1

Actuarial Mathematics I
SAS 2191: ACTUARIAL MATHEMATICS I
Pre-requisite: Financial Mathematics I

Course Purpose
The aim of this course is to provide grounding in the mathematical techniques which are of particular
relevance to actuarial work in life insurance, health and care and pensions.

Expected Learning Outcomes:

Objectives: At the end of the course, the learner should be able to:

1. Define, and use life table functions involving single lives to understand basic actuarial problems.

2. To solve actuarial problems by applying actuarial mathematics in life contingencies, and the
ability to apply the concepts of actuarial science in solving problems.

Course Description
Survival models and the life table including select mortality functions. Population theory: complete
and curtate expectation of life, central death rates, and stationary population including average ages.
Formulae for means and variances of assurances and annuities; commutation functions; premiums; net
premiums provisions.

Method of delivery: Lectures, interactive tutorials, and any other presentations/ demonstrations the
lecturer will deem fit towards enhancing understanding of the concepts taught in class. Special attention
will be given in the implementation of the methods taught using specialized software packages.

Instruction Materials/Equipment: Whiteboard and Handouts.

Course Evaluation: The final grade for the course will be based on a final examination at the end
of the semester (70%) and the Continuous Assessment Tests (CATs) (30%) which will be based on
assignments throughout the semester and at least Two sit-in CATs.

Prerequisites: Fundamentals of Actuarial Science

Core Text Books

1. CT5 Study material

2. W.F. Scott, Life Assurance Mathematics, Heriot-Watt University, 1999.

3. Neill, A. Life Contingencies, Heinemann 1977.

Other Textbooks

1. N.L. Bowers, H.U. Gerber, J.C. Hickman et al. Actuarial mathematics. 2nd edition. Society of
Actuaries, 1997. 753
2

2. H.U. Gerber Life Insurance Mathematics. 3rd edition. Springer; Swiss Association of Actuaries,
1997.

3. Batten & London. Life Contingencies. A Logical Approach to Actuarial Mathematics. 2005.

4. Fundamentals of Actuarial Mathematics, S.D. Promislow, ©2006 John Wiley and Sons, Ltd.
 LECTURE 1
The Life Tables
1.1 Introduction
The subject of Actuarial mathematics is an application of mathematics dealing with calculations in
respect of payments depending on human survivorship or death. Such calculations, which usually
include the elements of compound interest as well as mortality are required in connection with Life
assurance and pension funds.

Although the mortality experience of a group of persons in the future is not known, problems in life
contingencies can be solved using or by making assumptions on future mortality experiences and
numerical results are obtained with the aid of tables of functions based on that mortality.

1.2 The Mortality Table

Mortality tables are usually based directly or indirectly on the mortality which has actually been
experienced by a suitable group of persons.

If a large number nα of persons attaining the birthday α during a short period of time could be kept
under observation from that age until all of them are dead, the number of those surviving to each
birthday x = α + 1, x = α + 2, x = α + 3 and so on, would form a non-increasing sequence nx and
ultimately at some value of x equal to say ω, we should have nω = 0 when they have all died.

The number of persons who died between exact age x and exact age x + 1 would be

nx − nx+1 i.e. − ∆nx

∆nx
The ratio would be the proportion of persons dying between exact age x and exact age x + 1, out
nx
of a given number, (in this case nx ) attaining age x. This proportion can be used to measure the rate of
mortality at age x.

The fundamental function of the mortality table is a function known as lx . It is a positive non-increasing
function representing the number of lives who are expected to survive to age x out of lα lives who
attain age α, this being the youngest age for which lx is tabulated. For theoretical purposes it is
convenient to assume that lx is a continuous function, having values for all x and not just integers and
with unique finite differential coefficients to any required order at all points.

−∆lx /lx is the rate of mortality at age x assumed in the mortality table.

The mortality table can start from any age, thus

lx+t - represents the number of lives expected to survive to age x + t out of lx lives aged x.
lx+t
- is the probability of survival to age x + t in respect of an individual life aged x taken at random.
lx

3
4 LECTURE 1. THE LIFE TABLES

lx - is the fundamental function of the mortality table.

Normally the lx figure for the lowest age is taken as some convenient round number such as 1,000,000,
999,999, 500,000. This is called the radix of the table.

