Vol. 24, No. 2, 2007 Journal of Population Research

�

WHICH WOMEN STOP AT ONE CHILD IN AUSTRALIA?

Nick Parr,† Macquarie University

The decline in fertility in Australia in the 1990s reflected both decreased first-order birth rates and decreased second-order birth rates (Kippen 2004). Whilst childlessness has been studied extensively, little attention has been paid to the progression from one to two children. This study analyses which women stop at one child using data from 1,809 parous 40–54 year olds from Wave 1 of the Household Income and Labour Dynamics in Australia (HILDA) survey. Impor-tant early lifecourse predictors of whether a woman stops her childbearing at one child are shown to be a woman’s country of birth, highest level and type of schooling, and her father’s occupation. A woman’s marital status and her age at the time of the first birth are also shown to be significant predictors of her likelihood of not progressing to a second birth. The causes of trends over time are discussed.

Keywords: Fertility, cohort fertility, parity progression, births, one child, family formation, Australia, low fertility, Second Demographic Transition, migrants

Despite the recent small increase in the total fertility rate (from 1.73 in 2001 to 1.81 in 2005), the current low level of fertility remains a prominent issue in Australia’s public debate (ABS 2006a). Kippen (2004) has shown that the reduction in fertility in Australia between 1991 and 2000 was due largely to the reduction and the delay of first and second births, there having been little change in fertility rates for women with parity two or higher over this period. The percentage of women who by their 50th birthday had a completed family size of one increased from eight per cent to 11 per cent between 1991 and 2001 and is projected to rise to around 15 per cent in 2021, almost double the figure for 1991 (Kippen 2006). Whilst the demographic character-istics of the childless in Australia have been studied extensively (Merlo and Rowland 2000; Weston and Qu 2001; Parr 2005), it appears that little attention has been paid to women who cease childbearing after the first child. In some countries other than Australia, one-child families have become widespread. One example is China, which pursues a one-child policy, but other countries where a one-child pattern predomi-nates include Austria, Bulgaria, Italy, Romania, Spain, and the European countries formed from the former Soviet Union (Avdeev 2001; McDonald 2002; Greenhalgh 2003; Sobotka 2003). However, despite the increase in its prevalence during the 1990s, in contemporary Australia the one-child family is still something of a rarity; the two-child family remains the norm (McDonald 1998; Kippen 2006).

† Address for correspondence: Senior Lecturer in Demography, Department of Business, Division of Economic and Financial Studies, Macquarie University, Sydney, NSW 2109 Australia. E-mail: Nick.Parr@mq.edu.au.

2 Nick Parr

Since the 1960s the patterns of fertility and family formation in Australia have, broadly speaking, followed those associated with progression through the ‘Second Demographic Transition’ (Lesthaeghe 1995; Van de Kaa 1997; Carmichael 1998). Aus-tralia’s total fertility rate declined from a post-World War II peak (3.55) in 1961 to its lowest ever recorded level (1.73) in 2001, with the reduction in total fertility being particularly rapid in the early 1970s (ABS 2006b). Ages at childbearing and at mar-riage increased steadily from the early 1970s (ABS 2006b). The fertility trends were facilitated by the spread of access to new, more efficient methods of contraception, especially the contraceptive pill, and by the removal of adverse legal consequences from abortion (Santow 1991; Carmichael 1998). Some of the increase in age at mar-riage is linked to unmarried cohabitation becoming more prevalent, so much so that a period of unmarried cohabitation before a legal marriage has become the norm (ABS 2006c). Divorce has become more prevalent, particularly following a liberaliza-tion of divorce law which took effect in 1976 (Carmichael 1998). The percentage of children born outside a legal marriage has increased, reflecting not only the declining prevalence of the married but also the higher rates of childbearing among the unmar-ried (ABS 2006a).

The above changes in social and family life in Australia could well contribute to circumstances in which a woman is likely to have no more than one child in her lifetime. Furthermore, the progression from one to two children has ramifications for family budgets, parental time use, and the socialization and well-being of the existing child. Additional children cost money. However, the marginal direct cost (i.e additional expenditure required) has been found to be considerably greater for the first child than for higher-order children (Percival and Harding 2002; Henman et al. 2007). The first child has also been found to have a far greater marginal indirect cost, the forgone earnings of (usually) the female partner, than higher-order children (Chapman et al. 2001; Breusch and Gray 2004).

In Australia the economics of childbearing is also affected by the availability of a complex range of government benefits which are payable to the parents of children. The more significant benefits are means-tested on family income (Family Tax Ben-efit Part A) and on the income of the lower-earning parent (Family Tax Benefit Part B) (McDonald 2001). A means-tested benefit which partly covers the cost of child-care is also available. With effect from July 2004, the Australian Federal Government introduced a substantial, flat-rate payment to the mothers of all newly-born children, known as the Maternity Benefit, and increased the amounts and income thresholds for eligibility for Family Tax Benefits. In doing so it phased out a tax rebate based on the reduction of income following the birth of the first child, known as the Baby Bonus. The changes appear to have been at least partly motivated by its development of a concern that Australia’s birth rate was ‘too low’ (UNPD 2006). Australian Federal Treasurer Peter Costello’s widely publicized soundbite ‘If you can have children it’s a good thing to do – you should have one for the father, one for the mother and one for the country, if you want to fix the ageing demographic’ is evidence to this effect (Dodson 2004; Heard 2006).

There are time costs of children too. The time spent looking after their children is on average greater for women with more than with one child than for women with one child. As well as the additional time spent looking after their children being to the detriment of time spent in work, it also reflects a loss of time for personal care (sleeping, eating and drinking, bathing etc.) and for recreation, that is, recreation

Which Women Stop at One Child in Australia? 3

without children. However, the effects on time use of each additional child are less than the effects of the first child (Craig 2003, 2005, 2006; Craig and Bittman 2003).

Second, and higher-order, children should also have utility for the parents. Another reason why women may have no more than one child is that the marginal utility of higher-order children may not be as great as for the first child. For example the first child uniquely confers on the mother and father the status of parenthood, continues their bloodlines, and more frequently than for any other single birth passes on the father’s Y chromosome and family name (Blake 1979). Progression from one to two children not only affects the parents but also affects the well-being of the first child in numerous ways. Most obviously it brings to the first child the company of a sibling, but it may also deplete the material resources and parental time and energy which may otherwise have been spent on the first child, and this may contribute to the decision to stop at one child (Blake 1989; Parr 2006).

