S6 Ep52: Making banking safe

Episode Summary

Executive Summary: The episode examines why Silicon Valley Bank, Signature Bank, and First Republic failed despite post-crisis reforms, arguing that weak risk management, inadequate supervision, poor stress testing, and opaque accounting all contributed. The guests advocate tighter rules, larger capital and liquidity buffers, mark-to-market accounting, and better stress tests to make banking meaningfully safer.

Main Topics: Why SVB, Signature, and First Republic failed (Priority: 5/5): The discussion traces the collapses to a combination of bank-level risk failures, supervisory inaction, and resolution failures, with SVB used as the clearest case study. Stress testing failures (Priority: 5/5): The guests argue that stress tests were either not applied during rapid growth or failed to incorporate interest-rate risk, despite rising rates being the key vulnerability in 2021–2022. Uninsured deposits and bank runs (Priority: 4/5): The episode explains how concentrated, uninsured depositor bases made runs more likely and how the FDIC’s post-2008 resolution practice may have created expectations of broader protection. Accounting transparency and unrealized losses (Priority: 5/5): The speakers contend that hidden unrealized losses in securities portfolios obscured bank fragility and that mark-to-market accounting would have revealed problems earlier. Rules versus discretion in regulation (Priority: 4/5): The guests favor stronger rules-based regulation and larger buffers, while still allowing discretion, arguing that supervisors currently rely too much on judgment. Capital, liquidity, and risk-taking incentives (Priority: 4/5): They argue that higher capital and liquidity requirements increase owners’ skin in the game, reduce default risk, and ultimately support a safer banking system. Limits of making banking 'safe' (Priority: 3/5): The conversation closes by acknowledging that bank failures can never be eliminated entirely, but that regulation should reduce fragility and the need for extraordinary government intervention.

Key Arguments: SVB did not fail for one reason; it reflected multiple breakdowns: internal risk management, market discipline, supervisory enforcement, and resolution authority all failed to act in time. Signature and First Republic shared important vulnerabilities with SVB, especially weak capitalization, unrealized losses, and concentrated funding bases. Stress testing missed the crisis because some banks were not tested during rapid growth and because the tests ignored interest-rate risk in a rising-rate environment. Uninsured depositors were not sufficiently monitoring bank health, while FDIC resolution practices after 2008 made many of them effectively expect full protection. Banks became vulnerable because they paired runnable liabilities with risky assets, a combination that makes them highly exposed when asset values fall. Quarterly disclosures already contained clues—especially unrealized losses—that could have alerted investors and supervisors earlier. Mark-to-market accounting would improve transparency, push earlier recapitalization or resolution, and reduce the chance that hidden losses become sudden panic. More capital and liquidity should reduce, not increase, risky behavior by giving owners and managers more skin in the game and lowering default incentives. Rules should be strengthened so that supervision does not depend so heavily on discretionary judgment, which has repeatedly failed in practice.

Data Points: Silicon Valley Bank failure date: 10 March 2023 - Steve Caketti identifies the date of SVB’s collapse. First Republic takeover timing: During the weekend after the SVB and Signature failures - The transcript notes regulators spent the weekend finding a buyer before JPMorgan Chase took over First Republic. Discussion paper number: 18302 - The paper 'Making Banking Safe' is identified as CEPR discussion paper 18302. Time reference for Fed stress-test omission: 2021–2022 - Kim Schoenholz says interest-rate risk was the key vulnerability in this rising-rate period, yet stress tests did not include it. Historical reference to post-Lehman resolution practice: Since 2008 - The FDIC has often used resolutions that protect all deposits, not only insured ones. Estimated solvency of SVB: Barely solvent in mid-2022 - Steve says a more accurate accounting suggests SVB was just barely solvent in the middle of 2022. Lead time before failure: About 8 months - Steve notes SVB failed in March 2023 after appearing barely solvent in mid-2022.

Pivotal Quotes: "There were really four failures, and there's quite a lot of blame to go around." — Steve Caketti: Summarizing the causes of SVB’s collapse across the bank, markets, supervisors, and resolution authorities. "These banks did both. They took the runnable liabilities and put them into risky assets." — Steve Caketti: Explaining why funding structure and asset choice made the banks vulnerable to runs. "We’re relying too heavily on it, and we’ve seen obvious supervisory failures of the kind that we just discussed." — Kim Schoenholz: Arguing for stronger rules-based regulation and less dependence on supervisory discretion.

Implications: The episode argues for tougher bank rules, stronger capital/liquidity buffers, and transparent accounting to reduce hidden fragility. For regulators and depositors, the lesson is that safety depends less on promises and more on visible loss recognition and enforceable safeguards.