Using new data on county-level variation in alcohol prohibition from 1933 to 1939, we investigate whether the repeal of federal prohibition increased infant mortality, both in counties and states that repealed and in neighboring counties. We find that repeal is associated with a 4.0% increase in infant mortality rates in counties that chose wet status via local option elections or state-wide legislation and with a 4.7% increase in neighboring dry counties, suggesting a large role for cross-border policy externalities. These estimates imply that roughly 27,000 excess infant deaths could be attributed to the repeal of federal prohibition in this period.