We are at an interesting time in bank capital regulation. U.S. banks have performed very well thus far, but the current economic outlook is worse than the severely adverse scenario assumed in the ongoing stress test, and an even deeper and longer recession than currently projected is possible. Thus, the question has arisen how regulators should account for these developments.
In its supervision report, issued May 8, the Federal Reserve stated that “the current plan is to conduct the 2020 supervisory stress test as originally announced — to maintain the established process under the Federal Reserve’s stress test and capital rules — and also conduct a series of sensitivity analyses using alternative scenarios and certain adjustments to portfolios to credibly reflect current economic and banking conditions.”[1] In a parallel statement made a few weeks earlier, Vice Chair Quarles said on April 10, “I think the right thing to do is for us to continue our stress tests, but as part of them to analyze how banks’ portfolios are responding to real, current events, not just to the hypothetical event that we announced earlier this year.”[2]
This announcement came as little surprise; for months, many observers had opined that the Comprehensive Capital Analysis and Review (CCAR) stress scenario announced in March was out of date, and that the Fed should update it to reflect the current environment. Still, from both a substantive and procedural perspective, it is unclear exactly what the Federal Reserve intends to do, and what goals will inform its approach. This note describes the regulatory context in which the Fed will be performing its sensitivity analysis, explores possible routes that the Fed could take, and offers some cautions – against procyclicality and opacity and effectively jettisoning a regulatory regime that appears to be succeeding in its first wartime test.
THE REGULATORY CONTEXT
Before turning to the question of how the Fed might conduct an ad hoc sensitivity analysis, it is important to note how the existing regulatory regime is already stress-based – that is, how it already requires banks to capitalize rising risks.
The current legislative, regulatory and accounting regime is expressly designed to be countercyclical – to allow a bank to deploy in an economic downturn or time of financial instability capital that it built up in normal times.[3] To ensure bank safety and soundness, the current regulatory regime requires a bank to restrict or halt capital distributions as it draws down its capital buffers, and to face growth restrictions and other more severe sanctions to the extent that its capital levels fall below minimum capital requirements.
Furthermore, large banks are already recognizing potential future losses under both accounting and implementing capital rules. First, under the Current Expected Credit Losses (CECL) accounting standard, which took effect in the first quarter of this year, banks are required to establish a reserve for every loan they hold or make to reflect estimated losses over the life of that loan, under the current economic scenario. Large banks took $60 billion in charges in the first quarter, largely reflecting a downgrade in the economic forecast. At the end of June, they will be required to update their economic forecasts and take further charges if the forecast has worsened. The same will be true in the third quarter – meaning that they will have updated their economic forecasts three times prior to onset of the stress capital buffer (SCB) in October. (Lest one think that the forecasts are overly optimistic, note that these are securities filings subject to sanction; furthermore, most banks default to an average of scenarios published by Moody’s Analytics; for the rest, analysts are quite attentive to the assumptions.)
That said, on March 31, the federal banking agencies decided to delay the capital impacts of CECL for two years;[4] thus, accounting charges will not affect bank capital requirements for that period. Nonetheless, the allowance exercise will be transparent and thereby inform both investors and regulators of how bank balance sheets are performing.
It is worth noting that this relief would appear to conflict with any attempt to increase the SCB in light of a worsening economic outlook. Given the same worsened economic forecast, the agencies would be lowering the capital impact of that forecast through their CECL relief – a countercyclical action — and then raising it through their SCB increase – a pro-cyclical action. It would be a poor trade, given that even with the CECL relief, banks must continue to calculate and report their CECL allowance each quarter, so analysts and investors can determine exactly what the impact of the current relief is. In contrast, the sensitivity analysis apparently being considered by the Fed would presumably follow the CCAR tradition of secret models and methodology.
