The intergenerational transmission of first birth timing in Norway

Lars Dommermuth, Statistics Norway
Torkild Hovde Lyngstad, University of Oslo

The intergenerational transmission of family size and fertility timing between generations can play a role for fertility patterns, but the evidence for such transmission is still weak. The strong data requirements, i.e. detailed fertility histories of both the parents and their children, make it difficult to study this topic. Solid knowledge about any such effects may be of importance for development of policies, as the implications for future fertility are obvious. So far, studies on the intergenerational transmission of the timing of first birth only have used ordinary event history analysis. A common downside of these studies is that they are generally unclear about the status of causality. There are numerous other factors linked with both parents’ and children’s fertility behaviour that are not controlled, and which may in whole or in part generate the observed association and bias regression coefficients of the transmission effect. In our study we will address this problem in three steps and give new insight into the intergenerational transmission of first birth timing. We use Norwegian register data, including information from over 300.000 women born in 1954-64 and their mothers. We will first run ordinary Cox proportional hazard regressions to test for the existence of a transmission effect of timing of birth between mothers and their daughters. In a second step, we will include family fixed-effects to make within-family comparisons of how mother’s age at birth of daughter A affects daughter A’s age at first birth. The fixed-effects method allows us to control for all non-measured stable characteristics at the level of the family in a model of the transmission of timing of birth. In a third step we will examine how censoring and unequal censoring times within the family affect results in family fixed-effects models.