The likelihood of a woman stopping at one child may be related to variables whose values are determined early in the lifecourse. The socio-economic status of a woman’s family of origin, her ethnicity, number of siblings, and level and types of education affect her income and wealth in later life (Parr 2006). Thus the indirect costs resulting from her having children and her ability to contribute to covering the direct costs of children, both important parameters in fertility decision-making, will have been affected by these variables. The intactness of a woman’s parents’ union has been shown to correlate with her own likelihood of divorce and her willing-ness to commit to lasting relationships, and so may affect her fertility (Kiernan and Cherlin 1999; Weston and Qu 2001; Weston et al. 2004; Carmichael and Whittaker 2007). The number of siblings a woman has may have conditioned her preferences for her own fertility and may be related to her genetically inherited fecundity (Mur-phy and Knudsen 2002; Parr 2005). The type of education a woman had, as well as being related to parental wealth and religion (both of which may affect fertility), also may affect the type of education she herself, with her partner, aspires to provide for her children, and hence the prospective direct costs of children (Newman and Hugo 2006; Parr 2006). Fertility is likely to vary by country of birth not only because the socio-economic circumstances of migrant groups differ considerably but also because birthplace and ethnicity may condition attitudes to the importance of having sons and daughters, and expectations of the levels and types of parental involvement in children’s upbringing. Moreover, the act of migration may disrupt childbearing and unions (Abbasi-Shavazi and McDonald 2000). A woman’s age would be expected to be related to her fertility not only because even after the age of forty, the lower bound for the age-range analysed, there will be some variation by age in the numbers of children still to be produced, but also because the differing socio-economic circum-stances which women from different cohorts have experienced will have affected their fertility levels.

The analysis in this paper begins with examination of the relationships between early lifecourse variables (specifically parental occupations, the intactness of the parental union, numbers of siblings, the levels and types of education, ethnicity and age) and whether a woman with one child did not progress to a second child. The effects of variables measured at the time of the first birth are then considered, in par-ticular mother’s age at first birth, her marital status at this event, and the birthplace of the first child. The age at which the first child was born is likely to be an important determinant of her subsequent fertility, because it affects her remaining reproduc-

4 Nick Parr

tive lifespan and fecundity (Bongaarts and Potter 1983). The marital status at the time of first birth is related to the stability of and commitment to the union, and so may affect whether additional children are produced (De Vaus, Qu and Weston 2005; Liefbroer and Dourleijn 2006). Whether or not the first birth was outside Australia is considered because conditions affecting parity progression experienced outside Aus-tralia may differ from those experienced within it. The progression from one child to two children would also be affected by the varying values of a wide range of variables between a woman’s first birth and the cessation of her fecundity. However, the analysis of the effects of such variables is more complicated, since the statistical correlations between variables measured after the time of the first birth may reflect the effects of family size on the values of such variables as well as the effects of the variables on family size. Hence, the effects of variables measured after the first birth are not included in the models.

Data and methods

The data used are from Wave 1 of the Household Income and Labour Dynamics in Australia (HILDA) survey. This is a large-scale, nationwide, longitudinal survey of the household population of Australia conducted in 2001 by the Australian Common-wealth Government’s Department of Family and Community Services. A multistage, cluster sample design was used, and interviews were completed successfully with 13,969 men and women from 7,682 households and 488 census collection districts, which were stratified by state or territory and metropolitan or non-metropolitan area (Watson and Wooden 2002a). The sample was selected to be nationally representa-tive, except for the exclusion of some very remote areas. Differential non-response resulted in discrepancies between the distribution of respondents and the population sampled: residents of Sydney, unmarried persons and immigrants from non-English-speaking countries are underrepresented (Watson and Wooden 2002b). However the discrepancies are fairly minor and, hence, the analysis is unweighted. Data were col-lected on family formation and background, employment and unemployment his-tory and status, and income.

In this paper the analysis is restricted to the 1,809 female respondents who were recorded as aged between 40 and 54 years in the Wave 1 data and as having had one or more children. In three cases the only child had died. Almost all of the women with one child appear to have completed their childbearing, since when asked ‘how likely are you to have a child/more children in the future?’ only six (two per cent) of women rated the likelihood 6 or above on a scale from 0 to 10.

Two types of analysis are presented in this paper: first a descriptive analysis and then a multivariate analysis. Two logistic regressions of whether or not a woman with one or more children did not progress to having a second child are presented. Model 1 enters the early lifecourse variables and Model 2 adds variables measured at the time of the first birth. The early lifecourse variables entered are: birthplace (with countries grouped on geographical and linguistic lines); Aboriginal or Torres Strait Islander descent; highest level of education;1 type of school attended; number of siblings (grouped as an indicator of whether or not the respondent was an only child); father’s occupation, mother’s occupation, both measured using the Australian Standard Classification of Occupations (ABS 1997) and based on reports for age 14;2 intactness of the parental union at age 14; and age. The age at first birth,3 the marital

Which Women Stop at One Child in Australia? 5

status at that time, and whether the first birth occurred outside Australia are then introduced to the model.

Results

Descriptive analysisThe women with one child represent 12 per cent of women aged 40 to 54 and 13.5 per cent of such women with one or more children (Table 1). Table 1 shows the strength of the two-child norm: women with two children far outnumber those with three children, the second most common parity. The fertility of women aged 50 to 54 is clearly higher than that of women aged between 40 and 49. This pattern reflects the older group having spent its prime reproductive years during a period of higher fer-tility, and also is an effect of the incomplete fertility of the younger women, especially the 40-to-44-year-olds. The percentage with one child is greater among women 45 to 49 years old than among the other age groups. However, the difference from the 40-to-44 age groups is small and the comparison would be affected by delayed first and second births.

