The Fed Announces Updated Large Bank Capital Requirements
Five Quick Observations
The Federal Reserve Board this afternoon announced updated capital requirements for large banking organizations. This is an annual August announcement made by the Board following the confirmation of applicable stress capital buffers (SCBs) as determined by the annual supervisory stress test, the results of which were announced in June.
What follows are five quick observations about today’s announcement, focusing on U.S.-based firms.
A Number of the Very Largest Banks Face Increases in Capital Requirements
Increased Stress Capital Buffers
Of the eight U.S. global systemically important banking organizations, JPMorgan, Citigroup and Bank of America this year each saw significant increases in their SCBs. Effective October 1st JPM’s SCB will be 80 bps higher than currently, Bank of America’s SCB will be 90 bps higher than currently, and Citigroup’s SCB will be 100 bps higher than currently.
SCBs for the other U.S. G-SIBs stayed more or less flat,1 although of course in all of these cases the year-over-year comparison does not tell the whole story. For example, although Goldman Sachs was the only U.S. G-SIB to see an SCB decline, its SCB will remain the highest of the U.S. G-SIBs at 6.3%.2
Other U.S G-SIB stress capital buffers effective October 1, 2022: 5.8% (Morgan Stanley), 4% (Citi and JPM), 3.4% (Bank of America), 3.2 (Wells Fargo) and 2.5% (BNY Mellon and State Street).
Increased G-SIB Surcharges
The forthcoming SCBs are not the only capital increases facing certain U.S. G-SIBs in the near future. Under applicable U.S. capital rules, and unrelated to this summer’s stress tests, each U.S. G-SIB is required to calculate its systemic importance based on two methodologies and is then assigned an additional capital requirement (G-SIB surcharge) based on the result produced by the methodology which generates the higher requirement. This addition of a second methodology is different from how G-SIB surcharges are calculated in other jurisdictions, and it is this second methodology which generally drives U.S. G-SIB surcharges.3
The U.S. G-SIB surcharge rules operate with a lag, such that an increase in a banking organization’s G-SIB score in a given year is not reflected in the applicable surcharge until a few years down the line, provided that the increase in the G-SIB score persists.4 As a result, current U.S. G-SIB surcharges do not yet reflect the effects of the balance sheet and other growth experienced by these banking organizations in the aftermath of the COVID-19 pandemic.
They soon will, however.
For example, JPM disclosed in its 10-K earlier this year the following with respect to its G-SIB surcharge, which is currently 3.5%:
[A]n increase in the Firm’s GSIB surcharge to 4.0% … will be effective January 1, 2023. … The Firm’s estimated Method 2 surcharge calculated using data as of December 31, 2021 is 4.5%. Accordingly, based on the GSIB rule currently in effect, the Firm’s effective GSIB surcharge is expected to increase to 4.5% on January 1, 2024 unless the Firm’s Method 2 GSIB surcharge calculation based upon data as of December 31, 2022 is lower.
Recent regulatory filings indicate that Citigroup and Goldman Sachs are also preparing for increased G-SIB surcharges effective January 1, 2023, while Bank of America is expecting an increase effective January 1, 2024.
Citigroup: “[C]ommencing January 1, 2023, Citi’s GSIB surcharge will increase from 3.0% to 3.5%.”
Goldman Sachs: “Our G-SIB surcharge is 2.5% for 2022 and 3.0% for 2023. Based on financial data for 2021 and the six months ended June 2022, we are above the threshold for the 3.5% G-SIB surcharge. Based on the recent operating environment and our clients’ needs, we are currently targeting to be at the 3.0% G-SIB level as of December 2022 (which would be effective for our January 2024 G-SIB surcharge). We will continue to evaluate this target through the end of 2022.”
Bank of America: “The Corporation’s G-SIB surcharge, which is higher under Method 2, is expected to increase to 3.0 percent on January 1, 2024 unless its surcharge calculated as of December 31, 2022 is lower than 3.0 percent.”
For Slightly Smaller Banks, Mostly the Status Quo - But With a Few Notable Exceptions
Many of the regional banks which participated in this year’s stress test had their SCBs remain unchanged year-over-year, with most of these banking organizations continuing to be assigned SCBs at the SCB floor of 2.5%.5
A few regionals, however, saw increases in their SCBs, with the most notable increases for M&T Bank Corporation, which experienced the highest year-over-year change in SCB for any U.S. banking organization, going from 2.5% to 4.7%, and for Huntington Bancshares, which saw its SCB increase from 2.5% to 3.3%.
SCB Appeals Continue to Prove Difficult
Following the stress test results and its preliminary receipt of the unexpectedly high 3.3% SCB discussed above, Huntington filed a request for reconsideration. So too did Bank of America with respect to its 3.4% SCB.
