Board of Governors of the Federal Reserve System

Supervisory Stress Test Results

This section describes the Federal Reserve's projections of losses, revenues, expenses, and capital positions for the 33 firms in the December stress test under the severely adverse and alternative severe scenarios. Results are presented both in the aggregate and for individual firms. The aggregate results reflect the sensitivities of losses, revenues, and capital at these firms as a group to the stressed economic and financial market conditions contained in those scenarios. The range of results across individual firms reflects differences in business focus, asset composition, revenue and expense sources, and portfolio risk characteristics. The comprehensive results for individual firms are reported in appendix B.

Changes in supervisory stress test results across exercises reflect changes in

• firm starting capital positions;
• scenarios used for the supervisory stress test;
• portfolio composition and risk characteristics; and
• models used in the supervisory stress test.

While the results under the two scenarios described here are approximately similar, there are differences designed to reflect a broad set of severe but plausible risks. Generally, the alternative severe scenario exhibits a smaller initial decrease in economic activity but a more sluggish recovery relative to the severely adverse scenario. The different scenario paths lead to slightly larger PPNR and trading revenue projections under the severely adverse scenario as equity asset values recover faster. Additionally, the faster recovery leads to lower loan losses and lower provision expenses over the full 13 quarters under the severely adverse scenario. In the aggregate, each of the five capital and leverage ratios declines similarly but subsequently rises more under the severely adverse scenario.

Severely Adverse Scenario

Under the supervisory severely adverse scenario, the aggregate CET1 ratio is projected to decline to a minimum of 9.6 percent before rising to 10.2 percent at the end of nine quarters (see table 3). In the aggregate, each of the five capital and leverage ratios declines over the course of the projection horizon from their second quarter of 2020 levels, with third quarter of 2022 levels ranging from 1.2 percentage points to 2.2 percentage points lower than at the start of the projection horizon (see table 3).

The changes in post-stress capital ratios vary considerably across firms (see figure 11), and table 4 presents these ratios for each of the 33 firms. Differences in the declines in ratios across firms are primarily related to differences in the Federal Reserve's projections of losses, revenues, and expenses.

Projected Losses

The Federal Reserve projects that the 33 firms as a group would experience significant losses on loans and other positions under the severely adverse scenario. In this scenario, losses are projected to be $629 billion for the 33 firms in the aggregate over the nine quarters of the projection horizon.

These losses include

• $514 billion in accrual loan portfolio losses;
• $4 billion in securities losses;³⁸
• $95 billion in trading and counterparty losses at the 13 firms with substantial trading, processing, or custodial operations; and
• $16 billion in additional losses from items such as loans booked under the FVO (see table 3).

Losses on accrual loan portfolios account for 82 percent of the projected losses for the 33 firms, while trading and counterparty losses account for 15 percent (see figure 10).

Loan Losses

Total loan losses are $514 billion for the 33 firms in the December stress test. For the same firms, total loan losses were $433 billion in the June stress test.

Loan losses on consumer products (domestic residential mortgages, domestic junior liens and HELOCs, credit cards, and other consumer loans) represent a slightly smaller share of losses than losses on commercial products (domestic CRE, C&I loans, and other loans) (see table 3). Consumer and commercial products represent 38 and 44 percent of total projected losses, respectively. C&I loan losses and credit card losses are the two largest categories of loan losses at $121 billion and $158 billion, respectively.

The aggregate nine-quarter cumulative loss rate for all accrual loan portfolios is 7.7 percent, where the loss rate is calculated as total projected loan losses over the nine quarters of the projection horizon divided by average loan balances over the horizon. However, total loan loss rates vary significantly across firms, ranging between 1.7 percent and 21.3 percent across these institutions (see table 8 and figure 12).

Firms' loan loss rates reflect differences in the portfolios held by each firm and the characteristics of the loans within each portfolio. Loan portfolio composition matters because projected loss rates vary significantly for different types of loans. In the aggregate, nine-quarter cumulative loss rates vary from 2.1 percent on domestic first-lien mortgages to 22.3 percent on credit cards, reflecting both differences in typical performance of these loans and differences in the sensitivity of different types of loans to the scenarios. In particular, lending categories for which performance is sensitive to credit spreads or unemployment rates may experience high stressed loss rates due to the considerable stress on these factors in the severely adverse and alternative severe scenarios.³⁹

Projected loss rates on most loan categories show similar dispersion across firms (see table 8 and figures C.1 through C.7). There are significant differences across firms in the projected loan loss rates for similar types of loans. For example, while the median projected loss rate on C&I loans is 7.3 percent, the rates among firms with C&I loans vary from a low of 1.2 percent to a high of 27.0 percent. For credit card loans, the range of projected loss rates is from 13.3 percent to 35.8 percent, with a median of 23.7 percent.

