Silicon Valley Bank — interest-rate-risk and uninsured-deposit failure

1. Episode summary

Silicon Valley Bank (SVB) was the commercial-banking subsidiary of SVB Financial Group and the primary banker to roughly half of the venture-backed technology and life-sciences companies in the United States. Between 2020 and 2021, pandemic-era monetary accommodation and a venture-capital funding boom drove SVB's deposit base from about $62 billion to roughly $189 billion, overwhelmingly in balances above the FDIC insurance limit. Rather than hold the inflows short, SVB deployed about $91 billion into a held-to-maturity (HTM) portfolio of long-dated agency mortgage-backed securities and Treasuries, plus $26 billion in available-for-sale (AFS) securities, while unwinding most of the interest-rate swaps hedging the AFS book. When the Federal Reserve raised the federal-funds target by roughly 450 basis points between March 2022 and March 2023, the bank accumulated unrealised losses in its HTM portfolio exceeding $15 billion at year-end 2022 — about 89 percent of its common equity tier 1 capital — while venture-backed clients, no longer receiving fresh funding, drew down deposits. On 8 March 2023 SVB announced a $21 billion AFS sale at a $1.8 billion loss and a planned $2.25 billion capital raise; the announcement triggered an approximately $42 billion deposit outflow on 9 March, largely driven by venture-capital firms instructing portfolio companies over social media and private channels. The California Department of Financial Protection and Innovation closed the bank on 10 March 2023. The strategic question the episode turned on was whether a deposit-funded bank facing a visible rates regime-change would actively manage duration and hedging or let carry economics run uncorrected.

2. Sources

Primary:

1. Board of Governors of the Federal Reserve System, "Review of the Federal Reserve's Supervision and Regulation of Silicon Valley Bank" (Barr Report), 28 April 2023 — full report and Key Takeaways; sections on interest-rate-risk management, HTM portfolio, CRO vacancy, unaddressed supervisory findings.
2. Office of Inspector General of the Federal Reserve and CFPB, "Material Loss Review of Silicon Valley Bank," report 2023-SR-B-013, 25 September 2023.
3. SVB Financial Group, Form 10-K for fiscal year ended 31 December 2022 (filed with SEC February 2023), investment-portfolio footnotes on HTM/AFS classification and unrealized losses.
4. Basel Committee on Banking Supervision, "Report on the 2023 banking turmoil," Bank for International Settlements, October 2023, sections on SVB.

Secondary (with justification):

1. Federal Reserve Bank of St. Louis, "Interest Rate Risk, Bank Runs and Silicon Valley Bank," Regional Economist, May 2023 — peer-reviewed staff analysis synthesising balance-sheet evidence with rates data.
2. "Liquidity Risk (Mis)Management: The Failure of Silicon Valley Bank and the Liability-Driven Investment Episode in UK Gilt Markets," MIT Sloan / MIT Golub Center for Finance and Policy working paper, 2023 — academic post-mortem aggregating regulatory filings and market data.
3. Jha, Akhtar, Rana, "The Swift Rise and Sudden Fall: Examining the Collapse of Silicon Valley Bank," Global Business Review / SAGE, 2025 — peer-reviewed retrospective.
4. Nicolas Véron and others, INSEAD Knowledge, "Risks and Regulations: The Silicon Valley Bank Collapse," 2023 — policy-oriented synthesis of regulatory tailoring and supervisory record.

Tertiary (flagged):

1. "Collapse of Silicon Valley Bank," Wikipedia, 2023–2024 — used only as cross-reference index for dates and deposit-flow figures; not load-bearing.

