The regulator for the largest U.S. banks warned on Thursday against using industry-wide stress tests as a cornerstone of bank supervision going forward -- a view that diverges from a top Federal Reserve official.

Comptroller of the Currency John Dugan said the stress tests conducted last year were successful in shoring up confidence in the banking industry at a critical time.

But he cautioned that in the future, supervisors should not publicly disclose the results of stress tests that may not pay enough attention to the individual complexities of big banks.

Routine disclosure of negative supervisory information -- especially when that information is based on imperfect tests related to uncertain future events -- could generate the type of confidence problems that it is our job as supervisors to prevent, Dugan said in remarks prepared for a Federal Reserve Bank of Richmond symposium.

Dugan's comments come just two days after Fed Governor Daniel Tarullo said bank regulators should consider conducting routine, publicly disclosed stress tests.

The Fed and other bank regulators early last year conducted stress tests on the 19 largest U.S. financial firms and disclosed the results, a controversial decision even inside the central bank. The concern was that weaker banks might be harmed by the public disclosure.

It was a unique exercise in that it took a horizontal approach and tested how the banks would fare under uniform assumptions of economic conditions and loss rates.

The stress test exercise was largely seen as a success because it removed some of the uncertainty about the true condition of banks' balance sheets, encouraging private capital back into the banks.

Dugan acknowledged the exercise -- known as the Supervisory Capital Assessment Program, or SCAP -- was a success but said regulators should not go too far in relying on such tests going forward.

Dugan said using uniform assumptions across banks limits the ability to tailor stress-test details to firms' unique traits.

Modern financial institutions are complicated firms, and the finer details of their portfolios and business lines strongly influence the impact of a given set of stresses, he said.

He also said the stress test used a number of ad hoc and imprecise elements such as provisions for loan loss reserves, the payment of dividends, and the impact on other non-credit-related parts of the business that may not have been particularly telling when assessing the overall health of the bank.

Further, Dugan said the stress tests were conducted in an extreme environment when public disclosure of the bank weaknesses was necessary and revealed at a time when there was a backstop of public capital through bailout funds.

Since we will not in general have this kind of public backstop, I think we should not in general pursue this kind of transparency, he said.

Tarullo, speaking to the Council of Institutional Investors earlier this week, came to a different conclusion. He said releasing the information would help investors make informed decisions, and encourage public scrutiny of the regulators' methods.

Tarullo, however, acknowledged that there may be times when releasing the information could be unnecessarily destabilizing, particularly if the results showed banks needing capital in more normal times when the government is not providing emergency capital injections as it was last year. Major unpleasant surprises would be less likely with frequent, detailed disclosures, he said.

(Reporting by Karey Wutkowski, editing by Matthew Lewis)

Federal reserve