In this paper I document the fact that the relationship between human capital, as measured by education, and migration choices among Mexicans is U-shaped: the highest and lowest educated tend to migrate more than the middle educated. I provide an explanation for the U-shaped relationship based on the interaction of two forces. On the one hand, there is a loss of human capital faced by emigrants, due to imperfect transferability, that is progressive with education and causes the negative relationship. On the other hand, the altruism towards future generations and the transmission of human capital from one generation to the next drives the positive relationship. I calibrate the model to match relevant moments from the Mexican and US Censuses, and use the calibrated model for policy evaluation. I evaluate the long run effect of the Progresa policy on education and migration. I show that, by giving a monetary contribution to poor families that send their children to school at lower grades, the Mexican government will improve the educational distribution of future generations and this in turn will shift the composition of immigrants towards the higher educated. Overall it will lower emigration from Mexico attenuating the pressure, especially of illegal immigrants.