The Condition

Most institutions that examine governance failures are examining someone else. A regulator reviews a firm. An auditor reviews a company. A court reviews a board's conduct. The examining institution stands, structurally, outside the object of its inquiry.

In 2023, something less common occurred. Following the failures of Silicon Valley Bank and Signature Bank, the Federal Reserve and the FDIC – the primary federal supervisors of the two banks – each conducted a public review not only of what the banks had done wrong, but of what their own supervision had failed to convert into timely action. This is a different kind of document than an enforcement order or an examination report. It is an institution turning its own governance-visibility apparatus on itself.

What these reviews reveal is worth examining carefully, and for a specific reason: they are not third-party accounts constructed after the fact by critics with an interest in the outcome. They are the reviewing institutions' own characterizations of what they knew, when they knew it, and what happened between knowing and acting. That specific evidentiary posture – self-examination, independently corroborated by a structurally separate reviewing body – is what makes this territory worth Obraval's attention, and what requires unusual discipline in how it is read.

What Was Known

The Federal Reserve's own review of its supervision of Silicon Valley Bank states, in its own words, that supervisors "identified interest rate risk deficiencies in the 2020, 2021, and 2022" examination cycles. This is not a finding reconstructed by outside investigators piecing together fragments after the fact. It is the Federal Reserve's own account of its own examiners' own work product across three consecutive years.

What the same sentence goes on to say is the more precise and more useful fact: supervisors identified these deficiencies "but did not issue supervisory findings." The deficiencies were, in the report's words, "only communicated as written advisories or verbal observations" – a form of communication that exists below the threshold of the Federal Reserve's own formal escalation tools, the Matters Requiring Attention and Matters Requiring Immediate Attention designations that trigger tracked remediation timelines and visibility up the supervisory chain.

This is worth sitting with precisely because it resists a simpler, more dramatic reading. It would overstate the record to say supervisors "knew" the bank would fail and did nothing. What the record actually shows is narrower: examiners recognized a specific weakness, repeatedly, across multiple annual cycles, and each time routed that recognition through an informal channel rather than the formal one built to compel and track remediation.

The First Transition

Between recognizing a condition and compelling action on it sits a decision: whether to formalize the finding. For three years, in the specific matter of Silicon Valley Bank's interest rate risk management, that decision resolved toward the informal channel. The formal designation – the November 2022 Matter Requiring Attention – did not arrive until the bank had grown from a regional institution to one holding more than $200 billion in assets, and until, in the report's own words, the bank's condition had already deteriorated substantially.

This transition – from recognized-but-informal to formally-designated – is the first and, on this record, the longer of two separate gaps the Federal Reserve's review identifies. It is not a gap in information. The information existed in examiner work papers across three exam cycles. It is not, on this record, a gap in authority – the MRA and MRIA tools existed throughout and were, in fact, used 54 separate times against this same institution since 2019. It is a gap in the decision to invoke the formal mechanism for a specific, recurring, and eventually consequential weakness.

The Second Transition

A second, distinct transition follows the first, and the record here deserves care about mechanics. The same 2022 examination cycle that produced the November MRA also produced a separate supervisory conclusion: examiners planned to downgrade the bank's Sensitivity to Market Risk rating. These were related but procedurally distinct outputs of the same exam – an MRA is a specific finding requiring remediation; a CAMELS component rating is a periodic assessment of condition – not one mechanically triggering the other. What matters for this piece is not that one caused the other, but that both represent a formal supervisory conclusion reached in late 2022, and that the follow-through on that conclusion – the downgrade itself – was never finalized before the bank failed the following March. The gap here is not between a formal finding and its automatic procedural consequence. It is between reaching a formal conclusion and finishing what that conclusion was supposed to set in motion.

A Second Institution, a Structurally Similar Pattern

The FDIC's review of its own supervision of Signature Bank, published the same day as the Federal Reserve's report, describes a related but distinct version of the same underlying condition. The FDIC's own account states that examiners "documented and discussed repeat findings" across "multiple examination cycles" – a formal record of recurring concern – "without management effectively addressing the underlying supervisory concern." Separately, the report notes that the bank's Liquidity component rating was downgraded to "Fair" in 2017, while the bank's overall composite rating remained "Satisfactory" for six more years, until 2023.

This is worth reading precisely rather than dramatically. A component-level warning existed, formally, for six years, without producing a corresponding change in the aggregate assessment that typically drives the intensity of supervisory response. The FDIC's own conclusion is direct: "In retrospect, the FDIC could have acted sooner and more forcefully." The same report separately identifies a capacity constraint distinct from either information or authority: "failures in FDIC, due largely to staffing shortages, were a contributing factor" to delays in identifying and reporting the bank's weaknesses. This is neither an information gap nor an authority gap – it is a resourcing constraint on how quickly the existing authority could be exercised, offered by the institution itself as one contributing factor among several, not the whole explanation.

Independent Corroboration Across Time

A structurally separate institution – the U.S. Government Accountability Office, which reviews the regulators rather than being one of them – examined the same general condition both before and after the 2023 failures, which allows a check on whether this is a pattern specific to 2023's crisis narrative or a more durable institutional condition.

