Lessons Learned? A Report Card

We all make mistakes. But it’s also important to learn from those mistakes. Banking regulators frequently go thought a lessons learned exercise following bank failures and other events that expose weaknesses in the bank supervision system. We’ll look at some of the lessons learned from the 2023 bank failures and give out a report card indicating how well bank regulators have really learned their lessons.

A wave of bank failures, or even the failure of one or two major banks, usually sets in motion a series of post-mortems that look at what went wrong. These can come in the form of a material loss review by that agency’s Inspector General, peer reviews, Congressional hearings, or ad hoc internal reviews, such as the one done by Michael Barr following the failure of Silicon Valley Bank. The lifecycle of a lessons learned review follows this basic pattern:

What Went Wrong => Corrective Action => Backsliding

You can judge responses to lessons learned by looking at whether and to what extent the recommendations were implemented and then consider how long it took to see some backsliding. The worst responses are like the Bourbons: learned nothing and forgotten nothing.

Approaches can differ on whether to focus on missteps by individuals vs. institutional failures. Individuals made the mistakes, and those mistakes should have some consequences. The passive phrase, “lessons learned” itself gives off a distinct “mistakes were made” vibe. But tossing some (preferably mid-level) individuals under the bus may do little to address more fundamental weaknesses. It can also give truth to the phrase that [poop] rolls downhill. As shown here, a U.S. Senator and a Fed Governor can decry the lack of accountability at the Federal Reserve (was anybody fired?) while furiously disclaiming their own responsibility for legislation and regulation that helped enable SVB’s collapse.

A Look at Some of the Lessons

When we look at the 2023 failures, a few key issues stand out. They include the supervision of interest rate risk and liquidity; the use and misuse of tailoring; the need to act more quickly and decisively; and the need to maintain adequate staffing levels. The post-mortems applied to individual banks and supervisors, but the lessons learned can apply more broadly. The assessment considers responses across the banking agencies. Summary grades are provided below.

Need to Be More Attentive to Interest Rate Risk

What went wrong. The Fed’s SVB post-mortem notes that, “Supervisory responses for IRR were not rapid or severe enough given the fundamental issues in this area that actually drove poor decisions at SVBFG.” The report further noted that economic value of equity, a better indicator of structural IRR, received insufficient attention. The Fed’s I.G. likewise found that “Despite the warning signs, examiners did not sufficiently scrutinize the risks from rising interest rates on SVB’s HTM investment securities portfolio.” The FDIC echoed some of these criticisms in its review of the First Republic failure. However, as noted in an earlier post, the size and the cost of the First Republic bank failure should have merited a more self-critical tone.

Corrective Actions. Regulators have made no significant regulatory or policy changes regarding interest rate risk. While European regulators have specified a set of standardized rate scenarios and established numerical thresholds for IRR outliers, U.S. regulators have failed to follow suit. Public disclosures of IRR rarely mention EVE and generally lag reporting requirements for European banks. The FDIC last updated its examiner’s manual for IRR in 2018. The last updates by the Fed and OCC were in 2020. The only substantive proposal related to IRR was to include unrealized losses on available-for-sale (AFS) securities in regulatory capital for a wider group of banks. However, unrealized losses on held-to-maturity (HTM) securities played a bigger role in the cost of the bank failures.

Supervision appears to tell a different story. The confidential nature of supervisory information means that we can only make inferences based on available data. However, a look at public enforcement actions shows that provisions related to IRR have become much more common. Regulators also made changes to the Uniform Bank Performance Report to include more IRR-related information.

Grade. The response has been somewhat of a mixed bag here. The lack of a policy response is concerning but there seem to be some improvements in the supervision of IRR. The regulators merit a gentleman’s C.

Improve the Assessment of Liquidity Risk

What went wrong. Liquidity risk was the proximate cause of the SVB and Signature bank failures and played a significant role in First Republic’s failure as well. These failures exposed some critical weaknesses in the regulation and supervision of liquidity risk, including concentrations of uninsured deposits, failure to take account of the market liquidity of some “high quality liquid assets,” and inadequate contingency funding planning. Examiners did identify some key liquidity management weaknesses at SVB but failed to act quickly or decisively. This was also largely the case at Signature Bank. First Republic was a more thorough miss, with Liquidity assigned a “1” rating until March 2023.

Corrective Action. There has been more of a policy response to liquidity risk than to IRR. In July 2023, regulators addended a 2010 policy statement on liquidity risk management to stress “the importance of maintaining, assessing, and testing their contingency funding plans.” The agencies have also updated their examiner handbooks to incorporate some of the lessons learned from the 2023 bank failures. On the other hand, there have been no changes either to the application of the liquidity coverage ratio or to its calculation. As with IRR, we have seen more formal enforcement actions related to liquidity risk. On a less positive note, uninsured demand deposits are still treated as core deposits for purposes of both the UBPR and for the FDIC insurance assessment.

Grade: The policy response was a bit better than for IRR but with plenty of room for improvement. Supervision of liquidity risk appears to have improved. Regulators get a C+.

Tailoring Efforts in 2018-2019 Went Too Far

What went wrong. The SVB post-mortem found that the “Board’s tailoring approach in response to EGRRCPA and a shift in the stance of supervisory policy impeded effective supervision by reducing standards, increasing complexity, and promoting a less assertive supervisory approach.” This critique had less to do with the concept of tailoring than to its practice, which extended light touch supervision to a wider group of firms. The Fed’s I.G. made a similar critique but couched it in terms of not enough tailoring. A fast-growing, increasingly risky firm like SVB was treated in much the same way as smaller, less risky firms.

