Changing patterns of relationship formation:
implications for fertility

A complex array of interacting factors may affect an individual’s intentions about having children. The below-replacement level and still declining fertility rate of the last few decades has led to a great deal of attention being directed to the disincentives. Children are expensive to raise — an issue that is particularly likely to be important for those with limited financial resources. In addition, having a child often results in diminished income and career opportunities for women, along with difficulties in balancing work and family life, and the curtailment of freedom generally. Other reasons for childlessness include dislike of children and concerns about the responsibilities of parenting (Barnes 2001; McDonald 2000; Weston & Qu 2001).

While for some people childlessness is involuntary, contemporary childlessness is typically seen as voluntary. However, practical circumstances including difficulties in finding a suitable partner or maintaining a secure relationship, may hinder chances of having children. Thus, the distinction between voluntary and involuntary childlessness is at times difficult to make (Rowland 1998; Merlo and Rowland 2000).

Although the majority of people still get married, marriage rates have declined. Those who now enter into marriage do so later in life, often having lived in a sexual union with their partner before marrying (here called "cohabiting") (ABS 2001a). While the increasing prevalence of cohabitation has been accompanied by an increase in the proportion of births that are ex-nuptial and "paternity-acknowledged", most couples wait until they are married before they have children (ABS 1997a, 2001b). However, cohabiting relationships that end in separation can severely diminish the chances of some women ever having children.

In this paper, links between relationship formation, stability and fertility are explored through two studies. The first examines pathways to couple formation and the stability of relationships following different pathways (Qu & Weston forthcoming), while the second examines the apparent impact of changing personal relationships on intentions about having children (Qu, Weston & Kilmartin 2000).

I. Pathways to couple formation

For simplicity, the first set of analysis focuses on women. It examines how women’s first union started (whether it began with cohabitation or marriage) and how old they were when they entered their first union. In addition, it assesses the level of stability of unions that began with marriage compared with those that began with cohabitation, and the proportions of women born in different years who had a child within four years of the start of their first union.

Life course survey

The analysis of women’s family formation pathways is based on the Australian Life Course Survey, conducted by the Australian Institute of Family Studies in 1996. This study involved telephone interviews with a nationally representative sample of nearly 2700 respondents (around 1150 men and 1540 women) aged 25 to 70 years. Respondents provided a history of the pathways they had taken in family formation, beginning with the first time they had lived with a partner (here referred to as the "first relationship" or "first union"). This history included the dates of any periods of cohabitation, marriage, birth of children, and separation. In this study it was left to the respondents to define the start of their living together relationship.

Given the relatively small number of women in the sample who reported the birth of a child outside cohabitation or marriage (35 women), the different pathways followed by this group were not assessed. In order to identify changes in family formation trends, the final sample was divided into four groups: women born between 1967—1971, 1962—1966, 1957—1961, and 1952—1956 (N = 835). At the time of the survey, the oldest group was 40—44 years of age, while the youngest was 25—29 years.

The oldest and youngest groups had grown up in periods markedly different from each other. The youngest group of women, born in 1967-1971, typically began their first relationship in late 1980s to early 1990s. First relationships for the second youngest women, born in 1962-1966, mostly occurred in the mid 1980s, while the first relationships of the other two groups, born in 1957-1961 and 1951-1956, typically occurred in the late 1970s to early 1980s, and in the mid to late 1970s respectively.

The decline in the manufacturing industry beginning in the late seventies led to a dramatic fall in the availability of permanent full-time jobs for early school leavers. The restructuring of the economy resulted in the increased demand for skilled workforce and an increasing number of young people participating in further education after completing high school. As a result, young adults have increasingly remained financially dependent on their parents, regardless of their living arrangements. While continued participation in education and the increasing prevalence of premarital cohabitation have often been cited as reasons for the increase in age at marriage (for example, ABS 1997b, 2001a), little is known about whether or not cohabiting relationships have also been delayed.

Pathways to family formation

Four sets of analysis were conducted: (a) the timing of the first living together relationship, (b) whether the union began with cohabitation or marriage, (c) what happened after cohabitation — the proportion who married or separated within four years and who had a child within this period; and (d) what happened to those who married at the outset — the proportion who separated or had a child.

