Essays on the Economics of Flood Risk

Climate change yields heightened flood risk due to both sea level rise and an increased frequency of significant storms. While prior literature has estimated flood damages, it is difficult to measure how agents will respond to an increase in flood risk. Furthermore, it is important to understand

how policymakers can minimize these risks by investing in flood hazard protections. My dissertation studies the problem of valuing the cost of flood risk and the role of policymakers in response to flood risk. In the first part, I explore the relationship between flood risk and the housing market. This allows me to measure the cost of increased flood risk on housing markets using revealed preference

methods. In the second part of my dissertation, Chapters 2 and 3, I study why policymakers invest in flood hazard reduction and how flood hazard mitigation is valued by homeowners. My findings on what factors influence policymakers to invest in flood hazard protection at the local

level and how these investments are valued provide important insights for the federal government as it redesigns the National Flood Insurance Program (NFIP). My first chapter (joint with Nicholas Z. Muller) provides evidence that well-informed homeowners of coastal real estate respond to flood risk. Our innovations in this paper are two-fold. First, we parse homeowners according to the quality of information they possess about their flood risk, and second, we disentangle flood risk from property damage. Prior literature has found evidence that homeowners respond to flood shocks, however these papers focus mostly on local flood shocks. As such, it is difficult to separate the cost of damages from the effect of information about risk. Our paper circumvents this identification issue by testing whether non-local flooding events affect housing prices in coastal markets. Utilizing a difference-in-differences methodology, we test whether homeowners in high flood risk areas along the coast of New Jersey respond to non-local flooding. We use several well-publicized hurricanes and tropical storms that did not strike the Atlantic seaboard as non-local shocks. We find that prices of homes in areas of high flood risk

do not decrease after a storm, but rather increase. The literature has shown current and prospective homeowners do not always know their flood risk, so we further test if informed homeowners respond differently. We use participation in Community Rating System (CRS) public awareness activities as a measurement of homeowner’s information. We find that homes in high flood risk zones situated in towns that participate in public flood awareness activities incur a 7 to 16 percent decrease in price after the non-local shock. Further, we show that firms are more responsive to risk information than individuals and that markets exposed to such information are less adversely

affected by future disasters. For my second chapter I study the decision to participate in CRS, which is a federally run program that incentivizes local jurisdictions to undertake activities related to flood hazard mitigation in exchange for flood insurance discounts for their constituents. In this chapter I study the decisions of the government to participate in CRS from a static perspective. First, I build a model of the local government’s decision to provide a public good that mitigates hazard risk. Second, I use participation

in CRS in New Jersey to empirically test the hypotheses generated by the theoretical model in the context of flood hazard mitigation. Consistent with the model predictions, the empirical results show that an array of factors affect participation: income, population, housing values, risk, value of amenity access, information, and whether the local jurisdiction type is mayor-council. This paper further contributes to the literature on optimal public good provision by showing that incomplete information, weak government accountability, and lobbying can lead to inefficient levels of hazard mitigation. My third chapter further explores the decision of local governments to invest in flood hazard

mitigation and builds on my second chapter by using a dynamic structural model of the local government’s

decision, which accounts for the value of flood hazard mitigation to homeowners. Thus, I am able to consider the interdependency between changes in federal policy and the decisions of the local government. To this end, I estimate a homeowner’s marginal willingness to pay for flood

insurance and for the local government’s flood hazard mitigation actions based on housing sales in New Jersey from 1998 to 2018. I then use a dynamic discrete choice model of the local government’s investment decision to estimate their costs. I find that the spillover effects from mitigation are positive, insurance discounts are valued more than the actual savings, and that large initial perceived costs may prevent investments in hazard mitigation. Finally, I perform counterfactual analyses to consider the local government response to alternative federal policies. The counterfactuals suggest that either increasing the proportion of homes in federally designated high risk zones to account for climate change or raising the federally set flood insurance rates increase investment in flood hazard mitigation, and that implementing a cost subsidy rather than the current insurance discount policy may increase investment in municipalities with low property values.