I support strong capital requirements, and believe that placing limits on leverage is one of the most powerful tools bank regulators possess to promote a resilient banking system. However, I will vote against today’s proposal. I have concerns with the impact of excessive gold plating of international standards, I am skeptical of certain aspects of the underlying Basel standard, and I oppose unwinding the tailoring of the capital framework for large banks.
Following the 2008 financial crisis, and in response to a general consensus that the banking system was undercapitalized,1 U.S. regulators agreed to and implemented the Basel III reforms, which significantly increased the quantity and improved the quality of capital banks maintain to absorb losses. As a result, capital across the system grew substantially.2
The Basel III capital rule was a major piece of a vast post-crisis regulatory agenda, which comprised an extraordinary range of reforms at the largest banks, including standardized liquidity requirements, a long-term debt requirement, counterparty limits, margin requirements, clearing requirements, stress testing, enhanced governance standards, heightened risk management standards, enhanced reporting requirements, trading and activities restrictions, and CECL accounting for loan loss reserves, among many others.
In the late 2010s, when banking regulators reevaluated many of the post-crisis reforms, changes to the capital framework were limited and targeted, and regulators generally maintained the robust levels of capital in place.3 At the same time, for a number of years, banking agencies and other stakeholders have regularly and repeatedly commented on the strong capital levels at the largest banks and the resilience of the U.S. banking system.4
Meanwhile, lurking in the background, U.S. agencies agreed to another round of reforms to international capital standards now known as Basel III finalization, Basel III endgame, or, in some quarters, Basel IV.5 The goals of these agreements included, among other things, (1) limiting, however modestly, the ability of banks in other jurisdictions to use internal models to lower their capital requirements6 and (2) increasing the “risk sensitivity” of aspects of the standardized approach.7
Leading up to the final Basel agreement, international regulators emphasized that the new reforms were not intended to significantly increase capital requirements.8 In that vein, when the U.S. agencies began efforts to implement the Basel agreement several years ago, one of the goals was a loose concept referred to as “capital neutrality.”9 This goal reflected a general sense that the prevailing post-crisis level of capital in the system did a reasonably good job balancing the sometimes-competing objectives of safety and soundness, financial stability, and economic growth, and that large increases or decreases in required capital were generally undesirable.10
This goal also reflected an awareness that banks in the U.S. have higher levels of capital than those in many other jurisdictions, which is primarily driven by the ability of non-U.S. banks to use internal models to meaningfully reduce the amount of capital needed to meet their requirements.11 In addition, regulatory buffers in the U.S., including the stress capital buffer (SCB), gold-plated G-SIB surcharge, and gold-plated enhanced supplementary leverage ratio, tend to result in more capital than regulatory buffers in other jurisdictions, which include Pillar 2 add-ons12 and, in some cases, a positive countercyclical capital buffer.13
Today’s proposal rejects the notion of capital neutrality and takes a starkly different path, “gold plating” the new Basel standard in a number of ways and dramatically increasing capital requirements for banks with certain business models. To start, large banks in the U.S. would be subject to multiple “stacks,” and stacks within stacks,14 in determining their capital requirements. Large banks would continue to be subject to a “standardized” approach (or stack), but this would be different from what the Basel Agreement calls the “standardised” approach. Under the proposal, the standardized approach is effectively the risk-based capital rule that applies to most other U.S. banks,15 but in practice it is meaningfully more conservative given (1) it includes the new market risk framework – which would apply to less than ten out of 4,629 banks with less than $100 billion in assets – and (2) it includes the full slate of buffers,16 including the SCB and G-SIB surcharge, which do not apply to any banks below $100 billion in assets.
Large banks would also be subject to the so-called “expanded risk-based” approach, which is the U.S. version of the new, Basel standardized approach.17 Within this stack, a number of items are gold-plated from the Basel standard, including: within the credit risk framework, the risk weights for residential mortgages, retail exposures, exposures to banks and credit unions, and exposures to small businesses; within the operational risk framework, the floor for the Internal Loss Multiplier (ILM); and within the market risk framework, the requirement that banks use the standardized approach, rather than a modeled approach, for default risk charges. The U.S. is also declining to make several modifications that European jurisdictions have proposed,18 each of which further reinforces the relative conservatism of the U.S. approach.