Table 2 shows the variation in the numbers of women with one child expressed as a percentage of all women with one or more children, that is parous women. For reasons discussed above, the percentage of parous women aged 45–49 last birthday who have stopped at one child is slightly higher than the percentages of women aged 40–44 and 50–54 who did so. Of the variables which are usually determined early in the lifecourse, the greatest range of variation is between the country-of-birth groups. All the overseas-born groups are more likely than the Australia-born to have stopped their childbearing at one child, and there is a wide variation between the different overseas-born groups. Half the parous women who were born in an East Asian country have only one child; the women sampled from this group were mostly born in either China or Hong Kong. One quarter of the parous women who were born in a Northern, Western (excluding the UK and Ireland) or Eastern European country have just one child; the largest national subgroups are those born in the former republic of Yugoslavia, Germany, Netherlands and Poland. The percentages of women born in the main English-speaking overseas countries (UK, Ireland, New

Table 1 Percentage distributions of children ever born by five-year age groups of women

Age last birthday

Number of children ever born (%)

0 1 2 3 4 5+ N

40–44–4444 13.7 12.0 37.3 22.6 9.9 4.6 79145–49–4949 12.1 13.6 35.5 24.6 8.9 5.3 66250–54–5454 9.0 10.0 39.3 26.1 9.2 6.4 598Total 11.8 11.9 37.3 24.3 9.4 5.3 2,051

Source: Living in Australia (HILDA) Survey Wave 1.

6 Nick Parr

Table 2 Descriptive statistics. Women with one child as a percentage of women aged 40–54 with one or more children

Variables Percentage with one child N

BirthplaceAustralia 11.1 1,285Main English-speaking overseas countriesa 18.0 200Northern, Western or Eastern Europe 25.0 96East Asia 50.0 22South or Southeast Asia 18.5 92Other overseas 12.3 114

Aboriginal or Torres Strait IslanderNo 13.7 1,773Yes 8.3 36

Highest level of educationBachelor or higher 18.2 280Year 12 17.1 420Year 11 or lower 10.9 1,107

Type of school attendedGovernment 13.5 1,395Catholic non-government 10.5 276Other non-government 20.7 135

Number of siblingsOnly child 21.7 69One or more siblings 13.1 1,737

Father’s occupation at age 14Managerial and administrativeb 9.8 318Professional 13.7 190Associate professionals 16.2 198Tradespersons and related 14.7 415Advanced clerical and service 13.3 5Intermediate clerical, sales and service 13.7 131Intermediate transport and production 13.2 257Elementary clerical, sales and service 12.1 58Labourers and related 11.4 158No occupation 23.2 69

Mother’s occupation at age 14Managerial and administrativec 6.5 62Professional 18.9 169Associate professionals 17.8 101Tradespersons and related 12.9 116Advanced clerical and service 11.5 104

Which Women Stop at One Child in Australia? 7

Table 2 (continued)

Variables Percentage with one child N

Mother’s occupation (continued)Intermediate clerical, sales and service 16.8 197Intermediate transport and production 9.1 66Elementary clerical, sales and service 11.8 228Labourers and related 11.7 247No occupation 13.3 519

Intactness of family of origin at age 14Father absent or deceased 20.2 89Father not absent or deceased 13.2 1,720Mother absent or deceased 10.0 30Mother not absent or deceased 13.6 1,779

Age last birthday40–44 13.9 68345–49 15.5 58250–54 11.0 544

Age at first birthd

Below 20 5.5 29220–24 8.2 61625–29 12.6 53830–34 19.7 24135–39 46.8 10340+ 90.9 11

Marital status at first birthNot married 17.3 335Married 12.7 1,474

Place of first birth for migrantsOverseas 18.1 199Australia 20.3 325

Total 13.5 1,809

a United Kingdom, Ireland, New Zealand, Canada and United States of America.b 61 per cent of fathers in ‘managerial and administrative’ occupations were ‘farmers or farm

managers’.c 89 per cent of mothers in ‘managerial and administrative’ occupations were ‘farmers or

farm managers’.d Age at first birth was computed by subtracting the age in years of the eldest child from age

in years of the respondent. The unrecorded residual months of age were assumed to follow uniform distributions.

Source: Living in Australia (HILDA) Survey Wave 1.

8 Nick Parr

Zealand, Canada and USA) and in Southeast Asia who stopped at one child also are considerably higher than for the Australia-born. The percentage of Aboriginal or Torres Strait Islander women who stopped at one child is only about two-thirds the national average.

There are also wide variations by the level and type of schooling. The percentage that stopped at one child is considerably lower for women who did not complete the final year of senior secondary school, Year 12 (or the overseas equivalent), than for women who completed it. However, the difference is slight between women who completed a Bachelor’s degree or higher and those for whom Year 12 was the high-est level of education. Differences by type of schooling may reflect religious values, financial situation and educational aspirations A much higher percentage of parous women who attended non-government non-Catholic schools (most of which are fully independent but which also include schools for other religious denominations and other religions) stop at one child compared to women who attended government schools or Catholic schools. The percentage of women who attended Catholic schools who stopped at one child is less than the percentage who attended government schools. In Australia Catholic schools generally offer fee subsidies for larger families. This would reduce the cost of additional children for those who aspire to send their children to these schools, and so may reduce the disincentive for higher fertility.

Table 2 shows that the size, socio-economic status and intactness of the family of origin are also related to the likelihood of a woman stopping at one child. Interest-ingly, women who themselves grew up as an only child are much more likely them-selves to have just one child: 22 per cent, compared with 13 per cent of those with one or more siblings. The variation between women with other numbers of siblings is slight. There are also wide variations by the father’s and the mother’s occupations. Women who at age 14 had a father in a ‘managerial or administrative’ occupation are the least likely to stop at one child, whilst women who at age 14 had a father with no recorded occupation are the most likely to do so. A majority, 61 per cent, of the fathers recorded as being in ‘managerial or administrative’ occupations were ‘farmers or farm managers’, and the percentage of the daughters of ‘farmers or farm managers’ that stopped at one child is particularly low, seven per cent. Thus the low percentage of the daughters of ‘managerial and administrative’ fathers that stop at one child would have been affected by the higher percentage that had a rural upbringing.

