The future size and composition of populations are of critical concern to policy-makers, and the most uncertain element in population forecasts is fertility. The number of births in the United Kingdom fell from one million in 1965 to 630,000 in 1977, prompting claims that fertility would reach a level so low that the population would decline rapidly and its age structure become imbalanced. Similar fears were expressed in the 1930s, notably by Enid Charles in 'The Twilight of Parenthood'. She argued that Western Europe and the United States were facing an imminent, rapid and inevitable population decline. Her views found some support among demographers, who cited as evidence the fall in total fertility to about 1.7 in England and Wales in 1933. These fears proved unfounded, however: the birth rate soared in the postwar period.

At a joint CEPR/BSPS lunchtime meeting on 3 October, demographer Professor William Brass argued that present concerns about population decline and an unbalanced age structure could prove to have as little foundation as those voiced in the 1930s. Conventional fertility measures such as the annual number of births were unreliable, according to Brass, because they are subject to large transient fluctuations. He drew attention to other fertility measures that are less subject to such distortions and give a better indication of long-run population trends. These indicators, Brass noted, displayed only half the variation of conventional fertility measures and currently indicate that fertility is slightly above population replacement level. There is little or no support for the claim that Britain is facing a 'flight from childbearing', Brass concluded.

William Brass is Professor of Medical Demography and Director of the Centre for Population Studies at the London School of Hygiene and Tropical Medicine. He is currently President of the International Union for the Scientific Study of Population. He spoke at a CEPR lunchtime meeting, the first of a series organized jointly with the British Society for Population Studies, at which demographers present the results of recent research and their implications for economic and social policies. The meeting at which Brass spoke was sponsored by the Simon Population Trust.

Brass first discussed conventional measures of childbearing. These included not only the number of births per year, but also total fertility, the average number of children born per woman if the current age-specific birth rates persist for a group as it moves through its reproductive span. A total fertility of only 2.1 is sufficient to ensure population replacement, Brass noted, since under current conditions most females survive to reach childbearing age. Total fertility in Britain was below replacement level in 1977, at 1.7, although it rose a little to 1.8 in 1985. Many European countries showed more dramatic shortfalls from replacement levels; in 1983, for example, total fertility stood at 1.38 and 1.33 in Denmark and West Germany, respectively.

But the annual birth rate and the total fertility are not satisfactory indicators of long-run trends in fertility, Brass noted. They contain a transient component resulting from fluctuations in the timing of births, both from systematic changes in their distribution over the reproductive period and from comparatively short-term socioeconomic disturbances. Other fertility indicators are less affected by these transient fluctuations. The simplest of these is the cohort total fertility, the average number of births over the entire reproductive span of a 'cohort' of women born or married in the same year. In Britain this measure has shown considerably less variability than the conventional indicators.
{v20 b850}
Measuring cohort total fertility presents some difficulties, however, since childbearing is not effectively complete until about age forty. We know that the total fertility for the cohort born in 1945 was above replacement level. Later cohorts have experienced considerably lower birth rates in the earlier ages of reproduction, but we cannot predict with certainty whether this decrease is due to postponement of childbearing or smaller desired family sizes. Estimates of cohort total fertility can, however, be extrapolated from survey evidence on how many further children women expect to have. The 1979-83 General Household Survey reports the average family size expected by women in different cohorts. These figures only vary between 2.4 and 2.2 for the five-year birth cohorts between 1940 and 1964, and give little indication of any further fall in cohort total fertilities. Adjustments for non-response to the Survey lower the extrapolated total fertilities but leave them close to replacement levels.

Brass also outlined several techniques which could eliminate from period fertility rates the effects of systematic shifts in the timing of births. These techniques involved the use of parity cohorts, i.e. women moving from their (n)th to their (n+1)th birth. Such cohorts span only a few years, so that there is little difference between the cohort and period values. Brass had calculated the total period fertility (parity) measure for England and Wales over the period 1957-84, and noted that this measure displayed only half the range of variation of the conventional total fertilities and agreed with the cohort actual and expected values. In 1984, this alternative parity measure was still above replacement levels and indeed had shown little systematic movement since the late 1960s.

Brass drew attention to the another indicator of long-run fertility trends, the parity progression ratios; these are the proportions of women who had already had n births going on to have an (n+1)th. These ratios fell at higher orders (i.e. two to three, three to four, and over) until about 1975, but have since risen. Families of two children have become relatively more popular than families of one child. The most doubtful estimate is of the percentage that will remain childless. Marriage rates at younger ages have fallen steeply, but there has been a large increase in illegitimacy. The impact of these changes on first births and on completed family size is extremely hard to discern, but Kathleen Kiernan has reported that 90% of the women in the 1958 birth cohort expected to have children, according to the 1981 National Child Development Study. This is close to the estimated time period parity progression ratios for women who had not yet had children in the 1970s and 1980s. If family size expectations are fulfilled there will be a substantial increase in births in the 1990s, as numbers of the women in the reproductive period increases and as women who have postponed childbirth at last attempt to achieve their desired family size.

Parenthood still seems to be popular, Brass concluded. There is almost no evidence of a 'flight from childbearing' in the United Kingdom, nor of imminent population decline or age structure imbalance.

In the ensuing discussion, participants raised various questions concerning the timing of childbirth and the interaction of the desire for children with women's employment patterns. Whereas in the 1960s women completed their childbearing relatively early, the pattern now is for later first births and longer intervals between births.

One member of the audience drew attention to differences in fertility patterns across Europe. Britain seemed to have generally conservative attitudes to reproduction, according to Brass, varying little from one generation to the next. Whereas in Italy and the Netherlands there were lower parity progression ratios, and hence more childless and single-child mothers, this was not so in Britain or France. There seemed to be no single quantifiable factor explaining these cross-country differences, Brass noted.