In an effort to gauge how U.S. banks would perform in the worsening economic conditions due to COVID-19, the Federal Reserve conducted a second stress test in 2020. Typically, it does only one per year, but the first stress test scenarios had been articulated prior to COVID-19. The recently released results showed that the 33 banks that participated in the test are well-positioned to maintain capital levels above the minimum under two hypothetical scenarios.

In particular, the Fed updated its severely adverse scenario and provided an alternative severe scenario (see appendix for details). The results showed that in aggregate, for the 33 banks, their common equity Tier 1 (CET1) ratio declined from its starting point of 12.2% in June 2020 (20 basis points higher than where it was in December 2019) to a minimum of 9.6% under the severely adverse and 9.7% under the alternative severe scenario, well above the regulatory minimum of 4.5% (see table 1).

The results of December's test were more benign compared with the more dire possibilities presented in June's additional scenarios. As a result, the Fed is now allowing the banks that participated in the test to repurchase a limited amount of shares in the first quarter, based on the level of their recent income.

Despite greater projected loan losses, the results of the December test were not far worse than those in the severely adverse scenario of the June test. The June test showed a 2.1% CET1 ratio reduction to a minimum 9.9%, compared with December's 2.6% decline and 9.6% minimum in the severely adverse scenario. (For the results of the test in June, see "The Fed’s Latest Stress Test Points To Limited Bank Capital Returns," June 30, 2020.)

One of the reasons for that is capital ratio starting points were higher because of capital build in the first half of the year due to restricted shareholder payouts. Also, banks have built significant allowances, helping to cushion the Fed's larger assumed losses (greater than $600 billion in December's test versus $550 billion in June). As a result, projected provisions were lower in the December test than the June test. That said, the allowance build is partially offset by lower pre-provision net revenue assumptions in December's test (roughly $370 billion versus $430 billion in June).

Table 1

Less Severe Stress Than In June's W-Shaped And U-Shaped Scenarios

The results of December's stress test are in sharp contrast to pandemic-related overlay analyses the Fed conducted in June, in addition to its traditional severely adverse scenario. At that time, the Fed projected very sharp declines in capital in special U- and W-shaped economic recovery scenarios meant to take into account the possible impact of COVID-19. It projected aggregate minimum CET1 ratios of 8.1% and 7.7% in those scenarios, respectively, with some banks almost reaching their minimum capital requirement of 4.5%. Notably, the aggregate minimum CET1 ratio of 9.7% in December's alternative severe scenario is well above those levels.

We believe that's because banks and the economy have performed better than may have been feared early in the pandemic, and the type of stress incorporated in the U- and W-shaped scenarios now looks less likely. For instance, the W-shape had assumed a 12.4% peak-to-trough GDP decline and a loan loss rate of 9.9%. By comparison, December's severely adverse scenario has a 3.25% GDP decline and a loan loss rate of 7.7%, and the alternative severe scenario has an annualized GDP decline of 9% in fourth-quarter 2020 and growth of 2% in 2021.

By the end of the first quarter, the Fed will revisit its updated payout restrictions. Separately, as a result of banks' plan submissions, the Fed may elect to recalculate a bank's stress capital buffer (SCB)--the burndown, or decline, in a bank's capital ratios under the Fed's severely adverse stress scenario to minimum levels. The Fed has extended the deadline by which it can notify banks whether their SCB requirements will be recalculated to March 31, 2021. We calculate that 10 of 22 U.S.-domiciled banks' capital burndown in the December test was more than 50 basis points (bps) higher than in the June test. M&T showed the most significant change of a roughly 450 bps decline in capital, versus a 120 bps burndown in June (table 3). Alternatively, seven of the U.S.-domiciled banks' burndown was less in December than in June.

Our ratings reflect the resiliency of the U.S.-domiciled banks that participated in the December stress test. Of the 22 U.S.-domiciled banks that were included in the stress tests, about 80% have stable outlooks. Our negative outlooks on Ally Financial, Capital One, Discover Financial Services, Huntington Bancshares, and M&T Bank Corp. largely reflect either these banks' concentration in loan sectors we believe are potentially most vulnerable to COVID-19 (largely auto, credit cards, and commercial real estate), or merger-related activities.

In the aftermath of December's stress test results, we have lowered our loss rate assumption for our rated banks to 2.2% from our previous assumption of 3% over the next two years. This revision to our U.S. banking sector loss forecast is based on a combination of recent positive developments, including:

• Vaccine development and distribution in 2021;
• Greater stabilization in the economic outlook on the heels of the passage of additional fiscal stimulus;
• A third-quarter pause in bank allowance build; and
• Recent anecdotes from bank management teams suggesting a more sanguine outlook for credit losses in the quarters to come.

