Contents

2 Life tables i 26
2.1 Life table functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
hb
2.2 Fractional age assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
da

2.2.1 Uniform distribution of death between integral ages . . . . . . . . . . . . . . . . 29

Ed

2.2.2 Constant force of mortality or Exponential distribution between integral ages . . 31

2.3 Select–and–ultimate tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
ed

2.4 Moments of future lifetime random variables . . . . . . . . . . . . . . . . . . . . . . . . 38

m

2.4.1 Uniform distribution of death for all ages . . . . . . . . . . . . . . . . . . . . . . 39

ha

2.4.2 Constant force of mortality or Exponential distribution for all ages . . . . . . . 39

M

2.5 Sample exam questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

f.
o

sa
Pr

u.
ed
u.

I
ks

els
@

od
i
hb

lM
da

ica
ed

at
m

m
he
at
lM
ia
ar
tu
Ac

25
 Chapter 2

Life tables
i
hb
da

2.1 Life table functions

Ed

A life table is a concrete way to look at the survivorship random variable. A life table specifies a
ed

certain number of lives at a starting integer age x0 . Usually x0 = 0. This number of lives at age x0
m

is called the radix. Then for each integer x > x0 , the expected number of survivors is listed. The
ha

notation for the number of lives listed in the table for age x is ℓx .
M

Assuming for simplicity that x0 = 0, the random variable for the number of lives at each age, L(x),
is a binomial random variable with parameters ℓ0 and x p0 . so the expected number of lives is
f.
o

sa
Pr

ℓx = ℓ0 x p0 . Similarly, ℓx+t = ℓx t px .
u.
ed

More importantly, t px can be calculated from the table using t px = ℓx+t

ℓx
. A life table also lists the
u.

I
expected number of deaths at each age; dx is the notation for this concept. Thus
ks

els
dx = ℓx − ℓx+1 = ℓx − ℓx px = ℓx (1 − px ) = ℓx qx .
@

od
i
hb

Therefore,
lM
da

dx
.
qx =
ica
ed

ℓx
at
m

In other words the difference ℓx − ℓx+t is the expected number of deaths over the age interval of
m

[x, x + t). We denote this by t dx . It immediately follows that

he

t dx = ℓx − ℓx+t .
at
lM

We can then calculate t qx and m|n qx , by the following two relations:

ia

ℓx+t ℓx − ℓx+t t dx
t qx = 1− = =
ar

ℓx ℓx ℓx
tu

Pn−1
ℓx+m − ℓx+m+n n dx+m dx+m+k
= k=0
Ac

m|n qx = = = P((x) dies between age x + m and x + m + n).

ℓx ℓx ℓx
The Illustrative Life Table contains a lot of information.
https://www.soa.org/Files/Edu/2018/ltam-standard-ultimate-life-table.pdf (for i = 5%)
For now, you only need to know and use the first three columns: x, ℓx , and qx . For example, to obtain
the value q60 , simply use the column labeled qx . You should obtain q60 = 0.003398. It is also possible,
but more tedious, to calculate q60 using the column labeled ℓx we have
96, 305.8
q60 = 1 − = 1 − 0.996603 = 0.003397.
96, 634.1

26
 27

Below an excerpt of the ILT indicated in the above link

x 20 21 22 23 24 25
ℓx 100000 99975 99949.7 99924 99897.8 99871.1
qx 0.00025 0.000253 0.000257 0.000262 0.000267 0.000273

From the above table we have ℓ20 = 100000 and q20 = 0.00025, hence d20 = 25, we have
ℓx+1
qx = 1 − ⇐⇒ qx ℓx = ℓx − ℓx+1 ⇐⇒ ℓx+1 = ℓx − qx = ℓx (1 − qx )
ℓx
hence bi
ℓ21 = ℓ20 (1 − q20 ) = 100000 (1 − 0.00025) = 99975.0
h
da

and
Ed

ℓ22 = ℓ21 (1 − q21 ) = 99975 (1 − 0.000253) = 99949.70633

To get values of t px and t qx for t > 1, you should always use the column labeled ℓx . For example, we
ed

have
m

ℓ25 99871.1
= 0.99871 and 5 q20 = 1 − 5 p20 = 1 − 0.99871 = 0.00129.
ha

5 p20 = =
ℓ20 100000
M
.
of

Exercise 2.1.1 You are given the following excerpt of a life table:
.sa
Pr

x 20 21 22 23 24 25
u
ed

ℓx 100, 000 99, 975.0 99, 949.7 99, 924.0 99, 897.8 99, 871.1
u.

I
dx ? ? ? ? ?

els
ks

Calculate the following: a) 3 p20 , b) q24 , c) 4| q20 .

@

od
bi

lM
ah

Solution: To complete the table we can use either dx = ℓx − ℓx+1 or dx = ℓx qx

ica
d
ed

x 20 21 22 23 24 25
at
m

ℓx 100, 000 99, 975.0 99, 949.7 99, 924.0 99, 897.8 99, 871.1
m

dx 25 25.3 25.7 26.2 26.7 Unknown

he
at

a) 3 p20 = ℓℓ23 99924

= 100000 = 0.99924, b) q24 = dℓ2424 26.7
= 99897.8 = 0.0002672 or q24 = 1 − ℓℓ25 = 1− 99871.1
≃
lM

20 24 99897.8
ℓ24 −ℓ25 99897.8−99871.1 d24 26.7
0.0002672. c) 4| q20 = ℓ20 = 100000
= 0.000267. or 4| q20 = ℓ20 = 100000 = 0.000 267.
ia

t
Example 2.1.1 You are given S0 (t) = 1 − 100 for 0 ≤ t ≤ 100 and ℓ0 = 100.
r
ua

(a) Find an expression for ℓx , for 0 ≤ x ≤ 100 and find µx . (b) Calculate q2 . (c) Calculate 3 q2 .
t
Ac

x
Solution: a) We have ℓx = ℓ0 x p0 = ℓ0 S0 (x) = 100(1 − 100 ) = 100 − x. (UDD or De Moivre’s law),
1 Sx′ (t)
hence µx = 100−x . With a direct method we get µx+t = − Sx (t) for x = 0, we obtain
1
S0′ (t) 100 1
µt = − = t = .
S0 (t) 1 − 100 100 − t

You can also use

S0′ (x+t)
S ′ (t) S0 (x) S ′ (x + t)
µx+t =− x = − S0 (x+t) =− 0
Sx (t) S0 (x + t)
S0 (x)
 28

S ′ (x) 1
for t = 0 we get µx = − S00 (x) = 100−x
.
3
S0 (3) 1− 100 100−3
b) q2 = 1 − p2 = 1 − S2 (1) = 1 − S0 (2)
= 1− 2
1− 100
= 1− 100−2
= 0.01020. We can also use
ℓ3 100−3
q2 = 1 − p2 = 1 − = 1 − ℓ2
= 0.01020.
100−2
ℓ5 100−5
c) 3 q2 = 1 − 3 p2 = 1 − ℓ2 = 1 − 100−2 = 0.03061.

