is charged at the outset under the term assurance policy
for the mortality option. The company uses the North American method for pricing
options.
Give formulae for calculating the additional single premium charged at outset for the
mortality option. You may ignore expenses. Define all the symbols that you use,
where necessary.
[8]
9 A life insurance company sells with profit whole life policies, with the sum assured
and attaching bonuses payable immediately on the death of the life assured and with
level premiums payable annually in advance ceasing with the policyholder s death or
on reaching age 65 if earlier.
Simple reversionary bonuses vest under the policies at the end of each year.
The company prices the product using the following basis:

Renewal: 2% of second and subsequent years premiums
Claim: £150 at termination of contract
Bonuses: Simple: 6% of basic sum assured per annum
(i) Write down an expression for the gross future loss at the point of sale for one
of these policies, assuming it is sold to a life aged x exact (x < 65) at the
outset. Write the expression in terms of functions of the random variables T[x]
and K[x], which represent the exact future lifetime and the curtate future
lifetime of (x) respectively. [3]
(ii) Calculate the gross premium required for one of these policies for a sum
assured of £200,000 and issued to a life aged 40 exact at the outset, using the
equivalence principle. State any assumptions you make. [6]
[Total 9]

10 The following data are available from a life insurance company, relating to the
mortality experience of its term assurance policyholders.
x,d The number of deaths over the period 1 January 2000 to 30 September 2003,
aged x nearest birthday at entry and having exact duration d at the next
policy anniversary following the date of death.
Py ,e (n) The number of policyholders with policies in force at time n, aged y nearest
birthday at entry and having curtate duration e at time n, where n = 1.1.2000,
30.9.2000, 30.9.2002 and 30.9.2003.
(a) Develop formulae for the calculation of the crude select forces of mortality
corresponding to the x,d deaths.
(b) Derive the age and duration to which these estimates apply.
Assume that all months are of equal length. State all other assumptions that you make.
[11]
11 A special 3-year endowment assurance policy provides that the death benefit payable
at the end of year of death is £10,000 plus the endowment assurance net premium
reserve for that year that would have been held had death not occurred. £10,000 is
payable on survival to the end of the 3 years.
On the basis set out below, use a discounted cash flow method to calculate the level
annual premium payable in advance for a life aged 57 exact. The requirement is that
at the discount rate defined below the value of the annual emerging surpluses should
sum to zero.
Basis: Mortality: AM92 Select for experience and reserves

Expenses: 20% of the first annual premium
5% of subsequent premiums
Reserves: Value as a normal endowment assurance for a 3-year term
on a net premium basis using a valuation rate of interest of
4% per annum. Ignore the effect on reserving of the extra
death benefit defined above.
Interest earnings: 7% per annum on cash flow
Discount rate: 10% per annum
Ignore tax and any other items.

[12]
105 S2004 6
12 A pension scheme provides the following benefits in respect of a former male
member of the scheme who has just left service:
(a) A pension to him for life of £10,000 per annum if he survives to age 65: the
pension commences on his 65th birthday and is guaranteed payable for five
years in any event.
(b) A spouse s pension of £5,000 per annum, commencing immediately on the

death of the former member before his 65th birthday and payable for life to the
spouse.
A spouse s pension is payable on death in deferment if the former member is

married at the date of death.
Pensions are payable monthly in advance.
Pensions in payment and deferment are increased monthly in arrears at the effective
rate of 2.8846% per annum.
The former member is now aged 62 exact. You are not given any information as to
whether he has a spouse.
Calculate the expected present value of these benefits using the following basis:
Basis: Valuation rate of interest: 7% per annum
Mortality in deferment
and in retirement: PMA92C20 for the former member and
PFA92C20 for his spouse
Proportion of former
members with a spouse
at each age up to age 65: 90%
Age difference of spouses: Females are exactly 3 years younger than their
husbands
Assume that death before retirement occurs at the mid-point of the year of age
in respect of each year of age.
[12]

13 You are a member of a committee responsible for monitoring the trend in assured
lives mortality rates. You have been presented with the following ratios of actual to
expected mortality rates on the basis of a standard table constructed twenty years ago
( Standard Table A ) and the total expected deaths over the period 2000 2003 based
on this table.
Age Ratio of Actual to Expected Total Expected Deaths (000 s)

