International migration opportunities improve earnings capacity, but can also create parental separation from children. For example, approximately one in every four children in the Philippines, will not have a parent physically present in their life at some point due to parental migration. Both parental presence and monetary investments determine a child's human development and educational outcomes, so migration decisions necessarily involve tradeoffs. The net effect of migration on children depends on the evolving relative importance of money and parental presence, which parent migrates, and at what age they separate from the child. Therefore, who chooses to migrate, and when, potentially has significant welfare implications for children left behind. Such considerations, in turn, can affect parental decisions to migrate in the future.

I estimate a dynamic model of migrant households with an embedded age-specific child education production function by conducting a panel survey of Filipino migrants and combining it with newly assembled administrative data from the Philippines Department of Education and the Department of Migrant Workers. Substantial variation in the timing of parental time inputs and monetary investments into children arise from the systematic and widespread international migration of Overseas Filipino Workers (OFWs). For identification of the endogenous inputs in the education production function, I exploit shocks to the demand for male and female Filipino migrant workers in East Asia and the Middle East.

My instruments are motivated by the model of the migrant household where parents make repeated migration decisions and obtain utility from consumption and the academic performance of the children. This provides a framework to understand how parents trade off parental time and expenditures for their children when making migration decisions. The migration and expenditure allocation decision rules derived from the model are estimated, which offers insights into how migrants make repeated migration decisions as parents internalize the evolving needs of their children. I finally impose further structure onto the model and estimate its parameters to run counterfactual simulations and calculate the net impact of migration.

I find monetary resources play a considerably more significant role in shaping child human capital from the ages of 11-15, whereas both maternal and paternal time inputs are more critical between the ages of 6-10. A mother's time is always substantially more productive than a father's time across all ages. Parental time and monetary inputs are always complements. Father's and mother's presence are complementary when a child is 6-10, but become substitutes at later ages. Parents internalize the effects of their absence on their children and tend to migrate if they observe successful academic progress. In addition, parents are more inclined to migrate when they know their children possess a greater endowment of abilities that enhance academic outcomes. These findings hold significance for policymakers in developing countries who seek to enhance economic welfare by promoting temporary labor migration, which inevitably causes parental absence.