It has been a year since Silicon Valley Bank failed, and one of the major policy questions that emerged – what to do about deposit insurance? – remains unresolved.  Banking is an industry that needs public confidence to function, and we have become accustomed to relying on confidence-inspiring deposit insurance as our primary bulwark against bank runs.  But deposit insurance (at least in its current form) seems less like a silver bullet in the wake of SVB’s failure.

So what should be done? Some argue that we should dispense entirely with deposit insurance caps (currently set at $250,000 in the United States) to ensure that even the largest depositors have no incentive to run.  Proponents of this kind of policy point out that existing caps have been rendered largely meaningless – there are now expectations that uninsured deposits will be guaranteed after the fact, just as they were with SVB.  But others worry about unlimited deposit insurance creating incentives for riskier behavior by banks and about the cost of unlimited deposit insurance (even if we increase the premiums banks pay, taxpayers will remain the ultimate backstop) – not to mention the optics of what might look like a bailout for millionaires.

Both sides of this debate make extremely good points, and if they are both right, then perhaps we are simply asking deposit insurance to do too much and need to explore other policy options for responding to bank panics.  For example, we might need to dust off the concept of the “bank holiday” and consider whether authorities might ever need to implement one (or more limited transaction restrictions) in response to a banking panic.

I’m using the term “bank holiday” to mean a suspension of banking activity, as President Franklin Roosevelt used it in 1933 (not in the modern UK usage of “a day off work”).  The disruptions of a bank holiday would be extreme, to be sure, but we may nonetheless face circumstances in which one is necessary to buy time to develop and deploy other confidence-inducing measures.  Back in 1933, implementing a bank holiday could be more or less effected by keeping banks’ doors closed.  The way we bank has changed dramatically, though, and implementing any transaction restrictions would be much more complicated in this day and age.

Many accounts of the SVB run ascribe its speed to social media activity.  These accounts need to be taken with a grain of salt: In a bank run, the first movers have the advantage, and so rational depositors will keep their concerns off social media – at least until they’ve withdrawn all their funds.  While I readily concede that not all depositors will act rationally, there is contemporaneous reporting that many SVB customers did try to hush up their withdrawals. Incentives shifted once it was clear that SVB would not survive, though.  Then, the community of venture capitalists and startups served by SVB had incentives to convince authorities that guaranteeing SVB’s uninsured deposits was critical to preventing a broader banking panic.  From March 10-12, 2023, social media certainly seemed to be fanning concerns about such a broad panic.  Whether or not that was intentional, in our era of misinformation and generative AI, the possibility of bad actors using social media to instigate future bank runs cannot be discounted.

Even unlimited deposit insurance may not be able to protect against a coordinated attack on confidence in a bank’s liquidity or solvency.  And in the future, we may face new kinds of prompts for runs that deposit insurance is not even responsive to.  For example, as we move towards open banking, it may become much easier to switch from one bank to another: What if a widely used app recommends that all of its users switch funds to a bank that pays a higher interest rate? Depositors may have no concerns about the liquidity or solvency of their old banks but still move their deposits away.  It doesn’t matter how things kick off; once significant deposit outflows start, a bank (even one in excellent financial condition with fully insured deposits) will be forced to sell its best and most liquid assets to satisfy withdrawal requests, which could ultimately cause it to fail.

If confronted with an unfamiliar emergency, authorities may need to buy time to develop and implement new kinds of responses.  A digital bank holiday, or lesser transaction restrictions, could prove extremely valuable in such circumstances.  Some might fear that online and mobile banking have made withdrawals so speedy that no such intervention could ever be timely enough – but U.S. banks don’t allow their customers to make million-dollar withdrawals through instantaneous payment services like Zelle.  A withdrawal of that magnitude still requires a wire transfer, and wire transfers still take time to process (just read the contemporaneous reports of SVB depositors panicking as they waited for their wires to go through).  If U.S. authorities ever needed to implement a digital bank holiday, wires would be the transactions to target, and disabling banks’ Federal Reserve master accounts would be the most expedient way of preventing wires from being processed.

As former Treasury Secretary Tim Geithner used to say, “plan beats no plan,” and if there’s ever a chance we will need to implement some kind of emergency transaction restrictions, we should start thinking through the legal and operational aspects of implementing a digital bank holiday (or lesser restrictions) now.  The more quickly they are implemented the better – if rumors start to swirl that such restrictions are coming, damaging runs will only increase in the interim.  With my article, Digital Bank Holidays, I hope to kickstart advance planning for practical implementation and legal authority issues.  With regard to legal authority, although congressional authorization would be ideal, it’s quite possible that Congress will not be able to agree on such authorization in a timely fashion.  I therefore consider existing legal authorities that could support the Federal Reserve in acting unilaterally in an emergency, or in compelling other participants in the payments system to participate in transaction restrictions.

I want to stress again how dire conditions would need to be to justify these kinds of steps. Restrictions on Federal Reserve master accounts are already a political hot potato and digital bank holidays (or even lesser transaction restrictions) will inevitably have damaging unintended consequences.  But in the future, it’s possible that the only meaningful response to a fast-moving run or broader bank panic will be to hit the pause button – and plan beats no plan.

This post comes to us from Professor Hilary J. Allen at American University’s Washington College of Law. It is based on her recent article, “Digital Bank Holidays,” available here.