Sources added during §4 research: 5. Banking Dive / Legal Dive, "SVB had no risk chief through much of 2022, proxy statement shows," 13 March 2023 — contemporaneous reporting on Laura Izurieta's departure (April 2022 step-down, October 2022 formal exit) and Kim Olson appointment (announced 4 January 2023); used for CRO vacancy dates and Structure/Processes evidence. 6. Fortune, "Silicon Valley Bank had no official chief risk officer for 8 months while the VC market was spiraling," 10 March 2023 — corroborating account of the eight-month CRO gap; used for Structure evidence. 7. Federal Reserve Board, "Evolution of Silicon Valley Bank," federalreserve.gov/publications, April 2023 — Fed reconstruction of SVBFG's asset-growth trajectory (271% from year-end 2018 to year-end 2021) and concentration; used for Direction and Structure evidence. 8. Federal Reserve Board, "Supervision of SVBFG by Critical Risk Areas," federalreserve.gov/publications, April 2023 (companion section to Barr Report) — detailed findings on risk appetite statement limited to NII not EVE, Internal Liquidity Stress Test (ILST) failures from July 2022, and management's shift to less conservative stress-testing assumptions; used for Processes evidence. 9. Bloomberg, "Founders Fund Advises Companies to Withdraw Money From SVB," 9 March 2023; and "Thiel's Founders Fund Withdrew Millions From Silicon Valley Bank," 11 March 2023 — contemporaneous reporting on Founders Fund withdrawing its own holdings on the morning of 9 March and advising portfolio companies to do the same; used for Act/Culture evidence on the deposit-run mechanism. 10. Harvard Business Review, "Silicon Valley Bank's Focus on Startups Was a Double-Edged Sword," March 2023 — synthesis of SVB's structural identity as the VC ecosystem's bank and the cultural assumptions that followed from that positioning; used for Direction and Culture evidence. 11. NACDO / National Association of Corporate Directors, "Silicon Valley Bank: Key Takeaways and Questions for Board Risk Oversight," 2023 — analysis noting the Risk Committee had no chair in 2022 and no members with direct risk management experience; used for Structure evidence.

3. OTA narrative

Observe. The observation surface SVB needed to read was ordinary for a deposit-funded bank of its size: the FOMC's shift from a zero-lower-bound stance to an active tightening cycle, the mark-to-market impact of that shift on a long-duration securities book, the proportion of deposits that were uninsured and concentrated in a single funding-cycle-sensitive client segment, and the rate of venture-funding inflow that sustained those deposits. All of these signals were publicly visible and reported in the bank's own regulatory filings; the OIG Material Loss Review and the Barr Report both document that internal risk-limit breaches on interest-rate-risk metrics were occurring from 2021 onward, and that supervisors had identified thirty-one unaddressed safety-and-soundness findings by the time of failure. Observe was not a root cause in the failing sense — the apparatus produced the signals. Where Observe is partially implicated is at the governance layer: the Chief Risk Officer role was vacant from roughly April 2022 through January 2023, so the observation signals reached a risk function that had no permanent head during the window in which rates rose most sharply. Observe is therefore best read as a transmission step that was degraded but functional; it carried the signal to a reasoning layer that failed to act on it.

Think. The reasoning failure is the root cause. Management interpreted the post-2020 deposit surge as stable low-cost funding that could be termed out into long-duration HTM securities for carry, and interpreted the rising-rate environment of 2022 primarily as a threat to short-run net-interest-margin reporting rather than as a threat to the economic value of equity and to liquidity under a stress scenario. The bank actively unwound approximately $11 billion of interest-rate swaps in the first half of 2022, realising $517 million in gains, leaving only about $563 million of swap protection on the AFS book by year-end — a reasoning move that converted a hedged position into an unhedged one precisely as the hedging need intensified. The correct framework — duration gap management, stress-testing a rate shock against HTM marks, modelling deposit runoff under uninsured-deposit concentration — was standard for the large-regional-bank peer group and accessible in supervisory guidance and the bank's own capital-planning documents. The reasoning failure was therefore an Easy-Wrong Think: the interpretive task was routine for the Archetype peer group, and the bank chose the interpretation that optimised reported carry at the expense of economic resilience.

Act. Execution in 8–10 March 2023 was poorly timed and poorly communicated, but the outcome was already determined by the preceding Think failure. The decision to announce a $21 billion AFS sale, a $1.8 billion realised loss and a capital raise in a single press release on 8 March, without a pre-placed anchor investor and without coordinated regulator-and-client messaging, accelerated the deposit run that closed the bank within 40 hours. Social-media-mediated coordination among venture-capital firms advising portfolio companies turned what in a slower information environment might have been a multi-day outflow into a same-day $42 billion withdrawal. Act is a contributing-but-not-primary root-cause phase: classified Almost-wrong at the easy end of the task-difficulty axis, because crisis-communications sequencing for a distressed capital raise is a routine capability for a bank of this size and was not executed to that standard. The deeper causal weight, however, sits in Think — by the time Act ran, the balance sheet was already carrying the unrealised-loss and uninsured-deposit exposures that made a run rational for depositors.

