ANUx Introduction to Actuarial Science

Lesson 4

The Life Table

History of the Actuarial Profession
Were almost half way through the course and its time for a (very) short history lesson. Were going
to briefly have a look at the history of the actuarial profession. This material is largely based on the
short paper written by Chris Lewin and published in 2001 and a paper by Maurice Ogborn and
published in 1956. The Wikipedia article on Actuarial Science also has a range of information and
references on the history of the actuarial profession that have been utilised here. Youll see some of
the material in the Lewin paper pop up again in Lesson 5.
But lets get back to the history lesson. Insurance policies, or at the very least things that look very
similar to insurance policies, have existed for over 2,000 years. For example, in Roman times,
associations existed that collected a small, regular sum from their members (i.e. a premium) and,
upon the death of a member, paid a benefit to the family of members to cover burial costs. This
structure is not hugely different to what a not-for-profit insurer looks like today.
More formal types of insurance started to appear in medieval times, and in particular focused on the
protection of goods at sea. Trade between regions was an important part of the medieval economy,
with transport by ship being the most efficient method to move cargo long distances. However,
knowledge of the seas was limited, and travel by ship very dangerous compared to todays
standards. Hence there was a reasonable likelihood that the cargo sent by sea would never arrive at
its intended destination (it seems that people were not so worried about the loss of crew!).
Before medieval times, this risk was managed by those financing the journeys charging an exorbitant
interest rate on the money lent to those organising the transport of cargo between destinations.
Should the ship not successfully make the journey then the financier would not receive a repayment
of the loan. Hence, from the financiers point of view, the insurance on the cargo was effectively
incorporated into the interest rate charged.
Insurance in its current form first came into existence in 1350, and was arranged on a cargo of wheat
being sent from Sicily to Tunis. The insurers name was Leonardo Cattaneo, and he agreed to cover
the value of the cargo should it fall foul to some misfortune on the journey, in return for a premium
of 18% of the value of the cargo. Mr Cattaneo was not, however, involved in the financing of the
journey making this the first insurance policy we are aware of.
This type of insurance would today be known as non-life or general insurance. Marine insurance
became quite popular very quickly, but the application of these techniques to the protection of the
financial interests of lives (i.e. life insurance) would take some time.
One of the main reasons for this was a lack of data on mortality and its frequency in medieval
populations. How can you sell a life insurance policy to someone if youve got no idea how long they
are likely to live?
A first attempt at solving this problem came via the work of John Graunt, a haberdasher by trade,
who published the classic book, Natural and Political Observations made upon the Bills of
Mortality in 1662. This book was the first work to demonstrate the consistency in mortality
Adam Butt
Version 1 (2015)

Page 1

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

patterns across groups of people of the same age, which was the basis for the first life table, which is
the topic of this Lesson.
A life table is a representation of the mortality of a population. In Lesson 3, we considered models
of the transition intensity, or force, of mortality. In this Lesson we will extend these models into a
life table format. In fact, those of you who completed Extension Question 3.3 have already started
this process. By way of introduction, the key column in a life table is noted as lx . An example of the
structure of lx is shown below.
Age (x)

lx

30
31
32
33

79
80

100,000
99,923
99,844
99,763

80,611
79,196

The values of lx note the number of individuals aged x who would be expected to be alive given the
values of l at younger ages. For example, using the table above, if 100,000 individuals were alive at
age 30, we would expect 79,196 of the 100,000 to be alive 50 years later at age 80. Using the
notation described in Lesson 3, we can hence say that

50

p30 = 79,196/100,000 = 0.79196.

Finally, it is interesting to note that the term actuary was not in existence whilst John Graunt was
alive. In any case, John Graunt would in todays terminology be called a demographer, rather than
an actuary, as his work focused purely on population statistics and not the financial risks related to
the population statistics. In Lesson 5, we will start to integrate the financial material considered in
Lessons 1-2 with the mortality work considered in Lessons 3-4. The term actuary derives from the
Latin for clerk or account keeper, and was first used in its current context to describe the chief
officer of the Society for Equitable Assurances, the worlds oldest life insurer. The first professional
association of actuaries was the Institute of Actuaries, which was established in 1848 in London.