Differences in projected loss rates over time primarily reflect changes in loan and borrower characteristics and changes in the scenarios. The overall loan loss rate in the December stress test is higher than the June stress test due to the generally higher projected loss rates across portfolios, with larger increases for severely affected portfolios such as CRE. The increase in risk was offset somewhat by a shift in the composition of firms' loan portfolios away from portfolios with higher loss rates. In particular, consumer credit card balances have declined significantly since the start of the year as a result of reduced consumer spending and higher loan repayment rates.

Losses on Trading, Private Equity, SFT, and Derivatives Positions

The severely adverse scenario results include $95 billion in trading and counterparty losses generated from the global market shock and LCPD components. For the 13 firms subject to one or both components, losses ranged from $0.9 billion to $23 billion (see table 6).

The relative size of losses across firms depends on the specific risk characteristics of each firm's trading positions, inclusive of hedges. Importantly, these projected losses are based on the trading positions and counterparty exposures held by these firms on the same as-of date (June 30, 2020) and could have differed if they had been based on a different date.

Projected PPNR

In the aggregate, the 33 firms are projected to generate $371 billion in PPNR cumulatively over the nine quarters of the projection horizon, equal to 2.0 percent of their combined average assets (see table 3). PPNR projections are driven by the shape of the yield curve, the path of asset prices, equity market volatility, and measures of economic activity in the severely adverse scenario. In addition, PPNR projections incorporate expenses stemming from estimates of elevated levels of losses from operational-risk events such as fraud, employee lawsuits, litigation-related expenses, or computer system or other operating disruptions.⁴⁰ In aggregate for the 33 firms, operational-risk losses are $160 billion for the December stress test. For the June stress test, operational-risk losses were $144 billion.

Aggregate PPNR as a percent of combined average assets is lower under the severely adverse scenario in the December stress test than in the June stress test. Lower PPNR is partly due to lower net interest income, reflecting a sharper decrease in the term spread under the severely adverse scenario. There were also fairly sizable changes in firms' balance sheets between December 2019 and June 2020. All else equal, larger balance sheets result in higher revenue that is roughly offset by higher expenses in the stress test projections. Noninterest income remained higher than in the June stress test, supported by stronger trading revenues and investment banking fees. Stronger noninterest income partly dampened the decrease in PPNR.

The ratio of projected cumulative PPNR to average assets varies across firms (see figure 13). A significant portion of this variation reflects differences in business focus across the institutions. For instance, the ratio of PPNR to assets tends to be higher at firms focusing on credit card lending, reflecting the higher net interest income that credit cards generally produce relative to other forms of lending.⁴¹ Importantly, lower PPNR rates do not necessarily imply lower net income, because the same business focus and revenue risk characteristics determining differences in PPNR across firms could also result in offsetting differences in projected losses across firms.

Net Income and Regulatory Capital Treatment

Projected PPNR and provisions for loan losses are the primary determinants of projected pre-tax net income. The projected decline in pre-tax net income is 0.9 percent of average total assets, compared to a decline of 1.1 percent in the June stress test. Table 6 presents projections of the components of pre-tax net income, including provisions into the allowance and one-time income and expense and extraordinary items, under the severely adverse scenario for each of the 33 firms (see table 3 for aggregate). The projections are cumulative for the nine quarters of the projection horizon.

The Federal Reserve's projections of pre-tax net income under the severely adverse scenario imply negative net income at most of the 33 firms individually and for the firms as a group over the nine-quarter projection horizon. Projected pre-tax net income is an aggregate net loss of $173 billion over the projection horizon for the 33 firms. For the June stress test, projected pre-tax net income was a slightly larger loss of $177 billion even though CET1 declined less in June than in the December stress test. Further reducing CET1 are higher projected loan charge-offs combined with lower projected revenue over the projection horizon, which results in larger deferred tax assets. These assets are deducted from CET1 because they are not considered available to absorb losses.