The timing of the first living-together relationship

The timing of the first union was similar for the two oldest groups. They tended to enter their first union at an earlier age compared with the two youngest groups. For example, 59 to 63 per cent in the oldest two groups had entered their first union by age 22, compared with 45 to 51 per cent of the youngest two groups. However, by age 27, most women had started their first union. (Some women in the youngest group were under the age of 27 at the time of the survey.)

Cohabitation or marriage?

Figure 1 (a—d) shows the changing trends in the way women enter relationships: the more recently born the respondents, the more likely they were to start with cohabitation. For instance, by the time they were 25 years old, women in the youngest group were more than twice as likely to have started their union with cohabitation rather than with marriage. The exact opposite pattern is evident for the oldest age group: by age 25, these women were more than twice as likely to have begun their first union with marriage rather than cohabitation.

How cohabitation ends

For most couples, cohabitation was a temporary situation, ending in either separation or marriage. Figure 2 follows the pathways of couples who began their relationship with cohabitation. The figure indicates the proportions who had separated or married within four years of the start of this relationship.

It appears that marriage following cohabitation still remains a common pathway, but has become less so in recent times. Within four years from the start of their cohabiting relationship, around two-thirds of the oldest group and around one third of the youngest group had married. On the other hand, separation following cohabitation is becoming more common. Within four years, separation had occurred for less than 10 per cent in the oldest group and around 33 per cent in the youngest group. Overall, marriage and separation seemed equally likely to occur among the youngest group whereas marriage was more common than separation for the oldest women.

Ex-nuptial births

Consistent with ABS data, the Institute’s Australian Life Course data suggest that having children outside marriage is increasing but still applies to a minority. Within four years of cohabitation, around one in five women in the youngest group and only one in ten in the oldest group gave birth to a child while cohabiting. These trends appear to coincide with the growing acceptance of ex-nuptial births among young people: de Vaus (1997) found that only 30 per cent of people in their 20s disapproved of having children without marrying compared with 75 per cent of those aged 60—69 and 82 per cent of those aged 70 or older.

Married women’s pathways

Only a small number of women in the youngest group began their first living-together relationship with marriage, so trends for this group are not discussed. Compared with women who began their relationship with cohabitation, those who married at the outset were less likely to have separated and more likely to have had a child (Figure 3). No more than eight per cent in each group indicated that they had separated within four years of their marriage, while 62 to 67 per cent had given birth to a child within this period.

While there was little difference in the proportion of women in each group who had given birth to a child within four years of marriage, further analysis indicated that childbearing for younger women occurred later: within the first two years of marriage, 43 per cent in the oldest group compared to 25 per cent in the youngest group had given birth to a child.

What does "cohabitation" mean today?

The circumstances surrounding cohabitation can be diverse. Couples may decide to cohabit early in their relationship when first "going steady", or they may embark on cohabitation as a trial marriage or for practical reasons after having decided to marry. Others may opt for cohabitation, expecting to marry if and when they want to have children, while other committed, cohabiting couples may view marriage as redundant. Still others may decide to cohabit in an attempt to avoid all commitment — a "no strings attached" strategy. Some partners in couples may agree about the meaning of cohabitation, while others may hold quite different views, and the meaning for each partner may change during the course of cohabitation.

The present analysis suggests that the meaning of cohabitation may have changed. Cohabitation was predominantly a prelude to marriage for women born in the early 1950s and appears to have become progressively less so for younger generations. By contrast, childbearing while cohabiting was more prevalent for the youngest group.

At the same time, the chances of cohabitation ending in separation rather than marriage have increased in more recent times. Possibly, young couples today may be more likely to cohabit at an earlier stage in their relationship, when "going steady", rather than when considering marriage. Alternatively, young couples may now embark on cohabitation as a "trial marriage" but hold higher expectations about having their needs fulfilled in the relationship and be more prepared to separate if their needs are not met.

Summary

While the proportion of children born outside marriage has increased, most couples delay childbirth until they are married. However, most first unions now start with cohabitation rather than marriage. The Australian Life Course Survey suggests that these first unions are more likely to end in separation than in the past. Furthermore, first unions generally are beginning later in life. As will be seen below, delayed and disrupted relationships can have a profound impact on the chances of having children.

