Advanced Long-Term Actuarial Mathematics

Solutions to Sample Questions

Most of the sample questions for the ALTAM exam are from past MLC or LTAM exams. The
solutions given here are generally edited versions of the model solutions provided for the past
exams. In some cases, we include comments from the graders, where that could be helpful for
those learning the material. There may also be alternative solution methods that are not presented
here.
Different solutions show different levels of accuracy in intermediate results. These model
solutions are not intended to imply that this is the best rounding for each question. Graders do
not penalize rounding decisions, unless an answer is rounded to too few digits in the context of
the problem and the given information. In particular, if a problem in one step asks you to
calculate something to the nearest 1, and you calculate it as (for example) 823.18, you need not
bother saying “that’s 823 to the nearest 1”, and you may use 823.18 or 823 in future steps.
In the numerical solutions presented, there may be small rounding differences arising from the
fact that values used in the calculations are typically more accurate than the intermediate values
recorded.

Versions:
Feb. 26, 2023 Original set of 56 sample questions published for the ALTAM exam
Feb. 6, 2024 Added two sample Excel questions (Questions 57-58)
Jan. 20, 2025 Added one sample Excel question (Question 59) and two sample questions
(Questions 60-61)
Mar. 9, 2025 Changed Question 60 to a sample Excel question and updated the solution
of Questions 50 and 61
Apr. 10, 2025 Notation revision to Question 19 solution.

Current Version Dated April 10, 2025

Page 1
Question 1
(a)

t px µ x +t −t px ( µ x +t + µ x +t )
d
t px =
10 11 10 10 01 03

dt

x µ x +t −t px ( µ x +t + µ x + t + µ x + t )
d
t p=
11
x t p10 01 11 10 12 13

dt

=
Boundary Conditions: 10
0 px 0= 11
0 px 1

Graders’ Comments:
Most candidates achieved full credit for this part. Common errors included omitting the
boundary conditions, or omitting terms on the right hand side. A few candidates forgot the
derivatives on the left hand side.

d 10
(b) From the Kolmogorov equation for t p80 , we have
dt
10
−t p80
10
lim t +h
h →0
p80
h
=t p80 µ80+t −t p80
11 10 10
( µ8001+t + µ8003+t )
p10 − p10
So, for small h, t +h 80 t 80 ≈t p80
h
µ80+t −t p80
11 10 10
( µ8001+t + µ8003+t )

⇒ t +h p80 ≈t p80 + h  t p80

10 10
µ80+t −t p80
11 10 10
( µ8001+t + µ8003+t )
1
⇒ 1/3 p80
10
≈ 0 + (0.08) = 0.02667
3
1
10
and 2/3 p80 ≈ 1/3 p80
10
+ [ (0.90346)(0.08) − (0.02667)(0.13082) ] = 0.04960
3
1
1 p80 ≈ 2/3 p80 + [ (0.81652)(0.08) − (0.04960)(0.13186) ] =
10 10
0.06919
3

Graders’ Comments:
11
This part was quite well done. Many candidates calculated values for h p80 , not realizing
these had been given to them in the question.

(c) (i) The expected present value of the service fees is

= 8000 ( a8000 + a=
EPV 80 )
01
(8000)(5.5793 + 1.3813)
= 55,685

Page 2
 (c) (ii) The expected present values of the level 2 care costs is

=EPV (30, 000)a8002 + (10, 000)5| a8002

= (30, 000)a8002 + (10, 000)(v 55 p80
00
a8502 + v5 5 p80
01
a8512 + v5 5 p80
02
a8522 ) = 23, 005

Graders’ Comments: This part is designed to test understanding of the multiple state model
annuity factor notation, and ability to manipulate the probabilities and annuities to create term
annuity factors. It was one of the more challenging parts of the exam overall.

Many candidates achieved full credit for (c)(i). Those who did not tended to combine, for
example, a8000 and a8011 , which would require the person to be in state 0 and state 1
simultaneously at age 80.
In (c)(ii), most candidates did not take all three cases into consideration. Some answers
allowed appropriately for one or two cases, and these received partial credit. Some
02
candidates combined, for example, 5 p80 with a8002 , which is a more serious error, as it
indicates a lack of understanding of the functions involved.

Page 3
Question 2
(a) Let lx(τ+)t denote the value immediately before exits at exact age x + t , and lx(τ+)t + denote the
value immediately after. At t = 0.5, just before decrement 2 exits, we have
d (2)
′(2) = (τ )
q60
l60.5
(τ )
l60.5 ( 0.5
)
=1000 exp − ∫ 1.2t dt =(1000)(0.86071) =860.71
0

60
′(2)=
⇒ q60 = 0.0697
860.71

(b)
1

l = 800.71 ⇒ l = 800.71e ∫0.5 = 800.71(0.63763)

(τ ) (τ ) 1.2 t dt −
60.5+ 61 = 510.6
(τ )
= 510.6 − 45= 465.6
l61+
⇒ d 60(1) = l60(τ ) − d 60(2) − d 60(3) − l61+
(τ )
= 1000 − 60 − 45 − 466 = 429

Graders’ Comments: There are different ways of doing this part; full credit was
awarded for any correct method. Quite a few candidates gave the answer as 451, which
is obtained by ignoring decrement (2).

(c)
(i) If decrement 2 occurs at the start of the year, there are fewer lives exposed to force of
(1)
decrement 1, so q60 would be smaller.
(ii) If decrement 2 occurs at the start of the year, there are more lives exposed to force of
(2)
decrement 2, so q60 would be bigger.
(iii) Because all the decrement 3 exits happen at the end of the year, we have
d 60(3) d 60(3)
= (3)
q60 = ′
q60 (3)

l60(τ ) l61(τ )
(τ )
Where l61 is the expected number of in-force immediately before the decrement 3 exits at
′(1) p60
the end of the year. Also l61(τ ) = l60(τ ) p60 ′(2) and since the independent rates are unchanged,
(τ )
l61 is unchanged, which means that d 60(3) is unchanged, which means that q60
(3)
is unchanged.

Page 4
Question 3
(a) Death in Year 1:
= (1000 + G (1 + i ))v − G
L0 | Event = 1000=
v 943.4
Probability = 0.06

Withdrawal in Year 1:
L0 | Event = −G
Probability = 0.04

Death in Year 2:
L0 | Event= (1000 + G (1 + i ) + G (1 + i ) 2 )v 2 − G (1 + v)= 1000v 2= 890.0
= = 0.108
Probability (0.90)(0.12)

Survival in force to end of year 2:

L0 | Event = −G (1 + v) =−1.9434G
= (0.9)(0.88)
Probability = 0.792

In table form:
Value of L 0 , Given Probability
Event
that the Event Occurred of Event

Death in year 1 943.4 0.06

Withdrawal in year 1 −G 0.04

Death in year 2 890.0 0.108

Neither death or withdrawal −1.9434G 0.792

Graders’ Comments: Most candidates did well in this part. Amongst candidates who did
not achieve full marks, the most common problem was determining the amount of the
return of premiums benefit.

Page 5
 (b) (i)
= E[ L0 ] (943.4)(0.06) + (−G )(0.04) + (890.0)(0.108) + (−1.9434G )(0.792)
= 152.7 − 1.579G
=⇒ a 152.7 = and b 1.579
(ii)
= E[ L20 ] (943.4) 2 (0.06) + (−G ) 2 (0.04) + (890.0) 2 (0.108) + (−1.9434G ) 2 (0.792)
= 138,947 + 3.0312G 2

E[ L0 ]2 =(152.7 − 1.579G ) 2 =23,317 − 482.2G + 2.4932G 2

V [ L0 ] =E[ L20 ] − E[ L0 ]2 =115, 630 + 482.2G + 0.538G 2

=
⇒ c 0.538
= =
d 482.2 e 115, 630

Graders’ Comments: The table in part (a) was used by many candidates to answer part
(b), as the examiners intended. Standard variance formulas for level benefit term insurance
do not work in this case, and candidates who tried to use memorized formulas received no
credit for this part.

(c) For each policy:

E[ L0 ] =
a − bG = −52.57
V [ L0=
] cG 2 + dG + e= 187, 408

So for the aggregate loss

E[ Lagg ] =
(200)(−52.57) =
−10,514
= =
V [ Lagg ] (200)(187, =
408) 37, 481, 600 (6122.2) 2

 0 − (−10,514) 
⇒ Pr[ Lagg > 0] = 1 − Φ   = 1 − Φ (1.72)
 6122.2 
=1 − 0.9573 = 0.0427

Page 6
 Question 4
(a)
(i)

 
t px= exp  − At −

B x t
log c
( )
c c − 1  from formula sheet.

⇒ 0.5 p64 = e−0.00257678 = 0.9974265
⇒ 0.5 q64 = 0.0025735 ⇒ 100, 000 0.5 q64 =
257.35

(ii)
=4V EPV(Benefit) − EPV(Premiums)

=100,000 ( 0.5 ) (
q64 v 0.5 + 0.5|0.5 q64 v − 0.9(270) 1 + 0.5 p64 v 0.5 )
where =
0.5|0.5 q64 q64 −0.5 q64 = 0.005288 − 0.0025735
= 0.0027145

⇒ 4V 100, 000 ( 0.0025115 + 0.0025852

= ) − 243 (1.9733886 ) 30.14

Alternatively, by recursion
( 4.5V + 0.9 P ) (1.05) = 100, 000 0.5 q64.5 + 0.5 p64.5 5V
0.5

⇒= 272.15v − − 0.9(270)
= 22.5912
0.5
4.5V

( 4V + 0.9 P ) (1.05) =100, 000 0.5 q64 + 0.5 p64

0.5
4.5 V
⇒ 4.V =
30.14

Graders’ Comments: Most candidates knew to use either EPV or recursive formula to
compute the reserve. Common errors made by candidates include:
• Not properly handling the semi-annual premiums;
• Approximating semi-annual annuity using annual annuity; and
• Assuming uniform distribution of deaths or constant force of mortality.

 1 01  −0.06
p64 =exp − ∫ µ64+t + µ64+t dt  =e × p64 =0.9367845
00 03 SULT
(b) 1
 0 

Page 7
 (c)
=4V
(0)
100, 000 ( 0.5
03 0.5
p64 (
v + p64
03
− 0.5 p64
03
)) ( (
v − 0.9(270) 1 + 1 − 0.5 p64
03
) )
v 0.5

=
1 − 0.00264 =
03
0.5 p64 0.99736
So=
4V
(0)
100, 000 ( 0.002576 + 0.002697 ) − 243 (1.97332 ) = 47.80

Graders’ Comments: Candidates generally earned some credit. Typical mistakes

included:
• Did not consider all possible transitions when calculating the probability for the
year-end benefit;
• Did not realize premium is paid in State 0, 1, and 2 or improperly calculated the
EPV; and
• Did not consider state dependent reserves when using the recursive formula in a
multiple state model.

(d)
The expected time of each sojourn in “At risk”, in years, is

∞ ∞
(
− t µ10 + µ12 ) ∞
1
∫ t px dt ∫e = ∫=
−100 t
= 11
dt e dt = years 3.65 days
0 0 0
100

Graders’ Comments: Many candidates skipped this part.

Page 8
 Question 5
(a) (i) Let µ 13 = µt13 represent the constant transition force from State 1 to State 3. Then we
have
∞
a = ∫ t p0 e−δ dt
01 01 t
x
0
∞ ∞

∫= dt µ 13 ∫ t p001 e−δ t dt
01 13 −δ t
=A003 t p0 µt e
0 0
03
A
⇒ 0
01
µ
=
13

a
x

(ii)

∞ r
∫0 µt +s +µt +s ds
12 13

∫
−δ r − r
= = e=
11 11 11
a t r p e t dr where rp t e 0.3
0
∞
1
∫e
−0.34 r
⇒ at = dr = = 2.941
11

0
0.34

11
(iii) at11 < at11 since 𝑎𝑎𝑡𝑡 also includes the (positive) EPV of payments made on returning
to State 1 after visiting State 2.

Graders’ Comments: Most candidates completed Part i correctly. For all parts, some
candidates had minor expression errors where their variables and/or integral limits were
not consistent.
11
Almost all candidates were able to correctly justify that annuity factor 𝑎𝑎𝑡𝑡 is smaller
11
than 𝑎𝑎𝑡𝑡 .

(b)
03 01
(i) 22,000 𝐴𝐴0 = 22,000 (0.1) 𝑎𝑎0 from (a) (i)
= 22,000 (0.1) (2.930) = 6446

01
(ii) 12,000 𝑎𝑎0 = 12,000 (2.930) = 35,160

Page 9
 (iii) The EPV is the sum of EPV of additional costs during first sojourn in State 1
(from State 0) up to 4 months, plus the EPV of additional costs during returning
sojourns in State 1 from State 2 up to 4 months. That is:

 ∞ 00 01 11 −δ t ∞

EPV 8000  ∫ t p0 µt at:1 e dt + ∫ t p002 µt21 at11:1 e t dt 
−δ
=
0 
3 3
0
1
−0.34/3
3
1− e
∫ r p=
11 −δ r
=
a
11
t : 13 t e dr = 0.3151
0
0.34

(c) The expected time of the first transition to State 1 is unchanged.

i) Decrease
With a larger value of 𝜇𝜇𝑡𝑡13 , the expected time spent in State 1 will decrease.
With less time spent in State 1, the EPV of the cost in nursing care in State 1 will
decrease.

ii) Increase
With a larger value of 𝜇𝜇𝑡𝑡13 , the transition to State 3 (surgery) will happen
sooner, on average. With surgery happening earlier, the EPV of
the cost of surgery will increase.

Graders’ Comments: For part (i), most candidates were able to answer and justify that
higher µt13 will cause higher probability (and thus shorter time) from State 1 to State 3.
For part (ii), while most candidates were able to correctly answer that a higher µt13 will
increase EPV of surgery, they were not able to clearly express that the increased cost
was due to shorter travel time to State 3 (absorbing state) and thus the EPV with less
discounting is higher.

Page 10
Question 6
(a) Note that µ x01+t is the force of mortality under Makeham’s Law with parameters A, B and
=
c* , where =
A 0.004, B 0.015, and c* = e0.005 = 1.0050125
B
−5 A− ( c *) (( c *) −1)
60 5

5 p =e
00
60
log c*
= 0.8846922
Prob. of diagnosis = 1 − 5 p60
00
= 0.11531

(b)
t
px = ∫ r px µ x + r
12 11 12 22
t t −r px + r dr
0
t
= ∫ e−0.2 r (0.2)e−0.4(t − r ) dr
0

e −1  
(=
0.2 e ) ∫ e dr ( 0.2 e ) 
t 0.2 t
−0.4 t −0.4 t
= 
0.2 r

0.2
0  
−0.2 t −0.4 t
= e −e

(c) 5 x = 1 − 5 px − 5 px = 1 − e
p13 11 12 −5(0.2)
− e( −5(0.2)
−e
−5(0.4)
)=
0.399577

(d)
5 5
= ∫ t p e dt = ∫ e
11 11 −δ t −0.2 t −0.05t
(i) a x:5 x e dt
0 0

1 − e5(0.25)
= = 2.85398
0.25
5 5

(ii) ax12:5 = ∫ t p12

x e
−δ t
(
dt = ∫ e−0.2t − e−0.4t e−0.05t dt )
0 0

1 − e5(0.45)
= 2.85398
= − 0.86598
0.45
5

∫( ) −δ t
(iii) ax11: 5 + ax12: 5 + ax13: 5= t x + t px + t px
p11 12 13
e dt
0
5
= ∫ 1 e−δ t dt (law of complete probability)
0

= a5

Graders’ Comments: Candidates who attempted this part did well on this question in
general, with most receiving full marks. Part iii) required a rigorous demonstration.

Page 11
(e)
Under Option A the QL index is 0.5 ax11: 5 + 0.2 ax12: 5 + 1.0 ax13: 5 , where the annuities are
now calculated with 𝜇𝜇𝑥𝑥23 = 1.0.
ax:5 = 2.85398 from above.
11

t px is now = ∫ e
12 −0.2 r − −
(0.2)e (t r ) dr
0

 e0.8t − 1 
(
= 0.2 e− t ) 
 0.8 
= 0.25 e−0.2t − e− t( )
5

( )
So ax12:5 = 0.25∫ e−0.2t − e− t e−0.05t dt
0

 1 − e (1.05)5 
−
= 0.25  2.85398 − 
 1.05 
= 0.47665

=
Also ax13:5 a5 − ax11:5 − ax12:5
= 4.42398 − 2.85398 − 0.47665
= 1.09335

So QL = 0.5(2.85398) + 0.2(0.47665) + 1.0(1.09335)

A

= 2.61567

Under Option B the QL index is 0.55 ax11: 5 + 0.3 ax12: 5 + 1.0 ax13: 5 =
2.534 , where the
annuities are calculated as in part (d), with
ax12:5 =
0.86598 and ax13:5 =
4.42398 − 2.85398 − 0.86598 =
0.70402
⇒ QL = 2.5335
B

Based on the improved QL, the health authority should select Option A.

Graders’ Comments: This problem was not attempted by a majority of the candidates;
those that did generally received most of the available credit.

Page 12
Question 7
(a)
d
= t p0 µt − t p0 µt
00 01 10 00 01
t p0
dt

t p0 = t p0 µt − t p0 ( µt + µt )
d 01 00 01 01 10 12

dt
d
t p0 = t p0 µt
02 01 12

dt
Boundary Conditions:
= 00
0 p0 1= 01
0 p0 0= 02
0 p0 0

Graders’ Comments: Most candidates were able to give the differential equations, but
quite a few did not provide boundary conditions.
For full credit, candidates were expected to use the specific model given in the question,
which meant that subscripts and superscripts needed to correspond with the notation of the
question for full credit.

(b) (i)
The Euler equation for t + h p001 is:

t +h p001 =
t p0 + h
01
( t p000 µt01 − t p001 ( µt10 + µt12 ) )
so, using=t 0,=
h 0.5 gives:
⇒ 0.5 p001 = 0 p001 + h ( 0 p000 µ001 − 0 p001 ( µ010 + µ012 ) )
(
0 + 0.5 (1)(0.5) + (0) ( 0.2 + 20 ) =
= 0.25 )
(ii)
The Euler equation for t +h p002 is
t +h =
p002 t p002 + h ( t p001µt12 )
=
so, using =
t 0.0, =
h 0.5 and then =
t 0.5, h 0.5 gives
0 p0 + h ( 0 p0 µ 0 ) =
p002 = 0 + 0.5 ( (0)(20 ) ) =
02 01 12
0.5 0

0.5 p0 + h ( 0.5 p0 µ 0.5 ) =

and 1 p002 = 0 + 0.5 ( (0.25)(20.5 ) ) =
02 01 12
0.177.

Graders’ Comments: Many candidates’ solutions lacked clarity in this part, which
made it difficult to award partial credit for incomplete answers.

Page 13
(c)

(i) APV = 1000(v ×0.5 p002 + v 2 (1 p002 −0.5 p002 ) at 4%

=1000 (1.04) −1 (0) + (1.04) −2 (0.177 − 0)  =163.64

(ii) P(1 + v ×0.5 p000

= = 163.64
) APV
163.64
=⇒P = 95.08
1 + (1.04) −1 (0.75)

Graders’ Comments: This part was done well by most candidates who attempted it.
Note that all the probability values required were given in the question, so it was not
necessary to answer (b) to answer (c) correctly. Full credit was given for answers
using the rounded probability values given in the question.

(d) Use smaller h for greater accuracy.

Page 14
Question 8
(a)
10
=
ax0:10 ∫ px0 j ⋅ e −δ t ⋅ dt
j
0 t

10
⇒ ax00:10 + ax01:10 + ax= ∫ (t px00 + t px01 + t px02 ) ⋅ e −δ t ⋅ dt
02
:10 0

but t px00 + t px01 + t px02 =

1

10
⇒ ax00:10 + ax01:10 + ax02
=:10 ∫
0
e −δ t ⋅=
dt a10

Graders’ Comments: Many candidates did well on this part. Some candidates gave a
verbal explanation, rather than mathematical proof. Generally, this earned partial credit,
but full credit was awarded if the verbal proof was sufficiently detailed.

(b)
EPV Premiums: Pax00:10 = 4.49 P

1 − v10
EPV Disability Annuity: 1000ax01:10 = 1000 ( a10 − (ax00:10 + ax02:10 ) ) where a10 = = 6.32121
δ

⇒ EPV Disability Annuity =

471.21

EPV Death Benefit: 10000 Ax02:10 = 3871.0

3871 + 471.21
=
So the Premium is P = 967.1
4.49

Page 15
 (c) The Thiele equation at time t is

= δ ⋅t V (0) + P − µ x01+t ( t V (1) −t V (0) ) − µ x02+t (10, 000 −t V (0) )

d (0)
tV
dt
so at t = 3 we have
d
= tV
(0)
(0.1)(1304.54) + 967.1
dt t =3

− 0.04 ( 7530.09 − 1304.54 )

− 3(0.2) (10, 000 − 1304.54 )
= 326.80

Graders’ Comments: This part proved more challenging to many candidates. The responses
indicated that many candidates are memorizing Thiele’s formula rather than understanding
the intuition behind it.

(d) Let P* denote the new premium. The EPV of the return of premium benefit is
− δ
EPV=10 P*e 10 10 px00 , where
 10 01   10  −1.4
10 p exp  − ∫ µ x +t + µ x +t dt  =
=
00
x
02
exp  − ∫ 0.04 + 0.02t dt  =
e = 0.24660
 0   0 
=⇒ EPV 10 P*e−10
=×0.1
(0.24660) 0.90718 P*
3871 + 471.2
=⇒ P* = 1212.0
4.490 − 0.90718
(
⇒ P* − P =
244.9)
Graders’ Comments: A common error was to omit the discount factor, e−10δ , possibly on the
grounds that the premiums are returned without interest, but the discount factor reflects the
interest earned by the insurer, not the interest that might be returned to the policyholder. If the
premiums were returned with interest, the benefit would be in the form P* s10 instead of 10P*.

Page 16
Question 9
(a) The six common ADLs are
• Bathing
• Dressing
• Eating
• Toileting
• Continence
• Transferring

(b) One model is given below. Others may also be appropriate.

Active, ³5 ADLs Impaired Severely impaired

4 ADLs £3 ADLs
0 1
2

Dead

(c)
(i) Over the course of the policy, the insured was active for 12 + 12 + 12 + 6 + 4 + 12
+12 + 12 + 12 + 6 = 100, months, thereby paying a total of 100 × 150 = 15,000 in
premium. For the first disability, (6 - 3) × 1000 = 3000 in benefits were paid; for
the second disability, (8-3) × 2000 = 10,000 in benefits were paid. Since the
sum of the premiums paid exceeds the disability benefits paid out under the
policy, under the “return of premium" approach, the remainder of 15,000-13,000
= 2000 is added to the death benefit, for a total death benefit of 102,000.
(ii) As before, over the course of the policy, 13,000 in benefits were paid. This
amount is deducted from the sum insured of 100,000, leaving a death benefit
payment of 87,000.

Page 17
(d)
(i) With a 12 month off period, the second disability, which starts 9 months after the end of
the first disability period, is not subject to the 3 month waiting period. Hence, benefit
payments will commence immediately for this disability, increasing the long term care
benefits paid by 3×2000 = 6000. However, since the total long term care benefits paid
(19,000) would now exceed the total premium paid (15,000), there would no longer be a
return of premium added to the death benefit, lowering it by 2000. Hence, the total
benefits paid would increase by 4000.
(ii) Again, in this scenario, the long term care benefits paid would increase by 6000.
However, the death benefit paid would be 6,000 less than before. Hence, there is no net
change in total benefits paid by the policy.

Page 18
 Question 10

(a)
= 12
t + h px t x h px + t + t px h px + t
p11 12 12 22
(Markov property)
x h p x + t + t p x (1 − h p x + t − h p x + t )
p11 12 12 23 24
t (complete probability)

x + t px h px + t − t px h px + t − t px h px + t
p12 11 12 12 23 12 24
= t

x −t px
p12 12
 h p12   24
12  h p x + t
12  h p x + t 
23
⇒ t +h
= t p11
x 
x +t
−
 t x 
p −
 t x 
p 
h  h  
 h   h
p ij
and, where the transition intensity exists, µ xij+t = lim h x +t , and taking limits gives
h →0 h
d 12
t px = t px µ x +t − t px µ x +t − t px µ x +t
11 12 12 23 12 24

dt
(b)
t t

∫=
r p x µ x + r t − r p x + r dr ∫e
−0.24 r
= 12
t px
11 12 22
× 0.10 × e −0.34(t − r ) dr
0 0

e0.1t − 1
t
= 0.10(e −0.34t ) ∫ e0.1
= r
dr 0.10(e −0.34t ) = e −0.24t − e −0.34t
0 0.1

(c)
The EPV is 3000a9012 where
∞ ∞
a9012 = ∫ t p90 ∫ (e − e −0.34t ) e −0.05t dt
12 −δ t −0.24 t
=e dt
0 0
∞
= ∫ ( e −0.29t − e −0.39t ) dt = −
1 1
0
0.29 0.39
= 0.88417

So the EPV is 3000(0.88417) = 2652.5 .

(d)

1 − e−0.5×0.39
0.5 0.5

∫ p e dt ∫ e =
−δ t −0.34 t −0.05t
=
a 22
90:0.5 t=
22
90 e dt = 0.4543
0 0
0.39

Page 19
(e) The EPV of a benefit of 1 per year payable continuously after the waiting period is
∞

∫
0
r
11
p90 µ90
12
( 22 22
)
+ r a90 + r − a90 + r :0.5 e
−δ r
dr where a9022+ r =
1
0.39
=
2.5641

So the EPV is
∞ ∞
0.21099
∫0 e × 0.10 × ( 2.5641 − 0.4542 ) e dr = 0.21099∫0 e dr =
−0.24 r −0.05 r −0.29 r

0.29
=0.7276
So for a benefit of 3000 per year the EPV is 3000×0.7276=2182.7.

(f) Reason 1: Short term payments involve relatively high expenses. Once the illness has
extended beyond the waiting period, it is likely to be more significant and less costly
relative to the benefits
Reason 2: In some cases the policyholders will have other sources of income for short
term sickness, eg sick pay from employment.
Reason 3: Offer policyholders a choice, for the same premium they will receive higher
benefits for long term sickness in return for giving up benefit for shorter bouts.

