Flood risk poses a widespread and growing threat to economic activities in the US. Under perfect market design, both residential households and commercial businesses would gradually move out of the flood zones, which agrees with socially optimal resource allocation. The current flood assistance system, however, fails to enforce the flood insurance take-up by the commercial property owners, while providing government aid to these uninsured properties ex post. In other words, the commercial property owners in the flood zones are free riding the federal flood assistance system and take an advantageous position in the rental market. The low commercial rent, counterproductively, attracts businesses back to the flood zone, which leads to resource misallocation. This project looks back at the past two decades (1999--2020) and empirically shows businesses grow 10% faster in the central business districts after they are designated as flood zones, which manifests the pulling force of businesses due to the free riding problem in the current flood assistance system. We build a spatial model to interpret the empirical finding as business misallocation, and quantify the welfare loss.

To better understand how information affects the extent to which the housing market evaluates climate risks, we use a flood risk communication campaign in the municipality of Dordrecht, the Netherlands, as the context to analyze whether being exposed to enhanced flood risk information affects real estate behavior. Using a DiD framework, we find that due to the amenity effect, housing prices are 9.3%-13.6% higher in areas not protected by the dike and at higher risk of flooding. However, this price premium is observed only before 2010 when the municipality began sending annual letters informing residents of potential flood risks and mitigation measures in out-dike areas. After 2010, out-dike houses exhibit no significant price difference compared to those inside the dike. The days on the market for properties outside the dike exceed those for comparable properties inside the dike, with the difference increasing after 2010. Our results suggest that flood risk is not fully incorporated into real estate prices in the city of Dordrecht specifically, and in the housing market more generally. When new information is provided on flood risk and mitigation suggestions, individuals become more aware of the risks, leading to reduced market liquidity and lower prices in high-risk markets.

We study the effects of news about sea level rise on housing markets. We exploit

two natural experiments in Spain. In 2007 and 2014, Greenpeace published two alarming

reports predicting catastrophic consequences for La Manga, a tourist peninsula. These

reports were widely cited in the local news. We find that both reports caused immediate

and persistent drops in housing prices in La Manga but had no effect on housing rents.

The transaction numbers in La Manga remained unchanged post 2007 report but dropped

after the 2014 report. Difference-in-differences regressions with different control groups

confirm the results. We find persistent drop in average transaction price in La Manga

around 4% to 24% post publication of the reports. We also find a persistent drop of

around 5% to 10% in listing prices in La Manga post 2014 report. Transaction number

also drops by 9% post 2014 report. A set of robustness tests supports our findings.