Copy of The Great Misdiagnosis: Rethinking the Banking Crisis of 2023

When Silicon Valley Bank collapsed in March 2023, followed swiftly by Signature Bank and the First Republic, financial observers rushed to declare the culprits: interest rate hikes, uninsured deposits, and social media-accelerated bank runs. This conventional narrative has dominated policy discussions and academic analyses alike. However, according to economists Kelly and Rose, whose groundbreaking new working paper, 'Rushing to Judgment and the Banking Crisis of 2023, ' challenges established thinking, we have been looking at the crisis through the wrong lens.

They argue that the 2023 banking failures were not simply liquidity events triggered by interest rate mismatches but existential crises stemming from deeply flawed business models built around hyper-concentrated exposure to volatile economic sectors. The failed institutions were not victims of sudden panic but of their own strategic choices to serve a narrow clientele in cyclically sensitive industries—primarily venture capital and cryptocurrency.

"The standard narrative misses the forest for the trees," Kelly and Rose write. Their evidence is compelling: SVB's deposit outflows began well before March 2023, with the bank losing substantial funding as the venture capital sector contracted throughout 2022. Similarly, Silvergate Bank experienced a catastrophic deposit flight months earlier when the crypto market collapsed following FTX's November 2022 implosion.

What makes their analysis particularly convincing is the comparison with banks that survived despite similar balance sheet metrics. Numerous institutions carried unrealized securities losses exceeding 150% of equity—worse than SVB's position—yet avoided runs because they maintained diversified deposit bases across multiple economic sectors. The crisis struck with surgical precision, claiming only banks with concentrated exposure to technology and digital assets.

The timing also undermines conventional wisdom. If unrealized losses were the trigger, why did depositors wait until March 2023 when these losses had peaked six months earlier? If uninsured deposits were the key vulnerability, why did institutions like custody banks with similarly high proportions of uninsured deposits remain stable?

Geographic clustering further supports the sectoral contagion thesis. The authors note that 83% of failed banks were headquartered on the West Coast and deeply embedded in the technology ecosystem. When those sectors contracted, their banking partners inevitably followed. This suggests that the crisis was a series of isolated incidents and a systemic problem that affected banks with similar business models and geographical locations.

Kelly and Rose's findings have profound implications for regulatory policy. Current reforms emphasizing higher liquidity buffers and stress testing for interest rate risk address symptoms rather than causes. SVB maintained liquidity coverage ratios exceeding regulatory requirements but collapsed anyway because the fundamental viability of its business model came into question—not just its ability to meet short-term withdrawals. This underscores the urgent need for a shift in regulatory focus.

Equally problematic was regulatory tunnel vision. Bank supervisors cited SVB for interest rate concerns dozens of times yet failed to address its dangerous deposit concentration. While regulators developed specific guidance for crypto-exposed banks, they neglected analogous oversight for venture capital-focused institutions.

The authors also debunk the 'technological panic' hypothesis. This hypothesis claims that modern communication accelerated the bank runs, but the authors argue that this is largely unfounded. They point out that 92% of SVB's outflows came through traditional wire transfers by institutional clients, not panicked retail depositors using mobile apps. The authors argue that unprecedented speed in the bank runs reflected the inevitable culmination of long-developing sectoral contractions rather than a sudden panic accelerated by social media or other modern communication tools.

This crisis reframing challenges us to reconsider banking stability in an era of increasing specialization. As financial institutions pursue niche strategies to differentiate themselves, they may inadvertently create existential vulnerabilities by tying their fortunes to volatile economic sectors.

The experiences of 2023 demonstrate that traditional metrics—capital ratios, liquidity buffers, and asset quality—provide insufficient protection when a bank's fundamental business model is threatened. This calls for a new approach to assessing banking stability. Regulatory frameworks must evolve to evaluate balance sheet risks and the structural vulnerabilities embedded in specialized banking strategies.

The lesson is equally clear for investors and depositors: a bank's client concentration may matter more than its capital ratio. When evaluating financial institutions, we should ask, "How much capital do they have?" and "Whose money do they hold?"

As we navigate an increasingly specialized financial landscape, with banks serving ever-narrower market segments, the 2023 crisis serves as a warning. The most significant banking risks may not come from traditional sources like credit quality or interest rate mismatches but from the business models that differentiate modern financial institutions.

Kelly and Rose's analysis offers a compelling alternative to the hasty judgments that followed the banking failures. By shifting our focus from balance sheet metrics to business model vulnerabilities, they provide a more accurate diagnosis of what happened in 2023 and a roadmap for preventing similar crises in our increasingly specialized financial future.

Reference

Kelly, S., & Rose, J. (2025). Rushing to Judgment and the Banking Crisis 2023 (Working Paper No. 2025-04). Federal Reserve Bank of Chicago. https://doi.org/10.21033/wp-2025-04