Exercise 2.1.2 Use a radix of 1, 000, 000 at age 60 to complete the following mortality table

x 60 61 62 63 64 65
qx 0.001 0.002 0.003 0.004 0.005
i ℓx ? ? ? ? ? ?
hb
da
Ed

Solution: We shall recursively calculate ℓx , x = 61, . . . , 65 using ℓx+1 = ℓx (1 − qx ), thus

ed

x 60 61 62 63 64 65
m

qx 0.001 0.002 0.003 0.004 0.005 Unknown

ha

ℓx 1, 000, 000 999, 000 997, 002 994, 011 990, 035 985, 085
M
f.
o

sa
Pr

Exercise 2.1.3 You are given the following life table:

u.
ed

x 0 1 2
u.

I
ℓx 890
ks

els
dx 50
@

od
px 0.98
i
hb

Calculate 2 p0 . lM
da

ica
ed

at
m

ℓ2
Solution: We know that 2 p0 = ℓ0
, where ℓ2 is given in the table and ℓ0 = ℓ1 + d0 thus
m
he

ℓ2 ℓ2
2 p0 = =
at

ℓ0 ℓ1 + d0
lM

Now, we need to find ℓ1 , so using ℓ2 = ℓ1+1 = ℓ1 p1 , hence

ia

ℓ2 890
ar

2 p0 = = 890 = 0.92886.
tu

ℓ2
p1
+ d0 0.98
+ 50
Ac

Exercise 2.1.4 You are given:

(i) The probability that a person age 50 is alive at age 55 is 0.9.
(ii) The probability that a person age 55 is not alive at age 60 is 0.15.
(iii) The probability that a person age 50 is alive at age 65 is 0.54.
Calculate the probability that a person age 55 dies between ages 60 and 65.
 29

Solution: The required probability is


 − 10 p55
5 p55
P (5 ≤ T55 ≤ 10) = 5|5 q55 = 10 55 − 5 q55
q
5 p55 × 5 q60 .


From given information we have 5 p50 = 0.9, 5 q55 = 0.15, 15 p50 = 0.54. First observe that 15 p50 =
15 p50 0.54
5 p50 × 10 p55 , hence 10 p55 = 5 p50 = 0.9 = 0.6. Finally 5|5 q55 = 0.85 − 0.6 = 0.25.
Remark: You can also write
15 p50
5 q60 = 1 − 5 p60 = 1 − 5 p50+10 = 1 −
10 p50
i
hb
15 p50 15 p50
= 1− =1−
5 p50 5 p55 5 p50 (1 − 5 q55 )
da

0.54
Ed

= 1− = 0.29412.
0.9 (1 − 0.15)
ed

Hence 5|5 q55 = 0.85 × 0.29412 = 0.25.

m
ha
M

2.2 Fractional age assumptions

f.
o

sa
2.2.1 Uniform distribution of death between integral ages
Pr

u.

The Uniform Distribution of Death (UDD) assumption is extensively useful to calculate survival and
ed

death probabilities for fractional ages. The idea behind this assumption is that we use a bridge, denoted
u.

I
by U , to connect the (continuous) future lifetime random variable Tx and the (discrete) curtate future
ks

els
lifetime random variable Kx . Tx = Kx + U , hence U takes its values in [0, 1]. It is assumed that U
@

od
follows a uniform distribution over the interval [0, 1], and that U and Kx are independent. Then, for
i
hb

0 ≤ r < 1 and an integral value of x, we have

lM
da

ica
r qx = P (Tx ≤ r) = P (Kx + U ≤ r) = P (U ≤ r; Kx = 0) = P (U ≤ r) P (Kx = 0) = rqx .
ed

at
m

This means that under UDD, we have, for example, 0.3 q35 = 0.3q35 . The value of q35 can be obtained
m

straightforwardly from the life table: q35 = 0.000391. To calculate r px , for 0 ≤ r < 1, we use
he

r px = 1 − r qx = 1 − rqx . For example, we have 0.7 p40 = 1 − 0.7q40 .

at
lM

Remark 2.2.1 Observe that under UDD r px ̸= rpx . The equation r qx = rqx for 0 ≤ r < 1 and an
integral age x is equivalent to a linear interpolation between ℓx and ℓx+1 , that is,
ia
ar

ℓx+r = (1 − r)ℓx + rℓx+1 .

tu
Ac

Indeed

ℓx+r = ℓx r px = ℓx (1 − r qx ) = ℓx (1 − rqx )
= ℓx (1 − r (1 − px )) = (1 − r) ℓx + rℓx px = (1 − r) ℓx + rℓx+1 .

Exercise 2.2.1 Calculate under UDD 2.7 p30 using ILT for i = 0.05 (see LMS)
 30

Solution: We have under UDD

ℓ32
2.7 p30 = 2 p30 0.7 p32 = 2 p30 (1 − 0.7 q32 ) = (1 − 0.7q32 )
ℓ30
99, 663.2
= (1 − 0.7 × 0.000341) = 0.99912
99, 727.3
or alternatively

2.7 p30 = 2 p30 0.7 p32 = p30 p31 (1 − 0.7 q32 ) = (1 − q30 ) (1 − q31 ) (1 − 0.7q32 )
= (1 − 0.000315) (1 − 0.000327) (1 − 0.7 × 0.000341) = 0.99912
i
hb
Under UDD, we have the following equation for calculating the density function:
da

fx (r) = qx for all 0 ≤ r < 1.

Ed

Indeed fx (r) = d(dr

r qx )
= d(rq x)
ed

dr
= qx .
We can write also that
m
ha

qx = fx (r) = r px µx+r = (1 − r qx ) µx+r = (1 − rqx ) µx+r ,

M

therefore the force of mortality under UDD is given by

f.
o

sa
Pr

qx µx+r
u.

µx+r = ⇐⇒ qx = for all 0 ≤ r < 1 (2.2.1)

1 − rqx rµx+r + 1
ed
u.

I
Question: What if the subscript on the right–hand–side is not an integer? for example use ILT for
ks

els
i = 0.05 to calculate 0.2 p25.6 and 0.5 q30.7 under UDD.
@

od
Solution:
i
hb

0.8 p25 1− 0.8 q25 UDD 1 − 0.8q25 1 − 0.8 × 0.000273

lM
da

0.2 p25.6 = = = = = 0.99995

0.6 p25 1− 0.6 q25 1 − 0.6q25 1 − 0.6 × 0.000273
ica
ed

=1−
at
m

and 0.5 q30.7 0.5 p30.7 . First we have

m

1.2 p30 p30 1 − 0.2 q31 UDD

0.2 p31 1 − 0.2q31
he

0.5 p30.7 = = = (1 − q30 )

= (1 − q30 )
0.7 p30 0.7 p30 1 − 0.7 q30 1 − 0.7q30
at

1 − 0.2 × 0.000327
lM

= (1 − 0.000315) = 0.99984
1 − 0.7 × 0.000315
ia

Finally 0.5 q30.7 = 1 − 0.99984 = 0.00016.

ar
tu

Exercise 2.2.2 For a certain mortality table, you are given:

Ac

(i) µ80.5 = 0.020202, µ81.5 = 0.040816 and µ82.5 = 0.061856

(ii) Deaths are uniformly distributed between integral ages.
Calculate the probability that a person age 80.5 will die within two years.

