Bloomberg: Why Are Stablecoins Hard to Escape the "Run on the Bank" Curse?

Original Title: “Stablecoins Can Have Bank Runs”

Author: Matt Levine

Translation: BitpushNews Yanan

One point I often mention is that cryptocurrencies are rapidly absorbing lessons from traditional finance, and this process is quite interesting. First, it sometimes goes through some comical or even disastrous events, thereby relearning these lessons. Second, it often manages to grasp these experiences in a very clear, concise, and systematic way. It can be said that cryptocurrencies are an excellent teaching tool. For example, if you want to understand a credit crisis, you could choose to delve into the causes and processes of the 2008 global financial crisis, but that would undoubtedly be quite complex and chaotic. Alternatively, you could choose to study the causes and processes of the 2022 cryptocurrency financial crisis, which is relatively simpler and more straightforward, and also quite interesting. During this crisis, all parties involved were tweeting in real-time and giving interviews on YouTube. Many insights gained from the cryptocurrency crisis are helpful for understanding real financial crises, and they are presented in a simpler, more engaging, and more transparent manner. I once wrote: "Cryptocurrency is like putting those smart and ambitious interns from traditional financial firms into their own gaming market, letting them play freely and take charge." This is precisely why cryptocurrencies have such unique educational significance.

Stablecoins are a special type of cryptocurrency product that can be seen as an abstract representation of a bank. When you deposit 1 dollar with a stablecoin issuer, you receive a receipt, which is the stablecoin, stating "equivalent to 1 dollar." Subsequently, in the cryptocurrency environment, such as on a blockchain or within a cryptocurrency exchange, you can use this stablecoin as if it were 1 dollar. It has transactional functionality; for example, you can use 1 dollar's worth of stablecoin to purchase 1 dollar's worth of Bitcoin, while the seller of the Bitcoin will receive your stablecoin.

At the same time, the stablecoin issuer will hold your 1 dollar and invest it to seek profits. These profits are generally used to pay for operating expenses, executive salaries, and so on. You (or anyone holding the stablecoin) can usually redeem the stablecoin for your dollar within a short period. At this point, the stablecoin issuer must pay back that 1 dollar.

Overall, this process is quite similar to bank deposits, although there are some differences between the two. If you understand how banks operate, you can foresee some ways in which this model might encounter problems. Specifically, there are two main scenarios:

First, when banks or stablecoin issuers hold your funds, they have the right to invest and keep the profits. The more profit they make from depositors' funds, the more they keep for themselves. However, on the other hand, if they incur investment losses that lead to total loss of funds, depositors will have to bear most of the losses. Since most of the issuer's funds come from depositors, once losses occur, there will be no more funds to compensate them.

Thus, this creates an incentive for issuers to take risks: if the risky investments succeed, they can earn substantial profits; if the investments fail, the losses are borne by others.

In this case, a "run" may occur. The stablecoin issuer may invest all your funds in seemingly very safe projects, but some of these may be long-term investments. However, if all depositors demand to withdraw their funds on the same day, the issuer may not be able to meet this demand immediately. In such a case, they may have to urgently sell these long-term investments at a loss, ultimately leading to a shortfall in funds to repay all depositors.

Since this interconnected logic is well-known, it has a self-reinforcing characteristic: if you foresee a potential run, you will tend to withdraw your funds first for safety (before the issuer runs out of money). If everyone takes such action, a run will inevitably occur.

These two issues are often interconnected, as a common reason for bank runs is that banks have made poor investment decisions with depositors' funds. However, they do not always occur simultaneously. A bank may suffer significant losses on poor investments and go bankrupt before anyone realizes it and has time to queue up for withdrawals. At the same time, there may also be runs that are not directly triggered by investment losses but are entirely due to liquidity mismatches between the bank's immediate liabilities and illiquid assets.

In the banking sector, there are a series of SOPs to address various challenges. For investment issues, the main strategies are:

1. Prudent regulation: Regulatory agencies closely monitor banks' investment behaviors to prevent them from making unfavorable investment decisions.

