Anatomy of a banking crisis

Silicon Valley Bank's failure highlighted weaknesses throughout the regulatory system

Photo by Mariia Shalabaieva on Unsplash

The failure and subsequent rescue of Silicon Valley Bank (SVB) in March 2023 has reignited age-old debates about bank governance, capital and liquidity requirements, and the design of the deposit insurance system. Yet often lost amid the debates about these different regulatory tools is the fact that these tools were designed to work as part of a single system: with the success of each tool contingent on the effectiveness of the others, and the failure of any one tool potentially weakening the resilience of the system as a whole. It’s this system that failed in March 2023, not just SVB.

Dodd-Frank didn’t solve everything

On September 15, 2008, Lehman Brothers filed for bankruptcy and unleashed a financial maelstrom so catastrophic that it became known as the Global Financial Crisis. Millions of Americans lost their jobs, and the U.S. economy experienced its darkest days since the Great Depression. In response, Congress passed the most sweeping piece of financial legislation in generations—the Dodd-Frank Wall Street Reform and Consumer Protection Act—to improve the stability of the financial system. Fifteen years later, it’s clear that the task remains incomplete. 

Earlier this year, SVB experienced a bank run. With assets of roughly $200 billion, SVB was not viewed as a particularly big bank. Nor, despite its high profile and politically connected Silicon Valley clientele, was it viewed as particularly interconnected with the wider financial system or critical to the real economy. Nevertheless, the announcement on Friday, March 10, that SVB was being forced to close its doors triggered widespread panic amongst the bank’s depositors and, in the days that followed, the shareholders and creditors of many other “regional” banks. By Sunday, March 12, the panic was sufficiently threatening that senior officials at the Treasury Department, Federal Reserve, and FDIC felt compelled to invoke break-glass emergency measures to backstop the banking system, including full protection for uninsured depositors. 

This wasn’t supposed to happen—not like this, not so soon after the Global Financial Crisis. Lawmakers and regulators are now debating whether to impose greater accountability on bank executives, revamp deposit insurance, strengthen bank capital regulation, enhance prudential supervision, or make bank resolution more credible and effective. The answer is almost certainly “all of the above,” because a crucial lesson from the failure of SVB and subsequent policy intervention is that a banking crisis does not occur when one thing goes wrong. It occurs when multiple things go wrong. Indeed, bank regulation, supervision, and resolution can be understood as part of a sophisticated system of fail-safes or redundancies. The first component of the system is market discipline. The second is prudential regulation and supervision. The third is deposit insurance and, ultimately, resolution. And these mechanisms are related: the failure of one makes it harder for the next one to do its job. In this write-up, we explain the ways in which these fail-safes are supposed to work and why they failed.

Market discipline exists only in theory

The first line of defense against bank failure is market discipline. Where a bank is poorly managed, we should expect to observe this—at a relatively early stage—in the prevailing market price of both its publicly traded equity and debt. Yet in practice, shareholders may possess weak incentives to ferret out new information about poor management. And where debtholders are insured, or come to expect that policymakers will rescue failed banks and their creditors, they too may have little reason to impose discipline. As a result, bank shareholders and creditors have often come to view their investments as essentially reflecting a free put option: heads they win, tails everyone else loses.

A broadly similar, and even more transparently self-serving, logic can be observed in the seemingly contradictory positions taken by many Silicon Valley entrepreneurs during the March 2023 banking crisis. Indeed, the same billionaire libertarians that had for years fought against strict government regulation were amongst the first to call for immediate and powerful government intervention to rescue SVB, its customers, and ultimately themselves. And this is exactly what the government delivered on Sunday, March 12. 

To be clear, the Treasury Department, Federal Reserve, and FDIC came to the rescue not because they wanted to save the fortunes of billionaire libertarians but rather because there are severe economic consequences that stem from financial panics. Millions of jobs are at stake. Said differently, cutting off your nose to spite your face is not, and never will be, optimal macroeconomic policy. Yet it is precisely because everyone understands this dynamic that the market behaves as if it has a free put option. The first fail-safe exists only in theory, not in reality. This puts a greater burden on the remaining redundancies.

