Probative Inquiry into the
2023 Financial Sector Stress Events
The Barr Report
The Barr Report was conceived amid crisis. On March 13th, only three days after the closure of SVB — and while other institutions were failing or under acute stress — the Federal Reserve announced that Vice Chair for Supervision Michael Barr would lead a review of SVB’s failure.1 The review was expected to examine the regulations applicable to firms such as SVB, the supervisory regime under which the firm had operated, and whether supervisors possessed sufficient tools to identify and address the weaknesses that had emerged.
The Federal Reserve committed to publishing the results by May 1st 2023, and the Barr Report was issued on April 28th, less than seven weeks after the review began.
Directing the staff-led effort was [Board Staff 14], chosen for his deep supervisory experience — including five years leading supervision at the Federal Reserve Bank of New York — and a history of managing complex, system-wide initiatives. His subsequent secondment to the Board, moreover, had kept him at arm’s length from day-to-day supervision of SVB. [Board Staff 14] thus offered both the institutional expertise and independence needed to lead the review.
During Starling’s interview with [Board Staff 14], he recalls learning of SVB’s failure “just like everyone else did,” from the public announcement of the decision to exercise the Systemic Risk Exception. Within days, he had been asked to lead the review captured by the Barr Report. His response reflected the institutional mood: “When something comes up, everyone is all hands on deck, and you do what you can to help.” Once he accepted the assignment, he recalls, “it was full steam ahead.”2
Working with [Board Staff 14] was [Board Staff 2], a thirty-year Federal Reserve veteran with experience spanning community, regional, large-bank, and systemically important financial-institution supervision. In Starling’s interview with [Board Staff 2], he explained that he “ran the review of the supervision of SVB,” while [Board Staff 14] had responsibility for the overall Report.3
The broader review team, both men explained, was organized by subject-matter expertise and “risk stripe,” drawing specialists in liquidity, interest-rate risk, credit risk, governance, policy, and law from across the Federal Reserve System. A smaller synthesis group — centered on [Board Staff 14], [Board Staff 2], and supported by several members of the Fed’s Policy and Legal staff, overseen by Vice Chair Barr — worked to turn those specialist contributions into a coherent public account.
The production effort was intensive. [Board Staff 14] recalls “morning meetings every day for six weeks, lots of phone calls, lots of checking in.” The subject-matter teams were instructed to determine what had happened, from the perspective of their respective disciplines, and were given access to the relevant records. [Board Staff 14] and [Board Staff 2] then worked to identify the through-lines connecting their findings.
That was no small editorial undertaking. [Board Staff 14] estimates that the teams generated two or three times the volume of raw material that ultimately appeared in the roughly 100-page public Report. “A lot of what [Board Staff 2] and I did,” he explains, “was take this vast amount of stuff and funnel it into something that was hopefully manageable and useful.” [Board Staff 2]’s supervision team alone produced what he remembers to be approximately 200 pages of draft material. Much of it ultimately “hit the cutting room floor” as the team distinguished details essential to explaining SVB’s failure from those unlikely to illuminate it for a public audience.
Strengths of the exercise
The Report’s principal staff-authors describe an effort marked by diligence, technical depth, and considerable editorial independence. [Board Staff 2] recalls working from approximately 5:00 a.m. until midnight for several weeks. [Board Staff 14] remembers Board lawyers reviewing successive versions at 3:00 in the morning, with “people just grinding to get every bit done and checked.”
The intensity of that effort was matched by the quality of the people involved. [Board Staff 14] describes the Federal Reserve staff as being “at their best in the moment of most uncertainty.” The “deepness” and “breadth” of their subject-matter expertise, he says, were “stunning”: “I would put that group, from a technical sense, up against anyone,” [Board Staff 14] said.
Both [Board Staff 14] and [Board Staff 2] emphasize that the review team was permitted to reach and publish conclusions uncomfortable for the Federal Reserve itself. [Board Staff 14] recalls advising at the outset that, if he were to undertake the assignment, “I have to say what I think.” He reports that this condition was honored.
[Board Staff 2] describes [Board Staff 14] setting the same expectation for the wider group: “Listen, this is our report. If we want something in there, it’s going to stay in there.” In [Board Staff 2]’s telling, that message was directed not only to the working team, but also to the General Counsel, the head of Supervision and Regulation, Vice Chair Barr, and others. It established that material would not be removed merely because it was uncomfortable for the institution to place it into the public record.
The result was an unusually candid act of institutional self-examination. The Barr Report did not limit blame to the failed bank. It acknowledged serious shortcomings in Federal Reserve supervision, including failures to appreciate the extent of SVB’s vulnerabilities, delays in escalating identified concerns, and institutional practices that discouraged sufficiently timely and forceful action. The Federal Reserve also released a substantial body of material ordinarily treated as Confidential Supervisory Information, providing the public with uncommon visibility into the conduct of supervision.
