Dissertation

Fertility and social interaction: an economic approach

Abstract

Researchers are still in disagreement about one of the most profound questions of demographic change: Why fertility rates have fallen significantly in many societies and not in others, and why fertility transitions have been precipitous in some cases, and leisurely in others. Concepts of `diffusion' and `culture' are frequently invoked to explain these differences, but neither their theoretical underpinnings nor their dynamic implications are well understood. Economic fertility models can investigate these effects through explicit consideration of social interaction and positive externalities. The latter occur in the diffusion of information, the transformation of social norms and with respect to health or human capital investments. Couples' fertility choices are hence interrelated, and social interaction emerges as an important determinant for the onset, pace and pattern of fertility decline. Existing economic theories of fertility do not consider these effects adequately. They usually focus on individual decision making and disregard the path-dependent evolution of `local' environments. This dissertation attempts to overcome these limitations. It develops formal tools for the study of `socially embedded' fertility behavior, it provides microeconomic foundations for the `diffusion' of low fertility, and it investigates reproductive decisions interdependently with the transformation of social and economic institutions. The first analysis focuses on the dynamics of word-of-mouth communication. Women are uncertain about the merits of contraceptives, and they evaluate different methods through conversations with early adopters in their social network. This ability to learn from friends and neighbors is confounded due to unobserved heterogeneity in tastes and characteristics. Women therefore act as `econometricians' to infer the relative benefits of available methods. They decide on the basis of these estimates and their private information about personal characteristics. The analysis then examines how the long-term implications of social learning depend on the information revealed in conversations. Empirical studies of European or contemporary fertility declines often attribute the salient diversity of contraceptive patterns across communities and social strata to social learning. The theoretical analysis shows that this finding arises from `imprecise' conversations which focus on contraceptive choices rather than individuals' satisfaction. If communication is more informative, then diversity does not persist. Communities converge to an identical pattern of contraceptive use that differs insignificantly from the case of fully informed decisions. A second analysis investigates `social influence' in a broader perspective. Positive externalities induce multiple equilibria representing a `traditional' high fertility and a `modern' low fertility situation. The selection between these equilibria constitutes a coordination problem. The onset and speed of a transition do not only depend on historical conditions, but also on social interaction and community cohesion. Horizontal diffusion of low fertility occurs because innovators increase the incentives for low fertility in the remaining population. In this context, social networks accelerate the speed of a fertility transition because they inform parents about the behaviors and intentions of other couples. If social interaction facilitates a collective action within communities, it can also initiate a fertility transition that would otherwise be delayed or inhibited by the inertia of traditional environments. Moreover, social networks increase the effectiveness of media programs. The analysis hence provides an explanation why the most striking instances of sudden fertility transitions have occurred in comparatively small geographic areas with homogeneous populations. It also reveals when social forerunners in fertility decline constitute a nucleus for diffusion, and when separate demographic regimes across subpopulations are likely to prevail. (AUTHOR)