The daughters of women who had no occupation are only slightly more likely to have stopped childbearing at one child than the daughters of women with an occupation. However, there is wide variation between different types of maternal occupation. The divergence of the patterns for the two groups at the ‘high end’ of the occupational spectrum is striking: women with a mother in a ‘managerial or admin-istrative’ occupation are by far the least likely to have stopped at one child, whilst women whose mother was in a professional occupation are the most likely to have done so. The vast majority, 89 per cent, of the mothers in the ‘managerial and admin-istrative’ occupational category were ‘farmers and farm managers’, and the percent-age of the daughters of ‘farmers and farm managers’ who stopped at one child (six per cent) was particularly low. Thus rural upbringing appears to be a major explana-tory factor behind the low overall percentage in this category who stopped at one child. Women whose father was either absent from the parental home or deceased when they were aged 14 are about one-and-a-half times as likely as women whose father was present to have stopped childbearing at one child.

Which Women Stop at One Child in Australia? 9

There are wide variations in the percentages of women who ceased childbear-ing at one child by their characteristics at the time of the first birth. The relationship between the percentage of women who stopped at one child and the age at birth of the first child is particularly strong. The percentage that stopped at one child rises with increasing steepness as age at first birth increases. Almost all the women who first gave birth above the age of 40, and almost half the women who first gave birth between 35 and 39, stopped at one child. The percentage that stopped at one child is higher among women who were not married at the time of their first birth than for women who were married at the time of the first birth; no women were separated or widowed at the time of the first birth. There is little difference in the percentage that stopped at one child between migrants who had their first child outside Australia and those who had their first child within Australia.

Multivariate analysisTable 3 presents two models of the likelihood that a woman with one child did not progress to a second child. Model 1 includes only the early lifecourse variables. Model 2 adds controls for the effects of variables measured at the time of the first birth.

The largest early-lifecourse effects are those for birthplace. Having been born in East Asia greatly increases the likelihood of stopping at one child as opposed to Aus-tralian-born women. This is despite all the women in this group having been married at least once, over 90 per cent of the women in this group still being legally married, fewer than five per cent being unmarried at the time of the first birth, and the ages of marriage and first birth tending to be younger than average. Since half the women in this group were born in China, the one-child policy could be an explanatory factor. China-born women who had their first birth in China, approximately 80 per cent of all the China-born, are more likely to have stopped childbearing after one child than their counterparts who had their first child in Australia. The use of sterilization is widespread in China (Ross 1992); in some cases it may be due to coercive pressures (Short, Ma and Yu 2000). However, it is possible that the considerably increased like-lihood of China-born women stopping at one child also reflects a preference for this family size both in the context of China and in Australia. In traditional Chinese cul-ture the importance of a first child for continuation of the blood line and the family name is especially strong and this may lead to a selection into parenthood of couples who do not place such a high importance on other utilities of children (Fong 2004). In urban China a heavy parental investment in only children to promote their educa-tional success and cultural cultivation through a wide array of outside-school tuition and other expenditures has become widespread (Fong 2004; Nie and Wyman 2005). Consequently these parents perceive additional children as being very expensive. The pursuit of similar ambitions for their first (and only) child in Australia may have conditioned the attitudes of the China-born in Australia towards a preference for a one-child family.

It is also interesting to note that, as is the case for their China-born counterparts, the propensities of women born in Hong Kong and Taiwan to stop at one child also are very high, even though neither of these regions has been subject to the one-child policy. It may be that spousal separations caused by the male partner working in Hong Kong, Shanghai or elsewhere in East or Southeast Asia whilst the female part-ner and child are in Australia (so-called ‘astronaut families’) partly explains their increased likelihood of stopping at one child (Hugo 2006). Most Hong Kong-born

10 Nick Parr

Table 3 Logistic regression models of a woman stopping at one child

Variables

Model 1 Model 2

Coefficient SE Coefficient SE

BirthplaceAustralia (ref.) 0.00 0.00Main English-speaking overseas countriesa 0.50** 0.21 0.42+ 0.24Northern, Western or Eastern Europe 1.00*** 0.27 1.02*** 0.30East Asia 2.00*** 0.47 2.29*** 0.54South or Southeast Asia 0.42 0.31 0.33 0.36Other overseas 0.04 0.31 0.30 0.35

Aboriginal or Torres Srait IslanderNo (ref.) 0.00 0.00Yes –0.34 0.63 –0.07 0.64

Highest level of educationBachelor or higher 0.41** 0.21 –0.03 0.23Year 12 0.36** 0.18 0.17 0.19Year 11 or lower (ref.) 0.00 0.00

Type of school attendedGovernment (ref.) 0.00 0.00Catholic non-government –0.22 0.22 –0.24 0.24Other non-government 0.52** 0.24 0.49+ 0.26

Number of siblingsOnly child 0.44 0.32 0.46 0.34One or more siblings (ref.) 0.00 0.00

Father’s occupation at age 14Managerial and administrativeb –0.81+ 0.42 –0.89** 0.44Professional –0.68 0.43 –0.75+ 0.46Associate professionals 0.33 0.42 –0.42 0.44Tradespersons and related –0.26 0.39 –0.33 0.41Advanced clerical and service –0.44 0.86 –0.49 0.90Intermediate clerical, sales and service –0.41 0.46 –0.46 0.48Intermediate transport and production –0.32 0.41 –0.37 0.43Elementary clerical, sales and service –0.28 0.55 –0.42 0.59Labourers and related –0.48 0.44 –0.78 0.47No occupation (ref.) 0.00 0.00

Mother’s occupation at age 14Managerial and administrativec –0.49 0.56 –0.52 0.62Professional 0.31 0.26 0.27 0.27Associate professionals 0.21 0.31 0.09 0.33Tradespersons and related 0.03 0.32 0.10 0.33Advanced clerical and service –0.12 0.35 –0.18 0.37Intermediate clerical, sales and service 0.26 0.25 0.21 0.27

Which Women Stop at One Child in Australia? 11

Table 3 (continued)

Variables

Model 1 Model 2

Coefficient SE Coefficient SE

Mother’s occupation (continued)Intermediate transport and production –0.57 0.47 –0.49 0.49Elementary clerical, sales and service –0.09 0.25 –0.07 0.27Labourers and related –0.20 0.25 –0.15 0.27No occupation (ref.) 0.00 0.00