Assessment Question 4.1

Which of the following statements best describes what a life table is?
A)
B)
C)
D)

A flat bench which has the word life written on it

An equation that describes the transition intensity of mortality of an individual aged x
The group of executives who run a life insurance company
A representation of the mortality of a population by way of tracking the population of
individuals of an age through future years

Practice Question 4.1

Using the life table presented above, calculate the value of 2 p30 .
Adam Butt
Version 1 (2015)

Page 2

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

The Life Table

As described above, a life table is a representation of the mortality of a population. What we will do
first is look at the relationship between the force of mortality and the life table. Well do this using a
simple example, where the force of mortality, , is assumed to be constant at all ages and is equal
to 2% per annum. Well again assume a population of 100,000 individuals at age 30, i.e.
l30

100, 000 .

Calculations
0.02 , is constant, then the probability of mortality over a one year

If the force of mortality,

period, q x , is also constant and can be calculated as follows:

qx

ds

exp

0.02ds

exp

1 exp 0.02
0.019801

We also know that the probability of survival over a one year period, px , is just:

px

1 qx

0.019801

0.980199

If we have a population of 100,000 at age 30, then the population of individuals at age 31 is simply
equal to this 100,000 multiplied by the probability of survival whilst aged 30; i.e.:

l 31

l 30 p30 l 30 (1 q 30 )
100, 000 0.980199 100, 000 (1
98, 020

0.019801)

More generally this can be expressed as:

lx

lx

px

lx

(1

qx )

Since this is an iterative process, if we have the values of px or q x for all required ages x , and we
set the value l at some initial age, then we can calculate l at all subsequent ages. For example:
Adam Butt
Version 1 (2015)

Page 3

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

l 32

l 31 p31 l 31 (1 q 31 )
98, 020 0.980199 98, 020 (1
96, 079

0.019801)

The life table will hence look like the following:

x
30
31
32
33

79
80

lx
100,000
98,020
96,079
94,176

37,531
36,788

qx
0.019801
0.019801
0.019801
0.019801

0.019801

It is also worth noting that l can be expressed as follows:

lx

lx

n 1

lx

lx

px
px

lx

n 1

lx

(1

n 2

px

n 2

px

n 1

...

lx

px

px

... px

n 2

px

n 1

q )

n x

Practice Question 4.2

Assuming that
x

31

0.02 and l30

100, 000 , use a spreadsheet tool to calculate the value of lx from

80 . What is the value of l57 ? You should attempt to replicate the calculation of l57 by

hand.

Assessment Question 4.2

Assuming that
from x

21

0.012 and l20

500, 000 , use a spreadsheet tool to calculate the value of lx

100 . What is the value of l 82 ? You should attempt to replicate the calculation of l 82

by hand.

Adam Butt
Version 1 (2015)

Page 4

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

Introducing some new notation

Well now introduce some new notation that will be helpful in the probability calculations later on in
this Lesson. Well first show it on the life table above:

lx
100,000
98,020
96,079
94,176

37,531
36,788

Age (x)
30
31
32
33

79
80

qx
0.019801
0.019801
0.019801
0.019801

0.019801
0.019801

dx
1,980
1,941
1,902
1,865

743

You can probably guess by looking at the numbers in the table that dx is the number of deaths of
individuals aged x and is calculated as:

dx

lx

lx

lx

dx

lx

lx

lx (1

lx

px
px )

lx (1

px )

lx q x

As an aside, it is worthwhile noting that the above life table contains some significant simplifications
from the sorts of life tables that actuaries would use in practice. Some additional factors that
actuaries would typically allow for are:

Separate tables for males and females as males typically have a higher mortality rate than
females
Mortality rates that reduce over time for example the mortality rate for a 60 year old
today is much lower than it was for a 60 year old in 1960

Another factor that is of great interest to actuaries is whether the mortality rates of people who buy
insurance products will be consistent with the mortality rates of the general population. Imagine a
life insurance product that an individual purchases with a single premium, which pays the
policyholder a claim each year until the policyholder dies (this is called a life annuity). What do you
think the mortality rates of individuals who buy life annuities would be compared to the general
population? You might like to post your thoughts in the forum, where a thread has been created
specifically for this topic.

