Banks’ Progress in Capital Stress Testing Confronts Uncertain Future

Courtesy of Holly Presley

Donald Trump’s surprise election has led many to believe that regulatory relief is on the way for large banks. This optimism has led to a surge in bank stock prices, with the KBW bank index up over 20% since the election. While Treasury Secretary designee Steve Mnuchin has said the new administration wants to “strip back parts of Dodd-Frank,” the exact shape of any reforms to post-crisis financial regulation remains uncertain. Furthermore, it’s not entirely clear what specific changes banks are seeking. Large banks in particular have spent a significant amount of time and resources complying with the multitude of rules and regulations that emerged post-crisis. Surely, not every one of these rules can be bad. Such a sentiment is likely why J.P. Morgan Chase CEO Jamie Dimon recently stated, “We’re not asking for wholesale throwing out Dodd-Frank.”

One requirement that is unlikely to go away entirely is capital stress testing, which has been a key area of focus ever since Treasury Secretary Geithner launched the Supervisory Capital Assessment Program (SCAP) in 2009. Since then, large to midsize banks have invested heavily in improving their stress testing capabilities, largely to satisfy Dodd-Frank and supervisory requirements. As the incoming administration, along with Congress, begin to piece their financial reform agenda together, it is worth taking a step back to assess the progress banks have made in conducting capital stress tests, and the benefits of stress testing to banks and society at large. The scale of progress, and the benefits materialized to-date from stress testing, depend to a large degree on the size of the bank.

Initial Investments in Capital Stress Testing Post-Crisis

Following the 2010 Dodd-Frank Act, regulatory expectations regarding capital adequacy, assessment and planning were significantly heightened for institutions in the $50 billion and above asset size group. As a result of Dodd-Frank, banking institutions above $10 billion in consolidated assets, but particularly those banks with $50 billion or more in assets, have made long-term, multi-year investments in both modeling and governance capabilities to meet the post-crisis regulatory requirements of CCAR and DFAST.   [1]

First, banks devoted sizeable amounts of time and energy into developing the modeling and estimation methodologies needed to generate nine quarters of forward looking projections of credit losses, pre-provision net revenue, and balance sheets under the supervisory hypothetical stress scenarios. Even for a relatively small institution (let’s say $10 billion in assets), the primary endeavor to first review, catalog, and determine the materiality of all assets, liabilities and lending portfolios for the purposes of creating a comprehensive stress-testing framework is a significant exercise requiring senior executive leadership and bank-wide collaboration.

Next, the institution-wide planning process to designate individual balance sheet items for stress projections and develop appropriate methodologies for line items can easily be a multi-year process.   The timeline to establish stress-testing methodologies for various line items is determined by many facets (materiality, financial impact to the institution, available internal resources, and data availability) and the complete cycle can easily require a year’s lead time for a single line item. Thus, the investment and development time in the theoretical framework alone has been a sizeable undertaking.

As the regulatory regime matured following the financial crisis and banks successfully completed the development of an institution-wide stress testing framework, banks shifted their attention to building modeling and quantitative capabilities to generate the stress test results. These investments encompassed both the human capital resources necessary to setup and execute model-based projections, as well as the data and technology-rich investments that could supply granular lending information to the newly constructed internal modeling systems. Notably, these infrastructure investments and the design of future bank data platforms are long-term investments, requiring multiple years to design, test, and implement in test environments before going live. The build-out of stress modeling capabilities and data requirements were integrated into future operating and technology plans wherever possible.

Improvements in Governance

While DFAST and CCAR pushed banks to identify and quantify risks impacting capital adequacy, not all risks to banking institutions can be fully quantified with model methodologies, nor are all risks measured with reliable precision. The development of a robust governance framework overseeing models and enterprise risk management added value to institutions’ stress testing frameworks that could not be captured in mere quantitative models or figures.

As the expectations of Dodd-Frank became clear to banking institutions, executive management established parallel governance capabilities to oversee and bring coherence to the quantitative frameworks that were being set up. Board oversight and involvement became a new norm.  Governance frameworks have been, and continue to be, essential in helping banks meet the enhanced regulatory regime. For example, governance systems lend coherence to the overall stress testing exercise through systematic review of the assumptions, methodologies and approaches used throughout the bank. The heightened regulatory agenda prompted executives to begin embedding a new ethos across their institutions, both horizontally with senior management, and vertically through integration of modeling tools with firms’ risk appetite frameworks. The development of a robust governance framework also enhanced banks’ abilities to identify, plan for, and manage the various risk assumed in ordinary business operations.