Large U.S. banks would be required to determine their capital requirements under each of these stacks and maintain capital in excess of whichever stack produces the most stringent requirement. Taken together, the new framework would subject large banks to (1) a gold-plated “expanded risk-based” approach, which is meaningfully more conservative than the single standardized stack implemented in other jurisdictions, and (2) an enhanced version of the current U.S. standardized approach, which is both more conservative than the generally applicable standard in the U.S. and not applicable in any other jurisdiction … all while non-U.S. banks are still able to drive down their capital requirements through the use of internal models.19
Altogether, when also considering – among other things – the impact of (1) the new operational risk charge, which is entirely new and additive to what will often be the binding stack, and (2) the revised market risk charge, which is expected to increase market risk-weighted assets by more than double for the large banks most heavily engaged in capital markets activities, the proposal would have a substantial impact on how banks allocate capital. The result will be some combination of higher prices for consumers, less availability of products and services, migration of U.S. activities out of the regulated banking sector, migration of international activities out of U.S. banks, and more fragile financial markets.
* * *
As noted above, one of the objectives of the Basel agreement is to increase risk sensitivity. But risk sensitivity is a double edged sword, with deep pitfalls on each edge. On the one hand, too little risk sensitivity incentivizes risk taking, as banks can earn higher returns on capital by engaging in higher-risk activities. On the other hand, setting increasingly risk-sensitive standardized risk weights always involves choosing winners and losers, promoting homogeneity across the industry, and adding substantial complexity to the capital framework, along with the inevitability of misjudging and mispricing risk.
Nonetheless, there are some areas where additional risk sensitivity is justified. For example, in the proposal, incorporating loan-to-value (LTV) ratios into the risk weight treatment for residential mortgages is understandable. Market participants have been analyzing mortgage credit risk for many decades, spanning numerous business cycles, yielding a great deal of data and experience demonstrating the relationship between LTV and credit risk.
By contrast, operational risk is an amorphous concept, a catch-all category that encompasses a large and highly variable set of risks, ranging from fraud to bad behavior to overzealous enforcement agencies to cyber attacks to asteroids.
Under the proposal, the standardized operational risk charge would be calculated based on two factors. First, the business indicator (BI) component is in effect a proxy for size,20 but it is a very different proxy for size than that which regulators use for any other purpose, and one that is especially punitive for fee-income businesses. I understand part of the rationale is to capture certain off-balance sheet activities that present operational risk, but there are all sorts of operational risks not reflected on a bank’s balance sheet that the proposal does not capture – such as the quality and sophistication of a bank’s IT systems,21 the geographical distribution of its bank branches, or the nature of its personnel, to name just a few.
Second, the ILM component would adjust the operational risk charge based on operational losses over the prior ten years. The proposal states definitively, “Higher historical operational losses are associated with higher future operational risk exposure.”22 I suspect the proposal overstates our ability to understand and predict operational risk, and I wonder whether the operational risks for which we care most about capitalizing are those least likely to be captured by the ILM.
Overall, I understand the merits of having some consideration for operational risk in the capital framework, but am skeptical that a risk-sensitive operational risk charge is actually sensitive to the underlying risk.
* * *
Next, I’ll turn to the proposal’s treatment of tailoring. In 2018, Congress passed S. 2155, which, among other things, (1) raised the threshold for which firms were subject to enhanced prudential standards (EPS) from $50 billion to $250 billion in assets, while authorizing the Federal Reserve to apply enhanced standards to firms between $100 billion and $250 billion in assets, and (2) mandated that regulators tailor application of EPS for firms in scope. In 2019, the Federal Reserve, in part with the Office of the Comptroller of the Currency (OCC) and the FDIC, issued the tailoring rules, which implemented both of these mandates by establishing quantitative metrics for the application of standards to large banks.23 Categories I, II, and III included all firms with $250 billion or more in assets, plus other institutions that met specific criteria, while Category IV included the remainder of institutions with $100 billion or more in assets, who were in effect generally excluded from EPS.
