Demographic Analysis

Mortality:
The Life Table
Its Construction and Applications
 Mortality: Introduction

What is a life table?

A table that displays the life expectancy and the probability

of dying at each age (or age group) for a given population,
according to the age-specific death rates prevailing at that
time. The life table gives an organized, complete picture of
a population's mortality.

Source: Population Reference Bureau, Glossary of Demographic

Terms, http://www.prb.org/Publications/Lesson-Plans/Glossary.aspx

2
 Mortality: Introduction

What is a life table?

A table that displays the life expectancy and the probability

of dying at each age (or age group) for a given population,
according to the age-specific death rates prevailing at that
time. The life table gives an organized, complete picture of
a population's mortality.

Source: Population Reference Bureau, Glossary of Demographic

Terms, http://www.prb.org/Publications/Lesson-Plans/Glossary.aspx

3
 Mortality: Introduction

In this session, we will examine the key components of the

life table – a critical tool in demographic analysis
 Age-specific deaths rates (from deaths and population)
 Conversion of rates into probabilities or age-specific
mortality rates based on “separation factors”
 Implied survival proportions from a hypothetical birth
cohort (100,000)
 Total years lived by that cohort at each age
 Life expectancy at birth (and at other ages)

4
Life Table Construction: From nMx to ex
x n nMx nkx nq x lx nd x nLx 5Px Tx ex

0 1 0.1072 0.33 0.1000 100000 10000 93300 0.901 5630649 56.3

1 4 0.0034 1.56 0.0135 90000 1219 357023 0.983 5537349 61.5
5 5 0.0010 2.50 0.0049 88781 438 442809 0.996 5180326 58.3
10 5 0.0007 2.50 0.0037 88343 324 440903 0.994 4737517 53.6
15 5 0.0017 2.50 0.0084 88019 741 438242 0.988 4296614 48.8
20 5 0.0030 2.50 0.0147 87278 1283 433182 0.984 3858372 44.2
25 5 0.0036 2.50 0.0181 85995 1553 426092 0.978 3425190 39.8
30 5 0.0054 2.50 0.0268 84442 2266 416545 0.973 2999097 35.5
35 5 0.0054 2.50 0.0266 82176 2187 405411 0.952 2582553 31.4
40 5 0.0146 2.50 0.0705 79989 5635 385856 0.934 2177141 27.2
45 5 0.0128 2.50 0.0619 74354 4601 360265 0.907 1791286 24.1
50 5 0.0269 2.50 0.1262 69752 8802 326756 0.895 1431020 20.5
55 5 0.0170 2.50 0.0817 60950 4978 292305 0.864 1104264 18.1
60 5 0.0433 2.50 0.1954 55972 10935 252521 0.816 811959 14.5
65 5 0.0371 2.50 0.1699 45036 7651 206056 0.758 559439 12.4
70 5 0.0785 2.50 0.3281 37386 12266 156264 0.652 353383 9.5
75 5 0.0931 2.50 0.3775 25120 9483 101893 0.569 197119 7.8

5
 Life Expectancy

• Expectation of further life beyond age x

• The average number of additional years a
person could expect to live if current
mortality trends were to continue for the
rest of that person's life.
• Most common: life expectancy at birth or
e0
• Why would this be a useful measure?

6
 1. Age-specific Death Rates

We start the calculation of the life table with age-specific

death rates, nMx

Age-specific central death rates are calculated as

the number of deaths in a particular age group – typically
during a year – per 1,000 population in that same age
group (best measured at midyear):

nM x = nDx/nPx

7
 What is the 5M10?
Age Male
Group Deaths Male Population
0 57,200 1,260,897
1-4 18,526 6,005,011
5-9 9,005 8,730,750
10-14 6,400 8,284,907 5M10=0.772
per 1000
15-19 8,358 6,603,607
20-24 6,249 5,461,651

Source: 2011 Bangladesh Census of Population and Housing

8
 2. Probability of Dying – nqx

• Probability of dying within an interval of

length n that starts at age x and ends at
age x + n

9
 2. Probability of Dying – nqx
A probability requires calculation of an appropriate
population “at risk” for a certain event (in this case, a
death).
To calculate probability of dying within a certain age
interval, those at “at risk” of dying include:
1. Those who survived the interval (e.g. those
counted in a census at the next age interval)
2. Those who died during the interval

Such probabilities can be calculated from age-specific

death rates based on “separation factors” …

10
 2. Probability of Dying – nqx

Probability of dying between exact ages: nqx

In symbols:

n nMx
nqx =
1 + (n - nkx) n Mx

Where:

nMx is the age-specific death rate for ages

x to x+n; and

nkx is the separation factor of deaths for

ages x to x+n.

