Big banks operating in the U.S. might be able to stress out less over regulatory stress tests
|
The tide had already begun to turn in the past two years. Bank regulators in the U.S. had begun to ease on the toughest aspects of bank regulation.
And in Feb., 2019, the Federal Reserve in the U.S. amended some of the rules, requirements and expectations of stress tests administered to big banks after it had significantly changed requirements in May, 2018, with amendments to U.S. Dodd-Frank regulation.
In just the past two years, there are likely a hundred or so banks around the country quietly applauding such changes in stress-test requirements. In 2019, the Federal Reserve didn't abolish stress tests; it modified requirements by applying less pressure on banks to prove they will have substantial amounts of capital during prolonged periods of stress.
The stress tests for big banks still exist. The 2018 amendments to Dodd-Frank granted relief to medium-size banks: They are no longer subject to regulatory stress testing, although they should conduct them on their own.
The tests for the big banks (based on asset size) are administered and run by regulators using regulators' own models and scenarios. And big banks must still seek to pass the tests. In the latest episode of changing bank regulations, regulators promised not to apply hard pass-fail rules to the "qualitative" phases of the stress tests, the rules that make judgment calls on how banks assess whether future capital levels will be sufficient as they growth their businesses.
Stress tests of all kinds and conducted in any manner, based on assumptions and models to show how a bank would endure a period of systemic and economic stress, are still important. Dodd-Frank II permitted fewer banks to be subject to the Federal Reserve-administered tests. But bank regulators still expect banks to run their own tests and submit results to supervisors. Banks don't mind it when they manage their own tests, using their data and devising their own models. They become concerned if they are subject to tests based on others' models, especially when the details of such models are not widely available.
The Federal Reserve version of stress testing assumes big banks will be subject to a nine-quarter period of an macro-economic downturn (adverse case, severely adverse case, etc.). The regulator performs calculations of how each big bank will perform under the scenario. In most cases, banks will amass big losses. The regulator doesn't guess at the losses; it calculates them as precisely as possible. It determines whether the bank under review has sufficient capital to withstand the calculated losses and meet conventional capital requirements under Basel III and U.S. Dodd-Frank rules.
In banking (whether for regulators or for ongoing internal risk management), there are numerous kinds of stress tests: Stress tests for loan portfolios; stress tests for trading positions and investment portfolios; stress tests focusing on liquidity, funding and deposits; stress tests for operational risks (especially related to business continuity and cybersecurity), and stress tests related to country and political risks.
There is scenario-testing, as well: What happens to the loan book and deposits when interest rates rise or fall by 100 basis points? 200 basis points? What happens to trading positions when interest rates rise or equity markets fall by a certain percentage? What happens in trading activity when trading counterparties default or can't meet trading terms? What happens when the value of collateral in loan exposures or trading positions declines suddenly? The scenarios are countless in number. The bigger banks run models that look at thousands of variables and scenarios.
Most banks are well-equipped to develop the required sophisticated models to quantify such risks--whether the risks are overnight or are prolonged over a two-year period. They also develop models with the understanding that bank supervisors will have chances to review model integrity and second-guess the results. (Models are "back-tested" to evaluate whether they successfully predicted the amount of worst-case losses.)
Stress tests provide comfort for bank risk managers, bank investors and even regulators. They permit the bank to identify shortfalls in risk management, over-exposed business activities (in the loan book or in trading positions), and single out business units that aren't bringing in the rewards for the risks they absorb. They also help banks determine how much capital is necessary for today's balance sheet and for balance sheets in the periods to come. Stress tests contribute to bank decisions about whether they can increase dividend payouts or buy back shares.
Risk management in banking ultimately leads to the question of whether there are sufficient amounts of capital to absorb identifiable and unforeseen, hidden risks. Capital absorbs losses and prevents losses for creditors, depositors, trading counterparties, and ultimately central governments and deposit-insurance schemes that may feel the need to bail out failing banks.
Regulators obsess over capital adequacy. Banks obsess over calculating optimal capital levels. They want to make sure they have capital for today's operating environment and capital to pass tests. They embrace having excess amounts of capital, but they also don't want too much capital. Too much capital, they reason, implies wasted amounts of capital. They rationalize increased dividends or buy-back programs if there is excess beyond excess.
Stress tests, therefore, help regulators and banks in these calculations of what is the right amount of capital on an ongoing basis.
In the past year, the Federal Reserve reaffirmed the value of stress testing. But they provided relief for small banks (by not subjecting them to regulators' versions of the tests) and eased up on the pass-fail requirements for big banks.
Ease up or not, the last thing bank supervisors want is for big banks to learn how to mastermind or "game" the regulatory stress-test model to permit them to pass stress tests without difficulty.
Tracy Williams
See also:
Posts
No comments:
Post a Comment