The Fed's Report on the Failure of Silicon Valley Bank

This blog post sets out, in no particular order and with no particular claims about their importance, things I highlighted on a first read through of the Review of the Federal Reserve’s Supervision and Regulation of Silicon Valley Bank (the Review), as released by the Board this morning.¹

The FDIC also today released a report on the failure of Signature Bank. That report will be covered in a subsequent post.

Diminished Supervisors

The most interesting conclusions reached by the Review are also the most nebulous.

In his cover letter accompanying the Review, Vice Chair for Supervision Barr includes as one of his four key takeaways that “a shift in the stance of supervisory policy impeded effective supervision by reducing standards, increasing complexity, and promoting a less assertive supervisory approach.”

On this theme, the Executive Summary to the report includes a paragraph saying:

Over the same period, under the direction of the [former] Vice Chair for Supervision, supervisory practices shifted. In the interviews for this report, staff repeatedly mentioned changes in expectations and practices, including pressure to reduce burden on firms, meet a higher burden of proof for a supervisory conclusion, and demonstrate due process when considering supervisory actions. There was no formal or specific policy that required this, but staff felt a shift in culture and expectations from internal discussions and observed behavior that changed how supervision was executed. As a result, staff approached supervisory messages, particularly supervisory findings and enforcement actions, with a need to accumulate more evidence than in the past, which contributed to delays and in some cases led staff not to take action.

In a previous post on SVB, I wrote about this explanation and voiced a bit of skepticism.

Specifically, [Federal Reserve Board Governor Daniel] Tarullo wonders whether supervisors may have felt constrained by “instructions they were operating under." He asks, “Did the supervisors feel inhibited?”

I do not especially love this as an explanation because it feels frustratingly unfalsifiable. In particular, I am not sure it is possible to fairly evaluate whether supervisors did not do the right thing (1) because they simply erred or (2) because they wanted to do the right thing but felt constrained or inhibited by those above them.

On the other hand, I think Tarullo has a point that supervision is indeed in some sense about, for lack of a better word, vibes. So while I would be skeptical about this being the primary reason for the supervisory failures, it does seem like a legitimate topic for a portion of the Board’s inquiry.

Based on the Review, it looks like the Board sought to get at this "how do you prove it" issue in part by conducting a number of interviews with examiners, supervisors and other members of the staff. The interview transcripts themselves have not been provided,² so as an outsider I still sort of struggle with the but-for causation question. For instance, how should we view this episode described in a single sentence in the Review?

[S]taff informed SVBFG about a forthcoming MOU around information technology in 2021, but staff subsequently dropped the matter because they felt it would not be pursued by policymakers at that time.

What made the staff feel this way? What did they do to press their point? How did “policymakers” respond?

To be super clear, this is not me saying that the staff’s feeling was illegitimate — I have no idea. But I just would have liked the Review to have been a bit more explicit about the basis for the staff having this feeling.

The other thing I struggle with re: the disempowered supervisors thesis is the question of, “Compared to what?” Accepting as true that supervision in 2017-2021 was a little too relaxed, is it really the case that pre-2017 supervision was such that issues like those that brought down SVB would have been handled materially differently? Restrictions on confidential supervisory information make it difficult to say for sure, but the fact that during this pre-2017 time period a number of large banks were repeatedly failing to address supervisory issues, ultimately leading to significant public enforcement actions later in the decade, strikes me as at least partial evidence against that conclusion.

With those uncertainties expressed, for the remainder of this post, I will assume the validity of the disempowered supervisors thesis. I do this because, even taking that thesis as true, the Review makes clear that there were also other issues at play in the way SVB was supervised that cannot be traced solely to any late 2010s supervisory vibe shift.

Too much searching for consensus

The Review concludes that, overall, the Board’s supervisory approach was “too deliberative and focused on the continued accumulation of supporting evidence in a consensus-driven environment.”

The root causes for this are “difficult to ascertain,” but include “a focus on consensus-building and a perceived need to form ironclad assessments about what had already gone wrong and less on judgments with a more open mind about what could go wrong.”

