We are a large bank and required to report for CRA. I am wondering if large banks code their CRA loans in the core system so that it transfers into your CRA software correctly. There is purpose code, collateral code and call report codes that we all use but when you don't count certain types of loans for CRA and they transfer over as CRA reportable it can be reported wrong.

For example, we have chosen not to report letters of credit but they are included in my extract from our core system. There is nothing that distinguishes these loans from any other small business loan.

Another example is loan modifications - all of these are extracted as small business loans and may or may not count as CRA reportable.

I think if all our CRA loans are coded (maybe a purpose code) it would help.

Ah! You have identified all of the issues! Using an extract from your core system will work if you have confidence that the coding is correct. This means that the individuals that are responsible for the input know the importance of the data and how if will be used. Someone with equal experience needs to perform the callback.

As for Letters of Credit, you could establish a bank use field to identify these or have another means to identify them and pull them out. You could do the same for modifications. Talk to someone in your bank that has a good understanding of your core system. Some systems have "flex fields" that can be used for whatever you need.

Personally, I have never relied on others when it comes to the accuracy of CRA data. As the CRA Officer I am responsible for the data, so my department reviews every small business loan prior to submission in March. Yes, it's a lot of work depending on the size of the bank, but it's a lot less painful than doing a scrub at the direction of your regulator. Ditto for HMDA. Also, most audit firms will inquire about your data integrity testing, so you really can't rely 100% on other parties for the integrity of the data.

Because small business and small farm loans rely totally on Call Report classifications, and the Call Report is supposed to be right, I am a big believer in getting that right in the first place. I have worked at very large banks where, because there was a great emphasis on getting the loan purpose, collateral, and and Call Codes correct, the CRA reporting was NOT difficult. Yes, there was testing, but there was a great deal of testing and cleaning on the credit side to make sure that THEIR loans were being reported correctly on the call report...that then of course flowed to CRA reporting.

Spend time getting the loan amount fields right...pulling note amount and not outstandings for lines of credit for example.

Some banks also have added fields for the CRA categories so that items like letters of credit can be classified as CRA 9, etc.

If those loans are not being categorized correctly for the bank in general, the bank has reporting problems beyond CRA that can lead to problems.

Quality control for all coding is essential.....not just people slapping on codes as their best guess and then the codes stand. Someone else needs to go behind the person who assigns the codes and verify that the codes are correct.

Just as you state, we extract collateral, purpose and call report codes from the core system. We also have an alpha code in our core system that distinguishes the type of loan (e.g. – small business, small farm, non-profit, municipality, etc.).

We prefer to extract as many loans as possible then eliminate from there. For example, some shops do not extract loans to non-profits. We import them into our software to: 1) isolate for possible Community Development Loan consideration and 2) for possible consideration for small business reporting. Too often we see non-profit loans that are secured by an owner occupied structure such as a synagogue, church, or office, which would need to be reported as a small business loan. If we excluded loans to non-profits in the extract, there may be missed opportunities for community development lending credit and possible under reporting if the loan should be reported as a small business loan. Similarly for loans to cities and towns. Extract the loans, identify/isolate for possible community development consideration then delete from the small business file. This is also where purpose and collateral codes help.

We even extract a loan if the dollar amount is greater than a million for a small business or greater than $500,000 for a small farm. Knowing that we cannot report these, we review for community development consideration then deleted from the small business file. If we don’t remove the loans from the small business file, a systemic edit isolates these non-reportable loans, which then reminds the user to take some sort of action before submission.

We extract land development loans, which are not CRA reportable. The reason is that often loans secured by farmland are miscoded as land development loans. We look at the purpose, collateral, and possibly the loan file to determine if the loan is reportable as a small farm loan. All true land development loans are then deleted from the Small Business file.

Even though Letters of Credits are not CRA reportable, we still extract them from the core system for possible inclusion in “other lending” data for the regulators. These are removed from the small business file once they have been reviewed and considered for other lending data.

We have no way to distinguish our modifications either. In short, we touch every loan. We look for a change in the maturity date and if a loan has an extension in term we keep it in the small business file, otherwise the modification is deleted.

Hope this helps.

My only question, Tennismom, is if you are pulling out loans to non profits that are secured by non farm non residential real estate of $1 million or less. Those are small business unless you are ISB and moving to CDL.

But they should still be on call report as small business.

Kathleen - Miscoding of loans is commonplace in our shop. That why we extract and import ALL loans to non profits based on CALL Report line item #. We then review to determine if eligible for CDL credit, or to see if it is possibly secured by non farm non residential RE. If eligible for CDL credit the loan is deleted from the Small Business File and placed in the CDL File. If secured by non farm non residential RE the loan remains in the Small Business File as a loan type1. If the loan does not meet either of the above criteria it is deleted from the Small Business File.

Years ago discussions with our finance department went much like "We never have had anyone internal or external audit the CALL Report and so we don't worry about misclassified loans" With that mindset, all we can do is ensure that the LAR is correct and document the reason why we have a loan on the CRA LAR that is not accurately reflected on the CALL Report.

Ha ha. I have lived through a Call Report scrub and post scrub ongoing reconciliation that was required from an exam. It is a treat, believe me.

If the coding is not good, then you have no foundation to work with. I hope NAICS codes are right for concentrations, etc.

I often see bank data that makes no sense based on assigned codes. Bad coding adds so much extra work

It never ceases to amaze me when management chooses not to do something correctly simply because they have never been hurt by doing it incorrectly. Seems they should want to do everything correctly because doing it correctly is the right thing to do. Unfortunately, live and learn is usually the lesson to make this type of management finally conform.

@ Trying to comply - we are trying to get away from a manual process but I just don't see that happening. How big is your bank and how many people in your department are dedicated to scrubbing CRA loans?

Thank you everyone for your comments. It helps to know the problems I am seeing are commonplace.

Imagine a "really big bank" and Compliance trying to scrub all of the loans. Hence the need as a bank grows to make the business line or Credit Admin responsible for accurate coding. How much of a compliance staff is management will to fund?

I came from a fairly large bank with a large secondary market lending operation. There was NO way a staff of two could have looked at every application file. So individuals in the Credit Department were train specifically for coding and verified that all codes were correct. This worked well for us.

That is what keeps running through my head - how do the really big banks do it? How do they pass CRA Data Integrity unless they have a group of people dedicated to review loans. Exam results fall on CRA department so it seems prudent for us to scrub. We are in growth mode so if we don't do well for CRA exam then we could potentially not be allowed to continue acquiring banks.

We have a lot of training to do and hopefully that will improve our data integrity so that moving forward we will eliminate the need to review all CRA loans. In the meantime, gotta do what we gotta do!

Bad data results need to fall on the department charged with accurately boarding the loans. Accurate codes are part of loan set up/boarding!

Kathleen said that better than I did. We had a dedicated small group of trained individuals who verified the codes prior to booking.