Second, under risk-based capital standards, banks subject to the advanced approaches are required to model the risk of their assets, updating them as the risks rise. In the first quarter, banks saw an increase in risk-weighted assets. A significant portion of the increase in risk-weighted assets was attributable to the simple fact that there was a larger volume of loans, increasing the denominator of capital ratios and thereby lowering them.[5] At the same time, though, the nine banks subject to the advanced approaches capital measures also became more bound by risk-weighted assets calculated using their own models. For those banks, part of the rise in risk-weighted assets was attributable to downgrades of borrowers, higher mark-to-market losses and increased requirements for counterparty credit risk. Thus, under risk-based capital requirements, there was an increase in the denominator because of higher projected future losses.
In addition to these risk-based capital requirements pursuant to the Basel mandate, the Federal Reserve imposes its annual stress test. The stress test is designed to see how a bank performs under a financial crisis and severe recession as in 2007-09. The theory is that while each stress will be different, a bank with sufficient resilience to absorb one will likely be able to absorb others. (We have criticized this approach in the past and argued that the Fed should use multiple scenarios.[6])
The Federal Reserve and other federal banking agencies have provided only limited relief to these risk-based requirements.[7] And pursuant to CECL and risk-based requirements under the advanced approaches, capital requirements will continue to increase if economic conditions continue to deteriorate. Moreover, the largest banks face upward pressure on their GSIB surcharges given that banks have expanded their balance sheets to accommodate a $720 billion increase in lending to nonfinancial businesses and households[8] and absorption of a $1.5 trillion increase (and still rising) in reserve balances created by the Fed’s purchases of Treasury securities and Agency MBS.
It is worth nothing that unlike the U.S. banking agencies, other countries have been lowering risk-based capital requirements. Many other countries utilize “Pillar 2” requirements, which apply in addition to the minimum capital requirements and are intended to cover risks that are underestimated or not covered by such requirements.[9] While such requirements are normally calculated as a percentage of bank risk-weighted assets, in response to the current pandemic, the U.K. Prudential Regulation Authority (PRA) has lowered its Pillar 2 requirements to a “nominal amount” in order to alleviate “unwarranted pressure.”[10] In addition, as we will discuss further below, ten countries around the world have lowered their countercyclical capital buffers (CCyBs). The CCyB is not reduced to provide forbearance for bank losses, but rather to encourage banks to provide more credit. We know of no objections to these reductions.
THE FEDERAL RESERVE’S REGULATORY AND EXAMINATION CHALLENGE
Using the examination process to determine how banks are currently faring is a proper and not entirely unexpected step. While banks remain in compliance with regulatory capital requirements and appear likely to pass the 2020 CCAR stress test, the Federal banking agencies may issue a capital directive to an individual bank pursuant to 12 U.S.C. 3907 to the extent that it determines its capital condition to constitute an unsafe or unsound practice. Indeed, the SCB rule indicates that a bank holding company (BHC) must submit a new capital plan when the Board determines that a new stress scenario is needed because there have been changes in economic and financial conditions “…that could have a material impact on a bank holding company’s risk profile and financial condition.”[11] The question is how to make that determination, and whether to take a pro-cyclical capital action that might have adverse consequences for the economic recovery.
It does seem surprising that the sole capital focus of the Federal Reserve’s sensitivity analysis is the 34 banks subject to CCAR and really the approximately 20 banks whose SCB would likely land above the 250 basis point floor proposed for the SCB. Whether the goal is financial stability, lending capacity or bank solvency, one would have expected greater concerns about small banks. Small banks are not required to maintain capital buffers, in most cases are not required to establish a life-of-loan allowance under CECL, and in many cases are not subject to any risk-based capital requirements. Because small banks mark fewer of their assets to market than large banks, they have almost certainly seen a greater, artificial lift in their capital ratios as a result of the agencies’ decision to forebear on troubled debt restructuring. As noted in the Fed’s recent Financial Stability Report, small banks are also more exposed to the types of loans most likely to experience loss: “Smaller banks with high concentrations of lower-rated consumers, small and medium-sized businesses, and CRE were viewed as especially vulnerable.”[12] Unfortunately, a special stress test will provide no additional information about the resilience of small banks to current circumstances because they are exempt from the Fed’s stress testing regime.
