My dissertation consists of two chapters, which are both centered on the analysis of international economic policy. In the first chapter, ``Relocation Cost, Time Inconsistency, and the Temptation of Protection", I show that important aspects of agricultural support policy across developed economies emerge as features of the optimal policy of a government with a redistribution motive that lacks the ability to commit to its future policies. I document that, in the data, transfers to the agricultural sector are higher and more persistent over time in countries where the gap in productivity between agriculture and the rest of the economy is larger. This evidence is at odds with the benchmark with commitment, which prescribes a decreasing sequence of transfers to the low productivity sector to provide incentives to relocate. Without commitment, the government has a temptation to redistribute ex post, which depends on the gap in productivity between sectors. When the agricultural productivity gap is large, this temptation is strong, and the commitment outcome cannot be sustained. When the gap is small, the policy without commitment mimics the benchmark with commitment. In a quantitative exercise, I show that the policy without commitment can account for roughly 30% of the cross-sectional variance in the persistence of transfers to the agricultural sector over time. The second chapter, ``Migration and the State", joint with Zachary Mahone, studies the interaction between social insurance and migration policy. Governments in many countries help insure citizens against idiosyncratic risk. There is a long recognized tension between the potential gains from opening borders (increasing economic opportunity) and closing them (supporting state insurance). We develop a game-theoretic model of two countries that strategically interact in setting insurance and migration policies. We ask whether limits on mobility are a natural result of insurance provision, how equilibrium policies depend on the characteristics of the two countries and if these policies are efficient.