These are significant achievements, and the limitations discussed herein should not obscure them. The Barr Report was a serious, detailed, and unusually transparent effort, undertaken by experienced public servants, working under extraordinary time pressure and intense public scrutiny.
The constraints of a compressed inquiry
Yet the very conditions that make the Barr Report an impressive study demand that we exercise due care before unquestioningly accepting its conclusions. [Board Staff 2] offers a vivid account of the production timetable:
“You get about a week to figure out what happened, you get a week to write it, and then you get four weeks to edit it.” The team continued learning throughout those four weeks, but the timetable meant, as [Board Staff 2] acknowledges, that “you need to draw your conclusions early.” Asked whether the process might fairly be characterized as “ready, fire, aim,” he replied: “Yeah, a little bit. We’re not going to be way off target when we fire, but we do a lot of refining on the backend.”
That iterative process undoubtedly improved the final product. [Board Staff 2] credits Board Legal, Congressional Affairs, communications personnel, and other reviewers with supplying valuable questions and clarifications. [Board Staff 14] describes the Report as a broad Federal Reserve consensus document that was read and commented upon by dozens of Federal Reserve Board supervisory and Legal staff, under the direction of the Vice Chair for Supervision. But editing and refinement are not the same as beginning again from first principles. Once the principal causal narrative had been established, which appears to have been early in the production process, the remaining weeks were devoted primarily to refining and communicating that narrative.
This creates a risk familiar to any inquiry conducted under intense time pressure: an early organizing theory may become the frame through which later evidence is selected, interpreted, and offered. That does not in itself indicate confirmation bias. It does, however, create conditions in which confirmation bias may arise.
[Board Staff 14]’s own account provides an important counterweight: “This wasn’t a hypothesis testing exercise,” he explains. “I knew nothing about the firm going in.” He and the other reviewers read “document after document” and formed their views as they proceeded. Nevertheless, he also recalls that there was no substantive debate over the principal finding: “The primary conclusion was that the firm failed to manage its own risks, and there was no debate about that being the fundamental driver of what happened.”
[Board Staff 2] brought to the exercise precisely the seasoned examiner’s judgment that the review required. He explains that he joined [Board Staff 14] for many of the staff interviews because, in his own self-description, “I’m an old cranky examiner.” That experience was an asset: expert pattern recognition allows an examiner to find signal amid an overwhelming volume of information. But pattern recognition can also become pattern completion. An experienced examiner’s instinct to understand a bank failure first as the consequence of unsafe positions and unsound practices may shape which questions are pursued, which facts are treated as causal, and which competing explanations appearless salient out of hand.
Again, this is not to suggest that [Board Staff 2] approached the work in bad faith or consciously sought evidence supporting a predetermined conclusion. Nothing in our conversations with him or [Board Staff 14] would support such a charge. Rather, it is to recognize that the perspective of the “old cranky examiner” was both a source of strength and a potential interpretive lens through which the record was viewed, analyzed, and presented.
A substantial but one-sided testimonial record
The review team conducted interviews with approximately 35 people from across the Federal Reserve System. Those discussions generated 43 sets of interview notes, which Starling has yet to analyze in full. [Board Staff 2] indicates that he and [Board Staff 14] conducted most of the interviews with senior staff together. They spoke with current and former members of the SVB supervisory team; personnel involved in liquidity, interest-rate-risk, governance, audit, and other examinations; members of the Federal Reserve’s Risk function; and senior leaders in San Francisco and at the Board.
This was a substantial undertaking within the available time. But the witness pool was exclusively internal to the Federal Reserve. The review team did not interview SVB’s executives or directors. It did not issue document requests to them. The same appears to be true of the OIG and GAO postmortem studies.
When asked whether the Barr review team had spoken with SVB executives or directors, [Board Staff 14]’s answer was unequivocal: “We did not.” He explained that the record reviewed by his team included not only supervisory documents, but also internal firm materials that had been obtained in the ordinary course of supervision, including records and summaries concerning matters presented to the board of SVB. As such, he confirmed: “We did not talk to the management or the board of the firm.”
In a separate discussion outside the recorded interview, [Board Staff 14] explained that the review team did not believe it needed additional information from SVB’s management or directors. In his view, the records already held by the Federal Reserve provided his review team with all the information necessary to reach the Barr Report’s conclusions regarding SVB’s failure to manage its risks.
That explanation, however, presents a difficult tension. If the evidence already in the Federal Reserve’s possession was sufficient to make SVB’s risk-management failure unmistakable after SVB had failed, then the essential information was also available to Federal Reserve supervisors before SVB’s failure.
[Board Staff 14] describes the evidence as “crystal clear” and the firm’s risk-management failure as “the prime cause of what happened.” If so, the central question becomes less whether Federal Reserve supervisors possessed the information necessary to identify SVB’s vulnerability, than why it is that they failed to synthesize that information, to recognize the firm’s deepening vulnerability, and to force timely remediation well before the bank failed.
