This is my third consecutive post on the perils of loan loss reserves (first here; second here); and a fourth is going to follow very soon. My fourthcoming (I couldn't resist the pun J) post will be based on recent coverage of Chesapeake Energy Corp's natural gas reserves reporting; I believe there are lessons to be learned about how to do a better job of measuring loans from the experiences of energy exploration companies with estimates of their reserves.
But, let's not get ahead of ourselves. The motivation for this current post comes from an anonymous* reader of my previous post, who took umbrage with the following:
Dare I now hope that the FASB will ultimately rediscover the wisdom of measuring loans at market-based values? I would be more optimistic that reason might prevail if the newest three board members hadn't been specially chosen by the powers that be at the Financial Accounting Foundation to be dependable opponents of market values for loans.
That last bit in italics was what set my reader off. He suggested that I would think differently about at least one newbie Board member, R. Harold Schroeder, if I were to read the comment letter Mr. Schroeder submitted (before joining the Board, of course) to the exposure draft on accounting for loans, derivatives and hedging.
So I did.
On a Dual Statement Approach for Loans
Even though Mr. Schroeder and I disagree on a many aspects of loan accounting, we agreed on at least two things:
- The current loan accounting model is broken;
- Amortized cost proponents are wrong to claim that measuring market values for loans is impracticable.
Before addressing the portion of the comment letter that I take the most exception to – a mistaken rationale for the expected loss model – I need to discuss Mr. Schroeder's main recommendation: the FASB should explore a dual (or "side-by-side") financial statement presentation, whereby one full set of financial statements for a bank would be based on an expected loss model, and a supplemental balance sheet would be based on fair values.
Some kind of dual statement approach merits consideration, but I have two general concerns. First, as Mr. Schroeder stated, the SEC would probably have to amend Regulation S-X to set forth requirements for the new financial statement that Mr. Schroeder envisages, just as it sets forth the requirements for all other financial statements filed with the Commission. But, could the SEC handle the implications for the standard audit report of two balance sheets – that both could constitute a "fair presentation" of financial position? Also, which balance sheet (or "column") should articulate with the income statement?
(A slight digression—attaining resolution to any of these questions would require a collaborative approach to standard setting among the SEC, FASB and PCAOB, which hardly ever takes place. For this and other reasons, I have come to the conclusion that the FASB – like the PCAOB – should be directly responsible to the SEC, cutting the Financial Accounting Foundation out of the picture. Since S-OX took care of the funding of the FASB, I see no reason why, in regard to financial reporting for public companies, there needs to be an FAF.)
My second concern in regard to dual presentation is that Mr. Schroeder, as is typical of analysts, focuses on the 'information content' of financial statements and the effort it would take to recast a set of financial statements to his liking:
"The current proposal suggests a single-column, reconciliation approach that starts with amortized cost and follows with various fair value adjustments to arrive at fair value for each asset and liability covered by the proposal … [W]e find this format cumbersome and time consuming." [italics in original]
The letter does not address how analytical adjustments could be facilitated by appropriate XBRL coding; or more important, how dual reporting would affect decision making by managers, and consequently, shareholder value creation/destruction. Given the conventional wisdom (embraced by many corporate board members and executives) that the 'market' is fixated on earnings and financial leverage as portrayed in the financial statements as issued, information content should not always be the primary policy consideration. And surely, analysts are facile enough to recast the information to suit its tastes – so long as the data is available.
A Mistaken (and seemingly common) Rationale for the Expected Loss Model
Regardless of Mr. Schroeder's view of the role of fair (or market) value in financial reporting, his comments on the expected loss model contain a grievous error:
"Our long-held view is that there is an inherent risk of loss from the moment every dollar is lent. Some of those losses will result from a lender's initial underwriting or judgment errors, while others will result from subsequent changes in the economic environment, competitive landscape and other factors. We do not believe losses from any of these factors are 'unforeseeable.' Nor do we subscribe to the view that a lender would not make a loan unless they believe it will be fully collectible. Therefore, given the inherent nature of lending, all loans have some loss potential that should be given upfront recognition. …" [bold italics supplied]
To see the error in this reasoning, imagine that a bank lends $1,000,000 to a business at an interest rate of 10%, and that the interest rate charged by the bank corresponds to what any other bank would have charged for the risks that the loan presented.
Question: At origination, what should be the carrying amount – and market value – of the loan?
Answer: $1,000,000, which corresponds to the expected – i.e., not contractual – future payments from the borrower, discounted at 10%.
For some reason, everybody understands this logic when it comes to a traded bond, but it is not always carried forward to loans negotiated on a bank-to-borrower basis. Coincidentally, I had recently corresponded with a former Board member on just this question. Consistent with Mr. Schroeder's reasoning, he wrote this to me:
"I think an expected loss model makes sense logically. I always go back to the very simple example I was faced with in practice …. A bank client was getting into the credit card business and sent out a large number of applications with little credit checking. They said they were expecting a certain level of loss (say 3%) but were willing to take that in order to get into the business and begin earning what were then quite large interest fees on account balances. So when they made initial "loans" of, say, $10,000,000, they thought they should record a loss reserve of $300,000. But we told them that FAS 5 [Contingencies] didn't allow that and they had to wait until the balances aged sufficiently that they could demonstrate that there were losses actually incurred. This seemed to the client to lack common sense and I actually had to agree with them. But that was what GAAP said (and says)." [italics supplied]
I agree that a haircut to the principal amount of credit card loans is economically justified when the loans are specifically priced as a loss leader. In this case, the losses are associated with gains to be made on future business relationships with the 97% of new customers who are creditworthy. Even though the interest rate may be high, it may still be a windfall for some borrowers; thus the market interest rate could be even higher than the stated rate in the credit card contract.
But, the general case is my earlier example: a loan issued at the 'market' interest rate. I grant that bankers sometimes misestimate the market rate, but if one does, how could we expect that the same banker to be capable of producing a 'reliable' or 'reasonable' estimate the loss reserve for that same loan?
(And by the way, I also agree that a loan loss allowance on Day 1 could be justified for trade receivables, but that's only because the measurement basis is undiscounted expected future cash flows.)
* * * * *
I want my readers to know that I don't blog just for the thrill of being outrageous or to bash policymakers. I am passionate about accounting, and I write in the hope and expectation that my readers are willing to change their minds. Although I disagree with Mr. Schroeder on numerous points, I found that the views he expressed in his comment letter were the product of thoughtful analysis, which suggests that he will be considering other points of view on loan accounting with an open mind. So, thanks to the reader who suggested that I read it.
And, here's hoping that the extra time that the FASB is taking to contemplate the three-bucket approach to loan loss measurement will ultimately help them see the folly of the expected loss model.
*My longstanding policy is to not publicize private correspondence or conversations without permission. In contrast to a journalist, I take everything to be 'off the record' until I am told otherwise.