Solution: The required probability is 2 q80.5 = 1 − 2 p80.5 . First calculate 2 p80.5 . We have

2.5 p80 2 p80 0.5 p82 p80 p81 (1 − 0.5 q82 ) UDD (1 − q80 ) (1 − q81 ) (1 − 0.5q82 )
2 p80.5 = = = = .
0.5 p80 1− 0.5 q80 1 − 0.5 q80 1 − 0.5q80
 31
µx+r
Recall also that from (2.2.1) qx = rµx+r +1
thus

µ80.5 0.020202
q80 = = = 0.02
0.5 × µ80.5 + 1 0.5 × 0.020202 + 1
µ81.5 0.040816
q81 = = = 0.04
0.5 × µ81.5 + 1 0.5 × 0.040816 + 1
µ82.5 0.061856
q82 = = = 0.06
0.5 × µ82.5 + 1 0.5 × 0.06185 6 + 1

Therefore
i
(1 − q80 ) (1 − q81 ) (1 − 0.5q82 ) (1 − 0.02) (1 − 0.04) (1 − 0.5 × 0.06)
hb
2 p80.5 = = = 0.92179.
1 − 0.5q80 1 − 0.5 × 0.02
da
Ed

Finally 2 q80.5 = 1 − 2 p80.5 = 1 − 0.92179 = 0.07821.

ed

2.2.2 Constant force of mortality or Exponential distribution between

m
ha

integral ages
M

The idea behind this assumption is that for every age x and any 0 ≤ r < 1 we assume that µx+r = µ.
This implies that
f.
o

R1 R1
sa

px = e− 0 µx+u du = e− 0 µ du = e−µ .
Pr

u.

For any integral age x, we can write

ed
u.

Rr Rr
r

I
r px = e− 0 µx+u du
= e− 0 µ du
= e−µr = e−µ = prx ∀ 0 ≤ r < 1
ks

els
@

Moreover, for any integral age x and any r, u such that 0 ≤ r < 1 and r + u ≤ 1 we have

od
i
hb

− 0r µx+u+s ds
lM
R Rr
r
r px+u = e = e− 0 µ du = e−µr = e−µ = prx .
da

ica
ed

at
m

Exercise 2.2.3 Calculate 0.2 p25.6 and 0.5 q30.7 under CFM using ILT for i = 0.05 (see LMS).
he
at

Solution:
lM

CFM 0.2
0.2 p25.6 = p0.2
25 = (1 − q25 ) = (1 − 0.000273)0.2 = 0.99995
ia
ar

and 0.5 q30.7 =1− 0.5 p30.7 . First we have

tu

0.2
1.2 p30 p30p31 0.2 p31 CFM
Ac

0.5 p30.7 = = = (1 − q30 )0.3 (1 − q31 )0.2

= p30
0.7 p30 0.7 p30 p30
0.7

= (1 − 0.000315) (1 − 0.000327)0.2 = 0.99984.

0.3

Finally 0.5 q30.7 = 1 − 0.99984 = 0.00016.

Exercise 2.2.4 For a certain mortality table, you are given:

(i) µ80.5 = 0.020202, µ81.5 = 0.040816 and µ82.5 = 0.061856
(ii) Assume constant force of mortality between integral ages.
Calculate the probability that a person age 80.5 will die within two years.
 32

Solution: The required probability is 2 q80.5 = 1 − 2 p80.5 . First calculate 2 p80.5 . We have

2.5 p80 2 p80 0.5 p82 CFM p80 p81 p0.5

82 0.5 0.5
p81 p0.5


2 p80.5 = = = 0.5
= p80 82 = p81 (p80 p82 )
0.5 p80 0.5 p80 p80
Recall that
R1
p80 = e− 0 µ80+u du
= e−µ80.5 = e−0.020202 = 0.98,
R1
p81 = e− 0 µ81+u du
= e−µ81.5 = e−0.040816 = 0.96,
R1
p82 = e− 0 µ82+u du
= e−µ82.5 = e−0.061856 = 0.94.
i
hb
Therefore
da

2 p80.5 = (0.96) (0.98 × 0.94)0.5 = 0.92140.

Ed

Finally 2 q80.5 = 1 − 2 p80.5 = 1 − 0.92140 = 0.0786.

ed

Remark: In case we do not haveRUDD or CFMR how one can compute t px and t qx for any t > 0?
t x+t
m

One has to compute the integral 0 µx+u du = x µu du for a given force of mortality. Or use
Rt
ha

t qx = 0 fx (u) du whenever the density is given.

M
f.

2.3 Select–and–ultimate tables

o

sa
Pr

u.

Suppose you selected two 40–year–old men from the population. The first one was selected randomly,
ed

whereas the second one had recently purchased a life insurance policy. Would the mortality rate for
u.

both of these, q40 be the same? No. The second person was under written for a life insurance policy,

I
ks

els
which means his medical situation was reviewed. He had to satisfy certain guidelines regarding
@

weight, blood pressure, blood lipids, family history, existing medical conditions, and possibly even

od
i
hb

driving record and credit history. The fact he was approved for an insurance policy implies that his
mortality rate is lower than that of a randomly selected 40–year–old man. lM
da

ica
Not only would q40 be different. If both men survived 5 years, q45 would be different as well. A
ed

man whose health was established 5 years ago will have better mortality than a randomly selected
at
m

man.
m

International Actuarial Notation provides two-parameter notation for all actuarial functions. The
he

parameters are written as subscripts with a bracket around the first parameter and a plus sign between
at

the parameters. In other words, the subscript is of the form [x] + t. When t = 0, it is omitted. Thus
lM

the mortality rate for a 40–year–old who just purchased a life insurance policy would be written as
q[40] . The mortality rate for a 45–year–old who purchased a policy at age 40 would be written q[40]+5
ia
ar

Exercise 2.3.1 You are given: ℓ[45] = 1000, 5 q[45] = 0.04, 5 q[45]+5 = 0.05. Calculate ℓ[45]+10 .
tu
Ac

ℓ
Solution: We know 5 q[45]+5 = 1 − 5 p[45]+5 = 1 − ℓ[45]+10
[45]+5
Due to the effect of underwriting, a select death probability q[x]+t must be no greater than the
corresponding ordinary death probability qx+t · However, the effect of underwriting will not last forever.
The period after which the effect of underwriting is completely gone is called the select period. Suppose
that the select period is n years, we have

q[x]+t < qx+t for all t < n and q[x]+t = qx+t for all t ≥ n

The ordinary death probability qx+t is called the ultimate death probability and q[x]+t is called
the select death probability. A life table that contains both select probabilities and ultimate
 33

probabilities is called a select–and–ultimate life table. The following is an excerpt of a (hypothetical)

select–and–ultimate table with a select period of two years:
x q[x] q[x]+1 qx+2 x+2
40 0.04 0.06 0.08 42
41 0.05 0.07 0.09 43
42 0.06 0.08 0.10 44
43 0.07 0.09 0.11 45
It is important to know how to apply such a table. Let us consider a person who is currently age 41
and is just selected. Calculate q[41]+k for k = 0, 1, 2, 3, 4.
i
hb
Age 41 q[41] = 0.05
Age 42 q[41]+1 = 0.07
da

Age 43 q[41]+2 = q43 = 0.09

Ed

Age 44 q[41]+3 = q44 = 0.10

ed

Age 45 q[41]+4 = q45 = 0.11

m

Calculate p[42]+k for k = 0, 1, 2, 3. HOMEWORK

ha

To further illustrate, let us consider a person who is currently age 41 and was selected at age 40.
M

The death probabilities q[40]+k for k = 1, 2, 3, 4, 5 for this person are as follows:
f.