The distinction between anticipating the precise trigger and recognizing underlying vulnerability is important. Supervisors may reasonably have struggled to anticipate the speed, scale, and immediate catalyst of the run. But difficulty predicting the proximate trigger is not the same as an inability to recognize that SVB had become acutely vulnerable to a spike in interest rates, or some other potentially destabilizing event.
[Board Staff 2]’s reflections point toward this distinction. He describes supervisors as having focused on weaknesses in the processes by which SVB measured liquidity and interest-rate risk, without directly confronting the risk positions themselves: “We could see that the process had flaws, and so we cited the process. We did not cite the position.” That observation suggests that the raw facts may have been visible, while their combined prudential significance was not recognized.
The absence of interviews with SVB’s management and directors therefore matters for more than procedural completeness. Such interviews might have supplied relevant factual information, contested particular characterizations, or revealed additional evidence about governance and remediation. They might also have helped the Barr, GAO, and OIG review teams understand why supervisors and the supervised firm interpreted the same underlying information differently before and after the failure made its significance appear obvious.
The character of the interview evidence
A further methodological limitation of the Barr Report’s production arises from the manner in which review team interviews were recorded. The 43 interview records are not verbatim transcripts. They are summaries prepared by Board Legal staff who were involved in the discussions in at least a note-taking capacity.
There is no reason to presume that those summaries were inaccurately or improperly prepared. Legal participation may have improved discipline, clarity, and consistency. Nevertheless, a summary is an interpretive record. It necessarily reflects judgments about what to capture, what to condense, what to omit, how to characterize uncertainty, and which statements warrant emphasis.
The intervention of a legal note-taker, therefore, introduces an additional layer between the words of the interviewee and the narrative eventually presented in the Barr Report. It also creates a possible channel for a distinctly legalistic bias: a tendency to privilege documentary defensibility, chronology, formal authority, and supportable attribution over ambiguity, institutional culture, tacit understanding, and the informal dynamics through which supervisory judgments are actually reached.
This does not establish that the interview notes or the Barr Report were distorted. But where conclusions were formed rapidly, interviews were limited to Federal Reserve personnel, and the testimonial record consists of lawyer-prepared summaries rather than full transcripts, the possibility of cumulative interpretive bias is real. The Barr Report’s underlying interview notes must therefore be read alongside contemporaneous documents, draft work products, and other evidence rather than treated as neutral substitutes for the conversations themselves.
How the Barr Report’s findings should be read
The Barr Report articulates four principal findings: (1) that SVB’s board and management failed to manage the firm’s risks; (2) that supervisors did not fully appreciate the bank’s vulnerabilities; (3) that supervisors did not act with sufficient force or speed once problems were identified; and (4) that changes in the regulatory framework, caused by the enactment of EGRRCPA, and shifts in supervisory policy priorities under then-Vice Chair for Supervision Quarles, which emphasized due process, “impeded effective supervision.”4
[Board Staff 14] is emphatic that these findings were intended to appear in descending order of causal significance. Asked whether they appeared in order of importance, he answered: “Absolutely.” He also acknowledged that this hierarchy might have been made clearer: “Maybe it should have been explicit in the report.”
[Board Staff 2] goes further. With hindsight, he would have stopped after the first two findings: “The bank screwed up, we missed it, end of story.” Although he continues to stand by the substance of the Barr Report, [Board Staff 2] believes the fourth finding generated political and institutional “noise” that distracted from the Report’s key propositions. “Honestly,” he says, “I would’ve stopped at Key Takeaway 2.”
That formulation is refreshingly candid, but perhaps too neat. The first two findings cannot be treated as wholly independent propositions. If, as [Board Staff 14] argues, the records already held by the Federal Reserve made SVB’s failure to manage its risks “crystal clear,” then the proposition that “the bank screwed up” immediately raises the question of why the Federal Reserve “missed it.”
The first finding describes the condition of SVB. The second demands an explanation of the supervisory system that possessed the relevant information but failed to convert it into effective action.
The methodology of the Barr review may have encouraged the first proposition to settle early. The review was conducted through records gathered in the normal course of supervision and interpreted primarily by supervisors and other Federal Reserve personnel. It did not include interviews with those at SVB who might have contested or contextualized the emerging account. And interview evidence used to test that account was summarized by Federal Reserve lawyers, rather than preserved through verbatim transcripts.
These circumstances do not invalidate the findings of the Barr Report. They do, however, counsel against treating its causal narrative as settled history and help to explain why further inquiry is warranted.
[Board Staff 14] himself welcomes that effort. He describes Starling’s Probative Inquiry as “a great effort” that “should absolutely be done,” and he endorses “a one-step-back reassessment.” In that same spirit, the Barr Report is best treated as an indispensable starting point: a serious and unusually candid internal postmortem whose strengths give its conclusions weight, and whose constraints make continued examination necessary.
Join The Discussion
Sign in and be the first to comment.