The Federal Reserve projected in its recently released annual stress test that large banks have enough capital to withstand the stress scenario it devised prior to the COVID-19 pandemic, but they may face significantly greater pressure if the economy experiences a long-lasting or double-dip downturn because of the pandemic. As a result, the Fed has put in place limitations on shareholder distributions, which has already led to an announced dividend cut for one bank (Wells Fargo) and could also lead to dividend cuts for others.

This year the Fed has tested the banks not only under its traditional severely adverse scenario, but also under three COVID-19-related scenarios: V-, U-, and W-shaped recessions. The U- and W-shaped COVID-19 stresses are much harsher than the Fed's traditional severely adverse scenario, once thought to be the "as bad as it can get" scenario. Specifically, in this year's traditional severely adverse stress scenario, the aggregate bank common equity Tier 1 (CET1) ratio declines to 9.9%. But in a COVID-19 stress overlay built specifically for current economic conditions and assuming a W-shaped recession, the aggregate bank CET1 ratio falls to 7.7%, with some banks approaching their minimum required capital levels. The results raise questions about the harshness of the Fed's ongoing severely adverse scenario and whether future stress test scenarios need to be more severe, which in turn would require banks to hold higher levels of ongoing capital. But that's a question for another day--one to ponder after the COVID-19 stress is behind us.

For now, the release of this year's stress test results brings with it the introduction of a new method to calculate banks' required capital ratios with buffers. Finalized in March, the stress capital buffer (SCB) is calculated as the burndown or decline in a bank's capital ratios under the Fed's severely adverse stress scenario (not its COVID-19 overlay) plus four quarters of dividends. Starting Oct. 1, and in effect for the following year, the previous static 2.5% capital conservation buffer will be replaced with the SCB. According to our analysis of this year's results, nine of the 33 banks that participated in this year's test have a preliminary SCB above 2.5%, thus raising their required ongoing capital levels. The Fed will finalize bank SCBs in August.

Despite the use of the SCB as a building block in a bank's regulatory required capital ratio, this year we expect the Fed to use its pandemic-related scenarios in ongoing bank discussions, and they will likely have a greater influence in how banks manage capital. In previous years, with the release of stress test results, there was a big focus on banks' capital plans, which show how much capital they project to return to shareholders over the next year. But this year, bank capital plans have taken a back seat to the impact of COVID-19 on bank earnings and capital. Indeed, investors had little to no expectation that banks would be repurchasing shares or increasing their dividends. The bigger question this year has not been how much capital the banks could return to shareholders, but rather whether they can sustain their current dividends.

So far, the Fed has not publicly stated that any individual bank needs to cut its dividend. However, as part of its stress test results, it did provide a formula for whether banks can continue to pay current dividends. Specifically, the Fed mandated that dividend payouts are not to exceed the average of a bank's net income for the four preceding quarters, which for some may prove troublesome if consensus earnings for the second quarter or for the year come to fruition. The Fed said this mandate applies through the end of the third quarter, but it could extend it on a quarter-by-quarter basis. So far, Wells Fargo is the only bank to have announced a dividend cut.

The Fed used its discretion this year to stipulate that due to the economic uncertainties caused by the pandemic, banks will need to do the following, starting in the third quarter and continuing on a quarter-by-quarter basis as the economic situation evolves:

• Not increase dividends for at least the third quarter;
• Keep share repurchases suspended for at least the third quarter; and
• Conduct additional stress analyses later this year as more data become available and economic conditions evolve. Banks will need to resubmit capital plans to the Fed within 45 days after the Fed provides updated stress scenarios.

The Fed indicated it would retain the option to extend these parameters in subsequent quarters, depending on the trajectory of the economic recovery.

How We View Banks' Dividend Cuts From A Ratings Perspective

We addressed our ratings perspective on dividend cuts in a recently published commentary (see "How U.S. Bank Dividend Cuts Could Affect Ratings," June 3, 2020). At a high level, if a bank were to cut its dividends, this decision on its own would not necessarily lead to a negative rating action because such a decision is also supportive of creditors. Instead, we would evaluate the impact on a case-by-case basis after considering the circumstances that led to the dividend cut.

For example, some banks entered this period with higher dividend payout ratios than others. If a bank were to cut its dividend as a defensive move to bring its dividend payout in line with peers, we may regard the net effect on creditworthiness as supportive. But an idiosyncratic dividend cut by a bank to levels well below peer medians could indicate expectations of substantially weaker earnings capacity or disproportionately higher credit losses relative to peers, which could have negative rating implications.

We expect nominal minimum bank shareholder payouts this year

Given the Fed's suspension of share repurchases and the inability to raise dividends until economic conditions improve, we expect bank payouts to only comprise current dividends for the foreseeable future. Assuming all banks continue to pay their current dividends, we estimate that the median bank dividend payout will equate to about 67% of the consensus estimate of 2020 earnings, compared with 33% last year for the 12 U.S. banks that participated.