This afternoon, the Federal Reserve Board announced that both appeals have been denied, and thus re-affirmed the respective 3.3% and 3.4% SCBs preliminarily assigned to these banking organizations following the supervisory stress tests in June.
Today’s announcement of the two rejected appeals brings banking organizations to a collective 0/8 in the SCB appeals process since its inception in 2020. In a post a few months ago (second item) I suggested - wrongly, as it turns out - that in light of the standard of review for these appeals, firms might be dissuaded from filing appeals in the future.
The Board’s letters denying the requests for reconsideration indicate that the bar a request for reconsideration must clear is quite high. Specifically, the Board considers “whether the request identified any errors in the firm’s stress test results and whether each stress test model identified in the firm’s request is operating as intended, within the bounds of the Board’s published policies.”
As a result, setting aside instances where the Board just got the math wrong, most appeals by large banking organizations are likely to have a tough time. The most frequently raised complaints about the stress tests are not about whether the models are operating as intended and consistent with the Board’s published policies, but instead are about features of the models and policies in the first place.
The Board has not yet made public its letters denying the appeals from Huntington and Bank of America. When it does so,6 it will be interesting to see whether Huntington or Bank of America offered any arguments different to those that had been tried in previous years.
Next Year Is an Optional Year for Certain Regional Banks, but Most Won’t Really Have a Decision to Make
Under the Federal Reserve Board’s stress test rules, firms subject to the comparatively least stringent enhanced prudential standards, referred to as Category IV firms, are required to participate in the supervisory stress test only every other year. Next year is an off-year, so Category IV firms will not be required to participate. The rules do allow Category IV firms to opt-in to the stress test in off-years if they so choose, however, and thus in theory firms in Category IV will have a decision to make.
I say in theory because for many Category IV firms the decision will not be difficult. The majority of U.S. firms in Category IV will have SCBs floored at 2.5% and thus from their perspective can only make themselves worse off if they participate. Therefore, they almost certainly will decline to do so.
After excluding firms at the 2.5% floor, three U.S. firms in Category IV remain: M&T at 4.7%, Citizens at 3.4% and Huntington at 3.3%.
Even here, however, for two of these firms there will not actually be a choice to make in 2023. The SCB assigned to M&T this year did not take into account M&T’s acquisition of People’s United Financial. Similarly, the SCB assigned to Citizens this year did not take into account its acquisition of Investors Bancorp. The Federal Reserve Board has determined that these were material acquisitions and, as such, notwithstanding the general rule that Category IV firms can opt-out of participating in odd-numbered years, the Board will use next year’s stress test to calculate updated SCBs for M&T and for Citizens.
That leaves Huntington. As of this writing, Huntington has not offered any public indications as to what it plans to do in 2023. Given that Huntington was unhappy enough with its newly assigned SCB to appeal, it is possible that Huntington will elect to participate next year. This might particularly be the case if Huntington’s discussions with the Board in the context of the appeal - although fruitless in the immediate term - helped it to understand better what was driving the heightened SCB, such that it believes the SCB could be reduced meaningfully in the future.
On the other hand, participating in the stress test is a costly and time-intensive process. Add in the speculation that the Board under its new Vice Chair for Supervision is going to make the stress tests meaningfully more difficult, combined with the fact that under the Board’s rules Huntington must make its decision by January 15, 2023 (i.e., before the Board announces the scenarios for the 2023 test), and maybe that argues in favor of sitting out 2023. Note also that, based on the limited history we have to go on, the fact that a firm was upset enough to file an appeal in a prior year is not necessarily a definitive sign that the firm will elect to participate in the stress test the next year when not required to do so.7
A Footnote
In the Federal Reserve Board’s stress test results announcement in June there was an interesting footnote concerning MUFG Americas Holdings Corporation, the U.S. intermediate holding company for the non-branch U.S. operations of MUFG.
MUFG Americas is a Category IV firm and thus was required to participate in this year’s stress test. Unlike for almost all the other firms that participated,8 the Federal Reserve Board in June did not announce detailed projections for MUFG Americas, instead stating in a footnote the following:
Due to certain accounting reclassifications by MUFG Americas related to a pending divestiture, the Board determined not to include MUFG Americas’ detailed projections in the 2022 stress test. The Board conducted an evaluation and determined that MUFG Americas has sufficient capital to absorb losses and continue to serve as a credit intermediary under a severe recession.
Similarly today, in its table setting out updated capital requirements the Board included a footnote saying the following:
Due to certain accounting reclassifications by MUFG Americas related to a pending divestiture, the Board determined not to include MUFG Americas’ final stress capital buffer requirement in this publication.