Page 20
Question 11
(a)(i) EPV of future costs is
12 × 3000 × a65
(12)00
+ 12 × 7500 × a65
(12)01
+ 12 × 15000 × a65
(12)02
=
600,321.6

So F is 25% of the EPV = 150,080

(a)(ii) The monthly fee is M where

12 M (a65
(12)00
+ a65 =
(12)01
+ a65
(12)02
=
) 0.75(600321.6) ⇒ M 2854.95

(b)(i) The reserve is EPV future outgo – EPV future fee income, so

5 V (0) = 12(3000)a70
(12)00
+ 12(7500)a70
(12)01
+ 12(15000)a70
(12)02

− 12(2854.95) ( a70
(12)00
+ a70
(12)01
+ a7(12)02
0 )
= 183, 563

(b)(ii) Now we have

5 V (1) = 12 × 7500 × a70

(12)11
+ 12 × 15000 × a70
(12)12
− 12 × 2854.95 × ( a70
(12)11
+ a70
(12)12
)
= 712,340

(c) By recursion, we have

( 11
4 12 )
V (0) + M − 3000 (1.05) 12 = 1 p69
00 1

11 5V
(0)
12
+ 1 p69
12
01
11 5V
(1)
12
+ 1 p69
12
02
11 5V
(2)
12 12

5 V (2)
12 × (15000 − 2854.95) × a
= (12)22
70 =
1,340, 245
V (0) = ( 0.94937 (183,563) + 0.00906 ( 712,340 ) + 0.00003 (1,340, 245 ) ) v 12 + 145.05
1
⇒ 11
4 12

= 180,175

(d)(i) The equation of value for F is now

0.25 ( 600,321.6 + 0.5 FA65
F= (12)03
)=
150, 080 + 0.04501F
⇒F=
157,153
(d)(ii) The equation of value for M is now
12 M (a65
(12)00
+ a65
(12)01
a65
+= (12)02
) 0.75(600321.6 + 0.5 FA65
(12)03
)
⇒M = 2989.52

Page 21
Question 12
(a) 1) To better replicate the loss to the injured party (IP):
- lost wages, medical and other expenses due to injury, offset inflation (if
increasing annuity).
2) To relieve the IP from the investment/interest risk, or from the burden of managing
funds.
3) To reduce the dissipation risk:
- risk of running out of funds from squandering/over spending.

(b) 1) Tax :
The structured settlement annuity might not be taxed, or taxed at a lower rate than
salary. Less than 100% is needed to replace the net pre-injury earnings.

2) Incentive to return to work:

The IP may choose to return to work if that would increase his/her (net) income.

3) IP at fault:
The payments may be reduced if the IP is (partially) at fault.

Graders’ Comments: Performance on parts (a) and (b) was mixed. Some candidates
did not answer those two parts, others provided only one relevant reason for each part
and/or repeated the same reasons for both parts. A number of candidates discussed
pricing or underwriting issues, in one or both parts, for which no credit was awarded.

(c) The benefits are 90,000 when in States 0 or 2 plus 20,000 when in State 0 for up to 2
years. So,
( )
EPV = 90, 000 ax00 + ax02 + 20, 000 ax:2 .
Since returning to State 0 after leaving it is impossible, we have

ax : 2 =
ax − v 2 px ax + 2 =
00 00 2 00 00
0.543444
EPV 90, 000 ( 0.559 + 7.161) + 20, 000(0.543444)
⇒= = 705, 669

Page 22
(d)
(i) =
2V
(0)
90, 000 ax00+ 2 += (
ax02+ 2 698, 670 )
(ii) 2V (2) = 90, 000 ax22+ 2 = 956, 070

(iii) Let E 0 [ 2V ] denote the EPV at t = 0 of the policy value at t = 2 . Note that there
is no policy value in State 1 or State 3. Then
E 0 [ 2V ] = v +v = 564, 054
2 00 (0) 2 02 (2)
2 px 2V 2 px 2V

(iv) The EPV at t = 0 of payments in the first two years plus the EPV at t = 0 of the
policy value at t = 2 is equal to the EPV at t = 0 of all future payments. That is
EPV(First two years payments) = EPV − E0 [ 2V ] =
141, 615

Graders’ Comments: Most candidates did well on this part. Some candidates calculated
the EPV in (iv) directly, which was an acceptable alternative solution.

(e)
(i) The reserve at time 2 given State 2, 2V (2) will stay the same.
This reserve is conditional on IP being in State 2 and transitions out of State 2 are
not affected.

(ii) Increasing µ x01+t will lead to a decrease in t px00 as lives are more likely to move
from State 0 to State 1 in the interval from time 0 to time t. It will also lead to a
decrease in t px02 , as it means that lives are more likely to recover, and so less
likely to become permanently disabled. That means that 2V (0) will be smaller,
= E 0 [ 2V ] v 2 px 2V + v 2 px 2V will also be smaller.
2 00 (0) 2 02 (2)
and hence

Graders’ Comments: Most candidates determined that the reserve would stay the same
in part (i). Part (ii) proved to be more challenging. Candidates were expected to address
each of the terms in E 0 [ 2V ] .

Page 23
Question 13
(a) EPV(Death Benefit) = 50, 000 A55(12)02 = 24, 000
0.95G a55
EPV(Premiums − Commissions) = 00
= 8.3904G
⇒ 24, 000 − 8.3904G =
−0.1G
⇒G = 2894.9

(b) 10 V (0) = 50, 000 A65(12)02 − 0.95Ga65

00
= 16,805

( )
1
+ 0.95G (1.05)12 = + 121 p65 10121 V + 121 p65 (50, 000)
(0) 00 (0) 01 (1) 02
(c) (i) 10 V 1 p65 10 1 V
12 12

p65 =
1 − 121 p65 − 121 p65 =
00 01 02
1
12
0.99246
⇒10121 V = 19, 478
(0)

( )
1
(ii) 1
10 12 V (0) (1.05)12 = 1
12
00
p65 1 10 1 V
12 12
(0)
+ 121 p65121 10121 V + 121 p65121 (50, 000)
01 (1) 02

p65121 =
1 − 121 p65121 − 121 p65121 =
00 01 02
1
12
0.99238
⇒10122 V = 19,398
(0)

Graders’ Comments: The majority of candidates got part (i) correct. Fewer candidates
got part (ii) correct with the most common error being to include the premium.

(d) (i) 2
10 12 V (0) is the EPV of future benefits – the EPV of future premiums net of
commissions for a life in State 0 at t = 10 122 . The EPV of future benefits is
(0)
unchanged, the EPV of future premiums net of commissions is smaller, so 2
10 12 V
will increase.

(ii) 10122 V (1) is the EPV of future benefits – the EPV of future premiums net of
commissions for a life in State 1 at t = 10 122 . The change in commissions does not
affect this calculation, as no premiums are paid once the life transitions into State 1,
as there are no transitions back to State 0.

Graders’ Comments: Candidates did not do as well on this part. Many candidates
assumed that the reserve in part (i) would decrease. It is true that the accumulated past
premiums minus commissions would be smaller, but prospectively, the EPV of future
outgo minus income would be bigger. In (ii) most candidates correctly stated that the
reserve would not change but did not offer a sufficient explanation.

Page 24
Question 14
4
 1 (τ )   1 (1)    1 (1)  
(a) p (τ )
40 exp − ∫ µ40+t dt  =
= exp −4 ∫ µ40+t dt  =
 exp − ∫ µ40+t dt  
 0   0    0 
 1 (1)   B 40 
exp − ∫ µ40 + t dt  = exp − A − ′(1) 0.9988415
c (c − 1)  ==
p40
 0   log c 
(using Makeham's survival probability formula)
(τ )
( p40′(1) ) = 0.995374
4
⇒ p40 =

Graders’ Comments: Candidates did well on this part. The most common error was treating
decrements as independent instead of dependent. While this resulted in the same answer to four
decimal places, points were deducted for an incorrect calculation. Another common error was
assuming that the force of decrement was constant.

(b)
2 2
EPV 100, 000∑ v k +1 ⋅k =
= px(τ ) ⋅ qx(τ+)k 100, 000∑ v k +1 ( k px(τ ) − k +1 px(τ ) )
=k 0=k 0

(1 − 0.99537)v + (0.99537 − 0.99027)v 2 

EPV = 100, 000  
 + (0.99027 − 0.98462)v 3 
= 100, = 000(0.013916) 1391.60

(i)
EPV(premiums) =
P(1 + 0.99537v + 0.99027v 2 ) =
P(2.84618)

1391.60
=P = 488.93
2.846179

(c)
t

t q40 = ∫ s p40 µ40 + s ds

(1) τ
( ) (1)
(i)
0
t t

∫=
p µ ds 4 ∫
(τ ) (τ ) (τ ) (τ ) (1)
(ii) =q t 40 s 40 40 + s s p40 µ40+ s ds
0 0

1 (τ )
= 4 t q40 ⇒ t q40 =
(1) (1)
t q40
4

Page 25
 k =2
(iii) EPV(Disease 1 rider) = 50, 000∑ v
k +1 (1)
k| q40
k =0

50, 000 k = 2 k +1 (τ )
= ∑ v k| q40
4 k =0
50, 000 2 k +1 (τ )
= ∑
4 k =0
v (k px −k +1 px(τ ) )

50, 000 × 0.013916

= = 173.95 from (b)
4
173.95
Rider=
Premium P=
r
= 61.12 (denominator from (b))
2.8462

Graders’ Comments Part (i) was done well. In part (ii), candidates who assumed constant
forces of decrement were given partial credit. Many candidates skipped (iii), but those who
attempted it did well.

= (
r
)
(d) 0V + P + P (1.05 ) 100, 000q40 + 50, 000q40 + p40 1V
(τ ) (1) (τ )

⇒ 1V = 57.39

Graders’ Comments: Most candidates who attempted this part calculated 1V using the EPV
of future outgo minus income. This is a longer method and tricky to get right, but is an
acceptable alternative.

Page 26
Question 15
20, 000 Ax:10 = 3020
02
(a) EPV(Death Benefit):
EPV(Sickness Benefit): 1000 ax01:10 = 684
EPV(expenses): (
60 ax00:10 + ax01:10 = 391.68 )
EPV(Premiums – Commission): 0.95 Pax00:10 = 5.5518 P
4095.68
=
So P = 737.72
5.5518
(b)
Pr(death within 10 years without becoming sick) is
1 − e−
10 10 0.5

∫0 ∫0
− (0.03+ 0.02) t
= t p
00
x µ 02
x + t dt =
0.0 2 e d t 0.02 = 0.15739
0.05

(c) The elimination period or waiting period is the time between the beginning of a period of
disability/sickness and the start of the disability income payments.
1 1 1

(d) a = ∫ t p e dt
= 11
x:1
11
x
−δ t
∫=
e e dt ∫ e
− (0.01+ 0.05) t −0.07 t −0.13t
dt
0 0 0
−
1 − e 0.13
= = 0.93773
0.13
(e) The only change is the duration of the sickness benefit.

( )
10
= benefit) 1000 ∫ t px µ x +t ax +t − ax +t:1 e−δ t dt
00 01 11 11
EPV(Sickness
0

∞ ∞
1
∫ ∫e
−δ s −0.13 s
where =
ax11+t s px + t e =
11
ds =
ds = 7.6923
0 0
0.13

⇒ EPV(Sickness benefit)

( )
10
1000(7.6923 − 0.93773)1000 ∫ t px00 µ x01+t ax11+t − ax11+t:1 e−δ t dt
=
0

10
= 6754.58 ∫ t px µ x +t e dt = 6754.58 Ax:10 = 1182.05
00 01 −δ t 01

3020 + 1182.05 + 391.68

=⇒P = 827.43
5.5518

Graders’ Comments: This part was challenging, and few candidates achieved full
marks, but many candidates achieved partial credit for answers that indicated some
understanding.

Page 27
Question 16
A60:10 =
A60 − 10 p60 v A70 − 10 p60 v A70 =
02 02 00 10 02 01 10 12
(a) (i)
−10 −10
0.39077 − (0.75055)(1.05) (0.54335) − (0.13135)(1.05) (0.62237)
= 0.09022
(ii) 10V = 500, 000 A60:10 + 100, 000 A60:10 − 4850 a60:10
= 21,850.4
(0) 02 01 00

a60:10 =
a60 − 10 p60 v a70 =
00 00 00 10 00
7.1461
=A60:10 A=
60 − 10 p60 v A70
01 01 00 10 01
0.11397

= =
(1) 12
(iii) 10V 500, 000 A60:10 95, 700
= A60 −10 p60 v A70
12 12 11 10 12
where A60:10
−10
= 0.47904 − 90.75283)(1.05) (0.62237)
= 0.19140

(b)
d (0)
dt
tV= δ tV
(i )
(0)
+ P − µ x +t 100, 000 + tV − tV
01 (1) (0)
(
− µ x +t 500, 000 − tV
02 (0)
) ( )
At t = 10 :
d (0)
= tV ln(1.05)(21,850) + 4850
dt
− (0.00818)(100, 000 + 95, 700 − 21,850)
− (0.00724)(500, 000 − 21,850)
= 1032.2

(ii )
d (1)
dt
=tV δ tV − µ x +t 500, 000 − tV
(1) 12
((1)
)
At t = 10 :
d (1)
tV = ln(1.05)(95, 700) − (0.01811)(500, 000 − 97,500)
dt
= −2652.8

(iii) In State 0, the cash flows in are interest and premiums, and the cash flows out are
claims and reserves on transition to State 1, and claims on transition to State 2.

In the early years of the term insurance the State 0 reserve will gradually increase, as
premiums exceed expected outgo. In later years the reserve will decrease, as the greater
number of claims outweighs the premium and investment income. At time 10, the
reserve is still in the increasing phase.

Page 28
 If a policy is in state 1 it is effectively a single premium term insurance. There is a little
interest income, but no premium income. The outgo on claims, on average, depletes the
reserve, which will decline to 0 just before the end of the contract.

Graders’ Comments: Candidates found part (b) challenging. Many candidates skipped
part (iii).

(c)
20
t + 0.5
100, 000 ∫ t px µ x +t
00 01 11
0.5 px + t v dt
0

Graders’ Comments: Candidates who attempted this part received the majority of the
points.

(d)
(i) Under the Markov property, for a policy in State 1 at time t, the probability of moving
from State 1 to State 2 in any future time interval is independent of the history of the state
process before time t.

This means, that (for example) the probability of a life dying in the next 6 months for a
policyholder currently in State 1, is the same if they were diagnosed, say, in the last week
as if they were diagnosed several years ago.

(ii) This is inconsistent with the spike in mortality observed after diagnosis, as
this implies a higher mortality rate for lives recently arrived in State 2, compared with
lives who have been in State 2 for a long time.

Graders’ Comments: Many candidates skipped this part.

Page 29
Question 17
(a) The symbol qxy is the probability that at least one of two lives, currently age x and y, dies
within one year.

Graders’ Comments: Part (a) was well done by almost all candidates. For full credit,
candidates were required to specify the 1-year time period for the mortality probability.

(b) The function is

Graders’ Comments: Most candidates did well on this part. Candidates who sketched a non-
linear function for tpx, or who sketched a different probability, or who sketched a line from 1 to
0, received no credit for this part.

Page 30
(c)
s qxy =
1 − s p x ⋅s p y (independence)
= 1 − (1 − s qx )(1 − s q y ) = 1 − (1 − s ⋅ qx )(1 − s ⋅ q y ) (UDD)
= s (qx + q y ) − s 2 qx q y

Also qx + q y = qxy + qxy and qxy = qx q y

⇒ s qxy = s (qxy + qxy ) − s 2 qxy

s q=
xy s (qxy ) + ( s − s 2 )qxy

⇒ g (s) =
s − s2

Graders’ Comments: Performance on this part was mixed. Many candidates received full
credit. A number of candidates made a reasonable start but could not derive the final result.
Partial credit was awarded in these cases. The most common error was assuming that UDD
applied to the joint life status.

Page 31
Question 18
(a) The future lifetimes of ( x) and ( y ) are dependent, because the force of mortality for each
is different depending on whether the other is alive or not, as µ x01+t: y +t ≠ µ x23+t and
µ x02+t: y +t ≠ µ 13
y + t . That means that the distribution of the time to death of
( x) is different if
( y ) dies early than if ( y ) dies later (and vice versa), which means the future lifetime
random variables Tx and Ty are dependent.

Graders’ Comments: The key point, is that the force of mortality for each life is different
from the married state than from the widowed state. Many candidates wrongly claimed that
the lives are independent because there is no common shock transition.

(b) (i)
−t pxy00 ( µ x01+t: y +t + µ x02+t: y +t ) with boundary condition of
d
t p xy = pxy00 =
00
0 1
dt

(ii)
− r pxy00 ( µ x01+ r: y + r + µ x02+ r: y + r ) ⇒ − ( µ x01+ r: y + r + µ x02+ r: y + r )
d 1 d
r pxy00 = 00 r pxy00 =
dr r p xy dr

⇒
d
dr
ln ( r pxy00 ) =
− ( µ x01+ r: y + r + µ x02+ r: y + r )

Integrate both sides from 0 to t

t t t
d
∫0 dr ln ( r p xy ) dr = ∫ − ( µ x + r : y + r + µ x + r : y + r ) dr ⇒ ln ( t p xy ) − ln ( 0 p xy ) =
00 01 02 00 00
∫ − ( µ x+r: y +r + µ x+r: y +r )dr
01 02

0 0

from the boundary condition, ln ( 0 p 00

xy ) = 0, so
( )
t
ln ( t pxy00 ) =
∫ − ( µ x+r:y +r + µ x+r:y +r )dr ⇒ t pxy =
01 02 00
exp − ( µ x01+ r: y + r + µ x02+ r: y + r ) dr
0

Graders’ Comments: Most candidates scored partial credit for this part. Some candidates
wrote down a few steps, but missed the key parts of the proof. An acceptable alternative
approach was to plug the given solution into the Kolmogorov equation and demonstrate that
the integral equation for t pxy00 satisfies the Kolmogorov differential equation and the
boundary condition. This approach was awarded full points if done correctly.

Page 32
(c) (i) At time 10, (x) is age 60 and (y) is age 65. The value at time 10 of the joint and
reversionary annuities is
00
50, 000a60:65 + 30, 000a60:65
01
+ 30, 000a60:65
02

= (50, 000)(8.8219) + (30, 000)(1.3768) + (30, 000)(3.0175) = 572,924

(ii) The EPV of the deferred annuity for the case where both lives survive 10 years uses
the results from (i):

=
10 p50:55 ⋅ v ⋅ 572,924
00 10
= 302, 627
(0.86041)(0.61391)(572,924)

The EPV of the deferred annuity for the case where (y) survives 10 years but (x) does
not, is

10 =01
p50:55 ⋅ v10 (30, 000a6511 ) (0.04835)(0.61391)(30,
= 000)(10.1948) 9078

The EPV of the deferred annuity for the case where (x) survives 10 years but (y) does
not, is

10 =02
p50:55 ⋅ v10 (30, 000a6022 ) (0.08628)(0.61391)(30,
= 000)(11.8302) 18, 799

Let P denote the premium. Then the EPV of the benefit paid on the second death during
the deferred period is

=
3PA0350:55:10 3=
P(0.003421) (0.010263) P

⇒=P 302, 627 + 9078 + 18, 799 + 0.010263P

330,504
=
⇒P = 333,931
0.989737

Graders’ Comments: Only the top few candidates achieved full credit for this part. Many
candidates did not allow for the possibility that only one life would survive the deferred
period.

(d) (i)
10 V (0) = 572,924 from (c)(i)
=
10 V
(1)
=
30, 11
000a65 = 305,844
(30, 000)(10.1948)
=
10 V
(2)
=
30, = 354,906
000a6022 (30, 000)(11.8302)

(ii) For t ≥ 10 and δ = log (1.05 )

= δ tV (0) − 50, 000 − µ5001+t:55+t ( t V (1) − tV ( 0 ) ) − µ5002+t:55+t ( t V (2) − tV (0 ) )

d (0)
tV
dt

Page 33
(iii) We need µ60:65
01
=+ 0.009076 and µ60:65
A Bc 60 = 02
=+
A Bc 65 =
0.015919 . Then

t +h (
V (0) ≈t V (0) + h δ tV (0) − 50, 000 − µ50
01
+ t :55 + t ( t V
(1)
−t V (0) ) − µ50
02
+ t :55 + t ( t V
(2)
−t V (0) ) )
≈ 572,924 + 0.5{(0.048790)(572,924) − 50, 000 − (0.009076)(305,844 − 572,924)
− (0.015919)(354,906 − 572,924)}

= 564,848

Graders’ Comments: As in (c), only the strongest candidates achieved full credit for this
part. Many students correctly calculated the three policy calues, though some did not
apply the correct annuity values ti the state-dependent polict values.

Page 34
Question 19
(a) The equation of value is
Pa40:40:20 = 100, 000 A40:40
=
where A40:40 2=
A40 − A40:40 0.08157
and a40:40:20 = 2a40:20 − a40:40:20 =13.0842
⇒P=
623.42
Graders’ Comments: Candidates did very well on this part, most of them achieving full
credit.

(b)
• Common shock: both lives may die at the same time, e.g. as a result of an
accident.
• Common lifestyle: couples tend to share a common lifestyle, e.g. healthy
(exercising) or unhealthy (smoking) habits.
• Broken heart syndrome: mortality may increase temporarily after the death of
one’s partner.

Graders’ Comments: Almost all candidates stated at least one valid reason, and most
candidates received full credit for this part.

(c)
(i) A40:40 = 1 − d a40:40
=a40:40 a= 60:60
40:40:20 + 20 E40:40 a
SULT
17.6836
0.05
⇒ A40:40 = 1− (17.6836) = 0.15792
1.05
0.35912
(ii) 20 E40:40 =10 E40:40 × 10 E50:50 ⇒ 10 E40:40 = = 0.60570
0.59290
12.9254 − 8.0703
(iii) a40:40:20 = a40:40:10 + 10 E40:40 a50:50:10 ⇒ a50:50:10 = = 8.0157
0.60570
Graders’ Comments: Performance on this part was mixed. Many candidates did not
correctly incorporate the independent SULT functions applying after 20 years, and some
did not identify or utilize the relationship between the A and ä functions.

(d) We now have

= A40:40 2=
A40 − A40:40 0.08420
a40:40:20 = 2a40:20 − a40:40:20 =13.0616
⇒ P = 644.62
Page 35
(e)
(i ) E [ 10 L | only 1 life alive ] = 100,000 A50 − Pa50:10
= 13,738.6
(ii ) E [ 10 L | both alive ] = 100,000 A50:50 − Pa50:50:10 = 8, 223.3
(iii ) E [ 10 L | at least 1 alive ]
E [ 10 L | only 1 life alive ] × Pr [ only 1 life alive ] + E [ 10 L | both alive ] × Pr [ both alive ]
=
Pr [ at least one alive]
Pr [ Both alive ] = 10 p40:40 = 0.9866
Pr[only 1 alive] = 10 p40:40 − 10 p40:40 =
0.0114
Pr [ At least 1 ali
= ve ] =
p40:40 0.998
10

⇒ E [ 10 L | at least 1 alive ] =
8, 286.2
Graders’ Comments: Most candidates found this part challenging. Many received at
least partial credit for (i) and (ii). Only a few candidates received full credit for part (iii).

(f) The policy value used should correspond most closely with the information available.
The insurance company knows that no claim has been made, so it knows that both lives
have not died, but it does not know whether one or both lives are alive. Correspondingly,
it should hold the reserve based on the policy value in (iii), which is suitable for the
information that at least one life has survived. For each individual policy, this will be the
wrong number – the correct number will either be the value in (i) or the value in (ii), but
overall the value in (iii) will be most accurate in aggregate.
Graders’ Comments: Only the stronger candidates correctly identified the most suitable
value for the policy value and provided an adequate explanation.

Page 36
Question 20
(a) Common shock:
Simultaneous deaths represented by transitions from State 0 to State 4.
Broken heart syndrome:
Mortality of surviving life is higher than individual mortality when both are alive,
µ13>µ02 and µ23>µ01.

Graders’ Comments: Performance on this part was mixed. Many candidates identified
only one way that dependency is incorporated in the model.

(b)

( ) ( )
10 10 1.05
(i) 23
1 − exp − ∫ 1.05 µ50
10 p50 =
*
1 − exp − ∫ µ50
+ t dt =
*
+ t dt
0 0

= 1 −( 10 p50 )
1.05
where 10 p50 is from the SULT
= 0.020678

(ii) 10
00
p40:50 (
exp − ∫ µ40
= 01 10
+ t :50 + t + µ 40 + t :50 + t + µ 40 + t :50 + t dt
0
02 0 4
)
exp − ∫
= ( 10

0
(µ *
50 + t ) (
− 0.0005 + µ40
*
)
+ t − 0.0005 + 0.0005 dt )
=e0.005 10 p40 10 p50 = 0.977654

Graders’ Comments: Only the most well-prepared candidates achieved full or nearly
full credit on this part. Partial credit was awarded to candidates who showed some
understanding of the question by writing down some formulas relevant to the calculation
of these probabilities of transition..

00 03
(c) Note that we must use the given information, that 𝑎𝑎40:50:10 = 7.8487 and 𝐴𝐴40:50:10 =
0.00789. The equation of value is

=
Pa40:50:10 100, 000 A40:50:10 + 300, 000 A40:50:10
00 03 04

10 10
= ∫ t p40:50 µ40+t:50+t e− dt = 0.0005 ∫ t p40:50 µ40+t:50+t e− dt
04 00 04 δt 00 04 δt
A40:50:10
0 0

= 0.0005 a 00
40:50:10 = 0.00392435
⇒P=
250.53
Graders’ Comments: Most candidates were able to write down the formula needed to
calculate the premium. Some candidates were also able to write down a correct
04
expression for 𝐴𝐴40:50:10 but only a few recognized that it could be calculated from the
00
given value for 𝑎𝑎40:50:10 .

