Economic variables and the decision to have additional children: evidence from the Survey of Economic Opportunity.

Abstract

Data from the Office of Economic Opportunity's 1967 Survey of Economic Opportunity are analyzed. The full sample consisted of 5548 households, where the head is either black or white with wife between the ages of 14 and 64, and neither spouse married more than once. The dependent variable represents the outcome of the decision to have more children; the independent variables are income, value of the wife's time, wife's education, and residence. Subsamples based on wife's education and race of household were also analyzed. The primary finding is that value of wife's time and wife's education exert different effects on fertility across birth orders; hence use of wife's education as a proxy for value of wife's time can generate misleading results in birth order studies. The household's permanent income is found to have positive effects at lower birth orders and negative effects at higher ones, a confirmation of similar findings in several previous studies. The value of the wife's time has a negative effect at all birth orders in the full sample, as predicted by the opportunity-cost hypothesis, especially for households where the wife is white. Wife's education has both additive and interactive effects. In the full sample, it affects early birth orders positively and later ones negatively, apparently acting as a proxy for contraceptive efficiency, which reduces the variance around average family size. However, its effect is purely negative for white households. The effects of permanent income and value of wife's time are stronger for households with more-educated wives; moreover, the effect of wife's education across birth orders is less negative among households with more-educated wives, confirming Ben-Porath's hypothesis.