Why Better Loan-Loss Coverage Masks a Hidden Danger

WASHINGTON — One of the industry's biggest challenges in the crisis was storing reserves fast enough to keep pace with bad credits. But almost six years later, the industry's proportion of loss protection to noncurrent loans is finally gaining momentum.

The reason appears to have less to do with reserving strategy than steadily improving loan quality. Loan-loss reserves continue to fall but noncurrent loans are dropping at a faster clip, giving the industry's so-called "coverage ratio" a boost. The ratio of loan-loss reserves to noncurrent loans rose more than 2.5 percentage points in the first quarter to 67.8%, the sixth straight quarterly increase, the Federal Deposit Insurance Corp. reported last week. It is the longest streak since the ratio rose for 11 consecutive quarters between 2002 and 2005.

But a closer look reveals more questions about the industry's loss protection. Despite improvement in the coverage ratio over the past year and a half, it is still very low by historical standards. In the years leading up to the crisis, it had hovered around 150%. That suggests banks still need to raise their reserves for future downturns, especially as accounting regulators move toward a more forward-looking reserving model.

"If the ratio has gone up, that is taken as a good sign," but "the outlook is still cloudy," said James Barth, a finance professor at Auburn University and senior finance fellow at the Milken Institute. "It's difficult to draw strong conclusions about whether or not banks have a high enough coverage ratio."

 The coverage ratio had previously grown from quarter to quarter only four times from 2006 through 2011. Its lowest point in recent years was 57% in the third quarter of 2012. (It had reached as high as 179.3% in the late nineties.)

The FDIC's quarterly report on the industry's health indicated the coverage ratio's rebound is largely due to the persistent decline in noncurrent loans, and only because they fell more dramatically than reserves. Indeed, the largest positive factor on net income continued to be the release of reserves. The loss "provision" — or the amount set aside for reserves — was lower than that set aside a year earlier for the 18th consecutive quarter. Loss reserves dropped 2.6% from the previous quarter to $132.3 billion, the 16th straight decline.

But loans classified either as "nonaccrual" or at least 90 days past due fell at a faster rate. They also declined for the 16th straight quarter, dropping 5.8% from the previous quarter to just over $195 billion. Noncurrent loans previously had not dipped below $200 billion since the third quarter of 2008. The coverage ratio would also be significantly higher if adjusted for the fact that loans backed by Ginnie Mae, which can be classified as noncurrent, do not require banks to reserve against them because of their government guarantee.

"It's a positive sign," Robert Eisenbeis, vice chairman and chief monetary economist of Cumberland Advisors, said of the improving coverage ratio. "It shows that institutions are responding to a lot of the pressures and are restructuring their balance sheet."

But some observers said banks should continue to be aggressive in their reserving and not see the recovery as an opportunity to continually lower provisions. They noted that the sharp fluctuations in the coverage ratio between boom and bust periods is evidence that loss reserves should grow steadily even when the environment has calmed.

"When banks get into peacetime, they want to try to maintain some reserve coverage. But the past-dues are going down, the non-accruals are going down and people are paying off their loans. So here come the accountants and they want you to reduce the reserve, which is just ridiculous," said Timothy Long, a former senior official at the Office of the Comptroller of the Currency, where he was chief national bank examiner, who now heads up the regulatory risk practice at Protiviti.

Barth said the large swings between where the ratio stood before and during the real estate boom and where it is today prove that institutions should not be satisfied with their loss coverage.

"Despite it being quite high, we still had the worst financial crisis since the Great Depression. What appeared to be high wasn't high enough," Barth said. "Whatever it is, it is never high enough whenever we encounter serious financial difficulties."

Meanwhile, the entire system for how accountants measure loss reserves will soon likely be overhauled. After the crisis proved that predicting losses based on loans that already show some deterioration is inadequate, regulators and accounting standard-setters are moving toward adopting an "expected-loss" — rather than "incurred-loss" — reserving model. The Financial Accounting Standards Board has yet to finalize its December 2012 proposal on a more forward-looking model.

Movement toward the expected-loss model "should give banks more flexibility to maintain a strong reserve even in peacetime ... without the pressure of the accountants," said Long.

But others said being too cautious in a more forward-looking approach also carries risks.

"The question is: How many 500-year floods do you want to insure against?" said Eisenbeis.