This paper uses one of the strongest hurricanes to hit the US, Andrew in 1992, to define a quasi-random experiment that permits estimation of the responses of housing values to information about new hurricanes. Lee County, Florida did not experience damage from Andrew. The storm was a "near-miss." We hypothesize that Andrew conveyed risk information to homeowners in the county. A difference-indifferences (DND) framework identifies the effect of this information on property values in areas likely to experience significant storm damage. The DND findings indicate at least a 19 percent decline in property values. (c) 2005 Elsevier Inc. All rights reserved.