Life expectancy is calculated using a mathematical tool called a
'life table'. These are constructed by taking death rates from the
population in question (such as Australian males in 2006) and
applying them to a hypothetical cohort of persons. The life table
is then able to provide probabilities concerning the likelihood of
someone in this hypothetical population dying or surviving before
their next birthday. Life expectancy can be provided for any age in
the life table, by summing the number of person years (the total
number of years lived by all persons in the life table) and
dividing this by the number of persons still alive in the life
table.
Understanding how life tables work illustrates two important
points about life expectancy as an indicator. Firstly, the life
table models a synthetic population, not a real one. Life
expectancy is therefore only a guide to the health of a population
at a specific point in time - it is not a prediction about how long
people alive today will actually live. Secondly, changes in
mortality at younger ages will have a larger impact on life
expectancy than changes at older ages. This is because the total
number of years the population lives (person years) is used in the
life table. For example, infants surviving into adulthood and
living long lives will add a considerable number of person years to
the population when compared to the elderly living only a few years
extra.
It should also be noted that, since 1995, the ABS has published
life tables based on three-year averages (such as deaths and
populations over the period 2004-06). This is so that fluctuations,
especially in older ages where there is a small population at risk,
are kept to a minimum. However, as life expectancy is currently
increasing, taking a three-year average will (for some years)
slightly depress the life expectancy figure we would get compared
to if the latest single-year data were used.