Father absent or deceased at age 14No (ref.) 0.00 0.00Yes 0.38 0.33 0.37 0.35

Mother absent or deceased at age 14No (ref.) 0.00 0.00Yes –0.62 0.66 –0.42 0.67

Age last birthday40–44 (ref.) 0.00 0.0045–49 0.15 0.17 0.26 0.1850–54 –0.19 0.19 0.11 0.20

Age at first birth Under 20 (ref.) 0.0020–24 0.41 0.3125–29 0.94** 0.3330–34 1.54*** 0.3435–39 2.69*** 0.3640+ 4.78*** 0.83

Marital status at time of first birthNot married 0.88*** 0.21Married (ref.) 0.00

First birth outside AustraliaNo (ref.) 0.00Yes 0.02 0.28

Constant –1.85*** 0.41 –3.00*** 0.56–2 Log likelihood 1345.0 1207.6

*** p < 0.001, ** p < 0.01, * p < 0.05, + p < 0.10.a United Kingdom, Ireland, New Zealand, Canada and United States of America.b 61 per cent of fathers in ‘managerial and administrative’ occupations were ‘farmers or farm

managers’.c 89 per cent of mothers in ‘managerial and administrative’ occupations were ‘farmers or

farm managers’.d Age at first birth was computed by subtracting the age in years of the eldest child from age

in years of the respondent. The unrecorded residual months of age were assumed to follow uniform distributions.

Source: Living in Australia (HILDA) Survey Wave 1.

12 Nick Parr

and Taiwan-born migrants enter Australia under skills migration schemes (DIMIA 2002). The high percentages of Hong Kong-born and Taiwan-born women in Aus-tralia who work in professional occupations and a resultant greater cost of children in terms of childcare or forgone earnings, together with smaller means-tested family benefits, may also contribute to the explanation (Parr and Guo 2005). So too may strong parental aspirations for the educational success of their children and a result-ant restriction of family size to maximize the resources per child (Parr and Mok 1995). For all groups of migrants the location of grandparents, particularly grandmothers, in the origin country may make childrearing a more arduous prospect, thus discour-aging parity progression. This may be a particularly important consideration for East Asian women in view of cultural traditions which place a greater importance on grandmaternal child care. So too would the high costs in Australia of employing domestic workers, who are widely used by the middle classes to help to look after children in Hong Kong and Taiwan (Ho 2006).

Having been born in Northern, Western or Eastern Europe or in the main English-speaking countries for migrants to Australia also significantly increases the likeli-hood of stopping at one child relative to the Australia-born. For about two-thirds of the Northern, Western or Eastern Europe-born women and for three-quarters of the women born in the main English-speaking countries, the HILDA data show that the first birth occurred in Australia, and so their family formation is largely explicable in terms of their circumstances in Australia. The data also show that women born in Northern or Western Europe or the main English-speaking countries tend to be less traditional in their family formation patterns, that is, they are less likely to marry, more likely to cohabit unmarried, more likely to divorce or separate and more likely to be unmarried at the time of the first birth. They are also relatively likely to leave marriage and the birth of the first child to later in life. Thus union instability and infecundity, the so-called ‘biological clock’, appear to be contributory factors in their increased likelihood of stopping at one child. Women from these regions are relatively likely to participate in the labour force and to attain relatively high occupational sta-tus and incomes in Australia. Thus it may be that the greater indirect and childcare costs of additional children help to explain their higher propensity to stop at one. However, after the addition of the ‘at first birth’ variables to the model, women from these birthplaces still have a greater likelihood of stopping childbearing at one child than Australia-born women. The explanation of the increased likelihood of Eastern European women stopping at one child may be somewhat different. The largest sub-group is those born in the former Yugoslavia, and so the disruption to family life of the Balkans war of the early 1990s may be a factor. Fertility rates fell sharply across Eastern Europe following the downfall of the communist regimes (Sobotka 2003). The factors which gave rise to this, according to Sobotka (2003), include high unem-ployment, economic and social instability, and the increased pursuit of consumerist lifestyles; these factors may have contributed to the high percentages of women born in this region and now living in Australia who stopped at one child. The greater likelihood of stopping at one child among migrants than among the Australia-born may reflect the disruption of childbearing and, in some cases, of unions by the act of migration and the need of the migrants to establish themselves in the Australian labour market after their migration (Abbasi-Shavazi and McDonald 2000).

Higher levels of education significantly increase the likelihood of a woman ceas-ing childbearing at one child. The positive effects on the likelihood of stopping at one

Which Women Stop at One Child in Australia? 13

child of the completion of Year 12 or a Bachelors or higher degree appear to be linked to the later ages at first birth of the more educated women, since after controlling for age at first birth they become much smaller and are no longer significant (Table 3). Thus for the more highly educated women the progression to a second child is less likely largely because they are more likely either to plan for a later first birth and a small family size or because they, whether voluntarily or involuntarily, leave the first birth to such a late age that their family building is thwarted by infecundity. Higher levels of education increase a woman’s likelihood of participating in the labour force and entering a higher earning occupation, and this may affect the family size to which she aspires. The greater financial disincentive for having an additional child result-ing from the greater prospective loss of earnings in order to look after the additional child or children may be a factor (Breusch and Gray 2004). Alternatively, it may be the prospective greater expenditure on formal (non-parental) childcare and a lesser receipt of means-tested benefits and childcare subsidies, rather than forgone income, which defines the financial disincentive faced (Craig 2005). A time squeeze resulting from the combination of work and the demands of looking after additional children may be another factor, and one which may be felt more acutely by more educated women. More educated women tend to value the development of their children’s human capital more highly (Craig 2006). According to Craig (2005) mothers who work longer hours, who include disproportionate numbers of the more highly edu-cated, sacrifice time for housework, personal care, and child-free leisure to preserve time spent with children. Thus a greater loss of such time would be a part of the greater ‘opportunity cost’ of childbearing faced by the more educated. The greater likelihood of stopping at one child among the more highly educated may also reflect their more readily embracing family structures and sizes which are less traditional in Australia, including having just one child.

