By David Henry
NEW YORK (Reuters) - The newest stress tests for U.S. banks produced scores that are at odds with other measures of lenders' safety, in another sign that some institutions may be too big for regulators to understand and executives to manage.
For example, Citigroup Inc, which has been bailed out multiple times by the U.S. government, showed up on the score sheets posted by the Federal Reserve on Thursday as being clearly safer than JPMorgan Chase & Co.
That conclusion is at odds with the views of investors, bond analysts and credit-rating agencies, as well as when measured by a yardstick regulators themselves want to use in the future.
"At the end of the day, there is a legitimate question about the ability of regulators to fully evaluate $2 trillion institutions because of the complexity and exposures they have," said Fred Cannon, director of U.S. research at Keefe, Bruyette & Woods.
On Thursday, the Federal Reserve reported the latest results of the tests that began after the 2007-2009 financial crisis to determine if banks have enough capital to withstand a severe economic crisis. The Fed concluded that the banks are in "a much stronger position" than before the financial crisis in 2008.
While experts are not arguing with the fact that the banks are better capitalized now and that the system is safer than it was in the run-up to the financial crisis, some of the numbers the regulators published left analysts and bank executives groping for explanations. The test raises questions about the ability of regulators to head off the next big threat to the financial system because of the complexity of the institutions.
The results are also important as they will help the Fed decide how much capital banks can return to investors.
The report showed that Citigroup's capital, as tracked by the Tier 1 common capital ratio, would dip to 8.3 percent during two years of hypothetical stress. JPMorgan's would fall to 6.3 percent. Both numbers are better than the 5 percent minimum under current regulations, but they show Citigroup having a bigger cushion to weather losses.
That does not make a lot of sense to Kathleen Shanley, a bond analyst at GimmeCredit, a research service for institutional investors.
"I wouldn't say that Citi is safer than JPMorgan, for a variety of reasons, including its track record," Shanley said.
Citigroup has lower credit ratings than JPMorgan, and prices for credit default swaps show the market views JPMorgan as safer. Citigroup is the third-biggest U.S. bank by assets and JPMorgan is the biggest.
A Federal Reserve spokeswoman declined to comment, as did representatives for Citigroup and JPMorgan.
WHY CITI LOOKS SAFE
Citigroup's score came out better partly because it started the test with a better Tier 1 common ratio, 12.7 percent compared with JPMorgan's 10.4 percent.
The starting ratios were based on the banks' financial statements at the end of September. They were calculated based on a set of international regulations known as Basel 1, which the Federal Reserve intends to replace as inadequate with a pending new set known as Basel 3.
Under the expected Basel 3 rules, Citigroup has estimated its ratio was 8.6 percent at the end of the third quarter, about the same as the 8.4 percent JPMorgan estimated.
Among the reasons that Citigroup's ratio will fall so much under Basel 3 from the Basel 1 level is that the new rules will not treat as favorably Citigroup's deferred tax assets.
Citigroup expects those assets to allow it to pay lower taxes on future profits because it lost so much money when the financial crisis and recession hit. Also, Basel 3 will reduce the benefits of stakes Citigroup has in joint ventures, such as its brokerage with Morgan Stanley.
The Federal Reserve did not publish stress scores for the banks under Basel 3 because the regulators have not finalized those rules yet.
Analyst Cannon said there was one reason to think of Citigroup as being safer: its capital markets business is smaller than JPMorgan's. Regulators regard capital markets operations as riskier than consumer banking businesses.
ROOM FOR SUBJECTIVITY
The Fed's scoring is also at odds with results some of the banks calculated for themselves under the same scenarios, which shows there is room for subjectivity in the testing.
JPMorgan, for example, found that its ratio would fall to 7.6 percent, significantly better than the 6.3 percent reported by the Fed. Goldman Sachs Group Inc determined its low during the hypothetical stress period would be 8.6 percent, compared with the 5.8 percent reported by the Fed, with some of the difference related to its extensive capital markets activities.
Goldman declined to comment.
Wells Fargo & Co pegged its low at 8.3 percent compared with the Fed's 7 percent.
Wells Fargo said in a statement that it could not fully explain the difference because the Fed does not disclose all of the models it uses to score the banks. The bank said that for some securities, it takes into account more risk factors than the regulators do.
"It is primarily model-driven assumptions that will drive the differences," said Fernando De La Mora, who leads PricewaterhouseCoopers' banking and capital markets risk.
Last year, differences between scores by the banks and by the regulator were not disclosed, but people in the industry knew of significant disagreements over expected losses in some portfolios, said De La Mora.
This year, the Fed told the banks that it "will focus on the robustness" of each bank's testing.
For Citigroup, the Fed's ratio this year of 8.3 percent was nearly as high as the 8.4 percent the bank tallied for itself.
(Additional reporting by Rick Rothacker in Charlotte, North Carolina, and Lauren Tara LaCapra in New York; Editing by Paritosh Bansal, Martin Howell and Peter Cooney)