Interrelationships between childbearing and housing transitions in the family life course

Kulu, H. and Steele, F. (2013) Interrelationships between childbearing and housing transitions in the family life course. Demography (forthcoming)

Abstract

Research has examined the effect of family changes on housing transitions and childbearing patterns within various housing types. While most research has investigated how an event in one domain of family life depends on the current state in another domain, the interplay between them has been little studied. This study examines the interrelationships between childbearing decisions and housing transitions. We use rich longitudinal register data from Finland and apply multilevel event-history analysis to allow for multiple births and housing changes over the life course. We investigate the timing of fertility decisions and housing choices with respect to each other. We model childbearing and housing transitions jointly to control for time-invariant unobserved characteristics of women which may simultaneously influence their fertility behaviour and housing choices, and we show how joint modelling leads us to a deeper understanding of the interplay between the two domains of family life.

Further details

Housing transitions are closely interlinked with other life course events such as union formation and dissolution, births of children, and changes in employment. Research on the interplay between family events and housing transitions has received particular attention because of its importance in advancing our understanding of how couples plan their family lives. While the availability of longitudinal data has greatly enhanced our ability to disentangle the relationships between life course events, the study of these relationships is complicated because different event histories are likely to be influenced by a common set of factors, some of which may be unobserved. Previous research on the relationship between family and housing events has focused on either the effects of family changes on housing transitions or the impacts of housing changes on childbearing. However, considering only one side of the relationship ignores the possibility that family and housing events share unmeasured influences, potentially leading to biased estimates of the effect of one event on another.

In this paper, we examine the interrelationships between childbearing and housing changes. We extend previous research in the following ways. First, we use a simultaneous equations modelling approach to control for unobserved women’s characteristics (e. g. long-term plans and aspirations) that influence both fertility and housing transitions. Second, we model the risk of housing changes relative to the time since the birth of children and the hazard of births relative to the duration in the current house using flexible piecewise (linear) splines. This approach allows an investigation of how factors specific to current circumstances (e.g. short-term intentions and anticipation) influence fertility behaviour of couples and their housing choices. Finally, we also investigate whether and how the relative timing of childbearing and housing changes vary by context within a country. We contrast the patterns in large cities and those in small towns and rural areas. We are particularly interested in variation between settlements in the order of events, i.e. whether couples move in anticipation of or in response to a birth. Although both types of move may be seen as adjustment moves at first glance, an examination of the timing patterns by settlement may provide us with valuable information on how easy or difficult it is to adjust housing size to family size in various contexts.

We use register data from Finland for the following reasons. First, the large-scale longitudinal data allow a detailed examination of the timing of childbearing and housing transitions with respect to each other. Importantly, population data allow us to investigate whether the relationship between childbearing and residential mobility depends on the number and age of children, the current type of housing or place of residence; such fine-grain analyses would not be possible with smaller-scale survey data. Second, the Finnish register data allow us to study childbearing and housing transitions of both married and cohabiting couples, which is important in a context where childbearing in consensual unions is common; information on cohabiting couples is usually not available in population registers. Third, the data include detailed information on housing histories of couples including the exact date of housing changes. Information on date (to the nearest month) of moving is critical to investigate the timing of housing transitions relative to childbearing decisions. Finally, register data are (almost) immune to sources of selection bias such as attrition, item non-response and recall error.

The statistical model consists of four simultaneously estimated hazard equations: three for different moving to different types of house (single-family house, terrace and apartment) and one for conceptions leading to a live birth. The equations are linked by allowing for correlation between woman-specific random effects that represent unmeasured time-invariant influences on each event. Significant positive correlations were found between the residuals of the three housing equations and the residuals of the fertility equations, which indicates that women who were more likely to have children (net of their observed characteristics) were also more likely to move house whatever the destination type of housing. This suggests that their long-term family plans and housing aspirations were closely related: women who wished to have a large family were prone to move, possibly several times, in order to achieve their desired housing.

Our analysis showed that the birth of a child raised significantly the propensity of moving to a new house; further, the likelihood of moving to a single-family house increased with the number of children in the family. We also observed elevated fertility, particularly first and third birth rates, after a couple had moved to a new house, especially to a single-family house. Failure to account for unobserved characteristics of women which influenced both their childbearing and housing careers would have led us to overstate the effect of childbearing on housing-related moves and the fertility differences by housing type.

The analysis of the relative timing of childbearing and housing transitions revealed that many housing-related moves took place during pregnancy and that many children were conceived during the first months after the move to a new house. More precisely, three patterns of timing were observed: (i) a couple moved to a new house and the woman became pregnant after the move; (ii) the woman became pregnant, and the couple moved to a new house where the child was born; and (iii) a child was born and only then did the couple move to a new house.