Age 41 q[40]+1 = 0.06

o

sa
Pr

Age 42 q[40]+2 = 0.08

u.

Age 43 q[40]+3 = q43 = 0.09

ed

Age 44 q[40]+4 = q44 = 0.10

u.

I
Age 45 q[40]+5 = q45 = 0.11
ks

els
@

Calculate p[43]+k for k = 0, 1, 2. HOMEWORK

od
i
hb

We may measure the effect of underwriting by the index of selection, which is defined as follows:
q[x]+k lM
da

I(x, k) = 1 − .
ica
qx+k
ed

at
m

One can interpret this formula as follows: If the effect of underwriting is strong, then q[x]+k would
m

be small compared to qx+k , and therefore I(x, k) would be close to one. By contrast, if the effect of
he

underwriting is weak, then q[x]+k would be close to qx+k , and therefore I(x, k) would be close to zero.
at

For example, on the basis of the preceding table,

lM

q[41]+1 0.07 7 1
I(41, 1) = 1 − =1− = 1 − = = 0.125.
q42 0.08 8 8
ia
ar
tu

Example 2.3.1 Aicha was a newly selected life on 01/01/2000 and her age on 01/01/2001 is 21. Let p
Ac

denotes the probability on 01/01/2001 that Aicha will be alive on 01/01/2006. For a select–and–ultimate
mortality table with a 3–year select period:
x q[x] q[x]+1 q[x]+2 qx+3 x+3
20 0.09 0.11 0.13 0.15 23
21 0.10 0.12 0.14 0.16 24
22 0.11 0.13 0.15 0.17 25
23 0.12 0.14 0.16 0.18 26
24 0.13 0.15 0.17 0.19 27
Calculate p.
 34

Solution: Aicha is now age 21 and was selected at age 20. So the probability that White will be alive
5 years from now can be expressed as p = 5 p[20]+1 . We have

5 p[20]+1 = p[20]+1 4 p[20]+2 = p[20]+1 p[20]+2 3 p[20]+3 = p[20]+1 p[20]+2 p[20]+3 2 p[20]+4
Y5 5
Y

= p[20]+1 p[20]+2 p[20]+3 p[20]+4 p[20]+5 = p[20]+k = 1 − q[20]+k
k=1 k=1



= 1 − q[20]+1 1 − q[20]+2 (1 − q23 ) (1 − q24 ) (1 − q25 )
= (1 − 0.11) (1 − 0.13) (1 − 0.15) (1 − 0.16) (1 − 0.17) = 0.45887

Sometimes, you may be given a select–and–ultimate table that contains the life table function ℓ[x] . In
i
hb
this case, you can calculate survival and death probabilities by using the following equations:
da

ℓ[x]+u+t ℓ[x]+u+t ℓ[x]+u − ℓ[x]+u+t

t p[x]+u = and t q[x]+u =1− =
Ed

ℓ[x]+u ℓ[x]+u ℓ[x]+u

ed
m

Exercise 2.3.2 A select–and–ultimate table with a select period of 2–years is given as follows:
ha
M

x ℓ[x] ℓ[x]+1 ℓx+2 x+2

70 22507 22200 21722 72
f.

71 21500 21188 20696 73

o

sa
Pr

72 20443 20126 19624 74

u.

73 19339 19019 18508 75

ed

74 18192 17871 17355 76

u.

I
ks

els
@

od
i

1. Compute 3 q[73] and 3 q73 .

hb

lM
da

2. Compute the probability that a life age 71 dies between ages 75 and 76, given that the life was
ica
selected at age 70.
ed

at
m

3. Assuming UDD between integral ages, calculate 0.5 q[70]+0.7 .

m
he

4. Assuming CFM between integral ages, calculate 0.5 q[70]+0.7 .

at
lM

Solution:
ia
ar

1. We have
tu

ℓ[73]+3 17355 ℓ73+3 17355

3 q[73] =1− =1− = 0.1025906 and 3 q73 =1− =1− = 0.1614322
Ac

ℓ[73] 19339 ℓ73 20696

2. The probability that a life age 71 dies
between ages 75 and 76, given that the life was selected
at age 70 is given by P 4 ≤ T[70]+1 < 5 = 4| q[70]+1 , hence

4| q[70]+1 = 5 q[70]+1− 4 q[70]+1 = 4 p[70]+1 − 5 p[70]+1

ℓ[70]+5 − ℓ[70]+6 ℓ75 − ℓ76 18508 − 17355
= = = = 0.0519
ℓ[70]+1 ℓ[70]+1 22200
 35

3. Assume UDD, we have

1.2 p[70]
0.5 q[70]+0.7 = 1− 0.5 p[70]+0.7 =1−
0.7 p[70]


p[70] × 0.2 p[70]+1 p[70] 1 − 0.2 q[70]+1
= 1− =1−
0.7 p[70] 1 − 0.7 q[70]


p[70] 1 − 0.2 q[70]+1
= 1− (UDD)
1 − 0.7 q[70]
  
ℓ[70]+1 ℓ[70]+2
ℓ[70]
1 − 0.2 1 − ℓ[70]+1
i = 1− 
ℓ[70]+1

1 − 0.7 1 − ℓ[70]
hb
da

22200
1 − 0.2 1 − 21722



22507 22200
= 1− ≃ 0.0084200
Ed

22200


1 − 0.7 1 − 22507
ed
m

4. Assume CFM, we have from 3.

ha

0.2
M

p[70] 0.2 p[70]+1 p[70] p[70]+1

0.5 q[70]+0.7 = 1− =1−

0.7 p[70] p[70] 0.7
f.
o

sa


0.2
= 1 − p[70] 0.3 p[70]+1
Pr

u.

   
ℓ[70]+1 0.3 ℓ[70]+2 0.2
= 1−
ed

ℓ[70] ℓ[70]+1
u.