Capital Return Analysis

In the December stress test results, the Fed left in place its restriction on dividend increases but eased its prohibition of share repurchases starting in the first quarter of 2021. Specifically, the Fed will allow shareholder payouts (the sum of dividends and share repurchases) in the first quarter up to an amount equal to each bank's average quarterly net income in 2020. Based on this formula, we estimate that all of the U.S.-domiciled banks can opt to commence repurchasing shares in the first quarter but to varying degrees (see table 2). Notably, since the stress test results were announced, several banks have already declared their intention to repurchase shares in the first quarter. Still, total payout ratios (dividends and possible share repurchases in the first quarter) for many banks would constitute less than 100% of a run rate of projected fourth quarter-earnings.

The ability to repurchase shares aligns with our expectations even prior to the release of the stress test results, and explains why we had assumed capital ratios would decline in 2021.

In addition, the Fed is allowing all banks that participated in the test to make share repurchases that equal the amount of share issuances related to employee compensation. It's also allowing banks to redeem and make scheduled payments on additional Tier 1 and Tier 2 capital instruments.

Table 2

Stress Burndown Analysis

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.

Given the unique features of the Fed conducting a second stress test this year, we compared the results in December with those in June (see table 3). A few things to note:

• The burndown of capital for most banks under the December severely adverse and alternative severe scenarios was more severe than the June 2020 test due to the more stringent modelling assumptions.
• That said, some banks showed a lower burndown in the December test versus June. The U.S.-domiciled banks that fall into this category are Capital One Financial, Goldman Sachs, Morgan Stanley, Ally, JPMorgan, Truist, and Regions.
• Alternatively, some U.S.-domiciled banks fared worse in December's stress test versus June, with their capital burndown greater than 50 bps. These include M&T, Citizens Financial, Wells Fargo, Bank of America, Citigroup, Fifth Third, State Street, Northern Trust, Bank of New York Mellon, PNC, and U.S. Bancorp.
• Of all the U.S. domiciled-banks, M&T's CET1 ratio came closest to the minimum, totaling 5.0% at the end of the Fed's stress period, versus a 4.5% regulatory required minimum.

Table 3

Loan Loss Rate Analysis

The aggregate loan loss rate in December's severely adverse scenario totaled 7.7% (see table 4), with credit cards (22.3%) and CRE (12.6%) the sectors with the highest amount of loan losses. In December's alternative severe scenario, the aggregate loss rate totaled 7.3% (credit card and CRE are 20.3% and 11.9%, respectively). This compares with June's severely adverse scenario, which showed a loan loss rate of 6.3%, with credit cards at 17.1% and CRE at 6.3%. All individual banks' aggregate projected loss rates rose under the new scenarios, with a few banks' loss rates almost doubling from where they were in June. Notably, the aggregate commercial real estate loss rate almost doubled, mirroring our concerns for this asset class (see "U.S. Banks Face Long-Term Risks To Their Commercial Real Estate Asset Quality," Nov. 16, 2020).

Table 4

Updated Loss Rate Assumptions

S&P Global Economics expects U.S. real GDP to contract 3.9% this year and grow a modest 4.2% in 2021 (slightly better than the September forecast, which assumed a 4.0% contraction in 2020 and 3.9% growth in 2021). Given somewhat greater stabilization than we had projected at the outset of the pandemic, and growing promise regarding vaccine distribution, we have lowered our loss rate assumption for our rated U.S. banks to 2.2% from our previous projection of 3% over the next two years. We expect charge-offs to pick up in 2021 and likely decelerate in 2022. Notably, through the first three quarters of 2020, all banks we rate have taken provisions of 1.2% of loans, which suggests they are already well-positioned to absorb a little over half of our projected loss rates.

U.S. Banks Generally Remain Resilient

The Fed's stress test results and our own analysis suggest that most U.S. banks are heading into 2021 with relatively solid balance sheets and decent pre-provision revenue, all things considered. The Fed's explicit approval of a resumption of share repurchases also aligns with the largely stable outlooks we have maintained on our ratings on the largest banks. Although there is still uncertainty, we believe the stable outlooks indicate why rated banks should be able to handle additional potential shocks in the aftermath of the shutdowns that began in March 2020. U.S. banks so far have been more resilient--in contrast to the distress they experienced a decade ago.

Appendix

Table 5 summarizes the key features of December's severely adverse and alternative severe scenarios, as well as June's W-shaped scenario.

Table 5

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

• Despite Declining Loss Provisions, U.S. Banks Still Face Asset Quality Risks And Low Interest Rates, Nov. 19, 2020
• Earnings Among Large U.S. Banks Rebounded In Third Quarter, But Uncertainty Remains High, Nov. 17, 2020
• The Fed's Latest Stress Test Points To Limited Bank Capital Returns, June 30, 2020