There has been pushback on these findings from the industry’s advocates, pointing out, for example, that the Fed’s supervisory stress tests focused on credit risk rather than IRR and would not have flagged SVB as high risk. It should also be noted, though, that one rationale for lighter touch supervision for firms under $300 billion was that such firms posed little systemic risk. Actions to bail out uninsured depositors and SVB, Signature, and First Republic strongly suggest this was not the case.

Corrective Action. The initial Basel Endgame proposal would have extended more stringent capital requirements to a wider range of banks. However, that proposal met with furious resistance and seems to be going nowhere. More recently, both members of Congress and regulators have doubled down on tailoring, suggesting that lighter touch supervision should extend to more institutions not less. Taking a page from George Costanza, they looked at the lessons learned and did the opposite.

Initial efforts on the supervisory side were more significant. There has been a noticeable decline in the Fed’s supervisory ratings for large financial institutions (LFI). According to the Fed’s Supervision and Regulation Report, firms rated at least satisfactory on all components fell from about half in 2022 to about one-third by 2024. There has also been a sharp rise in the number of supervisory findings. There has been a more modest decline in CAMELS ratings, which apply to all banks. In response, Congress has proposed revisions to the CAMELS ratings system to focus on (as yet unspecified) “clear, objective measures.” The Fed has proposed changes to the LFI ratings framework, so a bank with a “deficient” governance and control rating could still be considered “well managed” for some other supervisory purposes. I kid you not.

Grade. There has been little substantive change in tailoring regulations, with Congress and regulators are now moving sharply in the opposite direction. Supervisors have taken a more assertive approach toward regional banks but appear now to be reversing course as well. Tailoring’s supporters have denied its role in the collapse of SVB and other banks, but asserting doesn’t make it so. The Tobacco Institute also strongly denied that smoking was bad for you. The overall grade here is C-, but it’s likely to get worse rather than better over the next few years.

Regulators Should Address Emerging Problems with a Greater Sense of Urgency

What went wrong. Supervisors not only missed red flags,but they also failed to address the problems they did identify in a timely fashion. The SVB post-mortem notes (on page 9) that the Fed’s supervisory approach to SVB was “too deliberative and focused on the continued accumulation of supporting evidence in a consensus-driven environment.” The report further noted “supervisory policy placed a greater emphasis on reducing burden on firms, increasing the burden of proof on supervisors, and ensuring that supervisory actions provided firms with appropriate due process… in some cases, the changes also led to slower action by supervisory staff and a reluctance to escalate issues.” The FDIC also acknowledged in its review of the Signature Bank failure that it “could have acted sooner and more forcefully to compel the bank’s management and its board to address these deficiencies more quickly and more thoroughly.”

Corrective Actions. We don’t have a lot to go on here as most bank supervision occurs outside the public view. Neither the Fed’s Annual Report nor its Supervisory Report mention timeliness. Exam timeliness has not been the subject of a recent Fed OIG audit. From what we do know, things seem to be moving in the opposite direction. Both Congress and senior regulators have advocated providing banks with a stronger hand in the appeals process, with some proposals also freezing future supervisory actions until resolution of the appeals. This is bound to slow down the process. The Wall Street Journal also reports that LFI ratings were delayed this year as Vice Chair Bowman sought additional layers of review.

Grade. There is simply too little information on timeliness to reliably assess the Fed’s response. Therefore, it receives a grade of I (Incomplete). The information that is available, however, suggests that the Fed and other regulators seem to be moving in the wrong direction.

Enhance Staffing Levels

What went wrong: Staffing levels, especially for larger banks, are inadequate. The FDIC’s Signature Bank post-mortem observed that “resource challenges with examination staff that affected the timeliness and quality of SBNY examinations.” The report went on to point out that “since 2020, an average of 40 percent of the LFI positions in the NYRO have been vacant or filled by temporary staff, and a number of large bank dedicated teams have significant vacancies.” The report noted staffing shortages in other FDIC regions as well.

Staffing shortages have also been a persistent problem at other agencies. In the early 2020s, for example, OCC renewed efforts to hire experienced examiners and reinstituted incentive relocation incentives, particularly for Large Bank Supervision in New York. The agency faced a shortage of experienced staff to execute its supervisory strategy.

Corrective Actions. The agencies are moving sharply in the opposite direction with unprecedented levels of staffing cutbacks. Rather than mere proposals, these cutbacks have already been implemented. News reports indicate that the FDIC is eliminating 1,250 jobs, or roughly 20% of its workforce. Staffing reductions at OCC have been even more extensive. The agency projects a year-over-year decline in staff of more than 30%.

OCC’s budget justification to Congress does little to explain how the agency will achieve its mission with cuts of this magnitude outside of vague references to “leveraging technology” and reference to a long-delayed Single Supervisory Platform. The rest of the narrative consists of the usual buzzword bingo: “modernization,” “one size fits all,” “tailoring,” and digitalization is a “transformation” not a trend. Sen. Elizabeth Warren asked incoming Comptroller Jonathan Gould (page 5) whether it was “important for OCC to have the necessary staff to execute its mission” and Gould completely ducked the question.

I’m not aware of any public comments made by the FDIC to explain its staffing reductions. On April 24, Sen. Warren and three other senators sent a letter to Acting FDIC Chair Hill requesting further details and justifications for staffing cuts at FDIC. Despite a May 8^th deadline for Hill’s response, there have been no further public comments from the FDIC.

Grade. As with tailoring, banking agencies have moved in the opposite direction of the lessons learned. However, the staffing response has been far worse. While the doubling down on tailoring is concerning, most of the suggested changes remain in the proposal stage and policy changes can be reversed at a later stage. The staffing reductions have already been put in place, and restoring staffing to their former levels could take years. Moreover, tailoring advocations at least offered a counternarrative, albeit an unconvincing one. The staffing cuts took place with virtually no justification or even explanation. Based on the foregoing, regulators get a failing grade of F.