Governance issues relating to the split of authority between the Board and its Reserve Banks may also play a role.

For example, the Board has delegated to the Reserve Banks supervisory authority for firms like SVBFG, including the authority to issue supervisory ratings, but in practice, Reserve Bank supervisors typically seek approval from or consensus with Board staff before making a rating change.

Similarly, this sort of need for coordination and desire to reach consensus, including with SVB’s state regulator, led to delays in taking enforcement action.

[I]t took more than seven months to develop an informal enforcement action, known as a memorandum of understanding (MOU), for SVBFG and SVB to address the underlying risks related to “oversight by their respective boards of directors and senior management and the Firm’s risk-management program, information technology program, liquidity risk management program, third-party risk-management program, and internal audit program.” SVBFG failed before the MOU was delivered.

The key point here, which the Board acknowledges toward the end of the Review, is that these concerns are not new.

Following the Global Financial Crisis in 2008 and 2009, the Federal Reserve Board conducted an evaluation of how it carries out its regulatory and supervisory responsibilities. […] Similarly, the Federal Reserve Bank of New York (FRBNY) commissioned an external review to draw on lessons learned from the Global Financial Crisis and make recommendations to the FRBNY. The non-final report, Report on Systemic Risk and Bank Supervision, focused on systemic risk issues but also had relevant insights for bank supervision that link to the SVBFG experience.

These relevant insights include “an excessive risk aversion and deference from supervisors, particularly during profitable periods” and “delay from a consensus-driven culture that smooths over complex issues.”

Or take the 2011 OIG review of failed state member banks between 2009 and 2011 that is described in today’s Review:

[T]he OIG noted that many “examiners identified key safety and soundness risks, but did not take sufficient supervisory action in a timely manner to compel the Boards of Directors and management to mitigate those risks. In many instances, examiners eventually concluded that a supervisory action was necessary, but that conclusion came too late to reverse the bank’s deteriorating condition.

Similarly, in the LISCC context the Board in November 2015 released a report that observed:

To evaluate whether adequate methods are in place for decision-makers to be aware of material matters that required reconciliation of divergent views related to the supervision of those firms, interviews were conducted with 122 current and former members of the dedicated supervisory teams. The review found that more than ninety-five percent of interviewees felt empowered to raise supervisory concerns and express divergent views. While staff expressed satisfaction in being able to raise divergent views, the review team noted differences among the dedicated supervisory teams in how staff were able to raise such views and to whom they could raise them to.

The Board’s report thus recommend that the LISCC OC:

Establish a requirement that Reserve Banks develop formal channels for supervisory staff to raise divergent views outside their immediate chain of command and implement a defined method for all Reserve Bank supervisory staff to raise supervisory concerns directly to Board staff.

SVB was not in the LISCC portfolio, so this recommendation is not directly applicable to its supervision, but even so this raises the question of what channels for dissent were made available to Reserve Bank staff in this case. Of course, one might retort that it makes no difference if an examiner can go around her immediate chain of command to raise concerns with the Board staff if the Board staff itself is operating under (real or perceived) pressure from above them.

I suppose there are two ways to interpret this discussion in the Review of prior findings of supervisory failings. One is that the Board and its Reserve Banks recognized these issues, took steps to solve them, and then saw their progress unwound by unnamed forces starting late in the 2010s, helping to lead to the failures described in the Review. Maybe so. But I wonder if the better way to read the conclusion to the Review is as saying that these issues were never fully addressed in the first place.

Limited scope reviews and reluctance to change recent ratings

As part of its case that supervision of SVB was too lax, the Review notes that the company was “subject to limited-scope liquidity reviews as part of the guidelines for smaller firms, despite its significant asset growth and idiosyncratic business model. A more thorough evaluation prior to joining the LFBO portfolio would have been beneficial, given the lag since the last in-depth examination and the heightened standards for a firm in the LFBO portfolio.”

Similarly, with respect to the company’s supervisory rating, although a new supervisory team “observed weaknesses in governance and risk management late in 2021, they were reluctant to issue a downgrade within seven months of the issuance of the prior rating without doing more examination work to support a change in view and related action.”