Mortality Rates
2000 2001 2002 2003 2000 2001 2002 2003
15 44 1.80 2.00 10 10
45+ 0.90 0.80 20 20
You have also been given details of the exposed to risk data in the two age groups
15 44 and 45+ corresponding to Standard Table A. The exposed to risk data are
described as Standard Population A .
(i) Define, giving a formula, the term Standardised Mortality Ratio . Define all
(ii) Show how the Standardised Mortality Ratio may be expressed as a weighted
average. Describe the function averaged and the weights. [3]
(iii) Calculate the Standardised Mortality Ratios for the periods 2000 2001 and
2002 2003 with reference to Standard Table A, using the data presented. [2]
(iv) The committee measured the change in mortality between the periods
2000 2001 and 2002 2003 by calculating a Comparative Mortality Factor
(CMF) for each period. This factor was calculated as
r1
, where
r2
r1 was the expected number of deaths for the period obtained by applying
the observed mortality rates to Standard Population A
r2 is the expected number of deaths in Standard Population A over a

two-year period based on Standard Table A
The CMF was 0.95 for the period 2001 2001 and 0.99 for the period
2002 2003, which led the committee to conclude that mortality was
deteriorating.
(a) Explain the difference between the results of your calculation of the
Standardised Mortality Ratios in part (iii) and these CMF figures.
(Hint: Express the CMF figures as weighted averages.)
(b) State, giving a reason, which set of figures you think provides the
better results.
105 S2004 8
(c) Comment on the conclusion of the committee that mortality was
deteriorating. [6]
[Total 13]
END OF PAPER
105 S2004 9
EXAMINATIONS
September 2004
EXAMINERS REPORT
Institute of Actuaries
Subject 105 (Actuarial Mathematics 1) September 2004 Examiners Report
In general, well prepared candidates did well on this examination which contained
reasonably standard questions. Indeed some students scored high marks testifying to the
fairly straightforward nature of the paper. The Examiners noted however that there were
many candidates who were just not well prepared for the examination and this resulted in a
large number being quite a few marks below the required pass level.
Questions without further comment below were those that were in general done well by
candidates.
1 The premium is given by:
(12)
P 10000a65:62
12 11
a65:62 a65 a62 a65:62 24
= 13.666 + 15.963 12.427 0.458
= 16.744
P = £167,440
2 The expected present value is given by:
50000
20 z M 50
ia z ia
R50
0.5*60* s49 s50 * D50
50000
20* 45392 363963
0.5*60* 9.031 9.165 *1796
£64,861
3 Let P be the weekly premium. P is given by
HS 0 / all HS 2 / all
52.18* 100 P a50 50a50 0.95*52.18 P * a50:15
100 P *0.456447 50*0.184025 0.95* P *9.516
P £4.25
Page 2
4 (i) Risk classification is used as an underwriting tool by life insurance companies.

The company divides policyholders into different risk groups according to
factors that affect mortality. The company s expectation is that policyholders
in the same risk group are homogeneous with respect to mortality risk. The
groups are defined by the use of rating factors, e.g., age, sex, smoking habit.
(ii) In theory the company should add rating factors to its underwriting system
until the all mortality differences are fully accounted for, apart from random
variation. In reality, the ability of prospective policyholders to provide
accurate responses to questions and the cost of collecting information limit the
extent to which rating factors can be used. In addition, from a marketing point
of view, proposers are anxious that the process of underwriting should be
straightforward and speedy.
In setting underwriting terms, companies compromise between the conflicting

requirements of risk classification and marketing and use a limited number of
rating factors. It is important for a company not to omit a significant rating
factor that is used by other companies in the market: otherwise, there would be
a risk of selection against the company.
Credit was given for other suitable points and description.
5 Let P =monthly premium

G = annual equivalent premium (=12P)
e = annual regular expenses
f = claim expenses
12
tV ' 100000 f A1x t:20 t
G e a
x t:20 t
A1x t:20 t
vq x t vp x t A1x t 1:20 t 1
12 12 12
and a a vp x t a
x t:20 t x t:1 x t 1:20 t 1
12 12
tV ' 100000 f vq x t vpx t A1x t 1:20 t 1
G e a vp x t a
x t:1 x t 1:20 t 1
12
100000 f vqx t G e a vp x t 100000 f A1x t 1:20 t 1
x t:1
12
G e a
x t 1:20 t 1
12
= 100000 f vq x t G e a vp x t t 1V '
x t :1
12
tV ' G e a
x t:1
1 i qx t 100000 f
Page 3
= px t * t 1V '
Many students attempted to just write down the relationship which was not satisfactory. To
score well the relationship had to be derived from 1st principles and the nature of the monthly
premium effect clearly brought out.
6 The gross premium prospective policy value is given by:
500, 200 A 1 1, 000*0.97 * a53:53:7