The Uncertain Future of Capital Stress Testing

Now that Republicans will be calling the shots, there are questions about what will become of capital stress testing. Will the regulatory requirements be changed depending on the size of the bank? Do the largest banks want or expect CCAR and/or DFAST to disappear or be rolled back?  These questions are currently being debated by policymakers, regulators, and industry professionals. Although answers are hard to come by, there are currently some proposals on the table, that if enacted, would significantly alter how stress testing is conducted in the future.

On December 1st, the House passed a bill that would eliminate the current systemically important (SIFI) threshold of $50 billion in assets and require regulators to assess several factors before classifying a financial institution as systemically important. If the bill were to become law, it would likely reduce the number of SIFI’s, and for those firms who are de-designated, it would free them from Dodd-Frank mandated “enhanced supervision and prudential standards.”[2] Although these banks would still be subject to DFAST and CCAR, the requirements would be less stringent. On the other hand, elimination of the $50 billion asset threshold would undoubtedly introduce an element of uncertainty for some institutions about whether they would be designated as a SIFI or not (clear alternative criteria for the SIFI designation have not been proposed). This uncertainty and guesswork will detract from prioritization and focus of the bank’s resources, priorities, and human resource allocation.

The annual CCAR exercise could change significantly if Congress passes The Financial Choice Act, which has already been passed out of the House Financial Services Committee. The Committee’s Chairman, Rep. Jeb Hensarling (R., Texas) is the bill’s main sponsor and was reportedly under consideration to be President-elect Trump’s Treasury Secretary. The bill would effectively free banks from most Dodd-Frank requirements provided they are willing to meet a 10% simple leverage ratio. For banks that do meet the 10% threshold, banking agencies would still be permitted to conduct stress tests but would be forbidden from limiting capital distributions in response to the stress test results. Such a provision would defang CCAR, which currently allows regulators to limit, or reject, share buybacks or dividends if a bank performs poorly on the exercise. The bill would also subject the assumptions underlying the stress test scenarios to public notice and comment, which banks have previously requested.

Even the Federal Reserve is considering changes to regulatory stress testing, albeit less drastic changes than those envisioned by House Republicans. In September, Fed Governor Daniel Tarullo announced a series of reforms to the annual CCAR exercise. Banks will now be required to hold a “stress capital buffer,” which will likely increase the capital requirements for the eight largest banks while lowering the requirement for all other banks. Other changes are designed to provide more transparency into the exercise. In addition, banks with less than $250 billion in assets will no longer be subject to CCAR’s qualitative review – only the quantitative review will apply.

The Value of Stress Testing

The above proposals seek to alter how regulatory stress testing is conducted and the size of firms subjected to the most stringent stress testing requirements. What is notable about the proposals however, is that they don’t eliminate regulatory stress testing entirely. There appears to be broad recognition, amongst policymakers, regulators and even banks, that stress testing provides value – to firms and to society – beyond just checking a required box.

The amount of value may depend on where a bank finds itself along the total asset size spectrum (since asset size also proxies for financial system impact and risk), the efficiency with which it executes the stress testing exercise, the efficacy of the bank’s risk management framework before and after the DFAST/CCAR requirements, and the institution’s interconnectedness in the financial system. There is no doubt, that for some banks, regulatory stress testing can be a significant burden. As PNC’s CEO, Bill Demchak, said of the stress testing exercises, “you bring the place [bank] to a grinding halt once a year.”  In the wake of the financial crisis however, this slowdown in business operations is the price for mitigating broad financial system risks. Under the current political changes set to take place in 2017, questions about the future regulatory agenda should be evaluated separately by asset size thresholds. One size is unlikely to fit all, and the Fed has already been moving towards differentiated expectations for CCAR filers, based on size and complexity through supervisory letters SR 15-18 and 15-19 and the elimination of qualitative reviews for banks under $250 billion in assets.

Impact on Large Banks

As the 7th annual CCAR submission deadline nears, most of the largest banks have made substantial progress in setting up a reliable framework and have taken long-term steps to integrate stress testing into business-as-usual (BAU) approaches of business management. Taking stock of the largest institutions shows that these banks have come far past the initial hurdle of setting up a sound model based, quantitative stress testing program (the recent CCAR objections against Santander and Deutsche Bank were for qualitative reasons). In fact, the largest banks have had several cycles to refine their quantitative and qualitative processes, thus making them not only more robust but also more efficient (and therefore less likely to bring the institution to a complete and grinding halt). Statistical models are reaching a maturing phase with additional sensitivity analyses; model overlays and model adjustments being the frequent points of focus rather than new model development.

