The paper presents a tractable econometric model of optimal migration decisions, focusing on expected income as the main economic influence. The model improves on previous work by considering optimal sequences of location decisions and allowing for many alternative locations. It is estimated using panel data from the NLSY on white males with a high school education. The main conclusion is that interstate migration decisions are influenced by income prospects. The results suggest that the link between income and migration decisions is driven by geographic differences in mean wages and a tendency to move in search of a better locational match when income in the current location is unfavorable. The paper develops a model of optimal sequences of migration decisions, focusing on expected income as the main economic influence. Migration decisions are reversible, and many alternative locations must be considered. The model is estimated using panel data from the National Longitudinal Survey of Youth on white males with a high school education. Structural dynamic models of migration over many locations have not been estimated before, presumably because the required computations have not been feasible. A structural representation of the decision process is of interest for the usual reasons: we are ultimately interested in quantifying responses to income shocks or policy interventions not seen in the data, such as local labor demand shocks, or changes in welfare benefits. The paper demonstrates that a fully specified econometric model of optimal dynamic migration decisions is feasible, and that it is capable of matching the main features of the data, including repeat and return migration. Although this paper focuses on the relationship between income prospects and migration decisions at the start of the life cycle, suitably modified versions of the model can potentially be applied to a range of issues, such as the migration effects of interstate differences in welfare benefits, the effects of joint career concerns on household migration decisions, and the effects on retirement migration of interstate differences in tax laws. The paper develops a model of optimal search process for migration. The basic assumption is that wages are local prices of individual skill bundles. Individuals know the wage in their current location, but to determine the wage in another location, they must move there, incurring a moving cost. Locations are distinguished by known differences in wage distributions and amenity values. The model aims to describe the migration decisions of young workers in a stationary environment. The wage offer in each location may be interpreted as the best offer available in that location. Although there are transient fluctuations in wages, the only chance of getting a permanent wage gain is to move to a new location. The paper discusses the empirical implementation of the model. A serious limitation of the discrete dynamic programming method is that the number of states is typically large, even if the decision problem is relatively simple. The model, with J locations and n points of support for the wage distribution, has J(n+1)^J states, for each person, at each age. The paper discusses the wage, moving costs, state variables and flow payoffs, transition probabilities, identification, data, estimation, and empirical results. The paper concludes that interstateThe paper presents a tractable econometric model of optimal migration decisions, focusing on expected income as the main economic influence. The model improves on previous work by considering optimal sequences of location decisions and allowing for many alternative locations. It is estimated using panel data from the NLSY on white males with a high school education. The main conclusion is that interstate migration decisions are influenced by income prospects. The results suggest that the link between income and migration decisions is driven by geographic differences in mean wages and a tendency to move in search of a better locational match when income in the current location is unfavorable. The paper develops a model of optimal sequences of migration decisions, focusing on expected income as the main economic influence. Migration decisions are reversible, and many alternative locations must be considered. The model is estimated using panel data from the National Longitudinal Survey of Youth on white males with a high school education. Structural dynamic models of migration over many locations have not been estimated before, presumably because the required computations have not been feasible. A structural representation of the decision process is of interest for the usual reasons: we are ultimately interested in quantifying responses to income shocks or policy interventions not seen in the data, such as local labor demand shocks, or changes in welfare benefits. The paper demonstrates that a fully specified econometric model of optimal dynamic migration decisions is feasible, and that it is capable of matching the main features of the data, including repeat and return migration. Although this paper focuses on the relationship between income prospects and migration decisions at the start of the life cycle, suitably modified versions of the model can potentially be applied to a range of issues, such as the migration effects of interstate differences in welfare benefits, the effects of joint career concerns on household migration decisions, and the effects on retirement migration of interstate differences in tax laws. The paper develops a model of optimal search process for migration. The basic assumption is that wages are local prices of individual skill bundles. Individuals know the wage in their current location, but to determine the wage in another location, they must move there, incurring a moving cost. Locations are distinguished by known differences in wage distributions and amenity values. The model aims to describe the migration decisions of young workers in a stationary environment. The wage offer in each location may be interpreted as the best offer available in that location. Although there are transient fluctuations in wages, the only chance of getting a permanent wage gain is to move to a new location. The paper discusses the empirical implementation of the model. A serious limitation of the discrete dynamic programming method is that the number of states is typically large, even if the decision problem is relatively simple. The model, with J locations and n points of support for the wage distribution, has J(n+1)^J states, for each person, at each age. The paper discusses the wage, moving costs, state variables and flow payoffs, transition probabilities, identification, data, estimation, and empirical results. The paper concludes that interstate