Estimation of Discrete Stochastic Dynamic Programming Models of Economic Behavior Using Monte Carlo Integration
The goal of this research is to develop new methods to solve and estimate discrete stochastic dynamic programming (DS-DP) models, and use these to study decision-making in areas such as human capital investment, occupational choice, investment in health, and others. In recent years it has become common in economics to model individuals who are making choices in dynamic environments as if they were solving a DS-DP problem to determine their optimal decisions. But empirical implementation of such models has been hampered because their solution and estimation requires that very high order numerical integrations be performed. This group investigated the use of simulation methods to circumvent these integration problems.
The research involves two projects. The first is a model of investments in health and decisions to buy private health insurance. For instance, in the U.S., senior citizens are covered by Medicare, but many buy supplemental private insurance (“Medigap”) to cover things that Medicare leaves uncovered. In many other countries a similar situation exists, except that the analogue of Medicare provides partial insurance for the whole population. When completed, this model will be useful for predicting how changes in what is covered by Medicare would affect health and program costs. The second project is a model of retirement behavior, in which the researchers model investments in human capital, savings decisions, and work decisions over the whole life cycle. When completed, this model will be useful for predicting how changes in Social Security benefit rules will affect not only retirement decisions but also asset accumulation and human capital investment decisions over the whole life cycle.