What the Evidence Can—and Cannot—Tell Us So Far

The evidence on completed U.S. bank mergers now supports a coherent judgment, but not a slogan. National banking consolidation is real, local mortgage markets can remain fragmented, and a merged lender’s average can change without its pricing rule changing. This checkpoint connects those ideas so readers can decide what the current evidence actually establishes—and see which questions about mortgage rates, access, performance, competition, selection, and regulation still require separate examination.

The contradiction disappears when the layers stay separate

The apparent contradiction begins with two observations that are both true. The U.S. has fewer active banks and thrifts than it did in 1985, and a much smaller number of institutions controls half of banking assets. At the same time, the mortgage market encountered by a borrower is organized locally and includes firms that are not banks. The Federal Reserve study’s county-level analysis finds many active mortgage originators, generally small individual shares, and substantial participation by nonbank lenders. There is no logical inconsistency in saying that the banking sector has consolidated while the local market for a particular mortgage remains competitive.

The same separation applies inside a merger. A legal combination joins two institutions, but it does not instantly erase the differences between their portfolios. One lender may originate more loans, serve a different borrower mix, retain more loans, or price differently from the other. Once the portfolios are reported under a single name, the combined average reflects both the institution’s behavior and the weights of the loans now being combined. A higher average can therefore be an observation about composition, not proof that the acquirer discovered new pricing power.

The causal question is narrower: after the merger, did the treated institution change relative to comparable lenders operating in the same county and quarter? The study’s event design makes that comparison rather than treating the legal completion date as a magic boundary. Its rate measure adjusts for observable borrower and loan characteristics and local market conditions. Its pro-forma construction keeps the target’s pre-merger activity attached to the combined entity, so the analysis does not mistake a change in reporting identity for a change in behavior.

The resulting causal chain is compact:

flowchart LR A["National consolidation"] --> B["Local mortgage market and lender alternatives"] B --> C["Merger type, target selection, and portfolio composition"] C --> D["Within-market comparison around completion"] D --> E["Rates, approvals, and loan performance"] E --> F["No systematic harm in the selected completed-merger sample"] F --> G["Limits: not universal safety or proven benefit"]

This chain also shows why the study’s findings cannot be read backward too quickly. Stable borrower outcomes do not prove that national consolidation is always benign. They are consistent with a setting in which many alternative lenders constrain the merged institution, in which deals are selected for scale or business-model fit rather than market capture, and in which transactions have already passed regulatory review. Those mechanisms make the result intelligible; they do not turn it into a rule for every market.