Banking Deserts and Local Credit Access

• In Event: Credit Access, Alternative Financing, and Homebuyer Outcomes

Abstract

In recent decades, the U.S. banking sector has experienced several waves of branch closures, raising concerns about access to mainstream financial services and local economic opportunity. This paper studies whether the formation of banking deserts reduces local access to credit, particularly for small businesses. This question is important because physical bank branches may still matter for relationship-based lending, trust building, and access to in-person financial services, especially for financially constrained populations. At the same time, it remains unclear whether reduced physical branch presence necessarily translates into reduced access to credit in an era of expanding digital and online banking. To study how banking desert formation affects local credit access, I use tract-level data from the Federal Reserve’s Banking Desert Dashboard from 2019 to 2025 and link these data to Community Reinvestment Act small business lending outcomes from 2015 to 2024. The banking desert data identify whether a census tract has no bank branch within the tract or within a distance threshold based on its metro status. Using this information, I identify census tracts that transition from having at least one active bank branch to having no nearby branch access, an event concentrated largely in suburban and some rural communities. The lending data allow me to examine changes in both the number and dollar amount of small business loan originations, including lending to firms with gross annual revenues below $1 million. I also plan to extend the analysis to household mortgage credit using Home Mortgage Disclosure Act data. To quantify how banking deserts affect local credit supply, I adopt a matched difference-in-differences approach that exploits variation in the timing of desert formation, defined as the year when a tract loses all bank branches within the relevant distance threshold. Using detailed tract-level demographic characteristics and county-level economic conditions, I construct a matching procedure to identify a set of comparison tracts that do not lose branch access. I then compare the evolution of lending outcomes in treated and comparison tracts around the time of desert formation. I also show that the results are robust to a non-matching difference-in-differences approach using the Callaway and Sant’Anna (2021) estimator. Findings show little evidence that banking desert formation reduces observed small business lending at the tract level. Estimated effects are generally close to zero, although some specifications suggest modest increases in the number and volume of loans, especially to very small firms. These findings complicate the common assumption that branch closures mechanically reduce local credit supply. Instead, they suggest that borrowers and lenders may adjust through alternative branches, digital banking, or shifts in lending composition thus promoting financial inclusion may require looking beyond branch counts alone and paying closer attention to how credit is delivered and which communities remain most dependent on physical branch access.

Author

• Angelica Maisuh, University of Delaware
  Presenting Author