II. Relationship status changes and fertility intentions

Surveys of fertility intentions suggest that Australian women, like their counterparts in most European countries, typically have fewer children than they intended having when they were young. Furthermore, it appears likely that many young women today who want to have children will never realise this ambition (see Barnes 2001; McDonald 1998).

While childless people appear to be better educated and more career oriented (Cameron 1990; Rovi 1994), the gap in the education level between those with and without children has narrowed, with childlessness becoming more widespread across different educational levels and labour force status groups (ABS 1999).

A recent study in the United Kingdom suggests that the decision to remain childless is a complex process, which takes place in the context of one's work, life experiences, personal health and relationships (McAllister and Clarke 1998). Intentions about having children can change in the light of evolving circumstances, values and aspirations. For example, some couples who decide to delay childbearing may eventually give up all ideas of having children.

Nevertheless, Schoen, Astone, Kim & Nathanson (1999) concluded that child-bearing intentions are remarkably persistent, and are thus strong predictors of outcomes. These authors also acknowledged, however, that the level of apparent stability diminishes as the interval between measured intentions and outcomes increases. According to Rovi (1994), the stability of intentions varies according to their nature: people who do not want children are less likely to change their minds than those who want children.

As Bracher and Santow (1991) have pointed out, assessment of fertility intentions or desires represents an appropriate starting point for examining steps in decision making regarding family formation and may shed light on the causes of fertility decline. In particular, it may be possible to examine factors that lead fertility outcomes to match or differ from fertility desires or intentions.

While financial and non financial costs of having children may lead women and men to have few if any children, it seems very likely that intentions about having children are also linked to expectations about relationship status. Those men and women who are already in an apparently stable relationship are likely to hold intentions based on an expectation that their relationship will continue. Some people who do not have a partner may intend to have children after they have found a suitable partner with whom to live. However, some relationships end before children are born and some single people do not find partners. Yet, as Merlo and Rowland (2000) pointed out, the impact on fertility decisions of external constraints such as relationship breakdown and the inability to find a suitable partner has not been explored adequately. These issues were explored in a follow-up study undertaken by the Australian Institute of Family Studies.

The Australian Family Formation Project

This analysis is based on data the Australian Family Formation Project, a national longitudinal conducted by the Australian Institute of Family Studies in 1981 (when respondents were 18 to 34 years old) and in 1990. Fifty-eight per cent of the original sample of 2,500 respondents were traced and agreed to participate in the second survey. Those who remained in the study tended to be better educated than those who did not participate, but these two groups varied little in terms of age and gender.

Attention is here directed to 783 respondents who did not have children in the initial survey (425 men and 358 women representing 52 per cent of the total sample who were re-interviewed in the second survey). Respondents were omitted from the analysis if they reported being a homosexual or that they did not intent having any children for medical reasons (N=9). Throughout this analysis, the results for men and women were combined because their patterns were very similar. While much of the literature on fertility focuses on women’s decisions, men’s views will also influence the decision to have children or not. They, too, can be concerned about financial and non-financial costs of raising children, and their views and those their partners appear to have much the same weight in influencing outcomes (Thomson, 1997).

The two issues

The following analysis focuses on intentions and outcomes regarding having children, covering nearly a decade. Two issues are examined: the relative stability of intentions to have or not have children, and the extent to which relationship status and changes in relationship status over the period influenced intentions and outcomes.

The stability of Intentions and fertility outcomes

Table 1 shows respondents’ intentions in wave 1 (1981) and outcomes and intentions in wave 2 (1990). In Wave 1, most respondents said they intended having children (89 per cent), while 6 per cent said that they would not have children, and 5 per cent indicated uncertainty.

Not surprisingly, respondents who in wave 1 said that they intended to have children were more likely than other respondents to have done so by the second wave. Fifty-seven per cent of those who wanted to have children had fulfilled their ambitions by this time. On the other hand, only a quarter of those who did not want any children or who were unsure about these matters in wave 1 reported having children by wave 2.

Only six per cent of those who had earlier intended to become parents subsequently changed their minds and about eight per cent became unsure, while just over half (52%) of those who intended to remain childless indicated the same intention almost a decade later. It is worth noting, however, that by wave 2, 42 per cent in this group either had children or intended to have children.

