This blog post is the second in a series that addresses the factual basis of a belief held by some opponents of bank mergers: that bank mergers drive branch closings that ultimately harm consumers. For instance, two prominent federal government officials (Consumer Financial Protection Bureau Director Rohit Chopra and Federal Trade Commission Chair Lina Khan) have articulated this view. In separate comment letters responding to the DOJ’s September 2020 request for public input on potential revisions to the 1995 Bank Merger Competitive Review Guidelines, they characterize bank mergers as a source of widespread branch closures.
In support of this notion, they cite a research paper authored by Hoai-Luu Q. Nguyen, a professor at the UC Berkeley Haas School of Business.[1] However, in drawing such an inference, they are misreading that paper. The referenced study finds only that merger-related closings most frequently occur where the merging banks’ branch networks closely overlap. Moreover, the Nguyen study is limited to just 13 large-bank mergers that took place between 2003 and 2007, a period during which the total number of bank branches in the U.S. increased more than 20 percent. Neither that study, nor the paper by Federal Reserve Board staff economists reviewed in the previous, Part 1 blog post, nor any other statistically rigorous study that this author is aware of indicate that bank mergers lead to broad-based and systematic branch closings.
Misguided Views
The comment letter submitted by now-CFPB Director Rohit Chopra (then an FTC commissioner) was co-authored and co-signed by Jeremy Kress, a professor of business law at the University of Michigan. [2] They state that:
Bank mergers, however, have led to widespread branch closures, inconveniencing customers who previously benefitted from proximity to bank offices.
The letter submitted by Federal Trade Commission Chair Lina Khan commented that:[3]
As banks have consolidated, merged entities have commonly closed branches.
In a separate response to the DOJ’s renewed request for public input on the merger review guidelines in December 2021, Professor Kress provided a marginal clarification of his view. He noted that the branches that close following mergers are typically located in close proximity to other branches of the merging banks:[4]
Bank consolidation has triggered merger-related branch closures throughout the country. As merging banks consolidate operations and cut overhead costs, they typically shutter branches in neighboring locations.
The Cited Study’s Approach and Findings
The Nguyen paper does not address whether mergers trigger branch closures per se. Instead, it investigates the relationship between the degree of branch overlap of merging banks and the likelihood of branch closure. The study finds that merger-related closings are much more likely to occur where the merging banks’ branch networks closely overlap. Specifically, in the first two years following a merger, the frequency of branch closure in census tracts where each of the merged banks had operated a branch is 27 percentage points higher compared to tracts that contained a branch of only one of the merged banks. In contrast, there is no perceptible difference in closure rates between these two area categories pre-merger.
This result is consistent with the main finding of the Board economists’ study, which showed less increase in total branches in markets where merging banks had more closely overlapping networks. It should be noted that the Nguyen study analyzes branch closings, while the Board economists’ study examines net change in the number of branches.
The Board economists’ study primarily defines closeness based on a geographic measure of distance between the branches of the merging institution within a given county. On the other hand, the Nguyen paper defines branches to be close if they are located within the same census tract.[5] As reported in the latter paper, the median census tract included in the analysis “is only 1.5 square miles—compared to 586 square miles for the median county.”
The Board economists’ study considers nearly 1,500 mergers that occurred between 1984 and 2010 and meet the criteria required by the paper’s analytical approach. In contrast, the analysis in the Nguyen study is limited to just 13 mergers that took place between 2003 and 2007. These mergers involved both the buyer and target banks holding at least $10 billion in pre-merger assets and having some degree of branch network overlap.
Thus, referencing the finding from the Nguyen study as evidence of “widespread branch closures” caused by mergers seems no more fitting than describing the results from the Board economists’ study as evidence that mergers induce systematic “branch destruction.” Neither characterization is accurate. There is no evidence in either study to suggest that bank mergers caused widespread branch closures or left communities deprived of conveniently located branches.
As seen in Figure 1 (reproduced from the Part 1), the total number of bank branches in the U.S. increased from about 70,000 to almost 86,000 during the 2003 through 2009 period to which the above-described result from the Nguyen study applies. This growth is certainly not indicative of widespread closures. As noted in the earlier post, often there is little justification for merging institutions to maintain pre-existing, duplicate branches in the same location. While a relatively small number of customers might have to travel a few extra miles or minutes longer to reach their nearest bank branch after the closure of redundant branch, most customers benefit from having access to the expanded branch network of the combined institution.
Figure 1: Number of Branches of FDIC Insured Depository Institutions, 1980-2014
Source: Federal Deposit Insurance Corporation
The Nguyen paper and the study by the Board economists use data from a mostly expansionary period for bank branches. However, since 2012 or so, the number of bank branches in the U.S. has been declining. Might this recent trend be driven by merger and acquisition activity? Once again, the answer is no. Branch closures have occurring at similar rates in banks involved in merger activity and those that were not, mainly driven by population trends and the expansion of mobile banking. Please refer to a BPI research note from January 2022 for details.[6]
[1] See Hoai-Luu Q. Nguyen, “Are Credit Markets Still Local? Evidence from Bank Branch Closings.” American Economic Journal: Applied Economics 2019, 11(1): 1–32. Available at: Are Credit Markets Still Local? Evidence from Bank Branch Closings (berkeley.edu). Beyond examining the effect of mergers on branch closing, the Nguyen study also investigates effects of merger-related branch closures on small business lending.
[2] See Comment of FTC Commissioner Rohit Chopra and Professor Jeremy C. Kress, p. 5.
[3] See Comment of FTC Chair Lina Kahn, p. 4.
[4] See Comment letter of Jeremy C. Kress, p. 8.
[5] In an appendix, the Board economists’ study confirms that net change in number of branches is inversely related to closeness of the merging institutions’ branch network when closeness is redefined as having branches within the same census tract.
[6] See Paul Calem, “Do Bank Mergers Create Banking Deserts? The Evidence Indicates No,” Bank Policy Institute, Jan 18, 2022. Available at Do Bank Mergers Create “Banking Deserts”? The Evidence Indicates No. – Bank Policy Institute (bpi.com).