THE DANGERS OF PRO-CYCLICALITY
A fundamental challenge to the design of bank regulation is that the requirements tend inevitably to be pro-cyclical. That is, the requirements are more likely to constrain banks when the economic situation deteriorates, which in turn reduces bank credit supply, which in turn worsens the downturn. In the stress tests, for example, when the economic situation worsens, banks’ loan portfolios weaken, and the “stressed” economic outlook deteriorates further; both changes contribute to higher expected losses and therefore require larger amounts of ex ante capital to pass the test.
The bank regulatory regime, both in the United States and internationally, was redesigned after the 2007-09 financial crisis to reduce pro-cyclicality using the Countercyclical Capital Buffer (CCyB). The CCyB is designed to be raised when financial stability risks are substantially above normal and lowered to boost bank credit supply when conditions return to normal. As with other capital buffers, a breach of the CCyB results in limits on banks’ ability to make capital distributions. Ten countries around the world have lowered their CCyBs; in most cases, the countries had raised their CCyBs in response to concerns about building systemic risks in their jurisdictions. The United Kingdom’s CCyB is 2 percent when conditions are normal; it was lowered to 1 percent out of concerns over Brexit, then reduced to 0 percent in light of the current crisis. Importantly, the CCyB is not reduced to provide forbearance for bank losses, but rather to encourage banks to provide more credit.
Unfortunately, as we discussed in a recent blog post (available here), in the United States, the CCyB cannot be reduced because it is currently zero. It is set at zero in normal times when systemic risks are moderate, and the financial stability assessment by the Federal Reserve was one of moderate risk, in large part because banks were so well capitalized and held such huge stockpiles of liquid assets. Moreover, as discussed in a recent speech by Vice Chair Quarles (available here), and quantified in a blog post by our own Francisco Covas (available here), it would be much more difficult for the United States to set its CCyB above zero in normal times than it is for other major jurisdictions because the U.S. regulatory regime’s baseline capital requirements impose significantly higher capital requirements than other jurisdictions. Nevertheless, it seems likely that if the current U.S. framework allowed for the CCyB to be reduced below zero at the present moment, the Fed would reduce it.
Both the CCyB and the SCB are components of the buffer banks are required to maintain above capital minimum requirements and at the end of the day, a buffer is a buffer. Therefore, it would not seem logical for the right policies to be lowering the CCyB and raising the SCB, just as it would not seem logical to defer the capital impact of CECL and raise the SCB. If the Fed is essentially just retrofitting the stress test with a new severely adverse scenario and an updated starting point, it is tightening capital requirements just when sound public policy (and the law)[13] calls for easing them. Such significant policy choices should not be driven by buffer labeling – if a “CCyB” should be cut, then a “CCAR” should not be tightened.
Moreover, the Fed might do well to consider the important distinction between “wartime” and “peacetime” stress testing as described in Schuermann (2016).[14] The objectives of stress testing are quite different between a period of economic stability versus an environment where the outlook is bleakly uncertain. “CCAR” stress tests in recent years have had the peacetime objective of evaluating the robustness of banks’ capital plans in relation to severely adverse scenario. However, the economy has transitioned to a “wartime” context, when the primary objective is to “to provide credible insight into the health of bank balance sheets” in order to ensure stability of banking and financial markets.
AN ALTERNATIVE WITH CONCEPTUAL AND HISTORICAL SUPPORT
Rather than conducting another stress test, the Fed could conduct an exercise akin to the 2009 “Supervisory Capital Assessment Program” (SCAP). The SCAP was the first “stress test” and the precursor to the current CCAR process, though with a different goal and design that seem appropriate now. It is widely considered a turning point in the 2007-09 financial crisis.[15]
SCAP was conducted in the spring of 2009, at the height of the 2007-09 financial crisis, in order to judge whether banks had sufficient capital to continue to lend even if then-current economic conditions worsened; any resulting capital shortfall was filled with a government capital injection from Troubled Asset Relief Program (TARP) funds. The designers of the SCAP recognized that the economic situation was, in fact, already severely adverse; they thus used a scenario that was only moderately worse than the expected path for the economy. Specifically, the “Alternative More Adverse” scenario included an unemployment rate that was only 0.5 percentage point higher than the rate in the baseline outlook. For reference, the severely adverse scenario released in February for this year’s CCAR stress test includes a path for the unemployment rate that was as much as 6.2 percentage points worse than the baseline.