53:53:7
m f
l60 l60
a53:53:7 a53:53 m
* f * v 7 * a60:60
l53 l53
7
9826.131 9848.431 1
16.716 * * *14.090
9922.995 9934.574 1.04
16.716 0.745975*14.09
= 6.205
m f
l60 l60
A 1 (1.04)1/ 2 * 1 da53:53:7 m
* f
* v7
53:53:7 l53 l53
= (1.04)1/2*(1 0.038462 * 6.205 0.745975)
= 0.015676
the gross premium policy value is:
500,200 * 0.015676 1,000 * 0.97 * 6.205 = £1822 to nearer £
Page 4
7 (i) Let l x t lx t 2 d x and l x t lx t 3d x
t px 1 t 2 qx and t qx t 2 qx
t px
t px x t 2tqx
t
1
(ii) Therefore aq x r px r px x r dr
0
1
= 1 r 3q x 2rq x dr
0
1
= qx 2r 2r 4 qx dr
0
1
2 2r 5
= qx r qx
5
0
2
= qx 1 qx
5
This question was done very poorly and few candidates derived satisfactory answers.
8 The expected present value of the benefits is given by
9 d 2 w
ad x t t 1
ad x 5t 1 5t
S v v Ax 5t , (I) where
t 0 al x t 1 al x
Ax S t px q x t vt 1
t 0
The expected present value of the premium income is given by
2 w
ad x 5t 1 5t
Px a x:10 Px 5t v ax 5t , (II) where
t 1 al x
9
a x:10 t
ap x
vt and a x t p x vt
t 0 t 0
Page 5
Px is the premium for the term assurance and Px 5 or Px 10 is the premium for the
whole life assurance at the date on which the option is effected.
The additional single premium is given by (I) (II).
A double decrement table is constructed for all lives that effect the term assurance
policy, with decrements of death and exercising the option, with the following
definitions:
d
ad x
, the number of decrements due to death aged x last birthday;
w w
ad x 4
and ad x 9
, the number of decrements due to exercise of the option at the
fifth policy anniversary and at the expiry of the 10-year term respectively; and
al x
, the number of lives aged exactly x in the double decrement table.
al x t
t
ap x al x
The dashed functions represent the mortality of those who have exercised the
option.
The above solution is just one of a number of possible approaches and credit was given to
candidates whose chosen method showed clear definitions. It was not totally necessary to
adopt a multiple decrement approach as movements took place at discrete points and again
credit was given for other methods.
T[ x ]
9 (i) L 250 S 1 0.06 K[ x ] 150 v 0.98 Pamin[1 K[ x ] ,65 x ]
0.02 P
(ii) Equivalence principle E L 0
Assume E T E K 1
2
250 0.94 S 150 A 40 0.06 S IA = 0.98 Pa 40 :25 0.02 P

40
1
2
250 1.04 0.94 200, 000 150 A 40 IA [40]
0.06 200, 000
= 0.98 Pa 40 :25 0.02 P
1
2
250 1.04 188,150* 0.23041 12, 000 *7.95835
Page 6
= P 0.98*15.887 0.02
1
250 1.04 2 43351.64 95500.2 P 15.58926
P £9, 099.32
10 x,d is classified as x nearest birthday at entry and duration d at policy anniversary

following death. Define a census taken at time t after the start of the period of
investigation (1.1.2000), P ' x,d t , of those lives having an in force policy at time t,
who were aged x nearest birthday at entry and will be duration d on the policy
anniversary following time t.
The Central Exposed to Risk is then given by
t 3.75
E xc,d P ' x,d t .dt
t 0
Then assuming that P ' x,d t varies linearly between the census dates (1.1.2000,
30.9.2000, 30.9.2002, 30.9.2003) the integral can be approximated by
1 *3 P'
2 4 x ,d 0 P ' x,d 0.75
1 *2 P'
2 x ,d 0.75 P ' x,d 2.75
1 *1 P '
2 x ,d 2.75 P ' x,d 3.75
However the censuses P ' x,d t have not been recorded. The recorded censuses
Px,d t have lives classified by x nearest birthday at entry and curtate duration d at
time t. We can write
P ' x ,d t Px,d 1 t
Substituting into the previous formula gives an expression for the required Central
Exposed to Risk.
x ,d
Then: x ,d estimates x d 0.5
Exc,d
because the average age at entry is x assuming birthdays are uniformly distributed
over the policy year, and the exact duration at the mid-point of the rate year (policy
year) of deaths is d 0.5 for all lives (no assumptions are necessary).
This question was generally done well by well prepared students but many did not appreciate
the relatively straightforward triangulation method.
Page 7
11 If St is surplus in year t per policy in force at begin year t then:
(t 1V+P Et)*1.07 = qt(10000 + tV) + (1 qt)*tV + St
Where tV etc is relevant reserve, P the required premium, Et is expenses for year t and
qt the relevant mortality for year t
So St = (P Et)*1.07 + t 1V *1.07 tV 10000* qt
We need to sum t-1 p[x]*St*vt at 10% for t = 1, 2, 3 and set to zero.
1V = 10000*(1 a[57] 1:2

