Federal Deposit
Insurance Corporation

June 27, 2005

Office of the Comptroller of the Currency
Board of Governors of the Federal Reserve System
Federal Deposit Insurance Corporation
Office of Thrift Supervision

Dear Sirs:

The purpose of this letter is to comment on the proposed changes to the supervisory framework for the classification of commercial credit exposures as discussed in OCC Bulletin 2005-8 and the attached narrative (70 FR 15681). I am speaking on behalf of Central Bancompany, a $7 billion community bank holding company based in Jefferson City, Missouri.

My responsibilities for the company include oversight of the internal loan review, internal audit, and compliance functions. My career in these functions spans twenty-three years.

On behalf of our company’s senior credit risk management team, I would respectfully suggest that the proposed changes should not be adopted.

The proposed two-dimensional rating system has existed in one form or another for a number of years, having been adopted by a number of large financial institutions and discussed regularly in trade publications such as the RMA Journal. As head of our company’s loan review function, I have studied these ratings systems with some interest.

There is a great deal of superficial appeal to these more-sophisticated systems. It is absolutely true that the risk of borrower default is a factor relevant the risk in the credit. It is equally true that each facility to a given borrower carries a different level of exposure. I am most familiar with the equation:

Probability of Default (PD) X Loss in the Event of Default (LED) = Expected Loss (EL).

Is this a better system? Discussions with some who have such a system and our own experiences tell us that it is not. The new system would:

• Create more work for analysts;
• Create more frustration for lenders; and
• Not result in more-reliable estimates of portfolio risk.

First, let me suggest an analogy and then address each of these issues separately. The proposal is much like the proposal when I was young that the United States adopt the metric system. At the time, I heard a farmer comment acidly that they could change his acreage to hectares if they wanted, but it would be the same amount of land and the yields wouldn’t be any better. The same should be said of this proposal: there will be the same amount of risk as before and the aggregate loss estimates won’t be any better. We come to these conclusions for the following reasons:

The proposal would create more work for analysts:

Analysts in this context would include not only bank credit analysts, but also analysts employed by the government; i.e. bank examiners. We believe that the required extra level of detail will create more work for both. Added to this will be the cost of the more sophisticated loan systems used to store and analyze the data. And to this will be the time and cost of studying and understanding all the “Issuances” that will inevitably be necessary to clarify and revise the original “Issuance” adopting the proposal.

Although as proposed, the requirements would pertain only to borrowers rated default, we think it probable that the requirement would eventually be expanded to include a larger segment of the portfolio, perhaps the entire portfolio. Every time the scope is expanded, “Issuances” will be necessary to clarify the revisions.

The proposal would create more frustration for lenders:

My experience is that most credit analysis is performed by “techies” and that lending is reserved for those with good business judgment. The techies of the world – the math majors, accounting majors, and others with strong numbers aptitude – tend to like more sophisticated ratings systems. They create beautiful spreadsheets and analysis based on historical information and impressive mathematical models. Then, when a recession hits, all of their numbers are shown to have relatively little predictive value. The building with a $5 million dollar appraisal is worth only $3 million when it becomes vacant. The machinery & equipment with a well-supported appraisal sells for only 50% (or less) of the appraisal. The impressive analysis fails to predict the losses that are eventually realized.

Lending, on the other hand is for those with good business judgment, but not necessarily as much mathematical sophistication. The best lenders do perform a careful review of a proposed customer’s financial condition before agreeing to a loan, but their basic analysis is much simpler:

• Do we want to do business with this person?
• Will he be able to repay the loan? and
• If he doesn’t, can we take the collateral back and be made whole?

A lifetime of business experiences is summed up in their answers to these three simple questions. When internal analysts and examiners start talking with lenders about the probability of default, the loss in the event of default, and the expected loss, the lenders of the world – even the good ones – will immediately want to get to the bottom line: what should “the” rating be?

If is for these reasons that we foresee the more sophisticated rating system creating more debates and frustration between lenders and examiners, rather than less. I think a lot of lenders will figuratively “throw up their hands” and say “you rate the credit.” This would be an unfortunate consequence of something that is theoretically very good.

The truth of the matter is this: Estimates of commercial loan credit risk are inherently and unavoidably very rough estimates. It is impossible to make them precise estimates because future economic events are inherently unknowable. There is very little value added when we try to make these estimates more precise than they are now. FASB 5 and FASB 114 along with related pronouncements have made this process about as exact as it can be.

The proposal would not result in more-reliable estimates of portfolio risk:

It is our experience that the most reliable indicators of credit losses are (a) the tolerance of loan area management for risk and (b) the condition of the local economy.

We have thirteen affiliate banks in our organization. In two instances, we had abnormally high credit losses shortly after management of the lending area took a more aggressive lending posture (and before internal loan review or company management could slow them down). We do not think that the proposed more-sophisticated rating system would have prevented – or even ameliorated – these situations.

In one additional instance, one of our affiliate banks was in a community that had a localized recession because the major industry of the community went through a serious downturn. Again, we do not think the loan losses we experienced would have been prevented by a more-sophisticated ratings system. The value of our collateral declined quickly and precipitously, an effect that would not have been predicted by separate facilities ratings.

* * * * *

In weighing these considerations against the positives we see in the proposal, we believe that scale balances in the direction of staying with the current rating system. The current system is simple and understandable and results in as good an overall evaluation of risk as would be the case with the new approach.

Thank you for the opportunity to provide input. We at Central Bancompany are vitally interested not only in the safety and soundness of our own institution, but of the entire banking system.

Respectfully,

Richard Popp
Sr. Vice President, Audit & Loan Review