With the economic environment radically different from when the stress test scenarios were designed this year, a sensitivity analysis conceptually similar to the SCAP seems appropriate to inform a meaningful discussion about banks’ capital plans. Would the bank be able to meet its capital requirements even if the economy were somewhat worse than the already grim outlook?
REALITY DOESN’T NECESSARILY BITE
We also note that if the Fed intends to do a sensitivity analysis to identify potentially higher losses as a result of the current economic scenario, it might also consider the potential for significantly lower losses attributable to the Global Market Shock (GMS) and operational risk. For some time, we and others have criticized both as dramatically overstating potential losses, and results to date during the current turmoil appear to have validated those criticisms. A blog post issued today analyzes how operational risk losses are likely, absent a sensitivity analysis, to overstate losses by approximately $100 billion; previous research, now seemingly vindicated, demonstrated in great detail how the GMS contains counterfactual and overly stringent assumptions and as a result overstates trading losses.[16]
CONCLUSION
Faith in the banking system is currently high. Fed Vice Chairman Richard Clarida, for instance, on May 21 stated “… as I sit here right today in the middle of May, the banking system is really a source of strength and a source of credit in the economy. And that’s important.”[17] The banks’ current strength in the face of adverse conditions is the result of a decade of effort by banks and regulators that had learned the lessons of the last crisis. Banks’ capital levels are twice as high, and liquidity levels are four times as high than in the past financial crisis.
Pulling these strings together, it is appropriate and prudent that the Fed supplement the 2020 stress test with additional analysis to evaluate banks’ ability to weather current circumstances. That analysis, like the 2009 SCAP that was instrumental in ending the last crisis, should avoid being overly pro-cyclical by considering a scenario that is only modestly worse than the current abysmal economic outlook. Moreover, the results should not automatically be translated into a mandatory tightening of capital requirements just when there is universal agreement that capital requirements should be eased – and in most major countries, are being eased — to ensure the supply of bank credit in the face of an unprecedented economic blow, and when the Fed is preparing to open a $600 billion program to convince banks to lend more and to riskier customers.
Disclaimer: The views expressed do not necessarily reflect those of the Bank Policy Institute’s member banks, and are not intended to be, and should not be construed as, legal advice of any kind.
[1] Board of Governors of the Federal Reserve System, Supervision and Regulation Report, 18 (May 2020) (hereinafter Supervision Report), https://www.federalreserve.gov/publications/files/202005-supervision-and-regulation-report.pdf
[2] See Andrew Ackerman, “This Year’s Stress Test For Big Banks Is Real: the Coronavirus Pandemic,” Wall Street Journal (April 10, 2020), https://www.wsj.com/articles/this-years-stress-test-for-big-banks-is-real-the-coronavirus-pandemic-11586553555
[3] See 12 U.S.C. 3907(a)(1) (“Each appropriate federal banking agency shall seek to make [] capital standards … countercyclical so that the amount of capital required to be maintained by an insured depository institution increases in times of economic expansion and decreases in times of economic contraction, consistent with the safety and soundness of the insured depository institution.”); Press Release, Federal Reserve Board approves final rule to help ensure banks maintain strong capital positions (Jul. 2, 2013), https://www.federalreserve.gov/newsevents/pressreleases/bcreg20130702a.htm (“The Federal Reserve Board … approved a final rule to help ensure banks maintain strong capital positions that will enable them to continue lending to creditworthy households and businesses even after unforeseen losses and during severe economic downturns.”); Board of Governors of the Federal Reserve System, Comprehensive Capital Analysis and Review 2020 Summary Instructions at 1 (March 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/bcreg20200304a3.pdf (describing with respect to the stress tests that “[t]he Federal Reserve expects firms to be able to continue operating and lending to households and businesses, even during times of economic and financial market stress.”).
[4] Regulatory Capital Rule: Revised Transition of the Current Expected Credit Losses Methodology for Allowances, 85 Fed. Reg. 17,723 (Mar. 31, 2020).
