The use of test-score-based "value-added" (VA) measures to evaluate teachers is highly debated, partly because there is little evidence on whether high VA teachers improve student outcomes in adulthood. We address this question by linking school district data from grades 3-8 to tax records for 3 million children. We find that being assigned to a high value-added teacher increases the probability of college attendance, raises earnings, reduces teenage pregnancy rates, and improves the neighborhood that the student lives in at age 25. We address concerns about sorting of students to teachers using previously unobserved parent characteristics as well as a quasi-experimental research design based on changes in teaching staff. The impacts of teacher VA are roughly constant across grades 3-8. A one standard deviation improvement in teacher VA in a single grade raises earnings by 0.6%. Replacing a teacher whose VA is in the bottom 5% with an average teacher increases students' earnings by up to $176,000 per classroom taught.