/ a[57]:3 ) = 10000*(1 (1+v*l59/l[57]+1)/2.873)
= 10000*(1 1.956/2.873) = 3191.79
2V=10000*(1 1/2.873) = 6519.32 and 3V=10000 using 4% interest.
The following table can now be completed:
Year end t 1 2 3
Prem-Expense 0.8*P 0.95*P 0.95*P

t 1V[57] 0 3191.79 6519.32
10000*q[57]+t 1 41.71 61.80 71.40
Interest 0.056*P 0.0665*P+223.43 0.0665*P+456.35
tV[57] 3191.79 6519.32 10000.00
St 0.856P 3233.50 1.0165*P 3165.90 1.0165*P 3095.73
t-1 p[57] 1.00000 0.99583 0.98967
t-1 p[57]*St 0.856*p 3233.50 1.0123*P 3152.70 1.006*P 3063.75
Therefore:
(0.856*P 3233.5)*v+(1.0123*P 3152.70)*v2+(1.006*P 3063.75)*v3=0 at 10%
i.e. 2.3706*P=7846.92
P = £3,310.10
Very few students produced a full answer here. Although most solutions attempted were as
above, it was also acceptable to take the 3rd year reserve as zero i.e. assuming the £10000
maturity value had been paid. This approach would have given a numerical answer of
£3287.7
Page 8
12 The inflation rate of 2.8846% p.a. combined with the valuation rate of 7% p.a. means
that all benefits can be valued at 4% per annum effective.
The expected present value of the member s pension is given by
Benefit a
l65 1 12
10000* * 3
*a
l62 1.04 65:5
12 12 l70 5 12
a a * v * a70 at i 4%
65:5 5 l65
12 i
a * a5 1.021537 * 4.4518 4.5477
5 12
d
l70 9238.134
0.957538
l65 9647.797
v5 0.82193
12
a70 11.562 0.458 11.104
12
a 13.287
65:5
the expected present value is given by
9647.797
10000* *0.889*13.287 116,607
9773.083
Benefit b
The expected present value of the spouse s pension on death before retirement is
given by
2
1 d 62 t 12
5000 t 0.5
* h62 t 0.5 * * a59 t 0.5
t 0 1.04 l62
12
a59.5 0.5* 16.982 16.652 0.458 16.359
Page 9
Similarly,
(12)
a60..5 16.024
(12)
a61.5 15.679
the value is given by
34.694 41.398
5000* 0.980581*0.9* *16.359 0.942866*0.9* *16.024
9773.083 9773.083
49.193
0.906602*0.9* *15.679
9773.083
= 866
the total expected present value is
116,607+866 = £117,473.
13 (i) The Standardised Mortality Ratio is the ratio of the actual deaths in a
population compared with the expected deaths, based on standard mortality
rates.
The formula is
Exc,t mx,t
x
, where
Exc,t s mx,t
x
E xc,t is the central exposed to risk in the population between ages x and
x t
mx,t is the central rate of mortality for the population between ages x and
x t
s
mx,t is the central rate of mortality for a standard population between ages x
and x t
Page 10
(ii) The Ratio may be written in the form
mx,t
Exc,t s mx,t s
x mx,t
E xc,t s mx,t
x
which is the weighted average of the age-specific mortality differentials

between the population being studied and the standard population.
mx,t
i.e. s
,
mx,t
weighted by the expected deaths in the population being studied based on

standard mortality.
i.e. Exc,t s mx,t
1.8*10 0.9* 20
(iii) The SMR for 2000 2001 is 1.2
30
2*10 0.8* 20
The SMR for 2002 2003 is 1.2
30
(iv) (a) A formula for the CMF is
s
Exc,t mx,t
x
s
Exc,t s mx,t
x
which may be written in the form
s mx,t
Exc,t s mx,t s
x mx,t
s
.
Exc,t s mx,t
x
This is simply a weighted average of
mx,t
s
,
mx,t
Page 11
weighted by
s
Exc,t s mx,t .
The differences between the SMR and CMF figures indicates that the
Standard Population A and the observed population have different
proportions in the two age ranges.
As the CMF<SMR, this indicates that Standard Population A is more

