Flooding remains a major global challenge, affecting millions of people and causing billions of dollars in damage yearly. Some 53 percent of natural disaster victims worldwide in 2012 were victims of floods (Guha-Sapir et al., 2013). In the United States, flooding caused an average of 95 fatalities and $8.2 billion in damages annually between 1982 and 2011. In order to reduce the impacts of floods in the US, the Federal Emergency Management Agency created the Community Ratings System (CRS) in 1990. The CRS is a voluntary rating system intended to encourage communities to engage in flood management by rating participating communities’ levels of flood management measures implemented. Community homeowners enjoy discounted premiums (up to 45%) on federally required flood insurance based on their community’s CRS score. Currently, over 1,000 communities participate in the CRS program, achieving a wide range of ratings.  Evaluating the CRS directly informs the intersection of federal and local policies for housing and community development in floodplains (an arguably larger global challenge as well).

It appears that communities participate in the CRS program in pursuit of flood risk reduction and for other (less noble) reasons.  As a result, we explore three research questions: (1) How much are participating communities taking advantage of or “gaming” the CRS scheme without appreciably reducing their flood risks? (2) What are the characteristics of communities that are gaming CRS? (3) What drives communities to game the system? This gaming involves communities behaving strategically in order to obtain CRS points that would put them in higher CRS classes (Zahran et al., 2010) and consequently afford them higher insurance premium discounts without substantially improving flood management. For example, communities might select a disproportionate number of less expensive (passive) activities relative to more expensive (active) activities (Brody et al., 2009) in order to get into a higher CRS class and obtain higher premium reductions. Such strategic behaviors of communities may be indicative of inefficiencies in the CRS incentive structure that could undermine the program’s effectiveness.

We answer these research questions using national data on historical CRS participation, previous Censuses, government characteristics from the Census of Governments, climate and topographical information, and other data sources. We use Regression Discontinuity (RD) approach to: (i) ascertain the extent of gaming within the CRS program; (ii) identify the characteristics of the communities that are gaming the CRS program; and (iii) test several competing hypotheses that might explain what is driving such gaming behaviors of participating communities. Some of these hypotheses follow Median Voter Theorem and special interests rationales, while others follow arguments about the size of local governments predicting the supply of management programs and community flood risks predicting greater demand for flood management. Preliminary results indicate that many communities are gaming the CRS program. More strategic communities differ from other CRS participants, although which characteristics differ depends on the intensity of participation. Our results have policy implications for community and housing development and flood risk reduction, both in the US and globally.