Page 37
 00
(d) (i) 𝑎𝑎𝑥𝑥:𝑦𝑦:10 : Stays the same.
The transition intensities leaving State 0 do not change and it is impossible
to return to State 0 after leaving it.
03
(ii) 𝐴𝐴𝑥𝑥:𝑦𝑦:10 : Would be higher.
If (y) dies first, no impact. If (x) dies first, (y) is likely to die sooner which
will increase the EPV of the insurance benefit.
02
(iii) 𝑎𝑎𝑥𝑥|𝑦𝑦 : This corresponds to 𝑎𝑎𝑥𝑥:𝑦𝑦 in this model. Would be lower.
This annuity is payable while (y) is alive after the death of (x). Since in
this case, (y) is likely to die sooner, the EPV of the annuity payments will
decrease.

(iv) Premium: Would be higher.

00 04 03
Since 𝑎𝑎𝑥𝑥:𝑦𝑦:10 and 𝐴𝐴𝑥𝑥:𝑦𝑦:10 would stay the same but 𝐴𝐴40:50:10 would
increase, the premium in (c) would increase.
Graders’ Comments: Most candidates received partial credit for correctly justifying the
impact of the change on some of the four given actuarial functions. A small number of
candidates adequately justified the impact of the change on all four functions.

Page 38
 Question 21
(a) Due to independence:
2
 l80   75, 657.2 
2

10=
p70:70 (10=
p70 ) 2 =   =  0.689972
 70 
l  91, 082.4 

(b) 𝐸𝐸[𝐿𝐿0 ] = 100,000𝐴𝐴70:70:10 + 400 − 𝐺𝐺(0.98 𝑎𝑎̈ 70:70:10 − 0.48) where

𝑎𝑎̈ 70:70:10 = 7.2329 (from SULT)
1 − da70:70:10 =
A70:70:10 = 1 − (0.04762)(7.2329) =
0.65557

Then 𝐸𝐸[𝐿𝐿0 ] = 65,557 + 400 − 10,658 (0.98(7.2329) − 0.48)

= −4,474

(c)

2
(i) 𝐴𝐴70:70:10 = 2𝐴𝐴70:70 + 10𝑝𝑝70:70 𝑣𝑣 20 (1 − 2𝐴𝐴80:80 )
= 0.30743 + 0.689972 (1.05)−20 (1 − 0.50165)
= 0.437

(ii) 𝐿𝐿0 = 100,000 𝑣𝑣 𝑚𝑚𝑚𝑚𝑚𝑚(𝐾𝐾70:70 +1,10) + 400 − 𝐺𝐺 �0.98 𝑎𝑎̈ 𝑚𝑚𝑚𝑚𝑚𝑚(𝐾𝐾70:70 +1,10) − 0.48�

  1 − v min( K70:70 +1,10)  

=L0 100, 000v min( K70:70 +1,10) + 400 − G  (0.98)   − 0.48 
  d  

0.98 𝐺𝐺 0.98
𝐿𝐿0 = �100,000 + � 𝑣𝑣 𝑚𝑚𝑚𝑚𝑚𝑚(𝐾𝐾70:70 +1,10) + 400 − 𝐺𝐺 � − 0.48)�
𝑑𝑑 𝑑𝑑

0.98 𝐺𝐺 2
𝑉𝑉𝑉𝑉𝑉𝑉[𝐿𝐿0 ] = �100,000 + � 𝑉𝑉𝑉𝑉𝑉𝑉�𝑣𝑣 𝑚𝑚𝑚𝑚𝑚𝑚(𝐾𝐾70:70 +1,10) �
𝑑𝑑

0.98 𝐺𝐺 2 2
= �100,000 + � ( 𝐴𝐴70:70:10 − ( 𝐴𝐴70:70:10 )2 ) = 737,080,000
𝑑𝑑

𝑆𝑆𝑆𝑆[𝐿𝐿0 ] = 27,150

Page 39
(d)
Let 𝐿𝐿0,𝑖𝑖 be the loss at issue for policy i, i=1,2,…,100.
100
The aggregate loss at issue is L = ∑ L0,i .
i =1

Then, 𝐸𝐸[𝐿𝐿 ] = 100 𝐸𝐸[𝐿𝐿0 ] = −447,400

𝑉𝑉𝑉𝑉𝑉𝑉[𝐿𝐿 ] = 100𝑉𝑉𝑉𝑉𝑉𝑉[𝐿𝐿0 ]
𝑆𝑆𝑆𝑆[𝐿𝐿 ] = 10𝑆𝑆𝑆𝑆[𝐿𝐿0 ] = 271,500

0 − (−447,400)
𝑃𝑃[𝐿𝐿 > 0] ≈ 1 − Φ � �
271,500
≈ 1 − Φ(1.648)
≈ 0.05

(e)

The total profit realized in this sub-portfolio will be greater than its expected value.

Justification:

There were two death benefits paid during the term of the contract and 8 endowment
benefits paid at maturity (time 10).

The expected number of death benefits in the sub-portfolio was 10 (1 − 10𝑝𝑝70:70 )

which represents 3.1 death benefits.

The EPV of the 2 death benefits was lower than expected.

Or
The mortality experience in this sub-portfolio was better than expected and the death
benefits were paid relatively late in the term of the policy (at times 6 and 9).

Consequently, the EPV of the benefits was lower than expected.

Graders’ Comments: The majority of candidates did well on parts (a) to (d) with many
receiving full credit. Part (e) proved much more challenging, with few candidates receiving
full credit.

Page 40
Question 22
(a)
The probability is:
Pr [1 alive ] =
2 × 20 p50 (1 − 20 p50 )

20 p=
50
l=
70 91, 082.4
= 0.92398
l50
98,576.4
⇒ (2)(0.92398)(1 − 0.92398) =
0.1405
(b)

a50:50:20
= 2a50:20
(2) (2)
− a50:50:20
(2)

1
a=
(2)
50:20
(1 − 20 E50 )= 12.8428 − 0.25(1 − 0.34824) = 12.6799
a50:20 −
4
 1  1
a50:50:20
= a50:50 − 20 E50 × a70:70 =  a50:50 −  − ( 20 p50 ) (1.05) −20  a70:70 − 
(2) (2) (2) 2

 4  4
−20
(15.8195 − 0.25) − (0.92398) (1.05) (9.9774 − 0.25) =
2
12.4396
⇒ a50:50:2
= (2)
0
(2)(12.6799) − 12.4396 = 12.9202

Graders’ Comments: Many students attempted this part, and most of them received
partial credit for correct intermediate steps. The most common errors were (i) incorrect
application of Woolhouse’s formula; (ii) using a joint endowment instead of a last
survivor when calculating the adjusted last survivor term annuity; (iii) Incorrect
Woolhouse coefficients (11/24 or ½ instead of 1/4)

(c)
EPV Benefits:
100, 000 A50:50 = 100, 000(2 A50 − A50:50=) 100, 000(2 × 0.18931 − 0.24669)
= 13,193

EPV Prems − Exp:

{
G ( a50:50:20
(2)
}
+ 0.1405v 20 a70 + 20 p50:50 v 20 a70:70 ) × 0.9 − 0.7
a70:70 =2 × a70 − a=
70:70 (2)(12.0083) − 9.9774
= 14.0392

Page 41
 {
− Exp G (12.9202 + 0.1405 × v 20 ×12.0083 + (0.92398) 2 (v 20 )(14.0392) ) 0.9 − 0.7
⇒ EPV Prems= }
= G {0.9 × 18.0734 − 0.7}= 15.5661G
13,193
=
⇒G = 847.55
15.5661
Graders’ Comments: Not many candidates received full credit, but most of those that
attempted this part received the majority of points for correctly setting up the EPV
formulas. Common mistakes included :
• Incorrect expense payment (60% vs. 70%)
• Missing the annuity portion due when exactly 1 person is alive
• Multiplying the first annuity term by 2 (incorrect treatment of semi-annual premium
payment)

(d)
(i) Let 20V(2) denote the reserve conditional on both surviving:

20 V (2) = 100000 A70:70 − (0.9)(847.55)a70:70

 0.05 
1 − da70:70 =
A70:70 = 1−   (14.0392) =
0.33147
 1.05 

or (2)A70 − A70:70 =(2)(0.42818) − 0.52488 = 0.33148

⇒ 20V (2) =
22, 437.6

(ii) Let 20V(1) denote the reserve conditional on exactly one surviving:

=
20V
(1)
100, 000 A70 − (0.9)(847.55)a70
= (100, 000)(0.42818) − (0.9)(847.55)(12.0083)
= 33, 658

Graders’ Comments: Candidates who attempted this part generally received the
majority of the credit. The most common mistake was neglecting the 90% coefficient on
the annuity portion of the calculation.

Page 42
(e) ( 19.5 V (2) + 0.9 × G / 2 ) (1.05 )
0.5
= (
0.5 q69.5 ) × 100, 000 + ( 0.5 p69.5 ) 20V
2

  
(2) 2

Both Insureds Die Both Insureds Live

+ 2 ( 0.5 q69.5 0.5 p69.5 ) 20V (1)

  
One Insured Lives and One Dies
The (2) in front is because either could die.

( p69 ) =
p69.5 = (1 − 0.009294)0.5 =
0.5
Also, 0.5 0.99534
⇒ 19.5V (2) =
21, 619

Graders’ Comments: Not many candidates attempted this part, and very few received full
credit. When they attempted to solve for the survival probability, students generally
performed the calculation correctly, the majority of mistakes came from the recursion
equation.

Page 43
Question 23
(a)

01
𝑎𝑎�xy is the actuarial value of a (whole-life) annuity paid to (y) starting on the death of (x),
payable continuously at a rate of 1 per year.

Or

A reversionary annuity paid to (y) starting on the death of (x), payable continuously at a
rate of 1 per year.

Graders’ Comments: Most candidates did very well. Some candidates didn't mention
"continuous", and a few candidates only provided a description that was not specific to the
given joint life model (e.g., "a continuous annuity payable at a rate of 1 per year while in
state 1 given that it is issued in state 0”)
(b)

𝑑𝑑 01 02
𝑝𝑝00 = − 𝑡𝑡𝑝𝑝𝑥𝑥𝑥𝑥
00
�𝜇𝜇𝑥𝑥+𝑡𝑡 + 𝜇𝜇𝑦𝑦+𝑡𝑡 �; 00
0𝑝𝑝𝑥𝑥𝑥𝑥 =1
𝑑𝑑𝑑𝑑 𝑡𝑡 𝑥𝑥𝑥𝑥

And

𝑑𝑑
𝑝𝑝01 = 𝑝𝑝00 𝜇𝜇 01 − 𝑡𝑡𝑝𝑝𝑥𝑥𝑥𝑥
01 13
𝜇𝜇𝑦𝑦+𝑡𝑡 ; 01
0𝑝𝑝𝑥𝑥𝑥𝑥 =0
𝑑𝑑𝑑𝑑 𝑡𝑡 𝑥𝑥𝑥𝑥 𝑡𝑡 𝑥𝑥𝑥𝑥 𝑥𝑥+𝑡𝑡

Graders’ Comments: Candidates did well on this part. A common error was incorrect or
missing subscripts (i.e., ages). A few candidates didn't provide the boundary conditions.

Page 44
(c)
We have
∞ ∞
00 00 −𝛿𝛿𝛿𝛿
𝑎𝑎�𝑥𝑥𝑥𝑥 =� 𝑡𝑡𝑝𝑝𝑥𝑥𝑥𝑥 𝑒𝑒 𝑑𝑑𝑑𝑑 = � 𝑔𝑔(𝑡𝑡) 𝑑𝑑𝑑𝑑
0 0
and i.e. 𝑎𝑎� is the integral, 𝑎𝑎̈ the sum
∞ ∞
𝑎𝑎̈ 𝑥𝑥00𝑥𝑥 = � 00
𝑘𝑘𝑝𝑝𝑥𝑥𝑥𝑥 𝑒𝑒 −𝛿𝛿𝛿𝛿 = � 𝑔𝑔(𝑘𝑘)
𝑘𝑘=0 𝑘𝑘=0
where
00 −𝛿𝛿𝛿𝛿
𝑔𝑔(𝑡𝑡) = 𝑡𝑡𝑝𝑝𝑥𝑥𝑥𝑥 𝑒𝑒 , 𝑔𝑔(0) = 1.

The derivative of g(t) using the product rule is

𝑑𝑑 00 −𝛿𝛿𝛿𝛿 00
𝑑𝑑
𝑔𝑔′(𝑡𝑡) = � 𝑝𝑝 � 𝑒𝑒 + 𝑡𝑡𝑝𝑝𝑥𝑥𝑥𝑥 � 𝑒𝑒 −𝛿𝛿𝛿𝛿 �
𝑑𝑑𝑑𝑑 𝑡𝑡 𝑥𝑥𝑥𝑥 𝑑𝑑𝑑𝑑

00 01 02
= − 𝑡𝑡𝑝𝑝𝑥𝑥𝑥𝑥 �𝜇𝜇𝑥𝑥+𝑡𝑡 + 𝜇𝜇𝑦𝑦+𝑡𝑡 00
�𝑒𝑒 −𝛿𝛿𝛿𝛿 + 𝑡𝑡𝑝𝑝𝑥𝑥𝑥𝑥 �−𝛿𝛿𝑒𝑒 −𝛿𝛿𝛿𝛿 �.

So, 𝑔𝑔′ (0) = −1 ∙ �𝜇𝜇𝑥𝑥01 + 𝜇𝜇𝑦𝑦02 � ∙ 1 + 1 ∙ (−𝛿𝛿 ∙ 1) = −�𝜇𝜇𝑥𝑥01 + 𝜇𝜇𝑦𝑦02 + 𝛿𝛿�

The approximation
∞ ∞ 1 1
� 𝑔𝑔(𝑡𝑡) 𝑑𝑑𝑑𝑑 ≈ � 𝑔𝑔(𝑘𝑘) − 𝑔𝑔(0) + 𝑔𝑔′(0)
0 𝑘𝑘=0 2 12

then gives

00 00 1 1
𝑎𝑎�𝑥𝑥𝑥𝑥 ≈ 𝑎𝑎̈ 𝑥𝑥𝑥𝑥 − 2 − 12 �𝜇𝜇𝑥𝑥01 + 𝜇𝜇𝑦𝑦02 + 𝛿𝛿 �.

Graders’ Comments: Candidates did fairly well, but not many received full credit. The
most common issue was the derivative, g'(t). A number of candidates mistakenly assumed
constant forces of mortality, receiving little credit.

(d)
Let B = annual annuity payment rate

00 01 02 )
1,000,000 = 𝐵𝐵 𝑎𝑎���������
70:70 = 𝐵𝐵(𝑎𝑎 �70:70 + 𝑎𝑎�70:70 + 𝑎𝑎�70:70
01 02
Since 𝑎𝑎�70:70 = 𝑎𝑎�70:70 = 𝑎𝑎�70|70 = 2.0317
00
1,000,000 = 𝐵𝐵 𝑎𝑎�70:70
�������� = 𝐵𝐵�𝑎𝑎
�70:70 + 2𝑎𝑎�70|70 �
Using the approximation in (c),

Page 45
 00 00
1 1
𝑎𝑎�70:70 ≈ 𝑎𝑎̈ 70:70 − − (2𝜇𝜇70 + 𝛿𝛿)
2 12
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆
1 1
= 𝑎𝑎̈ 70:70 − − (2(0.009881) + 0.04879) = 9.4717
2 12
1,000,000 1,000,000
 𝐵𝐵 = 9.4717+2(2.0317) = 13.5351
= 73,881.98

Graders’ Comments: Candidates did fairly well on this part. A few candidates mistakenly
applied the EMW approximation in Part c) for a last-survivor annuity. Note that the
formula for g(t) is different for different cases, the EMW approximation for a last-survivor
annuity would be different from c).

Page 46
Question 24
(a)

200,000 𝐴𝐴50:60:10 = 𝑃𝑃 𝑎𝑎̈ 50:60:10

𝑎𝑎̈ 50:60:10 = 7.9044
0.05
𝐴𝐴50:60:10 = 1 − 𝑑𝑑 𝑎𝑎̈ 50:60:10 = 1 − 1.05 (7.9044) = 0.6236

(200, 000) A50:60:10 (200, 000)(0.6236)

=P = = 15, 778.55
a50:60:10 7.9044

(b)

2
𝐴𝐴50:60:10 = 2𝐴𝐴50:60 + 𝑣𝑣 20 10𝑝𝑝50 10𝑝𝑝60 (1 − 2𝐴𝐴60:70 )

= 2𝐴𝐴50:60 + 10𝐸𝐸50 10𝐸𝐸60 (1 − 2𝐴𝐴60:70 )

= 0.3908335

Graders’ Comments: Candidates did well on these parts. Common errors were missing the
endowment part of the equation or using wrong discount factors for the endowment.

(c)

𝑃𝑃 2 2
𝑉𝑉𝑉𝑉𝑉𝑉[𝐿𝐿] = �𝑆𝑆 + � ( 𝐴𝐴50:60:10 − ( 𝐴𝐴50:60:10 )2 )
𝑑𝑑

= (531,349.55)2 (0.390834 − 0.62362 ) = (531,349.55)2 (0.001957)

= 552,524,514.1 = 23,505.842

Standard Deviation of L = 23,505.84

Graders’ Comments: Overall, candidates did well on this part also. Many candidates have
this formula memorized. Partial credit was awarded if there was an attempt to derive the
formula, if it wasn't memorized.

(d)
𝑆𝑆 𝐴𝐴50:60:10 124,720
𝐺𝐺 = = = 17,909.35
(0.9)𝑎𝑎̈ 50:60:10 − 0.15 6.96396

Page 47
 Graders’ Comments: Again, candidates did well on this part. The most common error was
the double counting first year expense (using a separate 0.25 instead and using .15)

(e)

𝑆𝑆 2𝐴𝐴50:60:10 200,000(0.390834)
(i) 𝐺𝐺 ∗ = (0.9) 2𝑎𝑎̈ = = 13,601.27
50:60:10 −0.15 5.747024

2 1− 2𝐴𝐴50:60:10
𝑎𝑎̈ 50:60:10 = 𝑑𝑑 ∗
= 6.552249

𝑑𝑑 ∗ = 0.1025/1.1025 = 0.0929705

A reduction of (17,909.35 − 13,601.27)/17,909.35 = 0.2405 or 24.05%

(ii) The percentage reduction would be less for a 10-year term insurance.
The endowment insurance also pays a benefit in case of survival to
time 10. The percentage reduction in the EPV of the survival benefit
will be larger than the percentage reduction in the EPV of the death
benefit (which is paid early, on average, and is therefore relatively
less affected by a change in the interest rate).

Graders’ Comments: For Part (i) the most common error was not calculating the correct
discount rate or calculating the net premium instead of the gross premium. For Part (ii), many
candidates did not really address the question answered and instead commented on expenses or
joint life versus single life as opposed to the difference between term and endowment.

Page 48
Question 25
(a)

The future lifetimes are not independent. The mortality of each life is different when their
partner is alive than it is when their partner is dead. This shows that the future lifetime of each
partner impacts the future lifetime of the other. OR: A necessary condition for independence
is that µ y = µ y and µ x = µ x . As this is not the case, the future lifetimes are not independent.
01 23 02 13

Graders’ Comments: Most candidates who correctly identified dependence gave good
reasoning. Many candidates failed to recognize dependence, assuming that no direct
transitions from State 0 to State 3 implied independence.

(b)
 t 01   t s 
=  ∫ x + r: y + r
− µ + µ =  − ∫ µ y + r + µ x + r − 0.005 dr 
00 02 s
p
t xy exp x + r: y + r dr  ex p
 0   0 
=e =
=> λ=
0.005t s s
t px t p y 0.005

Graders’ Comments: This is a “show that” question, where candidates need to be very precise
about their derivation. Overall candidates did well on this part. However, many candidates
assumed constant forces of mortality, which is incorrect and unnecessary.

(c)
∞ ∞
−δ t λ t −δ t
= ∫= ∫ px t = where δ * = δ − λ = 0.035
00 00 s s s
ax: y t px: y e dt t p y e e dt ax: y
δ*
0 0

⇒a = 19.3199
00
40:50

Graders’ Comments: The point of this part is to get candidates to work out the correct interest
rate to use. For full credit, they need to justify their choice. Many candidates overcomplicated
the question, spending time integrating, instead of using the table (this was especially true for
candidates who assumed constant forces of mortality).

(d)
∞
(i) A =
01
+A
x: y
02
x: y ∫
0
t
00
(
px: y µ y +t + µ x +t e dt
01 02
) −δ t

Page 49
 ∞

( )
*
−δ t
(ii) A=
x: y + Ax: y ∫ px t p y µ y +t + µ x +t − 0.005 e
01 02 s s s s
t dt
0

=− 1 − δ ax: y δ * − 0.005 ax: y δ *

Ax: y δ * 0.005 ax: y δ * =
s s * s s

1 − δ ax: y δ * =
= 1 − (0.04)(19.3199) =
s
0.227204

Graders’ Comments: Most candidates either skipped this part or assumed constant forces of
mortality.

(e)

=
500, 000 Ax: y + Ax: y
P =
( 01
(500, 000)(0.227204)
= 5,880.1
02
)
00
ax: y 19.3199

(f)
500, 000 500, 000 500, 000
=X = 11 s
= = 27, 791
a50 ax: y δ **=0.0425 17.9917

Graders’ Comments: Parts (e) and (f) were quite straightforward, relying on given answers
from earlier parts. However, many candidates who struggled with part (d) skipped these
parts.

Page 50
Question 26
(a)
=
EPV =
Benefits (100, = 15,161
000) A45 (100, 000)(0.15161)
= Pa
EPV Premiums 45 17.8162 P
=
EPV Expenses= 20a45 + 80 + 200 A45 + 0.1Pa45 + 0.65 P
= (20)(17.8162) + 80 + (200)(0.15161) + [(0.1)(17.8162) + 0.65]
= P 466.646 + 2.43162 P

15,161 + 466.646
P= = 1015.80
17.8162 − 2.43162

(b)
=
1V (100, 200) A46 − 0.9 Pa46 + 20a46
= (100, 200)(0.15854) − [(0.9)(1015.80) − 20](17.6706)
= 84.30

Graders’ Comments: Candidates could answer using the prospective approach or the
recursion approach.. The most common errors involved incorrect treatment of expenses.
Some candidates used the actual interest rate (1.07) instead of the pricing rate (1.05), which
led to partial credit if the rest of the solution was correct.

(c)
Interest Gain:
(10, 000)(0.07 − 0.05)[ P − (0.75 P + 100)] = 30, 790
Expense Gain:
(10, 000)[(0.75 P + 100) − (0.75 P + 105)](1.07) = −53,500
Mortality Gain: 0

Commentary:
Candidates were expected to indicate clearly whether the amount calculated is a gain or a
loss. A few candidates calculated only the total gain/loss; this received small partial
credit. One relatively common mistake was to use a factor of .07 instead of 1.07 when
calculating the expense gain and loss.

(d)
Interest: No gain or loss, as experience matches the assumption.
Expenses: Loss, as settlement expenses exceed the assumption.
Mortality: Expected deaths  10000 p45 q46  8.4 . There will be a loss, as actual deaths
exceed expected deaths

Page 51
Graders’ Comments: A few candidates compared the experience of year 2 with year 1. For
example, “interest in year 2 was lower than in year 1 so there was a loss due to interest''.
This did not directly answer the question and did not receive full credit. A few candidates
proposed that there would be a gain due to mortality because there were more deaths than
expected.

Page 52
Question 27
(a)
Pr2 = (1V + P − E )(1 + i2 ) − q51S − p51 ( 2 V )
= (400 + 1100 − 55)(1.02) − (0.00642)(100, 000) − (0.99358)(800) = 37.04

(b)
t t −1 V P Et It EDBt EMB EtV Prt
0 155 -155.00
1 0 1100 55 10.45 592 397.63 65.82
2 400 1100 55 28.90 642 794.86 37.04
3 800 1100 55 55.35 697 1092.33 0.00 111.02

The profit vector is the final column.

P is the premium
Et denotes expenses
It denotes interest on funds in year t
EDBt = 100, 000q50+t −1
=
In year 3, EMB3 p52 ×1100
EtV = p50+t −1 tV

Graders’ Comments: Many candidates received full credit for this part, and a larger number
received partial credit. The most common errors included (i) ignoring the maturity benefit and
(ii) incorrect use of probabilities in EtV. It is not necessary for candidates to define all terms, but
it can be helpful when graders are considering partial credit, especially if the values are
calculated in Excel.

(c) The profit signature at t is Π t where:

Π
= 0 Pr0 and = Π t t −1 p50 ⋅ Prt

⇒ (Π 0 , Π1 , Π 2 , Π 3 ) = (−155.00, 65.82, 36.82, 109.65)

3
⇒ NPV = ∑Π
k =0
k ⋅ v14%
k
= 5.08

Graders’ Comments: Most candidates did this part well. The most common errors were
incorrect survival probabilities when calculating the profit signature from the provit vector,
and incorrect discounting periods for the NPV.

Page 53
(d) The IRR of B is j where 155 = 210v 3j ⇒= j 10.65% .
and the IRR of A is greater than 14%, because the NPV at 14% is positive.
Hence IRR of B is less than IRR of A.

Product C has lower reserves in year 1,which allow an earlier release of surplus compared to
Product A, which give a higher NPV than A at an 14% hurdle rate, but does not necessarily
mean that C has a higher IRR.

The lower reserve in year 1 results in the following profit signature for C:
(-155.00, 165.23, -64.58, 109.65)

We note that the NPV of A is a little larger than 14%, because the NPV at 14% is close to
zero. Calculating the NPV for A and C at 16% gives −0.65 for A and 9.7 for C. Hence, the
IRR for C is greater than 16% , and for A is less than 16%.

That is ⇒ IRR( B) < IRR( A) < IRR(C )

Graders’ Comments: Many candidates evaluated the IRR for all three cases, presumably
using the financial functions on the BA2 calculator. This was awarded full credit if correct.
However, it was not necessary to determine the IRR to answer the question.
Candidates who demonstrated understanding of the relationship between the release of surplus
and the return to the insurer gained partial credit, even if the justification was incomplete.
No credit was awarded for the IRR ordering if there was no accompanying explanation or
justification

Page 54
Question 28
01 00 01 01 11
(a) 2𝑝𝑝60 = 𝑝𝑝60 𝑝𝑝61 + 𝑝𝑝60 𝑝𝑝61

= (0.9)(0.05) + (0.05)(0.2) = 0.055

Graders’ Comments:
1. This part was done correctly by almost all candidates.
2. For those who did not receive full credit, two common errors were to use an
incorrect annual probability of transition and to calculate only one of two terms.

