The (Un)intended Consequences of Stress Tests

Correction: An earlier version of this post incorrectly suggested that banks with $10 billion to $50 billion in assets (it used the word "capitalization") will be subject to the Fed's CCAR stress tests. They are instead required to take a separate set of stress tests under the Dodd-Frank Act, and the Fed does not publish the results for these banks as it does for those with $50 billion in assets and up. (The midsize banks will self-report instead.) The headline and certain passages have been updated to more clearly express the author's main point.

Now that the 2014 stress tests have been completed, we can consider the impact their results will have on the financial firms subjected to the testing. We can also leverage insights as to how midsize banks ($10 billion to $50 billion in assets, subject only to the stress tests under the Dodd-Frank Act) may well have a very different opinion of a passing or failing grade than the giants ($50 billion and up, subject also to the separate, better-known Comprehensive Capital Analysis and Review exercise).   

The 2013 list of 30 bank holding companies that participated in the CCAR resulted in 25 passes, and five capital plan rejections. Four of these were detailed as qualitative, one as quantitative. The CCAR report makes explicit that a qualitative rejection should be viewed as a request for a revised submission rather than an outright failure, and notes that banks taking the test for the first time are more liable to be rejected on qualitative grounds.

At this juncture, it is worth looking in some detail at the definition of "qualitative," as it is less clear cut than quantitative (where, simply put, the BHC's capital level is too low under the stresses). Within the report, a qualitative rating is explained as:

• The BHC's capital-adequacy assessment process is not sufficiently robust. This includes the corporate governance and controls around the process, as well as risk-identification, risk-measurement and risk-management practices supporting the process.
• The assumptions and analysis underlying the BHC's capital plan are inadequate.
• A BHC's capital adequacy process or proposed capital distributions would constitute an unsafe or unsound practice, or would violate any law, regulation, board order, directive or any condition imposed by, or written agreement with, the board.
• There are outstanding material, unresolved supervisory issues.

The midtier banks that take stress tests only under Dodd-Frank don't get graded on these, but they are required to disclose summaries of their results beginning in June of next year. This should be concerning, for reasons tied up with some other new rules that are rumbling through the regulatory pipeline, specifically but not only: single counterparty credit limits, non-reliance on third-party credit rating agencies and credit valuation adjustment.

These regulations require each bank to perform assessments of their counterparty's creditworthiness. Stress test results provide a clear starting point for such assessment, to the extent that any other model would need to at least take account of the findings, if not directly feed back into them. A poor showing for a bank in the DFA stress test would need to be reflected in a counterparty bank's internal rating system.

The knock-on effects of receiving a low rating in their peers' internal rating systems could include less access to market liquidity as a result of lower limits or higher CVA charges. These are outside of the mandated consequences of the stress test initiatives, which are based on non-approval of capital plans such as expenditure or dividend payment. Instead, these possible outcomes should be considered potentially unintended consequences.

Larger banks, who act as super-regionals or international trading counterparties, have the size and base to be somewhat unconcerned – but the same is not true for smaller BHCs.

The nature of the CCAR and Dodd-Frank stress test reports is such that the banking and trading books are stressed, and the data controls and calibrations are looked at, as well as the overall governance and risk control processes within the bank. It is, of course, possible to reduce or even eliminate derivative activity, which removes the issues around modelling and control of complex instruments. This still leaves the banking book, and the issues around data quality and calibration.

The bulk of the stress tests themselves are expressed in macro terms that require significant regression testing. This means they include GDP, inflation, house prices, etc., as well as the micro-factors that are typically used to price the banking and trading book, such as yield curve points, equity prices and other factors that are directly used by the valuation models. Of course, the lower the number of factors used by the bank, the simpler the data calibration process is.

Taken together, the stress tests could easily be seen as one more weapon to protect the system itself, at the expense of any individual BHC business line profitability. This makes sense in the context of regulation created as a direct result of a liquidity/credit crisis based, at least partially, on unreliable ratings. That being the case, smaller BHCs are faced with a potentially significant problem if their capital is judged insufficient for that regulatory environment, and that judgment becomes public in the middle of next year.

Whether or not these consequences are unintended, they are real, and financial firms need to take them into account when planning the risk governance committee, data control and internal risk management, as well as the stress tests themselves. Ultimately, the reporting process that is integral to the stress tests could prove to be one of the best-tasting carrots as well as one of the hardest sticks the Fed has to enforce in its raft of new rules.

Marcus Cree is vice president, risk solutions, capital markets at SunGard, a global software and technology services company.