We are now a large bank with new reporting criteria. Our first exam last September (as a large bank) was rather difficult - we had to re-do all of our loans while examiners were on-site. I am looking for suggestions to improve our data collection, manage reviews and reporting of correct information. This is not a topic I see discussed or training offered on and I get that it depends on what your institution decides to develop in terms of a CRA Program but I need some direction. What works for you?

My bank is at $4billion in assets and growing (acquiring) with 42 locations.

Currently, I review 100% of our CRA loans and we are at a 58% error rate due to incorrect info provided and input errors. Obviously this is unacceptable. What have you done to bring your institution up to par? I am working on a training program to address our data collection issues and we do use Centrax software to collect, review and analyze. There are 3 people in my department - the other 2 handle Community Development, Philanthropy and Services.

We just acquired another bank and will be on this track for many years to come. It is imperative that I get things in order now. Thanks in advance for your help and suggestions!

What type of data integrity errors are you experiencing? Are the loans coded incorrectly? The call code combined with loan amount drives whether a loan is small business or small farm.

A problem with that is a basic bank load boarding issue and NOT a CRA issue.

Identifying gross revenue and proceeds address are the primary errors.

Those truly are training and systems issues. Some banks, to avoid the proceeds address issue, decide to go with the main business address and let the chips fall where they may as to location of the borrower. It can save a great deal of time and thus money not chasing down addresses. If you do use a proceeds address, it should be collected in your system and not on paper.

Gross revenue is looking for exactly that, gross sales, not a calculated number. The only question then is for "small entrepreneurs" what to use.

Sounds like we need to make decisions on what we want to report and train on it. And as you say, let the chips fall where they may!

The only downside is if you end up with less geographic distribution of loans in low/mod tracts, but that is offset by no process criticisms, resubmissions, etc. for data integrity issues.