These results suggest that intentions to have children were fairly reliable or persistent over nearly 10 years, with the vast majority either having had, or still intending to have, children. However, this study would not have detected situations involving several changes in decision making during this interval.

As might be expected, uncertainty appeared to be the least stable of the three options: of those who felt unsure about having any children in 1981, only 13 per cent remained unsure, while 33 per cent had decided not to have children by wave 2. More than half the respondents who indicated uncertainty about having children either had children by the second wave or said they intended to do so in the future.

Although, as noted above, Rovi (1994) argued that those who do not intend to have children (called ‘negative intentions’) are less likely to change their minds than those who intend to have children (called ‘positive intentions’), the results in Table 1 suggest an opposite trend. Positive intentions appear to be more stable over time and more consistent with outcomes, compared with negative intentions.

Intentions and relationship status

As noted earlier, little attention has been paid to the impact on fertility decisions of relationship break-up or failure to find a partner. To examine this issue, the sample was divided into four basic groups. Two groups had experienced no change in relationship status, with one group remaining single over the entire period (N=131), and the other group living with the same partner continuously (either married or cohabiting) (N = 160). The other two groups experienced change in relationship status, with one group being single in wave 1 and living with a partner in wave 2 (N = 356), and the other group living with a partner in wave 1 only (N = 22). For simplicity, those who followed other pathways were omitted from this analysis (for example, the 33 respondents who had different partners in wave 1 and wave 2, and the 64 respondents who were single in both waves but had partnered at least once between waves).

Each of the four basic groups was then sub-divided further according to whether or not members indicated in wave 1 that they definitely intended having children. Because of the small numbers in some sub-groups, the 'undecided' were combined with those who did not intend having children for this analysis. As indicated above, these two groups were similar in terms of the proportion who, by wave 2, had children or did not intend having children.

Figures 4 and 5 respectively refer to those who with and without definite intentions to have children at the time of the first survey. It is noteworthy that the general pattern in Table 1 regarding the stability of ‘positive intentions’ held for each relationship status group represented in Figure 4. That is, regardless of stability or change in relationship status, the majority of respondents who in wave 1 said they would have children had, by the second wave, either achieved their ambition or still intended doing so.

Those who intended having children

Despite the general consistency between intentions and outcomes for those who, in wave 1, intended having children, Figure 4 suggests that relationship status contribute to outcomes. The most likely to have had children were those with the same partner in both waves (89%), followed by those who had found a partner by wave 2 (68%), then those who had separated (42%). Not surprisingly, very few who were continuously single reported having children by wave 2 (3%).

Those who had separated from their partner were the most likely to have changed their minds about having children (32%), followed by the continuously single (11%). Of all the four groups who intended having children, the most likely to express uncertainty about this matter were the continuously single (23%). Only 3 to 5 per cent in the other three groups indicated that they were uncertain about having children.

By wave 2, respondents aged 35 or more years who were continuously single were more likely than their younger counterparts to have decided against having children (27% compared with 3%), while around 22 percent in each age group had become uncertain about this matter. Such trends suggest that that changes in fertility intentions often reflect the process of adjustment to circumstances. Their apparent childlessness may not be as ‘voluntary’ as might appear on the basis of their expressed intentions. By contrast, age did not appear to affect intentions of those with new or continuous partners. The majority either had children or still intended to become parents.

Those with no clear intentions to have children

As noted above, a number of respondents who in wave 1 either did not intend having children or felt unsure about this matter had already had children by wave 2 or intended doing so. Those who had found a partner by wave 2 were the most likely to have changed their minds, with 56 per cent either having had a child or intending to have a child.

The extent to which their changed status or pressure from their new partners had led to a change in intentions is not known. However, the most likely to have had a child were those in a continuous relationship (42 per cent). Just over half in this group and in the continuously single group and only one third who had found a partner indicated in wave 2 that they did not intend having children.

Summary

The Australian Family Formation Project suggests that intentions to have children tend to be strong predictors of outcomes. Most respondents intended having children in wave 1, and regardless of their relationship status, most who held these intentions either had children by wave 2 or still intended to do so.

