Historically, bankers sought to establish broad branch networks to leverage the network effect, an empirically demonstrable phenomenon wherein such footprints capture a disproportionate share of market balances.
Larger branch networks captured greater balances per branch than smaller networks within a given market: an eight-branch network gained more than twice as much in deposits as a four-branch network. With consumers valuing the perceived convenience of broader networks, each successive branch provided a synergistic lift to the other branches.
In recent years, electronic channels have matured, and the pandemic forced branch-dependent consumers to consider other channels, raising questions of whether the network effect still holds or if the aforementioned developments weakened the effect. If the network effect holds, then branch deployment remains a race for institutions to build the broadest network in a market. If the effect has waned, it would justify branch-light distribution strategies.
To measure the network effect, Bancography studied the correlation between deposits per branch and the number of branches in the 50 largest U.S. metropolitan statistical areas (MSAs), using Spearman’s rank correlation coefficient, controlling for main-office biases and other outlier effects. Spearman’s statistic measures the degree to which the rankings of each variable align, minimizing the effect of outliers.
If the network effect holds, we would expect high correlation. The institution with the greatest number of branches in a market would show the highest deposits per branch; the next largest network would show the second-highest deposits per branch and so on.
Even with the rise of electronic channels and the COVID-driven inducement to utilize non-branch channels, the study found the network effect persists, with correlations significant at the .01 level in 48 of the top-50 metros, and at the .05 level in the other two. The average correlation between deposits per branch and branch count across the top 50 metros reached 65%.
In two-thirds of the markets, correlation increased from a 2017 study, indicating the network effect as stronger now than it was seven years ago. This suggests that as consolidation has reduced the number of institutions in a given market, there is more of a premium on network scale. As mid-tier banks are absorbed into larger entities, scale becomes an even more-prominent attribute to consumers.
The markets with lesser correlations reveal paths other than network ubiquity to realizing top-tier deposits per branch. In Austin, Texas, and Seattle, several leading banks by deposits per branch maintain smaller branch networks, employing a focused segment approach. In Austin, several specialists emphasize middle-market commercial banking, and in Seattle, multiple specialists focus on subsegments of the Asian-American community. These banks are forgoing the benefits of market-wide coverage, instead striving to dominate a specific product or affinity-driven segment.
However, for general retail-banking practitioners, industry consolidation has yielded an environment more dependent than ever on scale. The absorption of mid-tier competitors has left many U.S. markets approaching a Canadian-type competitive environment with a few dominant market-wide providers and a host of smaller localized community providers.
Electronic channels are not an equalizer for smaller banks. Although the ability to open accounts outside of a branch should negate disparities between large and small banks by reducing the need for costly branches, larger networks still capture a disproportionate share of market deposits.
The inability of online channels to level the environment reflects the homogeneity of those capabilities. Differentiators thus get arbitraged to table-stakes, leaving other factors, including brand awareness and perceived accessibility, as choice determinants — attributes impounded in the breadth of a bank’s branch network.
Thus, we can still summarize an effective distribution strategy as get big or get out.
Bankers must pursue broad market coverage to maximize efficiencies of their branch networks and scattered toehold presences remain ineffective. Given the capital to deploy four branches in Ohio, a bank should choose one market in which to concentrate all four branches versus placing one each in Cleveland, Columbus, Dayton and Cincinnati, since the concentrated approach would leverage the network effect.
Bankers should pare limited-scope deployments outside their primary markets, reallocating resources into core markets. Alternately, bankers can consider segment-specific or line-of-business models (wealth management, mortgage or middle-market commercial) in markets where they cannot attain market-wide coverage. Otherwise, banks can pursue a general retail-banking model, but at the corridor versus market level, as the network effect holds even in sub-corridors of a region.
Even as electronic channels have expanded the geographic reach of banks, deposit gains still accrue disproportionately to the banks with the broadest branch networks. As long as network breadth drives institution choice, branches remain relevant, and bankers must emphasize the markets where they can leverage the network effect, while paring investments in markets that lack a pathway to the density levels that enable top-tier efficiencies in deposit gathering.