(b)

(i) 2𝑉𝑉
(0)
= 50,000 𝐴𝐴02 00 01 00
62:8 + 150 𝑎𝑎̈ 62:8 + 150 𝑎𝑎̈ 62:8 − 5,000 𝑎𝑎̈ 62:8
= 50,000(0.46667) + 150(4.7328) + 150(0.2533) − 5,000(4.7328)
= 23,333.5 + 709.92 + 37.995 − 23,664.0 = 417.415

10 10
(ii) 2𝑉𝑉
(1)
= 50,000 𝐴𝐴12 11
62:8 + 150 𝑎𝑎̈ 62:8 + 150 𝑎𝑎̈ 62:8 − 5,000 𝑎𝑎̈ 62:8
= 50,000(0.4968) + 150(3.334 + 1.406) − 5,000(3.334)
= 24,840 + 711 − 16,670 = 8,881

(iii) � 2𝑉𝑉 (0) + 5,000 − 150�1.06 = 𝑝𝑝62

02 (50,000) 00
+ 𝑝𝑝62 3𝑉𝑉
(0) 01
+ 𝑝𝑝62 3𝑉𝑉
(1)

(417.415+4850)1.06=0.07(50,000)+0.88(1788) +0.05 3V(1)

(1)
(5583.46 − 3500 − 1573.44)
3 𝑉𝑉 = = 10,200.40
0.05

Alternatively
� 2𝑉𝑉 (1) + 0 − 150�1.06 = 𝑝𝑝62
12 (50,000) 10
+ 𝑝𝑝62 3𝑉𝑉
(0) 11
+ 𝑝𝑝62 3𝑉𝑉
(1)

(8881-150)1.06=0.12(50,000)+0.68(1788)+0.20 3V(1)
(1)
(9254.86 − 6000 − 1215.84)
3𝑉𝑉 = = 10,195.10
0.2

Graders’ Comments: Parts (i) and (ii) were done correctly by most candidates; for those
who did not receive full credit, a common error was to ignore maintenance expenses.
Although most candidates recognized that a recursive formula was needed to answer part
(iii), setting it up correctly proved to be challenging for many candidates.

Page 55
(c)
(0)
(i) 𝑃𝑃𝑃𝑃3 = � 2𝑉𝑉 (0) + 𝑃𝑃 − 60�1.057 − 𝑝𝑝62
00
3𝑉𝑉
(0) 01
− 𝑝𝑝62 3𝑉𝑉
(1) 02 (50,000)
− 𝑝𝑝62
= (417.415 + 5000 − 60)1.057 − (0.88)(1788) − (0.05)(10,200.4) −
(0.07)(50,000)
= 79.328
(1)
𝑃𝑃𝑃𝑃3 = � 2𝑉𝑉 (1) − 60�1.057 − 𝑝𝑝62
10
3𝑉𝑉
(0) 11
− 𝑝𝑝62 3𝑉𝑉
(1) 12 (50,000)
− 𝑝𝑝62
= (8881 − 60)1.057 − (0.68)(1788) − (0.2)(10,200.4) − (0.12)(50,000)
= 67.877

01
2𝑝𝑝60 = 0.055 from (a)
00
2𝑝𝑝60 = (0.9)(0.89) + (0.05)(0.69) = 0.8355
00 (0)01 (1)
π3 = 2𝑝𝑝60 𝑃𝑃𝑃𝑃3 + 2𝑝𝑝60 𝑃𝑃𝑃𝑃3
= (0.8355)( 79.328) + (0.055)( 67.877) = 70.012

(ii) 𝑁𝑁𝑁𝑁𝑁𝑁𝑡𝑡 = 𝑁𝑁𝑁𝑁𝑁𝑁𝑡𝑡−1 + 𝜋𝜋𝑡𝑡 𝑣𝑣 𝑡𝑡 = ∑𝑡𝑡𝑘𝑘=0 𝜋𝜋𝑘𝑘 𝑣𝑣 𝑘𝑘 ; 𝑁𝑁𝑁𝑁𝑁𝑁0 = 𝜋𝜋0 ; 𝑣𝑣 = 1/1.08

t π𝑡𝑡 𝑁𝑁𝑁𝑁𝑁𝑁𝑡𝑡

0 -200.00 -200.00
1 84.74 -121.54
2 80.35 - 52.65
3 70.01 2.93

Alternatively, with π3 = 70.291, we get 𝑁𝑁𝑁𝑁𝑁𝑁3 = 3.150

The Discounted Payback Period (DPP) is 3 years.

Graders’ Comments: For part (i), many candidates did not recognize that the expected
emerging profit depends on the state at the beginning of the period. Most candidates did
well on part (ii). Some candidates discounted the profit signature values at a rate other
than the hurdle rate.

Page 56
Question 29
(a) The emerging profit in year 3, conditional on being in State j at time 2 is
(𝑗𝑗) (𝑗𝑗) (𝑗𝑗) 𝑗𝑗𝑗𝑗
𝑃𝑃𝑃𝑃3 = � 2𝑉𝑉 (𝑗𝑗) + 𝑃𝑃3 − 𝐸𝐸3 �(1 + 𝑖𝑖𝑡𝑡 ) − � 𝐸𝐸 𝐵𝐵3 − � 𝐸𝐸3 𝑉𝑉𝑗𝑗𝑗𝑗
𝑘𝑘 𝑘𝑘
where 𝑖𝑖𝑡𝑡 = 0.06 .
For State 0,
(0)
2𝑉𝑉 = 12,000
(0)
𝑃𝑃3 = 30,000
(0) (0)
𝐸𝐸3 = (0.05)𝑃𝑃3 = 1500

𝐸𝐸 𝐵𝐵301 = 𝑝𝑝𝑥𝑥+2
01
600,000 = (0.014)(600,000) = 8400
02 02
𝐸𝐸 𝐵𝐵3 = 𝑝𝑝𝑥𝑥+2 1,000,000 = (0.014)(1,000,000) = 14,000
𝐸𝐸 𝐵𝐵303 = 𝑝𝑝𝑥𝑥+2
03
1,000,000 = (0.004)(1,000,000) = 4000

00 (0)
𝐸𝐸3 𝑉𝑉 00 = 𝑝𝑝𝑥𝑥+2 3𝑉𝑉 = (1 − 0.014 − 0.014 − 0.004)(9000) = 8712
01 01 (1)
𝐸𝐸3 𝑉𝑉 = 𝑝𝑝𝑥𝑥+2 3𝑉𝑉 = (0.014)(210,000) = 2940

So,
(0)
𝑃𝑃𝑃𝑃3 = (12,000 + 30,000 − 1500)(1.06) − 8400 − 14,000 − 4000 − 8712 − 2940
= 4878

Graders’ Comments: For those who did not achieve full marks, common errors
included omitting the value of the benefit payable in case of death after CI, incorrectly
including a value for the 500,000 benefit payable in case of transition from State 1 (at the
beginning of the year) to State 2, or using the wrong expenses for year 3.

(b) For State 1, we have

(1)
2𝑉𝑉 = 280,000
(1)
𝑃𝑃3 = 0
(1)
𝐸𝐸3 = 100
𝐸𝐸 𝐵𝐵312 = 𝑝𝑝𝑥𝑥+2
12
500,000 = (0.25)(500,000) = 125,000
𝐸𝐸3 𝑉𝑉 = 𝑝𝑝𝑥𝑥+2 3𝑉𝑉 (1) = (1 − 0.25)(210,000) = 157,500
11 11

(1)
𝑃𝑃𝑃𝑃3 = (280,000 − 100) (1.06) − 125,000 − 157,500 = 14,194

Graders’ Comments: Common errors included ignoring or using wrong expenses, and
adding a premium when none is paid while in State 1.

Page 57
 (0) (1)
(c) Π3 = 2𝑝𝑝𝑥𝑥00 𝑃𝑃𝑃𝑃3 + 2𝑝𝑝𝑥𝑥01 𝑃𝑃𝑃𝑃3

00 00 01 02 03
2𝑝𝑝𝑥𝑥 = 𝑝𝑝𝑥𝑥00 𝑝𝑝𝑥𝑥+1 = (1 − 𝑝𝑝𝑥𝑥01 − 𝑝𝑝𝑥𝑥02 − 𝑝𝑝𝑥𝑥03 )( 1 − 𝑝𝑝𝑥𝑥+1 − 𝑝𝑝𝑥𝑥+1 − 𝑝𝑝𝑥𝑥+1 )
= (1 − 0.01 − 0.008 − 0.004)(1 − 0.012 − 0.011 − 0.004) = 0.951594

01 01 11
2𝑝𝑝𝑥𝑥 = 𝑝𝑝𝑥𝑥00 𝑝𝑝𝑥𝑥+1 + 𝑝𝑝𝑥𝑥01 𝑝𝑝𝑥𝑥+1 = (0.978)(0.012) + (0.01)((0.75) = 0.019236

Π3 = (0.951594)(4878) + (0.019236)(14,194) = 4914.91

770 3536 4914.91

(d) 𝑁𝑁𝑁𝑁𝑁𝑁(3) = ∑3𝑘𝑘=0 Π𝑘𝑘 𝑣𝑣10%
𝑘𝑘
= −500 − 1.1
+ 1.12
+ 1.13
= 5414.96

Page 58
Question 30
(a) 𝐸𝐸𝐸𝐸𝐸𝐸(𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃) = 𝐸𝐸𝐸𝐸𝐸𝐸(𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝) + 𝐸𝐸𝐸𝐸𝐸𝐸(𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝) +
𝐸𝐸𝐸𝐸𝐸𝐸(𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒)
𝑃𝑃 ∙ (0.95 𝑎𝑎̈ 65:10 − 0.25 )
= (10 𝑃𝑃) 10𝐸𝐸65 𝐴𝐴 1 + (36,000) 20𝐸𝐸65 𝑎𝑎̈ 85 + 900 + 100 𝑎𝑎̈ 65
75:10

36,000(0.24381)(6.7993) + 900 + 100(13.5498)

𝑃𝑃 = = 10,259.385
(0.95)(7.8435) − 0.25 − (10)(0.55305)(0.65142 − 0.44085)

(b) 𝑃𝑃𝑃𝑃0 = −𝐸𝐸0 − 0𝑉𝑉 = −7000 − 500 = −7500

𝑃𝑃𝑃𝑃𝑡𝑡 = 𝑡𝑡−1𝑉𝑉 + 𝐶𝐶𝐶𝐶𝑡𝑡 − 𝐸𝐸𝑡𝑡 + 𝐼𝐼𝑡𝑡 − 𝐸𝐸𝐸𝐸𝐸𝐸𝑡𝑡 − 𝐸𝐸 𝑡𝑡𝑉𝑉 ; 𝑡𝑡 ≥ 1

where 𝐶𝐶𝐶𝐶𝑡𝑡 is the net cash flow received by the insurer at time t.

𝑃𝑃 = 10,259.385 𝑡𝑡 = 0,1, . . . ,9
𝐶𝐶𝐶𝐶𝑡𝑡 = � 0 𝑡𝑡 = 10,11, . . . ,19
−36,000 𝑡𝑡 = 20,21, . . .

𝑃𝑃𝑃𝑃1 = ( 0𝑉𝑉 + 𝑃𝑃 − (0.05𝑃𝑃 + 70))(1 + 𝑖𝑖) − 0 − (1 − 0.03)(1 − 0.9 𝑞𝑞65 ) 1𝑉𝑉

𝑃𝑃𝑃𝑃1 = (500 + (0.95)(10,259.385) − 70)(1.07) − (0.9648362)(10,150) = 1,095.68

𝑃𝑃𝑃𝑃12 = ( 11𝑉𝑉 − 70(1.02)11 )(1 + 𝑖𝑖) − 0.9𝑞𝑞76 (10)(10,259.385) − (1 − 0)(1 −

0.9 𝑞𝑞76 ) 12𝑉𝑉
𝑃𝑃𝑃𝑃12 = (143,035 − 70(1.02)11 )(1.07) − 0.0186012(102,593.85)
− (0.9813988)(151,210)
= 2,648.64

𝑃𝑃𝑃𝑃30 = ( 29𝑉𝑉 − 36,000 − 70(1.02)29 )(1 + 𝑖𝑖) − 0 − (1 − 0)(1 − 0.9 𝑞𝑞94 ) 30𝑉𝑉
𝑃𝑃𝑃𝑃30 = (155,745 − 36,000 − 70(1.02)29 )(1.07) − (0.8595532)(146,275) =
2,262.995

Page 59
(c)
Let 𝑁𝑁𝑁𝑁𝑁𝑁1 be the EPV at the start of year 2 (time 1) of future emerging profits per policy in
force.

1 00 2 00 3
𝑁𝑁𝑁𝑁𝑁𝑁1 = 𝑃𝑃𝑃𝑃2 𝑣𝑣0.1 + 1𝑝𝑝𝑥𝑥+1 𝑃𝑃𝑃𝑃3 𝑣𝑣0.1 + 2𝑝𝑝𝑥𝑥+1 𝑃𝑃𝑃𝑃4 𝑣𝑣0.1 +. ..
𝑛𝑛
𝑡𝑡
𝑁𝑁𝑁𝑁𝑁𝑁 = � 𝜋𝜋𝑡𝑡 𝑣𝑣0.1 = 8860
𝑡𝑡=0
00
where 𝜋𝜋0 = 𝑃𝑃𝑃𝑃0 and 𝜋𝜋𝑡𝑡 = 𝑡𝑡−1𝑝𝑝𝑥𝑥 𝑃𝑃𝑃𝑃𝑡𝑡 , 𝑡𝑡 ≥ 1.

2 3
𝑁𝑁𝑁𝑁𝑁𝑁 = 𝑃𝑃𝑃𝑃0 + 𝑃𝑃𝑃𝑃1 𝑣𝑣0.1 + ( 1𝑝𝑝𝑥𝑥00 𝑃𝑃𝑃𝑃2 𝑣𝑣0.1 + 2𝑝𝑝𝑥𝑥00 𝑃𝑃𝑃𝑃3 𝑣𝑣0.1 + ...)
= 𝑃𝑃𝑃𝑃0 + 𝑃𝑃𝑃𝑃1 𝑣𝑣0.1 + (EPV at time 0 of profits emerging in years 2, 3, …)
00 2
= 𝑃𝑃𝑃𝑃0 + 𝑃𝑃𝑃𝑃1 𝑣𝑣0.1 + 1𝑝𝑝𝑥𝑥00 𝑣𝑣0.1 (𝑃𝑃𝑃𝑃2 𝑣𝑣0.1 + 1𝑝𝑝𝑥𝑥+1 𝑃𝑃𝑃𝑃3 𝑣𝑣0.1 + ...)
00
= 𝑃𝑃𝑃𝑃0 + 𝑃𝑃𝑃𝑃1 𝑣𝑣0.1 + 1𝑝𝑝𝑥𝑥 𝑣𝑣0.1 𝑁𝑁𝑁𝑁𝑁𝑁1

1095.68 1
8860 = −7500 + + (1 − 0.03)(1 − 0.9𝑞𝑞65 ) 𝑁𝑁𝑁𝑁𝑁𝑁1
1.1 1.1

8860 + 7500 − 1095.68/1.1 15,363.9273

𝑁𝑁𝑁𝑁𝑁𝑁1 = = = 17,516.25
(0.9648362)/1.1 0.877124

Graders’ Comments: This was a challenging question part. Only a minority of

candidates achieved partial or full credit.

Page 60
Question 31
(a) EPV(payments + expenses)=100,000
2000 + 500 + (𝐵𝐵 + 25)𝑎𝑎 = 100,000 where 𝑎𝑎 = 𝑎𝑎10 + 10𝐸𝐸65 𝑎𝑎75
65:10 65:10

1−𝑣𝑣 10
𝑎𝑎 = + (0.55305)(9.3178) = 7.72173 + 5.15321 = 12.87494
65:10 𝑖𝑖

100,000 − 2500 − (25)(12.87494)

⇒ 𝐵𝐵 = = 7,547.848
12.87494

(b) Note that at age 73, the annuity is within the 10-year guarantee period. So only
states 0 and 1 are possible. Possible transitions are 0→0, 0→1, 1→1.

01
(i) 𝑝𝑝73 = 𝑞𝑞73 = 0.014664

02
(ii) 𝑝𝑝73 =0
Since there is no possible transition from State 0 at 73 to State 2 at age 74since
the annuity is within the guarantee period.

12
(iii) 𝑝𝑝73 = 0 Since payments would continue until age 76 if annuitant died before
73.

(c) At time 8, there are 2 years left in the guarantee period.

(0)
8𝑉𝑉 = 𝐸𝐸𝐸𝐸𝐸𝐸(future payments plus maintenance expenses)
(0)
8𝑉𝑉 = (𝐵𝐵 + 25)𝑎𝑎
65+8:10−8
= ( 7,547.848 + 25)𝑎𝑎 where
73:2
1 − 𝑣𝑣 2
𝑎𝑎 = 𝑎𝑎2 + 2𝐸𝐸73 𝑎𝑎75 = + 𝑣𝑣 2 2𝑝𝑝73 (9.3178)
73:2 𝑖𝑖

= 1.85941 + (0.879036)(9.3178) = 10.050091

⇒ 8𝑉𝑉 (0) = 76,107.81

Graders’ Comments: Many students did not answer this part, but those that attempted it did
well.

Page 61
(d)
In State 1, there are only 2 guaranteed payments to be made.

(1)
8𝑉𝑉 = (𝐵𝐵 + 25)𝑎𝑎2
= (7547.848 + 25)(1.85941) = 14,081.03

Graders’ Comments: As with Part (c), many students did not answer this part. Those that
did answer, did well.

(e)

(0)
(i) 𝑃𝑃𝑃𝑃9 = 8𝑉𝑉 (0) (𝑖𝑖9 − 𝑖𝑖)
= (76,107.81)(0.06 − 0.05) = 761.08

(1)
(ii) 𝑃𝑃𝑃𝑃9 = 8𝑉𝑉 (1) (𝑖𝑖9 − 𝑖𝑖)
= (14,081.03)(0.06 − 0.05) = 140.81

(0) (1)
(i) 𝜋𝜋9 = 8𝑝𝑝65 𝑃𝑃𝑃𝑃9 + (1 − 8𝑝𝑝65 )𝑃𝑃𝑃𝑃9
= (0.9295525)(761.08) + (0.0704475)(140.81)
= 717.38

Graders’ Comments: Candidates found this part to be very challenging.

Page 62
Question 32
(a)
For T50=11.8, we have K50=11 which means that 12 premiums would be returned without
interest.
Let S = 100,000 and P = 2000.

𝐿𝐿𝑔𝑔 = (𝑆𝑆 + 12𝑃𝑃 + 1000)𝑣𝑣 12 + 0.75 𝑃𝑃 − 0.95 𝑃𝑃 𝑎𝑎̈ 12

1 − 1.05−12
𝑎𝑎̈ 12 = = 9.306414
0.05/1.05

𝐿𝐿𝑔𝑔 = (100,000 + 12(2000) + 1000) 1.05−12 + 0.75(2000)

− 0.95(2000)(9.306414)

𝐿𝐿𝑔𝑔 = 69,604.6773 + 1500 − 17,682.187 = 53,422.49

Graders’ Comments: Candidates generally did well in this part. Common mistakes
were using 11.8 or 11 as the number of premiums or not properly including the return of
premium benefit.

(b) E  Lg  =+( S 1000 ) A50 + P 10 E50 (10 A60 + ( IA)60 ) + 0.05Pa50 + 0.75P − Pa50
= −248.53

Graders’ Comments: Many candidates did not correctly allow for the return of premium
benefit.

(c)
𝑔𝑔
10𝑉𝑉 = (𝑆𝑆 + 1000)𝐴𝐴60 + 10𝑃𝑃 𝐴𝐴60 + 𝑃𝑃(𝐼𝐼𝐼𝐼)60 − 0.95 𝑃𝑃 𝑎𝑎̈ 60

𝑔𝑔
10𝑉𝑉 = (100,000 + 1000)(0.029028) + 10(1000)(0.229028) + (2000)(6.63303)
−0.95(2000)(14.9041)
= 29,318.28 + 5805.60 + 13,266.06 – 28,317.79 = 20,072.15

Graders’ Comments: Candidates did better on this part than on Part (b) although the
return of premium benefit was still a stumbling block for many.

Page 63
(d)
(i) Total gain: using the actual experience (*) during year 16,
∗ ∗ ∗ ∗ ∗
1000( 15𝑉𝑉 𝑔𝑔 + 𝑃𝑃 − 𝑒𝑒15 )(1 + 𝑖𝑖15 ) − 1000 𝑞𝑞65 (𝑆𝑆 + 16𝑃𝑃 + 𝐸𝐸15 ) − 1000 𝑝𝑝65 16𝑉𝑉
𝑔𝑔

=1000(34,333.78+0.96(2000))(1.052)–7(100,000+32,000+2000) –
993(37,480.51)

= 38,138,976.56 – 938,000 – 37,218,146.43

= –17,169.87 i.e. a loss of 17,169.87

(ii) Gain by source:

Gain from expenses (E):
Expected expenses – Actual expenses, valued at year end
[1000(0.05𝑃𝑃)(1.05) + 1000 q 65 (1000)] − [1000(0.04𝑃𝑃)(1.05) + 1000q 65 (2000)]

= [105,000 + 1000(0.005915)(1000)] – [84,000 + 1000(0.005915)(2000)]

= 110,915 – 95,830 = 15,085  a gain from expenses of 15,085

Gain from interest (I):

Actual interest earned – expected interest earned, using the actual expenses
∗
1000( 15𝑉𝑉 𝑔𝑔 + 0.96𝑃𝑃)(𝑖𝑖15 − 0.05)= 1000 [34,333.78 + 0.96(2000)] (.002)
= 72,507.56a gain from interest of 72,507.56
So the gain from interest is 87,553.78 – 15,046.22 = 72,507.56

Gain from mortality (M):

Expected mortality cost – actual mortality cost, using the actual expenses and
interest
∗ )
[1000𝑞𝑞65 (𝑆𝑆 + 16𝑃𝑃 + 𝐸𝐸15 + 1000𝑝𝑝65 16𝑉𝑉𝑔𝑔 ]
∗
−[1000 (0.007)(𝑆𝑆 + 16𝑃𝑃 + 𝐸𝐸15 ) + 1000 (0.993) 16𝑉𝑉 𝑔𝑔 ]

= 792,610 + 37,258,812.78 – 938,000 – 37,218,146.43 = –104,723.65

 a loss from mortality of 104,723.65

Graders’ Comments: Gain by interest was done the best while gain by mortality was
done the worst. Some candidates answered part (d).ii before part (d).i and used the sum
of the gains from part (d).ii to answer part (d).i, if done correctly, this received full credit.
Some candidates forgot to multiply their gains by 1000 policies - a small deduction was
made for this.

Page 64
(e)

The orders MEI and EMI had the same interest gain as EIM.

This can be justified a number of ways.

• The mortality cost is incurred at year end and has no impact of the interest
earned during the year.
• Since actual premium expenses are different than expected, the gain from
expenses (E) must be calculated before that from interest (I) to get the same
value. So, E must precede I.
• Expenses affect the interest earned, so E must be calculated before I.

Graders’ Comments: Some candidates mistakenly commented on the total gain (i.e.
stated that regardless of the order, the total gain will not change) - this did not receive
any credit.

Page 65
Question 33

(a)
(i)

(𝐼𝐼𝐼𝐼)𝑥𝑥 = 𝑣𝑣 𝑞𝑞𝑥𝑥 + 2𝑣𝑣 2 𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+1 + 3𝑣𝑣 3 2𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+2 + 4𝑣𝑣 4 3𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+3 +. ..
= 𝑣𝑣 𝑞𝑞𝑥𝑥 + 𝑣𝑣 2 𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+1 + 𝑣𝑣 3 2𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+2 + 𝑣𝑣 4 3𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+3 +. ..
+ 𝑣𝑣 2 𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+1 + 2𝑣𝑣 3 2𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+2 + 3𝑣𝑣 4 3𝑝𝑝𝑥𝑥 𝑞𝑞𝑥𝑥+3 +. ..
= 𝐴𝐴𝑥𝑥
+ 𝑣𝑣 𝑝𝑝𝑥𝑥 {𝑣𝑣𝑞𝑞𝑥𝑥+1 + 2𝑣𝑣 2 𝑝𝑝𝑥𝑥+1 𝑞𝑞𝑥𝑥+2 + 3𝑣𝑣 3 2𝑝𝑝𝑥𝑥+1 𝑞𝑞𝑥𝑥+3 +. . . }
= 𝐴𝐴𝑥𝑥 + 𝑣𝑣 𝑝𝑝𝑥𝑥 (𝐼𝐼𝐼𝐼)𝑥𝑥+1

(ii) (𝐼𝐼𝐼𝐼)50 = 𝐴𝐴50 + 𝑣𝑣 𝑝𝑝50 (𝐼𝐼𝐼𝐼)51 = 5.8255

5.8255 − 𝐴𝐴50 5.8255 − 0.18931

(𝐼𝐼𝐼𝐼)51 = = = 5.92516
𝑣𝑣 𝑝𝑝50 0.998791/1.05

Graders’ Comments: Candidates did very well on this part. Where points were deducted, it
was primarily from part (i) not being sufficiently rigorous.

(b)

(200,000 + 1000)𝐴𝐴50 + 5000((𝐼𝐼𝐼𝐼)50 − 𝐴𝐴50 ) = 𝑃𝑃((0.95) 𝑎𝑎̈ 50:20 − 0.1)

(195,000 + 1000)𝐴𝐴50 + 5000(𝐼𝐼𝐼𝐼)50 = 𝑃𝑃((0.95) 𝑎𝑎̈ 50:20 − 0.1)

37,104.76 + 29,127.5 = 𝑃𝑃((0.95)(12.8428) − (0.1))

66,232.26
𝑃𝑃 = = 5473.44
12.10066

Graders’ Comments: Candidates did very well on this part with most getting full credit.