Today’s proposal repudiates these concepts, by “aligning” the capital rules for all banks with $100 billion or more in assets.24 This “alignment” includes application of the supplementary leverage ratio, the countercyclical capital buffer, and the standardized approach for counterparty credit risk, and the treatment of TLAC holdings, mortgage servicing rights, certain deferred tax assets, non-significant investments in the capital of unconsolidated financial institutions, minority interests in subsidiaries of banking organizations, and all other comprehensive income (AOCI). None of that is at all related to the new Basel agreement. In addition, the proposal would further apply the full scope of the new Basel standard to all institutions with $100 billion or more in assets. This includes applying the entire market risk framework to regional banks with de minimis trading activities, who have never previously calculated market risk capital, and who will now need to build out market risk compliance operations and systems despite market risk having virtually no impact on the institutions’ actual capital levels.
For purposes of the capital rules, the proposal effectively collapses Categories II, III, and IV into one category. The proposal undoes almost all of the tailoring of the capital framework for large banks, and is a repudiation of the intent and spirit of S. 2155. It is further a troubling sign for future policymaking, a signal that regulators intend to treat all large banks alike, in defiance of Congressional directives and in contradiction to the objective of a diverse banking sector with banks of varying sizes, niches, and business models.
Of course, I recognize that three Category IV banks failed earlier this year, at a substantial cost to the Deposit Insurance Fund, and, in one case, in quite spectacular fashion. I think there are lessons for us to consider from these failures around topics like interest rate risk, deposit insurance, contingency funding, and large bank resolution.25 But I also think it’s important that our reactions be thoughtful and targeted, and that we avoid the temptation to overregulate all Category IV banks in response to the unique circumstances of the spring failures. It’s worth noting that implementation of the new Basel agreement was expected to result in no increase in required capital at any of the three banks that failed,26 but would result in major increases at several other Category IV banks.
Finally, the proposal would require large banks to incorporate unrealized losses and gains on available for sale (AFS) securities in their capital requirements. This is a complicated topic,27 and I do not object to seeking comment on it. However, we should appreciate that while the proposed change may have reduced the likelihood of the specific catalyst that triggered the run on Silicon Valley Bank (SVB), it would have done little to address the underlying fundamental issue.28 Of SVB’s bond losses, 85% were on its held to maturity (HTM) portfolio and thus unaffected by the proposal – a percentage that may have been higher had the bank’s capital requirements incentivized such a shift. And unrealized losses on securities were not a major driver at Signature or First Republic.29 The fact that this change is only proposed for firms with $100 billion or more in assets, and only for AFS securities, I think reflects the challenges with using capital rules to address these issues.
I plan to vote against the proposal. I would like to recognize and extend my appreciation to the staff for their all hard work going back many years. I understand this can be a grueling process, and want to thank you for all of the long hours you put in.
1 See, e.g., Office of the Comptroller of the Currency, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, Notice of proposed rulemaking: Regulatory Capital Rules: Regulatory Capital, Implementation of Basel III, Minimum Regulatory Capital Ratios, Capital Adequacy, Transition Provisions, and Prompt Corrective Action, 77 Fed. Reg. 52792, 52800 (August 30, 2012) (“The recent financial crisis demonstrated that the amount of high quality capital held by banks globally was insufficient to absorb losses during that period. In addition, some noncommon stock capital instruments included in tier 1 capital did not absorb losses to the extent previously expected.”).
2 In the years after the crisis, regulators increased the Tier 1 capital ratio, imposed a new common equity tier 1 ratio (CET1), adopted a new supplementary leverage ratio, increased required capital for securitizations and derivatives exposures, and added a series of capital buffers on top of minimum capital requirements, among other changes.
3 See, e.g., Jelena McWilliams, Remarks at the Bipartisan Policy Center (February 3, 2022) (“We have continued to prioritize the importance of strong capital levels, particularly at our nation’s largest banks. The lead depository institutions of the eight U.S. G-SIBs grew their capital levels over the last several years despite stressful economic conditions, owing to rigorous capital requirements that were in place entering this period. The weighted average CET1 capital ratio at these institutions increased from 12.9% in 3Q2018 to 14.5% as of 3Q2021. Capital adequacy remains robust across the broader industry as well, including in the community banking sector.”).
4 See, e.g., Board of Governors of the Federal Reserve, Monetary Policy Report, p. 2 (June 16, 2023) (“Despite concerns about profitability at some banks, the banking system remains sound and resilient… [T]he broader banking system maintained substantial loss-absorbing capacity and ample liquidity.”); Board of Governors of the Federal Reserve, Supervision and Regulation Report, p. 1 (May 2023) (“The U.S. banking system is sound and resilient, with strong capital and liquidity.”); Board of Governors of the Federal Reserve, Financial Stability Report, p. 2 (May 2023) (“[T]he broad banking system remained sound and resilient. For the banking system as a whole, aggregate bank capital levels were ample.”).
