Bank Stress Tests: Overbuilding the Bridge
Yong Ye
May 20th, 2021
The past year has been full of surprises for everyone, including banks, which had to react to severe credit market fluctuations as the world entered and exited pandemic mode. According to the old saying, the only certainties in life are death and taxes. As we enter 2021, however, more regulation for banks of all sizes can safely be added to that list. Specifically, the new regulation will be enforced through review of stress test results. Indeed, in its most recent update to the Commercial Bank Examination Manual, the Fed’s Division of Supervision and Regulation emphasized that going forward, stress testing will play an essential role in the risk examination of all member banks regardless of size.
In general, a stress test is designed to show how a set of adverse, but not impossible, scenarios impacts the balance sheet, portfolio, and capital adequacy of a bank. For example, in its Commercial Bank Examination Manual, the Federal Reserve publishes a specific set of guidelines laying out the stress test requirements in the areas of market risk, counterparty credit risk, and liquidity risk. These tests are designed to ensure that banks have adequate capital relative to the prescribed stress events and also to create proper frameworks that ensure a bank remains viable over the long run.
However, it is the banks, not the regulators, who are responsible for the design and construction of stress tests, even for a stress test required by regulators. For example, in the Comprehensive Capital Analysis and Review (CCAR), the most stringent stress test the Fed imposes on the largest and most sophisticated banks, the Fed prescribes moves in factors such as the Dow Jones Industrial Average, VIX index, and Housing Price Index. It is up to the banks to translate these moves into the changes in the prices of other assets and products that the banks might hold, such as mortgages or loans of various forms. In doing such a translation, the bank’s view on market correlations (such as how VIX, or equity volatility, and interest rate volatility are related) and its own asset pricing models have significant impact on the projected valuation changes. More generally, stress scenarios are stated in terms of significant moves in macroeconomic factors, and the translation to portfolio profit/loss, balance sheet changes, and capital adequacy requirements are again up to the bank.
The requirement of stress testing places a significant burden on banks that can only be fulfilled with the allocation of considerable resources and effort. Larger banks have the capacity to dedicate the time and effort of a team of economists and quantitative researchers working in close collaboration. However, the small-to-medium sized banks are placed in something of a bind, because they lack such resources. Instead, the prevailing practice appears to represent something of a suboptimal compromise that can perhaps be described in analogous engineering terms: When there is not the ability or capacity to do the structural engineering required to build a bridge, the builders simply over build it. That is, they pour so much steel and cement into the bridge that there is no possible way it can fail, even without any sophisticated structural calculations. So, the requirement of building a bridge is fulfilled, but, the downside is that perhaps ten times more materials and labor went into its construction than were required.
The translation of this engineering analogy to the current practice of moderately sized banks is that they apply very simple and very conservative models to the stress test framework. This passes the testing regime, but leads to exaggerated capital requirements and unnecessary restrictions on the scope of business, with corresponding effects on revenue and profit. Like the practice of overbuilding the bridge, this represents a considerable drag on bank operations.
However, an alternative solution that addresses these inefficiencies has recently emerged, namely the outsourcing of regulatory modeling and analytics. This is yet another instance where outsourcing increases the quality of the product while at the same time reducing its cost. Well known examples include cloud computing, outsourced tax preparation and advice, outsourced IT and operations management, and so on. In the current situation, since the provider is already proficient in stress testing and has ready built systems and analytics to address the required calculations, the marginal cost to provide stress testing services to a new client is not substantial. At the same time, the expertise in regulatory calculation translates into a more expert product. The bank is left to focus on its primary business, which is the prudent generation of revenue and profit, while fulfilling regulatory requirements at modest cost.
