As banks grow, tiered regulation places them in stricter categories to address their greater risks. In new research, Pedro M. Batista finds that transition periods between tiers might delay when some higher standards become mandatory, potentially leaving growing banks exposed while known risks remain unresolved, as happened with Silicon Valley Bank. Federal Reserve reforms should address this gap between growing risks and delayed safeguards.
Silicon Valley Bank (SVB) nearly tripled in size between 2019 and 2021. The Government Accountability Office found that its assets grew by 198 percent, compared with median growth of 33 percent among 19 peer banks. As the bank expanded, it moved into the Fed’s supervisory group for larger institutions and reached the threshold for tougher safety requirements. But rules on transitioning between regulatory tiers delayed when SVB had to comply with some of those requirements by several months to several years. The bank could therefore keep operating without all the safeguards intended for its larger size, even though its existing systems for controlling risk had already fallen behind its growth. SVB collapsed on March 10, 2023, after a run on deposits following its announcement two days before that it had sold securities at a loss to cover withdrawals.
The Fed is now reconsidering how it supervises such transitioning banks. In her October 14 speech, Vice Chair for Supervision Michelle Bowman argued that supervisors had recognized SVB’s vulnerabilities but failed to act effectively. She called for clearer authority to make decisions and accountability for acting in time. Her October 6 announcement also set out plans to review the size cutoffs used to group banks for supervision and the rules that adjust requirements to their size and risk. Those reforms should address both how quickly new safety requirements apply and what supervisors do about risks they have already identified.
In my article, “Escalation without enforcement: hidden grace periods in tiered bank supervision,” I examine how these delays developed during SVB’s move into large-bank supervision. The additional time was not limited to regulatory deadlines. Supervisors also postponed formal ratings of the bank’s condition and management, and treated some weaknesses as problems to improve gradually. Such decisions can delay demands for corrective action even when officials have already identified risks and have powers to address them.
Banks may reasonably need time to install reporting systems or recruit qualified staff. But before giving a bank more time to make those improvements, supervisors should assess whether the risks they have already identified can be adequately controlled in the meantime. The need to improve its systems does not, by itself, justify allowing a bank to keep operating with known weaknesses. If those weaknesses remain unresolved, further growth can make the risks harder to contain.
At SVB, the move into more intensive supervision did not immediately bring a new supervisory team. The Fed’s 2023 supervisory review records that the bank formally entered its large-bank supervisory group in February 2021. But the incoming team was not due to take over until the existing regional team completed its annual supervisory cycle in July.
Transition rules also delayed when SVB had to comply with some of the stricter safety requirements for larger banks. Its parent company crossed the relevant $100 billion average-assets threshold in June 2021. But the new requirements to test its ability to meet heavy demands for cash and maintain a reserve of assets it could quickly turn into cash did not apply until July 2022. Its first Fed-run capital stress test, designed to assess its ability to absorb losses, was scheduled for 2024, after the bank’s failure in March 2023. The transition timetable therefore allowed SVB to continue operating before it had to meet safeguards intended for its larger size, even though supervisors had already identified weaknesses in its existing risk controls.
The Fed also delayed its formal assessment of how SVB was managed. Such ratings matter because a poor assessment can lead to stronger demands for improvements. According to the Fed’s review of critical risk areas, Federal Reserve Board staff allowed supervisors to postpone the parent company’s 2021 large-bank rating while they gathered more evidence to support a lower rating for the way its board and senior management oversaw risk. Initial ratings were presented to the company’s board in July 2022 and were formally issued in August.
Even after SVB’s transition period for the new liquidity requirements ended, supervisors continued to allow time for improvements. The bank’s own tests showed that it could not turn enough assets into cash to meet projected demands during 30 days of financial stress. Management and supervisors treated this as a problem with its arrangements for raising cash that could be corrected over time. But the Fed’s review identified an apparent breach of the rule and concluded that supervisors should have required immediate action to cover the shortfall and lowered the bank’s liquidity rating. The emphasis on gradual improvement during SVB’s move to a higher tier may have encouraged supervisors to keep granting time after compliance became mandatory. This is a possible explanation for the delay, although the review does not establish that connection.
Shortening grace periods would require banks to meet new standards sooner. But it would not resolve delays that occur after those standards already apply. Reform must also address decisions to postpone supervisory ratings or allow more time to correct weaknesses, especially where the bank’s condition calls for immediate action.
Bowman presents the preliminary independent review she commissioned as rejecting claims that changes to the laws and regulations governing banks of different sizes caused supervisors’ failure to act. Her emphasis on their failure to use existing powers addresses an important part of the problem. But the fact that supervisors could have acted does not show that giving a bank time to adjust played no part in their decision to wait. Officials may have authority to require corrective action yet continue waiting for the bank to complete planned improvements or waiting for another examination to strengthen the case for intervention.
