Neighborhood effects on family size

Janna Bergsvik, Statistics Norway
Sara Cools, BI Norwegian Business School
Rannveig K. Hart, Statistics Norway and University of Oslo

Peer effects on fertility behavior have been found for various network types, such as friends, colleagues and siblings. While it is well known that fertility behavior is spatially clustered, knowledge of fertility contagion among neighbors remains scarce. We aim to fill this gap in the literature by estimating peer effects on family size among neighbors. To handle self-selection of women who intend to have large families into high fertility neighborhoods, we use the sex composition of the two eldest children of neighboring families as an Instrumental Variable for neighbors’ family size. Data on full residential and childbearing histories are drawn from Norwegian administrative registers. Individual neighborhoods are defined by way of time-varying GIS coordinates on place of residence. Our study sample consists of females aged 25 to 35 who gave birth to a second child in 2004 or 2005 (N ~ 30 000). Preliminary results indicate that the children sex mix instrument indeed creates exogenous variation across neighborhoods in the distribution of women with at least three children. We find no significant effects of neighbors’ family size on women’s propensity to have three children. This indicates that self-selection, rather than peer effects, drives the geographical clustering of fertility behavior.