Are you in compliance?

Don't miss out! Sign up today for our weekly newsletters and stay abreast of important GRC-related information and news.

Why the Current Proposal on Loan-Loss Accounting Is Flawed

Scott Taub | May 21, 2013

In January 2011, FASB and IASB issued a proposal regarding loan losses that suggested separating loans into a “good book” and a “bad book” based on whether each loan had an indication of a credit loss or not. The 2011 proposal would have required a reserve for all expected losses on the bad book and for some expected losses on the good book. The approach would have eliminated the need for a loss to have actually occurred before it was recognized.

At the time, I wrote about my concern that FASB seemed to be abandoning the concept of recording things as they happen, not before (or after) they occur. When FASB decided not to move ahead with the 2011 proposal I was at first optimistic, but this was entirely unwarranted. Not only did the replacement proposal, issued late last year, keep the goal of recognizing losses before they occur, but it actually doubled-down on that notion, requiring that all expected losses be recorded, even for “good” loans.

If the 2012 proposal is finalized, it will require companies to recognize all expected losses for every financial instrument not carried at fair value. They will have to calculate expected losses at the discounted value of any contractual cash flows not expected to be received. In making this estimate of expected losses, the rules will require an entity to consider the possibility that cash flow may not be received, even on highly-rated assets. As such, a reserve of zero would never be appropriate for any financial asset.

This new model addresses concerns that current requirements result in recognizing losses too late. But it does so at the cost of abandoning concepts that have underpinned Generally Accepted Accounting Principles (GAAP) for decades. The figures reported would not faithfully represent the economics of the financial assets. Not only are the proposal's requirements decidedly conservative and biased, they would foster earnings management by financial institutions, since they require the build-up of rainy-day reserves based on highly subjective estimates, with no conceptual anchor or limit.

There must be a better way.

It is true, of course, that every loan or debt security, even those with a creditworthy counter-party, could result in some losses. The lender considers the possibility of such losses when deciding what interest rate to charge, just as investors do when they decide how much to pay. If the expectation of loss has not changed since the origination or purchase, it is already accounted for in the cost basis of the asset. Still, the 2012 proposal would require the holder to recognize the expected losses again, by setting up a reserve. As a result, the rule would require a company to carry a loan or debt security whose expected losses have not changed since it recognized the asset at less than fair value.

FASB justifies this result in part by noting that a lender or investor is exposed to a risk of loss that it wasn't previously exposed to immediately when it originated or purchased the instrument. This proposal would change the timing of loss recognition for loans and debt securities from when the loss occurs to when the entity becomes exposed to the risk of loss. While this concept might sound OK at first blush, it is one that FASB has rejected numerous times. Just consider the blunt language from FASB Standard No. 5: “In the board's judgment, however, the mere existence of risk, at the date of an enterprise's financial statements, does not mean that a loss should be accrued.”

“This new model addresses concerns that current requirements result in recognizing losses too late. But it does so at the cost of abandoning concepts that have underpinned Generally Accepted Accounting Principles for decades.”

In addition, the proposal conflicts with the accounting for pretty much every other asset. If we apply the same principles to inventory, for example, a retailer would recognize expected theft losses when the inventory is acquired, not when the theft occurs, because it is upon acquisition that the retailer is exposed to the risk of inventory shrink. Similarly, a company that owns and operates a fleet of trucks would record losses to account for expected damage to the trucks from accidents that haven't happened yet. If these suggestions seem silly to you, you are beginning to understand why I don't want to record expected credit losses before they happen.

And what about the upside? While an entity is exposed to the risks of a loan or debt security, it is also exposed to rewards from that instrument in future interest payments. The proposal would require the entity to record losses related to the risks with no consideration of the rewards. This one-sided consideration of the economics of the investment seems to conflict with the conceptual framework.

Concepts Statement No. 8 states that financial information should be neutral in order to be representationally faithful. I don't know how accounting that requires that future losses be recognized currently while future income related to the same instruments is ignored could be considered neutral. Similarly, I don't know how systematically valuing an asset below fair value (and below cost) could be considered neutral.

Again, there simply must be a better way.

How to Fix It

There are basically two problems with today's incurred loss model. First, we don't reserve for a loss until we think it is “probable” that a loss has happened. Because “probable” in GAAP is generally interpreted to mean something in the neighborhood of a 75-to-80 percent chance of occurring, the current loan-loss model essentially doesn't record losses until the entity is almost certain they have been incurred. This leads to the justified criticism that the current model records losses too late.

This is easily fixed. We could simply require recognition of a loss when it is “more likely than not” that a loss has been incurred. This would increase reserves by requiring a balanced assessment instead of one that is biased toward not recognizing losses, while also clarifying that losses should be recognized even if a payment default is not imminent.

The second problem with the incurred loss model is that it is hard to know when a loss on a loan actually occurs. Indeed, even in situations where the event that caused a loss is clear in hindsight, the lender might not know about the event in real time. The 2012 proposal would solve this by ignoring the question of when the loss actually happens. It wouldn't matter that we have imperfect information about incurred losses, because we'd record the loss in the financial statements before the actual loss anyway.

I don't believe this is the right resolution. Instead of spending any more time justifying the move away from time-tested principles and concepts, FASB should focus on providing guidance to help with the “incurred but not reported” aspect of credit losses. That approach would encourage consistency, and allow FASB to again stress that “incurred” doesn't mean “wait until a payment isn't received,” which, if certain complaints about the incurred-loss model are to be believed, is what some institutions do today.

FASB could also consider other options for dealing with the issue of losses that are incurred but not reported (IBNR). The 2011 proposal required ratable recognition of certain losses and a floor on reserves. Both could be used as methods of dealing with the IBNR issue. Ratable recognition of expected losses could be explained as dealing with the reality that many events taken together typically lead to a loss. Perhaps the best estimate of when a loss actually occurs is ratably over the life of the loan. A floor, on the other hand, could be explained by noting that, on average, a period of time elapses between the time a loss on a loan occurs and the time the lender becomes aware of the loss. To account for this delay, we could assert that a 12-month look-forward (or some other time frame) reserve is necessary in order to capture IBNR losses.

While ratable recognition and a floor wouldn't be my preferred methods of getting at the IBNR losses, they could still be explained and used within a model that retains the logical goal of recognizing losses when they occur. This would be far better than moving to expected losses. The implications of embedding a principle into any part of the accounting literature to recognize something before it happens are troubling, and abandoning concepts that have long underpinned our accounting model would do far more harm than any good that could come out of having “rainy day reserves” set up before the next economic downturn.

The need for larger loan-loss reserves in general and so called “rainy-day reserves” in particular is often voiced by those concerned with bank regulation. The incurred-loss model has also been called insufficient by the same people. I don't see what relevance that has to FASB's considerations. It shouldn't surprise us that what works for a financial reporting system focused on efficient allocation of capital may not work for prudential regulation focused on safety of the banking system. This does not imply a problem in either system. It merely acknowledges what should be obvious—different measures are sometimes necessary to achieve different aims. Bank regulators can set whatever requirements they believe are appropriate and can use or disregard GAAP measurements as they deem appropriate. There is no need to mold financial reporting to the needs of prudential regulators, and doing so is dangerous.