Hurricanes imposed a heavy toll on the U.S. in 2004 and 2005: damages from the four Florida hurricanes in 2004 exceeded Hurricane Andrew, while Hurricane Katrina was the costliest natural disaster in U.S. history. Researchers have spent years devising plans and mitigation measures to reduce damages from hurricanes. The lack of a reliable database on mitigation efforts has hampered assessment of the effectiveness of these measures [Mileti DS. Disasters by design. Washington, DC: Joseph Henry Press; 1999]. We propose a new test of the effectiveness of mitigation using a past landfalling hurricane as a proxy, since conventional wisdom holds that communities often implement mitigation after a disaster [Cutter S. Living with risk: the geography of technological hazards. London: Edward Arnold Press; 1993]. We find that a prior landfalling hurricane – and by implication mitigation – can significantly reduce damages, by the equivalent of about a one category reduction on the Saffir–Simpson scale of hurricane intensity.