Adam Butt
Version 1 (2015)

Page 5

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

Assessment Question 4.3

Assuming that
table for x

30

0.005(x

20) and l30

100, 000 , use a spreadsheet tool to produce the life

40 . What is the value of d 38 ? You should attempt to replicate the calculation of

d 38 by hand.

Extension Question 4.1

Return to the spreadsheet solution you created in Extension Question 3.3, a solution for which is
available for download in the relevant Courseware of the edX version of the course. Assuming that
l20

100, 000 , use the spreadsheet tool to produce the life table, using the

Extension Question 3.1 (i.e.

5 10

5 0.085x

2 10 e

attempt to replicate the calculation of d97 by hand.

Adam Butt
Version 1 (2015)

Page 6

and

from

). What is the value of d97 ? You should

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

Calculating Probabilities using the Life Table

In some instances the life table may already have been provided without the need to analyse
transition intensities of mortality. In this case we can calculate desired probabilities without the
need to perform any complex mathematical calculations on transition intensity formulae.
For the examples in this section, we will work with the Male lx from the English Life Table No. 16
2000-02 life table (ELT16), a condensed version of which is available for download in the relevant
Courseware of the edX version of the course and is also reproduced below.

Age (x)

lx

Age (x)

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

100,000
99,402
99,358
99,333
99,316
99,301
99,288
99,276
99,264
99,253
99,242
99,230
99,217
99,203
99,186
99,165
99,138
99,101
99,049
98,983
98,908
98,830
98,753
98,675
98,597
98,517
98,436
98,353

28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55

lx

Age (x)

98,268
98,181
98,091
97,997
97,900
97,798
97,693
97,582
97,466
97,344
97,215
97,080
96,936
96,783
96,619
96,441
96,247
96,034
95,799
95,538
95,249
94,933
94,587
94,212
93,808
93,374
92,906
92,395

56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83

lx
91,835
91,214
90,523
89,754
88,904
87,969
86,946
85,834
84,631
83,328
81,915
80,379
78,708
76,889
74,905
72,743
70,391
67,850
65,125
62,224
59,162
55,957
52,630
49,207
45,701
42,127
38,503
34,852

Age (x)
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109

lx
31,204
27,594
24,080
20,722
17,575
14,683
12,078
9,777
7,773
6,055
4,607
3,412
2,451
1,705
1,149
750
474
289
168
93
49
25
12
5
2
1

A quick reminder about calculating notation

We have only been given lx in the table above. The other notation we have used can be calculated
as:
dx

Adam Butt
Version 1 (2015)

lx

lx

Page 7

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

px

qx

lx 1
lx

lx

lx
lx

dx
lx

A quick reminder about calculating probabilities

Recall that that the probability of an event occurring can be calculated as the number of times an
event occurs divided by the total number of equally possible outcomes. For example, when rolling
two six-sided dice, the total number of possible outcomes (the event space) is 6 x 6 = 36. The
number of times that the sum of the two dice is equal to 11 within this event space is 2 (i.e. {5,6} and
{6,5}). Therefore the probability of the sum of the two dice being equal to 11 is 2/36 = 1/18.
This result will be useful in calculating the probabilities below.

An example of calculating probabilities

Noting all of the above, we can use the life table to calculate any mortality probability we might be
interested in calculating.