The pending divestiture in question is, of course, MUFG’s pending sale of Union Bank to U.S. Bancorp. 9
The SCB assigned to Goldman Sachs decreased by 10 bps, while Morgan Stanley and Wells Fargo saw their SCBs increase by the same amount. SCBs assigned to BNY Mellon and State Street remained at the 2.5% SCB floor.
There was a funny (so far as these things go) exchange on Goldman’s earnings call last month as to whether these facts taken together make Goldman a relative winner or a relative loser:
Mike Mayo: … On capital, I mean, you talked about a broader, diverse, resilient, more annuity-like business stream, but then we get the Fed stress test results and you guys finished in last place, albeit better than last year but still last place. So I guess the regulators don’t see it the same way that you do. …
David Solomon: … I mean I would highlight, while it’s a fair comment that our SCB is higher than everyone else’s because of our current business mix, if that’s what you refer to as last place, I would comment for the stress test we came in first place in that we were the only firm that actually saw their SCB decrease and many institutions saw their SCB increase very meaningfully. Now I don’t take that as a victory lap. We’re early in the journey of taking assets off our balance sheet…
A sympathetic law firm summary from 2015:
The first method (‘Method 1’), which is consistent with the Basel G-SIB Framework, is based on the aggregation of a BHC’s “systemic indicator scores” reflecting size, interconnectedness, cross-jurisdictional activity, substitutability, and complexity. … The second method (‘Method 2’), which is not part of the Basel G-SIB Framework, replaces the substitutability indicator with a measurement of the G-SIB’s reliance on short-term wholesale funding (‘STWF’). It is intentionally ‘calibrated to result in significantly higher surcharges’ than those calculated using Method 1—a further example of the Federal Reserve implementing more stringent, or ‘super equivalent,’ requirements than under certain internationally-agreed BCBS frameworks.
See 12 CFR 217.403(d)(1): “An increase in the GSIB surcharge of a global systemically important BHC will take effect … on January 1 of the year that is one full calendar year after the increased GSIB surcharge was calculated.”
To put this in practical terms, a firm takes its balance sheet at the end of Year 1 and must by the end of Year 2 use that Year 1 balance sheet to calculate its G-SIB score. Even if the G-SIB score would result in an increased surcharge, because the increased surcharge does not take effect until January 1 of a full calendar year after a G=SIB score calculation is performed, nothing changes in Year 3, and the increased surcharge takes effect starting January 1 of Year 4.
Decreased surcharges, on the other hand, take effect more quickly: “A decrease in the GSIB surcharge of a global systemically important BHC will take effect … on January 1 of the year immediately following the calendar year in which the decreased GSIB surcharge was calculated.” See 12 CFR 217.403(d)(2).
These nuances of the rule explain why there is some degree of hedging language in the excerpted disclosures quoted in the main text of the post - it is in theory possible that certain of these forthcoming increases will not come to pass.
Also notable were the results for Ally Financial and Discover, each of which this year decreased their SCBs to the 2.5% floor, compared to SCBs of 3.5% and 3.6%, respectively, in the prior supervisory stress test in which these firms participated.
The 2020 and 2021 appeal denial letters were eventually made public, sometimes with redactions, so I assume the Board will take the same approach here and make the letters public once the firms have in question have been offered an opportunity to propose redactions of any sensitive information.
In 2020, Citizens was preliminarily assigned an SCB of 3.4%. It immediately issued a press release taking issue with the “inaccurate result” produced by the Board’s models and stating that it would file a request for reconsideration. That request was denied, leaving the 3.4% SCB in place for 2020. Citizens then issued another press release saying that it was “disappointed with [the Board’s] conclusion” and that it looked forward to working with the Board to “improve the accuracy of [the Board’s] models.” Nonetheless, when in 2021 Citizens had a chance to opt-in to the stress test and take another shot at reducing its SCB, it declined to do so, choosing instead to stick with the 3.4% 2020 SCB.
Consistent with past years, the other firm for which detailed results were not disclosed was DWS USA, a subsidiary of Deutsche Bank holding the firm’s U.S. asset management business. See footnote 2 in the upper-left hand table here. (For the backstory as to why DWS USA exists and why it is required to participate in the stress tests despite being well below the relevant asset threshold on a standalone basis, see this letter.)
As described in MUFG’s 20-F: “The assets and liabilities related to the business of MUFG Union Bank that are expected to be transferred to U.S. Bancorp were reclassified as assets held for sale and liabilities held for sale and were included in Other assets and Other liabilities, respectively, as of March 31, 2022.”