I
 0.3  0.2
22200 21722
ks

els
= 1− ≃ 0.008437.
22507 22200
@

od
i
hb

lM
da

Exercise 2.3.3 A select–and–ultimate table with a select period of 4–years is given as follows:
ica
ed

at
m

x q[x] q[x]+1 q[x]+2 q[x]+3 qx+4 x+4

m

40 0.00101 0.00175 0.00205 0.00233 0.00257 44

he

41 0.00113 0.00188 0.00220 0.00252 0.00293 45

at

42 0.00127 0.00204 0.00240 0.00280 0.00337 46

lM

43 0.00142 0.00220 0.00262 0.00316 0.00384 47

44 0.00157 0.00240 0.00301 0.00367 0.00445 48
ia
ar
tu

q[x]+k
1. The index of selection x is defined by I(x, k) = 1 − . Calculate the index at age 44 for k = 0,
Ac

qx+k
1, 2, 3.

2. Construct the select life table function of table ℓ[x]+k for x = 40, 41, 42 and k = 0, 1, 2, 3,4
given that ℓ[40] = 10, 000

3. Calculate i) 3 q[41]+1 , ii) 3|2 q[41] .

Solution:
 36

1. We have
q[44] 0.00157
I(44, 0) = 1 − =1− = 0.38911,
q44 0.00257
q[44]+1 0.00240
I(44, 1) = 1 − =1− = 0.18089,
q45 0.00293
and
q[44]+2 0.00301
I(44, 2) = 1 − =1− = 0.10682,
q46 0.00337
q[44]+3 0.00367
i I(44, 3) = 1 − =1− = 0.044271
q47 0.00384
hb
da
Ed

2. The select life table function of table ℓ[x]+k for x = 40, 41, 42 and k = 0, 1, 2, 3,4 is given by the
following table using the formulas
ed

ℓ[40]+k+1 = 1 − q[40]+k ℓ[40]+k and ℓ[41]+4 = ℓ45 = (1 − q44 ) ℓ44 ,
m
ha

moreover
M

ℓ[41]+k+1
ℓ[41]+k = for k = 0, 1, 2, 3 and ℓ[42]+4 = ℓ46 = (1 − q45 ) ℓ45 ,
f.

1 − q[41]+k
o

sa
Pr

u.

ℓ[42]+k+1
and then use again ℓ[42]+k = 1−q[42]+k
for k = 0, 1, 2, 3.
ed
u.

x ℓ[x] ℓ[x]+1 ℓ[x]+2 ℓ[x]+3 ℓx+4 x+4

I
ks

els
−→ −→ −→ −→
@

od
40 10000.00 9989.90 9972.42 9951.97 9928.79 44
i
hb

←− ←− ←− ←−
lM
da

41 9980.20 9968.92 9950.18 9928.29 9903.27 ↓ 45

ica
←− ←− ←− ←−
ed

42 9958.74 9946.09 9925.80 9901.98 9874.25 46

at
m

m
he

3. By the definition in terms of the select life table function we have

at
lM

ℓ[41]+4 9903.27
3 q[41]+1 =1− =1− = 0.006585
ℓ[41]+1 9968.92
ia
ar

and
ℓ[41]+3 − ℓ[41]+5 9928.29 − 9874.25
tu

3|2 q[41] = 3 p[41] − 5 p[41] = = = 0.005415.

ℓ[41] 9980.20
Ac

Exercise 2.3.4 You are given the following select–and–ultimate life table with a two year select period.

x ℓ[x] ℓ[x]+1 ℓx+2

90 t+1
and q[x]+t = qx+t
91 1250 920 3
92 1000 900
Calculate ℓ[90]+1
 37

Solution:
Step 1 By definition
ℓ[90]+2 ℓ92 ℓ92 1 ℓ93 1 920
ℓ[90]+1 = = = 2 = 2 = 2 .
1 − q[90]+1 1 − q[90]+1 1 − 3 q91 1 − 3 q91 1 − q92 1 − 3 q91 1 − q92

We need to backup q91 and q92 .

Step 2 Find q91 . For t = 0, q[91] = 13 q91 thus q91 = 3q[91] .
Step 3 Find q[91]

ℓ[91]+1 1 ℓ[91]+2 1 920

q[91] = 1 − =1− =1− .
ℓ[91] ℓ[91] 1 − q[91]+1 1250 1 − 32 q92
i
hb
da

Step 4 Find q92    

ℓ[92]+1 900
Ed

q92 = 3q[92] =3 1− =3 1− = 0.3.

ℓ[92] 1000
ed

Therefore
m

 
1 920
q91 = 3q[91] =3 1− = 0.24.
ha

1250 1 − 0.6
3
M

Finally
f.

ℓ[90]+2 ℓ92 1 ℓ93 1 920

o

sa

ℓ[90]+1 = = = = × ≃ 1565.
Pr

2 2
1 − q[90]+1 1 − q[90]+1 1 − 3 q91 1 − q92 1 − 3 × 0.24 1 − 0.3
u.
ed
u.

I
Exercise 2.3.5 Select mortality rates for [45] are half of the Illustrative Life Table’s (i = 0.06)
ks

els
mortality rates for a selection period of 3 years. Calculate 2|3 q[45] .
@

od
i
hb

Solution: We know that

lM
da

ica


2|3 q[45] = 2 p[45] 3 q[45]+2 = 2 p[45] 1 − 3 p[45]+2 ,
ed

at
m

so we need 2 p[45] and 3 p[45]+2 . So

m
he

2 p[45] = p[45] p[45]+1 = 1 − q[45] 1 − q[45]+1
at
lM

and



3 p[45]+2 = p[45]+2 p[45]+3 p[45]+4 = 1 − q[45]+2 1 − q[45]+3 1 − q[45]+4 .
ia
ar

For q[45] and q[45]+1 we use half of the ILT rates. Then
tu

2 p[45] = (1 − 0.5q45 ) (1 − 0.5q46 ) = (1 − 0.5(0.004))(1 − 0.5(0.00431) = 0.995849,

Ac

moreover q[45]+2 = 0.5q47 but q[45]+3 = q48 and q[45]+4 = q49 since the selection period ends after 3 years.
Mortality for duration 3 and so on is not different from standard mortality.

3 p[45]+2 = (1 − 0.5(0.00466))(1 − 0.00504)(1 − 0.00546) = 0.98722

The answer is


2|3 q[45] = 2 p[45] 3 q[45]+2 = 2 p[45] 1 − 3 p[45]+2 = 0.995849(1 − 0.98722) = 0.012727.
 38

2.4 Moments of future lifetime random variables

First, let us focus on the moments of the future lifetime random variable Tx . We call E [Tx ] the
complete expectation of life at age x, and denote it by e̊x . We have
Z ∞ Z ∞
E [Tx ] =: e̊x = tfx (t) dt = − tSx′ (t) dt.
0 0

Using integration by parts and the fact that limt−→∞ tSx (t) = 0, we can show that
Z ∞
i e̊x = t px dt. (2.4.2)
0
hb
Note that if there is a limiting age, we replace ∞ with ω − x: that
da
Ed

Z ω−x
e̊x = t px dt.
ed

0
m

The second moment of Tx can be expressed as

ha

Z ∞ Z ∞
M

t2 Sx′ (t) dt
 2 2
E Tx = t fx (t) dt = −
0 0
f.
o

sa

Using integration by parts and the fact that limt−→∞ t2 Sx (t) = 0, we can show that the above formula
Pr

u.

