| Abstract: | There is growing evidence that flood mitigation is often inefficient because individuals misestimate flood risk. The propensity to misestimate flood risk is expected to rise because climate change ensures the past will be a poor predictor of the future. Greater reliance on downscaled climatological and hydrological forecasts has been suggested to address these information failures. This article combines stochastic dynamic programming with historical data and climate‐driven streamflow projections to determine how changes in flood risk forecasts influence optimal investments in flood mitigation infrastructure. Using upgrades in California's levee system as an example, we show that climate change is causing benefit–cost analysis to become increasingly biased in favor of flood mitigation infrastructure projects. We also show that using downscaled hydroclimate forecasts to achieve more accurate estimates of flood risk can decrease the efficiency of flood mitigation infrastructure investments, if flood risk is currently overestimated. JEL Classification: D81 Q25 Q54 |
