The CRA proposal's 85.4% asset figure cannot establish LMI farm-loan coverage

No. The agencies' 85.4% figure says that 86 banks satisfying the proposed large-bank test would hold 85.4% of total assetsassetsAssets are the resources reported on a bank's balance sheet, and the FFIEC transmittal file records the bank's prior year-end Call Report assets in thousands of dollars. https://www.ffiec.gov/sites/default/files/data/cra/flat-files/24FlatTransSpecs.pdf at the banks in the OCCOCCThe Office of the Comptroller of the Currency is the federal agency responsible for CRA supervision of national banks and federal savings associations. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf-FDICFDICThe Federal Deposit Insurance Corporation is the federal agency responsible for CRA supervision of state-chartered banks that are not Federal Reserve members and state savings associations. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf analysis. It does not say what share of small-business or small-farm loans those banks reported, what share was in low- and moderate-income tracts, or how the share differs between metropolitan and nonmetropolitan places. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf

The distinction is like measuring a library by the floor space it keeps rather than by which books remain on the shelves. A regional bank can represent little of the industry's assets while supplying many observations from places that larger banks serve less intensively. A large banklarge bankUnder the proposal, an OCC- or FDIC-supervised bank would be a large bank if it had more than $10 billion in assets at both of the preceding calendar-year ends. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf can hold vast assets without supplying the same proportion of every kind of local loan record. The base used for the percentage determines what the percentage can prove.

The FFIECFFIECThe Federal Financial Institutions Examination Council coordinates work among federal financial regulators and publishes the combined CRA data used in this analysis. https://www.ffiec.gov/data/cra/findings-from-2024-data-fact-sheet's published size comparison demonstrates the mismatch. Across all three CRA regulators, 157 reporters with at least $10 billion in assets supplied 88.3% of 2024 small-business originations by count but 70.7% by dollars. That calculation includes Federal Reserve-supervised banks, uses one year of assets and excludes loan purchases, so it is not an estimate of the OCC-FDIC proposal. It does show that even two loan measures from the same bank-size group can differ sharply. https://www.ffiec.gov/data/cra/findings-from-2024-data-fact-sheet

The latest nationwide CRA tables cover lending during 2024 by 731 reporting institutions supervised by the OCC, FDIC or Federal Reserve. They separate metropolitan and nonmetropolitan locations and sort census tracts by income relative to the surrounding metropolitan area or the nonmetropolitan part of the state. https://www.ffiec.gov/data/cra/findings-from-2024-data-fact-sheet

Combining the low- and moderate-income rows, nonmetropolitan places accounted for 47.4% of reported small-farm loans by count and 55.0% by dollars. The corresponding shares for small-business lending were 5.6% by count and 6.4% by dollars. https://www.ffiec.gov/sites/default/files/data/cra/tables/CRA-25-tables-1-5.pdf

Here is the arithmetic behind the lead finding. Nonmetropolitan low- and moderate-income tracts had 10,176 reported small-farm loans, compared with 21,480 across metropolitan and nonmetropolitan locations combined. Their reported farm lending totaled $777.139 million out of $1.413428 billion. The FFIEC tables state dollar amounts in thousands, and the percentages were calculated from the unrounded values. https://www.ffiec.gov/sites/default/files/data/cra/tables/CRA-25-tables-1-5.pdf

The result is not limited to lower-income tracts. Across all tract-income groups with known locations, nonmetropolitan places supplied 59.0% of reported small-farm loans and 65.8% of farm-loan dollars. The corresponding small-business shares were 8.3% and 9.7%. https://www.ffiec.gov/sites/default/files/data/cra/tables/CRA-25-tables-1-5.pdf

Nonmetropolitan is not a synonym for rural. It means outside a metropolitan statistical areametropolitan statistical areaA metropolitan statistical area is a group of counties centered on an urban area of at least 50,000 people and linked by commuting and other economic ties. https://www.census.gov/programs-surveys/metro-micro/about/glossary.html, so the category can include farms near small cities and micropolitan communities. The chart establishes where reported loans were located under the official classification, not that every loan financed a remote rural operation. https://www.census.gov/programs-surveys/metro-micro/about/glossary.html

The OCC and FDIC released the proposal on July 31, 2026. A bank would generally be a large bank only if it had more than $10 billion in assets on December 31 of both preceding calendar years. Using 2024 and 2025 Call ReportCall ReportA Call Report is the regulatory financial report from which the agencies took the year-end bank asset figures used for the proposed size test. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf data, the agencies estimated that 86 banks, or about 2.4% of the banks in their analysis, would qualify and would hold 85.4% of total industry assets. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf

Large banks would still have to collect loan-level information and report annual aggregated small-business and small-farm data for each census tractcensus tractA census tract is a small, relatively permanent statistical subdivision of a county used to publish local data. https://www.census.gov/programs-surveys/geography/about/glossary.html where they originated or purchased loans. Proposed small and intermediate banks generally would be excluded from those requirements unless they chose an evaluation framework that carries the large-bank data requirements. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf

The proposal's impact analysis estimates that 126 OCC-supervised banks and 250 FDIC-supervised banks currently treated as large would be reclassified as intermediate banks. The agencies describe this as a compliance-cost reduction, but those bank counts still do not reveal how much LMI farm lending the reclassified institutions currently report. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf

The Federal Reserve BoardFederal Reserve BoardThe Federal Reserve Board oversees CRA compliance at state-chartered banks that belong to the Federal Reserve System, and it did not join this OCC-FDIC proposal. https://www.federalreserve.gov/supervisionreg/consumer-compliance.htm is not an issuing agency for this proposal. The proposal therefore would not itself alter the Board's CRA reporting rules for state member banks, even though those banks appear in the nationwide 2024 FFIEC totals used for context here. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf https://www.federalreserve.gov/supervisionreg/consumer-compliance.htm

A proposal-specific estimate must classify banks before adding their loans. For a counterfactualcounterfactualA counterfactual is a structured hypothetical calculation of what the 2024 data would have contained if the proposed eligibility rule had already applied. applied to 2024 lending, the analyst would test whether each OCC- or FDIC-supervised reporter exceeded $10 billion at both preceding year-ends, then total the qualifying banks' reported loans separately by loan type, tract-income category and metropolitan status. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf

The public files contain the necessary building blocks. The transmittal file provides a respondent identifier, regulator code, Federal Reserve identifier and prior year-end assets. The disclosure file provides respondent and regulator identifiers, originations and purchases, small-business and small-farm records, income groups, locations, loan counts and dollar amounts. A second asset year must be joined from an earlier transmittal file or the Call Report data. https://www.ffiec.gov/sites/default/files/data/cra/flat-files/24FlatTransSpecs.pdf https://www.ffiec.gov/sites/default/files/data/cra/flat-files/24FlatDiscSpecs.pdf

The useful outputs would be separate coverage ratios for business and farm loans, counts and dollars, low- and moderate-income tracts, and metropolitan and nonmetropolitan places. The agencies' notice reports institutional classifications and asset coverage but does not report those loan-coverage ratios. Until that calculation is published or independently completed, the proposal's exact disclosure loss remains unknown. https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf

The 55.0% result identifies where the current LMI farm-loan dollars sit, not which banks supplied them. If the banks that would lose mandatory-reporting status contribute disproportionately to nonmetropolitan LMI farm lending, a national asset statistic could conceal a concentrated data gap. If those banks contribute little such lending or continue reporting voluntarily, the eventual loss could be small. Both outcomes remain consistent with the published asset figure.

There is a useful historical precedent for doing the harder work. When regulators considered another CRA size-threshold increase in 2005, Federal Reserve researchers used public lending data to identify banks whose status would change and the local banking markets most exposed to that change. That study analyzed a different proposal and data year, so its results cannot be carried forward, but its method shows why bank and market location matter. https://www.federalreserve.gov/pubs/bulletin/2005/spring05_research.pdf

Voluntary reportingVoluntary reportingVoluntary reporting means a bank supplies CRA loan data even though its size does not require it to do so. https://www.ffiec.gov/data/cra/findings-from-2024-data-fact-sheet may preserve some observations. In 2024, 61 of the 731 CRA reporters had assets below that year's mandatory threshold and reported voluntarily or because they elected large-bank evaluation. That experience proves voluntary reporting occurs, but it does not tell us how many newly reclassified banks would continue after this proposal. The FDIC's paperwork analysis assumes that 10% of reclassified FDIC banks would report voluntarily for burden-estimation purposes, which is an agency modeling assumption rather than an observed outcome. https://www.ffiec.gov/data/cra/findings-from-2024-data-fact-sheet https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf

If a bank stops filing CRA loan data, the immediate measured effect is less public information. It does not follow that the bank stopped making loans. Reporting incentives and lending incentives can move together, separately or in opposite directions, so the two outcomes must be measured independently.

The 2024 CRA files describe completed originations and loan purchases. They do not contain unsuccessful applications, precise borrower revenue, borrower demographics or the overall demand for credit in a place. The FFIEC specifically warns that different loan volumes can reflect different local credit demandlocal credit demandLocal credit demand is the amount and type of borrowing that businesses and farms in a place want, whether or not lenders approve their applications. https://www.ffiec.gov/data/cra/findings-from-2024-data-fact-sheet. https://www.ffiec.gov/data/cra/findings-from-2024-data-fact-sheet

Section 1071Section 1071Section 1071 is a separate small-business credit reporting rule that covers applications, including applications that do not become loans, with compliance now scheduled to begin on January 1, 2028. https://www.consumerfinance.gov/1071-rule/ is intended to provide broader small-business application data, including information on applications that do not become loans. The CFPB's May 2026 revision moved its compliance date to January 1, 2028. The OCC and FDIC say they hope eventually to use Section 1071 data to reduce duplicated CRA reporting, but data not yet collected cannot establish the immediate proposal-specific coverage result. https://www.consumerfinance.gov/1071-rule/ https://www2.occ.gov/news-issuances/news-releases/2026/nr-ia-2026-64a.pdf

A credible test of a lending effect would wait for post-rule observations and compare lending changes at affected OCC- and FDIC-supervised banks with changes at similar unaffected banks, while accounting for local conditions and credit demand. Even then, the comparison would require assumptions about whether the groups would otherwise have followed similar paths. This article performs no such causal test.

• Publish the share of current small-business and small-farm counts and dollars attributable to banks that pass the proposed two-year asset test.
• Report those shares separately for metropolitan and nonmetropolitan low- and moderate-income tracts.
• Show the result with and without Federal Reserve-supervised reporters so the proposal's actual regulatory scope is visible.
• State and test assumptions about voluntary reporting by reclassified banks.
• After any final rule takes effect, measure disclosure changes separately from changes in lending.

The defensible answer is therefore narrow. The 85.4% figure accurately describes asset concentration in the agencies' proposed large-bank group. It does not establish how much geographically detailed LMI farm-loan reporting would remain. Current data show why the missing calculation matters, but they do not tell us its result.