The likelihood of a woman stopping at one child is also increased by attendance at a non-government, non-Catholic school. This is despite the higher incomes and wealth and, hence, a capacity to afford more children in this group (Parr 2006). It may be that women who attended non-government, non-Catholic schools are more likely to aspire to ‘higher quality’ children and, perhaps, to send every one of their children to such schools. Thus, even though discounted fees for siblings of existing children may be offered by some such schools, they may perceive additional children as more costly (Blake 1979; Parr 2006). It is likely a higher percentage of those attending non-government non-Catholic schools were boarders who had been separated from their siblings for extensive periods. Thus these women’s early life experiences may have led them to see the provision of a sibling for their child as less important.

Table 3 shows that having a father in a ‘managerial or administrative’ occupation reduces the likelihood of a woman ceasing her childbearing at one child. The size of the effect is changed only slightly by the introduction of variables measured at first birth. The effect of a father in a ‘managerial or administrative’ occupation may reflect the daughters being better able to afford to have more children owing to the greater wealth they attain (Parr 2006). It would also reflect that a high percentage of the fathers in such occupations were ‘farmers or farm managers’, and hence a higher proportion of these women were brought up in rural or semirural locations. In rural areas of Australia fertility rates tend to be higher and the disruption of unions by divorce or separation less likely than in urban areas (Carmichael and McDonald 2003). After controlling for the other variables, women whose fathers had no recorded

14 Nick Parr

occupations had the greatest likelihood of stopping at one child. This may reflect the adverse effect of the loss of the father’s income on the living standards of the family, and hence a reduced ability to afford having more than one child (Parr 2006). The lack of significance of the effect of the father being absent or deceased would reflect not only the small number of women in this category but also the significant correla-tion (0.46) with the ‘no occupation’ category.

After controlling for the other early lifecourse variables, the differences between birth cohorts in the likelihood of a woman ceasing childbearing at one child are not significant. This reflects some of the differences between birth cohorts, shown in Table 2, being accounted for by differences in the levels and types of education of the different cohorts, particularly the lower levels of education and the smaller percent-age educated in non-government non-Catholic schools of the 50–54-year-olds.

Women who were not married at the time of the first birth are considerably more likely to stop at one child than women who were married at the time of first birth, after controlling for the other variables. This reflects the greater fragility of unmar-ried relationships, both cohabiting and non-cohabiting, compared to married rela-tionships, and either the discontinuation of childbearing directly because of the relationship breakup or a reduced willingness to have subsequent children because of a lack of commitment from the partner, or fears that the relationship in the future may break up (Liefbroer and Dourleijn 2006). The effects of age at the birth of the first child on a woman’s likelihood of stopping at one child are highly significant and increase steeply as the age increases. This may reflect the reduced remaining reproductive lifespan and fecundity with increased age at first birth. Aspirations for a smaller completed family size among women who first give birth at a later age may be another factor. The correlation of age at first birth and the likelihood of stopping at one child may also reflect both variables being affected by fecundity (before the first birth), the effectiveness of use of contraception, and attitudes to abortion.

Discussion

The increase in the propensity of Australian women not to progress from a first birth to a second or beyond during the 1990s may be seen as a part of the intercon-nected array of changes in fertility and family formation which has been dubbed the ‘Second Demographic Transition’ (Lesthaeghe 1995; Van de Kaa 1997; Carmichael 1998). This paper has found individual-level links between the likelihood of stopping childbearing at one child with the postponement of childbearing, with having a pre-marital first birth and with prolonged female participation in education. During the course of the ‘Second Demographic Transition’ marriage and first births are increas-ingly postponed and cohabitation, premarital childbearing, and marital dissolution become more prevalent. A rising status of women and a secularization of society tend to accompany these demographic changes (Lesthaeghe 1995). Thus, whilst nei-ther Lesthaeghe (1995) nor Van de Kaa (1997) explicitly mentions it, for Australia at least, a reduced progression from a first to a second child should therefore be seen as part-and-parcel of this transition. It seems plausible that the growing reproduc-tive individualism associated with the ‘Second Demographic Transition’ may also have encouraged more couples to deviate from the two-child norm by choosing to have only one child, and the emergence of more effective contraception (the ‘Second

Which Women Stop at One Child in Australia? 15

Contraceptive Revolution’) has facilitated this choice (Lesthaeghe 1995; Van De Kaa 1997).

The cumulative effect over generations of parental investment in their children ‘getting ahead’ (as Marjoribanks 2002 terms it) may also help to explain the increased propensity of Australian women not to progress to a second child. Only children gain higher than average educational achievements, income and wealth (Marks 2006; Parr 2006). This reflects the time, energy and material resources which their parents can give to them being undiluted by the competing demands of siblings (Blake 1989). According to a survey by Weston et al. (2004) roughly a fifth of both male and female Australians aged 20 to 39 cited ‘whether having another child would reduce the opportunities available to other children’ as an important factor influencing their having children. This paper has also shown that more-educated women, and women who were educated in non-Catholic non-government schools, are less likely to progress from a first to a second child. Thus some of the increase over time in the propensity of Australian women to stop at one child may be linked to past increases in the retention of females in education and to past increases in the percentage of female school students who attend non-government non-Catholic schools, a trend which reflects past increases in parents’ willingness to invest in the human capital development of their female children (ABS 2003).

Immigration too has contributed to the increased propensity of Australian women to stop childbearing at one child. The rate of immigration to Australia is one of the highest in the world. This study shows that migrant women generally have a higher propensity to stop childbearing at one child than do the Australia-born. Moreover, the four largest country-of-birth groups of new migrants to Australia during the 1990s, New Zealand, United Kingdom, Hong Kong and China, all have relatively high propensities to stop childbearing at one child (DIMA 2001, 2006). The former Yugoslavia and a number of Southeast Asian countries were also significant sources of migrants to Australia in the 1990s which also have relatively high propensities to stop at one child. Of the various migrant groups, the very high propensity of East Asia-born women, especially the China-born, not to progress to a second birth is par-ticularly interesting. What would happen to fertility in China in the event of a relaxa-tion or even a removal of the one-child policy is arguably one of the most important demographic questions not only for China but also for the world. The fertility pat-terns of Chinese emigrants living in Australia (and in other overseas countries) may offer a ‘test case’ for the extent to which the one-child family may be expected to continue in China following the removal of some or all of the elements of the one-child policy. Further research is needed to establish whether this reflects one child as a preferred family size for this group even in Australia, whether it is a legacy of steri-lization in China, or whether it is explicable by other circumstances, such as social or economic barriers, aspirations for child educational success, or a lack of extended family support. The relatively high prevalence of one-child families among migrants should facilitate the upward social mobility of the second generation in Australia (Parr 2006).