The pre-tax net income projections incorporate loan losses through provisions, which equal projected loan losses plus the amount needed for the allowance to be at an appropriate level at the end of each quarter. The $429 billion in total provisions includes $514 billion in net charge-offs, with the remainder being the reserve release. These amounts are cumulative over the projection horizon and do not fully reveal variation in the allowance during the course of the nine quarters. Specifically, the projected allowance increases during the early quarters of the projection horizon, given the increased economic stress in the severely adverse scenario, and then declines as the economic stress abates.

While loan losses are higher for the December stress test than in the June stress test, the projected reserve release is due to the sharp increase in firms' loan-loss reserves since the beginning of the year. As firms simultaneously responded to the economic contraction and the adoption of CECL, loan-loss reserves more than doubled in the first two quarters of 2020.⁴² Without the increase in loan-loss reserves, projected provision expenses would increase by $100 billion.

The ratio of pre-tax net income to average assets for each of the 33 firms ranges from −3.3 percent to 3.3 percent (see figure 14). Projected cumulative pre–tax net income for most of the firms (29 of 33) is negative over the projection horizon. Differences across the firms reflect differences in the sensitivity of the various components of net income to the economic and financial market conditions in the supervisory scenarios. Additional variation in projected net income results from the effect of the global market shock and LCPD components that affect 13 of the 33 firms.

Firms that are required to include AOCI in regulatory capital and those that opt in to including it are also affected by OCI (see table 6). OCI is driven by unrealized gains and losses on AFS securities in the supervisory stress test. The severely adverse scenario features a lower initial level for the 10-year Treasury yield, leading to a relatively smaller decline in the yield in the first projection quarter and subsequently lower projected unrealized gains for AFS securities. The interest rate path and credit spreads assumed in the scenario result in $5.2 billion of OCI over the nine quarters of the projection horizon for firms required to include AOCI in regulatory capital and those that opt in to including it.

Table 2. Applicable capital ratios and calculations for firms in the December stress test

Table 3. 33 participating firms* Projected stressed capital ratios, risk-weighted assets, losses, revenues, net income before taxes, and loan losses Federal Reserve estimates: Severely adverse scenario

 

Capital ratios, actual 2020:Q2 and projected 2020:Q3–2022:Q3

Projected loan losses, by type of loan, 2020:Q3–2022:Q3

Risk-weighted assets, actual 2020:Q2 and projected 2022:Q3

Projected losses, revenue, and net income before taxes through 2022:Q3

Table 4. Projected minimum common equity tier 1 capital ratio under the severely adverse scenario, 2020:Q3–2022:Q3 33 participating firms*

Table 5.A. Capital ratios, actual 2020:Q2 and projected 2020:Q3–2022:Q3 under the severely adverse scenario:Risk-based Category I, II, and III firms

Table 5.B. Capital ratios, actual 2020:Q2 and projected 2020:Q3–2022:Q3 under the severely adverse scenario: Risk-based Category IV firms*

Table 5.C. Capital ratios, actual 2020:Q2 and projected 2020:Q3–2022:Q3 under the severely adverse scenario: 33 participating firms

Table 6. Projected losses, revenue, and net income before taxes through 2022:Q3 under the severely adverse scenario: 33 participating firms*

Table 7. Projected loan losses by type of loan for 2020:Q3–2022:Q3 under the severely adverse scenario: 33 participating firms

Table 8. Projected loan losses by type of loan for 2020:Q3–2022:Q3 under the severely adverse scenario: 33 participating firms

Alternative Severe Scenario

The projected maximum declines in post-stress capital ratios in the alternative severe scenario are approximately similar to those under the severely adverse scenario. In the aggregate, each of the five capital and leverage ratios declines to a similar minimum as the severely adverse scenario, but the ratios rise less over the remaining projection horizon, reaching a lower ending value than the severely adverse scenario. The levels at the end of the projection horizon in the third quarter of 2022 range from 1.4 percentage points to 2.3 percentage points lower than at the start of the projection horizon (see table 9) for the five capital and leverage ratios. The changes in post-stress capital ratios vary considerably across the 33 firms (see figure 16 and table 10).

Generally, the slightly lower ending capital and leverage ratios reflect the sluggish recovery and relatively higher persistence of stress in the alternative severe scenario compared to the severely adverse. The alternative severe scenario features a less-severe initial drop in global economic activity relative to the severely adverse scenario but a more sluggish and severe subsequent recovery. Unemployment, volatility, and corporate spreads stay elevated for a longer period of time, while the term spread recovers slightly faster.

