Boom or gloom? Cohort fertility data from the 2006 census.

Author:Heard, Genevieve

Data from the 2006 census show a continued decline in cohort fertility. This is especially true of women aged 40 to 44 in 2006 and thus towards the end of their reproductive years. Declines in the number of children born to younger women are also clear. Cohort fertility is considered with regard to population composition data also available from the census. Urbanisation, lower rates of marriage, and immigration all militate against recovery in cohort fertility.


From an all-time low of 1.73 babies per woman in 2001, Australia's total fertility rate (TFR) has recovered somewhat to 1.81 babies per woman in 2005. (1) In newspaper reporting, the increase has been described as a 'boom'. (2) With the exception of some more cautious journalism, (3) this reporting has created the impression of a problem resolved: we have 'turned the corner on fertility'; (4) 'fertility slump ends'. (5) Invariably, the credit goes to the Maternity Payment ('baby bonus') introduced by the current Government in July 2004. (6)

Claims made in academic forums, such as this journal, are more circumspect. Nevertheless, McDonald argues that the TFR has stopped falling and predicts it will remain steady at 1.8 babies per woman for the next decade. (7) Jackson is more sceptical in her interpretation of recent births data. (8) One of her 'points of caution' is that the completed fertility of women currently reproducing--that is, how many babies they end up having--may show no change.

The distinction between the TFR and cohort fertility is an important one. To read the newspapers one might imagine that families are growing in size, when in fact this is not necessarily implied by an increased TFR: a TFR increase, whether or not described as a 'boom', is a cross-sectional phenomenon, and is not necessarily reflected in cohort completed fertility. One would not yet expect to see the current TFR increase reflected in completed fertility statistics, due to the necessary lag between current births and the collection of data on completed fertility. More importantly, the increase may never be evident in completed fertility data, because the TFR--the simplest and most widely used 'snapshot' measure of current fertility--is a synthetic measure, subject to distortions caused by delayed or accelerated childbearing. These caveats are presumably too technical for a general audience and are therefore absent in most media commentary.

Of course, the measure of most interest depends on the concern. The annual TFR is important, regardless of (indeed, because of) delayed or accelerated childbearing, because the number of babies born in any given year matters to the age structure of the population. However completed cohort fertility provides a better measure of generational replacement.

This article considers cohort fertility, both complete and incomplete, in the light of new data from the 2006 census. The census is the primary source of information on cohort fertility; however, the question on children ever born is asked every alternate census only. Therefore, this once-in-a-decade opportunity to study actual cohort fertility is eagerly awaited by Australian demographers. The first wave of 2006 census data was released in June 2007, including results on the number of children ever born to female respondents. When presented as part of a time series, the data allow an assessment of change in the average number of children born to successive cohorts of women over successive years.

This paper further considers cohort fertility with regard to certain aspects of population composition, examining the average number of children per woman by social marital status, by country of birth, and by capital city versus balance of state. For other key variables, such as educational attainment, we must wait for second wave data (with additional variables), due for release in October. Change across these limited dimensions nevertheless provides contextual information in light of which the potential for recovery in completed cohort fertility (in addition to the increase in TFR) might be assessed.


Despite recent increases in birth numbers, cohort fertility data from the 2006 census give little cause for excitement. The average number of children ever born per woman remains in long-term decline (Table 1). This decline is evident in each age group across the reproductive years. Because data on children ever born is collected in every alternate census only, it is difficult to assess the pace of decline--for example, we cannot tell whether, between 2001 and 2006, the decline slowed relative to the preceding five years. Nevertheless, women of all ages in 2006 had fewer children, on average, than did their counterparts in 1996.

Completed cohort fertility

Change in completed cohort fertility is easiest to assess. Women aged 40 to 44 years in 2006 had, on average, 2.05 children. Since women of this age are nearing the end of their reproductive lives, we may assume that this cohort (born 1962 to 66) will fall just short of 'replacement level' fertility (usually taken to be 2.1 babies per woman). (9) By contrast, women aged 40 to 44 years in 1996 exceeded replacement level with an average 2.23 children each. In other words, completed fertility continued its long-term decline over the decade to 2006. (10)

Table 2 shows that the decline continues across the spectrum of family size, but is particularly marked in the growing proportion of women with no children (16 per cent in 2006) by age 40 to 44 years. Given the greater tendency for women without children to omit answering the census question on children ever born, (11) this figure may under-represent the true level of childlessness. (12)

Meanwhile, the proportion of women with three children by age 40 to 44 years continues to decrease, as does the proportion with four or more. McDonald argues these higher order parity progressions are particularly important to the sustainability of Australian fertility levels. (13) Two remains the most common number of children per woman, and the proportion of women with two children by age 40 to 44 years remains relatively stable at 38 per cent.

Incomplete cohort fertility

The primary disadvantage of using actual cohort fertility data is that, for younger women, childbearing is incomplete. Yet trends among younger women are arguably of most interest, given the claimed success of recent pronatalist initiatives, and in light of the fact that the recent TFR increase largely...

To continue reading