can be rewritten as Z ∞
ed

 2
E Tx = 2t t px dt
u.

I
0
ks

els
which is generally easier to apply. Again, if there is a limiting age, we replace ∞ with ω − x: that is
@

od
i

Z ω−x
hb

 2
E Tx = 2t t px dt
lM
da

0
ica
ed

In the exam, you may also be asked to calculate E [Tx ∧ n] = E[min(Tx ; n)]. This expectation is known
at
m

as the n–year temporary complete expectation of life at age x, and is denoted by e̊x:n . It can be shown
m

that Z n
he

e̊x:n = t px dt.
at

0
lM

Exercise 2.4.1 You are given µx = 0.01 for all x ≥ 0. Calculate the following: a.
ia

e̊x:n , b.
ar

Var (Tx ∧ n) .
tu

We have:
Ac

Z n Z n
1 n

e−0.01t dt = 1 − e−0.01n = 100 1 − e− 100


e̊x:n = t px dt =
0 0 0.01
and Z n Z n
E (Tx ∧ n)2 = 2te−0.01t dt = 20000 1 − e−0.01n − 200ne−0.01n
 

2t t px dt =
0 0
thus
n
2
Var (Tx ∧ n) = 20000 1 − e−0.01n − 200ne−0.01n − 1002 1 − e− 100 .


 39

2.4.1 Uniform distribution of death for all ages

Recall that the uniform distribution on [0, κ] has a mean of κ2 , which is its median and midrange. Its
2
variance is κ12 . It is traditional to use the letter ω to indicate the upper limit of a mortality table. For
a uniform model, age at death T0 is uniformly distributed on (0, ω).
While the uniform distribution has memory–after all, you can not live beyond ω, so the older one
is, the less time until certain death–It has the following property that simplifies calculation: if age at
death is uniform on (0, ω], then survival time for (x) is uniform on (0, ω − x).
Here are the probability functions for Tx if T0 is uniform on (0, ω): After working out several
problems, you should recognize the uniform distribution on sight.
For the uniform distribution of death (UDD) over (0, ω − x) or De Moivre’s law ℓx = ω − x for
i
hb
0 < x < ω. Ee have: The density of the Tx given
da

( 1
if 0 < t < ω − x
Ed

fx (t) = ω−x
0 if t ≥ ω − x.
ed
m

hence the c.d.f. ifs given by

t
ha

Fx (t) = t qx = .
ω−x
M

Therefore
f.

ω−x−t t
=1−
o

Sx (t) = t px = .
sa

ω−x ω−x
Pr

u.

The force of morlatity under UDD is

ed
u.

I
fx (t) ω−x 1 1
ks

µx+t = = = , that is µx = .

els
ω−x−t
Sx (t) ω−x−t ω−x
@

ω−x

od
i
hb

The expectation and the variance are obviously given as lM

da

ica
ed

ω−x (ω − x)2
E [Tx ] = and Var(Tx ) = .
at
m

2 12
m

Assume that UDD holds, we have Tx = Kx + U , where U follows a uniform distribution over the
he

interval [0, 1]. Taking expectation on both sides, we have the following relation: e̊x = ex + 12 .
at
lM

2.4.2 Constant force of mortality or Exponential distribution for all ages

ia
ar

If the force of mortality is the constant µ, the distribution of survival time is exponential. Survival
tu

probabilities are then independent of age; t px does not depend on x. The constant force of mortality
Ac

µ is the reciprocal of mean survival time; in other words, a life with constant force of mortality µ has
expected future lifetime µ1 , regardless of the life’s current age.
Here are the probability functions for Tx if Tx has constant force of mortality µ. Notice that none
of the functions vary with age x

∀ x, ∀ t, µx+t = µ, Sx (t) = t px = e−µt , Fx (t) = t qx = 1 − e−µt , fx (t) = µe−µt ,

consequently
1 1
E [Tx ] = and Var(Tx ) = 2 .
µ µ
 40

1. Recursive formulas
e̊x = e̊x:n + n px e̊x+n , e̊x:n = e̊x:m + m px e̊x+m:n−m , for m < n
ex = ex:n + n px ex+n = ex:n−1 + n px (1 + ex+n )
ex = px + px ex+1 = px (1 + ex+1 )
ex:n = ex:m + m px ex+m:n−m = ex:m−1 +
m px (1 + ex+m:n−m ), for m < n
ex:n = px + px ex−1:n−1 = px 1 + ex−1:n−1 .

For any integral age x, n ≥ 1, 0 ≤ r < 1 and 0 ≤ r + t < 1

bi
Function Uniform distribution of deaths Constant force of mortality
h
ℓx+r ℓx − rdx ℓx prx
da

r qx rqx 1 − prx
Ed

r px 1 − rqx prx
rqx
r qx+t 1 − prx
ed

1−tqx
qx
µx+r − ln (px )
m

1−rqx
r px µx+r qx − r px ln (px )
ha

e̊x:n ex:n + 0.5 n qx

M

e̊x:1 px + 0.5qx
.

e̊x ex + 0.5
of

.sa
Pr

u
ed

2.5 Sample exam questions

u.

I
els
ks

Exercise 2.5.1 You are given that µx = 0.001x. Calculate e35.3:3

@

od
bi

Solution: By definition we have e35.3:3 = 3k=1 k p35.3 . So, we shall first calculate the three survival
lM
P
ah

probabilities that we need.

ica
d
ed

 Z t 
µx+s ds = exp −0.0005 (x + t)2 − x2 = exp (−0.001xt) exp −0.0005t2




at

t px = exp −
m

0
m
he

thence
at

p35.3 = e−0.0005((36.3) −(35.3) ) = 0.964833

2 2
lM

−0.0005((37.3)2 −(35.3)2 )
2 p35.3 = e = 0.929973
ia

−0.0005((38.3)2 −(35.3)2 )
p
3 35.3 = e = 0.895476
r
ua

Finally we get
t
Ac

e35.3:3 = 0.964833 + 0.929973 + 0.895476 = 2.790282.

If future lifetime is exponential, the sum t px is a geometric series.
Exercise 2.5.2 The force of mortality for (x) is the constant µ = 0.01 Calculate the curtate life
expectancy of (x).
Solution: The survival probability is k px = e−0.01k . Then
∞
X e−0.01
ex = e−0.01k = = 99.5008
k=1
1 − e−0.01
 41

Exercise 2.5.3 For (30), calculate e̊30 given


0.01 0 < t ≤ 10
µ30+t =
0.02 t > 10.

Solution: Using formula (2.4.2),

Z ∞ Z 10 Z ∞
e̊30 = t p30 dt = t p30 dt + t p30 dt
0 0 10
Z 10 Z ∞
−0.01t
= e dt + t p30 dt
i
hb
0 10
da

For t > 10
Ed

 Z t 
t p30 = exp − µ30+u du
0
ed

Observe that
m

Z t Z 10 Z t
ha

µ30+u du = µ30+u du + µ30+u du = 0.1 + 0.02(t − 10).