[5] See Supervision Report at 6.
[6] Robert Lindgren and Bill Nelson, “An Even Better Way To Combine Stress Test Results,” (June 19, 2019), https://bpi.com/an-even-better-way-to-combine-stress-test-results/; Francisco Covas and Bill Nelson, “The Federal Reserve’s stress capital buffer proposal presents an opportunity for improvement,” (May 11, 2018), https://bpi.com/the-federal-reserves-stress-capital-buffer-proposal-presents-an-opportunity-for-improvement/
[7] In March, the banking agencies issued an interim final rule permitting use of a transition to mitigate CECL’s impact on regulatory capital. See 85 Fed. Reg. 17723 (March 31, 2020). In two related interim final rules the agencies modified the definition of eligible retained income for purposes of the capital rules, including the total loss absorbing capacity rule. See 85 Fed. Reg. 15,909 (March 20, 2020; 85 Fed. Reg. 17,003 (March 26, 2020). The Federal banking agencies issued an Interagency Statement on Loan Modifications and Reporting for Financial Institutions Working with Customers Affected by the Coronavirus (Revised) on April 7, 2020 to clarify that financial institutions generally do not need to categorize COVID-19-related modifications as troubled debt restructurings (TDRs).
[8] Federal Reserve Statistical Release H.8, All Banks, Not Seasonally Adjusted, Loans and Leases minus Other Loans, May 13, 2020 minus January 29, 2020.
[9] We strongly support the U.S. regulators for eschewing a wholly subjective and effectively lawless Pillar 2 approach.
[10] Bank of England, Prudential Regulation Authority, Conversion of Pillar 2A Capital Requirements from RWA Percentage to a Nominal Amount (May 7, 2020), https://www.bankofengland.co.uk/-/media/boe/files/prudential-regulation/publication/2020/conversion-of-pillar-2a-capital-requirements.pdf?la=en&hash=2A5E282730858C7A4C4E165E08C6F513F7709D0B.
[11] 12 CFR 225.8(e)(4).
[12] Board of Governors of the Federal Reserve System, Financial Stability Report, 59 (May 2020), https://www.federalreserve.gov/publications/files/financial-stability-report-20200515.pdf
[13] Section 616 of the Dodd-Frank Act directs the Agencies to seek to make capital requirements “countercyclical so that the amount of capital required to be maintained by a company increases in times of economic expansion and decreases in times of economic contraction, consistent with the safety and soundness of the company.’’ Pub. L. No. 111-203, § 616, 124 Stat. 1376 (July 21, 2010).
[14] “Stress Testing in Wartime and in Peacetime,” Til Schuermann, Forthcoming in Stress Testing and Macroprudential Regulation: A Trans-Atlantic Assessment CEPR eBook edited by R. W. Anderson, March 2016, https://fic.wharton.upenn.edu/wp-content/uploads/2016/11/16-01.pdf
[15] See, e.g., Governor Daniel K. Tarullo, Speech at the Federal Reserve Third Annual Stress Test Modeling Symposium, “Stress Testing After Five Years,” (June 25, 2014), https://www.federalreserve.gov/newsevents/speech/tarullo20140625a.htm; Governor Daniel K. Tarullo, Speech at the Federal Reserve Board International Research Forum on Monetary Policy (Mar. 25 2010), https://www.bis.org/review/r100330e.pdf; see also Regulations Q, Y, and YY: Regulatory Capital, Capital Plan, and Stress Test Rules, 85 Fed. Reg. 15,576 (March 18, 2020).
[16] SIFMA, “Global Market Shock and Large Counterparty Default Study Recommendations for Reforms Based on a Statistical Analysis of Stress Testing Scenarios,” (Aug. 2019), https://www.sifma.org/wp-content/uploads/2019/09/SIFMA-GMS-LCD-Study-FINAL.pdf
[17] See Discussion with Vice Chair Clarida on the U.S Economic Outlook and Monetary Policy, video available: https://www.federalreserve.gov/newsevents/speech/clarida20200521a.htm.
