Responding to risk: Information and decision making in the floodplains of St. Louis County, Missouri

This dissertation is an economic study of individual and firm decision making regarding flood risk in the United States. It is composed of three papers, all of which focus on St. Louis County, located at the confluence of the Missouri and Mississippi Rivers. The first paper asks if a severe flood causes homeowners to update their assessment of flood risk as seen in the capitalization of risk into property prices. Properties in 100-year floodplains were discounted 3.2% to 4.5% before the flood, but there was no price discount for properties in 500-year floodplains before the flood. A repeat sales model and property fixed effects model are used to estimate changes in the discount post-flood. For 100-year floodplains there was no change in price, but prices fell by approximately 3% post-flood for property in 500-year floodplains. In addition, prices fell significantly post-flood for all properties in municipalities where most of the damage occurred. The second paper looks at the demand for flood insurance among residential homeowners using data on all the policies-in-force in St. Louis County for the years 2000-2006. The paper examines the contract choices and retention rates of policyholders and explains variation in take-up rates and average coverage amounts by census tract in terms of economic, demographic, and geographic variables. Take-up rates are much lower in St. Louis County than in coastal areas. Risk variables, including the extent of 100-year and 500-year floodplains, being on a major river, and the presence of a levee, as well as wealth variables, are significant predicators of take-up rates and average coverage amounts. The third paper shifts focus to firms and uses a simple game-theoretic model to explain the decision of firms to locate in a hazardous area through the effects of signals (proxies firms look to as an indication of risk levels) and spillovers (positive or negative externalities from one firm to a neighboring firm). The hypotheses derived concerning the type of firm that is first to invest in a risky location and the dynamic of development over time are compared to development patterns in the floodplain of Chesterfield, Missouri.