Whilst the trend up until 2001 was of a reduction in the propensity of Australian women to progress from a first to a second birth, a subsequent continuation of this trend cannot be presumed. Progression to a second birth would logically have been encouraged by changes to family-related benefits4, improvements in the availabil-ity of parental leave and family-friendly working arrangements, strong economic

16 Nick Parr

growth, and the apparent strengthening of a ‘halo effect’ associated with parenting (Baird 2005; Heard 2006; ABS 2007). The introduction of the new maternity benefit, provided its effect is not cancelled out by the inflation of the cost of childcare, may logically be expected to have a greater effect on progression beyond a first birth than it does on progression to a first birth, because it compares more favourably to the (lesser) additional costs of second and higher-order children. Future waves of HILDA should facilitate the monitoring and analysis of such trends.

Acknowledgments

An earlier version of this paper, entitled ‘Who Stops at One: Evidence from Aus-tralia’ was presented to the 2006 Biennial Conference of the Australian Population Association held in Adelaide, South Australia 5–8 December 2006 and to the 2007 Annual Meeting of the Population Association of America held in New York, USA 29–31 March 2007.

This paper uses confidentialized unit record files from the Household, Income and Labour Dynamics in Australia (HILDA) survey. The HILDA Project was initi-ated and is funded by the Commonwealth Department of Family and Community Services (FaCS) and is managed by the Melbourne Institute of Applied Economic and Social Research (MIAESR). The findings and views reported in this paper, however, are those of the author and should not be attributed to either FaCS or the MIAESR.

The author wishes to acknowledge the helpful comments of three anonymous referees.

Notes1 Women who had completed a ‘certificate’ or ‘diploma’ were evenly spread over the edu-

cation categories. An investigation of this group of women as a separate category found small and insignificant results.

2 Missing or unknown occupation of parents is classified as ‘no occupation’.3 Computed by subtracting the age in years of the eldest child from age in years of the

respondent and assuming that the unrecorded residual months of age follow uniform dis-tributions.

4 Including, with effect from 1 July 2004, the introduction and staged increase over time of the new universal, flat-rate maternity benefit and increases to family tax benefits.

ReferencesAbbasi-Shavazi, M.J. and P. McDonald. 2000. Fertility and multiculturalism: immigrant fertil-

ity in Australia 1977–1991. International Migration Review 34(1): 215–242.Australian Bureau of Statistics (ABS). 1997. Australian Standard Classification of Occupations

(ASCO) Second Edition. Catalogue Number 1220.0. Canberra. Australian Bureau of Statistics (ABS). 2003. Schools. Catalogue Number 4221.0. Canberra.Australian Bureau of Statistics (ABS). 2006a. Births 2005. Catalogue Number 3301.0. Can-

berra.Australian Bureau of Statistics (ABS). 2006b. Australian Historical Statistics. Catalogue Number

3105.0.65.001. Canberra.Australian Bureau of Statistics (ABS). 2006c. Marriages 2005. Catalogue Number 3306.0.55.001.

Canberra.

Which Women Stop at One Child in Australia? 17

Australian Bureau of Statistics (ABS). 2007. Australia at a Glance. Catalogue Number 1309.0. Canberra.

Avdeev, A. 2001. How fertility has fallen in Russia and the reasons for the fall. Paper presented to the IUSSP Seminar on International Perspectives on Low Fertility: Ttrends, Theories and Policies, Tokyo, 21–23 March.

Baird, M. 2005. Who’s rocking the cradle? Making the Link 16: 32–38.Blake, J. 1979. Is zero preferred? American attitudes towards childlessness in the 1970s. Journal

of Marriage and the Family 41(May): 245–257.Blake, J. 1989. Family Size and Achievement. Los Angeles: University of California Press.Bongaarts, J. and R.G. Potter. 1983. Fertility, Biology and Behavior. New York: Academic Press. Breusch, T. and E. Gray. 2004. New estimates of mothers’ forgone earnings using HILDA data.

Australian Journal of Labour Economics 7(2): 125–150. Carmichael. G.A. 1998. Things ain’t what they used to be! Demography, mental cohorts, moral-

ity and values in Post-War Australia. Journal of the Australian Population Association 15(2): 91–114.

Carmichael, G.A. and P. McDonald. 2003. Fertility trends and differentials. Pp. 40–76 in S-E. Khoo and P. McDonald (eds), The Transformation of Australia’s Population �970–2030. Syd-ney: UNSW Press.

Carmichael, G.A. and A. Whittaker. 2007. Choice and circumstance: qualitative insights into contemporary childlessness in Australia. European Journal of Population 23(2):111–143.

Chapman, B., Y. Dunlop, M. Gray, A. Liu and D. Mitchell. 2001. The impact of children on the lifetime earnings of Australian women: evidence from the 1990s. Australian Economic Review 34(4): 373–389.

Craig, L. 2003. The time cost of children: a cross-national comparison of the interaction between time use and fertility rate. Paper presented to the 25th IATUR Conference on Time Use Research ‘Comparing Time’, Brussels, 17–19 September.

Craig, L. 2005. Where Do They Find the Time? Social Policy Research Centre Discussion Paper No 136. Sydney: UNSW.

Craig, L. 2006. Parental education, time in paid work and time with children: an Australian time-diary analysis. British Journal of Sociology 57(4):553–575.

Craig, L. and M. Bittman. 2003. The time costs of children in Australia. Paper presented to Rethinking Expenditures on Children: Towards an International Research Agenda, ANU, Canberra, 15–16 January.

Department of Immigration and Multicultural Affairs (DIMA). 2001. Immigration: Federation to Century’s End. Canberra.