Page 66
(c)
Recursively,

(5473.44)(0.85)(1.05) − (0.001209)(201,000)
1𝑉𝑉 = = 4647.6
0.998791
[4647.6 + (5473.44)(0.95)](1.05) − (0.001331)(206,000)
2𝑉𝑉 = = 10,078.9
0.998669

(d)
(i) Total gain: using the actual experience (*) during year 11,
∗ ∗ ∗ ∗ ∗
1000( 10𝑉𝑉 + 𝑃𝑃 − 𝑒𝑒10 )(1 + 𝑖𝑖10 ) − 1000 𝑞𝑞60 (250,000 + 𝐸𝐸10 ) − 1000 𝑝𝑝60 11𝑉𝑉

=1000(63,208 + 0.94(5473.44)) (1.06) – 6 (250,000+ 1100) – 994(71,217)

= 1000[72,454.2156 -1506.6 -70,789.698]

= 157,917.62 i.e. a gain per policy in force at 10 of 157.92

(ii) Gain by source:

Gain from interest:

Actual interest earned – expected interest earned,

1000(63,208 + 0.95(5473.44)) (1.06) − [1000(63,208 )

+ 0.95(5473.44)) (1.05)]

= 1000(68,407.768)(1.06−1.05)

= 684,077.68 i.e. a gain from interest of 684,077.68

Gain from mortality:

Expected mortality cost – actual mortality cost, using the actual interest
[1000𝑞𝑞60 (250,000 + 𝐸𝐸10 ) + 1000𝑝𝑝60 11𝑉𝑉 𝑔𝑔 ]
∗ ∗ 𝑔𝑔
−[1000 𝑞𝑞60 (250,000 + 𝐸𝐸10 ) − 1000 𝑝𝑝60 11𝑉𝑉 ]

= 852,898 + 70,975,004.6 – (1,506,000 + 70,789,698)

= - 467,795.37 i.e. a loss from mortality of 467,795.37

Page 67
 Gain from expenses:
Expected expenses – Actual expenses, valued at year end, using actual interest &
mortality
∗
(1000( 10𝑉𝑉 𝑔𝑔 + 𝑃𝑃 − 𝑒𝑒10 )(1.06) + 1000 𝑞𝑞60 (250,000 + 𝐸𝐸10 ))
∗ ∗ ∗
−[1000( 10𝑉𝑉 𝑔𝑔 + 𝑃𝑃 − 𝑒𝑒10 )(1.06) + 1000 𝑞𝑞60 (250,000 + 𝐸𝐸10 )]
= 1000(0.05P-0.06P)(1.06) + 1000(0.006)(251,000 – 251,100)
= –58,618.46 i.e. a loss from expenses of 58,618.46

Graders’ Comments: A common error was not doing the calculations in the order given or
not incorporating some of the actual experience.

Page 68
Question 34
(a) Pr0 = –3000 – (0.15) P = –3000 –(0.15)(80,000) = –15,000

Graders’ Comments: Candidates generally did well here. Common mistakes included
calculating a reserve when the question specifically stated V=0, or recalculating P when
the question explicitly gave a gross premium.

(b) Pr2 = (0.95P – 100)(1.07) – S qx+1

= (0.95(80,000) – 100)(1.07) – 1,000,000 (0.1)

= 81,213 – 100,000 = –18,787

Graders’ Comments: Candidates generally did well here. Some candidates incorrectly
applied survival probabilities that should have been applied in part (c).
(c)

(𝜏𝜏)
Π = (−15,000; 31,213; 𝑃𝑃𝑃𝑃2 1𝑝𝑝𝑥𝑥 )
= (−15,000; 31,213; – 18,787 (1– 0.05)(.9))
= (−15,000; 31,213; – 16,062.885)

Graders’ Comments: The most common mistake here was not treating the decrements as
independent for the profit signature at time 2. Some candidates also forgot to include one
of the two decrements entirely.
(d)

2
NPV = −15,000 + 31,213 𝑣𝑣0.2 – 16,062.885 𝑣𝑣0.2
= – 15,000 + 26,010.83 – 11,154.78 = –143.95

(e)

Let j=IRR
−15,000 + 31,213 𝑣𝑣𝑗𝑗 – 16,062.885 𝑣𝑣𝑗𝑗2 = 0

31,213 ±�31,2132 −4(15,000)(16,062.885) 31,213 ±3237

 𝑣𝑣𝑗𝑗 = =
2(16,062.885) 32,125.77
 𝒋𝒋 = −6.7% 𝑜𝑜𝑜𝑜 𝟏𝟏𝟏𝟏. 𝟖𝟖𝟖𝟖%  Keeping the positive rate, IRR= 14.83%.

Page 69
(f)
The NPV is being calculated at a risk discount rate of 20% while the IRR uses a discount
rate of 14.83. The 20% discount rate results in a negative because the discounted profits
at the end of one year and at the end of two years when discounted at 20% are not
sufficient to cover the pre-contract expenses. The NPV is positive for some rates of
return less than 20%. The IRR>0 for discount rates less than IRR=14.83% (rates between
-6.7% and 14.83%).

Graders’ Comments: Many candidates simply restated the question here and could not
point to how it's possible to have a positive IRR but a negative NPV. Some mentioned the
difference in discount rate vs IRR, but we were really looking for commentary of the
timing of the cash flows.

(g)
(𝜏𝜏)
With a 15% lapse rate, 1𝑝𝑝𝑥𝑥 = (1 − 0.05)(.85) = 0.8075

The profit signature becomes

(𝜏𝜏)
Π = (−15,000; 31,213; 𝑃𝑃𝑃𝑃2 1𝑝𝑝𝑥𝑥 )
= (−15,000; 31,213; – 18,787 (1– 0.05)(.85))
= (−15,000; 31,213; – 15,170.50)

And NPV at 20% is

2
NPV = −15,000 + 31,213 𝑣𝑣0.2 – 15,170.50 𝑣𝑣0.2
= – 15,000 + 26,010.83 – 10,535.07 = 475.76

(h)
Lapse rates are much less predictable than mortality rates. A contract that is only
profitable if people lapse is very vulnerable to the risk that policyholders lapse at much
lower than the expected rates. The growth of the third party market for insurance makes
lapsing much less attractive for policyholders, as a rational policyholder would not lapse
a policy that could have positive value on the third party market.

Graders’ Comments: Very few candidates noted that the decision to lapse is the
policyholder's, and there are many factors outside the control of the insurer to impact it.

Page 70
Question 35
(a)
- To set premiums
- To set reserves
- To measure profitability
- To stress test profitability
- To determine distributable surplus (for participating contracts)

Graders’ Comments: Most candidates that tried this part received full credit for it.

(b)
𝑛𝑛
2𝑉𝑉 = 𝐸𝐸𝐸𝐸𝐸𝐸 𝑜𝑜𝑜𝑜 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 − 𝐸𝐸𝐸𝐸𝐸𝐸 𝑜𝑜𝑜𝑜 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃
= 𝑆𝑆 𝑣𝑣 𝑞𝑞𝑥𝑥+2 − 𝑃𝑃 = 1,000,000(1.05−1 )(0.3) − 177,313.20
= 285,714.29 − 177,313,20 = 108,401.09

Graders’ Comments: Most candidates received full credit for this part. Common errors
involved an incorrect discount rate, including expenses, or including withdrawals.

(c)
(i) 𝑃𝑃𝑃𝑃0 = −4000 − (0.15)𝐺𝐺
= −4000 − (0.15)(210,000) = −35,500

(ii) Lapse rate: w=0.1

𝑃𝑃𝑃𝑃1 = (0.95𝐺𝐺 − 𝐸𝐸0 )(1 + 𝑖𝑖 ∗ ) − 1,000,000 𝑞𝑞𝑥𝑥 − 𝑝𝑝𝑥𝑥 (1 − 𝑤𝑤) 1𝑉𝑉 𝑛𝑛
= (0.95(210,000) − 100)(1.07) − 1,000,000(0.1) − (0.9)(0.9)(95,754)
= 213,358 − 100,000 − 77,560.74 = 35,797.26

Graders’ Comments: Part (i) was done well by most candidates. In part (ii), many
candidates failed to include withdrawals.

(d)
(i) Π0 = 𝑃𝑃𝑃𝑃0 = −35,500
Π1 = 𝑃𝑃𝑃𝑃1 = 35,797.26
(𝜏𝜏)
Π2 = 𝑝𝑝𝑥𝑥 𝑃𝑃𝑃𝑃2 = (0.9)(0.9)(37,766) = 30,590.46
(𝜏𝜏) (𝜏𝜏) (𝜏𝜏)
Π3 = 2𝑝𝑝𝑥𝑥 𝑃𝑃𝑃𝑃3 = 𝑝𝑝𝑥𝑥 𝑝𝑝𝑥𝑥+1 𝑃𝑃𝑃𝑃3 = (0.9)(0.9)(0.8)(0.9)(29,347) = 17,115.17

Page 71
 𝑡𝑡
(ii) 𝑁𝑁𝑁𝑁𝑁𝑁 = ∑𝑡𝑡 Π𝑡𝑡 𝑣𝑣0.12

35,797.26 30,590.46 17,115.17

= −35,500 + + + = 33,030.61
1.12 1.122 1.123

𝑁𝑁𝑁𝑁𝑁𝑁
(iii) 𝑃𝑃𝑃𝑃 = 1 (𝜏𝜏) 2 (𝜏𝜏)
𝐺𝐺�1+𝑣𝑣0.12 𝑝𝑝𝑥𝑥 +𝑣𝑣0.12 2𝑝𝑝𝑥𝑥 �

33,030.61 33,030.61
= =
210,000(1 + 0.7232143 + 0.4649235) 495,508.93
= 0.07188 𝑜𝑜𝑜𝑜 7.19%

Graders’ Comments: Many candidates failed to include withdrawals, and some candidates
did not use the correct discount rates.

(e) The NPV in terms of G is (v=v0.12)

𝑁𝑁𝑁𝑁𝑁𝑁 = −4000 − (0.15)𝐺𝐺
+{(0 + 0.95𝐺𝐺 − 100)(1.07) − (1,000,000)(0.1) − (0.9)(0.9)(95,754)}𝑣𝑣
+{(95,754 + 0.95𝐺𝐺 − 100)(1.07) − (1,000,000)(0.2)
− (0.8)(0.9)(108,401.09)} (0.9)2 𝑣𝑣 2
+{(108,401.09 + 0.95𝐺𝐺 − 100)(1.07) − (1,000,000)(0.3)} (0.9)2 (0.9)(0.8) 𝑣𝑣 3

𝑁𝑁𝑁𝑁𝑁𝑁 = 𝐺𝐺�−0.15 + 0.95(1.07)(𝑣𝑣 + (0.9)2 𝑣𝑣 2 + (0.9)3 (0.8)𝑣𝑣 3 )�

−4000 − 177,667.74 𝑣𝑣 − 142,316.19 𝑣𝑣 2 − 107,377.52 𝑣𝑣 3
𝑁𝑁𝑁𝑁𝑁𝑁 = 1.8359304 𝐺𝐺 − 352,514.70

1.8359304 𝐺𝐺 − 352,514.70
𝑃𝑃𝑃𝑃 = = 0.1
2.18814 𝐺𝐺

352,514.70
 𝐺𝐺 = 1.8359304−0.218814 = 𝟐𝟐𝟐𝟐𝟐𝟐, 𝟗𝟗𝟗𝟗𝟗𝟗. 𝟔𝟔𝟔𝟔

Graders’ Comments: Very few candidates attempted this part of the

problem. Candidates who attempted to write the NPV in terms of the new premium were
rewarded significantly, even if the attempt was incomplete/incorrect.

Page 72
 Question 36
l96
(a) (i) 1q[95] =
1− =
0.45
l[95]

l96
(ii) 1q95 =−
1 =0.2143
l95

(iii) Typical select tables used for life insurance pricing assume lighter mortality during
the select period, as the underwriting process selects lives that are healthier than average
at inception. In this table, the mortality during the select period is heavier than ultimate
mortality.
Graders’ Comments: Candidates generally did well on part A. We awarded partial
credit for candidates calculating the probability of survival, not death. On part iii we
were looking for commentary on the qx s themselves, and not necessarily the structure of
the table (since select and ultimate tables come in many varieties).
(b)

(i)

t t-1V Prem Annuity Exp Int EtV Prt

0 300 −300
1 0 110,000 50,000 100 4,193 57,750 6,343
2 105,000 -- 50,000 100 3,843 56,000 2,743

=E=
1V p=
[95] × 105, 000 57, 750; =
E2V p96 × 80, 000 56, 000

(ii)
Π 0 =Pr0 =−300; Π1 =Pr1 =6,343; Π 2 =p[95] Pr2 =1,509;
=
Π3 2 p=
[95] Pr3 130;=
Π4 3 p=
[95] Pr4 82

550 385 195

=
where p[95] = , 2 p[95] =
, and 3 p[95]
1000 1000 1000
NPV (1) = NPV (0) + Π1 v12% = 5,363; 

NPV (2) = NPV (1) + Π 2 v = 6,566;
2
12% 

NPV (3) = NPV (2) + Π 3 v12% = 6, 659;
3


= (4) NPV (3) + Π 4 v12% = 6, 711
4
NPV 

Page 73
(iii) The profit margin is 6,711/110,000=0.0610

Graders’ Comments: Candidates also generally did well here.

(c) We now have

t t-1V Prem Annuity Exp Int EtV Prt

1 0 P -- -- 0.07P 30,195 1.07P-30,195
2 54,900 -- -- -- 3,843 56,000 2,743

All the Πt values are the same as in (a), except for the first year:
Π1= Pr1= 1.07 P − 30,195

NPV =0.061P =6711 − 6343v12% + (1.07 P − 30,195)v12%

⇒P=
28,973

Graders’ Comments: Many candidates did not attempt this question. For those who did,
some incorrectly attempted to use recursion to calculate a new premium based on then
new reserve, which doesn't work in part because of the constraint to maintain the same
profit margin.

(d) Disadvantages:
Under the deferred annuity, the policyholder would pay 78,973 (annuity plus premium) in
the first year, 50,000 in the second year, conditional on survival, and nothing thereafter.
The total cost for a survivor, is 128,973. Under the annuity-due, the total cost is
110,000. So, for a 1-year survivor, even allowing for interest, the deferred annuity is
more expensive.

If death occurs in the first two years, the annuitant and/or beneficiaries will receive
nothing, losing the single premium.

Advantages:
An advantage of the deferred annuity is that it is less costly for lives who die in the first
year, whose outlay is 78,973 compared with 110,000 for the annuity-due, thereby leaving
more available for bequest under the SPDA, in the event of early death.

Page 74
Question 37
(a) Reasons include:
• To compete for new employees
• To retain employees in productive years
• To facilitate turnover of employees at older ages
• To offer tax efficient remuneration
• As a tool in negotiations with unions (or other employee collective bargaining units)
• To fulfill responsibility to provide for long-serving employees.
• To improve morale of employees

(b) Let S x denote the salary earned in year of age x to x + 1. We have

S x = (50, 000)(1.03) x −38 .
The EPV of the death benefit is:
(2S62 )q62 (1.04) −1 + (2S63 )1| q62 (1.04) −2 + (2S64 ) 2| q62 (1.04) −3
=
Where =
q62 0.08; 1| q62 = 0.0828;
(0.92)(0.09) =2| q62 = 0.08372
(0.92)(0.91)(0.10)

So the EPV of the death benefit is

(2)(50, 000)(1.03) 24 [0.08(1.04) −1 + (1.03)(0.0828)(1.04) −2 + (1.03) 2 (0.08372)(1.04) −3 ]

= 47, 716

(c) The EPV of the retirement benefit is

(0.03)(27)( FAS )(3 p62 )(1.04) −3 a65

Where FAS is the final average salary =

 (1.03) 24 + (1.03) 25 + (1.03) 26 
FAS =
50, 000   104, 719
 3 

So EPV is
(0.03)(27)(104, 719)(0.75348)(0.88900)(4.7491) = 269,833

Graders’ Comments: Parts (b) and (c) were answered well. The most common minor error was
counting years of service incorrectly, which resulted in a small penalty for part (b), and none for
(c) if the answer was consistent with (b).

Page 75
 Question 38
(a) Let S x denote the salary in year of age x to x + 1 . The projected salaries in the final three
years are:

= =
S62 (80, 000)(1.04)17 155,832
= =
S63 (80, 000)(1.04)18 162, 065
= =
S64 (80, 000)(1.04)19 168,548
155,832 + 162, 065 + 168,548
FAS = 162,148
3

So the projected benefit per month is

25[(0.02)(100,000)+(0.03)(62,148)]/12=8050.92

(12)(8051)
(b) The Replacement Ratio (RR) is = 57.3%
168,548

Graders’ Comments: This part was done well. The most common error was using the final
average salary instead of the final year’s salary in the denominator.

(c) The EPV is

EPV = (8051)(12)a65
(12)

1 − A65(12) 1 − 0.470
a65
= (12)
= = 10.885
d (12) 0.04869

⇒ EPV =
1, 051, 620

Graders’ Comments: Many candidates gained full credit on this part. The most common error
was forgetting to multiply the monthly benefit by 12

(d) We require (0.8-0.573) = 0.227 replacement ratio for the DC plan.

The cost of adding 22.7% RR is

(0.227)(168,548)a65
(12)
= 416, 464

The accumulated contributions to age 65, where c is the contribution rate, are

Page 76
cS 45 ( (1.07) 20 + (1.04)(1.07)19 + ... + (1.04)19 (1.07) )
 (1.07) 20 − (1.04) 20 
=
cS 45   (4, 789,500)c
 1 − (1.04) / (1.07) 

Equating the contributions and the benefit value gives

c = 8.70%

Graders’ Comments: This part was less well done, although a number of strong candidates
scored full marks. The most common errors, which earned partial credit, were miscalculating
the geometric series, or omitting either the 4% or 7% terms in the series.

Page 77
 Question 39
(a) RR denotes the replacement ratio, FAS denotes the final average salary, S x denotes salary in
the year of age x to x + 1 . Then

(9.5)(900) + (15.5)( FAS )(0.03)

RR =
S64
= =
S64 (30, 000)(1.02) 24 48, 253

 1.0222 + 1.0223 + 1.0224 

FAS =
(30, 000)   47,313
 3 

30,551
⇒ RR
= = 63.3%
48, 253

(b)
900n + (25 − n)(47,313)(0.03)
RR ≥ 0.65
48, 253

35, 485 − 519.4n

⇒ ≥ 0.65 ⇒ n ≤ 7.9 years
48, 253

(c) The total accrued benefit based on the first 15 years of employment is
 S + S53 + S54 
(7)(900) + (8)(0.03)  52 = 15, 615
 3 
The required benefit is (0.65)(S64 ) = 31,364
So the annuity payments are 15,749 per year, requiring premium of P where
Pa55:10 = (10 E55 )(15, 749)(a65 )
⇒ P(8.0192) = (0.59342)(15, 749)(13.5498)

⇒P=
15, 791

Page 78
Question 40
(a) The individual has a Defined Benefit plan. A Defined Contribution pension plan specifies
how much an employer will contribute, usually as a percentage of salary, into a plan. The
plan may allow employees to also contribute. The contributions are accumulated in a
notional account which is available to the employee when they leave the company. The
contributions may be set to meet a target benefit level, but the actual retirement income
may be well below or above the target, depending on investment experience.

(b)
To complete Parts (b) and (c), we note that the individual will retire at age 60 or 61.
Therefore we will need to know how much benefit has been accrued for both 60 and 61.
We will also need to know the monthly annuity values at age 60 and 61. Using the 2-term
Woolhouse approximation we have
11 11
a60
(12)
= a60 − = 14.4458 and a61
(12)
= a61 − = 14.1908
24 24

(i)
Under the Projected Unit Credit cost method, the actuarial accrued liability is the actuarial
present value of the projected benefit. The projected benefit is equal to the final average
salary at the decrement date multiplied by service as of the valuation date and by the
accrual rate.

We have the following information.

Projected Final Average Salary at 60
50, 000[(1.03)3 + (1.03)4 + (1.03)5 ] / 3 =
56, 292
Projected Final Average Salary at 61 50, 000[(1.03)4 + (1.03)5 + (1.03)6 ] / 3
= (56, = 292)(1.03) 57,981
Service at valuation date 25
Accrual Rate 0.016
Projected Benefit for retirement at 60 (56,292)(0.016)(25) = 22,517
(PB60)
Projected Benefit for retirement at 61 (57,981)(0.016)(25) = 23,192
(PB61)

Page 79
The actuarial accrued liability is the actuarial present value (as of the valuation date) of the
projected benefit and is given by
exact
(12) r60
During
(12) r60 + i60 + l61(τ ) 6
AAL55 = PB60 ⋅ a60 ⋅ (τ ) ⋅ v5 + PB61 ⋅ a61 ⋅ v
l55 l55(τ )
 27,925.6   6187.6 + 61.9 + 58, 699.9 
(22,517)(14.4458)   (1.05)−5 + (23,192)(14.1908)   (1.05)−6
 104, 687.7   104, 687.7 
= 220,351

(ii)
We now need the accrued benefit at age 56
Projected Benefit for retirement at 60 (PB60) (56,292)(0.016)(26) = 23,417
Projected Benefit for retirement at 61 (PB61) (57,981)(0.016)(26) = 24,120

The actuarial accrued liability at December 31, 2016 is given by

r60exact 4 During
(12) r60 + i60 + l61(τ ) 5
AAL56 = PB60 ⋅ a60
(12)
⋅ ⋅ v + PB ⋅ 
a ⋅ v
l56(τ ) l56(τ )
61 61

 27,925.6   6187.6 + 61.9 + 58, 699.9 

(23, 417)(14.4458)   (1.05)−4 + (24,120)(14.1908)   (1.05)−5
 102,307.9   102,307.9 
= 246, 221

Then

t V + Ct EPV of benefits for mid-year exits + v ⋅1 px(τ ) ⋅t +1 V where:

Ct Normal Cost for year t to t + 1 and tV is the Actuarial Accrued Liability at time t

Note that EPV of benefits for mid-year exits is zero. Then:

t V + Ct EPV of benefits for mid-year exits + v ⋅1 px(τ ) ⋅t +1 V

 102,307.9 
220,351 + Ct =0 + (1.05) −1   (246, 221)
 104, 687.7 

Ct = 8815

Page 80
(c) (i)
Under the Traditional Unit Credit cost method the actuarial accrued liability (AAL) is the
actuarial present value of the accrued benefit on the valuation date.

The formula for the accrued benefit, B, is

 1 + (1.03)−1 + (1.03)−2 
B55= 0.016 ⋅ 25 ⋅ 50, 000 ⋅  = 19, 423
 3 
and thus

r60exact 5 during
(12) r60 + i60 + l61(τ ) 6
AAL55 =B55 ⋅ a60
(12)
⋅ ⋅ v + B ⋅ 
a ⋅ v
l55(τ ) l55(τ )
55 61

 (12) r60exact 5 during

(12) r60 + i60 + l61(τ ) 6 
 
= B55  a60 ⋅ (τ ) ⋅ v + a61 ⋅ v 
 l55 l55(τ ) 

  27,925.6   6,187.6 + 61.9 + 58, 699.9  

= 19, 423 (14.4458)   (1.05)−5 + (14.1908)   (1.05)−6 
  104, 687.7   104, 687.7  

= 186, 248

(ii)
For the NC, we must calculate the expected accrued benefit, B61, one year after the
valuation date.
 1.03 + 1 + 1.03−1 
B56= 0.016 ⋅ 26 ⋅ 50, 000 ⋅  = 20,806
 3 

 (12) r60exact 4 during

(12) r60 + i60 + l61(τ ) 5 
 
AAL56 = B56 ⋅  a60 ⋅ (τ ) ⋅ v + a61 ⋅ v 
 l56 l56(τ ) 

  27,925.6   6,187.6 + 61.9 + 58, 699.9  

20,806 (14.4458)   (1.05)−4 + (14.1908)   (1.05)−5 
  102,307.9   102,307.9  

= 214,358

Then:

t V + Ct EPV of benefits for mid-year exits + v ⋅1 px(τ ) ⋅t +1 V where:

Ct Normal Cost for year t to t + 1 and tV is the Actuarial Accrued Liability at time t

Note that EPV of benefits for mid-year exits is zero. Then:

Page 81
 t V + Ct EPV of benefits for mid-year exits + v ⋅1 px(τ ) ⋅t +1 V

 102,307.9 
186, 248 + Ct =0 + (1.05) −1   (214,358)
 104, 687.7 

Ct = 13, 262

(d)
Under both funding approaches, the contribution rate tends to increase as the member
acquires more service and gets closer to retirement. The TUC contributions start smaller
than the PUC contributions and, rise more steeply, ending at considerably more than the
PUC contribution. This is true because the TUC contributions do not project future salary
increases or future service credit while PUC contributions are based on salary with projected
future salary increases. Therefore, the TUC contributions in any given year must reflect the
full increase in the salary and the additional year of service now reflected in the accrued
benefit at the end of the year of service which were not reflected in the accrued benefit at the
end of the year of service. The PUC contribution will only reflect the additional year of
service; it does not need to reflect the salary increase as this is already allowed for in the
accrued liability at the beginning of the year.

Page 82
Question 41
(a)
0.036 (64−30)(12) (12)
𝑆𝑆64 = 40,000 �1 + � 𝑆𝑆1 𝑖𝑖=1.00312−1 = 138,044.47
12

The monthly pension, B, is:

𝐵𝐵 = (0.02)(35)𝑆𝑆64 /12 = 8052.594

Graders’ Comments: Many candidates did well on this part, especially those who drew a
timeline to show the increasing monthly earnings. For those who did not receive full
credit, the most common error was to calculate the final year salary, S64, incorrectly, for
which a small deduction was applied.

(b)
(12)
𝐸𝐸𝐸𝐸𝐸𝐸 = (0.02)(35)𝑆𝑆64 𝑎𝑎̈ = (0.02)(35)(138,044.47)(13.38208) = 1,293,125.93
65:10

where
(12) (12) (12) (12)
𝑎𝑎̈ = 𝑎𝑎̈ 10 + 10𝐸𝐸65 𝑎𝑎̈ 75 = 𝑎𝑎̈ 10 + 10𝐸𝐸65 (𝑎𝑎̈ 75 − 11/24)
65:10
1−1.05−10
= �12(1−1.05−1/12 )� + (0.55305)(10.3178 − 11/24) = 13.38208

Graders’ Comments: Common errors included calculating the EPV at the wrong date,
using annual payments, incorrectly calculating the EPV of the 10-year guaranteed
annuity and using UDD instead of the Woolhouse formula.