In 2019, four years before Silicon Valley Bank failed, GAO found that Federal Reserve "data for escalation of matters requiring attention to matters requiring immediate attention and enforcement actions were collected in a manner that made it difficult for [GAO] to determine the extent to which escalation occurred." This is a striking finding in its own right: an independent oversight body, examining the Federal Reserve's own records years before any specific failure was in question, could not itself determine how often or how reliably concerns moved from a lower-severity designation to a higher one.

In its 2024 review of the specific SVB and Signature Bank matters, GAO reached a closely related conclusion: "The Federal Reserve's procedures for moving from a lower-level concern to an enforcement action often weren't clear or specific, which may have delayed enforcement actions." And examining Silicon Valley Bank's supervisor specifically, GAO found that the San Francisco Federal Reserve Bank had, by its own internal documents, already concluded that an informal enforcement action was warranted – yet "did not recommend the issuance of a single enforcement action despite the bank's serious liquidity and management issues before the bank's failure."

Read together across five years and two separate reviews, these findings converge on something more precise than "regulators were slow." They converge on a specific, recurring institutional condition: the criteria and procedures governing the transition from a recognized concern to a more forceful response were not, themselves, clearly specified – independently identified before either bank's failure and independently reconfirmed after it.

The Distinction, Refined

Read across these sources, the governing distinction is not simply information possession against institutional action, stated as a single before-and-after. The evidence supports something more precise: at least two separate transitions exist between recognizing a condition and completing action on it, and each is a distinct point where delay can occur, largely independent of whether authority to act exists.

The first transition is from recognition to formal designation – the decision to invoke the mechanism built to compel and track a response. The second is from formal designation to completed action – the process, once triggered, actually finishing before the underlying condition has moved past the point where the action still matters. Authority, on this record, was present at every stage. What varied was whether, and how quickly, that authority was formally invoked, and whether the process it triggered could complete in time.

Complication and What This Does Not Establish

It would overstate this evidence to read it as proof that different or faster supervisory action would certainly have prevented either bank's failure. None of the reviewing institutions make that claim, and this piece does not make it either. The Federal Reserve's own report attributes the delay to "a complex combination of many factors," including the pace of the bank's growth relative to a supervisory framework built around asset-size tailoring, and a cultural emphasis on "consensus and the continued accumulation of evidence" that the report itself identifies as part of the problem rather than a neutral description of careful process.

It would also overstate the evidence to treat 2020-2022 uncertainty as though it were obviously resolved at the time. The interest rate environment, the bank's unusual deposit concentration, and its transition between supervisory categories as it grew were all genuinely evolving conditions, not facts sitting fully formed and ignored. The Federal Reserve's own report situates its criticism specifically at the point where deficiencies had already been identified through the exam process – it is examining the space between recognition and formalization, not asserting that the underlying risk was obvious from the outset to anyone, including supervisors, before examiners themselves identified it.

Finally, it is necessary to hold apart the regulator and the firm it supervises. A bank's own board and management are held, by these same reports, to a different and in the Federal Reserve's own words more direct standard of failure – the report's very first finding is that "Silicon Valley Bank's board of directors and management failed to manage their risks." The supervisory findings examined in this piece describe a distinct institutional layer, operating under different authority, different information access, and a different accountability structure than the firm itself. The pattern identified here – information existing without producing timely completed action – appears at this supervisory layer in a form that resembles, but is not identical to, similar patterns that can appear inside the firms these regulators oversee. The resemblance is worth naming. It is not evidence that regulators and regulated firms are institutionally equivalent, and this piece does not treat them as such.

The Obraval Interpretation

The aperture through which this piece examines governance is wider than usual – an institution examining not a company, but itself, and specifically its own capacity to convert what it knows into what it does. The underlying questions, however, are the same ones Obraval asks at any scale: what was visible, what remained uncertain, who held the authority to act, and where – precisely – did the space between recognizing a condition and completing action on it actually open.

What this specific evidence suggests, read at this larger institutional scale, is that the space between information and action is rarely a single gap. It is more often a sequence of discrete transitions – recognition to formalization, formalization to completion – each of which can independently stall, each of which can do so even where authority to act was never in question, and each of which a sufficiently rigorous self-review can identify with real precision, after the fact, using the institution's own records.

The Executive Implication

For any leadership team examining a condition already visible inside its own organization – not a supervisory institution's condition, but any organization's own – this evidence suggests a narrower and more useful question than "do we know about this" or "do we have the authority to act." The more precise questions are these: has this recognized condition actually been formally escalated, or is it still being handled through informal, undocumented channels that carry no compelled timeline? And if it has been formally escalated, is the process that escalation triggers actually completing – or has the decision to escalate become, itself, a stopping point rather than a step toward resolution?

Close

An institution capable of examining its own governance failures is, in that specific respect, further along than one that cannot. The Federal Reserve's and FDIC's own reviews do not exonerate the supervisory function; they document, in the institutions' own words, a real and specific gap between recognizing a condition and finishing action on it – a gap independently corroborated, before and after these particular failures, by a separate reviewing body examining the same underlying procedures.

What remains open, and what this piece does not resolve, is whether that gap – identified here at the scale of federal bank supervision – is a condition specific to this supervisory apparatus, or a more general feature of any institution, at any scale, that must convert recognized conditions into completed action before those conditions move past the point where action still matters. The aperture can widen or narrow. The question underneath it does not change.