Abstract 2. Data and method of analysis

1. Introduction

Changing family structures together with greater instability of marriages in most industrialised countries require that we revise our traditional concept of families and, in particular, the concept of stepfamilies. Pre-marital childbearing and cohabitational unions are no longer the exception to families in the traditional sense. Childbearing is not restricted to a single-marriage union. In fact, it is not even restricted to a marital union at all. This is of great importance in relation to the concept of completed fertility over one's life course, since divorce and re-marriage typically occur during the main reproductive years. To take into account these changes in traditional family structures the term stepfamily, formerly restricted to marriages only, has been extended to include cohabitational unions involving a child of only one partner (cf. e.g. [Bumpass et al. 1995]). This extended definition of stepfamilies takes into account the fact that divorced or separated partnerships are often followed by a cohabitational union and not necessarily by a second or higher-order marriage.

Demographic developments in Austria during the past few decades provide evidence that an increasing number of men, women, and children experience the formation of a stepfamily.

Illegitimate (non-marital) fertility has become very pronounced in Austria. In 1998, 30 per cent of all children were born out of wedlock, as compared to 18 per cent in 1980 and 13 per cent in 1970 [Austrian Central Statistical Office 2000]. At the same time, the total divorce rate increased from 18 per cent in 1970 to 26 per cent in 1980 and reached a level of 39 per cent in 1998. This indicates that 39 per cent of all current marriages will sooner or later end in divorce if the duration-specific divorce rate observed in 1998 remains the same in the future. Among divorced couples in 1998, 34 per cent had no children, 30 per cent one child, 27 per cent two children, and nine per cent three or more children. About 63 per cent of the children were below the age of 14. Moreover, a cross-tabulation of marriages in Austria by the family status of each partner at the time of marriage in 1998 (Table 1.a) shows that in three out of ten marriages, at least one of the partners was divorced or widowed. This latter figure had risen from 20 per cent in 1961. Summing up, these trends provide clear evidence for the potential for stepfamilies in Austria (see also [Note1] and [Buber and Prskawetz 1999]).

A large number of studies in the European literature focus on the stability of stepfamilies and the consequences for the life course of children. Alternatively, one can take the stepfamily as one possible environment in which to investigate demographic events such as childbearing. This is of particular interest since childbearing behaviour in stepfamilies relates to childbearing of higher-order parity (at least for one partner), and it will depend in particular on whose parity progression ratio one considers (the female's or the male's). The question then arises "whether first children in new unions arrive at the same rate as first children in first unions, or at the relevant rate of children that are of higher order in a lifetime perspective, or at some other rate" [Vikat et al. 1999]. When studying stepfamily fertility, the focus has to be shifted from the couple's relationship to the broader family, which includes children already present at the time of union formation (see [Wineberg 1990]).

Our analysis contributes to the growing literature on stepfamily fertility, which up to now has been based mainly on American data (see the extensive literature review in [Vikat et al. 1999]). European studies are restricted mainly to four countries: Sweden [Vikat et al. 1999], Czechoslovakia [Kucera 1984], Germany [Heekerens 1986], and France [Toulemon 1997]. In line with these studies, we investigate the determinants of childbearing in stepfamilies by focusing on the effect of pre-union children on fertility in second unions (and more specifically, on the conception of the first shared child) [Note 2]. We disentangle the effects of pre-union children by their number, their place of residence (in the household vs. out of household) at the time of the formation of the second union, by distinguishing between a woman's and a man's pre-union children, and by the age of the youngest child at the time of the formation of the second union. Additionally, we also control for the history of the previous union, characteristics of each partner at the time of the formation of the second union, changes in characteristics of the second union as recorded each month during the observation since union formation, calendar time period, and time elapsed since the formation of the second union.

Since we focus on the conception of the couple's first shared child in a second union, we try to control for the characteristics of both partners. While information on the respondent's characteristics is generally rather complete in most surveys on union and birth histories, corresponding information for the respondent's partner is usually missing or incomplete. Though the Austrian Family and Fertility Survey, on which we base our analysis, is of exceptionally high quality in terms of information on the partner's union and birth histories, we nevertheless have to restrict part of our analysis to male or female respondents only.

As outlined in [Bulatao 1981], several studies have shown that the effect of a couple's characteristics on childbearing may vary with parity. Hence, if pre-union children are present, the effect of these common characteristics on the first shared child is complicated by the fact that this child might be of higher-order parity for at least one partner. To control for such parity-specific effects we shall test for interactions between variables related to the number of pre-union children and variables referring to the couple's characteristics.

