Originally published on the LPE Blog
March 23, 2023
SVB has put “interest rate risk,” usually a background player, at center stage. The bank had sound, indeed high quality, assets on its balance sheet – Treasury securities and other fixed rate bonds. But those high quality assets, precisely because they were fixed rate promises, lost value when interest rates rose. As the Fed raises interest rates, the purchase price of such assets drops: investors recalculate that price so that the yield will rise, producing an effective interest rate comparable to the higher rates available on new securities.
When it appeared that SVB might have to “mark to market” bonds that it intended to hold to maturity – thus valuing those bonds at their new and lower price, the potential for significant capital losses became clear, and depositors ran. “Interest rate risk” is not confined to SVB. According to a recent paper by Erica Jiang et al, the market value of assets in the U.S. banking system is $2 trillion lower than their book (par or face) value. In the face of that phenomenon, the Fed opened a new liquidity facility for banks, lending them cash at the higher par value of their bonds rather than the lower market value.
This emerging story line is surely right. But it’s also a Russian doll. Inside the “interest rate risk” is another unappreciated risk: the “uninsured depositors risk.” The Jiang paper also teaches us that SVB was not an outlier with respect to unrecognized losses or (in)adequacy of capitalization. Instead, it was anomalous because 92.5% of its depositors were uninsured. Because they are exposed to the danger that they will lose everything if a bank fails, uninsured depositors are much more likely to run at the first sign of trouble. Their flightiness produces self-fulfilling doom: as customers withdraw their cash, they force a bank into a fire-sale that drives down the value of its assets. Approximately 43% of domestic deposits are uninsured today, some $7.7 trillion dollars. Moreover, these uninsured deposits are spread unevenly, exposing some banks (like SVB) to particularly high risk of runs. Facing that issue, the government intervened again, guaranteeing that it would protect uninsured depositors.
But we should open up “uninsured depositors risk” as well. The FDIC insures deposits up to $250,000, so the “uninsured” were large money-holders. Who were those uninsured depositors? Undoubtedly, some were businesses managing payrolls and other business expenses. But it appears that others were individuals or entities with large cash pools that they failed to put in safer places. One hypothesis is that the beneficiaries of venture capital funds (mostly tech start-ups), at the behest of those firms, disproportionately banked at SVB so that the VC firms would more easily monitor the money flows of those investees. SVB became a lender to individual beneficiaries, who yanked their money quickly, again at the behest of VC funders. A similar trend exists more generally in private banking: large lenders finance borrowing by rich individuals collateralized by equity they own, and those individuals hold those loan proceeds at the lending bank. A related story might be told about the run on Signature Bank, which apparently held large pools of cash for cryptocurrency investors. And finally, a significant number of wealthy individuals left large cash balances in banks in the (heretofore) low-interest environment created by quantitative easing, a monetary policy that itself disproportionately fed the fortunes of the wealthy as it fueled a rise in asset prices
The danger of contagion follows. Uninsured depositors are a small percentage of the American population who, by definition, hold large quantities of cash. They need not pull any strings to get a rescue. Rather, their concentrated wealth means that a tiny number of people – Thiel and his network in the case – can have an outsized and destabilizing impact on the banking system. Contagion in that world happened, literally, by word of mouth. Financialization and escalating inequality thus contribute to the top-heavy architecture we have built, one that benefits the wealthy and is backstopped by the rest.
The modern banking ecosystem grew from a world in which retail money creation by entrepreneurs irrigated a world parched for liquidity. We institutionalized banks not because they were expert at credit allocation, but because they were effective at amplifying the money supply through the provision of private credit. That approach, and the exclusivity of bank privilege that it entailed, is neither sustainable nor any longer necessary. We need to think about structural reforms that would reduce the danger of destabilizing runs in the short term and support more equitable diffusion of money and credit in the longer term, from FedAccounts to direct-issue dollars, narrow banking to The Narrow Bank, and creative channels for credit allocation, from old to new.
