Poster #127 - Extreme Heat and the Economic Vulnerability of Downtown Businesses: Evidence from Nine States

• In Event: Friday Poster Reception

Abstract

Downtown business districts are the economic and civic anchors of communities across the urban-rural continuum. Unlike the automobile-oriented commercial corridors that dominate modern retail geography, downtown businesses depend on walkability, pedestrian-scale access, and outdoor public space. However, these same characteristics make them structurally exposed to extreme heat in ways their suburban counterparts (e.g., strip malls and big-box stores) are not. As heat events intensify in frequency and severity, state and local policymakers face urgent questions about which places and businesses bear the greatest economic burden from climate shocks, and whether existing patterns of vulnerability vary across states in ways that should shape targeted adaptation investment.

We leverage high-resolution, business-level foot traffic data from 2019-2025—comprising nearly 300 million establishment-day observations across more than 150,000 points of interest in nine states—matched to spatially interpolated daily heat index measures derived from 2,000+ weather stations. Our study spans four climatically distinct regions across nine states, from the Southern Plains and Carolinas to the Pacific Northwest and Northern New England. Extreme heat is defined locally, as days exceeding the 99th percentile of each community's own climatological norm, ensuring that identification reflects genuinely anomalous conditions rather than chronic regional temperature levels.

Our core empirical strategy is a triple-difference design that compares downtown versus non-downtown establishments within the same community on the same day, across extreme heat days versus normal days, absorbing citywide heat effects through city-by-date fixed effects and seasonal business patterns through establishment-by-month fixed effects. We subject this design to a systematic battery of robustness checks motivated by recent advances in difference-in-differences econometrics. We first assess whether our estimates are sensitive to treatment effect heterogeneity across the staggered, multi-state panel using the decomposition approach of Goodman-Bacon (2021) and the heterogeneity-robust estimator of Callaway and Sant'Anna (2021). We then turn to the continuous treatment framework of de Chaisemartin and D'Haultfoeuille (2024), which is particularly well-suited to our setting given that heat exposure varies continuously across communities rather than switching cleanly on and off. Rather than collapsing a complex, multi-region treatment process into a single average effect, these approaches allow us to characterize the full distribution of heat vulnerability across community types, climate regions, and levels of exposure.

We find substantial regional heterogeneity in downtown heat vulnerability with direct policy implications. In Northern New England and the Pacific Northwest, downtown businesses experience significant declines in foot traffic on extreme heat days relative to non-downtown counterparts, with effects concentrated in general merchandise retail and personal services. In chronically warmer states, no significant differential effect is detected, suggesting that behavioral and infrastructural adaptation may buffer communities already acclimatized to heat. These findings imply that state variation in climate adaptation policy is not merely appropriate but necessary: the communities facing the greatest downtown economic exposure to heat shocks are often those least equipped to invest in protective infrastructure such as shade structures, urban greening, and cooling amenities.

Author

• Andrew J. Van Leuven, University of Vermont
  Presenting Author