0 0 10
M

therefore t p30 = e−0.1−0.02(t−10) = e0.1−0.02t finally

f.
o

sa
Z 10 Z ∞
Pr

−0.01t
e0.1−0.02t dt = 100 1 − e−0.1 + 50e0.1 e−0.2



e̊30 = e dt +
u.

0 10
ed

−0.1
= 100 − 50e = 54.75813.
u.

I
ks

els
One ca use also the fornula e̊x = e̊x:n + n px e̊x+n
for x = 30 and n = 10, hence
@

od
Z 10 Z ∞
i
hb

e̊30 = e̊30:10 + 10 p30 e̊40 = t p30 dt + 10 p30 t p40 dt

lM
da

0 0
Z 10 Z ∞
ica
ed

−0.01t −0.01×10
= e dt + e t p40 dt,
at
m

0 0
m

Rt
Recall that t p40 = e− 0 µ40+u du
= e−0.02t , thus
he
at

Z 10 Z ∞
−0.01t −0.01×10
e̊30 = e dt + e e−0.02t dt = 54.75813
lM

0 0
ia

0.5
Exercise 2.5.4 Future lifetime of (20) is subject to force of mortality µx = , x < 100. Calculate
ar

100−x
e̊20
tu
Ac

Solution: After computations we get we have ω = 100, x = 20 and by definition of e̊20

Z 80
e̊20 = E [T20 ] = t p20 dt,
0

where
 Z t t
1 t 1
  Z  Z  
0.5
t p20 = exp − µ20+s ds = exp − = exp − ds ds
0 0 100 − (20 + s)
2 0 80 − s
     1
1 t 1 t 2
= exp [ln (80 − s)]0 = exp (ln (80 − t) − ln (80)) = 1 − .
2 2 80
 42

therefore 1
Z 80 Z 80 
t 2 160
t px dt = 1− dt = = 53.333333.
0 0 80 3
α t

α
In general if µx = ω−x
, x < ω, this means the survival has the from t px = 1 − ω−x
, hence

ω−x ω−x  α "  α+1 #ω−x

(ω − x)
Z Z
t t
e̊x = t px dt = 1− dt = − 1−
0 0 ω−x α+1 ω−x
0
"   #0α+1
(ω − x) t ω−x
= 1− = .
α+1
i ω−x α+1
ω−x
hb
da

Exercise 2.5.5 Ahmed and Aicha are age 30. (i) Ahmed’s future lifetime is uniformly distributed
Ed

with ω = 100. (ii) The force of mortality for Aicha is

ed

α
m

µ30+t =
70 − t
ha
M

(iii) Aicha’s survival probability to age 80 is twice as high as Ahmed’s. Determine Aicha’s expected
(Aicha)
future lifetime (e̊30 ).
f.
o

sa
Pr

Solution: Probability Ahmed survives to 80 is

u.
ed

ω − 80 100 − 80 2
50 p30
= =
=
u.

ω−x 100 − 30 7

I
ks

els
α
Aicha’s survival is beta, so t px for her is ω−x−t


, and here ω = 100, x = 30, and t = 50, so from (iii),
@

od
ω−x
i
hb

α
lM
  
70 − 50 2
=2
da

70 7
ica
ed

ln( 4 )
which gives α ln( 27 ) = ln( 47 ) hence α = ln( 72 ) = −0.559615 = 0.44670 . Then since Aicha’s survival is
at
m

7
−1.252762
m

beta, the expected value is

ω − 30 70
he

e̊30 = = = 48.38598.
α+1 1.4467
at
lM

1
Exercise 2.5.6 Future lifetime of (20) is subject to force of mortality µx = , x < 100
ia

100−x
ar

1. Calculate e20 2. Calculate e20:50 3. Calculate Var (K20 )

tu

Solution: 1. For uniform mortality,

Ac

ω − 20 100 − 20
e̊20 = = = 40.
2 2
Since e20 = e̊20 − 0.5, thus e20 = 39.5.
2. Computations give
Z 50 Z 50  
t 1  2 50
e̊20:50 = t p20 dt = 1− dt = 50 − t 0
0 0 80 160
2
50 275
= 50 − = = 34.375.
160 8
 43

Then
1 15
e20:50 = e̊20:50 − 50 q20 = 34.375 − = 34.0625
2 28
3. Now, we calculate the second moment.
79
X 79
X 79
X  
 2 2 k+1 k
E K20 = k k| q20 = k2 ( k+1 q20 − k q20 ) = k 2
−
k=0 k=0 k=1
100 − 20 100 − 20
79  
1 X 2 1 (79)(80)(159)
= k = = 2093.5
80 k=1 80 6
i
hb
because n
X n(n + 1)(2n + 1)
da

k2 = .
6
Ed

k=1

The variance is Var (K20 ) = 2093.5 − (39.5)2 = 533.25

ed
m

Exercise 2.5.7 You are given: (i) Deaths are uniformly distributed over each year of age.
ha

(ii) e̊55.2:0.4 = 0.396. Calculate µ55.2 .

M

q55
Solution: We have µ55.2 = and
f.

1−0.2q55
o

sa
Pr

Z 0.4 Z 0.4 Z 0.4

0.2+t p55 1
u.

e̊55.2:0.4 = t p55.2 dt = dt = (1 − 0.2+t q55 ) dt

0.2 p55 0.2 p55
ed

0 0 0
Z 0.4
u.

I
= (1 − (0.2 + t) q55 ) dt
ks

els
0.2 p55 0
@

 Z 0.4 
1

od
i

= 0.4 − q55 (0.2 + t) dt

hb

1 − 0.2q55 0
lM
da

0.4 − q55 × 0.16

= = 0.396.
ica
ed

1 − 0.2q55
at
m

0.04950
Therefore q55 = 0.04950. Thus µ55.2 = = 0.049995.
m

1−0.2×0.04950
he

Exercise 2.5.8
at
lM

1. A survival function is given by

ia

(10 − t)2
ar

S0 (t) = for 0 ≤ t < 10,

100
tu
Ac

Calculate the difference between the force of mortality at age 1, and the probability that a life age
(1) dies before age 2.
t 0.4
2. For life aged 60, the survival probability is t p60 = (1 − ) . Calculate µ80 .
60
3. Assume that the future lifetime T20 is subject to force of mortality
1
µx = for 0 ≤ x < 100.
100 − x
a. Calculate e20 , b. Calculate e20:50 , c. Calculate Var(K20 ).
 44

Hint
n
1 1 X n(n + 1)(2n + 1)
e̊x:n = ex:n + n qx , e̊x = ex + and k2 = .
2 2 k=1
6

Solution:
−Sx′ (t) S0 (x+t)
1. Recall that µx+t = Sx (t)
where Sx (t) = S0 (x)
then

−S0′ (x + t) 2 (10 − (x + t)) 2

µx+t = = 2 =
i S0 (x + t) (10 − (x + t)) 10 − x − t
hb
therefore
da

−S0′ (1) 2 2
µ1 = = = = 0.22222
Ed

S0 (1) 10 − 1 9
and the probability that a life age (1) dies before age 2 is given by
ed
m

S0 (2) 82
ha

P (0 ≤ T1 ≤ 1) = q1 = 1 − p1 = 1 − =1− 2
S0 (1) 9
M

So µ1 − q1 = 0.22222 − 0.20988 = 0.01234.

f.
o

sa
0.4
Pr


t
u.

