Becker Friedman Institute
for Research in Economics
The University of Chicago

Research. Insights. Impact. Advancing the Legacy of Chicago Economics.

The Paradox of Frozen Liquidity

How unusable liquidity can prevent bank runs—a tale of sunspots and taxicabs

Imagine you are lying in bed at 6:30 in the morning and you hear the newspaper land at your front door. You get up and look at the front page and see the headline: “Bank Run Today, see page 2.”

You know that everybody else in town gets the paper, so you wonder if you should put your shoes on and run down to the bank. You start to fear for the safety of your deposits. Should you get in front of the line before the bank opens to ensure that you get your money out?

But then you decide to read page 2 to get more information. And the headline on page 2 reads: “Marathon this Morning, Sponsored by First National Bank.” That’s a relief, you think. But then you pause again, because you don’t know how many other people have taken the time to read page 2. Maybe a lot of them panicked when they saw the front page and were already running to the bank. If enough of them run, then your deposits are still at risk, even if the bank is actually in good financial shape.

What do you do?

That’s the “sunspot” story that Douglas Diamond, Merton H. Miller Distinguished Service Professor of Finance at the University of Chicago Booth School of Business, tells his MBA students to illustrate the fear that takes hold in a bank run. Sunspot in this case means an event that all can see, but all do not interpret the same way. When it comes to the newspaper headlines, it’s not the fear that a bank is in trouble that fuels a bank run, it’s the fear of everyone else’s fear, or the fear of fear itself.

That insight into bank panics informs the recent work of Diamond and Anil Kashyap, Edward Eagle Brown Professor of Economics and Finance at Chicago Booth. Diamond and Kashyap have developed a model that shows how liquidity regulation can prevent bank runs. They produced this model in response to forthcoming liquidity rules that are meant to protect banks during runs. Diamond and Kashyap find these new rules lacking and misdirected, and they offer a solution that directly addresses—and resolves—the fear that takes hold during panics. Take away the fear, and there is no need to panic.

The Last Taxi at the Basel Train Station

Diamond and Kashyap’s work is in response to forthcoming Basel III rules that stipulate liquidity requirements meant to protect banks against runs. (Basel III is the third installment of voluntary regulations issued by the Basel Committee on Banking Supervision in response to the financial crisis of 2008.) The Basel III proposal includes two key elements regarding liquidity:

The net stable funding ratio (NSFR), which requires “banks to maintain a stable funding profile in relation to the composition of their assets and off-balance sheet activities” (Basel Committee on Bank Supervision (2014)). Broadly, the NSFR forces banks to match long-term assets with long-term funding. Diamond and Kashyap, though, interpret this requirement to mean that the bank is free to violate the requirement temporarily in the future (with a long period allowed to rebuild its liquidity holdings), so it is not always a binding restriction. And if it were not binding, then why would depositors believe it?

The second part of Basel’s liquidity plan is called the liquidity coverage ratio (LCR), which requires “that banks have an adequate stock of unencumbered high-quality liquid assets (HQLA) that can be converted easily and immediately in private markets into cash to meet their liquidity needs for a 30 calendar-day liquidity stress scenario” (Basel Committee on Bank Supervision (2013).

From the authors’ perspective, an LCR is superior to an NSFR because it can require a certain amount of unusable liquidity levels against deposits. The problem with Basel’s LCR is that it is not inviolable. Within that given 30-day stress period a bank may tap into its reserve and, therefore, the amount of liquidity will be unknown. Such uncertainty can cause panic (remember, it’s not necessarily the fear of the bank’s liquidity position that matters, but the fear of everyone else’s fear).

So what you need is a scheme that offers a concrete guarantee to depositors that there will always be enough liquidity on hand. Such assurance will do more than calm depositors during a financial panic—it will prevent panics from occurring in the first place. How? Consider the last taxi problem, a metaphor variously ascribed to a number of economists over the years, including Milton Friedman, and illustrated here by Charles Goodhart:

“The weary traveler who arrives at the railway station late at night, and, to his delight, sees a taxi there who could take him to his distant destination. He hails the taxi, but the taxi driver replies that he cannot take him, since local bylaws require that there must always be one taxi standing ready at the station.” [Goodhart, 2008]

This may sound inefficient at best, and crazy at worst, the authors acknowledge, but leaving a metaphorical last taxi at the station—in this case, unused liquidity on the bank’s books—ensures that the bank will always have enough liquidity on hand. Why is that useful? Because if depositors know that there will always be unusable or extra liquidity at the bank, then that means the bank must, in addition, have sufficient funds to “back up” the unusable portion—or the last taxi. If some deposits are withdrawn, a fraction of the required liquidity can be used, but some liquidity must remain.  If there is always a last taxi at the station, and people know it, then there will never be a reason to panic about the existence of taxis. This is the paradox of unusable liquidity: something that can be used to stem panic is actually partially frozen to prevent that panic from ever occurring.

But what about that poor guy who is told that the last taxi can’t take him home? That’s the last loan that can’t be made, and it’s a cost to the individual and to the broader economy. But it’s a small cost compared to the damage wreaked by a financial panic. Imagine a scenario where many hundreds or thousands of people are rushing toward that last taxi in fearful desperation. That’s what the Diamond-Kashyap model avoids. It’s better to leave our unfortunate traveler to find another way home then to risk such mayhem.