These decisions to wait can arise from the practical demands of supervision. Bank examiners work with incomplete information and limited time, and must coordinate with colleagues and defend judgments that banks may challenge. A newly assigned team may want to investigate further before intervening, while a downgrade may require supervisors to explain why their assessment differs from an earlier favorable rating. But if supervisors wait for further evidence and agreement before requiring corrective action, weaknesses they have already identified may remain unaddressed throughout the review.
A transition can reinforce these pressures by making further delay seem reasonable. Management is preparing for new requirements, while the incoming supervisory team is still completing its assessment. Officials may therefore decide to wait for this work to finish before requiring corrective action. Acting sooner would require them to defend an intervention that the bank might challenge, whereas the consequences of waiting remain uncertain and responsibility for them may be spread across several decision-makers.
The risks at SVB had been developing for years. In a recent ProMarket article, Niccolò Comati draws on research with Eric Rosengren to trace the bank’s reliance on largely uninsured deposits from similar businesses and its investments in long-term securities. Many depositors stood to lose money if the bank failed and could withdraw at short notice, while rising interest rates could reduce the value of its investments. Comati argues that supervisors responded too much to losses already recognized and too little to the risks that preceded them. He also identifies transition provisions as part of the problem. Earlier detection offers limited protection if supervisors then postpone requiring the bank to address those risks.
Rapid growth gives supervisors a further reason to question how much adjustment time a bank should receive. In “Failing Banks,” published this year in the Quarterly Journal of Economics, Sergio Correia, Stephan Luck, and Emil Verner study American bank failures across roughly 160 years. Their account of the findings identifies rapid asset growth as a recurring precursor to failure. The pattern often involves aggressive lending followed by losses, a shrinking capital cushion to absorb further losses, and greater reliance on funding that can disappear when confidence weakens.
This evidence does not establish that transition periods cause failures. SVB’s rapid growth came largely from a surge of deposits that it invested in securities, while the typical pattern in the study involved an expansion of lending. But the findings support treating rapid growth as a warning that warrants closer scrutiny. Consider two banks of similar size: one has expanded gradually while investing in risk management; the other has doubled quickly, relying on depositors whose similar businesses may lead them to withdraw money at the same time, and systems built for a smaller institution. Their size alone cannot tell supervisors whether the same adjustment period would be appropriate for both.
Extra time to address weaknesses may also give a bank a competitive advantage. A bank that builds adequate controls before expanding pays for them upfront. A competitor allowed to grow first and improve its controls later can expand without bearing those costs immediately. The extent of any effect on competition would need to be established through evidence. But regulators deciding how much adjustment time to allow should consider the possible disadvantage to banks that paid to prepare before growing.
Supervisors should start preparing for more demanding oversight when a bank’s growth makes that need foreseeable. Its rate of expansion, dependence on similar depositors, and changes in its business can help identify the need for closer attention before it crosses a size threshold. The incoming supervisory team could examine the bank and its unresolved weaknesses before the formal handover. Management could also be asked to explain how its systems for controlling risk would keep pace with its planned expansion.
This preparation should not amount to enforcing new rules before their published compliance dates. If those dates allow too much time, agencies should change the regulations through the required rulemaking process. They should not impose an earlier deadline through private instructions to an individual bank. But existing requirements and powers to address unsafe banking practices remain available during the adjustment. Where those powers justify corrective action, the fact that a bank is preparing for other requirements should not excuse delay.
The time allowed for improvements should depend on whether the bank can control its risks while the work is done. If serious weaknesses cannot be corrected immediately, supervisors should assess what protection is needed in the meantime. Depending on the evidence and their legal authority, they might require the bank to reduce excessive dependence on one type of investment, arrange reliable access to cash in an emergency, or limit the activity creating the risk until it can manage that activity adequately.
Every decision to allow more time should record why waiting is justified, which official approved it, how the risks will be controlled, and when the decision will be reviewed. That should apply to waivers, extended deadlines, and decisions to postpone action on a serious concern. When another supervisory team takes over, it should receive the history of unresolved weaknesses and earlier decisions, so that the handover does not become another reason to defer action.
The Government Accountability Office has called for clearer, measurable procedures for escalating supervisory action: moving to stronger measures when weaknesses persist. Applying that approach to decisions about additional time would help reviewers see whether a series of apparently reasonable extensions had left the same weakness unresolved for too long. Further time might still be justified, but the decision should account for the full period of delay and the risks accumulated during it.
The Fed’s review should therefore address more than the size at which a bank enters a new category. It should also ask what protects depositors and the financial system while the bank is given time to meet higher standards. When existing controls are already inadequate, supervisors should use their existing powers to contain those risks while the bank makes the necessary improvements.
Author’s Disclosures: The author reports no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
