Conference Agenda

Flood risk is increasing with climate change, intensifying pressure on communities to invest in protective infrastructure such as levees. This paper studies household demand for levees using a stated preference discrete choice experiment conducted among homeowners in flood-prone areas of the Susquehanna River Basin in Pennsylvania. We combine subjective flood risk perceptions with objective, location-specific risk information and embed a randomized information intervention that corrects misperceptions prior to the choice experiment. We document widespread inaccuracies in flood risk perceptions, particularly for probabilistic measures such as cumulative flood risk over a 30-year mortgage horizon. Despite these misperceptions, providing objective flood risk information has no statistically or economically meaningful effect on willingness to pay (WTP) for levees, nor does it affect respondents with low baseline knowledge or those who under- or overestimate risk. Households do not significantly differentiate between levees offering different levels of flood protection. WTP declines when levees impose downstream flood damages, indicating that respondents internalize negative spillovers. Heterogeneity in WTP is driven primarily by personal flood experience rather than information or demographics.