---

stage: 4 case_id: F-021 case_title: Silicon Valley Bank — interest-rate-risk and uninsured-deposit failure prepared: 2026-06-04 researcher: Researcher agent (T-368 re-rating) methodology_version: METHODOLOGY-ota-scoring-v4.md (v4.4, locked 2026-06-03)

4. Modality evidence

Direction. SVB's strategic direction was defined by a specific, dated, and attributable set of choices that concentrated the bank's business model in a single client segment and then deployed the resulting deposit windfall in a manner optimised for carry income rather than balance-sheet resilience. The foundational directional choice — becoming the primary commercial bank to venture-backed technology and life-sciences companies — was not itself a failure; it had been SVB's identity for four decades and had produced a dominant market position (the Barr Report and the Fed's "Evolution of Silicon Valley Bank" both document that the bank served roughly half of all US venture-backed companies). The directional failure arrived when management responded to the 2020–2021 deposit surge — driven entirely by the same VC ecosystem concentration — by making a further directional bet: deploying approximately $91 billion into long-dated held-to-maturity agency securities, classifying the majority as HTM to avoid mark-to-market capital volatility, while simultaneously choosing not to hedge the resulting duration exposure through the rising-rate cycle. This was a directional choice in the methodology's sense: specific (a deliberate HTM classification decision and hedge unwind, documented in the 10-K and the OIG Material Loss Review), datable (the HTM portfolio was assembled through 2020–2021; the swap unwind occurred in the first half of 2022), and attributable to identifiable management and board actors (SVB Financial Group CEO Greg Becker and the board's Risk Committee, which approved the investment policy statement). The HBR post-mortem and the Barr Report both characterise the deposit-concentration-plus-duration-concentration strategy as an integrated directional posture, not a series of incremental operating decisions. Direction also surfaces in the framing of what threat the rate cycle represented: management's risk appetite statement, per the Barr Report, was calibrated to net-interest-income (NII) sensitivity over a 100 bps ramp, explicitly excluding economic value of equity (EVE) — a directional choice about what the bank was trying to protect that shaped every downstream risk-management decision.

Structure. Three structural features created the conditions in which the directional failure could proceed unchecked. First, the Chief Risk Officer position was vacant from Laura Izurieta's step-down in April 2022 (formal departure October 2022) through Kim Olson's appointment on 4 January 2023 — an eight-month gap across the period in which the Federal Reserve was executing its fastest rate-tightening cycle in four decades (Banking Dive and Fortune, March 2023; Barr Report). The CRO vacancy meant the risk-function escalation path to the board was structurally degraded precisely when it was most needed. Second, the Risk Committee of the SVB Financial Group board had no permanent chair in 2022 and included no members with direct risk-management experience (NACDO, 2023), a structural gap that the Barr Report directly links to the board's failure to receive or act on EVE and liquidity stress-test findings. Third, the risk appetite framework was architecturally incomplete: the board-approved risk appetite statement covered NII sensitivity but not EVE, meaning the structural mechanism that would have routed the EVE breach data to the board simply did not exist as a formal governance channel (Barr Report; OIG Material Loss Review). The consequence, documented in the Barr Report, was that there is no evidence the full board was aware that EVE limits were being breached for years while the HTM portfolio accumulated. This is a structural failure in the methodology's precise sense: the formal channels, reporting lines, and governance mechanisms did not allow the right information to reach the right people — not because of cultural suppression, but because the channels were absent or mis-scoped.

Processes. The bank's asset-liability management processes failed at three identifiable operational nodes. First, the internal liquidity stress-testing process: from July 2022, when SVB first became subject to enhanced prudential standards under Regulation YY after crossing the $100 billion threshold, the bank repeatedly failed its own Internal Liquidity Stress Test (ILST), and management responded by switching to less conservative stress-testing assumptions rather than restructuring the portfolio — a process-level manipulation documented in the Barr Report and the Fed's "Supervision of SVBFG by Critical Risk Areas." Second, the interest-rate-risk management process: the decision to unwind approximately $11 billion of interest-rate swaps in the first half of 2022, realising $517 million in gains and leaving only $563 million of swap protection on the AFS book by year-end, was an action taken through the ALCO (asset-liability committee) process at precisely the moment the hedging need was intensifying (OIG Material Loss Review; SVB 10-K FY2022; GFMI / Finalyse analyses). The correct ALM process — running duration-gap analysis, stress-testing a rate shock against HTM marks, and modelling deposit runoff under uninsured-deposit concentration — was standard for the large-regional-bank peer group and was specified in supervisory guidance and the bank's own capital-planning documents (Barr Report; Basel Committee report, October 2023). Third, the supervisory-findings escalation process: the Barr Report documents thirty-one unaddressed safety-and-soundness supervisory findings at the time of failure — triple the peer-bank average — indicating that the internal process for receiving, tracking, and remediating regulatory observations had broken down systematically rather than in any single instance. The OIG Material Loss Review confirms that supervisors had identified interest-rate risk deficiencies in the 2020, 2021, and 2022 CAMELS exams and communicated them as advisories and verbal observations that produced no corrective action before failure.

