Selectivity in higher-order childbearing in Sweden

In this note, we present relative risks of giving birth for mothers with different numbers of children. We use Swedish register data and study the propensity to continue the childbearing beyond child number two. We pay special attention to births of those higher orders that are studied only rarely in conventional demographic analyses. As the parity increases to higher numbers, we expect to find some kind of selection so that the group of mothers increasingly consists of very birth prone women. For births after a fourth child, we indeed find such an effect in that the relative risk of giving birth to an additional child then increases with the birth order. In our intensity-regression models, we also check whether this selection effect of increasing birth risks can be picked up by the inclusion of a specific factor for unobserved heterogeneity in the mothers’ propensity to give birth. We find that the positive gradient in the propensity to give birth indeed disappears when such a factor is included into our model.