Large banks are facing a new but familiar challenge—ensuring clean and accurate data reporting—as they adjust to revisions in the Community Reinvestment Act (CRA), which includes the collection of small business and small farm data.
Banks and regulators have been dealing with similar reporting issues for several years under the Home Mortgage Disclosure Act (HMDA) and have gradually improved their data accuracy. However, testing the integrity of HMDA and CRA data is still at the forefront of compliance examination concerns for large banks.
In reviewing a bank's performance under the CRA regulation, the examiners primary tool is the data reported on the loan application register for HMDA and the loan register for CRA. The CRA regulation's focus on lending—and in particular, lending to low-to-moderate individuals and areas—gives the data produced from lending records utmost importance. Credibility with examiners, as well as community groups that may analyze the data, could be compromised if inaccuracies occur.
A recent OCC advisory letter (98-16) describes common mistakes banks make when recording data. They include:
1. Human error, especially keystroking mistakes. Simply pressing the wrong button on the computer keypad can cause the incorrect reporting of key information such as the borrower's address or income—thus misrepresenting the neighborhood or income category to which the loan was made. In addition, if the person keying in the loan does not enter all of the required data, analysis and submission are likely to be difficult or delayed.
2. Misinterpretation of the actual reporting requirements. In the case of a small business or small farm, the loan recipient's address is often reported incorrectly. CRA regulations say, "The institution should record the loan by either the location of the business headquarters or the location where the greatest portion of the proceeds are applied, as indicated by the borrower." Often, the lender will record a home address of the small-business borrower or another location, which is neither where the headquarters are located nor where the proceeds of the loan will be used.
3. Geocoding. All loans should have addresses that can be geocoded to indicate the location (i.e., the census tract or block numbering area) of the loan. Borrowers' addresses in rural areas frequently consist of only rural route numbers or post office boxes. However, the new CRA Q&As state that "prudent banking practices dictate that an institution know the location of its customers or loan collateral. Therefore, institutions will typically know the actual location of their borrowers or loan collateral." Once an institution has this information available, it should assign census tracts or block numbering areas to the location (geocode) and report it.
The Q&As do give an exception to this "if the institution cannot determine the borrower's street address and does not know the census tract or block numbering area." The institution should then report "the borrower's state, county, MSA, if applicable, and 'NA,' for 'not available,' in lieu of a census tract or block numbering area code." This option should be used only after all efforts to geocode have been exhausted, including the use of census tract maps.
4. Reporting the renewal of a small-business loan as an origination. In the case of a credit-line increase, many institutions report the total amount of a line of credit instead of the increase alone.
When it comes to preventing data integrity problems, the stakes are high. In fact, some institutions have spent upwards of a million dollars cleaning up their HMDA data. Ensuring your institution's data is clean will prevent you from spending unnecessary time and money "scrubbing" HMDA and CRA data before submission.
Ultimately, the potential for errors is large, and data integrity issues will persist unless bank management gives proper attention to the problem and takes corrective action.
Keep up with what’s new and noteworthy at the St. Louis Fed. Sign up now to have this free monthly e-newsletter emailed to you.
Fed in Print: An index of the economic research conducted by the Fed.