A new theory of the economic determinants of fertility.

Abstract

My new theory of the economic determinants of fertility is that greater human capital induces parents to have fewer, better-educated children due primarily to the way in which children's human capital is produced within the household. I posit that parents' human capital shifts the production function for children's human capital in an asymmetrical way, inducing higher-skill parents to choose lower numbers of children. In this model, higher wages (holding human capital constant) induce higher fertility, but higher human capital (holding wages constant) induces lower fertility, both at an individual household level, and in the aggregate, in cross-section and over time. This theory is consistent with prior, apparently paradoxical, findings that fertility increases with income in time-series data but fertility declines with income in cross-sectional data, if the primary source of variation in time-series data is variation in the wage per unit of effective labor and the primary source of variation in cross-sectional data is variation in human capital. I also present new estimates that offer support for the theory. Using international panel data, I find that fertility tends to decline in lagged secondary school enrollment, and rise with GDP, in a variety of specifications. In US Census data, including specifications using instrumental variables, I find that fertility tends to rise with income when controlling for education, and fall with education controlling for income. All of the estimated effects are very small, and often the confidence intervals are quite large, but I argue that even small effects can be quite important when effects are compounded over generations out into the future. Subtle differences in the modeling of fertility can produce large changes in projections of future per-capita consumption levels, or important policy evaluations, such as the solvency of Social Security.