大西洋飓风对商业活动的影响

Policy Research Working Paper The Impact of Atlantic Hurricaneson Business Activity Sumit AgarwalSmarajit Paul ChoudhuryMingxuan FanLeora Klapper Policy Research Working Paper11217 Abstract This paper quantifies the short-run economic impact of21 Atlantic hurricanes on U.S. local business activity from2017 to 2024 using anonymized Mastercard transactiondata aggregated by ZIP code. On average, hurricanesreduce merchant sales by 12.4 percent during the prepara- across nearby areas or large online platforms is limited, indi-cating widespread local consumption declines. Economicdisruption varies more by industry than storm intensity,with independent stores hit harder than chains. Local busi-nesses with larger online presence face smaller, shorter sales The Impact of Atlantic Hurricanes on Business Activity BySUMITAGARWAL,SMARAJITPAULCHOUDHURY,MINGXUANFAN,ANDLEORAKLAPPER∗ JEL: G20, G50, D22, O31, Q54 Keywords: Digital payments, Extreme weather, Consumer spending, Small- andmedium-sized enterprises ∗Agarwal, National University of Singapore,bizagarw@nus.edu.sg; Choudhury, Mastercard Economics Institute,smarajit.PaulChoudhury@mastercard.com; Fan, National University of Singapore,mingxuan.fan@nus.edu.sg:Klapper, Corresponding author, World Bank,lklapper@worldbank.org. We are grateful to the MastercardEconomics Institute for providing controlled access to data based on aggregated and anonymized transaction data aswell as technical support. We thank Joel Alcedo, Ethan Harris, Adrienne Manns, and Michelle Meyer for helpfulcomments. We acknowledge financial support from the World Bank Research Support Budget (RSB). The findings,interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily 1.Introduction Extreme weather events have increased in frequency and severity in recent years, placingabout 1.2 billion people at high risk from weather-related hazards (World Bank, 2024). Whilethe long-term macroeconomic consequences remain uncertain, recent research points tosignificant economic risks. Empirical studies consistently find that rising temperatures reduceaggregate productivity, particularly in lower-income countries, where weather vulnerability is At the microeconomic level, weather-related risks place significant financial strain onboth households and firms. Households experience sharp declines in spending followingextreme weather shocks (Bui et al. 2014; Gallagher and Hartley 2017; Deryugina, Kawano, andLevitt 2018), with disproportionately large impacts on lower-income households (Sawada andShimizutani 2008; van den Berg 2010). Elevated temperatures have also been shown to reducelabor productivity (Hugon and Law 2018; Cevik and Miryugin 2023) and exacerbate financing Despite this growing literature, relatively little is known about how localized businessactivity responds to hurricanes—one of the most destructive and increasingly frequent types of extreme weather—especially across different industries. This paper addresses that gap by usinganonymized high-frequency credit and debit card transaction data from Mastercard,1aggregated affected by hurricanes between 2017 and 2024, offering a uniquely granular view of real-timeeconomic responses to weather-related shocks. This study makes three main contributions to the economics literature. First, we providenew high-frequency evidence on the short-run economic costs of hurricanes, which capturesdaily sales activity across both physical and online sales channels, including small independent Second, we analyze sales at local merchants (in-store and online), as well as totalspending by local consumers at local and nearby business locations and online marketplacesduring a defined time period. This allows us to test for substitutions in spending acrossgeographic locations and sales channels. We find little evidence of such substitution: spending Third, we use a stacked difference-in-differences design to estimate the dynamic impactsof 21 hurricanes—ranging from Category 1 to 5—and quantify the changes in businessperformance during the hurricane’s preparation, impact, and recovery periods. We findsignificant short-term revenue losses, equivalent to approximately two days of typical sales, Our results reveal that variation in the economic impact of hurricanes is more pronouncedacross industries than across storm categories or firm types. Hotels and gas stations oftendemonstrate more resilience and, in some cases, temporary boosts in demand. Gas stationslikely benefit from the increase in sales as people prepare for potential evacuations. For hotels,this may reflect a complex interplay of canceled reservations and increased demand from Overall, our findings contribute to a deeper understanding of the localized and industry-specific impacts of extreme weather events. By leveraging high-frequency transaction-leveldata, we shed new light on the short-run dynamics of economic disruption and recovery. From T