2. Given S60 (t) = 1 − , we have

60
ed
u.

I
−S60 (t) 0.4 60 0.4
ks

µ60+t = = = ,

els
S60 (t) 60 60 − t 60 − t
@

od
i
hb

then
µ80 = µ60+20 =
0.4
=
0.4
= 0.01. lM
da

60 − 20 60 − 20
ica
ed

at
m

3. a. We know e20 = e̊x − 0.5,

m
he

Z 100−20 R 20+t − du
100−u = e[ln(80−t)−ln(80)] = 1 −
t
at

e̊20 = t p20 dt and t p20 = e

20

0 80
lM

therefore
ia

Z 80  
t
e̊20 = 1− dt = 40, thus e20 = 40 − 0.5 = 39.5.
ar

0 80
tu

b. and
Ac

Z 50 Z 50   Z 50  
t t 275
e̊20:50 = t p20 dt = 1− dt = 1− dt = = 34.375
0 0 80 0 80 8
Therefore using the formula
1 1
e̊20:50 = e20:50 + 50 q20 = e20:50 + (1 − 50 p20 )
2 2
we get
1 1 50
e20:50 = e̊20:50 − (1 − 50 p20 ) = 34.375 − = 34.063
2 2 80
 45

2
c. By definition Var(K20 ) = E [K20 ] − (E [K20 ])2 and
80 80   80  
X X k 1 X 1 1 + 80 79
E [K20 ] = k p20 = 1− = 80 − k = 80 − 80 = = 39.5
k=1 k=1
80 80 k=1 80 2 2

and
80 80  80 80
2
X X k 1 X X
2k 2 − k
 

E K20 = (2k − 1) k p20 = (2k − 1) 1 −
= (2k − 1) −
k=1 k=1 k=1
80 80 k=1
   
1 + 80 2 80(80 + 1)(160 + 1) 1 1 + 80 4187
= 2 i 80 − 80 − + 80 = = 2093.5
2 80 6 80 2 2
hb
da

Finally Var(K20 ) = 2093.5 − (39.5)2 = 533.25.

Ed

Exercise 2.5.9
ed
m

x
1. The force of mortality of a life is µx = 1000
. Calculate e35.5:3
ha

2. The force of mortality for life (x) is the constant µx = 0.01. Calculate the curtate life
M

expectancy of (x).
f.
o

sa

Exercise 2.5.10 For a select–and–ultimate mortality table with a one–year select period, you are
Pr

u.

given:
ed
u.

x ℓ[x] d[x] = q[x] ℓ[x] e̊[x]

I
ks

85 1000 100 5.225

els
@

86 850 100

od
i
hb

Assume deaths are uniformly distributed over each year of age.

lM
da

ica
ed

1. Calculate p[85] , p[86] and p86 .

at
m

R2
2. Use the decomposition t p[85] = 1+t−1 p[85] to calculate 1 t p[85] dt
m
he

R∞ R∞
3. Calculate e̊[85] − 2 t p[85] dt and deduce the value of 2 t p[85] dt.
at
lM

4. Calculate e̊[86]
ia

Solution:
ar
tu

1. We have
Ac

d[85] 100 d[86] 100 15

p[85] = 1 − =1− = 0.9, p[86] = 1 − =1− = = 0.88235,
ℓ[85] 1000 ℓ[86] 850 17

and
ℓ87 ℓ[86]+1 ℓ[86] − d[86]
p86 = = =
ℓ86 ℓ[85]+1 ℓ[85] − d[85]
850 − 100 750 5
= = = = 0.83333.
1000 − 100 900 6
 46
R2
2. We want to calculate 1 t p[85] dt.
Z 2 Z 2 Z 2 Z 2
t p[85] dt = 1+t−1 p[85] dt = p[85] t−1 p[85]+1 dt = p[85] t−1 p86 dt.
1 1 1 1

Now, we use the change of variable t − 1 = r, so when t = 1, r = 0 and when t = 2, r = 1, and

dt = dr, thus
Z 2 Z 1
t p[85] dt = p[85] r p86 dr
1 0
i Z 1 Z 1
= p[85] r p[85]+1 dr = p[85] (1 − r q86 ) dr
hb
0 0
da

Z 1  Z 1 
UDD
= p[85] (1 − rq86 ) dr = p[85] 1 − q86 r dr
Ed

0
    0
1 1 1
ed

= p[85] 1 − q86 = p[85] 1 − (1 − p86 ) = p[85] (1 + p86 )

2 2 2
m

 
1 5
ha

= 0.9 1 + = 0.825.
2 6
M
f.
o

sa
Pr

3. We know that
u.

∞ ∞ ∞
ed

Z Z Z
e̊[85] − t p[85] dt = t p[85] dt − t p[85] dt
u.

I
2
Z0 2 2
ks

Z 1 2

els
Z
= t p[85] dt = t p[85] dt + t p[85] dt.
@

od
0 0 1
i
hb

And lM
da

ica
1 1 1
ed

Z Z Z



t p[85] dt = 1 − t q[85] dt = 1 − tq[85] dt
at
m

0
Z0 1 0
m

Z 1



= 1 − t 1 − p[85] dt = (1 − t (1 − 0.9)) dt = 0.95
he

0 0
at

R∞
lM

Therefore e̊[85] − 2 t p[85] dt = 0.95 + 0.825 = 1.775. We deduce that

ia

Z ∞
t p[85] dt = e̊[85] − 1.775 = 5.225 − 1.775 = 3.45
ar

2
tu
Ac
 47

4. We have
Z 1 Z ∞ Z 1

Z ∞
e̊[86] = t p[86] dt + t p[86] dt = 1 − tq[86] dt + p[86] t−1 p[86]+1 dt
0 1 0 1
Z 1 Z ∞


= 1 − tq[86] dt + p[86] t−1 p[86]+1 dt
0 1
Z 1 Z ∞



= 1 − t 1 − p[86] dt + p[86] t p87 dt
0 0
Z 1   Z ∞
15
= 1−t 1− dt + p[86] t p87 dt
0
i 17 0
Z ∞
16 15 1 t+2 p[85]
hb
= + t+2 p[85] dt (t p87 = t p[85]+2 = )
17 17 2 p[85] 0 2 p[85]
da

Z ∞
16 15 1 16 15 3.45
Ed

= + t p[85] dt = + 9 5 = 5.
17 17 p[85] p86 2 17 17 10 6
ed
m
ha
M
f.
o

sa
Pr

u.
ed
u.

I
ks

els
@

od
i
hb

lM
da

ica
ed

at
m

m
he
at
lM
ia
ar
tu
Ac