Why Do Banks Fail?

Driven by private credit, financial technology, cryptocurrency, and artificial intelligence, the U.S. banking system is changing rapidly. While the spring 2023 banking panic now feels distant, its lessons remain critical. A recent Federal Reserve Bank of New York white paper, reviewing data from 1863 (before FDIC insurance) to the spring 2023 banking panic, argues that bank failures are highly predictable and stem primarily from two drivers: (i) Solvency and (ii) Liquidity. Our data supports this.  

For simplicity, we measure (i) solvency, using adjusted tangible common equity as a proxy (adjusting for unrecognized bond and loan losses), and (ii) liquidity risk by the proportion of non-interest-bearing and uninsured deposits to total deposits. The table below, of notably stressed banking institutions in the spring of 2023, provides a summary of their solvency, liquidity, and status.

Status of Notably Stressed Banks in the Aftermath of March 2023 Failures

(Banks Sorted by Assets in Millions)

Sources: S&P CapIQ, Klaros Capital Analysis, Federal Reserve Bank of New York, Staff Reports, Failing Banks No. 1117 Revised June 2025.. NIB = non-interest-bearing. Adjusted tangible common equity (“TCE”) reflects unrecognized after-tax securities portfolio (held-to-maturity) and loan fair value marks.

At Klaros, we believe bank failures are a last resort for the FDIC. Only if the agency has no option will it step in and allow a bank to fail. Even if a bank is facing insolvency, as long as it is liquid, the FDIC can, and in our experience, does, play for time, delaying intervention for as long as possible while seeking potential buyers. If a bank is solvent but faces a liquidity crunch, the Fed discount window and similar back-up sources of liquidity can similarly permit the FDIC to buy time. Playing for time, often referred to as “forbearance” by bank nerds like us, can be controversial, as healthy banks claim that strict regulatory capital, liquidity, and solvency rules are not being applied evenly. But only when a bank faces a combination of both solvency and liquidity simultaneously is failure a likely outcome. 

Despite the rapid pace of deposit outflows during the 2023 cycle, the swift actions by the Federal Reserve and Treasury contained the damage, resulting in only four bank failures. While much has changed since I was at the Federal Deposit Insurance Corporation (FDIC) leading resolutions for large bank failures, the fundamentals of why banks fail remain the same. The evidence clearly demonstrates that bank failures occur when insolvency is combined with liquidity risk, while banks that are solvent and maintain low liquidity risk generally do not fail.