Both of these observations are intended to support an argument that banks like SVB should be subject to heightened supervision and, more to the point, should be subject to that supervision without unduly prolonged transition periods. I agree with that, but I also think it’s worth stepping back when examining these two observations.

It is true that SVB was subject to limited-scope liquidity reviews as part of the guidelines for smaller firms. But this is in part because, as the Review explains, “[i]n terms of liquidity, SVBFG was rated “Strong-1.”³ In other words, the comparatively light-touch supervision of SVB was not just a function of its size, but was also a function of supervisory mistakes by the Board and Reserve Bank in assessing SVB's liquidity risk profile in the first place.

Similarly, I have no argument with the point that perhaps supervisors are too anchored to, and too reluctant to change, recently issued ratings. But even if that is so, this was not the root cause of the issue. Even if politics or senior Board figures made it unduly difficult to change the supervisory rating once SVB was in the LFBO portfolio, the bigger issue is the too-forgiving supervisory rating that was handed out to SVB as an RBO in the first place.

Recommended Changes

The Review endorses or previews various changes to regulation or to supervisory approaches, while also saying that others should be considered.

Basel III, stress testing and, long-term debt proposals coming soon

In his cover letter, VCS Barr says that the Board intends to “seek comment … soon” on the following proposals: “implementation of the Basel III endgame rules; the use of multiple scenarios in stress testing; and a long-term debt rule to improve the resiliency and resolvability of large banks.” No surprises here, as all of this has been recently previewed by VCS Barr or Board staff.

Probably just me, but I can’t help but reminded here of the SLR proposals that the Board said in March 2021 that it would “soon be inviting public comment on.” I get the sense that “soon” in this context implies more decisive action than that 2021 statement did.

Changes to the liquidity rules and the AOCI opt-out

The VCS Barr cover letter also mentions other changes that would come over a longer time horizon. I am not sure whether to read anything into the “should require” language here re: the AOCI filter vs. the slightly less committal “re-evaluate” or “consider” language on the LCR/NSFR.

In addition, we are also going to evaluate how we supervise and regulate liquidity risk, starting with the risks of uninsured deposits. Liquidity requirements and models used by both banks and supervisors should better capture the liquidity risk of a firm’s uninsured deposit base. For instance, we should re-evaluate the stability of uninsured deposits and the treatment of held to maturity securities in our standardized liquidity rules and in a firm’s internal liquidity stress tests. We should also consider applying standardized liquidity requirements to a broader set of firms. Any adjustments to our liquidity rules would, of course, go through normal notice and comment rulemaking and have appropriate transition rules, and thus would not be effective for several years.

With respect to capital, we are going to evaluate how to improve our capital requirements in light of lessons learned from SVB. For instance, we should require a broader set of firms to take into account unrealized gains or losses on available-for-sale securities, so that a firm’s capital requirements are better aligned with its financial positions and risk.

Pillar 2(??)

Perhaps more interesting still is the sentence in the cover letter, never elaborated upon in Review itself, saying that:

[T]he Federal Reserve generally does not require additional capital or liquidity beyond regulatory requirements for a firm with inadequate capital planning, liquidity risk management, or governance and controls. We need to change that in appropriate cases.

I am maybe out over my skis here but the first thing I thought of when reading this was something like the Pillar 2 process under the Basel rules, which has never really been applied in the same way in the United States.

Changes intended to address interest rate risk

The Review also acknowledges possible issues with the Board’s current approach to interest rate risk, saying that one of the reasons supervisors were impeded in their assessment of IRR is a “lack of IRR regulations, and the high-level nature of existing guidance (SR letter 10-1).” The conclusions of the Review do not put forward any specific recommendations on this topic, saying instead that “[s]upervisors could reconsider how to best reflect interest rate risk in regulatory capital assessments.”

Incentive compensation

The Review concludes that:

The incentive compensation arrangements and practices at SVBFG encouraged excessive risk taking to maximize short-term financial metrics. SVBFG’s compensation practices also did not adequately reflect longer-term performance, nonfinancial risks, or unaddressed audit or supervisory issues. Nor did they include sufficient opportunities for SVBFG’s internal control functions to provide feedback or challenge.