5 See Basel Committee on Banking Supervision, Basel III: Finalising post-crisis reforms (December 7, 2017) (“Final Basel III reforms”); Basel Committee on Banking Supervision, Minimum capital requirements for market risk (January 14, 2019); Basel Committee on Banking Supervision, Minimum haircut floors for securities financing transactions (January 26, 2021).
6 Basel Committee on Banking Supervision, High Level Summary of Basel III Reforms, p. 11 (December 2017) (“Consistent with the original floor, the revised floor places a limit on the regulatory capital benefits that a bank using internal models can derive relative to the standardised approaches. In effect, the output floor provides a risk-based backstop that limits the extent to which banks can lower their capital requirements relative to the standardised approaches.”).
7 Final Basel III reforms, supra note 5, p. 1 (“The revisions to the regulatory framework set out in this document will help restore credibility in the calculation of RWAs by: (i) enhancing the robustness and risk sensitivity of the standardised approaches for credit risk and operational risk, which will facilitate the comparability of banks’ capital ratios; (ii) constraining the use of internally-modelled approaches; and (iii) complementing the risk-weighted capital ratio with a finalised leverage ratio and a revised and robust capital floor.”).
8 See, e.g., id., at p. 1 (“[T]he [Basel] Committee conducted a comprehensive and rigorous assessment of the impact of these revisions on the banking system and the wider macro economy. As a result of this assessment, the Committee focused on not significantly increasing overall capital requirements.”); Global Heads of Supervision (GHOS), Media Conference (December 7, 2017) (GHOS Chair Mario Draghi stating, “This will reduce excessive variability of risk-weighted assets without significantly increasing capital in the aggregate of course”, and “The focus of the exercise was not to increase capital. As a matter of fact, the GHOS almost a year ago endorsed this review by the Basel Committee, provided it wouldn’t create a significant capital increase in the aggregate of the banking system.”); Financial Stability Report Question and Answer document, Bank of England, pp. 16-17 (November 30, 2016) (then-Governor of the Bank of England and Chairman of the Financial Stability Board Mark Carney stating, “[W]e have said, and we said at the equivalent press conference a year ago actually made this point and then we’ve made it subsequently and had agreement of all the members of the G20 and the steering committee of the Basel Group, that there would be no significant increase in overall capital requirements as a consequence of finishing this process of Basel III.”).
9 See, e.g., Randal K. Quarles, Between the Hither and Farther Shore, Thoughts on Unfinished Business (December 2, 2021) (“A major issue that we are grappling with is how to implement these reforms, which reduce the role of bank internal models on bank capital requirements, while maintaining the overall level of aggregate capital requirements… What policymakers will need to do as they implement the Basel III reforms is determine whether adjustments to other parts of the capital framework are necessary to ensure that we do not unduly increase the level of required capital in the system.”).
10 This was also consistent with policy that generally maintained the robust levels of capital in place. See, e.g., supra note 3.
11 The 72.5 percent output floor in the final Basel standard will require non-U.S. banks to raise capital levels to at least 72.5 percent of the new international standardized minimum – but still well below minimum standards for U.S. banks.
14 For example, under the proposal, banking organizations that receive approval to use internal models to calculate market risk-weighted assets would be required to determine market risk capital requirements using both internal models-based and standardized approaches. See infra note 20, at section III.H.1.b.
15 The remaining U.S. banks are those that have elected to use the Community Bank Leverage Ratio.
16 Mechanically, the buffers will be calculated based on whichever risk-based stack is binding, but in effect, this is equivalent to all buffers applying to both risk-based stacks.
17 The proposal describes the expanded risk-based approach as replacing the internal models-based approach for Advanced Approaches banks. However, under the proposal, the expanded risk-based approach is expected to often be the binding stack, while the internal models-based approach today is rarely binding and, therefore, generally irrelevant to capital levels.
18 See, e.g., Bank of England, Implementation of the Basel 3.1 standards, Consultative Paper, note 3 (November 2022) (“[T]he [Prudential Regulatory Authority] is not consulting in this [Consultative Paper] on the implementation of minimum haircut floors for securities financing transactions (SFTs) in the capital framework – one of two approaches envisaged in the [Financial Stability Board]’s report Regulatory framework for haircuts on non-centrally cleared securities financing transactions.”); id., at section 3.99 (“A corporate entity would not need to have securities outstanding on a recognised exchange to be assessed as [investment grade].”).