Capability. SVB possessed the technical capabilities standard for a large regional bank: it had access to duration-gap modelling tools, ILST infrastructure, interest-rate-swap instruments (it had used them and unwound them), and a treasury function capable of executing large-scale securities purchases and sales. The 10-K disclosures and the OIG report confirm that the bank generated the HTM unrealised-loss figures and knew their magnitude — the $15.1 billion unrealised loss on the HTM book, representing roughly 89 percent of CET1 capital, appeared in the year-end 2022 regulatory filings (SVB 10-K FY2022). This indicates the data-generation and financial-modelling capabilities were present; the numbers were computed and disclosed. The capability gap the episode surfaces is narrower and more specific: SVB lacked the institutional risk-management capacity — experienced senior risk leadership on the bench, board-level risk expertise, and a functioning escalation culture between the first and second lines of defence — to convert correct modelling outputs into corrective action. This is better characterised as a capability gap in the risk-governance domain rather than in financial engineering. Notably, the eight-month CRO vacancy (Structure) interacts with Capability here: the absence of a permanent CRO was both a structural gap (no one held the role) and a capability signal (the organisation could not promptly fill a critical risk-leadership position during a stress period). The Capability contribution is real but thinner than Structure or Processes as a standalone explanation, because the requisite technical capabilities — swaps, duration modelling, stress testing — were demonstrably available and partially deployed.

Scoring note (zero-modality rationale): the capability described in this subsection is recorded at zero per cent in the modality weights on the rationale of insufficient causal weight — the §4 evidence establishes that Silicon Valley Bank possessed the technical and operational capability the situation required; the failure mechanism was located in Direction, Structure, Processes rather than in a capability gap. The capability is acknowledged as present in the narrative but does not carry standalone weight in the failure attribution. Categorisation under METHODOLOGY-ota-scoring-v4.md §5 "Zero-modality rationale rule": insufficient causal weight.

Culture. The cultural dimension of the SVB failure is partially evidenced but requires careful boundary-testing against Structure and Processes. The clearest cultural signal is the management response to adverse internal data: when stress tests produced uncomfortable results, management switched to less conservative assumptions (Barr Report; Fed "Supervision of SVBFG" companion section) — this is a Culture manifestation in the methodology's sense, because the formal machinery (stress testing) existed and was run, but management subverted it rather than acting on its outputs. A second cultural marker is the bank's deep integration into the VC ecosystem and the identity assumptions that followed: the HBR analysis and contemporaneous reporting describe SVB's culture as defined by closeness to the startup community, an optimism about VC-cycle permanence, and a disposition to prioritise founder-and-investor relationships over conservative balance-sheet management — norms that appear to have shaped the willingness to term out deposits into long-duration carry positions without the hedges that a more conventionally risk-anchored culture would have maintained. The deposit-run mechanism on 9 March 2023 also carries a cultural dimension: the concentration of SVB's client base in a tightly networked VC community meant that when Founders Fund advised portfolio companies to withdraw (Bloomberg, 9 March 2023) and the signal amplified across private channels and social media, the cultural cohesion of the client community — which had been a business-model asset — became a systemic liquidity liability. This is partly exogenous to SVB's internal culture, but it reflects the bank's directional choice to build a client culture of concentrated VC dependency. The cultural evidence is thin on named internal actors and documented dissent-suppression (the primary evidence is regulatory commentary rather than whistleblower or congressional-record testimony of the type present in the Boeing case), which should be flagged as a thin-evidence condition: the Culture modality is supported but the primary-source depth for internal-norm dynamics is lower than for Structure and Processes.

Scoring note (zero-modality rationale): the cultural evidence in this subsection is acknowledged in the narrative but is not load-bearing for the strategic failure causation of the episode — the §4 evidence itself characterises it as partially evidenced but requires careful boundary-testing compared with the modalities that carried the failure causation (Direction, Structure, Processes). Culture is therefore recorded at zero per cent on the rationale of modality acknowledged in narrative but not load-bearing for the strategic value created in the episode. Categorisation under METHODOLOGY-ota-scoring-v4.md §5: modality acknowledged in narrative but not load-bearing.

---