This project would bring together in the same model a number of dimensions of behavior that heretofore have been analyzed separately. The behavioral model we propose to construct and estimate includes the following features: 1. Saving and retirement are jointly explained. 2. The model is structural, including separately the preferences and constraints guiding individual behavior. 3. The analysis is dynamic, following retirement outcomes from full time to part time work and/or retirement, and following saving over the life cycle. 4. People are forward looking in their decision making, although time preference is distributed heterogeneously. This means that the response rates to future rewards from wages or from postponing Social Security or pension claiming differ among the population. It also means, consistent with the distribution of wealth, that some will be well prepared for retirement and others poorly prepared. 5. The analysis is stochastic. People may reduce their work effort overtime, then subsequently increase it. 6. The saving and retirement decisions of married couples are interdependent. Outcomes are jointly determined in the context of a bargaining model with limited cooperation, and are influenced by elements in each spouse's opportunity set as well by the interaction of preferences. 7. Different measures of retirement will be used, blending self reports of retirement status and of hours. 8. Many jobs typically held in the prime working years and paying higher wages do not allow partial retirement, or require full time work or none at all. The resulting minimum hours constraints are incorporated in the analysis, with partial retirement usually outside the main job. 9. The wage offer will depend on tenure and on hours of work. 10. Linked data used in calculating the opportunity set include detailed pension plan descriptions obtained from employers and Social Security records. 11. Saving and retirement are modeled on the assumption that one cannot borrow on future income or Social Security. Thus liquidity constraints are incorporated in the analysis. 12. Layoffs are incorporated in the analysis. The econometric analysis will be based on the Health and Retirement Study. Various measures of health and health change will be incorporated into the analysis.