Department of Immigration and Multicultural Affairs (DIMA). 2006. Immigration Update. Can-berra.

Department of Immigration and Multicultural and Indigenous Affairs (DIMIA). 2002. Austral-ian Immigration Consolidated Statistics. Canberra.

De Vaus, D., L. Qu and R. Weston. 2005. The disappearing link between premarital cohabitation and subsequent marital stability 1971–2001. Journal of Population Research 22(2): 99–118.

Dodson, L. 2004. The mother of all spending sprees. Sydney Morning Herald 12 May: 1.Fong, V.L. 2004. Only Hope: Coming of Age Under China’s One-Child Policy. Stanford: Stanford

University Press.Greenhalgh, S. 2003. Science, modernity and the making of China’s one-child policy. Population

and Development Review 29(2): 163–201.Heard, G. 2006. Pronatalism under Howard. People and Place 14(3): 12–24.Henman, P., R. Percival, A. Harding and M. Gray. 2007. Costs of children: research commis-

sioned by the Ministerial Taskforce on Child Support. Occasional Paper No. 18. Canberra: Department of Families, Community Services and Indigenous Affairs.

Ho, C. 2006. Migration as feminization? Chinese women’s experiences of work and family in Australia. Journal of Ethnic and Migration Studies 32(3): 497–514.

18 Nick Parr

Hugo, G.J. 2006. Immigration responses to global change in Asia: a review. Geographical Research 44(2): 155–172.

Kiernan, K.E. and A.J. Cherlin. 1999. Parental divorce and partnership dissolution in adult-hood: evidence from a British cohort study. Population Studies 53(1): 39–48.

Kippen, R. 2004. Declines in first and second birth rates and their effects on levels of fertility. People and Place 12(1): 28–37.

Kippen, R. 2006. The rise of the older mother. People and Place 14(3): 1–11.Lesthaeghe, R. 1995. The Second Demographic Transition in Western countries: an interpreta-

tion. Pp. 17–62 in K.O. Mason and A. Jensen (eds), Gender and Family Change in Industrialized Countries. Oxford: Clarendon Press.

Liefbroer, A.C. and E. Dourleijn. 2006. Unmarried cohabitation and union stability. Demogra-phy 43(2): 203–221.

Marjoribanks, K. 2002. Family and School Capital: Towards a Context Theory of School Outcomes. London: Kluwer Academic Publishers.

Marks, G.N. 2006. Family size, family type and student achievement: cross-national differ-ences. Journal of Comparative Family Studies 37(1): 1–24.

McDonald, P. 1998. Contemporary fertility patterns in Australia: first data from the 1996 cen-sus. People and Place 6(1): 1–12.

McDonald, P. 2001. Family support policy in Australia: the need for a paradigm shift. People and Place 9(2): 14–20.

McDonald, P. 2002. Low fertility: unifying the theory and the demography. Paper presented to the 2002 Annual Meeting of the Population Association of America, Atlanta, 9–11 May.

Merlo, R. and D. Rowland. 2000. The prevalence of childlessness in Australia. People and Place 8(2): 21–32.

Murphy, M. and L.B. Knudsen. 2002. The intergenerational transmission of fertility in contem-porary Denmark: the effects of number of siblings (full and half), birth order, and whether male or female. Population Studies 56(3): 235–248.

Newman, L.A. and G.J. Hugo. 2006. Women’s fertility, religion and education in a low- fertility population: evidence from South Australia. Journal of Population Research 23(1): 41–66.

Nie, Y. and R.J. Wyman. 2005. The one-child policy in Shanghai: acceptance and internaliza-tion. Population and Development Review 31(2): 313–338.

Parr, N. 2005. Family background, schooling and childlessness in Australia. Journal of Biosocial Science 37(2): 229–243.

Parr, N. 2006. Do children from small families do better? Journal of Population Research 24(1): 1–25.

Parr, N. and F. Guo. 2005. The occupational concentration and mobility of Asian immigrants in Australia. Asian and Pacific Migration Review 14(3): 351–380.

Parr, N. and M. Mok. 1995. Differences in the educational achievements, aspirations and val-ues of birthplace groups in New South Wales. People and Place 3(2): 1–8.

Percival, R. and A. Harding. 2002. All they need is love and around $450,000. The AMP-NAT-SEM Income and Wealth Report Issue 3. Australia: AMP <http://www.amp.com.au/au/3column/0,2338,CH5306%255FSI56,00.html>. Accessed 25 January 2007.

Ross, J.A. 1992. Sterilization: past, present, future. Studies in Family Planning 23(3): 187–198.Santow, G. 1991. Trends in contraception and sterilization in Australia. Australian and New

Zealand Journal of Obstetrics and Gynaecology 31(3): 201–208.Short, S.E., L. Ma and W. Yu. 2000. Birth planning and sterilization in China. Population Studies

54(3): 279–291.Sobotka, T. 2003. Re-emerging diversity: rapid fertility changes in Central and Eastern Europe

after the collapse of the Communist regimes. Population (English Edition) 58(4-5): 451–486.United Nations Population Division. 2006. World Population Policies 2005. New York.Van de Kaa, D. 1997. Options and sequences: Europe’s demographic patterns. Journal of the

Australian Population Association 14(1): 1–30.

Which Women Stop at One Child in Australia? 19

Watson, N. and M. Wooden 2002a. The Household, Income and Labour Dynamics in Australia (HILDA) Survey: Wave 1 survey methodology. Hilda Project Technical Paper Series No 1/2. <http://www.melbourneinstitute.com/hilda/>. Accessed 1 May 2007.

Watson, N. and M. Wooden 2002b. Assessing the quality of the HILDA Survey Wave 1 Data. Hilda Project Technical Paper Series No. 4/02. <http://www.melbourneinstitute.com/hilda/>. Accessed 20 August 2007.

Weston, R. and L. Qu. 2001. Men’s and women’s reasons for not having children. Family Mat-ters 58 (Autumn): 10–15.

Weston, R., L. Qu, R. Parker and M. Alexander. 2004. It’s Not for Lack of Wanting Kids …: A Report on the Fertility Decision Making Project. Research Report No. 11. Melbourne: Aus-tralian Institute of Family Studies.