Bank Runs or Not:
An Interpretation of FTX
Bank runs are again a topic of conversation.1 Two events, probably unrelated, have brought this issue to the forefront of public discussion. First, the Nobel Prize in Economics for 2022 was shared by Douglas Diamond and Philip Dybvig (with Ben Bernanke) for their paper on bank runs. Second, there was the FTX collapse, that many have interpreted as a bank run.
Bank Runs ala Diamond and Dybvig
This famous paper in economics makes two important points.
First, it makes clear the important role of illiquidity in the intermediation process. Essentially banks (and near banks) invest in long term, relatively illiquid assets. These investments though are funded by shorter term liabilities, such as demand deposits.
Second, these banks are fragile. Given a deposit contract (an important given), there can be two outcomes. One arises when depositors are optimistic: if a single depositor wished to make a withdrawal, then the bank would be sufficiently liquid to meet that demand. A second outcome is driven by pessimistic: depositors do not believe that banks are sufficiently liquid and thus all run to the bank to withdraw.2
Both outcomes are self-fulfilling: i.e. they are Nash Equilibria given the deposit contract and the investment profile of the bank. This means, absent any deposit insurance and without a way to limit withdrawals, deposit arrangements with banks are vulnerable to runs. They may not always happen, but they could happen.
The contribution of the Diamond-Dybvig paper was to formalize these points, particularly making clear the conditions for the fragility of banks highlighted by the self-fulfilling nature of bank runs. That analysis has since been built upon and extended to study a wide range of closely related intermediation activities.
Having formalized this problem, Diamond-Dybvig offered a solution to stabilize the banking system, again working through the beliefs of depositors. If there was a government commitment to the provision of deposit insurance, then even if depositors thought the bank was illiquid, they would not need to withdraw their deposits. If everyone else ran, thus draining the banks of resources, the government would always be there to provide deposit insurance. In this was, deposit insurance stabilizes the banking system.
There are a number of issues left open. First, of course, beliefs are only one source of fragility of banks. Their analysis does not imply that, say, information about the returns to a bank cannot also create instability. Second, their analysis does not make clear what might set off a bank run in the absence of deposit insurance. That is, what triggers a switch from optimism to pessimism? Third, what exactly is a bank? In their model it is a just an intermediary that has an imbalance between short term liabilities and long term assets.
Other instances of Fragility
This type of instability where beliefs are so powerful also pertain to other markets. These are not versions of bank runs but rather interactions which are also prone to the adverse effects of pessimism.
A leading example, relates to the value of one currency in terms of another.3 In this case, the run is not against a bank but against a currency supported by a government. Imagine that a government wishes to maintain the value of its currency, say the peso, in dollars, by targeting a price of the peso in dollars. To maintain this target, the government has available resources which can be used to buy pesos and thus support the currency even when everyone else wants to sell them at the targeted price. But these resources are finite.
There are investors who may (or may not) be dubious about the ability of the government to support its currency. These investors are like the depositors in the Diamond-Dybvig model. If the investors are confident that the government will be successful in keeping the peso at its target price, then they have no need to sell pesos for, say, dollars. But if investors are pessimistic, then they fear the government will be unable to maintain the peg. As a consequence they are led to sell pesos for dollars, thus driving down the price of pesos. The government can attempt to support pesos at its targeted price but in the end, with scarce resources, maintaining the high value of the currency may be impossible. If it is, then there is a currency run.
There is another type of “run” that is common to economic models that entail fiat currency. By definition, fiat currency retains value only in exchange for goods and services (or other monies). That is, in a fiat currency system there is no explicit backing of the currency by the government. In contrast, say, a gold system entails a promise of a government to exchange gold for paper money. The dollar, the euro, the RMB are all fiat monies.
There is a certain fragility in the valuation of fiat money. If everyone believes that everyone believes that ….. the US dollar is a worthless piece of green paper, then that will be self-fulfilling. Think about it this way, if no one else is willing to give up goods or perform a service in exchange for dollars, then why would you?
And then there is another phenomenon called bubbles.4 We often hear about real estate bubbles where the price of a house goes up and up, in part because of anticipated capital gains from the next sale. As long as there is that next sale, the bubble continues. Until it doesn’t.
It is not that unlike fiat money. Once people stop thinking it has value, then the pessimism reduces the value of the object. When it is fiat money, the value of the paper is close to zero. For a house, that value is the service flow from living in the house. For sure that is not zero, but it is less than the inflated price of the asset, in this case a house, sustained by the bubble.
And then there are Ponzi schemes. These are a bit different in that they often involve an element of fraud. Here is how it goes. Someone offers you a return higher than the market return. You cannot resist and invest your savings with that individual. Your investment, no surprise, just earns the market return but you are promised more. The difference is made up by the deposits into the fund of the next group of investors. This process can continue as long as there is another group of investors. Once that pool dries up, the returns are gone.
FTX
The recent FTX experience has elements of all these models. To the extent that FTX operating as an intermediary by taking in funds and lending them out, then it was prone to the same “runs mentality” as a bank in the Diamond-Dybvig world.
There was also a crypto currency component to FTX. And these currencies like others are also prone to wide fluctuations in value.
But unlike the emphasis in the vast economics literature on bank runs, the FTX crash was not be based on pessimism alone.5 Looking at the various accounts of this episode there were certainly signals along the way that the fundamentals were not right. It was not just a question of liquidity but also solvency.
But once the power of beliefs are taken into account, a small change in fundamentals can have a big impact. That is, when we talk about the Diamond-Dybvig bank or a currency run, it is not clear what switches the minds of agents from optimism to pessimism. One obvious candidate is a tiny bit of news that alters perceptions and sets in motion a process that in the end is a pessimistic outcome.6
What should be done?
And, we should add, by whom?
We know two things for sure:
some people lost money from the FTX crash;
Congress is holding hearings.
On the first point, welcome to the world of investing! There are always people losing money and others winning. It is not much different than going to the racetrack or buying stock shares recommended by your broker. Sometime you win, most often you don’t.
This is not problematic or by itself a cause for government intervention. Unless, of course, fraud is involved. And that may have been the case: FTX founder Sam Bankman-Fried was charged with some serious crimes.
There is of course the connection to the exchange of crypto currencies. Does this episode with FTX indicate some instabilities in that industry that warrants some additional government oversight?
From the evidence thus far, the FTX unraveling seems pretty standard. The combination of a little optimism by investors, a little bit of blind faith and a whole lot of moral hazard in the traditional form of “ … take the money and run” seems like a recipe to explain these events.
It is a bit like how you should react to emails from “prospective business partners” around the word, offering a high return if you just wire some funds into their accounts. Sure, no problem. As usual, investor beware!
There is yet no definitive account of these events. From that perspective, imagine being an. investor in the midst of this crisis.
The issue here is liquidity not solvency of the bank.
Maurice Obstfeld is a key contributor to this literature.
I found the essay by George Selgin a great read on these events.
This sounds good but as far as I know is not formally part of these models.