19 Earlier in this process, another goal was to simplify capital requirements; it would be hard to argue that goal has been met. See, e.g., McWilliams Speech, supra note 3 (“I am hopeful that revisions to the capital framework will reduce complexity and rationalize the many calculations necessary to determine capital adequacy.”).
20 See Office of the Comptroller of the Currency, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, Notice of proposed rulemaking: Regulatory capital rule: Amendments applicable to large banking organizations and to banking organizations with significant trading activity (“Proposal”), section III.F.2 (“The business indicator would serve as a proxy for a banking organization’s business volume and would be based on inputs compiled from a banking organization’s financial statements.”).
21 Relatedly, industry surveys often find that the #1 operational risk concern is cyber risk, a risk to which the framework is unlikely to be particularly sensitive. See, e.g., Risk.net staff, Top 10 Operational Risks for 2023 (March 8, 2023) (“Cyber risk – in one form or another – has dominated the top 10 op risk concerns since the survey’s inception nearly 15 years ago.”).
22 See Proposal, supra note 20, at section III.F.3.
23 The metrics categorized banks based on size, complexity, and risk profile, and the agencies have used the tailoring categories for application of the capital and liquidity frameworks, among other rules. See Board of Governors of the Federal Reserve System, Final Rule: Prudential Standards for Large Bank Holding Companies, Savings and Loan Holding Companies, and Foreign Banking Organizations, 84 Fed. Reg. 59032 (November 1, 2019); Office of the Comptroller of the Currency, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, Final Rule: Changes to applicability thresholds for regulatory capital and liquidity requirements, 84 Fed. Reg. 59230 (November 1, 2019).
24 See, e.g., Proposal, supra note 20, at section I.a. (“For banking organizations subject to Category III or IV capital standards, the proposal would align the calculation of regulatory capital – the numerator of the regulatory capital ratios – with the calculation for banking organizations subject to Category I or II capital standards, providing the same approach for all large banking organizations.”).
26 This analysis estimated the impact of the U.S. implementation of “Basel III endgame,” including U.S. gold-plating and multiple stacks, but does not include the proposed change to the treatment of AOCI, which was not part of the new Basel agreement and is discussed below.
27 See Hill speech, supra note 25 (“[O]ne much-discussed way to try to address this problem would be to require banks to hold capital against some – or all – unrealized losses on their bond investments. It is possible that moving aggressively in this direction would have reduced the likelihood of SVB’s failure, as it may have forced the bank to address its core problem sooner: either by raising more capital or by reducing the maturity of its assets. But these types of proposals also have well-known downsides, including, for example, (1) the tendency for market prices to exaggerate fluctuations in value during times of stress, and (2) the incongruence of banks’ capital requirements being driven by changes in the market value of securities, while ignoring changes in the value of loans.”).
28 If SVB had been required to hold capital against the unrealized losses on its AFS portfolio, the bank may have been less likely to announce a capital raise at the same time that it sold its AFS portfolio, which was the event that sparked the run. To argue that this may have prevented the bank’s failure perhaps raises a deeper question of whether the run at SVB was more about a poorly timed announcement or the fundamental hole in its balance sheet. For reference, see Michael S. Barr, Holistic Capital Review (“Realizing the losses from these securities, without adequate capital to protect from those losses, was an important part of the set of events that triggered the run on Silicon Valley Bank (SVB). If the bank had already been required to include those losses in its reported capital, it is less likely that the market and depositors would have reacted the same way.”); Martin Gruenberg, Remarks on the Basel III Endgame (“Had the unrealized losses on available for sale securities on the balance sheet of SVB, that were realized once sold, been required to be recognized in capital, as the Basel III framework would do, it might have averted the loss of market confidence and the liquidity run.”).
29 As of year-end 2022, SVB’s total mark to market bond losses were $17.7 billion, compared to $3.2 billion at Signature and $5.2 billion at First Republic. Of First Republic’s bond losses, 90% were on its HTM portfolio, while 76% of Signature’s were on its AFS portfolio; the latter figure presumably would have been lower were the proposal in effect.