Family labour supply and fertility: a two-regime model.

The central hypothesis of the "new micro-economics of fertility" was examined, i.e., because children are female-time intensive, fertility is inversely related to the mother's wage. A theoretical model of lifetime family labor supply and completed fertility which forms the basis for the subsequent empirical investigation is presented. Empirical estimates of the model are provided using 1971 Canadian census data. The 1-period model of completed family size and lifetime labor supply indicates that fertility and labor supply behavior of families where the wife participates in the labor market will differ from that of families with never-participating wives. The systematic differences arise because families with nonworking wives face the additional binding constraint of an inflexible supply of wife's time to the home. A comparison of these 2 regimes allows inferences concerning the female-time intensity of children and the substitutability of husband's and wife's time. In the context of the 1-period model, the relevant criterion for separating families into the 2 regimes is the lifetime rather than current participation status. The evidence suggests that when this procedure is followed, much doubt is cast on the conventional wisdom that children are female-time intensive. The usual inverse relation found between wife's wage or education and completed family size appears to be the result of a complex interaction between wage rates, lifetime labor force participation, and fertility. Some support is found for complementarity between the home time of the husband and wife, but the evidence is not totally consistent. Results of the female participation equation strongly suggest complementarity; proxies for the husband's lifetime labor supply suggest weak or zero complementarity. When the sample is split by current participation of the wife rather than lifetime participation, then the results support conventional wisdom concerning the female-time intensity of children and replicate Kneisner's finding of complementarity between husband and wife's time. Reconciling these differences within the framework of the 1-period model is difficult since it yields no predictions concerning current variables. Yet, since the wife's lifetime participation is the correct criterion for allocating families to the 2 regimes under the 1-period model, doubt must remain concerning the female-time intensity of children until the current lifetime results are reconciled.