11
 2. Probability of Dying – nqx

Probability of dying between exact ages: nqx

In symbols:

n nMx
nqx =
1 + (n - nkx) n Mx

Where:

nMx is the age-specific death rate for ages

x to x+n; and

nkx is the separation factor of deaths for

ages x to x+n.

12
 3. Separation Factors nkx
The “separation factor” is the average number of years
lived during an age interval by persons who die within
that age interval.
• For most age groups, set to n/2 (or 2.5 if using 5-year
age groups)
• Different for youngest (ages 0, 1-4) and oldest ages
(open ended age group)

13
 Separation Factors at Youngest and
Oldest Ages
However, at very young and very old ages, the likelihood of death within
a given interval is not the same.
For instance, during infancy, the risk of death is especially high during
the first month. Thus, for example, from birth to age 1, a separation
factor of 0.33 would indicate that, on average, infants who died had lived
for 1/3 of a year (or four months).
Problem with the open ended age interval
• Can’t take n/2.
• Generally use 1/∞Mx

14
 Now let’s calculate nqx

Age n nMx nax nqx

10 5 0.00077 2.500 0.00386
5q15 =
15 5 0.00127 2.500 0.00631

20 5 0.00114 2.500 0.00570

25 5 0.00116 2.500 0.00578
5q30 =
30 5 0.00135 2.500
0.00673
35 5 0.00187 2.500 0.00931

15
 Difference between nMx and nqx

• Time interval
– At most ages, nqx will be about n times that of nMx.
– nqx are measured across an n-year age span
– nMx indicate average rate of death at each year in the
span
• Denominator
– nMx: the denominator of the death rate is based on mid-
year counts of survivors,
– nqx: the denominator is based on those starting an age
interval

16
Infant Mortality Rate (IMR) – a special case
For infants, those “at risk” of dying consist of births.
When vital registration data are available, the infant
mortality rate is thus often calculated as the ratio of
the number of deaths of infants under 1 year of age
(D0) to the number of live births occurring that year
(B), times 1,000,
For example, the IMR for Chile in 1986 is
obtained as follows:
D0 / B * 1,000.
(5,220 / 272,997) x 1,000 = 19.12
(infant
deaths) (births)

There were 19 infant deaths per 1,000

live births in Chile in 1986.

17
 Life Table Construction: nqx
x n nMx nkx nq x lx nd x nLx 5Px Tx ex

0 1 0.1072 0.33 0.1000 100000 10000 93300 0.901 5630649 56.3

1 4 0.0034 1.56 0.0135 90000 1219 357023 0.983 5537349 61.5
5 5 0.0010 2.50 0.0049 88781 438 442809 0.996 5180326 58.3
10 5 0.0007 2.50 0.0037 88343 324 440903 0.994 4737517 53.6
15 5 0.0017 2.50 0.0084 88019 741 438242 0.988 4296614 48.8
20 5 0.0030 2.50 0.0147 87278 1283 433182 0.984 3858372 44.2
25 5 0.0036 2.50 0.0181 85995 1553 426092 0.978 3425190 39.8
30 5 0.0054 2.50 0.0268 84442 2266 416545 0.973 2999097 35.5
35 5 0.0054 2.50 0.0266 82176 2187 405411 0.952 2582553 31.4
40 5 0.0146 2.50 0.0705 79989 5635 385856 0.934 2177141 27.2
45 5 0.0128 2.50 0.0619 74354 4601 360265 0.907 1791286 24.1
50 5 0.0269 2.50 0.1262 69752 8802 326756 0.895 1431020 20.5
55 5 0.0170 2.50 0.0817 60950 4978 292305 0.864 1104264 18.1
60 5 0.0433 2.50 0.1954 55972 10935 252521 0.816 811959 14.5
65 5 0.0371 2.50 0.1699 45036 7651 206056 0.758 559439 12.4
70 5 0.0785 2.50 0.3281 37386 12266 156264 0.652 353383 9.5
75 5 0.0931 2.50 0.3775 25120 9483 101893 0.569 197119 7.8

