From Ivan Dixon
Your interesting piece on the “population bombshell”
(Inside Science, No 112, 11 July)
reminded me of a correlation I’ve never seen mentioned in any article
on population. I came across it accidentally when testing a graph-plotting
program some years ago.
The data I used came from the World Bank database that was then located at
the University of Bath. It consisted of tables of annual data for the 27 years
stretching from 1960 to 1986, for each of 120 countries, for various variables,
including some population variables.
Plotting the fertility ratio (number of children per woman) against life
expectancy at birth showed a very strong relationship between the two variables.
Fertility always reduced as life expectancy went up.
Whether it was industrial countries, agricultural, First World, Third World,
Catholic or non-Catholic, individual countries or all countries lumped together,
annual data or 27-year means, the exact correlation might vary but there was
always a very strong relationship. For the pooled data, the crossover point
where fertility went below the replacement rate of 2.1 children per woman was a
life expectancy of 79.5 years.
Advertisement
Finding a correlation between two variables does not prove that one drives
the other (both could be driven by something else), but it would make a kind of
sense if life expectancy did influence the fertility ratio. If life is short and
people tend to die young, there would be strong psychological pressure to have
children as soon as possible. Conversely, if people survive for many years, then
children can be postponed. It seems probable that the older people are when they
start having children, the fewer children they are likely to have.
If fertility really is driven by life expectancy, this has profound
implications for aid programmes and population control. It would mean that it’s
not the aluminium smelter or the big dam that matters, but rather the
unglamorous sewage systems, clean water, health services, affordable food,
decent shelter, and anything else that leads to an increase in life span.
Unfortunately the data I have end in 1986. I would be most interested to see
comparable data for more recent years. I would really like to know if the
relationship has remained as strong.
Exeter, Devon