The first age at which lx becomes negligible is called the limiting age and is denoted by ω.

lω = 0

ω = is the youngest age to which no one survives. Theoretically ω = ∞.

Age lx dx 1000qx
0 1,000,000 1580 1.58
Life Table 1 998,420 680 68
2 997,740 485 0.49
3 997,255 435 0.44
In Table 1.2 we have presented an excerpt from a typical life table. In such a table the column lx denotes
the number of lives which have survived to age x. For this to make sense, we have to assume a starting
population l0 . In our case l0 = 1, 000, 000, but any value would have sufficed. As we shall explain,
lx
it is the ratio of entries in the table, not the individual numbers, which is important. In particular
l0
represents the probability of surviving from birth to age x. As a general function of x, it is called the
survival function and is denoted S(x).

Note that in Table 1.2, l1 = 998, 420. This means that 1580 lives have died in the first year of life,
and this is the entry d0 . In general, dx denotes the number of lives, out of those aged x, which do not
survive to age x + 1. Thus,

dx = lx − lx+1 . (1.1)

qx denotes the probability that a life aged x will not survive to age x + 1. Thus,
dx
qx = . (1.2)
lx
0.49 485
In our case, the final column tells us that q2 = = 0.00049, and = 0.00049 as well.
1000 997, 740
Using a life table we can compute numerous probabilities concerning survival.

1.2.1 Probabilities of living and dying

We introduce the important notations:

t px = P{(x) survives to age x + t}

Note: (x) is a shorthand notation for “a life aged (exactly) x”.

t qx = P{(x) dies between ages x and x + t}

When t = 1, it may be omitted, so that

px = P{(x) survives for at least a year}

qx = P{(x) dies within a year}
1.2. THE MORTALITY TABLE 5

qx is called the (q-type) rate of mortality of x. (The m-type, or central rate of mortality, mx , is considered
later).

It is important to note that, for all t ≥ 0.

t px + t qx =1

t px can be expressed in terms oflx+t and lx as follows;

lx+t
t px =
lx
and
lx+t lx − lx+t
t qx = 1 − t px = 1 − =
lx lx
Note that, when t = 1;
lx+1
px =
lx
lx+1 lx − lx+1 dx
qx = 1 − px = 1 − = =
lx lx lx
where;

dx = lx − lx+1 or dx = −∆lx
= the (expected) number of deaths at age x last birthday among lx lives aged x.

EXAMPLE 1.1
Consider the extract of the mortality table below:
x lx dx
40 80,935 455
41 80,480 481
42 79,999 511
43 79,488 546
44 78,942 585
45 78,357 625
.. .. ..
. . .
50 74,794 844
Probabilities of death and survival can be obtained directly from the columns of lx and dx .

Calculate the following probabilities (i) p40 , (ii) q40 .

Solution.
l40+1 l41 80, 480
p40 = = = = 0.9944
l40 l40 80, 935
d40 455
q40 = = = 0.0056
l40 80, 935
Alternatively;
q40 = 1 − p40 = 1 − 0.9944 = 0.0056
l40+5 l45 78, 357
5 p40 = = = = 0.9681
l40 l40 80, 935
6 LECTURE 1. THE LIFE TABLES

t px can also be looked at as a product of successive probabilities

lx+1 lx+2 lx+3 lx+t
t px = · · ··· ·
lx lx lx lx+t−1
= px · px+1 · px+2 · px+3 · · · px+t−1

t qx = Probability that (x) will die within t years

lx − lx+t 1 x+n−1
t qx =
lx
= ∑ dy = 1 − n pn
lx y=x

so
80, 935 − 78, 357
5 q40 = = 0.0319
80, 935
or
455 + 481 + 511 + 546 + 585
5 q40 = = 0.0319
80, 935
or
5 q40 = 1 − 5 p40 = 1 − 0.9681 = 0.0319
Note:
px = 1 px and qx = 1 qx

Question 1.1. Use Table 1.1 to find the following:

(a) The probability that a newborn will live to age 3.

(b) The probability that a newborn will die between age 1 and age 3.

Question 1.2. Find an expression for each of the following:

(a) The probability that an 18-year-old lives to age 65.

(b) The probability that a 25-year-old dies between ages 40 and 45.

(c) The probability that a 25-year-old does not die between ages 40 and 45.