It goes on to note that “[s]tronger or more specific supervisory guidance or rules on incentive compensation for firms of SVBFG’s size, complexity, and risk profile—or more rigorous enforcement of existing guidance and rules—may have mitigated these risks.”⁴

The Review leaves it ambiguous, however, as to what the Board thinks is the best step forward. It says, “the Board, together with five other federal financial regulatory agencies, issued proposals in 2011 and 2016 to implement the incentive compensation provisions in section 956 of the Dodd-Frank Act. An implementing rule, however, has not yet been finalized.” The Board does not indicate whether it believes the 2016 proposal would have been sufficient, or whether it now believes something stronger is needed.

Complexity of the regulatory structure

The Review concludes that a root cause of supervisory failure here was a “a complex oversight program that involves multiple categories, triggers, phase-in periods, rule sets, runways, and supervisory expectations.” Although “undoubtedly warranted in parts, [this] is also an impediment to both firms and their supervisors as they navigate through a challenging rule set with discrete cliff effects.”

In light of that, the Review suggests various ways to make oversight and the tailoring framework more efficient and effective. Suggestions include “greater clarity on portfolio expectations, well-defined internal governance over ratings, [and] an explicit supervisory plan for firms transitioning between portfolios.”

The Review also says that “reduced complexity of the regulatory structure” could be beneficial as it would allow for the shifting of some bandwidth “away from the supervisory process and more toward understanding and effectively managing the fundamental risk itself.”

I was not sure how to interpret this. Is this about reducing the complexity of how the Board decides which regulations apply to which firms, or is this the Board saying that the complexity of the regulations themselves should be reduced? If the latter, I think almost everyone would agree with that. But the trouble is they would disagree violently on the details.

Partial Answers to Previous Questions and Other Points of Interest

The Review addresses a number of issues that had been previously highlighted on this blog or elsewhere as things that could be interesting for the report to cover.

Internal Liquidity Stress Test

In a previous post, I had wondered how SVB fared on its internal liquidity stress test, a requirement that — unlike the LCR and NSFR — applies to all firms with $100 billion or more in total consolidated assets, and thus (unlike the LCR and NSFR) a requirement to which SVB was subject before its failure.

As I hope is clear, I am not asserting that these requirements eliminate the need for LCR or NSFR requirements. I do think it is interesting context, though, and as policymakers sift through the aftermath of these events I hope we learn more about what SVB's liquidity stress tests looked like in the quarters leading up to last week’s events, whether they indicated any major signs of risk and, if so, how management and regulators responded.

The Review supplies an answer, and the answer is that these tests did not go well!

SVBFG showed foundational weaknesses in its liquidity risk management, including both its liquidity position and its ability to manage risk through its internal liquidity stress tests (ILST), limits, and contingency funding plans (CFP). For example, beginning in July 2022 when SVBFG first became subject to enhanced prudential standards (EPS) under Regulation YY as a consequence of exceeding the $100 billion threshold, SVBFG repeatedly failed its own ILST. Management responded by increasing funding capacity, but the funding capacity actions were not rapidly undertaken or fully executed by March 2023. Management also switched to using less conservative stress testing assumptions, which masked some of these risks.

Discount window issues and other troubles finding sources of liquidity

As noted in the above quote, management knew (vaguely and reluctantly, but still) that SVB could face liquidity troubles. According to the Review, the company tried to address this, but in hindsight (or even in foresight) probably could have tried harder.

SVBFG’s ILST shortfall remediation plan from July 2022 cited the need to expand capacity and options for repo funding, including increased bilateral relationships, FICC direct membership, tri-party, and the Federal Reserve’s Standing Repurchase Agreement facility, among other sources. These efforts were not complete by March 2023. […]

SVBFG was not able to monetize (immediately raise funds against) its investment securities. SVBFG had not arranged for enough access to repo funding and had not signed up for the Federal Reserve’s Standing Repurchase Agreement facility. SVBFG had limited collateral pledged to the Federal Reserve’s discount window, had not conducted test transactions, and was not able to move securities collateral quickly from its custody bank or the FHLB to the discount window.