19
 4. Survivors at exact age x: lx

The next life table function to be calculated is the number

of persons surviving to each exact age, or lx
l0 is called the radix and usually starts with 100,000 births
(at exact age 0).
The number of survivors from birth to each exact age is
obtained using the probabilities, nqx, as estimated above.

20
4. Survivors at exact age x: lx

Survivors at exact age x, lx

In symbols:

lx+n = lx (1 - nqx)

Where:
nqx is as defined above; and

The first lx is l0 and usually is

defined as 100,000.

21
Life Table Construction: lx
Figure III -3. Selected Life Table Functions

22
 Let’s calculate lx

x n nMx nax nqx lx

- - - - - -
0 1 0.03471 0.139 0.03370 100,000
1 4 0.00309 1.757 0.01226 l1 = 96,630

5 5 0.00103 2.500 0.00514 95,446

10 5 0.00077 2.500 0.00386 l10 = 94,955

15 5 0.00127 2.500 0.00631 94,589

20 5 0.00114 2.500 0.00570 93,992
25 5 0.00116 2.500 0.00578 93,456
30 5 0.00135 2.500 0.00673 92,916

23
 5. Life table deaths: ndx
Since lx represents the number of persons alive at each
exact age x, the difference between two consecutive
values (for example, lx and lx+n) represents the number of
deaths between the corresponding ages (x and x+n, in this
case). This number of deaths between two exact ages is
symbolized by ndx in the life table.

24
Life Table Construction: ndx

Deaths between two exact ages: ndx

In symbols:

ndx = lx - lx+n

Where:

lx is as defined above.

25
 Let’s calculate ndx

x n nMx nax nqx lx ndx

- - - - - - -
0 1 0.03471 0.139 0.03370 100,000 3,370
1 4 0.00309 1.757 0.01226 96,630 1,184
5 5 0.00103 2.500 0.00514 95,446 491
10 5 0.00077 2.500 0.00386 94,955 366

15 5 0.00127 2.500 0.00631 94,589 597

20 5 0.00114 2.500 0.00570 93,992 536
25 5 0.00116 2.500 0.00578 93,456 540
30 5 0.00135 2.500 0.00673 92,916 625

26
 6. Person Years Lived: nLx
The total number of years that the single cohort of
100,000 would live between two specific ages, x to
x+n.
Area under the lx curve

27
Life Table Construction: lx
Figure III -3. Selected Life Table Functions

28
 6. Person Years Lived: nLx
The nLx function can be computed as the sum of person-
years lived by those that survive to age x+n and the
person-years lived by those that die between those ages:

nLx = n * lx+n + nkx * ndx

29
 Let’s Calculate nLx

x n nMx nkx nqx lx ndx nLx

- - - - - - - -
0 1 0.03471 0.139 0.03370 100,000 3,370 97,100
1 4 0.00309 1.757 0.01226 96,630 1,184 383,864
476,001
5 5 0.00103 2.500 0.00514 95,446 491
10 5 0.00077 2.500 0.00386 94,955 366 473,859
15 5 0.00127 2.500 0.00631 94,589 597 471,452
20 5 0.00114 2.500 0.00570 93,992 536 468,620
25 5 0.00116 2.500 0.00578 93,456 540 465,930
30 5 0.00135 2.500 0.00673 92,916 625 463,017
35 5 0.00187 2.500 0.00931 92,291 859 459,307

30
 7. Person years of life remaining: Tx
The person years of life remaining may be calculated by
summing all the nLx values

Person years of life remaining for ages x and above: Tx

In symbols:

w
Tx =  nLi
i=x

Where:

nLx is as defined above; and

w represents the oldest age.