heavily weighted to the older age group.
(b) In my opinion, use of the SMR gives better results for comparing the
population in each of the two periods. The mortality experience in the
two periods is compared using Standard Population A exposed to risk
in the CMF calculations and the observed population exposed to risk in
the SMR calculations. Standard Population A appears to have a
significantly different composition from the observed population.
Therefore, using the Standard Population A exposed to risk in the
weight calculations could introduce differences in the results which
have nothing to do with underlying mortality differences. Use of the
observed population exposed to risk removes this difficulty and results
should be more reliable.
(c) I disagree with the committee s conclusion. The SMR figures

indicate that the mortality experience has not changed between
2000 2001 and 2002 2003.
In part (iv) other acceptable comments were given credit.
END OF EXAMINERS REPORT
Page 12

Examinations: 18 April 2000 (Am)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Examinations: 18 April 2000 (Am)

Uploaded by

Copyright:

Available Formats

Faculty of Actuaries Institute of Actuaries

18 April 2000 (am)

Subject 105 — Actuarial Mathematics 1

Time allowed: Three hours

INSTRUCTIONS TO THE CANDIDATE

2. Mark allocations are shown in brackets.

3. Attempt all 17 questions, beginning your answer to each question on a

Graph paper is not required for this paper.

AT THE END OF THE EXAMINATION

Hand in BOTH your answer booklet and this question paper.

In addition to this paper you should have available

2 In a select mortality investigation, θx,r corresponds to the number of deaths

4 A 25 year annual premium endowment assurance policy was sold to a life

Basis: Mortality: A1967–70 Select

(a) Write down a recursive relationship between the gross premium

(b) Explain the meaning of this formula. [3]

Let the random variable g(T) take the following values:

(ii) Express E[g(T)] as concisely as possible in the form of an annuity

10 The number of people sick with a new disease is expected to increase

11 A multiple decrement table is subject to two forces of decrement α and β.

Calculate q αx and q βx . [3]

105—3 PLEASE TURN OVER

13 A pension scheme provides the following benefit to the spouse of a member,

Basis: a(55) Ultimate mortality at 8% per annum interest [8]

(a) the mortality, at different times, of the population of a given

(b) the mortality, at a certain time, of two different occupational

Occupation A Whole Country

(iii) Without performing any further calculations, describe what other

105—5 PLEASE TURN OVER

Rate of interest on unit investments: 8% per annum

(ii) Without performing any further calculations, state with reasons

(a) the risk discount rate were 10% per annum

Mortality: A1967–70 Ultimate

Subject 105 — Actuarial Mathematics 1

1 “Prospective service benefit” means a benefit not dependent on either past or

Examples include — lump sum death benefit of 4 ´ salary or spouse’s pension

Many candidates confused “prospective” with “future”.

2 x next birthday at entry Þ x - ½ on average at entry assuming birthdays

r at policy anniversary after death means exact duration r - 1 at the

The force estimated is m[x-½]+r-½ , so y = x - ½, t = r - ½.

5 (a) ( tV ¢ + GP - et) (1 + i) = qx+t (S) + px+t (t+1V ¢)

where tV ¢ = gross premium reserve @ time t

i = interest rate in premium/valuation basis

é æ .04 ö 2487.2117 ì æ .04 ö üù

= ½[1 - .75773 - (.24297)(.61758)]

7 Future service = 18 + 14 past Þ total = 32 > max of 30.

æ2ö æ 4.28 ö æ 35846 ö

8 (i) A continuous annuity of £1 p.a. payable for a minimum of n years and

The deductions include expenses, cost of benefits, tax, transfers to shareholders,

Only a minority of candidates seemed familiar with the logistic model.

11 Under UDD in single decrement table

( aq )=x = qx= (1 - ½qx> ) = qx= - ½qx= qx> = 0.2

( aq )>x = qx> (1 - ½qx= ) = qx> - ½qx= qx> = 0.05

Þ qx= - qx> = 0.15 Þ qx= = qx> + 0.15

(qx> + 0.15) - ½(qx> + 0.15) qx> = 0.2

Þ - ½(qx> )2 + .925qx> + 0.15 = 0.2

OR ( qx> )2 - 1.85qx> + 0.1 = 0

1.85 m 1.852 - 0.4

qx> = 0.05573 and qx= = 0.20573

1st iteration qx= = 0.2 ÷ [1 - (.5)(.05)] = .205128

qx> = 0.05 ÷ [1 - (.5)(.2)] = .055556

Similarly, 2nd iteration qx= = .20571, qx> = .05571

= 100,000 A65 – (.95)(5,530.53) a65

= G(K30+t) + (.025)(990) amin[ K