Extreme Weather Events and Financial Markets: The Impact of North Atlantic Hurricanes

An Industry Comparison of the Impact of Extreme Weather Events on Stock Returns: Scientists forecast increasing extreme weather events over the coming decades due to climate change. We examine the effect of North Atlantic hurricanes on U.S. stock returns over the period January 1990 to December 2014. For industry portfolios spanning the entire U.S. stock market, we find that the abnormal effect on stock returns is differentiated according to industry classification, hurricane intensity, landfall location, and estimated ex post financial damages. The market portfolio exhibits negative mean daily standardised abnormal returns of -0.0064 in the prelandfall period, -0.0015 on the event day and subsequent day, and is approximately -0.006 out to 30 days post-landfall. Aggregate figures for the market, however, conceal a wide variation in the abnormal effect by industry: in the pre-landfall
period, abnormal returns range from -0.075 (Restaurants) to +0.071 (Oil) and over the 30-day post-landfall period, from -0.053 (Real Estate) to +0.041 (Gold). Many manufacturing-based and consumer-focused industry portfolios exhibit consistently negative abnormal returns over the event window. Conversely, portfolios of Gold, Oil, Hardware, Software, Banks and Trading stocks exhibit a consistently positive abnormal effect, possibly due to investors seeking stocks not negatively exposed to the effects of a hurricane. We find no evidence that market efficiency has increased over the period of our study, i.e., that market participants have incorporated the expected financial implications of hurricane events into their stock price expectations.

Idiosyncratic Volatility and Return Predictability: The Impact of Extreme Weather Events : We examine the impact of North Atlantic hurricanes on the idiosyncratic volatility of portfolios of U.S. stocks sorted by realised volatility. We find that the magnitude of the abnormal effect on idiosyncratic volatility increases with underlying portfolio volatility in absolute terms and decreases in relative terms. The magnitude of the effect is further related to hurricane intensity. In out-of-sample tests, we find weak evidence that hurricane events are a predictive variable for one-day ahead returns on an equal-weighted portfolio of stocks for the lower-volatility portfolios and the aggregate market portfolio after controlling for the business cycle. Finally, using portfolios of stocks sorted by industry classification, we find a differential abnormal effect on idiosyncratic volatility across industry portfolios.

Extreme Weather Events as Distress Risk Shocks: Evidence of Size and Value Effects : Hurricanes represent exogenous shocks to firm distress. Using firm size (ME) and book-to-market equity ratio (BE/ME) as ex ante proxies for firm distress, we investigate how the abnormal effects due to landfall hurricanes on stock returns, volatility, and illiquidity, vary across decile portfolios of stocks sorted by ME and BE/ME. Abnormal returns are negatively related to firm size, with small-firm stocks exhibiting a larger negative abnormal effect than big-firm stocks. For portfolios sorted by BE/ME, Value and Growth stocks at the extremes of the decile range exhibit the most negative abnormal return effect. Abnormal variations in idiosyncratic volatility are greatest for the smallest stocks and Value stock decile portfolios. Small-firm stocks and Value stocks at the extremes of the decile range experience a significant increase in illiquidity. The magnitude of the abnormal effects also vary with maximum pre-landfall hurricane intensity, with the more severe hurricanes producing a more adverse abnormal effect. Although the SMB and HML factors may proxy for distress risk, they do not appear to capture the specific distress risk shocks posed by hurricane events.