What Depositors Don’t Know Can Hurt Them

Bank liquidity is opaque, and not only for depositors but also for regulators. This lack of transparency is a key driver of Diamond and Kashyap’s approach.  With complete information available at all times, a bank would be forced (by the market) to hold enough liquidity to deter runs. However, complete information about bank liquidity is unrealistic, for several reasons, the authors argue.

First, if disclosure (or a regulatory requirement) regarding liquidity only applies on some dates (such as the end of an accounting period), the bank can distort the disclosure. Second, even if a liquidity disclosure (or requirement) applies on all dates, it is plausible that the bank knows more about its customers’ liquidity needs than anyone else, which makes it difficult for others to determine if a given level of liquidity is sufficient to make the bank run-free.

Third, it is difficult to interpret the kind of accounting data that must be parsed to decide whether to join a run. Disclosures that are made on liquidity positions typically occur with a delay and are periodic (such as at the end of a quarter or a fiscal year). Fourth, this inference problem for depositors can be compounded by the temptation for banks to engage in window dressing of their accounting information.

Finally, only the bank knows the normal fraction of depositors who seek withdrawals for typical reasons. These withdrawals vary seasonally, cyclically, and for other reasons. This is an important consideration for depositors and regulators, for without such knowledge, depositors can never be sure about the bank’s liquidity levels preceding or during a run. As the authors state: “If there is no way to communicate what the bank knows about this, and self-interest does not automatically make it stable, then runs will occur unless the bank discloses enough liquidity to stay solvent even in the worst case: the largest possible fraction of normal withdrawals plus the fraction withdrawn in a run.”

Depositors who worry about runs but who do not know how many others are withdrawing for normal reasons will have even more incentive to withdraw in times of uncertainty. This means that a liquidity disclosure rule that is not sufficient to make a bank run-free for the highest level of normal withdrawals can lead to runs at all levels of withdrawals. Why? Because if a worried depositor doesn’t know for sure what the normal level of withdrawals is and she is concerned about a panic, then she should always assume the worst.

Those worried depositors are essential to understanding Diamond-Kashyap’s model. The authors make the important point that all runs are partial; not all depositors withdraw their funds.

Who are these depositors? They are the sophisticated or institutional investors, those with an incentive to pay close attention to signals of possible problems, or sunspots, and whose interpretation of those sunspots can actually cause a run. These are the depositors that Diamond and Kashyap include in their model. The authors show that if these worried, sophisticated depositors can be convinced that a bank has extra liquidity—if they can see that these banks are always liquid—then they will have no incentive to run. And if worried depositors don’t run (remember, all runs are partial), then banks are run-free.

Theory into Practice

Broadly speaking, under Diamond and Kashyap’s plan, regulators need only measure how much liquidity per deposit remains after normal withdrawals occur. There must be a positive fraction of liquidity left unused after these normal withdrawals. This is the last taxicab that must never leave the station, and this is the key difference between Diamond and Kashyap’s model and their interpretation of the NSFR. Under Diamond-Kashyap and the LCR, regulators don’t have to know why withdrawals happen, because if the extra liquidity is held, only normal withdrawals will occur.

Diamond and Kashyap introduce another element that distinguishes their LCR-type model from the Basel plan: the integration of a lender of last resort that would hold excess liquidity. Should the need ever arise, banks would have access to excess liquidity by borrowing against it from their country’s lender of last resort, similar to the Federal Reserve’s discount window. Though this would effectively violate the bank’s liquidity requirement, a sufficient penalty—such as reduced executive compensation and a limitation on dividends—would provide disincentives to borrow against it to fund normal withdrawals. The liquidity would be there only to finance unexpected withdrawals in a run.  But the key point is that simply having access to this liquidity via a lender of last resort would act to further calm worried depositors and prevent runs.

Incentives Matter More Than Buffers

We can have a run-free banking system, according to Diamond and Kashyap, which sounds like a startling claim, especially given the financial crisis of 2008 and the fear that coursed through financial markets. That does not mean that banks couldn’t fail, but they wouldn’t do so in such dramatic fashion that induces further panic. Extra liquidity would allow for gradual, more orderly closure. It would be less likely that one would need crisis meetings over sleepless weekends, followed by bombshell announcements on Sunday night in advance of market openings.

Most analysis of liquidity requirements begins with the question: How much liquidity do we need to buffer against extreme withdrawals (as in a crisis)? But trying to answer that question leads to ineffective policies, especially because depositors can never be sure what the normal level of withdrawals is at a given time, let alone what extreme withdrawals might be. Rather, the goal of liquidity regulation should be to provide incentives for banks to hold liquidity in excess of the required amount. If a bank wants to use some liquidity, then it must have more than is required.

This is not just a buffer against runs that happen. It’s a buffer that prevents runs from ever happening. Get the incentives right, and the buffer will follow.

Diamond and Kashyap’s work is a reminder that liquidity regulation serves a critical function in market stability. And just because Basel III is already written and ready for implementation through 2019, doesn’t mean that this debate is closed. Countries and their central banks can still decide how to implement the new rules, and there will probably be a Basel IV. Diamond and Kashyap’s model serves as a basis for subsequent discussions on designing optimal liquidity requirements. There is still time to park some taxis at the train station.

— David Fettig

Download paper

Diamond and Kashyap in Banking Perspectives