The Community Reinvestment Act (CRA) aims to ensure that lending institutions provide small businesses with critical access to funding and support the needs of the communities in which they operate. Lenders who meet a certain asset-size threshold for two consecutive years must submit their previous year’s lending records for evaluation by March 1 of the following year. Banks with more than $1.564 billion in assets in each of two consecutive years are subject to CRA requirements; this threshold changes annually.

Assessing the Challenge

To pass a CRA examination, accuracy is paramount. Institutions must correct and resubmit CRA small business and small farm data when at least five percent of the data collected and maintained were recorded incorrectly. Institutions must also resubmit if data for five percent or more of the loans do not meet the definition of community development.

ADI’s client, a Bank with nearly $5 billion in assets, exceeded the five percent threshold and was therefore subject to reexamination with only a few weeks to prepare.

Designing the Approach

Our client’s LAR showed errors in three CRA fields, and therefore, ADI performed three types of audits, or scrubs, for the client:

  • Business/Farm Gross Annual Revenue
  • Loan Amount
  • Census Tract

To compare the information in each borrower’s loan file with the data recorded on the Bank’s CRA LAR, we deployed a custom online platform designed specifically for CRA audits. ADI imported the lender’s CRA data into the database, and the client securely transferred imaged loan files to ADI.

Using the platform, data analysts recorded data values from the imaged loan files. Then, ADI compared the analysts’ findings to the Bank’s reported field values. When results did not match, we conducted second and sometimes third reviews. ADI then provided a report to the client, in which we identified the loans and data fields with inaccurate entries and their correct values.

The third scrub, in which we reviewed the Census Tract data for locations at which each loan’s proceeds were used, required a unique solution. The Federal Financial Institutions Examination Council (FFIEC) geocode map, which provided accurate Census Tract data for loans originated in 2022, 2023 and 2024, based on the 2020 Census.

However, our client submitted an older dataset that used 2010 Census Tract information unavailable on the FFIEC map at the time. Using publicly available mapping software, we downloaded data from the United States Census Bureau and provided visual evidence of the property address and its 2010 Census Tract for each loan file.

Implementing the Solution

For the Business/Farm Gross Annual Revenue scrubs, conducted on CRA data submissions for three different years, ADI reviewed each loan file’s credit approval memo. Reviewers then looked for documentation to support reported gross annual revenues (GAR). We consulted with the lender to understand its process for determining GAR, and we scrutinized the loan files to find documentation to support the reported value (Less than $1 million; More than $1 million, or Unknown).

Because CRA examiners look for tax documents such as Form 1120 for entities and Form 1040 Schedule C, E or H for individually held businesses, we asked the client to provide this documentation. Determining the revenues for start-up businesses required extra scrutiny, as the Bank did not consider pro-forma and projected revenues as part of the credit decision.

For the Loan Amount audit, which included a single year’s filing, the challenge was to differentiate among loans for new money, renewals, and changes in maturity dates or principal balances, using the promissory note or Change in Terms documents found in each loan file.

ADI’s consultants reviewed the client’s data as originally submitted on the Bank’s CRA LAR, as well as source documents such as loan approval memos and tax documents. Analysts captured the field values found in the loan files and again compared those values to the ones in the CRA resubmission data file. The project manager ran reports that returned exceptions for the client to review.

If the client disagreed with ADI’s findings, the client had the opportunity to provide additional evidence to support the original data. However, in this case, the client agreed with our findings and the LAR required no further changes.

ADI’s timely, targeted reviews provided the client a way to resubmit three clean CRA LAR files within a short time frame.