The missing CRO

In a different previous post on SVB, I mentioned that the company was required by Reg YY to maintain a Chief Risk Officer, but had failed to do so for much of 2022. This was bad, clearly, but in that prior post I wondered whether it was technically a violation, given that the regulation does not provide parameters on how long (if at all) the CRO role can be allowed to remain vacant.

Neither the Board’s regulations nor, from what I can tell, any other guidance from the Board describe how long is acceptable to have the CRO role sit vacant. Obviously transition periods, gardening leaves, and so on mean that it is unrealistic to expect the CRO role to never be vacant, but the way that SVB Financial handled this particular transition does not seem like best practices. I’ll be curious to see if the Board’s upcoming report on what went wrong from a supervisory perspective provides any more detailed guidance or recommendations for other firms in this position.

The Review does not provide a concrete answer to this question, although it does say that this something is supervisors “could” have used as the basis for an MRIA. Instead, however, “[i]n consultation with Board staff, supervisors decided not to issue the violation since the firm was actively searching for a CRO with the appropriate skills and experience.”

Board review of applications

Another thing previously discussed on this blog was SVB’s 2021 acquisition of Boston Private, which the Board approved in a unanimous order saying, among other things, that SVB would not “pose significant risk to the financial system in the event of financial distress” or “exhibit an organizational structure, complex interrelationships, or unique characteristics that would complicate resolution of the firm.”

The Review includes a section on the Board’s approval (either itself or under delegated authority) of applications filed by SVB or its holding company in recent years. The commentary in the Review on the Boston Private approval is factual only, and does not explain much about how these conclusions were reached or if, in hindsight, the Board would have done anything differently.

The Board’s Division of Research and Statistics (R&S) is responsible for completing the financial stability analysis related to applications acted on by the Board. R&S staff concluded that the proposed merger would not result in meaningfully greater or more concentrated risks to the financial stability of the United States.

Maybe more interesting is the discussion in this same section of the Review of notices filed by SVB under Regulation K, the Board’s regulation governing certain overseas investments by U.S. banking organizations. Here, the Board says that an October 2022 $1.8 billion investment by SVB in its UK entity, SVB UK Ltd, was opposed by both the supervisory central point of contact at the Reserve Bank and an analyst in the LFBO group at the Board.

The supervisory CPC highlighted supervisory issues that SVB needed to remediate at the time of the October 2022 notice and recommended that it not be approved. The Board LFBO analyst had a similar recommendation due to recent liquidity risk management issues and outstanding information technology and European exchange rate mechanism issues.

But ultimately, “staff decided that there were not sufficient grounds to object to the notice.”

The Review does not offer anything more by way of explanation.

SVB IDI plan

SVB Financial (the holding company) was not required to file a living will with the Board/FDIC because it had not yet crossed the threshold at which that requirement applies under the 2019 tailoring rules. SVB (the bank) was, however, required to file an IDI resolution plan with the FDIC, and did so for the first time in December 2022.

I’ve wondered previously how that went. We have yet to hear directly from the FDIC itself on this point, but the interim GAO report that was also released today suggests the FDIC was not impressed by what it saw.

SVB submitted its first resolution plan on December 1, 2022. (SVB had exceeded $100 billion in total assets in 2021.) According to FDIC staff, they were still reviewing the plan at the time of the bank’s closing. They told us that their reviews of resolution plans typically take 5–6 months. […]

According to FDIC officials, their preliminary findings were that the bank’s initial resolution plan was not thorough. For example, according to FDIC staff, the resolution plan did not list potential acquirers for a whole bank purchase, specific portfolios, and franchise components. The plan did not detail crisis communication, liquidity needs, liquidity resources, or processes for determining liquidity drivers.

It is not clear from the GAO report how much of this is being said with the benefit of hindsight, and how much of it was apparent from the FDIC’s review even before the bank’s failure.