Given the Fed's stipulation that dividends cannot exceed the average of net income over the preceding four quarters, it's likely that some banks will need to cut their dividends this year after releasing second-quarter results (the earnings season kicks off in mid-July). Using consensus second-quarter earnings estimates, we estimate Capital One and Wells Fargo may be above their four-quarter averages, therefore necessitating cuts in their dividends. In fact, after the stress test results were made public, Wells Fargo announced it expects to reduce its dividend from its current quarterly rate of $0.51 per share to an amount to be announced when it reports second-quarter earnings on July 14. The announcement of a dividend cut didn't have an immediate ratings impact. We will await Wells' second-quarter results and in the aftermath assess how the bank performs compared with similarly rated peers to further assess the rating. Looking further out to year-end, Ally Financial Inc., Citizens Financial Group Inc., Discover Financial Services, Huntington Bancshares Inc., and KeyCorp also may need to cut their dividends by the fourth quarter if consensus 2020 earnings estimates hold.

If economic conditions become more favorable later in the year or next year than banks currently project, according to the construct of the SCB final rule, banks would have more flexibility to increase share payouts than they previously had after the stress test. Under the final rule, a bank will be required to provide the Fed with notice within 15 days after making any capital distributions in excess of those included in its capital plan. A bank that has resubmitted its capital plan or received a qualitative objection to its capital plan must continue to request approval from the Fed for incremental capital distributions.

Table 1

Assessing the large banks' new required capital ratios with buffers

In March, the Fed finalized changes to capital rules for the largest banks (those with greater than $100 billion in assets) (see "Credit FAQ: The Fed's New Rules Change Capital Management Dynamics For U.S. Banks," March 19, 2020). A bank's SCB is calculated as: the difference between its starting and minimum projected CET1 capital ratios under the severely adverse scenario in the supervisory stress test, plus four quarters of planned common stock dividends as a percentage of risk-weighted assets. To determine the SCB, the Fed assumes that banks maintain a constant level of assets over the planning horizon. The SCB requirement has a floor of 2.5% of risk-weighted assets.

In table 2, we show our estimates of banks' SCBs and their new required capital ratios (equal to the sum of minimum capital ratios, the SCB, and a global systemically important bank [GSIB] surcharge when applicable). We note that the SCBs are not final (they will be finalized in August), and therefore our estimates are preliminary but in line with those banks that disclosed their SCBs after the results of the stress test were made public. Notably, this year, a bank may also request reconsideration of its SCB requirement by submitting a written request and rationale to the Fed within 15 days of receipt of its SCB requirement. If the Fed agrees to a reconsideration, a bank's SCB could change. So far, Citizens Financial and BMO Financial Corp. are pursuing this route.

For now, our initial analysis suggests that most of the banks subject to the stress test have current capital ratios that are equal to or above their new required minimums including buffers (the SCB and GSIB surcharge). We presume that banks whose existing CET1 ratios fall short of the new estimated required minimums (including the SCB and GSIB surcharge) will aspire to build to those levels by some combination of retaining earnings, reducing risk-weighted assets, or cutting dividends. Only two banks appear below or close to their proposed minimums by our calculation: Goldman Sachs and JPMorgan.

Table 2

Ranking the banks according to the severely adverse scenario

One useful analytical exercise in the aftermath of the stress test results is to rank the banks by their capital declines, without taking into account their capital payout requests. To do so, we look at a bank's starting CET1 ratio and its minimum CET1 ratio during the nine quarters of the stress test. The smaller the decline from the beginning ratio to the minimum ratio (or burndown), the stronger the bank's capital and earnings position is relative to its risk, according to the Fed's stress test. A few things to note:

• The burndown of capital for most banks was lower this year than last year (and lower than in 2018 for banks that didn't participate in last year's test). We believe revised assumptions in this year's test with respect to banks maintaining current dividends or keeping balance-sheet growth static accounted for the lower burndown. Still, some banks' results showed higher capital burndown in this year's test. These variations could be due to either changes in the risk characteristics of a bank's portfolio, declines in its preprovision revenue, or Fed parameters that modeled a more negative impact on a bank's business model than in previous years.

• With respect to capital burndown, Northern Trust, American Express, TD Group US, U.S. Bancorp, State Street, Bank of New York Mellon, and PNC performed best, with the lowest burndown. This is due to a combination of low loan losses and higher after-stress preprovision revenue for these banks. Not coincidentally, these are among the highest-rated banks in the U.S.

