Structural Transformation, Capital-Skill Complementarity, and the Endogenous Fertility Decline

Fertility in the rich world has fallen below the level needed to sustain a population, and in the United States it has been falling for nearly two decades without recovering. Most economic accounts of this decline take as given the very things that need explaining, namely how much women earn relative to men, how costly it has become to raise a child, and how attractive the alternatives to raising one now are. This paper builds a general equilibrium model in which all three emerge together from a single familiar engine, the long migration of economic activity out of manufacturing and into services. As an economy becomes a service economy, where women hold a comparative advantage and where the fast-growing forms of capital are most useful, women's wages rise toward men's and an hour of a parent's time comes to command ever more of what the market offers. Both forces raise the real cost of children even as they make couples more equal. Parents respond by having fewer children, and a growing share have none. The model reproduces not only the falling average but the actual distribution of family sizes, including the large share of families that stop at one child, a pattern simpler models cannot generate. New cross-country evidence supports the mechanism where the model says it should operate. Within countries, cohorts of women whose childbearing years passed in a more service-intensive economy, facing a dearer relative price of services and a narrower gender wage gap, completed fewer births, and the parity distribution shifted toward one-child families. Left alone, and where child-rearing cannot be purchased on the market, the decline does not settle at a lower but stable level. It continues, gradually, toward ever fewer children, so that below- replacement fertility is the early and visible stage of a much longer contraction rather than a temporary dip. The paper then asks what offsetting this force would require, focusing on publicly funded childcare. Because institutional care draws on the same advancing technology that makes parental time expensive, neutralizing the time cost of children is fiscally feasible in the model, at a bounded share of output. Feasibility is not suffciency, however, and the cross-country record is sobering, since the economies that have socialized childcare most are today at record-low fertility. Sustained below-replacement fertility is a predictable consequence of the structure of a service economy and does not reverse on its own, and the most-discussed policy lever offsets only part of that structure.