(d) The probability that a 30-year-old dies before age 60.

Question 1.3. A scientist studies the mortality patterns of Golden-Winged Warblers. She establishes
the following probabilities of deaths: q0 = 0.40, q1 = 0.20, q2 = 0.30, q3 = 0.70 and q4 = 1. Starting
with l0 = 100, construct a life table.
x


Question 1.4. Given lx = 1000 1 − 105 , determine each of the following:

(a) l0 (b) l35

(c) q20 (d) 15 p35 (e) 15 q25 .
√
Question 1.5. Let 100 − x.

(a) Calculate 28 p36

(b) Henry and Ritta are both 19 years old. Find the probability that Henry lives at least 17 years,
Ritta lives at most 45 years, and at least one of them survives for 32 years.
1.2. THE MORTALITY TABLE 7

1.2.2 Deferred Probabilities

The symbol m | indicates deferment for m years; for example,

m |n qx = P{(x) will die between ages x + m and x + m + n}

By elementary probability, this equals

= P{(x) will die before age x + m + n} − P{(x) will die before age x + m}
= m+n qx − m qx
= m px − m+n px
lx+m − lx+m+n
=
lx
That is;
lx+m − lx+m+n lx+m lx+m − lx+m+n
m|n qx = = ·
lx lx lx+m
= m px · n qx+m

This may be remembered by the following rule: of lx lives aged x, lx+m − lx+m+n is the (expected)
number of deaths aged between ages x + m and x + m + n.

If n = 1 it may be omitted, so we have;

m |qx = P{(x) will die between the ages x + m and x + m + 1}

dx+m
=
lx
= m px · qx+m

EXAMPLE 1.2
Calculate the following probabilities
l40+3 − l40+2+3 79, 488 − 78, 357
3|2 q40 = =
l40 80, 935
1, 131
= = 0.0140
80, 935

d44 585
2| q42 = = = 0.0073
l42 79, 999

Note: Since
dx
qx = ⇒ d x = q x · lx
lx
dx = lx − lx+1 ⇒ lx = dx + lx+1
lx+1
px = ⇒ lx+1 = lx · px
lx
Therefore given one of the functions, the other ones can be easily obtained.
8 LECTURE 1. THE LIFE TABLES

EXAMPLE 1.3
Given the mortality table below, answer the following.
Find the probability of a person aged 30;

(i). surviving to age 40

(ii). dying before age 40

(iii). dying between the ages of 60 and 80.

Age x lx
0 10,000,000
1 9,929,000
2 9,911,725
3 9,896,659
.. ..
. .
10 9,805,870
.. ..
. .
20 9,664,994
.. ..
. .
30 9,480,358
.. ..
. .
40 9,241,359
.. ..
. .
50 8,762,306
.. ..
. .
60 7,698,698
.. ..
. .
70 5,592,012
.. ..
. .
80 2,626,372
.. ..
. .
90 468,174
.. ..
. .
99 6,415
100 0

For this table the lx lowest age i.e., Radix = 10, 000, 000, i.e., l0 = 10, 000, 000.
1.3. FORCE OF MORTALITY 9

Solution.
l40 9, 241, 359
10 p30 = = = 0.9748
l30 9, 480, 358

l30 − l40 238, 999

10 q30 = = = 0.0252
l30 9, 480, 358
or
10 q30 = 1 −10 p30 = 1 − 0.9748 = 0.0252

l60 − l80 5, 072, 362

30|20 q30 = = = 0.5350
l30 9, 480, 358

EXAMPLE 1.4
If a mortality table is represented by the function
√
lx = 1000 100 − x

Find

(i). the probability of a life surviving from birth to age 19.

(ii). the probability of a life aged 36 dying before age 61.

