Do climate risks predict US housing returns and volatility? Evidence from a quantiles-based approach

Abstract

This paper analyzes the ability of textual-analysis-based daily proxies of physical (natural disasters and global warming) and transition (US climate policy and international summits) climate risks to predict daily movements in the US housing market in the entire conditional distribution of housing returns and volatility. Using data for the period 2 August 2007 to 29 November 2019, a nonparametric causality-in-quantiles test is used, accounting for nonlinearity and structural breaks between housing returns and climate risk factors. The Granger causality analysis shows that climate risk factors (and the associated uncertainty) predict housing returns and volatility across the entire conditional distribution. These results are robust to alternative daily data of aggregate housing prices for the US and 10 major metropolitan statistical areas. Insights from our findings are beneficial for the decision-making of investors, policymakers and the academic research community.