Uncertain Hazards, Insurance, and Consumer Choice: Evidence from Housing Markets

ent with the expected utility model when considering potential losses from earthquakes in the Los Angeles and San Francisco areas. Shilling, et al. (1985) illustrate that the housing market in Baton Rouge, Louisiana, values single family dwellings less when they are located in an area with an enhanced probability of flooding. Thayer, et al. (1985) estimate the reductions in housing prices that resulted in Mammoth, California, when the USGS issued earthquake and volcano hazard alerts. Dunn (1986) investigates the impacts on real estate in Institute, West Virginia, following a leakage of hazardous chemicals. All of these studies use residential property values to study consumer behavior with respect to probabilistic events that cause losses in utility. In like fashion, we investigate behavioral responses to a natural hazard (flooding) by examining residential property values. This study differs from Shilling, et al. in that we examine the theoretical aspects of the location decision. Moreover, this paper complements Brookshire, et al. by including insurance considerations. This type of research is important for at least two reasons. First, the results of the investigations cited above can be used to develop benefit/cost studies which attempt to assess the economic merits of policies that change the likelihood or magnitude of an event. Policies such as special building codes for earthquake prone areas, the construction of pumping stations to divert flood waters, tighter enforcement and more stringent controls in the chemical industry, and regulating urban land use are very costly public projects. While the costs of these projects are usually available, the benefits to consumers must be estimated. What are consumers willing to pay to reduce the likelihood or magnitude of an event? The residential housing markets may provide an avenue for estimating these values since the choice of where to live often includes the choice of hazard level. To the extent that consumers realize this, differentials in property values will arise that reveal marginal willingness to pay for reductions in hazard levels. By statistically holding all other influences on property values constant, these differentials can be calculated.