Solution.
√
l19 1000 100 − 19 9
19 p0 = = √ = = 0.9
l0 1000 100 10

√ √
l36 − l51 1000 64 − 1000 49 1000 1
15 p36 = = √ = = = 0.125
l36 1000 64 8000 8

1.3 Force of Mortality

The force of mortality denoted by µx represents the instantaneous rate of mortality at a certain age
measured on an annualized basis. This may be defined as the “instantaneous rate of mortality”, i.e.,
h qx
µx = lim
h→0+ h
Since lx is a continuous function, it can be differentiated and the ratio that the rate of decrease of lx at
age x bears to the value of lx at that age is called the force of mortality at age x and is represented by
the symbol µx .
 
h qx lx+h − lx
µx = lim = lim −
h→0+ h h→0+ hlx
1 d(lx ) l ′
=− =− x
lx dx lx
10 LECTURE 1. THE LIFE TABLES

Recall,  
d lx+h − lx lx+h − lx
(lx ) = lx′ = lim and h qx = −
dx h→0+ h lx
Therefore;
 
h qx lx+h − lx
µx = lim = lim −
h→0+ h h→0+ hlx

lx′
=−
lx
l′
Since µx = − x , we can write it as
lx
d d l′
µx = − log lx since loge lx = x
dx e dx lx
Since h qx = lx − lx+h , this is then written as −(lx+h − lx ) hence the negative sign in the formula.
d
∴ µx = − log lx
dx e
µx is also given as
1 dlx d
µx = − = − loge lx
lx dx dx
This is a measure of mortality at the precise moment of attaining age x expressed in the form of an
annual rate.

The negative sign makes µx positive since lx is a decreasing function.

1.3.1 Expressing lx in terms of µx

1 dlx
µx = −
lx dx
Rewriting the above formula
dlx
= −lx µx
dx
and integrating over the range α to the limiting age of the table, x being written as α + t.
dlα+t
Z ω−α Z ω−α
dt = − lα+t µα+t dt
0 dt 0

Z ω−α
lα+t ]ω−α
0 =− lα+t µα+t dt
0

Z ω−α
∴ l ω − lα = − lα+t µα+t dt
0

Since lω = 0.

Therefore; Z ω−α
lα = lα+t µα+t dt
0
1.3. FORCE OF MORTALITY 11

Since lx = 0 for all x > ω, ∞ is often substituted for ω − α in such integrals, and writing x for α we
obtain, Z ∞
lx = lx+t µx+t dt
0
If the limits of integration above had been taken as 0 and 1, we would have obtained
Z 1
[lα+t ]10 = − lα+t µα+t dt
0
Z 1
∴ lα+1 − lα = − lα+t µα+t dt
0

and lα+1 − lα = −dα ,

Therefore; writing x for α, we have;

Z 1
dx = lx+t µx+t dt
0

1.3.2 Expressing px in terms of µx .

Recall, from the above equation
d
µx = −
log lx
dx e
Integrating from age α to α + n, again writing α + t for x, we obtain.
Z n Z n
d
µα+t dt = − loge lα+t dt
0 dt0
= − [loge lα+t ]n0
= − loge lα+n − loge lx
lx+n
= − loge
lα

lx+n n Z
i.e., loge =− µα+t dt
lα 0
lx+n Rn
∴ = e− 0 µα+t dt
lα
 Zn 
= exp − µα+t dt
0

and writing x instead of α Rn

n px = e− 0 µx+t dt

and Rx
lx = l0 e− 0 µt dt

since
l0+n Rn
= e− 0 µt dt
l0
Therefore; Rx
lx = l0 e− 0 µt dt
Rn
The formula for n px = e− 0 µα+t dt bears an obvious and important similarity to the compound interest
formula:
12 LECTURE 1. THE LIFE TABLES
Rn
Present value of 1 due n bears PV = e− 0 δt dt , where δt is the varying force of interest; thus the force
of mortality is analogous to the force of interest.

The probability n qx can then be expressed as

Rn
n qx = 1 − e− 0 µx+t dt

since  Zn 
n px = exp − µα+t dt
0
An alternative formula can be derived by integrating the formula
dlx
= −lx µx
dx
from age α to α + n when x is rewritten α + t.
Z n Z n
dlα+t
dt = − lα+t µα+t dt
0 dt Z0n
∴ [lα+t ]n0 = − lα+t µα+t dt
Z0n
∴ lα+n − lα = − lα+t µα+t dt.
0

and rewriting x for α and dividing by lx

Z n
lx − lx+n 1
n qx = = lx+t µx+t dt
lx lx 0

or
Z n
n qx = t px µx+t dt
0

When n = 1,
Z 1
qx = t px µx+t dt
0

This form of expression is particularly useful when extended to multiple life functions, and can be
roughly described as ‘keep the life alive until duration t; kill him off at that instant; then add (i.e.,
integrate) these probabilities over the required time range’.

