Branch banking in the USA: a bank level analysis, 2003–2023

Purpose This research seeks to understand the microeconomic behavior of banks around the long-term decline in bank branches in the USA. Design/methodology/approach Federal Deposit Insurance Corporation data from 2003 to 2023 are used to characterize banks by net annual changes in the number of branches, asset size, market competition, market demographics, extent of traditional banking activities as well as urban status. Regressions are estimated for changes in the absolute number of branches and qualitative changes, which allow for asymmetric effects of independent variables on expansion and contraction. Findings Strong evidence is found for churning, with banks likely to expand in some years also likely to contract in others. Churning is associated with the number of prior branches, bank size, non-urban markets and banks focused on traditional banking activities in terms of loans and deposits. Churning undercuts standard hypotheses positing linear or symmetric effects for independent variables. Nonetheless, as hypothesized, poverty is related to contraction, but African Americans and Hispanics are not. Also contrary to predictions, monopoly power is associated with stability, with competition related to churning, as forseen by J.R. Hicks in 1935. Finally, the link between branch expansion and both market competition and bank size weakened or disappeared entirely by the 2020–2023 period, consistent with on-going branch contraction. Originality/value Findings on the frequency of branch churning behavior are novel, as are the findings of decreased importance of market competition and bank size starting in 2020.