Volcker

There is a kind of odd section of the Review in which the Volcker rule is discussed. First the Board says, pretty definitively, that “[t]he activities that led to SVBFG’s failure were not the activities that the Volcker rule was intended to address.”

But, you know, some people really like the Volcker rule, and whether for that reason or for other reasons, in the subsequent paragraphs the Board tsks with mild disapproval at changes made to the rule during the Trump Administration.

Unlike certain other changes discussed in the Review, the Review stops short of calling outright for these changes to be reversed, instead opting to just kind of shrug and say the changes may have made it more difficult to say for sure whether SVB was fully compliant.

SVBFG held investments in certain venture capital funds that may have been covered funds subject to the restrictions of the Volcker rule. The Volcker rule excludes “qualifying venture capital funds,” as defined by the SEC regulations from the restrictions of the covered fund provisions.

SVBFG was presumed to be in compliance with the Volcker rule because it had limited trading assets and liabilities, and SVBFG had no obligation to affirmatively demonstrate compliance with the regulation on an ongoing basis. This presumption, along with the reduced recordkeeping requirement for SVBFG’s fund investments, resulted in limited documentation that Federal Reserve staff could review to determine whether SVBFG would have been in compliance with the Volcker rule or met the requirements of any applicable exceptions, including without the presumption of compliance or absent the changes to the regulations.

The Board’s own evaluation of the Federal Reserve Bank of San Francisco

On page 31 of the Review, the Board describes how it has consistently rated the FRBSF’s supervision of RBOs and LBOs as either “Strong” or “Effective” (the two highest ratings on a four-factor scale).

For instance, in 2022 staff of the Board’s Division of Supervision and Regulation concluded that, “with respect to the SVBFG transition, that supervisory planning [at the Reserve Bank] had been effective and necessarily agile.” In addition, in the view of S&R staff, SVB’s Dedicated Supervisory Team at the Reserve Bank had “demonstrated superior ability.”

The Review does not go into detail on what the Board believes went wrong in its approach to evaluating the FRBSF, but evidently something in that evaluation did not match the facts on the ground, given that the Board a few pages later in the Review notes that “[t]he transition of SVB from the RBO portfolio to the LFBO portfolio lacked a defined plan and process.”

Overall, this section of the Review, confined to a three quick paragraphs, felt a little underdeveloped to me. Especially given the fact that it is tough to chalk this up to a one-time failure of the FRBSF RBO supervisory program, in light of the mess earlier this year with Silvergate Bank.

Bad consulting

Most of the criticism in the report is reserved for the Board’s own supervision and regulation, as well as the failures of SVB’s management. But an unnamed consulting firm does come in for its own brief criticism.

The board, management, and chief risk officer (CRO) all failed to recognize that their year-long program for their risk-management framework to meet EPS was ineffective, until supervisors started identifying issues in late 2021. Consultants who did the initial 2020 EPS gap assessment with respect to SVBFG practices and helped execute the plan to close those gaps also failed to design an effective program.

Thanks for reading! As always, thoughts, challenges, criticisms are always welcome at bankregblog@gmail.com. I am sure there are interesting things this post missed.

1. The Board also made certain supervisory materials made publicly available in connection with the Review.

2. I would have loved to have read them, but I cannot really blame the Board for not releasing the interviews, particularly given that in many of them staffers are evidently criticizing their superiors.

One thing to watch may be whether the various House committees seeking to conduct their own review will try to obtain copies of the transcripts.

3. This is on page 7 of the Executive Summary, but see also page 53.

Due in part to SVB’s “Strong-1” Liquidity rating and the perceived low level of inherent risk, the examination of liquidity risk-management practices during the annual CAMELS and BHC exams was not extensive. RBO “risk-focusing guidelines” led staff to conduct lighter reviews of areas where either inherent risk was considered low or risk-management practices were satisfactory. Typically, one person would cover multiple assignments (e.g., liquidity, interest rate risk, and the investment portfolio).

4. Consistent with this conclusion, VCS Barr’s cover letter says that regulators “should consider setting tougher minimum standards for incentive compensation programs and ensure banks comply with the standards we already have.”