• Among U.S.-domiciled banks, Goldman Sachs, Capital One, and Morgan Stanley fared worst in terms of burndown. That said, all of these banks have among the highest starting capital levels, indicating significant cushions to absorb sizable declines in their regulatory capital ratios. Morgan Stanley's capital burndown, while high, also improved substantially, at only 530 basis points (bps) this year, compared with 800 bps in the 2019 test. It is difficult to ascertain the precise cause of the improvement based on Fed disclosures, and the inclusion of dividends in the 2019 results explains only a portion of the change.

Table 3

Loan loss rates are another useful input

The median loan loss rates as a percentage of average loan balances improved marginally in aggregate in this year's stress test compared with the previous two years. Specifically, the median loss rate was 5.7% this year, compared with 6.1% when combining 2019 and 2018 stress test data. Capital One and Discover Financial showed the largest increases in loss rates from previous years; Goldman Sachs, KeyCorp, and M&T showed the biggest improvements.

We also analyzed the reported gap between an individual bank's loss rate by loan type and the aggregate loss rate for that loan category. A bigger gap could point to riskier lending than peers.

Table 4  |  View Expanded Table

Preprovision revenue is also a telling sign of how banks will fare under stress

A bank that can maintain strong preprovision revenue (PPNR), even under stress, can better withstand credit losses, given the strength of its business model. Since banks are already experiencing stress from COVID-19, we find it useful to assess how their actual PPNR holds up compared with stressed PPNR, according to the Fed. To do so, we annualized first-quarter results and compared them with the Fed's severely adverse stress PPNR. Most banks in the first quarter posted annualized PPNR significantly above the Fed's assessment. But for a few banks (Wells Fargo, Fifth Third, and Ally), first-quarter PPNR was already running below the Fed's assessment. Second-quarter PPNR could be much lower than the first-quarter figure because it more fully incorporates COVID-19 stress.

Further, the Fed-modeled PPNR changed for banks in this year's stress test. Some banks reported a large increase in PPNR (Morgan Stanley, Fifth Third, and JPMorgan) over the previous year's results. Conversely, Wells Fargo, Northern Trust, Huntington, Regions, Ally, M&T, and Goldman showed declines in PPNR from previous tests. While it is possible that these variations in PPNR from year to year could correlate with changes in earnings potential, we remain cautious about drawing firm conclusions from this analysis because of insufficient details on the Fed's assumptions for individual banks and the related potential for model-driven risk.

Table 5

How the stress tests results figure into our ratings process

Although the stress test is not a direct input into our ratings, we incorporate insight from the test results into our assessments in many ways. We consider the stress test's burndown analysis and loan loss rates to identify outliers in our own independent credit reviews. Large differences between a bank's internal stress test results (which are published separately) and the Fed's results are also useful to explore with bank management teams. We incorporate the loss rates derived from the Fed's stress in our own internally run stress test (see "For Large U.S. Banks, Loan Loss Expectations Will Be Key To Ratings," May 5, 2020). Last but certainly not least, we expect to use the bank capital plans divulged in the aftermath of the test to help guide our proprietary projections for banks' risk-adjusted capital ratios.

Some aspects of the stress test burndown analysis and loan loss rates may align with our ratings view. In other words, many banks that perform well on the stress test are also high rated. That said, variations are possible based on factors that are not addressed by the Fed's stress tests. Our rating on a bank could be lower than the stress test results might suggest due to, for example, a more negative view of a lack of revenue diversification by location or business, or due to our expectations with respect to the likelihood of support from a foreign parent.

Appendix

What happens if a bank breaches its capital buffer?

Since the SCB doesn't go into effect until October, a bank technically has until the end of December to ensure its capital ratio is above its new required minimum with the SCB. If a bank were still in breach of its buffer by year-end, restrictions on dividends and executive bonuses could come into play. The extent of the buffer breach is an important consideration, as are a bank's earnings for the last four quarters.

Guidance from the Fed suggests regulators will calculate the percentage of "eligible retained income" permissible to be paid out as shareholder distributions and executive bonuses (see table 6 for how regulators mandate permissible percentage payouts based on the size of the buffer breach). They calculate eligible retained income as the higher of the following two items to determine the maximum distribution:

• Net income over a cumulative four quarters, net of distributions and associated tax effects not already reflected in net income
• The average of a banking organization's net income over the preceding four quarters (that is, gross of dividends)

Table 6

One development that will help banks build capital or stay above their required buffers is regulators' delaying of the full impact of the current expected credit losses (CECL) methodology on CET1 capital ratios. Specifically, the Day One impact of CECL is not included in banks' CET1 ratios, nor is 25% of their allowance build starting in the first quarter. As a result, bank capital ratios will not be as harshly affected as they otherwise would have been as credit losses further materialize.

