This project will investigate how fertility adjusts over the life cycle in response to the family's accumulated demographic events. A major part of the research effort will be devoted to the development and testing of new methodologies for the analysis of hazard rates. The main innovations over earlier methodologies are the use of a discrete time framework, the explicit incorporation of important biological determinants, and corrections for heterogeneity and endogeneity. Intensive use of Monte Carlo experiments during the development of the empirical model will greatly improve the interpretability of estimates. Although the methodological improvements will be of considerable value in many applications, this project will mainly be confined to a study of fertility transition in rural France. The hypothesis of natural fertility can be reduced to testable assertions about the internal dynamics of family-building, without reference to any variables outside the family's demographic history. The methods developed in this project are uniquely well-suited to making full use of fertility histories. Family reconstitution data, like that constructed for France by the Institut National d'Etudes Demographiques and used in previous work by the present investigators, provide abundant and comparatively complete data on fertility histories, generally without any additional information. In combination, the data and methods to be used here will produce powerful tests of the fertility hypothesis. When used in combination with other data sets linking fertility histories to other variables, the new methods can test more specific theories about the behavioral and biological determinants of fertility.