While the findings about the influence of pre-union children on fertility in subsequent higher-order unions are uniform across the different studies in the literature (negative with increasing number of pre-union children), the results are more diverse concerning more detailed characteristics of pre-union children, such as their place of residence (in the household vs. out of household) and whether they are the woman's or the man's pre-union children. Since these differing results might well depend on the availability and quality of the data, the comparatively high quality of information on the partner's birth and union histories in Austria allows us to gain more insight into these important factors.

We structure our analysis by testing various hypotheses about the effects of pre-union children on the intensity of the conception of a first shared child in higher-order unions. More specifically, we refer to the relative importance a first shared child might have for unions where pre-union children are present as opposed to unions where no pre-union children are present (see [Griffith et al. 1985] and [Vikat et al. 1999]). As outlined in [Bulatao and Arnold 1977] (p. 141) "the values or satisfactions individuals expect to attain through having children" in general and through having children with the current partner, "as well as the disvalues or costs they expect to incur", influence the fertility pattern and fertility preferences. As "children of different parities may serve different functions for their parents and entail different material and non-material costs as well" ([Bulatao 1981], p. 23ff), the value of a first shared child in stepfamilies will depend on the number of pre-union children.

As outlined in the literature, a child might confer a union commitment and a parenthood effect. Since the birth of a first shared child "reflects the couple's commitment to each other" ([Vikat et al. 1999], p.213), the rate at which a first child arrives in a second union should be independent of the number of children born to either partner before the second union. The parenthood effect is more closely connected to the fertility history of each partner separately, i.e., to the importance of childbearing for the individual's status as opposed to the status of the marital union (see [Griffith et al. 1985], p.74). It should be present for those partners for which the first shared birth in a second union is also their first birth ever. For them the risk of a first birth in a second union should be independent of the number of pre-union children of their partner. If it is the case that stepchildren can be a substitute for biological children, then the parenthood effect may not be observed when pre-union children are present. The extent to which a stepchild is a substitute for one's own child certainly depends on emotional and personal relationships between the stepparent and the stepchild (see [Bulatao and Arnold 1977]).

Besides the union-commitment and parenthood effect, a first shared child in stepfamilies may also confer a sibling effect. To the extent that pre-union children will act as half-siblings for children in new unions and substitute for biological children, we can argue as follows. The strong two child norm that is present in Austria [Note 3] may imply that stepfamilies with only one pre-union child may exhibit a higher intensity to conceive a first shared child as opposed to stepfamilies with two or more pre-union children.

Whether pre-union children will act as half-siblings depends among other things on the age of the youngest pre-union child. Since the length of intervals between successive births may be prolonged if births occur in various unions (see [Griffith et al. 1985]), we might generally observe lower fertility in stepfamilies as a consequence. We expect couples, and more specifically mothers, of young pre-union children at the time of union formation to have higher intensities of conception of a first shared child in the new union. The underlying assumption is that long extensions of childbearing periods are avoided as mothers do not wish to re-enter the childbearing phase later for economic or career reasons. Hence, life course experiences may put some restrictions on continued childbearing in higher-order unions. As further noted in [Griffith et al. 1985] the age of the youngest pre-union child may have different effects, depending on the number of pre-union children.

The age of the youngest pre-union child may also influence the parenthood effect and thereby have the opposite effect of the sibling effect. The younger the pre-union child at the time of the formation of the new union, the more likely it might be accepted by the step-parent and possibly substitute for a shared biological child. We would therefore expect increasing intensities of the conception of a first shared child, the older the youngest pre-union child is at the time of union formation.

Summing up, our interest in this paper is to identify how the value of a first shared child changes in the presence of pre-union children of at least one partner. In particular, we test for the following values that a first shared child in higher-order unions may confer: the union-commitment effect, the parenthood effect, and the sibling effect. We test for these effects controlling for the number of pre-union children and the age of the youngest pre-union child. Since one or both partners may have already experienced childbearing, the change in the value of the first shared child may not be due only to the fact that (a) it is already a higher-order child for at least one partner, but also to the fact that (b) the values themselves have changed through the experience of earlier childbearing.


Abstract 2. Data and method of analysis

Fertility in second unions in Austria: Findings from the Austrian FFS
Isabella Buber, Alexia Prskawetz
© 2000 Max-Planck-Gesellschaft ISSN 1435-9871