Example 1 Calculate the probability that an individual currently aged 25 will survive until age 58
p25 . We did an example equivalent to this right
at the start of this Lesson. Rewording the above, the probability of an event occurring can be
calculated as the number of people meeting the event criteria divided by the total number of people
alive at the starting age. Hence our denominator is the number of people alive at age 25 and our
numerator is the number of people alive at age 58:

In terms of the notation we saw in Lesson 3, this is

33

p25

l58
l25

90, 523
98, 517

33

0.91886

Example 2 Calculate the probability that an individual currently aged 55 will die between ages 85
and age 90
Apart from Extension Question 3.1, we havent seen notation for this type of probability before. It is
a deferred mortality probability, which we can express as

q , where n is the deferral period that

n |t x

the individual must survive during before dying within the t years after that deferral period. In this
instance we are calculating

q . Again we will calculate this as the number of people meeting the

30|5 55

event criteria divided by the total number of people alive at the starting age, in which case our
denominator is the number of people alive at age 55. The numerator is the number of people who
die between age 85 and 90:
Adam Butt
Version 1 (2015)

Page 8

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

d85

d86

30|5 55

d87
l55

d88

d89

l85

l90
l55

27, 594 12, 078

92, 395

0.16793

Assessment Question 4.4

Using the ELT16 life table presented above, calculate q 70 .

Assessment Question 4.5

Using the ELT16 life table presented above, calculate the probability that a newborn (i.e. aged 0) will
die in the first 10 years of their life.

Assessment Question 4.6

Using the ELT16 life table presented above, and available for download in the relevant Courseware
of the edX version of the course, calculate the deferred probability of death for a newborn

q for

n |1 0

each value of n. At which age is a newborn most likely to die?

Assessment Question 4.7 (Hard)

A new life table, which is available for download in the relevant Courseware of the edX version of
the course, contains a life table with only the lx values included. Calculate the expected number of
years that an individual will live from birth (i.e. age 0).
Hint: the expected number can be calculated as:

Number of years lived Probability of living for that number of years

The Probability of living for that number of years is simply

q which you know how to calculate

n |1 0

from Assessment Question 4.6. You can assume that individuals who die whilst age x live for exactly
x 0.5 years (i.e. they die halfway through the year). All you need to do then is sum (i.e. the
symbol) the multiplied values together across all potential ages of death (do a google search of your
spreadsheet tool if you dont know how to sum multiple cells together).

Extension Question 4.2

Return to your answer to Extension Question 4.1 and, using the techniques described for calculating
deferred probabilities of mortality using the life table, calculate the probability that an individual
currently aged 30 will die between age 75 and 80. Confirm that this result matches the result you
calculated by hand in Extension Question 3.1.

Adam Butt
Version 1 (2015)

Page 9

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

Extension Question 4.3

Assume a constant transition intensity of mortality
0 at all ages. Calculate the expectation of
life in terms of , but by hand rather than using a spreadsheet tool. Try performing the same
calculation in a spreadsheet tool, assuming that people die on average halfway through the year. Do
you get the same answer? Why or why not?
Hint: the approach is essentially the same as for Assessment Question 4.7 but replacing the discrete
summation

with a continuous summation

. Without giving too much away, youll need some

pretty sophisticated integration techniques to solve this question.

Adam Butt
Version 1 (2015)

Page 10

ANUx Introduction to Actuarial Studies Lesson 4 The Life Table

Summary of the Lesson 4 Material

A life table is a representation of the mortality of a population where lx is the number of

individuals aged x who would be expected to be alive given the values of l at younger ages.

Other columns typically in a life table are q x , the probability that an individual alive at age x
will die during that year, and dx , the number of individuals who are expected to die at age
x.

Given lx , other values can be calculated as:

dx

lx

px

lx 1
lx

qx

lx

lx

lx
lx

dx
lx

A deferred mortality probability, which we can express as

individual aged x will die sometime between age x

q , is the probability that an

n |t x

n and age x

t.

Probabilities in the life table can be calculated as the number of people meeting the event
criteria divided by the total number of people alive at the starting age.

Adam Butt
Version 1 (2015)

Page 11