The aim of this project is to empirically examine how changes in labor markets for young workers have affected their marriage, fertility, and housing decisions. The 1980s and 1990s have witnessed significant changes in the age of first marriage and the percentage of young households that own a house. Understanding these developments requires an understanding of how career developments for young individuals affect these other significant life decisions. There is a growing literature which documents that the 1980s involved a dramatic increase in income inequality in the United States. It is argued that declining real income prospects for workers at the outset of their careers do not appear to fully explain the new trends in marriage, fertility, and housing decisions. For example, using Census of Population data it is shown that the declines in home ownership for young men between 1980 and 1990 occurred across the income and education spectrum. To reconcile this, other broad changes in labor market outcomes are examined. This research will add to the emerging literature which argues that earnings variability increased in the 1980s. The trends in earnings variability are documented and analyzed using panel data from the National Longitudinal Surveys and matched Current Population Surveys. The recent investment literature emphasizes the importance of the option value of delay for investments that involve irreversibilities and must be made in an uncertain economic environment. A general prediction is that the option value of delay increases with the degree of uncertainty regarding the costs/payoffs from the investment. These option based models apply equally to an individual's decisions regarding marriage, fertility, and housing. All of these decisions involve significant transactions costs and other forms of irreversibilities. Increasing variability of labor market earnings increase the option value of delaying these decisions. To date, little empirical work has been carried out to test this prediction. This project will take an important step in the direction of filling in this gap. The research will also significantly advance the general statistical modeling used to study housing tenure decisions. Most of the existing empirical literature on housing tenure decisions is estimated with cross sectional data. It is argued that using panel data and transition models provides a much tighter integration between economic theory and measurement. These models can be simulated to provide meaningful estimates of the stock of homeowners at any given age. It is shown that for variables such as the user cost of housing which vary considerably over time, the traditional estimation methods lead to significantly biased inferences. It is also shown how the econometric framework can be extended to handle multiple housing transitions, and allowing household formation, marriage, and fertility decisions to be endogenous. u»