PPNR is slightly lower compared to the severely adverse scenario due to lower trading revenue from the slower recovery in equity asset values. This lower revenue is only partially offset by higher interest income as a result of wider net interest margins compared to the severely adverse scenario. While larger balance sheets result in higher revenue projections, the increased revenue is roughly offset by higher expenses in the stress test projections.

Losses over the full 13-quarter projection horizon are higher in the alternative severe scenario, leading to higher projected loan loss provisions.⁴³ Loss rates in the first nine quarters are lower in the alternative severe scenario, as the sluggish recovery pushes losses further out into the projection horizon than in the severely adverse scenario. Other losses and gains are slightly higher in the alternative severe scenario, reflecting the larger markdowns for loans held under the FVO as spreads stay elevated for a longer period of time. The alternative severe projections imply lower aggregate pre-tax net income, which slightly reduces post-stress ending capital ratios.

Projected Losses

The Federal Reserve projects that the 33 firms as a group would experience significant losses on loans and other positions under the alternative severe scenario. In this scenario, losses are projected to be $612 billion for the 33 firms in the aggregate over the nine quarters of the projection horizon.

These losses include

• $491 billion in accrual loan portfolio losses;
• $4 billion in securities losses;⁴⁴
• $95 billion in trading and counterparty losses at the 13 firms with substantial trading, processing, or custodial operations; and
• $22 billion in additional losses from items such as loans booked under the FVO (see table 9).

Losses on accrual loan portfolios account for 80 percent of the projected losses for the 33 firms, while trading and counterparty losses account for 16 percent (see figure 15).

Loan Losses

Total loan losses are $491 billion for the 33 firms (see table 9). For the June stress test, total loan losses were $433 billion for the same firms. Consumer and commercial products represent 36 and 44 percent of total projected losses, respectively. C&I loan losses and credit card losses are the two largest categories of loan losses at $121 billion and $144 billion, respectively.

For the full group of 33 firms, the nine-quarter cumulative loss rate for all accrual loan portfolios is 7.3 percent, where the loss rate is calculated as total projected loan losses over the nine quarters of the projection horizon divided by average loan balances over the horizon. However, total loan loss rates vary significantly across firms, ranging between 1.6 percent and 19.5 percent across these institutions (see table 14 and figure 17).

Losses on Trading, Private Equity, SFT, and Derivatives Positions

The global market shock and LCPD components in the alternative severe scenario are the same as those reported in the severely adverse scenario.

Projected PPNR

In the aggregate, the 33 firms are projected to generate $363 billion in PPNR cumulatively over the nine quarters of the projection horizon, equal to 1.9 percent of their combined average assets* (see table 9). The Federal Reserve's PPNR projections are driven by the shape of the yield curve, the path of asset prices, equity market volatility, and measures of economic activity in the alternative severe scenario. In addition, the PPNR projections incorporate expenses stemming from estimates of elevated levels of losses from operational-risk events such as fraud, employee lawsuits, litigation-related expenses, or computer system or other operating disruptions.⁴⁵ In aggregate for the 33 firms, operational-risk losses are $162 billion for the December stress test. For the June stress test, operational-risk losses were $144 billion.

Net Income and Regulatory Capital Treatment

Projected PPNR and losses are the primary determinants of projected pre-tax net income. The projected decline in pre-tax net income is 1.1 percent of average total assets, which is the same as the decline of 1.1 percent in the June stress test. Table 12 presents projections of the components of pre-tax net income, including provisions into the allowance and one-time income and expense and extraordinary items, under the alternative severe scenario for each of the 33 firms (see table 9 for aggregate). The projections are cumulative for the nine quarters of the projection horizon.

The Federal Reserve's projections of pre-tax net income under the alternative severe scenario imply negative net income at most of the 33 firms individually and for the firms as a group over the nine-quarter projection horizon. Projected pre-tax net income is an aggregate net loss of $198 billion over the projection horizon for the 33 firms. The projected pre–tax net income incorporates $440 billion in total provisions.

The ratio of pre-tax net income to average assets for each of the 33 firms ranges from −3.5 percent to 3.5 percent (see figure 19). Projected cumulative pre–tax net income for most of the firms is negative over the projection horizon.