(c)
The accumulated value of contributions, AV , is
40,000
𝐴𝐴𝐴𝐴 = (0.06) � 12
� x [1.008419 + (1.003)(1.008418 ) + ⋯ + (1.003419 )(1.0080 )]
1.003 420
419
1 − �1.008 �
= 200 �(1.008 )� 1.003 �� = (200)(28.181425)(176.627679)
1 − �1.008 �

= 995,523.94

Graders’ Comments: Most candidates found this challenging. Those who used a
timeline to show the contribution amounts found it helpful in answering this part.

Page 83
(d)
Since the lump sum of 1,293,126 is larger than the value of contributions at 9.6%
(995,524), the IRR that earned by taking the lump sum is more than 9.6% convertible
monthly.

Graders’ Comments: Many candidates tried to answer the question in general terms
instead of comparing the options using the results from parts b) and c).

(e)
The revised monthly benefit, B*, is such that
(12)
𝐸𝐸𝐸𝐸𝐸𝐸 = 1,293,125.93 = (12 𝐵𝐵 ∗ ) 𝑎𝑎̈ 65 = (12 𝐵𝐵 ∗ )(𝑎𝑎̈ 65 − 11/24)

1,293,125.93
𝐵𝐵 ∗ = = 8,231.35
11
12 �13.5498 − 24�

Graders’ Comments: The candidates who attempted this part did well. Some candidates
used the AV of contributions instead of the EPV of benefits, for which no credit was
given.

(f)
(i) Adverse selection in insurance refers to a situation where a policyholder makes a
decision based on asymmetric information about the risk, i.e. information known to the
policyholder but unknown to the insurer. Policyholders who represent higher risk tend
to buy more insurance.
Here, adverse selection would occur whenever an employee would choose the most
advantageous option at the time of retirement based on his/her health and other risk
factors unknown to the pension plan sponsor.

(ii) Adverse selection is likely to increase the cost to the plan. The healthiest employees
mostly choosing the life annuity and those in poor health choosing the lump sum.

Graders’ Comments: Well-prepared candidates provided a coherent rationale as to why

the costs to the plan would increase due to adverse selection. The candidates who
discussed the pricing issues rather than the impact on the plan’s costs received no credit
for part (ii).

Page 84
Question 42
𝑟𝑟 (12) 𝑟𝑟 (12) 𝑟𝑟 (12)
(a) 0𝑉𝑉 = 𝐴𝐴𝐴𝐴0 = (𝑇𝑇𝑇𝑇𝑇𝑇)63 ∙ 𝛼𝛼 ∙ �𝑙𝑙63 𝑣𝑣 0.5 𝑎𝑎̈ 63.5 + 𝑙𝑙64 𝑣𝑣 1.5 𝑎𝑎̈ 64.5 + 𝑙𝑙65 𝑣𝑣 2 𝑎𝑎̈ 65 �
63 63 63

4515.2
𝐴𝐴𝐴𝐴0 = (2,500,000)(0.02) � (1.05)−0.5 (13.514)
47,579.3
4061.0 38,488.3
+ (1.05)−1.5 (13.231) + (1.05)−2 (13.086)�
47,579.3 47,579.3

= (50,000)(1.251550 + 1.049600 + 9.601498) = 595,132.40

Graders’ Comments: Performance on this part was mixed. Candidates either did very
well, earning full credit, or not at all well. The most common error was to incorrectly
value the liability for mid-year exits.

00
(b) 𝐴𝐴𝐴𝐴0 + 𝑁𝑁𝑁𝑁 = 𝐸𝐸𝐸𝐸𝐸𝐸(benefits for mid-year exits) + 𝑣𝑣 𝑝𝑝63 𝐴𝐴𝐴𝐴1
where
𝑟𝑟63 0.5 (12)
𝐸𝐸𝐸𝐸𝐸𝐸(benefits for mid-year exits) = ((𝑇𝑇𝑇𝑇𝑇𝑇)63 + 0.5 𝑠𝑠63 ) 𝛼𝛼 𝑣𝑣 𝑎𝑎̈ 63.5
𝑙𝑙63
= �2,500,000 + (0.5)(160,000)�(0.02)(1.25155)
= 64,579.98
𝑟𝑟64 0.5 (12) 𝑟𝑟65 1 (12)
𝐴𝐴𝐴𝐴1 = ((𝑇𝑇𝑇𝑇𝑇𝑇)63 + 𝑠𝑠63 ) ∙ 𝛼𝛼 ∙ � 𝑣𝑣 𝑎𝑎̈ 64.5 + 𝑣𝑣 𝑎𝑎̈ 65 �
𝑙𝑙64 𝑙𝑙64

00
𝑟𝑟64 1.5 (12) 𝑟𝑟65 2 (12)
𝑣𝑣 𝑝𝑝63 𝐴𝐴𝐴𝐴1 = ((𝑇𝑇𝑇𝑇𝑇𝑇)63 + 𝑠𝑠63 ) ∙ 𝛼𝛼 ∙ �
𝑣𝑣 𝑎𝑎̈ 64.5 + 𝑣𝑣 𝑎𝑎̈ 65 �
𝑙𝑙63 𝑙𝑙63
= (2,500,000 + 160,000)(0.02)(1.0496 + 9.601498) = 566,638.41

𝑁𝑁𝑁𝑁 = 64,579.98 + 566,638.41 − 595,132.40 = 36,085.99

Alternatively,
𝑟𝑟63 0.5 (12) 𝑟𝑟64 1.5 (12) 𝑟𝑟65 2 (12)
𝑁𝑁𝑁𝑁 = 𝑠𝑠63 ∙ 𝛼𝛼 ∙ �(0.5) 𝑣𝑣 𝑎𝑎̈ 63.5 + 𝑣𝑣 𝑎𝑎̈ 64.5 + 𝑣𝑣 𝑎𝑎̈ 65 �
𝑙𝑙63 𝑙𝑙63 𝑙𝑙63

= (160,000)(0.02)[(0.5)(1.25155) + 1.0496 + 9.601498] = 36,085.99

Graders’ Comments: Partial credit was awarded to candidates who showed some
understanding of the question by writing down some relevant formulas before evaluating
them numerically.

Page 85
Question 43
(a)
(i) Final Average Salary (FAS) between ages 64 and 65 is projected to be

𝑆𝑆39 (1.03)54−39 (1.025)64−54 = 110,000(1.557967)(1.2800845)

= 219,376.23

(12) (12)
𝐴𝐴𝐴𝐴 = 219,376.23(0.02)(20) 25𝐸𝐸40 𝑎𝑎̈ 65 = (87,750.49) 25𝐸𝐸40 𝑎𝑎̈ 65

where
25
25𝐸𝐸40 = 𝑣𝑣 25𝑝𝑝40 = 20𝐸𝐸40 5𝐸𝐸60 = (0.36663)(0.76687) = 0.281158

(12) 11 11
𝑎𝑎̈ 65 = 𝑎𝑎̈ 65 − = 13.5498 − = 13.09147
24 24

⇒ 𝐴𝐴𝐴𝐴 = (87,750.49)(0.281158)(13.09147) = 322,989.50

(12)
𝐴𝐴𝐴𝐴0 = 322,989.50 and 𝐴𝐴𝐴𝐴1 = 219,376.23(0.02)(21) 24𝐸𝐸41 𝑎𝑎̈ 65
𝐴𝐴𝐴𝐴0
𝑁𝑁𝑁𝑁 = 𝑣𝑣 𝑝𝑝40 𝐴𝐴𝐴𝐴1 − 𝐴𝐴𝐴𝐴0 = = 16,149.48;
20
𝑁𝑁𝑁𝑁 16,149.48
𝑁𝑁𝑁𝑁 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 = 𝑆𝑆40
= 110,000(1.03) = 0.142537

Graders’ Comments: Candidates did very well on both parts, up to the normal cost. Many
candidates made errors concerting the normal cost to a normal cost rate.
(b)
(i) Per unit of FAS, the EPV of the retirement benefits are

(12) (12)
- With partner at retirement, 𝐸𝐸𝐸𝐸𝐸𝐸 𝑝𝑝 = 𝑎𝑎̈ 65 + 0.5 𝑎𝑎̈ 65|65
= 13.09147 + 0.93335 = 14.02482

where
(12) (12) (12)
𝑎𝑎̈ 65|65 = 𝑎𝑎̈ 65 − 𝑎𝑎̈ 65:65
11 11
= �𝑎𝑎̈ 65 − � − �𝑎𝑎̈ 65:65 − �
24 24
11 11
= �13.5498 − � − �11.6831 − � = 1.8667
24 24

Page 86
 (12)
- Single, 𝐸𝐸𝐸𝐸𝐸𝐸 𝑠𝑠 = 1.05 𝑎𝑎̈ 65 = 13.74604

𝐴𝐴𝐴𝐴 = 219,376.23(0.02)(20) 25𝐸𝐸40 �(0.9)𝐸𝐸𝐸𝐸𝐸𝐸 𝑝𝑝 + (0.1)𝐸𝐸𝐸𝐸𝐸𝐸 𝑠𝑠 �

= 219,376.23(0.02)(20)(0.281158)�(0.9)(14.02482) + (0.1)(13.74604)�
= 219,376.23(0.02)(20)(0.281158)(13.99694)
= 345,329.04

(ii) The revised NC rate is

=
NC AL = =
/ 20 345,329.04 / 20 17, 266.45

NC 17, 266.45
=
NC rate = = 0.1524
S 40 110, 000(1.03)

Graders’ Comments: This part was challenging. It required candidates to use the
information in a new context. Many candidates did not use the 90% married at
retirement information. A number discounted the benefit using 25 E40:40 instead of 25 E40
, as if the member and partner had to survive to age 65 to receive any retirement benefit.

(c)
The EPV of the reversionary partner benefits would decrease.

Sample explanation:
Since 𝑎𝑎̈ 𝑥𝑥|𝑦𝑦 = 𝑎𝑎̈ 𝑦𝑦 − 𝑎𝑎̈ 𝑥𝑥:𝑦𝑦 where 𝑎𝑎̈ 𝑦𝑦 does not change, the change in 𝑎𝑎̈ 𝑥𝑥|𝑦𝑦 would be the
opposite of the change in 𝑎𝑎̈ 𝑥𝑥:𝑦𝑦 .
Fixing the time of death of (x) and considering the two cases:
- (x) dies before (y): positive correlation means (y) will now tend to die younger than
under the independence case but since payments stop at the death of (x) => no
change
- (y) dies before (x): positive correlation means (y) will now tend to die older (closer
to (x)). Here payments stop at the death of (y) => EPV increases on average.
Conclusion: 𝑎𝑎̈ 𝑥𝑥|𝑦𝑦 decreases.

Graders’ Comments: Many candidates used the broken heart syndrome as an example of
positively correlated lifetimes and argued that the value of the reversionary annuity
would decrease. This is answer (if complete) was awarded full credit. A few presented a
more general argument to justify the decrease in EPV when lifetimes are positively
correlated.

Page 87
Question 44
(a)
(12)
𝐴𝐴𝐴𝐴 = 20 × 8 × 45,000 × 0.02 × 30𝐸𝐸35 × 𝑎𝑎̈ 65
(12)
+5 × 25 × 62,000 × 0.02 × 5𝐸𝐸60 × 𝑎𝑎̈ 65
(12)
+32,000 × 𝑎𝑎̈ 70
(12) 11 (12)
𝑎𝑎̈ 65 = 𝑎𝑎̈ 65 − = 13.0915; 𝑎𝑎̈ 70 =11.5500; 30𝐸𝐸35 = 0.21981
24
⇒ 𝐴𝐴𝐴𝐴 = 20 x 20,719 + 5 x 311,223 + 369,599
= 414,380 + 1,556,115 + 369,599 = 2,340,094

(b)
NC for age 35 group:
9
414,380 × �1.025 × − 1� = 63,452
8
NC for age 60 group:
26
1,556,111 × �1.025 × − 1� = 102,704
25

Total Salary
45,000 × 20 × 1.025 + 62,000 × 5 × 1.025 = 922,500 + 317,750 = 1,240,250

Normal Contribution Rate

63,452 + 102,704
= 13.397%
1,240,250

(c) Using the numbers calculated in part (b), we have

102,704
NC = = 32.32%
317,750

(d) Under PUC the curve of contribution rates by age is less steep, so the NC at age 60 will be
closer to the NC at age 35, implying that the change will be smaller. The reason is that the
PUC pre-pays for future pay increases on accrued benefits, while the TUC does not. That
means that TUC NC rates at older ages must pay for the new accrued benefit, and in addition
must pay to upgrade all past accruals for the most recent pay rise. This creates a very steep
curve of contribution rates at older ages.

Comment: Candidates who used a sketch of the PUC and TUC conrribution rates to support
their justification were generally given full credit (assuming the sketch was correct).

Page 88
Question 45
(a)
𝑟𝑟 (12) 𝑟𝑟 (12) 𝑟𝑟 (12)
i) 𝐴𝐴𝐴𝐴0 = 𝛼𝛼 (𝑇𝑇𝑇𝑇𝑇𝑇)63 𝑥𝑥 � 𝑙𝑙63 𝑣𝑣 0.5 𝑎𝑎̈ 63.5 + 𝑙𝑙64 𝑣𝑣 1.5 𝑎𝑎̈ 64.5 + 𝑙𝑙65 𝑣𝑣 2 𝑎𝑎̈ 65 �
63 63 63
4515
= (0.02) (2,400,000) ∙ � 1.05−0.5 (13.5139) +
47,579

4061 38,488
+ 1.05−1.5 (13.2312) + 1.05−2 (13.087)�
47,579 47,579

= 48,000 {1.2515 + 1.0496 + 9.6022} = 571,358.40

𝑟𝑟63 (12) 𝑟𝑟 (12) 𝑟𝑟 (12)

ii) 𝑁𝑁𝑁𝑁 = 𝛼𝛼 𝑆𝑆63 ∙ �0.5 𝑣𝑣 0.5 𝑎𝑎̈ 63.5 + 𝑙𝑙64 𝑣𝑣 1.5 𝑎𝑎̈ 64.5 + 𝑙𝑙65 𝑣𝑣 2 𝑎𝑎̈ 65 �
𝑙𝑙63 63 63

4515 4061
⎧0.5 1.05−0.5 (13.5139) + 1.05−1.5 (13.2312)⎫
⎪ 47,579 47,579 ⎪
= (0.02) (170,000)
⎨ 38,488 ⎬
⎪ + 1.05−2 (13.087) ⎪
⎩ 47,579 ⎭

1.2515
= 3,400 � + 1.0496 + 9.6022� = 38,344
2

Alternatively,

𝑁𝑁𝑁𝑁 = 𝑣𝑣𝑝𝑝63 𝐴𝐴𝐴𝐴1 − 𝐴𝐴𝐴𝐴0 + 𝐸𝐸𝐸𝐸𝐸𝐸(mid-year retirements)

1 42,805
= �1.05� �47,579� (638,995.53) − 571,358.40 + 62,199.55 = 38,346
where
𝑟𝑟64 0.5 (12) 𝑟𝑟65 1 (12)
𝐴𝐴𝐴𝐴1 = 𝛼𝛼 [(𝑇𝑇𝑇𝑇𝑇𝑇)63 + 𝑆𝑆63 ] ∙ � 𝑣𝑣 𝑎𝑎̈ 64.5 + 𝑣𝑣 𝑎𝑎̈ 65 �
𝑙𝑙64 𝑙𝑙64
4061 13.2312 38,488 13.087
= (0.02)(2,570,000) � � 0.5
�+ � ��
42,805 1.05 42,805 1.05

= 51,400 {1.22502 + 11.2068} = 638,995.53

𝑟𝑟63 0.5 (12)

𝐸𝐸𝐸𝐸𝐸𝐸(mid-year retirements) = 𝛼𝛼 [(𝑇𝑇𝑇𝑇𝑇𝑇)63 + 0.5 𝑆𝑆63 ] ∙ � 𝑣𝑣 𝑎𝑎̈ 63.5 �
𝑙𝑙63
= (0.02) [2,485,000] {1.2515} = 62,199.55

Page 89
(b)

𝑑𝑑 𝑑𝑑64
i) 𝐴𝐴𝐴𝐴0 = 4 𝛼𝛼 (𝑇𝑇𝑇𝑇𝑇𝑇)63 𝑥𝑥 � 𝑙𝑙 63 𝑣𝑣 0.5 + 𝑣𝑣 1.5 �
63 𝑙𝑙63

213.9 215.1
= 4(0.02) (2,400,000) 𝑥𝑥 �47,579 1.05−0.5 + 47,579 1.05−1.5 �

= 192,000 {0.004387 + 0.004202} = 1649.09

𝑑𝑑63 𝑑𝑑64
ii) 𝑁𝑁𝑁𝑁 = 4𝛼𝛼 𝑆𝑆63 𝑥𝑥 �0.5 𝑣𝑣 0.5 + 𝑣𝑣 1.5 �
𝑙𝑙63 𝑙𝑙63

213.9 215.1
= 4(0.02) (170,000) 𝑥𝑥 �0.5 1.05−0.5 + 1.05−1.5 �
47,579 47,579

0.004387
= 13,600 { + 0.004202} = 86.98
2

Alternatively,
𝑁𝑁𝑁𝑁 = 𝑣𝑣𝑝𝑝63 𝐴𝐴𝐴𝐴1 − 𝐴𝐴𝐴𝐴0 + 𝐸𝐸𝐸𝐸𝐸𝐸(mid-year deaths)

1 42,805
= �1.05� �47,579� (1008.26) − 1649.09 + 872.20 = 86.95
where
𝑑𝑑64 0.5
𝐴𝐴𝐴𝐴1 = 4𝛼𝛼 [(𝑇𝑇𝑇𝑇𝑇𝑇)63 + 𝑠𝑠63 ] ∙ �
𝑣𝑣 �
𝑙𝑙64
215.1 0.5
= 4(0.02)(2,570,000) � 𝑣𝑣 �
42,805
= 1008.26

𝑟𝑟63 0.5
𝐸𝐸𝐸𝐸𝐸𝐸(mid-year deaths) = 4𝛼𝛼 [(𝑇𝑇𝑇𝑇𝑇𝑇)63 + 0.5 𝑆𝑆63 ] ∙ � 𝑣𝑣 �
𝑙𝑙63
= 4(0.02)(2,485,000){ 0.004387} = 872.14

Page 90
Question 46
(a)
𝑙𝑙 2
(i) 7.5𝐸𝐸57.5:57.5 = 𝑣𝑣 7.5 ( 7.5𝑝𝑝57.5 )2 = 𝑣𝑣 7.5 �𝑙𝑙 65 �
57.5
2
94,579.7
= 1.05−7.5 � � = 0.6551081
(0.5)(97,435.2 + 97,195.6)

𝑤𝑤 𝑙𝑙58.5 𝑤𝑤 1 (0.5)(97,195.6+96,929.6)
(ii) 𝑎𝑎57.5 = 𝑣𝑣 𝑎𝑎58.5 = 1.05 (10.5804) = 10.050395
𝑙𝑙57.5 (0.5)(97,435.2+97,195.6)

Alternatively,

𝑤𝑤 (12) (12)
𝑎𝑎57.5 = 7.5𝐸𝐸57.5:57.5 𝑎𝑎̈ 65:65 + 7.5𝐸𝐸57.5 (1 − 7.5𝑝𝑝57.5 )𝑎𝑎̈ 65
(12) (12) (12)
= 7.5𝐸𝐸57.5 𝑎𝑎̈ 65 + 7.5𝐸𝐸57.5:57.5 �𝑎𝑎̈ 65 − 𝑎𝑎̈ 65:65 �
= (0.674057)(13.0807) + (0.6551081)(13.0870 − 11.2158)
= 8.8213838 + 1.0258383 = 10.04722

where
94,579.7
7.5𝐸𝐸57.5 = 1.05−7.5 � � = 0.674057
(0.5)(97,435.2 + 97,195.6)

Graders’ Comments: Candidates generally did well on part (i) but struggled with part
(ii). Common errors were misapplied recursive annuity formula and using an annuity-
due instead of an annuity-immediate. The annuity is deferred to age 65 so there is a
payment at t=0.

(b)
𝑤𝑤57 0.5 𝑤𝑤 𝑤𝑤58 1.5 𝑤𝑤 𝑤𝑤59 2.5 𝑤𝑤
𝐴𝐴𝐴𝐴 = (0.018)(35)(100,000) � 𝑣𝑣 𝑎𝑎57.5 + 𝑣𝑣 𝑎𝑎58.5 + 𝑣𝑣 𝑎𝑎59.5 �
𝑙𝑙57 𝑙𝑙57 𝑙𝑙57
1976 1929.9
= (63,000) � 1.05−0.5 (10.04722) + 1.05−1.5 (10.5804)
99,960.2 99,960.2
1884.3
+ 1.05−2.5 (11.1456)�
99,960.2
= (63,000){0.193825 + 0.1898566 + 0.1859744} = 35,888.33

Graders’ Comments: In general, if candidates applied first principles to the contingent

benefit they did well. Trying to shoehorn memorized formulas did not work well. Some
candidates applied mortality decrements from SULT instead of retirement decrements
from Standard Service Table, and many candidates missed that there were 3 possible
withdrawal ages.

Page 91
 (c)
𝑤𝑤57 0.5 𝑤𝑤 𝑤𝑤58 1.5 𝑤𝑤 𝑤𝑤59 2.5 𝑤𝑤
𝑁𝑁𝑁𝑁 = (0.018)(100,000) �(0.5) 𝑣𝑣 𝑎𝑎57.5 + 𝑣𝑣 𝑎𝑎58.5 + 𝑣𝑣 𝑎𝑎59.5 �
𝑙𝑙57 𝑙𝑙57 𝑙𝑙57

= (1,800){(0.5)(0.193825) + 0.1898566 + 0.1859744} = 850.94

Or

NC= vp57 AL1 + EPV (mid − year − exits ) − AL0

= 24,353.85 + 12,35.46 − 35,888.33 = 850.94

 w  w  w 

vp57 AL1 w
(36)(0.018)(100, 000)  58  v1.5 a58.5 +  59  v 2.5 a59.5 
 l57   l57  

= (64,800){0.1898566 + 0.185974}
= 24,353.85

w 
EPV (mid − year − exits ) = w
(35.5)(0.018)(100, 000)  57  v 0.5 a57.5
 l57 

= (63,900)(0.193825)
= 12,385.46

Graders’ Comments: Many candidates tried to use the PUC shortcut formula for a TUC
plan. Those who used the TUC shortcut did not allow for the benefit being contingent on
withdrawal.

(d)

Value of settlement pre-divorce:

𝑤𝑤
(0.018)(35.5)(100,000)𝑎𝑎57.5 = (63,900)(10.04722) = 642,017.36

Value of settlement post-divorce:

(12) (12) 𝑋𝑋 4𝑋𝑋
7.5𝐸𝐸57.5 𝑎𝑎̈ 65 𝑋𝑋 + 7.5𝐸𝐸57.5 𝑎𝑎̈ 65 3
= (0.674057)(13.0870) 3
= 11.761845 𝑋𝑋

For the two values to be equal, X=54,584.75

Page 92
Question 47
(a)
F: account balance
25
𝐹𝐹 = � 𝑐𝑐 𝑆𝑆35+𝑡𝑡 𝑒𝑒 𝛿𝛿(25−𝑡𝑡) 𝑑𝑑𝑑𝑑
0
25
𝐹𝐹 = � (0.12) (50,000) 𝑒𝑒 0.02𝑡𝑡 𝑒𝑒 0.06(25−𝑡𝑡) 𝑑𝑑𝑑𝑑
0

0.06(25)
1 − 𝑒𝑒 −0.04(25)
= 6000 𝑒𝑒 � � = 424,945.17
0.04

Graders’ Comments: Some candidates treated the salary as an annual lump sum. Also,
some confused the projection of the salary between enrollment and payment, with the
accumulation from payment until retirement.

(b)
(i) X: monthly income
(12) (12)
424,945.17 = 12𝑋𝑋 �𝑎𝑎̈ 10 + 10𝐸𝐸60 𝑎𝑎̈ 70 �
11
= 12𝑋𝑋 �7.92949 + (0.57864) �12.0083 − 24��
= 12𝑋𝑋 [7.92949 + (0.57864)(11.54997)]
= 12𝑋𝑋 (14.61276)

(12) 1−𝑣𝑣 10 1−1.05−10

𝑎𝑎̈ 10 = = = 7.92949
𝑑𝑑 (12) 0.04869

 X = 424,945.17/[(12)(14.61276)] = 2423.37

(ii) Final 1-year salary is

25
� (50,000)𝑒𝑒 0.02𝑡𝑡 𝑑𝑑𝑑𝑑 = 81,617.17
24

𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 29,080.42

𝑅𝑅 = = = 0.3563
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑖𝑖𝑖𝑖 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑡𝑡𝑡𝑡 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 81,617.17
R=35.63%

Graders’ Comments: Common mistakes were using UDD instead of Woolhouse to

calculate the monthly annuity or using an annual annuity factor. Additionally, the
replacement ratio compares the first year in retirement to the last year in employment.
Many candidates used the rate of salary instead of the salary during the last year.

Page 93
(c)
(12) 2 (12) (12)
750,000 = 12𝑋𝑋 �𝑎𝑎̈ 60:60 + �𝑎𝑎̈ 60:60 − 𝑎𝑎̈ 60:60 ��
3
(12) 2 (12) (12) (12)
= 12𝑋𝑋 �𝑎𝑎̈ 60:60 + ��2𝑎𝑎̈ 60 − 𝑎𝑎̈ 60:60 � − 𝑎𝑎̈ 60:60 ��
3
4 (12) 1 (12)
= 12𝑋𝑋 � 𝑎𝑎̈ 60 − 𝑎𝑎̈ 60:60 �
3 3
4 11 1 11
= 12𝑋𝑋 � �𝑎𝑎̈ 60 − � − �𝑎𝑎̈ 60:60 − ��
3 24 3 24
4 11 1 11
= 12𝑋𝑋 � �14.9041 − � − �13.2497 − ��
3 24 3 24
= 12𝑋𝑋 (14.9972)

X=4167.44

Graders’ Comments: There were many ways to calculate the annuity values in this part,
but many candidates are not comfortable with annuities whose payment depends on a
two-life status.