Key details in assessing the Fed's COVID-19 overlay

Given the uncertainties in the economy, the Fed constructed a COVID-19-specific stress test to go along with its original severely adverse test. For the scenario analysis, the Fed showed the results only in aggregate, providing quartile analysis for minimum capital ratios (see table 7). According to the Fed, no bank breached its minimum required capital ratio of 4.5%; that said, some banks seemed to come precariously close. In addition, the capital depletion that would result from common equity distributions over the projection horizon was not included. The inclusion of common equity distributions during the first half of 2020 would have resulted in a reduction in aggregate capital ratios of approximately 50 bps.

Table 7

The Fed constructed its COVID-19 sensitivity analysis using three alternative downside scenarios:

• V-shaped recovery: In this scenario, the economy regains much of the output and employment lost by the end of 2020. The Fed stated that this scenario is akin to its severely adverse scenario (see below for particulars of this year's severely adverse scenario). This scenario maintains the same level of financial stress as in the severely adverse scenario but concentrates the macroeconomic stress early in the projection horizon.
• U-shaped recovery: This is a much slower recovery, in which only a small share of lost output and employment is regained in 2020. The path of the economy in this scenario is consistent with the possibility of prolonged social distancing to combat ongoing outbreaks of the virus across the country.
• W-shaped recovery: This scenario incorporates a double-dip recession with a short-lived recovery, followed by a severe drop in activity later this year due to a second wave of COVID-19, which leads to a second increase in unemployment and a drop in GDP.

In the V-, U-, and W-shaped scenarios, the Fed used net charge-off ratios of 8.2%-10.3%, 1.2x-1.5x greater than what was experienced in the 2008 global financial crisis.

Notably, none of the scenarios from the Fed takes into account the possible positive impact from all of the government stimulus programs and the recent market rebound.

Table 8

Other differences between the Fed's severely adverse scenario and its COVID-19 overlay

Besides worse economic declines, the Fed's COVID-19 scenario analyses also included two other more severe factors as compared with the severely adverse scenario:

• Unlike in its severely adverse scenario, in which bank balance sheets remain static, in the COVID-19 scenario analyses, the Fed assumes corporate loans for each bank grow by 12%, similar to levels experienced in the first quarter. This adjustment not only increases banks' balance sheets but also their risk-weighted assets.
• To capture stress on certain businesses, the Fed assumes the internal ratings on borrowers in affected industries--such as the retail, travel-related, and food industries--are lowered by one notch.

Changes to this year's stress test from last year's

The following changes were incorporated in this year's stress test:

• A COVID-19 overlay was included.
• In this year's severely adverse scenario, the Fed published results for 33 banks, whereas last year it included only 18.
• The adverse scenario was removed.
• The Fed assumed a bank maintains a constant level of assets over the projection horizon (last year it assumed balance-sheet growth) and that a bank will not pay any common dividends (last year it assumed that current dividends will be continued).
• A stress leverage buffer requirement was excluded.
• Banks can increase their planned capital distributions in excess of the amount included in their capital plans without prior approval from the Fed, so long as they remain above their new minimum risk-based ratios.
• Banks can appeal their SCB calculation.
• Qualitative objections were removed for all banks except foreign banks. (Last year, the Fed removed the qualitative objection for the smaller banks.) Still, banks with weak practices may be subject to a deficient supervisory rating and potential enforcement action.

Key assumptions in this year's severely adverse scenario

This year's severely adverse scenario included a severe global recession, accompanied by a period of heightened stress in commercial real estate (CRE) and corporate debt markets; a peak in the U.S. unemployment rate of 10% in third-quarter 2021; real GDP contraction of 8.5% from its prerecession peak, reaching a trough in third-quarter 2021; three-month Treasury bills falling near zero and remaining there through the end of the scenario; the 10-year Treasury yield dropping to 0.75%; equity prices falling 50%; and house prices and CRE prices experiencing large overall declines of 28% and 35%, respectively. The global market shock component included widespread defaults on a variety of debt instruments by business borrowers, with the leveraged loan market coming under considerable pressure, and a counterparty default component that involves the instantaneous and unexpected default of the firm's largest counterparty.

Related Criteria

• Group Rating Methodology, July 1, 2019
• Risk-Adjusted Capital Framework Methodology, July 20, 2017
• Banks: Rating Methodology And Assumptions, Nov. 9, 2011

Related Research

• Capital Markets Revenue Should Be A Bright Spot For Banks In A Tough 2020, Article Says, June 23, 2020
• How U.S. Bank Dividend Cuts Could Affect Ratings, June 3, 2020
• For Large U.S. Banks, Loan Loss Expectations Will Be Key To Ratings, May 5, 2020
• Most Banks Breezed Through The Fed's Stress Test--And Can Now Return Significant Capital To Shareholders, July 1, 2019