31
 8. Life Expectancy: ex
• Expectation of further life beyond age x

• The average number of additional years a person

could expect to live by those who are alive at each age
x.
• The number of years that the life table population will
live from age x up to the point when all have died,
divided by the number of persons alive at exact age x,

32
 Life Table Construction: ex

Life expectancy at age x: ex

In symbols:

Tx
ex =
lx

Where Tx and lx are as defined above.

33
 Let’s Calculate ex

x n nMx nax nqx lx ndx nLx Tx ex

- - - - - - - - - -

0 1 0.03471 0.139 0.03370 100,000 3,370 97,100 7,558,096 75.58

1 4 0.00309 1.757 0.01226 96,630 1,184 383,864 7,460,996 77.21

5 5 0.00103 2.500 0.00514 95,446 491 476,001 7,077,132 74.15

10 5 0.00077 2.500 0.00386 94,955 366 473,859 6,601,130 69.52

15 5 0.00127 2.500 0.00631 94,589 597 471,452 6,127,272 64.78

34
 Life Table Construction: ex
x n nMx nkx nq x lx nd x nLx 5Px Tx ex

0 1 0.1072 0.33 0.1000 100000 10000 93300 0.901 5630649 56.3

1 4 0.0034 1.56 0.0135 90000 1219 357023 0.983 5537349 61.5
5 5 0.0010 2.50 0.0049 88781 438 442809 0.996 5180326 58.3
10 5 0.0007 2.50 0.0037 88343 324 440903 0.994 4737517 53.6
15 5 0.0017 2.50 0.0084 88019 741 438242 0.988 4296614 48.8
20 5 0.0030 2.50 0.0147 87278 1283 433182 0.984 3858372 44.2
25 5 0.0036 2.50 0.0181 85995 1553 426092 0.978 3425190 39.8
30 5 0.0054 2.50 0.0268 84442 2266 416545 0.973 2999097 35.5
35 5 0.0054 2.50 0.0266 82176 2187 405411 0.952 2582553 31.4
40 5 0.0146 2.50 0.0705 79989 5635 385856 0.934 2177141 27.2
45 5 0.0128 2.50 0.0619 74354 4601 360265 0.907 1791286 24.1
50 5 0.0269 2.50 0.1262 69752 8802 326756 0.895 1431020 20.5
55 5 0.0170 2.50 0.0817 60950 4978 292305 0.864 1104264 18.1
60 5 0.0433 2.50 0.1954 55972 10935 252521 0.816 811959 14.5
65 5 0.0371 2.50 0.1699 45036 7651 206056 0.758 559439 12.4
70 5 0.0785 2.50 0.3281 37386 12266 156264 0.652 353383 9.5
75 5 0.0931 2.50 0.3775 25120 9483 101893 0.569 197119 7.8

35
Life Table Construction: Survival Ratios
Finally, we have the life table survival ratio, denoted
either as 5Px or 5Sx.

Survival ratio: 5Px or 5Sx

In symbols:

5Lx+5
5Px =
5Lx

Where:

5Lx is as defined above.

36
Life Table Construction: Survival Ratios
For survival from birth to ages 0 to 4 and for the open-
ended age group, we have:

Survival ratio from birth to Survival ratio of open-ended

ages 0
1 to 4: 5 Pb age group: Pw

5 L0
5 Pb = Tw
5 x l0 Pw-5 =
Tw-5
Where:
Where:
5 L0 and l0 are as defined above.
T is as defined above; and
w represents the oldest age.

37
Life Table Construction: Survival Ratios
Figure III -3. Selected Life Table Functions

38
 Applications for Life Table Construction
• PAS: LTPOPDTH.xls
– User enters population and deaths by ages
– Allows for alternative infant mortality input
– Allows for smoothing of nMx values
• MORTPAK: LIFTB
– User enters nMx, nqx, or nlx
• PAS: LTMXQXAD
– User enters nMx or nqx
– Allows users to incorporate adjustment factors to
correct nMx or nqx

39
 Exercises

• Construct a life table for your country

based on population and death data by
age using LTPOPDTH.

40
 Exercises

• (Optional) Use another application and

compare the results to see why the results
may be different.

41