(d)
(i) (1) (4167.44) + 1000 = 5167.44
(ii) (2/3)(4167.44) + 1000 = 3778.29
(iii) (2/3)(4167.44) + 500 = 3278.29

(e)
4 (12) 1 (12)
750,000 = 12𝑋𝑋 � 𝑎𝑎̈ 60 − 𝑎𝑎̈ 60:60 �
3 3
4 (12) 1 (12)
− 12 (1000) �3 𝑎𝑎̈ 65 − 3 𝑎𝑎̈ 65:65 � 5𝐸𝐸60:60 (i.e. both alive at 65)

2 (12)
− 12 (1000) 2 �3 𝑎𝑎̈ 65 � 5𝐸𝐸60 ( 5𝑞𝑞60 ) (i.e. only one alive)
4 (12) 1 (12)
where 3 𝑎𝑎̈ 60 − 3 𝑎𝑎̈ 60:60 = 14.9972 from part c).

750,000 = 12𝑋𝑋 (14.9972)

4 1
− 12 (1000) �3 (13.5498 − 11
24
) − 3 �11.6831 − 11
24
�� (0.75057)

2
− 12 (1000) 2 �3 (13.5498 − 11
24
)� (0.76687)(0.02126)

= 12𝑋𝑋 (14.9972) − 123,516.35 − 3414.95

=> 𝑋𝑋 = 4872.74

Page 94
5𝐸𝐸60:60 = ( 5𝐸𝐸60 )2 (1 + 𝑖𝑖)5 = (0.76687)2 (1.05)5 = 0.75057
94,579.7
5𝐸𝐸60:60 = 𝑣𝑣 5 ( 5𝑝𝑝60 )2 = (1.05)−5 (96,634.1)2 = (1.05)−5 (0.97874)2 = 0.75057

94,579.7
5𝑞𝑞60 = 1 − 5𝑝𝑝60 = 1 − 96,634.1 = 1 − 0.97874 = 0.02126

Graders’ Comments: Few attempted this part and very few got it right. The key is to
consider all the possible outcomes when the couple (jointly or singly) reaches 65..

Page 95
Question 48
(a)

(i) The EPV at age 63.5 is

(12)
𝐸𝐸𝐸𝐸𝐸𝐸 = 12.5 × 0.016 × 𝐹𝐹𝐹𝐹𝐹𝐹63.5 × (1 − 18 × 0.004) × ä63.5
where 𝐹𝐹𝐹𝐹𝐹𝐹63.5 = 0.5 × 72,100 + 0.5 × 70,000 = 71,050
𝐸𝐸𝐸𝐸𝐸𝐸 = 13,186.88 × 13.5139 = 178,206.18

(ii) The EPV at 1 Jan 2022 of the mid-year exits is

𝑟𝑟63 4,515.2
𝑣𝑣 0.5 (178,206.18) = 47,579.3 1.05−0.5 (178,206.18)
𝑙𝑙63
= 16,503.92 (or 16,503.90 if using 178,206)
(b)

(i) The projected actuarial liability (AL) at 1/1/2023 is the value at 1/1/2023 of the exit
benefits for ages 64.5 and 65, based on 13 years of past service.
The AL at 1/1/2022 is the value at 1/1/2022 of the exit benefits for ages 63.5, 64.5
and 65, based on 12 years of past service. This gives us
12 12 𝑙𝑙
𝐴𝐴𝐴𝐴 = 12.5 × 16,503.92 + 13 × 𝑣𝑣 × 𝑙𝑙64 × 191,309
63

12 12 42,805.0
= × 16,503.92 + × 1.05−1 × × 191,309
12.5 13 47,579.3

= 15,843.76 + 151,307.50 = 167,151.26

(ii) The normal cost is the value at 1/1/2022 of the mid-year exits plus the value at
1/1/2022 of the projected 1/1/2023 AL minus the 1/1/2022 AL. That is
𝑙𝑙
𝑁𝑁𝑁𝑁 = 16,503.92 + 191,309 × 𝑙𝑙64 × 𝑣𝑣 − 167,151.26 = 13,269.11
63

0.5 1 𝑙𝑙64
𝐴𝐴𝐴𝐴 = × 16,503.92 + × 𝑣𝑣 × × 191,309
12.5 13 𝑙𝑙63

0.5 1 42,805.0
= × 16,503.92 + × 1.05−1 × × 191,309
12.5 13 47,579.3

= 660.16 + 12,608.96 = 13,269.11

Page 96
(c)

(i) The AL using TUC will be smaller than the PUC AL.
• The TUC AL does not project the FAS to retirement, using the current FAS
instead. As salaries are increasing, this means that the TUC AL is always less than
the PUC AL, except at the very start of the member’s employment, when service
n = 0, so both ALs are 0, and at the very last retirement date, when the FAS is the
same under both PUC and TUC.

(ii) The TUC NC will be greater than the PUC NC.

• The TUC normal cost each year starts below the PUC, but around 2/3 through the
max employment period, the lines cross. In B’s case, as they are near to
retirement, the TUC NC will be greater than the PUC NC.

(d)
(𝐴𝐴𝐴𝐴2022 + 𝑁𝑁𝑁𝑁2022 )(1.051) = (167,151.26 + 13,269.11)(1.051) = 189,621.81

13 × 0.016 × 72,100 × (1 − 12 × 0.004) × 13.3735

= 14,276.95 × 13.3735 = 190,932.84

189,621.81 − 190,932.84 = −1311.03

The plan made a loss of 1311 from the early retirement.

Graders’ Comments: Very few candidates attempted this part. When a question is
unfamiliar, returning to first principles is usually the best strategy.

Page 97
 Question 49

(a) =
AVt ( AV ){t − 1} + Pt − Et − COIt ) (1 + itc )
where COIt = ADBt × qx +t −1 × viq , ADBt = 150, 000, and i q = 0.04;
⇒ AV3 = (8166 + 4800(0.95) − 377.88) (1.04) = 12,842
The total DV in year 3 is DB3 = ADB3 + AV3 = 162,842

(b) The profit test table is:

Year, t AVt-1 Pt Et It EDBt ESVt EAVt Prt

0* 0 800 -800.00
1 0 4800 1200 288 230.47 114.528 3277.377 265.63
2 3647 4800 336 649 237.25 1331.2 6523.001 668.43
3 8166 4800 336 1010 244.26 12822.74 573.40

AVt −1 is the account value brought forward.

Pt is the premium paid at the start of the t-th year.
Et represents the expenses incurred at the start of the tth year.
It = ( AVt −1 + Pt − Et ) (0.08) is the interest earned during the tth year, for a policy
in force at the start of the year.
= ( ADBt + AVt ) 0.0015(150, 000 + AVt )
EDBt qx( d+t)−1=
qx( w+t)−1 ( AVt − SCt ) where q35
ESVt =
+ ( w)
=
0.1 1 − q35
(d )
(( w)
; q36 = )
0.2 1 − q36
(d )
( )
( w)
and q37 ( (d
)
= 1 − q37 ) . SCt is the surrender charge in year t.
(τ ) (τ )
=EAVt p= x + t −1 AVt where px + t −1 1 − qx( d+t)−1 − qx( w+t)−1
Prt is the profit vector.

(c)
Year, t Prt Πt
0 -800.00 -800.00
1 265.63 265.63
2 668.43 600.68
3 573.40 411.61

Where=
Π 0 Pr0 , and
= Πt t −1 px(τ ) Prt .
⇒ NPV = −800 + 265.63v + 600.68v + 411.61v at 10%
2 3

= 247.2

Page 98
(d) The profit margin is

NPV 247.2 247.2

= =
((τ )
)
(τ ) 2
EPV Premiums at 10% 4800 1 + px v + 2 px v 11569.0
= 2.14%

Page 99
Question 50
(a) For each year, we assume first that the corridor factor does not apply, and then check to
see whether the Death Benefit: Account Value ratio is sufficient. For the first year we
have
= AV1 ( 50, 000(0.8) − 75 − (0.025)(100, 000 − AV1 )v4.5% ) (1.065)
⇒ AV1 (1 − 0.025v4.5% (1.065)) = 39,972
⇒ AV1 = 41, 017
DB 100, 000
= = 2.44 > CF1
Check corridor factor:
AV1 41, 017
Since the ratio of the DB to the AV is greater than the corridor factor, not adjustment to
the DB is required.
For the second year, we have
=
AV2 ( 41, 017 + 50, 000(0.92) − 75 − (0.03)(100, 000 − AV2 )v4.5% ) (1.0575)
⇒ AV2 (1 − 0.03v4.5% (1.0575)) = 88,906
⇒ AV2 = 91, 689
DB 100, 000
Check corridor factor: = = 1.09 < CF1
AV2 91, 689
The ratio is too small, and the corridor factor applies. We must recalculate the account
value.
A=V2 ( 41, 017 + 50, 000(0.92) − 75 − (0.03)(0.4 × AV2 )v4.5% ) (1.0575)
⇒ AV2 (1 + 0.03(0.4)v4.5% (1.0575)) =
91,9416
⇒ AV2 = 90,838

Graders’ Comments: The most common errors were ignoring the corridor factor, or
assuming the corridor factor applied in both years.

(b) The EPV of the annuity is

Page 100
 Thus, 𝐸𝐸𝐸𝐸𝐸𝐸 = 4𝑄𝑄(15.0759)

Set this equal to AV2 to get 𝑄𝑄 = 1506.34.

Graders’ Comments: Most candidates used Q instead of 4Q throughout.

(c)
(d) It is uncommon to offer surrender values for annuities because of adverse selection. Lives
who are in poor health are more likely to surrender, reducing the pool available for
survivors. The objective of an annuity pool is that any excess funds arising from early
deaths are used to support the continuing payments to the surviving lives.

Graders’ Comments: Most candidates seemed unaware that traditional annuities cannot
generally be surrendered, and few mentioned the adverse selection problem.

Page 101
Question 51
(a) AV2= ( AV1 + 0.9 P2 − 10 − 200v ) (1.06)= 918.3

(b) Now AV2 = (165 + 0.9 P2 − 10 ) (1.06) − 200 = 0.954 P2 − 35.7

=
AV3 ( AV2 + 0.9 P3 − 10 ) (1.06) − 300
=( 0.954 P2 − 35.7 + 0.9 P3 − 10 ) (1.06) − 300
= 1.0112 P2 + 0.954 P3 − 348.4
=
⇒ a 1.0112;
= b 0.954;= c −348.4

(c) There are four possible outcomes for AV3 , and hence for the death benefit and surrender
values. Note that the surrender value cannot be negative.

P1 P2 AV3 DB3 SV3 Prob

1000 1000 1616.8 101,616.8 1516.8 0.36

1000 200 853.6 100,853.6 753.6 0.24
200 1000 807.8 100,807.8 707.8 0.08
200 200 44.6 100,044.6 0 0.32

So
(i) E [ DB3 ] = 101, 616.8 × 0.36 + ... + 100, 044.6 × 0.32 =
100,866
(ii) E [ SV3 ]= 1516.8 × 0.36 + ... + 0 × 0.32 =783.5

(d) Let AV10 = 5114 denote the account value at time 10 if all the premiums are paid,
(1)

AV10(2) , and AV10(3) , denote the account values if the 3rd premium and the 10th premium,
respectively, are omitted. Then
AV10(2) = AV10(1) − 1000(0.9)(1.06)8 =3679.5
AV10(3) =
AV10(1) − 1000(0.95)(1.06) =
4107.0
⇒ E [ AV10 ] = 0.5 × 3679.5 + 0.5 × 4107.0 = 3893.3

Graders’ Comments: In part (c)(ii), many candidates subtracted the surrender charge
from E [ AV3 ] , which does not work because of the floor of 0 on the surrender value.
Many candidates omitted part (d), but those who attempted it scored well.

Page 102
Question 52
(a) AV1= ( 2500(0.95) − 0.005(100, 000)v6% − 30 ) (1.06)= 1985.70
⇒ CV=
1 0.2 AV=
1 397.1

=
(b) Pr1 ( P − E ) (1 + i e ) − EDB1 − ESV1 − EAV1
= =
P 2500; E 120; i e = 0.11;
EDB1 0.004 (100, 000 + AV
= =1 + 200 ) 408.7
ESV1 0.1(0.996) ( CV
= =1 + 100 ) 49.5
= =
EAV1 0.996(0.9) AV1 1780.0
⇒ Pr1 =
403.6

Graders Comments: Part (a) was done well, although a few assumed that this was a
Type A policy, when it is a Type B (fixed additional death benefit). In part (b), the most
common errors were in calculating the dependent withdrawal and survival probabilities,
failing to include the account value in EDB, and using the contract expense assumptions
instead of the profit test expense assumptions.

(c)=
NPV (1) Pr0 + Pr1 v14% = −200=
+ 403.6v 154.0

(d) Let * denote values under the revised assumptions. Then we have
NPV =+ Pr0 Pr1 v + px(τ ) Pr2 v 2 + 2 px(τ ) Pr3 v3 + ... at i =
0.14
Prt* = Prt for t = 0, 2, 3, ...
E= =
DB1* EDB1 408.7
ESV1 = 0.2(0.996)497.1 = 99.0
*

EAV1 = 0.996(0.8) AV1 = 1582.2

*

⇒ Pr1* = ( 2500 − 120 ) (1.11) − 408.7 − 99.0 − 1582.2 = 551.9

0.8 (τ )
Also, for k = 1, 2, ..., k px(τ )* =
k px
0.9
0.8
⇒ NPV
= *
( NPV − Pr0 − Pr1 v ) + Pr0 + Pr1* v
0.9
0.8
=( 2000 − 154 ) − 200 + 551.9v =1925.0
0.9
Graders’ Comments: The change in assumptions impacts Pr1, which most candidates
correctly allowed for, but it also changes the survival probabilities used for all the profit
signature values from the 2nd contract year, which most candidates did not allow for.

Page 103
 Question 53

(a) Universal life is regulated as an insurance product. The corridor factor ensures that the
insurance benefit is significant throughout the term of the contract, differentiating the
policy from a pure investment contract.

(b)
(i) AV1 = (15, 000(0.99) − CoI ) (1.05)
CoI = 1.2q50 (100, 000 − AV1 ) v4% =139.62 − 0.0013962 AV1
15, 445.9
⇒=
AV1 = 15, 468.6
0.998534
⇒ CoI =118.02
(ii) CoI = 1.2 q50 (1.2 AV1 ) v4% = 0.001675 AV1
15,592.5
⇒ AV1 = =
15,565.12
1.001759
⇒ CoI =26.07
= =
(iii) CoI max (118.02, 26.07 ) 118.02

(c) AV1 = 15, 468.6 from (b)(i)

=
ADB 1 DB1 − AV1 =100, 000 − 15, 486.6 =84,513.4
= CV1 15, 486.6 − 0.05(15, 000) = 14, 736.6

(d) The value of the term insurance is

= 1
100, 000 A70:20 =
100, (
000 A70:20 − 20 E70 29, 778 )
⇒ The reserve is 29, 778 − 20, 000 = 9, 778.

Graders’ Comments: Many candidates memorized the CoI formulas for part (b). If the
formulas used were correct, full marks were awarded, as the question did not ask for a
derivation. Some candidates answered the question as if the contract were a Type B. This
earned little credit as it significantly simplifies the work, and because understanding the
difference between Type A and Type B contracts is important.

Page 104
Question 54
(a)
=
AV1 ( 0.4(5000) − 200 − 0.006(100000)v5% ) (1.05)
= 1290.0

Graders’ Comments: A number of candidates incorrectly used information from the

`Assumptions' table in this part. It is important in UL to understand the different roles of
the policy terms and conditions (used to project account values) and the valuation
assumptions (used to determine profitability).

(b)
(τ ) 2 (τ ) 3
NPV (3) =
−1206v12% + 374 p60 v12% + 400 2 p60 v12%
(τ ) (τ )
p60 =−
1 0.004 − 0.1 =0.8960; 2 p60 =0.8960(0.9450) =0.8467
⇒ NPV (3) =
−568.6

(c)
(i) Revised profit test table for year t = 2 :

V
t −1 P E I EDB ESV E tV Prt
690 1,000 60 97.8 507.5 65 1,228.5 −73.2

1 V = 1290 − 600 = 690

Premium, P = 1000
Expenses, E= 0.05P + 10= 60
Interest=
,I 0.06 ( 690 + 1000 −=
60 ) 97.80
EDB =q61
(d )
× (100000 + AV2 ) =507.5
ESV =q61
( w)
× (1500 − 200) =65
(τ )
EtV =p61 × (1500 − 200) =1228.5

(ii) In year 2 the beginning of year reserve, brought forward with interest during year
2 is much lower using the cash value. The end year reserve is also reduced, but
the change is smaller as the surrender charge is lower. Hence, the surplus
emerging in year 2 decreases.

(iii) Overall the total present value of profits will increase, because the profits are
released earlier when reserves are smaller. This increases the NPV whenever the
risk discount rate is greater than the assumed earned rate.

Graders’ Comments: Many omitted part (c). Of the candidates who attempted it, many
answered (c)(i) correctly or with a minor error (such as omitting AV2 from the expected
cost of death benefit).
Page 105
Parts (c)(ii) and (iii) proved to be challenging conceptual questions. A few strong
candidates correctly answered (c)(ii), and a handful correctly (and impressively)
answered (c)(iii). Some candidates offered one-word answers (`increase' or `decrease'),
which received no credit.

Page 106
 Question 55
Note: This question is on a new topic. It is particularly suited to the use of Excel for calculations.
The candidate would be expected to include all the detail given here, at a minimum, for full
credit. It will not be sufficient to write down the numerical answers without explanation. Excel
workbooks will not be submitted or graded.

(a)
t Ft MCt F t+1
0 5000.0 125.0 4192.5
1 4192.5 104.8 3883.3
2 3883.3 97.1 3937.7

=
Where MCt 0.025=
Ft and Ft +1 ( Ft − MCt )(1 + Rt ) , and Rt is the return on assets in the
t-th year.

(b)
t Income Expenses Interest EDBt EMBt Prt
0* 200.0 -200.0
1 475.0 100.0 18.8 40.4 353.4
2 104.8 71.9 1.6 67.0 -32.5
3 97.1 68.8 1.4 74.4 988.0 -1032.7

Explanation:
• The income is the MC from part (a), plus 350 policy fee in the first year.
• Expenses at time 0 are the 200 pre-contract costs; in year 1 the expenses are 2%
of F0; in year 2 and 3 the expenses are 1% of Ft. plus 30.
• Interest is 5% of income – expenses.
• The expected death benefit cost, EDBt, is qx +t −1 ( 5000 − Ft )
+

• The expected maturity benefit cost, EMBt is 0 in the first two years. In the third
year is is px + 2 ( 5000 − F3 )
+

• The Profit vector value for the t-th year is

Prt = Incomet − Expensest + Interest t − EDBt − EMBt
(c)
t Prt t-1px Πt NPV(t)
0 -200.0 -200.0 -200
1 353.4 1 353.4 121.2
2 -32.5 0.95 -30.8 95.8
3 -1032.7 0.893 -922.2 -597.1

Explanation:
• Prt is the profit vector from (b)
• t −1 px is the survival probability to the start of the t-th year

Page 107
 • The profit signature at t is Π t = t −1 px Prt for t ≥ 1; Π 0 =Pr0 .
• NPV(t) is the partial NPV:
NPV (0) = Π 0 ;
NPV (t ) = NPV (t − 1) + Π t vrdr
t

Where the discount factor uses the risk discount rate of 10%.

The NPV is NPV(3)= –597.1.

(d) The Black Scholes Price for a put option on a premium of P with a guarantee of kP, is

(
BSP(t ) = P ke− r (T −t ) Φ ( −d 2 ) − ξ Φ ( −d1 ) )
where ξ is the expense deduction factor, and
log(ξ / k ) + (r + σ 2 / 2)(T − t )
d1 = , d=
2 d1 − σ T − t .
σ T −t
In this case:
P = 1, k= 1, ξ= (1 − 0.025) = 0.92686
3

log(0.92686) + (0.04 + 0.252 / 2)3

=d1 = 0.318228
0.25 3
=d 2 0.318228 − 0.25 3 = −0.114785
= e−
⇒ BSP Φ ( 0.114785 ) − 0.92686 Φ ( −0.318228
= ) 0.13627
3(0.04)

(e) The GMDB option cost is

qx (446.7) + 1| qx (590.8) + 2| qx ( 5000 × 0.13627 ) = 98.60

=
= =
=
as qx 0.05, 1| qx 0.95 × 0.06 0.057, and 2| qx 0.95 × 0.94
= × 0.07 0.0625

The GMMB cost is

3 px × 5000 × 0.13627 =
565.85

So the total cost is 664.45

(f)
t Income Expenses Interest EDBt EMBt Prt
0* 200.0 -200.0
1 475.0 765.0 -14.5 -304.5
2 104.8 71.9 1.6 34.5
3 97.1 68.8 1.4 29.7

NPV = −200 − 304.5v + px (34.5)v 2 + 2 px (29.7)v3 =

−429.8, at 10%

Page 108
(g)
Advantage: Hedging the guarantees limits the downside risk; we see from this stress test
that we are replacing the uncertain and potentially very serious future loss from
guarantees, with the certain cost of buying the options.

Disadvantage: The cost of the options is substantial. The income for this policy is 350
from the policy fee, plus around 1.5% of the fund annually (the management charge
minus the fund management costs) , which even on optimistic assumptions is unlikely to
be greater than 550 in total. This is insufficient to cover the hedge costs. (This is a sign
that the guarantees are under-priced, rather than an argument against hedging.).

Page 109
 Question 56

(a) π (0) E 0 e ( kP − F10 ) I(Tx > 10)  where P is the premium, k is the proportion of
−10 r +
= Q
 
premium guaranteed under the GMMB, r is the risk free rate, F10 is the fund value at time
10, and I(Tx >10) is an indicator random variable taking the value 1 if (x) survives 10
years, 0 otherwise. The Q superscript indicates that we are taking expectations under the
risk neutral measure, and the 0 subscript indicates that we are valuing based on the
information available at time 0.
In our case, k = 1 , and Ft = P(0.95)e−12(0.002 )t St , for t = 0, 112 , 2 12 , 312 , ,10.
We assume that the Tx and Ft random variables are independent, and that the Q and P
measure expectations for the future lifetime are the same. Then
E 0 [ I=
(Tx > 10) ] Pr
= [Tx > 10]
Q
10 px .
=
Also (
let ξ 0.95 e−120*0.002
= ⇒ F10 Pξ S10)
Then π (0) = 10 px EQ0 e−10 r ( P − Pξ S10 ) 
 1 
= 10 px P ξ EQ0 e−10 r  − S10  
 ξ 
= 10 px P ξ BSP 1 ( ξ ) as required, with ξ = 0.74730 and K = ξ * −1
.

(b)
= 10 p60 =
0.94255; =
P 100, 000; ξ 0.747230

10 p60 P= ( ξ)
ξ BSP 1 10
1 −
ξ

p60 Pξ  e 10 r Φ ( −d 2 ) − Φ ( −d1 ) 

= 10 p60 P e ( −10 r
Φ ( −d 2 ) − ξ Φ ( −d1 ) )
log(ξ ) + (r + σ / 2)(10)
2
d1 = = 0.53279; d 2 = d1 − σ 10 =
−0.25778
σ 10
⇒ π (0) = 0.94255 (100, 000 ) e ( −10×0.04
Φ ( 0.25778 ) − 0.74723Φ ( −0.53279 ) )
=0.94255(100,000)(0.181325)=17,091
The normal cdf values are found using Excel.

(c) The bond part of the hedge at t = 0 is 10 p60 Pe

−10(0.04)
Φ ( −d 2 ) =38017 .
The stock part of the hedge at t = 0 is − 10 p60 Pξ Φ ( −d1 ) =−20926 .
After one month, the bond part of the hedge is worth 38, 017e0.04/12 = 38,144
After one month, the stock part of the hedge is worth −20, 926(0.97) = −20, 298
The total hedge value before rebalancing is 17,846.

Page 110
(d) Let c denote the annual rate of payment, and let m = 12 × 0.002 denote the annual
management charge rate. The equation of value, under the Q measure, is
c
12
−r
( 

) c
12
−2 r
17, 091 = EQ0  F112 e 12 I T60 > 1  + EQ0  F212 e 12 I T60 > 2  + ...
12 12( 

)
c
( 
)
c
... + EQ0  F91112 e 12 I T60 > 911  + EQ0  F10 e− r I (T60 > 10 ) 
 12
−911 ×r
12  12
10 

 c 
 12 
{
=   P (0.95) 112 p60 E 0  S 112 e 12 e−  + 212 p60 E 0  S 212 e 12 e 12  + 
Q −r m /12 Q

−2 r −2 m


+  + 91112 p60 EQ0  S91112 e

−91112×r −91112×m
e   }
 + 10 p60 EQ0  S10 e−10 r e−10 m 


Note that under the risk neutral measure, E 0  St e  = S0 , and in this case S0 = 1 , so we
Q − rt

have
17, 091
c
12
{
( 0.95) P 112 p60 e− 12 + 212 p60 e− 12 + 312 p60 e− 12 +  + 10 p60e−10 m
m 2m 3m

}
c
( )
= ( 0.95 ) P 12a60:10 at i = em − 1 = 2.429%
12
(12)

17, 091
=
⇒c = (12)
0.0207
0.95 Pa60:10
0.0207
The monthly rate of deduction is thus 12
= 0.0017.

Page 111
Question 57

This question is an Excel question. Please see the ALTAM Sample Excel Solution file for
details. The content of the Excel file for this question is provided below for reference.

