An estimable model of illegal Mexican immigration

This dissertation analyzes the migration decisions of illegal immigrants from Mexico to the US. I develop a dynamic general equilibrium overlapping generations model of migration decisions in a village economy characterized by local labor markets. Each period an agent located in Mexico receives a job offer from the US with some probability that depends on his previous experience in the US and the number of his fellow villagers already in the US, and decides whether to migrate or stay in Mexico. The model allows for multiple spells of migration, up to a terminal age when the agent retires to Mexico. The wage in Mexico is derived in equilibrium depending on the size of the village population residing in Mexico. Recent US migration policy has focused on increasing the cost of crossing the border to discourage illegal migration from Mexico. This type of policy does not necessarily imply lower migration once the probability of migrating depends on the size of the community of migrants in the US. Moreover a policy trying to discourage migration will affect the opportunity in the origin, through the labor market. My model is able to quantify these effects and to compare the above policy with an alternative policy aimed to reduce the wage of migrants through tighter controls on illegal immigrants employers. The data used to estimate the model comes from the Mexican Migration Project (MMP), a joint effort by the University of Pennsylvania and the University of Guadalajara, and is gathered in Mexico during Christmas time. This type of data is especially suited to minimize the selection problem. The model is parameterized and structurally estimated. Estimation is by the Method of Simulated Moments, matching sample moments with moments produced by model simulation. For each evaluation of the objective function it is necessary to calculate the model equilibrium. The model is estimated on a selected sample of Mexican villages, and simulations show that the pattern of migration observed in the data are well replicated. Using the estimated parameters, I evaluate the effect on migration flows of different types of immigration policies. In particular the increase in the cost of crossing the border does not seem to have a big impact on the stock of illegal Mexican immigrants in the U.S