Thus the probability n|m qx may similarly be expressed.

Z n+m
n|m qx = t px µx+t dt.
n

Note: The graph of the function µx lx is called the curve of deaths.

EXAMPLE 1.5
An investigation has revealed that 70 percent of newborn babies survived to exact age 40 years
and 40 percent of newborn baby survived to exact age 60 years. Assuming that force of mortality
µ is constant at ages over 20 years. Calculate µ.
Solution
1.3. FORCE OF MORTALITY 13

40 p0 = 20 p0 e−20µ = 0.7
60 p0 = 20 p0 e−40µ = 0.4

Dividing the first equation by the second

e20µ = 1.75
∴ µ = 0.02798

EXAMPLE 1.6
1
Find lx if µx = .
100 − x
Rx 1
Solution. We know lx = l0 e− 0 100−t dt
where l0 is the arbitrary radix.
 Zx 
1
∴ lx =l0 exp − dt
0 100 − t
x
= l0 exp [loge (100 − t)]
0
= l0 exp [loge (100 − x) − loge 100]
  
100 − x
= l0 exp loge
100
 
100 − x
= l0
100

EXAMPLE 1.7
Find the differential coefficient of t px with respect to t and x.
Solution. We know
lx+t
t px =
lx
∴ loge t px = loge lx+t − loge lx
and differentiating with respect to t.
1 d 1 d
(t px ) = · (lx+t ) = −µx+t
t px dt lx+t dt
d
∴ (t px ) = −t px µx+t
dt
Differentiating with respect to x instead of t.
1 d 1 d 1 d
(t px ) = · (lx+t ) = − (lx )
t px dx lx+t dx lx dx
= −µx+t + µx
d
∴ (t px ) = t px (µx − µx+t )
dx
These two formulae are worth remembering.
14 LECTURE 1. THE LIFE TABLES

EXAMPLE 1.8
A life is subject to a constant force of mortality of 0.039221. Find the probability that

(a). that he will live 10 years

(b). that he will die within 15 years

Solution.
 Z 10 
10 px = exp − (0.039221)dt
0
= exp −[0.039221t]10
 
0
= exp [−0.39221]
= 0.67556

We want 15 qx = 1 − 15 px
 Z 15 
15 px = exp − (0.039221)dt
0
h i
= exp −[0.039221t]15
0

= exp [−0.039221(15)]
= exp [−0.588315]
= 0.555262

Therefore 15 qx = 1 − 15 px = 1 − 0.555262 = 0.4447.

EXAMPLE 1.9
The force of mortality is given by
µx
50 < x ≤ 60 0.010
60 < x ≤ 70 0.015
70 < x ≤ 80 0.025
80 < x ≤ 85 0.050
x > 85 0.080
Calculate the probability that a life aged exactly 65 will die between exact ages 85 and 88.
Solution
20|3 q65 = 5 p65 × 10 p70 × 5 p80 × (1 − 3 p85 )
= e−5(0.015) × e−10(0.025) × e−5(0.05) × (1 − e−3(0.08) )
= e0.575 × (1 − e−0.24 )
= 0.1201

Practice Exercise 1.1

∂
1. Differentiate partially with respect to x the following t px .
∂x
1.3. FORCE OF MORTALITY 15

2. Show by integrating the formula Z n

n qx = t px µx+t dt
0
d
using the result dt (t px ) = −t px · µx+t , that n qx = 1 − n px .

3. A life aged 50 is subject to a constant force of mortality of 0.048790. Find the probability that
he will die between the ages of 70 and 80.

1.3.3 Estimation of µx from Mortality Tables

When lx is tabulated and any underlying mathematical formula is not known, values of µx can be found
only approximately, the formula demonstrated below being the most useful.
R1
px = e − 0 µx+t dt
Z 1
∴ µx+t dt = − loge px
0
sometimes expressed as = cologe px

where ‘colog’ is written for ‘− log’. The integral represents the mean value of µx+t between x and
x + 1 and if it is assumed that this approximates to µx+ 1 , we have
2

µx+ 1 ≃ − loge px
2

However in actual experience µx varies rapidly as x increases from 0 to 1, and its rarely possible to find
a satisfactory value for µ0 if the only information available is the value of lx , at integer ages.