[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries
.

Fertility assumptions

In the first 1970-based population projection for England and Wales, an elaborate matrix of fertility rates reflecting the latest estimates of parity progression within marriage was used to estimate fertility in future years, in conjunction with rather sweeping assumptions as to fertility in second marriages and extra-marital fertility. A completed family size of 2.4 children for married women was a long term assumption (after 1985). Added to this were a constant annual number of 65,000 illegitimate births and ‘about 20,000 a year’ for children born of remarried women.
This forecast performed quite well in the short term, anticipating some fall in the birth rate in the early 1970s. But after 1973 conditions changed and there was a considerable fall in the birth rate that had not been anticipated. The growth in extra-marital births made it necessary to fall back on a more primitive methodology that did not distinguish between intra-marital and extra-marital parity progression. The fall in marriage rates was associated with other changes and was itself a reason for the fall in the birth rate. Also, an increase in divorce led some couples to cancel plans to have children.
For more than ten years (1975 to 1985), even after the birth rate had fallen considerably, the official assumption was of a recovery in the total fertility rate (TFR) in the long term to 2.1 in the 1975 to 1983-based projections and to 2.0 in the 1985 to 1989-based projections. The sequence of long-term assumptions for the TFR for successive projections is shown in figure 1, lagging behind the parallel fall in the observed birth rate. It seems the forecasters had faith in a spontaneous recovery in the TFR to around replacement level and ignored the major cultural and economic changes leading to fertility decline. In 1973 there was an increase in government promotion of free hormonal contraceptives for married and unmarried women. The 1967 Abortion Act opened the way for a large increase in legally induced abortions, free of charge from the National Health Service, reported in the 1970s.
For England and Wales there is a curiously close fit between the shortfall below replacement level of the birth rate and the abortion rate. This is shown in figure 2.
It seems that modern fertility control has taken away the recovery in the birth rate intended by the prospective parents and anticipated by the demographers, but the ostensible rationale for assuming a recovery back to replacement level in the birth rate was that modern fertility control would enable couples to have the children they intended to have. Women interviewed in surveys such as the General Household Survey indicated their intentions to have more than two children. ‘The expected family size has not changed significantly over the five years that the GHS has been operative’, and for the 1976-based projection it was decided to disregard current fertility rates as a guide to the long-term fertility trend. Although other countries, such as West Germany, were showing a marked decline, the pattern was not clear and replacement level was again assumed to apply in the long term. This stance was maintained throughout the 1980s, as shown in figure 1.
From the 1991-based projection, long-term fertility assumptions were lowered to a compromised level, to TFR 1.9. The current TFR was still considered to be too low to use for the long term and there continued to be an assumption of future recovery.
Most recently the committee determining the fertility assumptions used by the government actuary have had more success. The 1994- (and subsequent until 1998) based projections assumed a long-term TFR of 1.8 in England and Wales and in 2005 the observed TFR did recover to 1.8 in England and Wales (see figure 1).

Scotland
Age-specific information on fertility and parity progression in Scotland was less detailed and the approach in the first projection was to set the fertility assumptions for Scotland in relation to those for England and Wales using assumed ratios of Scottish fertility to English at the relevant quinquennial age groups. For example, for ages 25 to 29 the ratio was assumed to be 1.08 or Scottish fertility was assumed to be 8% more than English. In the first 1970-based projection it was assumed that these ratios would apply in all future years.
Scottish fertility was then higher than English and it was assumed that this pattern would continue in the event there was a greater decline in Scottish fertility than English right through the 1970s and 1980s. More recently, this Scottish deficit has continued in the 1990s and now into the 21st century. It has grown to be about 0.2 or one-fifth of a child. As a result, forecasts of fertility for Scotland have proved to be even more optimistic than for England.
After 1981, when it was apparent that Scottish fertility had fallen below English, the assumption of long-term fertility for the 1981-based forecast was the same for Scotland as for England. Not until the 1996-based forecast was a lower future fertility rate assumed for Scotland. The sequence of assumed future rates and observed rates for Scotland is shown in figure 1.
It seems as though the demographers on the fertility committee were reluctant to accept the divergence in fertility between Scotland and England. Scotland has a more working-class culture that was more severely affected by the demise of traditional industries and the privatisation of council housing. Whereas miners had large families, in modern conditions it is easier for women than men to find employment. The financial strains of owner occupation are a deterrent to family formation. Some cultural changes have also been greater in Scotland with a higher proportion of births outside wedlock tending to reduce parity progression.
In Scotland the committee determining the fertility assumptions used by the government actuary have also had more success recently. The 2000-based and subsequent projections assumed a long term TFR of 1.6 in Scotland and in 2005 the observed TFR did recover to 1.6 (see figure 1).

Northern Ireland
For Northern Ireland even less detail on age-specific fertility was available than for Scotland and a constant ratio for Irish to English fertility was assumed across the whole age range of women from 15 to 44 in the first 1970-based projection. The ratio of Northern Irish birth rates to English had been 1.3 from 1968 to 1970. It was assumed that a ratio of 1.3 would decline gradually to 1.2 over 40 years into the future.
In the event there was a greater decline in Northern Ireland than elsewhere in the UK but it still has the highest birth rate. For the 1983-based projection it was assumed that Northern Ireland would have the same birth rate in the long term future as the rest of the UK, ie 2.1 (replacement level). But in successive population projections it has generally been assumed that the Northern Ireland birth rate would continue to be higher than England. The sequence of assumed future rates and observed rates for Northern Ireland is shown in figure 1.
The assumption in the 1980s of near replacement level fertility in the long run has not proved to be so wide of the mark in respect of Northern Ireland as for other parts to the UK. There is some cultural affinity between Northern Ireland and Scotland in respect of factors influencing fertility and similar economic changes have occurred since 1970, such as the decline of traditional industries and privatisation of council housing. The 1967 Abortion Act, however, has never been extended to Northern Ireland and this continues to be a factor that goes some way to explain higher fertility in Northern Ireland than in Scotland.
In Northern Ireland the committee determining the fertility assumptions used by the government actuary have also had more success recently. The 1996 to 1998-based assumption of a long-term TFR of 1.85 for Northern Ireland and the 2000-based and subsequent projections that assumed a long-term TFR of 1.8 in Northern Ireland are quite near the observed TFR of 1.82 for 2005 (see figure 1).

Beyond 2007 for all of the UK
In the 21st century there may now be some stability in fertility rates. Whereas a continuation of under-replacement fertility is hardly sustainable in the long run and has serious consequences, this is only slowly being acknowledged, even in Scotland, where there is most reason for concern at a low birth rate. Neo-Malthusian thinking, whereby a limit on population growth is seen as important to economic success, still predominates among policy makers. For governments to address the issue and take effective measures to reverse the decline in fertility would require a reversal of policies in many areas, which cannot be anticipated in the near future.
There are also to be recognised some 21st-century factors pointing to a further decline in the birth rate. Higher house prices are much discussed but the negative implications for fertility are less acknowledged. Women are claiming a greater share of educational and career opportunities. Universities are more accessible and a higher proportion of students in higher education are now women. More young women are graduating with larger student debts to repay. Men will have to accept a smaller share of the best-paid posts. Here again, demographers are failing to step forward to say what implications these trends bring for the birth rate.

07_04_03.pdf