Simply dialing-up regulation will never be enough

Since the adoption of the first Basel Accord in 1988, regulators have focused their energies on setting minimum standards for bank capital. In a nutshell, bank capital includes equity and retained earnings that can absorb losses while a bank is still a going concern, thus providing them with something of a “cushion” in times of severe institutional and systemic stress. 

The logic and limits of this capital cushion were on full display in March as depositors came to the realization that SVB was technically insolvent. At the time, SVB held a large portfolio of U.S. Treasury securities. Under U.S. accounting rules, the bank was not required to mark these securities to market on its balance sheet if its intention was to hold them to maturity. Nevertheless, in a period of steadily rising interest rates, the market value of these securities had experienced a significant decline—so much so that, had SVB been forced to sell them, the resulting losses would have entirely wiped out the bank’s capital cushion. That’s when depositors became jittery, and the run began.

In the wake of SVB, regulators are once again focused on revising capital standards, this time under the auspices of what is known as Basel III Endgame (not to be confused with Avengers: Endgame). The banking industry is furious, as capital requirements are expected to increase significantly. The Bank Policy Institute, a leading bank advocacy group, has even produced a TV ad warning about the consequences of higher capital requirements for “Everyday Americans.” Regardless of what one thinks about the merits of increasing these capital requirements, the reality is that they can only really decrease the probability of bank failure ex ante and, even then, only at the margin. Ex post, once a bank faces an existential threat, and this threat metastasizes into a full-blown bank run, no amount of capital will save it from insolvency. It’s like trying to catch a falling knife.

The limits of bank capital were exposed by the Global Financial Crisis. Since then, regulators have also focused on developing bank liquidity standards as a complement to conventional capital standards. These liquidity standards require banks to hold more liquid assets that can be easily converted into cash to meet heightened redemption requests. The idea is that having stockpiled these more liquid assets will allow a bank to survive for a longer time during periods of institutional or systemic stress—possibly even long enough to outlast the storm. 

Like bank capital requirements, the problem with liquidity requirements is that they were not designed to withstand a death spiral. The run on SVB took out the bank in a matter of days. The magnitude and speed were unprecedented. Accordingly, while bank regulation can and does play a valuable role in promoting bank stability and limiting the probability and impact of destabilizing runs, there are clear limits on the ability of these mechanisms to make banks safe.

Supervisors didn’t sound the alarm

As our experience with bank capital and liquidity standards makes clear, regulation is never perfect. It’s impossible to predict every contingency ex ante and guard against them with bright-line rules. This is why we have bank supervisors who have access to on-the-ground, real-time confidential data about the banks they supervise, along with the authority to compel banks to refrain from risk-taking that they deem to be excessive. In theory, these supervisors should be able to spot troubling developments or trends and then take prompt corrective action designed to protect banks, their depositors, and the wider banking system. 

In practice, however, things aren’t always so simple. In the case of SVB, it turns out that bank supervisors failed to sound the alarm. Why? There are several possible reasons. The first is a lack of information. Yet much of the information needed to identify the problems at SVB was already in the public domain: including rising interest rates, the bank’s large holdings of U.S. Treasury securities, and growing questions about the composition of its depositor base and their propensity to run. The second is supervisory capture. In this respect, it’s worth observing that the CEO of SVB had been on the board of directors of its supervisor, the Federal Reserve Bank of San Francisco, since 2019. The third is leadership. Here, decisions by the Federal Reserve Board, San Francisco Fed, and other banking agencies around things like the intensity of prudential supervision, the allocation of scarce institutional resources, and investments in human, technological, and other institutional capital seem likely to have played, at the very least, a contributing role.

Like bank regulation, supervision is far from perfect. While we can and should continue to learn the lessons from supervisory failures like SVB, the fact that these failures are inevitable provides a strong justification for two further fail-safes: deposit insurance and resolution.

