Women in the United States are much more likely to become mothers as teens than those in other rich countries. Teen births are particularly likely to be reported as unintended, leading to debate over whether better information on sex and contraception might lead to reductions in teen births. We contribute to this debate by providing causal evidence at the population level. Our causal identification strategy exploits county-level variation in the timing and receipt of federal funding for more comprehensive sex education and data on age-specific teen birth rates at the county level constructed from birth certificate natality data covering all births in the United States. Our results show that federal funding for more comprehensive sex education reduced county-level teen birth rates by more than 3%. Our findings thus complement the mixed evidence to date from randomized control trials on teen pregnancies and births by providing population-level causal evidence that federal funding for more comprehensive sex education led to reductions in teen births.