The failure of SVB has prompted renewed attention to how the government regulates and supervises the thousands of depository institutions that issue most of the money in our economy. Already, it is clear that many of the legal and regulatory changes made since 2017 to loosen restrictions on banks with assets between $50 and $250 billion were misguided. Depository institutions with more than $50 billion in assets are critical infrastructure and their disorderly failure can be catastrophic for the communities they serve, as well as for the broader monetary system. These institutions should be subject to enhanced prudential standards including rigorous annual stress testing.
But they should also be subject to meaningful, ongoing safety and soundness supervision. In SVB’s case, it shouldn’t have taken enhanced prudential standards to avert a collapse. SVB’s failure must therefore also be examined in the context of a much broader and deeper wave of de-supervision that transformed the financial system in the 1990s and early 2000s. This wave, which I examined in an article titled, “Too Big To Supervise: The Rise of Financial Conglomerates and the Decline of Discretionary Oversight in Banking,” and recently discussed in an episode of Odd Lots, reconceptualized the purpose of bank supervision away from making independent judgments about bank balance sheet configurations. Instead, it focused supervisors on enforcing a regime of bright line rules known as capital requirements and overseeing bank processes and disclosures around risk. Champions of this transformation, like Alan Greenspan at the Federal Reserve, argued that the combination of capital rules and procedural oversight would ensure that shareholders had the necessary information and skin-in-the-game to regulate bank managers and ensure that their risk taking was appropriate. In effect, the government shifted from public-sector oversight of bank risk taking to a market discipline approach.
This approach failed catastrophically in 2008, and for an obvious reason: bank shareholders benefit from levels of risk taking that are harmful to depositors, employees, and the broader financial system. In the aftermath of the 2008 crisis, bank regulators, especially Dan Tarullo at the Federal Reserve, rebooted and reinvigorated supervision, especially for the largest banks. The most important enhancement was the stress testing regime, which served as a way for supervisors to evaluate bank balance sheet configurations and hold bankers accountable for excessive risk taking.
Outside of the stress testing regime, however, there is evidence that the banking agencies have remained overly focused on process. We saw this, for instance, in 2012 with the London Whale losses at JPMorgan Chase. The bank used its excess deposits to engage in flagrantly unsafe and unsound financial speculation. Supervisors failed to halt the speculation. And when the bank’s bets resulted in a multibillion dollar loss, the banking agencies pursued enforcement actions that cited primarily procedural problems (inadequate risk management processes and the like) even though the most significant issue was not the process, it was the investment decisions that came out of the process.
There is obviously a lot we don’t know yet about what happened at SVB. I suspect we will learn that supervisors were aware of the excessive risks taken by management long in advance of the bank’s collapse. We may also discover that, in the wake of Tarullo’s reforms, supervisors were willing to question the business judgment of bank executives, issuing confidential letters to bank executives known as Matters Requiring Attention. But what we are not likely to find was appropriate escalation when SVB failed to address supervisory concerns. The SVB collapse, then, suggests that the banking agencies are still too reluctant to override the business judgment of bank managers. To put it bluntly: Congress did not task supervisors with simply writing letters. It charged them with eliminating unsafe and unsound banking practices using a variety of binding legal tools. Depositors and the public generally would be better served if they used them more.
No one wants to hear this, but SVB operated within one of the most regulated areas of Silicon Valley finance. Every day, financial technology or “fintech” companies, data brokers, and banks collaborate to evade even light-touch regulations. Most relevantly, these companies combine forces to create new forms of “shadow money”—“deposit equivalents,” which avoid substantive banking law and lack critical consumer protections.
The companies claim to disrupt finance. Although the data-driven collaborations are indeed distinct from past partnerships, the hubris is traditional and predictable. Like the first financial institutions to use telephones and index cards, computers and spreadsheets, and the internet, they have successfully argued policymakers should trust but not verify their promises of stability. But as in the past, this is perilous. For instance, a run on the uninsured balances held in fintech software applications (apps), often known as “digital wallets,” could pose problems on a more expansive, but also more granular scale than the SVB run.
Consider an everyday example—if you transfer funds from your bank to your Venmo app (thanks to a data broker, Plaid), the FDIC no longer insures the money, and you would have no real redemption rights during a run on Venmo. For this reason (among many others), some legal scholars refer to Venmo funds as a new type of shadow money. Venmo also pools customer balances, places them in chartered banks, and can yank them from the banks at any time. My colleague Nathan Tankus rightfully argues that certain uninsured “brokered deposits” may also function as shadow money.