Deposit insurance wasn’t designed for this

If market discipline, capital and liquidity regulation, and supervision fail to prevent a bank’s collapse, bank regulation also ensures that its depositors do not directly bear the resulting costs. The principal mechanism for achieving this objective is deposit insurance, under which the government guarantees depositors’ money up to a specific amount, currently $250,000 per account. Following the paralyzing bank runs of the early 1930s, Congress passed the Banking Act of 1933—better known as the Glass-Steagall Act—which created the FDIC and instituted federal deposit insurance. This deposit insurance has the dual benefits of protecting depositors from losses and, thereby, reducing the probability of destabilizing runs. After all, why rush to withdraw your money from the bank if the government guarantees it in the event of your bank’s failure? 

Unfortunately, deposit insurance delivers neither of these benefits where a bank has a substantial number of uninsured depositors. Imagine a business that deposits $10 million at a bank. Only $250,000 of that is insured; the other $9.75 million are uninsured. In March, over 40 percent of total deposits at U.S. commercial banks were uninsured, according to the FDIC’s Call Reports data. Of the three U.S. banks that failed earlier this year, SVB had an uninsured deposits ratio of 93.9 percent, Signature Bank had 89.7 percent, and First Republic Bank had 67.7 percent. When the percentage of uninsured depositors is that high, it’s easy to see why depositors still run. Their money isn’t safe—not by a mile.

Resolution: zero credibility

In the aftermath of the Global Financial Crisis, everyone agreed on the mantra of “no more bailouts.” As a result, the Dodd-Frank Act gave us the final fail-safe: resolution planning. The underlying idea is that large banking groups would write their own “living wills,” blessed by regulators. If a banking group were nearing its last breath, regulators would then follow the steps laid out in this living will to restructure and break up the group: all without triggering a panic and—importantly—without the need for a government rescue package. That was the aspiration.

Silicon Valley Bank filed such a plan with the FDIC in December 2022, a few months before it collapsed. We reproduce a paragraph of the bank’s resolution plan here for illustration.

SVB’s CIDI Resolution Plan 

The resolution planning rule promulgated by the Federal Deposit Insurance Corporation (FDIC) and subsequent guidance (together, the CIDI Rule) require each covered insured depository institution (CIDI) with $100 billion or more in total assets to periodically submit a plan (a Resolution Plan) to guide the FDIC, as receiver, in the unlikely event of resolution of the CIDI (see 12 C.F.R. § 360.10(a)). A Resolution Plan is intended to permit the FDIC to efficiently resolve the CIDI under Sections 11 and 13 of the Federal Deposit Insurance Act, 12 U.S.C. § 1821 and 1823, in a manner that (a) ensures that depositors receive access to their insured deposits within one business day of the CIDI’s failure (two business days if the failure occurs on a day other than Friday), (b) maximizes net present value return from the sale or disposition of its assets, and (c) minimizes the amount of any loss realized by the creditors in the resolution (see id.).

The idea underpinning resolution planning is that, if and when the fateful day arrives, the plan will be implemented using the bank’s own financial resources—not the government’s. So much for that idea. As we have already seen, when staring down the barrel of a severe economic shock, policymakers are understandably inclined to simply tear up these resolution plans and throw banks a very expensive life jacket. This inclination destroys the credibility of the resolution planning process. It also means that the resolution tools designed for exactly these types of emergencies remain untested.

Systemic failures require systemic fixes

If there’s one lesson to be learned from the past six months, it’s that banking crises reflect systematic failures in bank governance, regulation, supervision, and resolution. This makes perfect sense given that the key components of this system are designed to complement each other. Bank regulation and supervision are designed as a counterweight to ineffective market discipline. Capital and liquidity requirements are designed to reduce the probability of bank failure, thereby reducing the burden on deposit insurance and resolution frameworks. Resolution frameworks are designed to impose losses on uninsured creditors, thereby promoting more effective market discipline. And behind the scenes, supervision plays a vital role in gathering and analyzing the information needed to identify and monitor new threats to safety and soundness, ensure that the price of deposit insurance accurately reflects the relevant risks, and inform the ongoing process of improving technocratic rulemaking. These components are part of an interdependent system. In closing, while policymakers should continue exploring the best ways to reform each of these components, they should also be asking whether this system, as it works in practice, is currently fit for purpose.