« Good introductory textbooks | Gene Expression Front Page | Implications of chromosomal sex determination »
June 27, 2003


This is about birth rates.

The crude birth rate is the number of births per head of the population in a given period, usually a year. It is an objective statistic, and it is useful for some purposes, but it is seriously affected by the age structure of the population.

To avoid this problem what is usually quoted as the ‘birth rate’ is the Total Fertility Rate. This is the number of children that a women would have in her
reproductive life if at any given age she had the same fertility as an average woman of that age at present.

For comparative purposes the TFR is less misleading than the crude birth rate, but it is not sufficiently understood that the TFR can also be misleading in its own way. It is a statistical construct based on the experience of a heterogeneous population of women over a short period of time. It does not measure the fertility of any actual cohort of women in the past, and it does not accurately predict the fertility of any actual cohort of women in the future.

One problem is that the TFR will fluctuate according to temporary circumstances, such as economic recession. If in a particular year women are less likely to have babies because of (say) economic uncertainties, then the TFR may fall sharply, but it is likely to bounce back. More seriously, if there is a long term trend for women to have babies at a different stage of their reproductive life, this will distort the TFR upwards or downwards, and it may take a decade or more for the true picture (the actual lifetime fertility of a cohort of women) to become clear. Notably, if women are postponing having babies from their twenties to their thirties, this will immediately reduce the TFR, but the TFR will eventually rise again when they have the babies they postponed earlier.

So, for example, when you read that the TFR in Japan has fallen to 1.3, this doesn’t necessarily mean that any cohort of Japanese women will on average only have 1.3 children. Maybe they will, maybe they won’t, but at least part of the fall in the TFR is likely to be a temporary distortion due to a time-shift in the pattern of child-bearing.

TFRs for immigrant groups can also be very misleading, as they are often constructed out of the fertility experience of women of different ages, partly in their country of origin and partly in the host country, under different demographic regimes.

I’m not suggesting that TFRs are useless, just that it is important to understand their limitations.


Posted by David B at 09:52 AM

Dear David:

I write about birthrates sometimes and I like to use the Total Fertility Rate because readers can visualize it a lot better. I think thought that it can provide only a partial picture because the average generation length matters too. Perhaps you could help me develop a rule of thumb for thinking about this. For example, if, say, the average age at which an Asian-American woman gives birth is 32 and a Hispanic woman is 24, and the TFR of Asians is 2.1 (the replacement rate)and the TFR of Hispanics is 3.15 (50% higher), how much faster would the Hispanic population grow, all else being equal?

Posted by: Steve Sailer at June 28, 2003 11:11 PM