Substantive banking regulation doesn’t cover Venmo. Moreover, financial data governance law is so antiquated and weak that it doesn’t govern how Plaid sorts, stores, scores, and shares our data—the actual prize of most of the fintech industry. Multiple forms of arbitrage make these arrangements attractive to both Wall Street and Silicon Valley. In fact, a few days before the SVB run, I presented a paper on the long list of intertwisted financial and informational harms that could flow from this business of “data-brokered deposits.”
Post-SVB, what might a hypothetical run on digital wallets look like? SVB experienced a run on $42 billion of funds. PayPal, Venmo’s parent company, has nearly $40 billion in customer accounts payable. According to one survey, nearly one-third of Millennials and more than one-fourth of Zoomers use a fintech app as their primary checking account. App usage has accelerated throughout the pandemic. Without regulatory reconstruction, the future bank runs of informational capitalism could be much more painful for the general public than the SVB explosion.
If you were to look at the Federal Deposit Insurance Act and related regulations on their face today, you would be forgiven for thinking that the $250,000 cap on individual deposit account insurance that is visible in the corner of almost every chartered bank in the country actually means something.
In practice, however, the cap has been all but dead letter for decades, and in almost every instance of notable bank failure in recent memory – excluding a few small local banks, which serve as the exceptions that prove the general rule – uninsured depositors were made whole through a combination of emergency measures, and prophylactic sales of failing banks to other financial institutions as part of the receivership process.
Until this month, these de facto guarantees were sufficiently submerged behind a de jure commitment to let large-value depositors fail to maintain the political fiction of market discipline on the liability side of the banking system.
Of course, that fiction also carried all sorts of costs, including encouraging the rise of shadow banking, and preventing a more constructive discussion about the appropriate ways to regulate banking activity. But it served its core ideological purpose, which was to keep at bay the deeply disconcerting question lurking behind our entire bank regulatory framework: if the depository system is truly too big to fail, why do we let it continue to operate like a private enterprise at all?
The failure of Silicon Valley Bank, and the quick response by the administration, Treasury, Federal Reserve, and the FDIC to guarantee all depositors, has finally killed the zombie idea that bank deposits are a private good. Instead, even the most casual finreg tourist can now see that bank deposits are, and always have been, a form of public money, issued by banks in their capacity as franchisees of the sovereign money power. Consequently, their safety and soundness is and must necessarily be guaranteed as a matter of public policy, not private markets, in exchange for far more explicit regulation and public accountability for the kinds of credit and asset-generating activity banks engage in on the upswing.
Drawing on Katharina Pistor’s brilliant formulation of the “Paradox of Finance,” we might describe what we are witnessing as the “Paradox of Financial Market Discipline”: In good times, regulators must convince everyone that they will let banking institutions fail in order to establish market discipline, but in crises, such discipline must necessarily be relaxed in order to avoid the very outcomes the initial discipline was ostensibly generated to prevent.
The difficult lesson of this crisis, and indeed so many before it, is that it’s a paradox because it doesn’t work. We aren’t generating market discipline, all we’re doing is rationalizing the same old story of private gains, socialized losses, with the added bonus of intergenerational amnesia. “What have we done! What have we done again?! What have we continued to do!?” we cry, while repeating the same mistakes over and over, and expecting different results.
My father, a federal bureaucrat in Australia, has an old saying that has become increasingly salient to me in recent years: there’s never time to do it right, but there’s always time to do it again. Now is the time to do banking regulation right: accept the inevitable, which is that bank money is a public good, and that the appropriate place for regulatory and institutional governance is on the asset side of banking, not the liability side. Lean into the uncomfortable implications of this fact, and start working on creating a new kind of banking system – one that embraces, rather than shrinks away from, the implications of new digital fiat currency technologies – where people can rest easy knowing their money will always be 100% safe even when individual banks, or indeed bankers writ large, mess up.