Large banks are also making execution of the annual exercise more efficient while better linking the risk modeling framework with BAU processes. As such, the long-term modeling and quantitative investments are also likely evaluated for their potential to optimize future business operations and return focus on profitability measures.

Impact on Medium-Sized Banks

For banks with $10-50 billion in assets, a wide set of challenges and questions persist around meeting the post-crisis regulatory regime. These institutions are also subject to the DFAST annual company-run stress test exercise and thus require the fundamental model and governance infrastructure to execute stress test results. Yet, fewer submission cycles have passed for this filing group and banks have less clarity around the scope and extent of regulatory (FDIC and OCC) expectations, such as line items for quantitative approaches versus qualitative approaches, governance frameworks, and how to strike the appropriate balance for line items with regard to sophistication of quantitative methodologies, data limitations, costs/resources/timeline constraints. Smaller banks in particular can find the efforts to develop and maintain new quantitative and governance frameworks to be challenging in regard to both financial costs and the identification and retention of human resources (which must balance regulatory related duties with BAU duties). Moreover, externally-based quantitative options for small banks such as vendor support models and systems, are often disproportionately costly and come with an opaque black-box nature from their proprietary restrictions. Overall, the stress testing regime remains less established for smaller banks, and many banks remain deep within the process of standing up robust and repeatable systems that can be executed annually.

Getting the Balance Right

Post-crisis regulatory standards have heightened capital risk management processes across the banking sector, especially for the largest banks and bank holding companies above $50 billion in assets. More important to the layperson, regulatory stress testing has forced banks to hold more capital, making them more resilient and better able to withstand future crises. Like any regulation, there is a cost to comply, and these costs vary depending on the size of the institution. Large banks have had the time and resources to incorporate elements of regulatory stress testing into their day-to-day risk management practices. But for banks with $10-50 billion in assets, significant challenges, and uncertainty, remain regarding the establishment and execution of sound stress testing frameworks, methodologies, and governance systems that are appropriately tailored for their size, risk, and complexity. How should the current calls and proposals for some measure of regulatory relief for this group be evaluated, and using what criteria? The challenge is to tailor the regulatory expectations and standards in such a way that banks gain a better understanding of their own risks, through quantitative and qualitative processes, while still yielding net value to society, and arguably each individual institution itself.





Holly Presley is a Senior Manager in the financial services sector and has assisted various financial sector firms on topics related to model risk governance, model risk management frameworks, quantitative methodology support, and stress testing practices for regulatory requirements.  The opinions expressed herein are solely those of the author and are for the purpose of furthering discussion and debate about current policies in financial services regulation.

[1] The Comprehensive Capital Analysis and Review (CCAR) is an annual exercise for bank holding companies (BHCs) with $50 billion or more in consolidated assets and is conducted by the Federal Reserve to assess whether the largest BHCs operating in the United States have sufficient capital to continue operations in times of economic and financial stress. CCAR, through its qualitative component, also examines whether BHCs have a robust, forward-looking and comprehensive capital planning process to capture specific risks and vulnerabilities of each BHC.  The CCAR quantitative exercise uses each BHC’s planned capital actions, the BHC internal stress tests and supervisory scenarios and the BHC’s own internal scenarios. The CCAR Supervisory post-stress capital analysis however uses the Federal Reserve’s quantitative models, supervisory scenarios, and each BHC’s planned capital actions under the hypothetical scenarios.

Dodd-Frank Act stress testing (DFAST) is a complementary and related exercise to CCAR, with important differences. Under DFAST, there are supervisory-run stress tests and company-run stress tests depending on the bank’s total consolidated asset size. BHCs must publish results of the company-run stress tests each year.  For BHCs with $50 billion or more in consolidated assets, the supervisory-run DFAST stress tests conducted by the Federal Reserve use a standardized set of assumptions about capital action distributions as set forth in the Dodd-Frank Act rules (different than those used under CCAR). As set out in SR 14-3, BHCs above $10 billion but less than $50 billion are subject to different stress test requirements. Notably for the smaller institutions, banks are not subject to supervisory-run stress tests, CCAR, or the specific data collection provisions. For this group only company-run stress tests are required and are submitted to the depository institution’s applicable regulator (FDIC or OCC).

[2] Section 165 (a) of the Dodd -Frank Act requires the Fed to establish “enhanced supervision and prudential standards “for bank holding companies with more than $50 billion assets that are both (1) stronger than the standards applicable to smaller institutions and (2) increase in strength based on an evaluation of each bank holding company’s unique riskiness.


Leave a Reply

Leave a Reply

Your email address will not be published. Required fields are marked *