Firms that are required to include AOCI in regulatory capital and those that opt in to including it are also affected by OCI (see table 12). OCI is driven by unrealized gains and losses on AFS securities in the supervisory stress test. The interest rate path and credit spreads assumed in the alternative severe scenario result in a loss of $3.1 billion of OCI over the nine quarters of the projection horizon for firms required to include AOCI in regulatory capital and those that opt in to including it.

Table 9. 33 participating firms* Projected stressed capital ratios, risk-weighted assets, losses, revenues, net income before taxes, and loan losses Federal Reserve estimates: Alternative severe scenario

 

Capital ratios, actual 2020:Q2 and projected 2020:Q3–2022:Q3

Projected loan losses, by type of loan, 2020:Q3–2022:Q3

Risk-weighted assets, actual 2020:Q2 and projected 2022:Q3

Projected losses, revenue, and net income before taxes through 2022:Q3

Table 10. Projected minimum common equity tier 1 capital ratio under the alternative severe scenario, 2020:Q3–2022:Q3 33 participating firms*

Table 11.A. Capital ratios, actual 2020:Q2 and projected 2020:Q3–2022:Q3 under the alternative severe scenario:Risk-based Category I, II, and III firms

Table 11.B. Capital ratios, actual 2020:Q2 and projected 2020:Q3–2022:Q3 under the alternative severe scenario: Risk-based Category IV firms*

Table 11.C. Capital ratios, actual 2020:Q2 and projected 2020:Q3–2022:Q3 under the alternative severe scenario: 33 participating firms

Table 12. Projected losses, revenue, and net income before taxes through 2022:Q3 under the alternative severe scenario: 33 participating firms*

Table 13. Projected loan losses by type of loan for 2020:Q3–2022:Q3 under the alternative severe scenario: 33 participating firms

Table 14. Projected loan losses by type of loan for 2020:Q3–2022:Q3 under the alternative severe scenario: 33 participating firms

 

References

 

 38. For firms that have adopted ASU 2016-13, the Federal Reserve incorporated its projection of expected credit losses on securities in the allowance for credit losses, in accordance with FASB, Financial Instruments–Credit Losses (Topic 326), FASB ASU 2016-13 (Norwalk, Conn.: FASB, June 2016). Prior to the adoption of ASU 2016-13, securities credit losses were realized through other-than-temporary impairment (OTTI). Return to text

 39. Additionally, losses are calculated based on the EAD, which includes both outstanding balances and any additional drawdown of the credit line that occurs prior to default, while loss rates are calculated as a percent of average outstanding balances over the projection horizon. Return to text

 40. These estimates are conditional on the severely adverse scenario and conservative assumptions. They are not a supervisory estimate of the firms' current or expected legal liability. Return to text

 41. As noted, credit card lending also tends to generate relatively high loss rates, so the higher PPNR rates at these firms do not necessarily indicate higher profitability. Return to text

 42. For firms electing the CECL transition provision in 12 C.F.R. pt. 217, subpt. G, the starting capital position used to project the path of regulatory capital includes eligible transitional amounts intended to mitigate the adverse effect of differences between allowances under CECL and the incurred loss methodology. Projected capital levels do not incorporate changes in the transitional amounts during the transition period, which has a duration of three or five years depending on the transition election. Return to text

 43. The Federal Reserve assumes that the allowance at the end of each quarter covers projected loan losses for four quarters into the future. The allowance at the end of the ninth quarter equals projected loan losses from quarters 10 through 13. Return to text

 44. For firms that have adopted ASU 2016-13, the Federal Reserve incorporated its projection of expected credit losses on securities in the allowance for credit losses, in accordance with FASB, Financial Instruments–Credit Losses (Topic 326), FASB ASU 2016-13 (Norwalk, Conn.: FASB, June 2016). Prior to the adoption of ASU 2016-13, securities credit losses were realized through OTTI. Return to text

 45. These estimates are conditional on the alternative severe scenario and on conservative assumptions. They are not a supervisory estimate of the firms' current or expected legal liability. Return to text

*Note: The Federal Reserve revised this report on June 24, 2021: the sentence "In the aggregate, the 33 firms are projected to generate $363 billion in PPNR cumulatively over the nine quarters of the projection horizon, equal to 2.0 percent of their combined average assets (see table 9)." has been revised to "In the aggregate, the 33 firms are projected to generate $363 billion in PPNR cumulatively over the nine quarters of the projection horizon, equal to 1.9 percent of their combined average assets (see table 9)."