Prior to SVB’s collapse, few would have guessed that the failure of a regional, state-chartered bank would inspire such a reckoning with the structure of banking in the United States. However, the fact that the Fed and the FDIC felt compelled to protect SVB’s uninsured deposits in order to prop up the financial system went against much of the underpinning of Dodd-Frank, which relies on a distinction between large systemically-important banks and less important financial institutions. More generally, the government’s response raises fundamental questions about the role that size plays in our decision to bail out banks, one that is playing out politically at this moment.
Canada learned a similar lesson in 1985, when two relatively small regional banks (together amounting to only around 1% of the total banking system) failed, leading the government and the Bank of Canada to rescue their uninsured depositors and to launch various other efforts to stave off financial contagion. What followed was an independent inquiry, after which the government consolidated various financial regulations and created a single office to oversee all federally-regulated financial institutions.
What these experiences suggest is that maybe what makes banks important is not their size but their role. (And this was an issue explicitly raised, though then dismissed, in the Canadian inquiry on those bank failures.) While banks are generally seen as private enterprises that should receive government support only if and when their scale and importance warrant it, another view sees them as public instrumentalities, under a kind of franchise arrangement with the state – designed, from their inception, to carry out the very public task of issuing most of the money we use in our daily lives. This is a view that defines banks more qualitatively, than quantitatively.
Yet if we accept such a view—accept, that is, that what marks the systemic importance of banks is not their size but their role—then this moment does not merely offer a new opportunity for new regulatory measures (which, of course, are absolutely crucial), but also an opportunity to consider broader structural changes to our monetary system. For example, this may be a good opportunity to consider the consolidation of banking regulations: after all, are there good prima facie reasons to exempt small banks and regional banks from the same prudential regulations warranted for larger ones? If we are prepared to bail them all out, then this distinction seems tenuous. This also provides an opportunity to clearly define banks qualitatively in federal law – a position long advocated by some scholars concerned with financial stability.
This should also provide impetus to advocate for more public banking initiatives, such as postal banking, FedAccounts, state public banks or other public money initiatives. For, once banks are recognized as state instrumentalities, the very idea of tasking them (and, for the most part, only them) with issuing our money comes into question. The key underlying question here should be: why accept that the public task of issuing our money should be placed almost exclusively in the hands of undemocratic, for-profit banks at all?
In the immediate aftermath of the collapse of SVB, OpenAI’s Sam Altman observed that “the speed of the world has changed. things can unwind fast. people talk fast. people move money fast.” He was far from the only one talking about a “digital bank run.” While I’m not sure how novel SVB really was in terms of withdrawals—after all, online banking was readily available in 2008—when it comes to panic spreading, online communications channels like Twitter and Slack do seem to have contributed to speed of the run on SVB.
We need to be careful not to overstate their role: the speed of the SVB run was also driven by the close-knit nature of the SVB depositor community, and by VCs advising their portfolio companies to withdraw their funds from SVB. Still, we know that technology does speed things up, and the typical playbook for restoring confidence may not always work quickly enough in this day and age. Moreover, panics aren’t the only way that problems can jump from bank to bank. In a new paper, I explain how increased interoperability of bank systems and shared usage of third-party services like clouds can open up new channels of technological contagion between banks (contagion channels that cannot be addressed by deposit insurance). Faced with all of these possibilities, it’s past time for banking regulators to start thinking about using some kind of “circuit breaker” to slow things down in a crisis.
A bank holiday is a euphemistic term for temporarily suspending banking transactions. Roosevelt deployed one for a week in 1933 to save banks from their depositors’ withdrawals, and to buy time for the government to put in place measures to stabilize confidence in those banks. But what would a bank holiday look like in 2023, when most bank transactions are done online? We need a conversation among banking and technology personnel at banks, the Clearing House Interbank Payments System, banking regulators, Federal Reserve personnel working on master accounts, FedWire, FedNow – and surely many others. There will be technical dimensions to this conversation, but also distributional ones. Will everyone have their transaction accounts suspended, or will some be favored over others? The California energy provider PG&E had to make these kinds of decisions during the 2019 California wildfire seasons, and they were highly controversial. If we are to avoid formulating financial policy in the fog of war, we need to begin conversations about digital bank holidays as soon as possible.