However, some people changed their intentions and such changes were often linked with personal relationships. Those who had separated from their partners were the most likely to change their minds and decide against having children, followed by those who were continuously single. Nonetheless, this change of mind for the continuously single appeared to become more entrenched with age. Such trends may well reflect respondents’ adjustment to the fact that time was running out. In this case, there appears to be a blurring of 'voluntary' and 'involuntary' childlessness.

Intentions not to have children were less stable than intentions to have children, and were more likely to be reversed by single respondents who subsequently partnered than by respondents who remained with the same partner over the decade. The extent to which such changes arose from enhanced opportunities to have children or home pressures from the new partner is unclear.

Conclusion

Childlessness applies to a heterogeneous group and its reasons can be complex. The inability to find a suitable partner or the experience of relationship breakdown can prevent individuals from fulfilling their ambitions to have children. The apparent increasing tendency for first unions to begin at a later age, in tandem with the apparent increased fragility of cohabitation as a family form, seems to be partly responsible for the falling fertility rate. Such practical circumstances can blur the distinction between voluntary and involuntary childlessness.

Thus, intervention strategies that successfully enhance the quality of relationships thereby preventing separation may have the added advantage of helping couples who want a child or who want to have more children to achieve their aims.

References

ABS (1997a), Australian Demographic Trends, Australian Bureau of Statistics, Catalogue No. 3102.0, Canberra.

ABS (1997b), Social Trends, Catalogue No. 4102.0, Canberra.

ABS (1998), Social Trends, Catalogue No. 4102.0, Canberra.

ABS (1999), Births Australia 1998, Cat no. 3301.0.

ABS (2001a), Marriages and Divorces, Australia 2000, Australian Bureau of Statistics, Catalogue No. 3310.0, Canberra.

ABS (2001b), Births, Australia 2000, Australian Bureau of Statistics, Catalogue No. 3301.0, Canberra.

Barnes A (2001), Low fertility: a discussion paper. Canberra, ACT: Department of Family and Community Services (Occasional Paper no. 2).

Bracher, M., & Santow, G. (1991), ‘Fertility desires and fertility outcomes’, Journal of the Australian Population Association, vol. 8, no.1, pp. 33—49.

Cameron J. (1990), 'Why have children? A New Zealand case study', University of Canterbury press.

de Vaus, D. (1997), "Family values in the nineties", Family Matters, no 48, pp. 5-10.

McAllister F. with Clarke L. (1998), 'A study of childlessness in Britain', Family Policy Studies Center, Joseph Rowntree Foundation.

McDonald, P. (1998), 'Contemporary fertility patterns in Australia: First data from the 1996 Census', People and Place, vol. 6, no. 1, pp.1-12

McDonald, P. (2000), ‘Low fertility in Australia: Evidence, causes and policy responses’, People and Place, vol.8, no. 2, pp. 6-21.

Merlo R. & Rowland, D. (2000), 'The prevalence of childlessness in Australia', People and Place, vol. 8, no. 2, pp.21-32

Qu, L., Weston, R. & Kilmartin, C. (2000), "Effects of changing personal relationships on decisions about having children", Family Matters, No. 57, pp. 14—19.

Qu, L., and Weston, R. (forthcoming), ‘Starting out together through cohabitation or marriage’. Family Matters no.60.

Rovi S.D. (1994), 'Taking 'no' for an answer: using negative reproductive intentions to study the childless/childfree', Population Research and Policy Review, vol. 13, pp. 343-365

Sarantakos, S. (1996), Modern Families, Macmillan Education Australia, Melbourne.

Schoen, R., Astone, N.M., Kim, Y.J., & Nathanson, C.A. (1999), 'Do fertility intentions affect fertility behaviour?', Journal of Marriage and the Family, Vol 61 (August), pp. 790-799

Thomson E. (1997), 'Couple childbearing desires, intentions, and births', Demography, vol. 34 no. 3, August 1997, pp. 343-354.

Weston, R. and Qu, L. (2001), "Men's and women's reasons for not having children", Family Matters No.58 Autumn 2001, pp. 10-15.

Australian Institute of Family Studies, Level 20, 485 La Trobe Street, Melbourne Vic 3000, Australia.
Tel: (03) 9214 7888. Fax: (03) 9214 7839. URL: http://www.aifs.gov.au/

Copyright, privacy, disclaimer | Site map