(a)
Time (t) µ5001+t µ5002+t µ50
12
+t t p5000 t
01
p50 t p5002
0 0.0200 0.0012 0.0023 1.0000 0.0000 0.0000
0.083 0.0200 0.0012 0.0023 0.9982 0.0017 0.0001
0.167 0.0200 0.0012 0.0023 0.9965 0.0033 0.0002
0.250 0.0200 0.0012 0.0024 0.9947 0.0050 0.0003
0.333 0.0200 0.0012 0.0024 0.9930 0.0066 0.0004
0.417 0.0200 0.0012 0.0024 0.9912 0.0083 0.0005
0.500 0.0200 0.0012 0.0024 0.9895 0.0100 0.0006
0.583 0.0200 0.0012 0.0024 0.9877 0.0116 0.0007
0.667 0.0200 0.0012 0.0025 0.9860 0.0132 0.0008
0.750 0.0200 0.0012 0.0025 0.9842 0.0149 0.0009
0.833 0.0200 0.0012 0.0025 0.9825 0.0165 0.0010
0.917 0.0200 0.0013 0.0025 0.9807 0.0182 0.0011
1.000 0.0200 0.0013 0.0025 0.9790 0.0198 0.0012
1.083 0.0200 0.0013 0.0026 0.9773 0.0214 0.0013
1.167 0.0200 0.0013 0.0026 0.9755 0.0230 0.0014
1.250 0.0200 0.0013 0.0026 0.9738 0.0247 0.0015
1.333 0.0200 0.0013 0.0026 0.9721 0.0263 0.0017
1.417 0.0200 0.0013 0.0026 0.9703 0.0279 0.0018
1.500 0.0200 0.0013 0.0027 0.9686 0.0295 0.0019
1.583 0.0200 0.0013 0.0027 0.9669 0.0311 0.0020
1.667 0.0200 0.0014 0.0027 0.9652 0.0327 0.0021
1.750 0.0200 0.0014 0.0027 0.9635 0.0343 0.0022
1.833 0.0200 0.0014 0.0028 0.9617 0.0359 0.0023
1.917 0.0200 0.0014 0.0028 0.9600 0.0375 0.0025
2.000 0.0200 0.0014 0.0028 0.9583 0.0391 0.0026
2.083 0.0200 0.0014 0.0028 0.9566 0.0407 0.0027
2.167 0.0200 0.0014 0.0028 0.9549 0.0423 0.0028
2.250 0.0200 0.0014 0.0029 0.9532 0.0439 0.0029
2.333 0.0200 0.0014 0.0029 0.9515 0.0454 0.0031
2.417 0.0200 0.0015 0.0029 0.9498 0.0470 0.0032
2.500 0.0200 0.0015 0.0029 0.9481 0.0486 0.0033
2.583 0.0200 0.0015 0.0030 0.9464 0.0501 0.0034
2.667 0.0200 0.0015 0.0030 0.9447 0.0517 0.0036
2.750 0.0200 0.0015 0.0030 0.9430 0.0533 0.0037
2.833 0.0200 0.0015 0.0030 0.9413 0.0548 0.0038

Page 112
 Time (t) µ5001+t µ5002+t µ50
12
+t t p5000 t
01
p50 t p5002
2.917 0.0200 0.0015 0.0031 0.9396 0.0564 0.0040
3.000 0.0200 0.0015 0.0031 0.9380 0.0579 0.0041
3.083 0.0200 0.0016 0.0031 0.9363 0.0595 0.0042
3.167 0.0200 0.0016 0.0031 0.9346 0.0610 0.0044
3.250 0.0200 0.0016 0.0032 0.9329 0.0626 0.0045
3.333 0.0200 0.0016 0.0032 0.9312 0.0641 0.0047
3.417 0.0200 0.0016 0.0032 0.9296 0.0656 0.0048
3.500 0.0200 0.0016 0.0032 0.9279 0.0672 0.0049
3.583 0.0200 0.0016 0.0033 0.9262 0.0687 0.0051
3.667 0.0200 0.0017 0.0033 0.9245 0.0702 0.0052
3.750 0.0200 0.0017 0.0033 0.9229 0.0718 0.0054
3.833 0.0200 0.0017 0.0034 0.9212 0.0733 0.0055
3.917 0.0200 0.0017 0.0034 0.9195 0.0748 0.0057
4.000 0.0200 0.0017 0.0034 0.9179 0.0763 0.0058
4.083 0.0200 0.0017 0.0034 0.9162 0.0778 0.0060
4.167 0.0200 0.0017 0.0035 0.9146 0.0793 0.0061
4.250 0.0200 0.0018 0.0035 0.9129 0.0808 0.0063
4.333 0.0200 0.0018 0.0035 0.9112 0.0823 0.0064
4.417 0.0200 0.0018 0.0036 0.9096 0.0838 0.0066
4.500 0.0200 0.0018 0.0036 0.9079 0.0853 0.0068
4.583 0.0200 0.0018 0.0036 0.9063 0.0868 0.0069
4.667 0.0200 0.0018 0.0037 0.9046 0.0883 0.0071
4.750 0.0200 0.0018 0.0037 0.9030 0.0897 0.0073
4.833 0.0200 0.0019 0.0037 0.9014 0.0912 0.0074
4.917 0.0200 0.0019 0.0038 0.8997 0.0927 0.0076
5.000 0.0200 0.0019 0.0038 0.8981 0.0942 0.0078

(b)
Death Premium
Discount Benefit Annuity
Time (t) t p5000 t
01
p50 t p5002 Factor EPV Expenses EPV
0.000 1.0000 0.0000 0.0000 1.0000 0.00 505.00 0.0833
0.083 0.9982 0.0017 0.0001 0.9950 47.78 4.97 0.0828
0.167 0.9965 0.0033 0.0002 0.9901 48.00 4.95 0.0822
0.250 0.9947 0.0050 0.0003 0.9851 48.21 4.92 0.0817
0.333 0.9930 0.0066 0.0004 0.9802 48.43 4.90 0.0811
0.417 0.9912 0.0083 0.0005 0.9754 48.65 4.87 0.0806
0.500 0.9895 0.0100 0.0006 0.9705 48.87 4.85 0.0800
0.583 0.9877 0.0116 0.0007 0.9657 49.09 4.83 0.0795
0.667 0.9860 0.0132 0.0008 0.9609 49.31 4.80 0.0789
0.750 0.9842 0.0149 0.0009 0.9561 49.54 4.78 0.0784
0.833 0.9825 0.0165 0.0010 0.9513 49.76 4.75 0.0779

Page 113
 Death Premium
Discount Benefit Annuity
Time (t) t p5000 t
01
p50 t p5002 Factor EPV Expenses EPV
0.917 0.9807 0.0182 0.0011 0.9466 49.99 4.73 0.0774
1.000 0.9790 0.0198 0.0012 0.9419 50.21 4.70 0.0768
1.083 0.9773 0.0214 0.0013 0.9372 50.44 4.68 0.0763
1.167 0.9755 0.0230 0.0014 0.9326 50.67 4.66 0.0758
1.250 0.9738 0.0247 0.0015 0.9279 50.91 4.63 0.0753
1.333 0.9721 0.0263 0.0017 0.9233 51.14 4.61 0.0748
1.417 0.9703 0.0279 0.0018 0.9187 51.38 4.59 0.0743
1.500 0.9686 0.0295 0.0019 0.9141 51.61 4.56 0.0738
1.583 0.9669 0.0311 0.0020 0.9096 51.85 4.54 0.0733
1.667 0.9652 0.0327 0.0021 0.9051 52.09 4.52 0.0728
1.750 0.9635 0.0343 0.0022 0.9006 52.33 4.49 0.0723
1.833 0.9617 0.0359 0.0023 0.8961 52.57 4.47 0.0718
1.917 0.9600 0.0375 0.0025 0.8916 52.81 4.45 0.0713
2.000 0.9583 0.0391 0.0026 0.8872 53.06 4.42 0.0709
2.083 0.9566 0.0407 0.0027 0.8828 53.31 4.40 0.0704
2.167 0.9549 0.0423 0.0028 0.8784 53.55 4.38 0.0699
2.250 0.9532 0.0439 0.0029 0.8740 53.80 4.36 0.0694
2.333 0.9515 0.0454 0.0031 0.8697 54.05 4.33 0.0690
2.417 0.9498 0.0470 0.0032 0.8653 54.31 4.31 0.0685
2.500 0.9481 0.0486 0.0033 0.8610 54.56 4.29 0.0680
2.583 0.9464 0.0501 0.0034 0.8567 54.82 4.27 0.0676
2.667 0.9447 0.0517 0.0036 0.8525 55.07 4.25 0.0671
2.750 0.9430 0.0533 0.0037 0.8482 55.33 4.23 0.0667
2.833 0.9413 0.0548 0.0038 0.8440 55.59 4.20 0.0662
2.917 0.9396 0.0564 0.0040 0.8398 55.85 4.18 0.0658
3.000 0.9380 0.0579 0.0041 0.8356 56.12 4.16 0.0653
3.083 0.9363 0.0595 0.0042 0.8315 56.38 4.14 0.0649
3.167 0.9346 0.0610 0.0044 0.8274 56.65 4.12 0.0644
3.250 0.9329 0.0626 0.0045 0.8232 56.92 4.10 0.0640
3.333 0.9312 0.0641 0.0047 0.8191 57.19 4.08 0.0636
3.417 0.9296 0.0656 0.0048 0.8151 57.46 4.06 0.0631
3.500 0.9279 0.0672 0.0049 0.8110 57.73 4.04 0.0627
3.583 0.9262 0.0687 0.0051 0.8070 58.00 4.01 0.0623
3.667 0.9245 0.0702 0.0052 0.8030 58.28 3.99 0.0619
3.750 0.9229 0.0718 0.0054 0.7990 58.56 3.97 0.0614
3.833 0.9212 0.0733 0.0055 0.7950 58.83 3.95 0.0610
3.917 0.9195 0.0748 0.0057 0.7910 59.12 3.93 0.0606
4.000 0.9179 0.0763 0.0058 0.7871 59.40 3.91 0.0602
4.083 0.9162 0.0778 0.0060 0.7832 59.68 3.89 0.0598
4.167 0.9146 0.0793 0.0061 0.7793 59.97 3.87 0.0594

Page 114
 Death Premium
Discount Benefit Annuity
Time (t) t p5000 t
01
p50 t p5002 Factor EPV Expenses EPV
4.250 0.9129 0.0808 0.0063 0.7754 60.25 3.85 0.0590
4.333 0.9112 0.0823 0.0064 0.7716 60.54 3.83 0.0586
4.417 0.9096 0.0838 0.0066 0.7677 60.83 3.81 0.0582
4.500 0.9079 0.0853 0.0068 0.7639 61.12 3.79 0.0578
4.583 0.9063 0.0868 0.0069 0.7601 61.42 3.77 0.0574
4.667 0.9046 0.0883 0.0071 0.7563 61.71 3.75 0.0570
4.750 0.9030 0.0897 0.0073 0.7525 62.01 3.74 0.0566
4.833 0.9014 0.0912 0.0074 0.7488 62.31 3.72 0.0562
4.917 0.8997 0.0927 0.0076 0.7451 62.61 3.70 0.0559
5.000 0.8981 0.0942 0.0078 0.7414 62.91

EPV = 3294.96 (sum of Death Benefit EPV column)

(c) Premium
= (3294.96 + (sum of Expenses column) / (sum of Premium Annuity EPV column) / 12
= (3294.96 + 759.08) / 4.12 / 12 = 81.93

(d)
(1) A greater force of transition to the Disabled state will reduce the chances of being in
the Healthy state and hence the expected number of premium payments. So the
premium will need to be higher to compensate for the fewer expected number of
premium payments.
(2) The force of transition to the Dead state is higher from the Disabled state than from
the Healthy state, hence there will now be a larger probability of death, increasing the
EPV (Benefits) and hence the gross premium.

Page 115
Question 58

This question is an Excel question. Please see the ALTAM Sample Excel Solution file for
details. The content of the Excel file for this question is provided below for reference.

(a)

Reserve Expected Expected Profit Profit

at time cost of Reserve Vector Signature
Time (t) t-1 Premium Expenses Interest Claims at t (a) (i) (a) (ii)
0 300.0 -300.0 -300.0
1 3000 250.0 165.0 1041.3 940.1 933.6 933.6
2 1000 3000 252.5 224.9 1167.0 1408.4 1397.0 1313.3
3 1500 3000 255.1 254.7 1308.1 1875.1 1316.4 1162.0
4 2000 3000 257.7 284.5 1466.4 2340.2 1220.3 1009.9
5 2500 3000 260.4 314.4 1644.0 2803.1 1106.8 857.4
6 3000 3000 263.1 344.2 1843.3 3263.7 974.0 705.0
7 3500 3000 266.0 374.0 2066.8 3721.5 819.8 553.3
8 4000 3000 268.9 403.9 2317.5 4175.9 641.6 402.9
9 4500 3000 271.8 433.7 2598.4 4626.6 436.9 254.6
10 5000 3000 274.9 463.5 2913.2 5072.8 202.6 109.3
11 5500 3000 278.0 493.3 3265.8 5804.0 -354.6 -176.3
12 6000 3000 281.2 523.1 3660.7 5780.4 -199.1 -95.8
13 6000 3000 284.5 522.9 4102.5 5753.8 -618.0 -286.4
14 6000 3000 287.9 522.7 4596.8 5724.2 -1086.1 -482.7
15 6000 3000 291.3 522.5 5149.3 5691.0 -1609.1 -682.3
16 6000 3000 294.8 522.3 5766.5 5654.0 -2193.0 -882.0
17 6000 3000 298.5 522.1 6455.4 5612.7 -2844.4 -1078.1
18 6000 3000 302.2 521.9 7223.7 5566.6 -3570.6 -1265.9
19 6000 3000 306.0 521.6 8079.8 5515.2 -4379.3 -1440.5
20 6000 3000 309.9 521.4 9032.6 0.0 179.0 54.1

(b) Present value of Profit Signature column at 12%:

= –300.0 + 933.6×1.12-1 + 1313.3×1.12-2 … = 3441.59

(t )
(c) Profit margin = 3441.59 / (Sumproduct(Premium, t −1 p70 ,PV factors @ 12%) × 1.12
= 19.24%

(d) It is not advisable to have negative emerging profit in the later years of the contract. That
requires the insurer to acquire additional cash from elsewhere, which may not be feasible.

Page 116
 (e) (i)

Reserve Expected Profit Profit

Time at time cost of Expected Vector Signature
(t) t-1 Premium Expenses Interest Claims Reserve at t (a) (i) (a) (ii)
0 300.0 -300.0 -300.0
1 0 3000 250.0 165.0 1041.3 0.0 1873.7 1873.7
2 0 3000 252.5 164.9 1167.0 0.0 1745.3 1640.8
3 0 3000 255.1 164.7 1308.1 1407.1 194.4 171.6
4 1501 3000 257.7 254.6 1466.4 3031.4 0.0 0.0
5 3238 3000 260.4 358.7 1644.0 4692.7 0.0 0.0
6 5022 3000 263.1 465.5 1843.3 6381.3 0.0 0.0
7 6843 3000 266.0 574.6 2066.8 8085.1 0.0 0.0
8 8690 3000 268.9 685.3 2317.5 9789.2 0.0 0.0
9 10549 3000 271.8 796.6 2598.4 11475.3 0.0 0.0
10 12402 3000 274.9 907.6 2913.2 13121.0 0.0 0.0
11 14226 3000 278.0 1016.9 3265.8 14699.0 0.0 0.0
12 15195 3000 281.2 1074.8 3660.7 15328.3 0.0 0.0
13 15911 3000 284.5 1117.6 4102.5 15641.2 0.0 0.0
14 16310 3000 287.9 1141.4 4596.8 15567.1 0.0 0.0
15 16317 3000 291.3 1141.6 5149.3 15018.1 0.0 0.0
16 15833 3000 294.8 1112.3 5766.5 13884.5 0.0 0.0
17 14734 3000 298.5 1046.1 6455.4 12026.4 0.0 0.0
18 12856 3000 302.2 933.3 7223.7 9263.7 0.0 0.0
19 9985 3000 306.0 760.7 8079.8 5360.0 0.0 0.0
20 5831 3000 309.9 511.3 9032.6 0.0 0.0 0.0

(ii) NPV = –300.0 + 1873.7×1.12-1 + 1640.8×1.12-2 + 171.6×1.12-3 = 2803.113

Page 117
Question 59

This question is an Excel question. Please see the ALTAM Sample Excel Solution file for
details. The content of the Excel file for this question is provided below for reference.

(a) The account value (AV) at age 90 is 84726.91.

Mort Cor
Age Year Rate DB Premiums Expenses iq ic Fact AV
60 1 0.003398 50,000 1,600 345 3% 5% 1.50 1,114.54
61 2 0.003792 50,000 1,600 153 3% 5% 1.48 2,469.14
62 3 0.004234 50,000 1,600 153 3% 5% 1.46 3,873.03
63 4 0.004730 50,000 1,600 153 3% 5% 1.44 5,327.55
64 5 0.005288 50,000 1,600 153 3% 5% 1.42 6,834.04
65 6 0.005915 50,000 1,600 153 3% 5% 1.40 8,394.04
66 7 0.006619 50,000 1,600 153 3% 5% 1.38 10,009.29
67 8 0.007409 50,000 1,600 153 3% 5% 1.36 11,681.81
68 9 0.008297 50,000 1,600 153 3% 5% 1.34 13,413.91
69 10 0.009294 50,000 1,600 153 3% 5% 1.32 15,208.40
70 11 0.010413 50,000 1,600 25 3% 5% 1.30 17,204.81
71 12 0.011670 50,000 1,600 25 3% 5% 1.28 19,280.25
72 13 0.013081 50,000 1,600 25 3% 5% 1.26 21,441.01
73 14 0.014664 50,000 1,600 25 3% 5% 1.24 23,694.94
74 15 0.016440 50,000 1,600 25 3% 5% 1.22 26,051.82
75 16 0.018433 50,000 1,600 25 3% 5% 1.20 28,523.89
76 17 0.020668 50,000 1,600 25 3% 5% 1.18 31,126.65
77 18 0.023175 50,000 1,600 25 3% 5% 1.16 33,879.73
78 19 0.025984 50,000 1,600 25 3% 5% 1.14 36,808.14
79 20 0.029132 50,000 1,600 25 3% 5% 1.12 39,943.93
80 21 0.032658 50,000 1,600 25 3% 5% 1.10 43,328.34
81 22 0.036607 50,000 1,600 25 3% 5% 1.08 46,980.20
82 23 0.041025 50,000 1,600 25 3% 5% 1.06 50,829.91
83 24 0.045968 50,000 1,600 25 3% 5% 1.04 54,901.66
84 25 0.051493 50,000 1,600 25 3% 5% 1.02 59,225.88
85 26 0.057665 50,000 1,600 25 3% 5% 1.00 63,840.92
86 27 0.064554 50,000 1,600 25 3% 5% 1.00 68,686.72
87 28 0.072237 50,000 1,600 25 3% 5% 1.00 73,774.80
88 29 0.080798 50,000 1,600 25 3% 5% 1.00 79,117.29
89 30 0.090326 50,000 1,600 25 3% 5% 1.00 84,726.91

Page 118
(b)
- DB at end of 15th year = 50000.00
- DB at end of 20th year = 50000.00
- DB at end of 25th year = 60410.40

(c)
- CV at end of 1st year = 0.00
- CV at end of 2nd year = 669.14

(d) The main purpose of the surrender charge is to ensure that the insurer receives enough to
pay the acquisition charges if the policy were to lapse.

(e) The annual premium is 957.94.

(f) The corridor factors do not affect the premium. This can be seen by noting that the AV
based on a death benefit using the corridor factors is always larger than without the
corridor factors. This is reasonable because with the lower premium, the cash value times
the corridor factor is always less than 50,000.

(g) The year of termination is 27.

(h) A no lapse guarantee guarantees that if the policyholder pays a certain minimum
premium each year, then the policy will stay in force even if the account value goes
negative. The guarantee could apply if the expense charges and/or the mortality charges
increase sufficiently to make the minimum premium insufficient.

Page 119
Question 60

This question is an Excel question. Please see the ALTAM Sample Excel Solution file for
details. The content of the Excel file for this question is provided below for reference.

a) The final salary is (3.698 + 3.643 + 3.589)/(3 × 3.186) × 100,000 = 114,354

b) See the table below:

Retirement Annuity Final Discount Actuarial

Age (y) Factor Avg Salary Factor Probability Liability
60 14.4414 106,152 0.78353 0.26675 136,172
60.5 14.3149 106,947 0.76464 0.05910 29,405
61.5 14.0559 108,548 0.72823 0.05324 25,138
62.5 13.7889 110,175 0.69355 0.04793 21,462
63.5 13.5139 111,828 0.66053 0.04313 18,298
64.5 13.2312 113,507 0.62907 0.03879 15,576
65 13.0870 114,354 0.61391 0.36765 143,556
ANSWER: 389,607
c) The normal contribution is 15,584, which is simply 1/25 of the actuarial liability.
d) The purpose is to ensure that employees receive a fair result for the amount that they have
contributed compared to other employees. It should ideally be calculated to make the
actuarial value of their benefits at the date of retirement approximately equal to the
present value of the benefits if they retire at the normal retirement age.
e) See the table below.
f) See the table below:
Replacement Rate
Age Reduction w/ reduction w/o reduction
60 29.0% 35.7% 50.3%
60.5 26.6% 37.5% 51.1%
61.5 21.5% 41.4% 52.8%
62.5 16.0% 45.7% 54.4%
63.5 10.0% 50.5% 56.1%
64.5 3.5% 55.8% 57.8%
65 0.0% 58.6% 58.6%

Page 120
Question 61
𝑆𝑆2 𝑆𝑆2
(a) Fund value at time 2 is �𝑆𝑆0� 10,000(0.972 ) = 9409 �𝑆𝑆0�
GMDB during year 2 is 10,000𝑒𝑒 2(0.06) = 11,274.97.
𝑆𝑆 11,274.97
Those are equal if � 2 � = = 1.1983. Anything higher, and the death benefit will be the
𝑆𝑆 0 9409
fund value.

(b) The Death Benefit for death in year 1 is

𝑆𝑆1
max(10,000𝑒𝑒 0.06 , 𝐹𝐹1 ) = max �10,000𝑒𝑒 0.06 , 0.97(10,000) �
𝑆𝑆0
𝑒𝑒 0.06 𝑆𝑆1
= 9,700 max � − , 0�
0.97 𝑆𝑆0
𝑒𝑒 0.06
which is 9,700 times the payoff of a put option with strike price 0.97
= 1.09468.

(i)
(1 − 0.03)1 0.302
�𝑙𝑙𝑙𝑙 � � + �0.04 + � (1)�
𝑒𝑒 0.06(1) 2
𝑑𝑑1 (0,1) = = −0.018197
[(0.30)10.5 ]
(ii)
𝑑𝑑2 (0,1) = 𝑑𝑑1 (0,1) − �(0.30)10.5 � = −0.318197
𝑣𝑣(0,1) = 9,700[1.09468 𝑒𝑒 −0.04 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁. 𝑆𝑆. 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷(0.318197,1)
− 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁. 𝑆𝑆. 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷(0.018197,1)] = 1,454.13
Contribution to hedge portfolio = 𝑞𝑞85 (1,454.13) = 0.057665(1,454.13) = 83.85

(iii)
(1 − 0.03)2 0.302
�𝑙𝑙𝑙𝑙 �0.06(2) � + �0.04 + 2 � (2)�
𝑑𝑑1 (0,2) = 𝑒𝑒 = −0.025735
[(0.30)20.5 ]
𝑑𝑑2 (0,2) = 𝑑𝑑1 (0,2) − �(0.30)20.5� = −0.449999
𝑣𝑣(0,2) = 10,000(0.972 )�1.094682 𝑒𝑒 −0.04(2) 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁. 𝑆𝑆. 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷(0.44999,1)
− 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁. 𝑆𝑆. 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷(0.025735,1)� = 2,210.27
Contribution to hedge portfolio = 1|q85 (2,210.27) = 0.060831(2,210.27) = 134.45
Hedge portfolio = 83.85 + 134.45 = 218.30

(iv) (See solution to Example 17.6 for how this formula arises)
218.30 = 10,000 𝑐𝑐 (1 + (1 − 0.03)𝑝𝑝85 ) = 10,000(1 + 0.97(0.942335)) 𝑐𝑐
= 19,140.65 𝑐𝑐
𝑐𝑐 = 0.0114

(v) The cost would not change. Chapter 17 of AMLCR uses the convention that 𝑆𝑆0 = 1, but 𝑆𝑆0 is
an index of stock prices. (As noted on page 632, “We interpret the stock price process as an
index for the fund assets; as an index. We can arbitrarily set 𝑆𝑆0 to any convenient value.)
(c)
(i)
(1 − 0.03)1 0.302 0.5
𝑑𝑑1 = �𝑙𝑙𝑙𝑙 � � + �0.04 + � (1)� / �(0.30)1 � = −0.018197
𝑒𝑒0.06(1) 2

Page 121
 Exactly the same as before. The exact value of the assets on the calculation of the hedge
portfolio does not appear, because it is irrelevant. Asset values follow the same lognormal
price process from that point, and the GMDBs (the strike prices) are the same multiples of
the starting assets)

So, the ratio of the 𝑑𝑑1 ’s is 1.000000000…

(ii) Similarly, 𝑑𝑑2 is the same as before.

The assets at the end of year 1 are 10,000(2)(0.97) = 19,400
The Reset Death Benefit for death in year 1 is
𝑆𝑆2
𝑚𝑚𝑚𝑚𝑚𝑚(10,000𝑒𝑒 0.06 , 𝐹𝐹2 ) = 𝑚𝑚𝑚𝑚𝑚𝑚 �19,400𝑒𝑒 0.06 , 0.97(19,400) �
𝑆𝑆1
𝑒𝑒0.06 𝑆𝑆2
= 18,818 𝑚𝑚𝑚𝑚𝑚𝑚 � − , 0�
0.97 𝑆𝑆1
𝑒𝑒 0.06
which is 18,818 times the payoff of a 1-year put option at time 1 with strike price =
0.97
1.09468
𝑣𝑣(1,1) = 18,818[1.09468 𝑒𝑒 −0.04 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁. 𝑆𝑆. 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷(0.318197,1)
− 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁. 𝑆𝑆. 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷(0.018197,1)] = 2821.02

Contribution to hedge portfolio is 𝑞𝑞86 (2,821.02) = 0.064555(2,821.02) = 182.11

𝑣𝑣(1,2) = 18,818(0.97)[1.094682 𝑒𝑒 −0.04(2) 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁. 𝑆𝑆. 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷(0.44999,1)

− 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁. 𝑆𝑆. 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷(0.025735,1)] = 4,287.93

Contribution to hedge portfolio is = 1|q86 (4,287.93) = 0.067574(4,287.93) = 289.76

Hedge portfolio = 182.11 + 289.76 = 471.87

(d) Your colleague is wrong. Expected values are additive, whether or not the underlying events are
independent.

Page 122