ABSTRACT
We combined five data sources to inform the analysis. Data on CRS participation is obtained from the 2013 CRS Coordinator’s Manual (FEMA, 2013) and the FEMA. Underlying flood risk data from the US Department of Transportation (US DOT) (1996) offers very high resolution (1 km grid cell) rankings of flood risk (on a 0-100 scale) that use underlying topography and hydrography of the area. This flood hazard rank variable derives from a formula that equally weights annual flooding frequency ranked from 0-100 (which itself is an area-weighted average of flooding by soil map units within the 1km grid cell) and their potential scour depth ranked from 0-100. Scour depth reflects erosion hazard based on 100-year flood flow, sediment size, and stream shape characteristics (Williams, Carreon, & Bradley, 1992). Thus this flood hazard risk variable captures both the frequency and the intensity of flooding. This flood risk measure has three advantages: (1) it derives from data and computations that largely predate the start of the CRS program, (2) it offers a rich quantitative scale for flood risk, and (3) it provides spatial resolution much smaller than cities or counties, which allows better distributional characterization of flood risk1. The Natural Amenities Index, which contains data on the physical characteristics of counties like topography, climate, and water coverage, are taken from the USDA’s Economic Research Service. Information about the population and housing stock of communities is obtained from 1990 block-group level Census data from the United States Census Bureau. Finally, information about government expenditures and revenues is taken from the 1992 Census of Governments, the earliest available Census data on local governments’ finances.
