Spatial spillovers in the pricing of flood risk: Insights from the housing market

In this study, we analyse how, and to what extent, direct and indirect effects (spatial spillovers) matter when estimating price effects for a property located in a flood zone. Unlike the previous literature, we show the importance of indirect effects resulting from a neighbourhood being situated in a flood zone. Additionally, the types of indirect effects (global vs. local) need to be determined theoretically or empirically using an appropriate spatial model comparison approach. Using the Bayesian model comparison for data related to the flood-prone city of Dresden, Germany, we find strong evidence for a spatial Durbin error model which controls for local spillover effects. These indirect price effects amount to -6.5% for houses and -4.8% for condominiums. However, direct effects diminish when controlling for spatial spillovers. Our results are generally robust across different model specifications, urban areas, and risk-adjusted prices that include future insurance costs, thus providing evidence of the importance of addressing indirect effects in the form of local spillovers in the analysis of flood zone effects. Ignoring indirect flood effects when formulating policy can lead to flood management that is inefficient and not cost-effective, as the economic consequences of flood are underestimated.