We provide new evidence of one channel through which circular labor migration has long-run effects on origin communities: by raising completed human capital of the next generation. We estimate the net effects of migration from Malawi to South African mines using newly digitized census and administrative data on access to mine jobs, a difference-in-differences strategy, and two opposite-signed and plausibly exogenous shocks to the option to migrate. Twenty years after these shocks, human capital is 4.8-6.9 percent higher among cohorts who were eligible for schooling in communities with the easiest access to migrant jobs.