2. Capital regulation: Banks are required to hold a certain proportion of additional capital so that even if investments incur losses, they can use non-depositor funds to absorb the losses.

For run issues, the main solutions are:

1. Liquidity regulation: Banks are required to maintain sufficient cash reserves to meet withdrawal demands that depositors may make at any time.

2. "Lender of last resort" system: If a bank has high-quality illiquid assets but faces simultaneous withdrawal demands from all depositors, the Federal Reserve will provide loans to that bank, as it believes the bank will ultimately be able to repay.

3. Deposit insurance: The government promises, "If the bank really fails, we will refund your deposits up to a certain limit, so please do not participate in a bank run."

In the stablecoin sector, overall, are these measures largely absent? Indeed, there are some proposals in the industry that cover some of these measures. However, we have discussed Tether, the largest stablecoin issuer, which got into trouble in 2019 due to extremely risky investments of customer funds, with its published accounts showing a capital ratio of only 0.2% (though this has since improved). Additionally, we discussed TerraUSD, whose investment strategy was almost equivalent to "a large number of highly correlated risky assets," resulting in a complete collapse during a run in 2022.

So, what elements should an ideal regulatory solution for stablecoins have? "Deposit insurance and access to the Federal Reserve's discount window" sounds like an interesting and practical requirement, and as I write this, I suddenly realize this could very well become a reality in the first year of the next Trump administration.

A more direct and core answer is: "Stablecoin issuers should invest funds in relatively safe and highly liquid projects, and they should hold a certain amount of their own capital. This way, even if investment projects incur losses, the issuers will still have enough funds to repay depositors." Of course, on this basis, further refinement of specific implementation details is needed.

Gordon Liao, Dan Fishman, and Jeremy Fox-Geen co-authored a paper titled "Risk-Based Capital for Stable Value Tokens," all from the stablecoin issuing institution Circle Internet Financial. Although Circle has its own interests in lobbying for looser stablecoin regulations, this paper still offers some profound insights into the relationship between stablecoins and the banking industry. One point is that stablecoins are more transparent than banks in many ways, which is undoubtedly a positive factor, especially for cryptocurrency enthusiasts who are skeptical of the opacity of traditional finance. However, this opacity does not exist without reason.

Liao, Fishman, and Fox-Geen point out that, all else being equal, tokenization increases the likelihood of coordinated run risks facing stable value debt. This is because tokenization allows the underlying value to be transferred and traded outside the complete control of the token issuer. This trading creates secondary market prices, providing the market with an observable signal. The real-time visibility of market prices may exacerbate investors' reactions, making issuers more susceptible to run risks. If the ledger is public, transfer activities will also become observable. The information revealed through observable public signals may lead to "overreactions" in financial markets and prompt market participants to coordinate run behaviors in a global game (Morris & Shin, 2001). In other words, if token holders observe a significant drop in secondary market prices or a large number of redemptions, they may panic and choose to sell or redeem tokens, disregarding the fundamentals of asset backing.

In traditional banking, the main reason for bank runs is often that you believe others will also run on the bank. What is the basis for this judgment? Oh, there are many sources: rumors, poor earnings announcements, panicked television appearances, etc. A drop in a bank's stock price may be seen as a signal of deposit problems. But this is, after all, an imprecise science. In contrast, stablecoins trade on public markets, and their prices directly reflect people's confidence in them. For example, if a stablecoin's trading price is 1.0002 dollars, that may indicate that no run is currently happening; but if the trading price is 0.85 dollars, that may suggest a run is occurring.

The main solution to this issue is that stablecoin issuers should maintain a high level of liquidity for most of their funds.

Due to the increased risk of runs, methods for managing financial risk also need to differ from traditional banks. For example, fiat-backed stablecoins typically hold highly liquid assets, with minimal maturity mismatches compared to banks and relatively lower credit risks. Therefore, considering the resilience of the asset pool isolated for the benefit of token holders, the capital buffer of stablecoins to absorb financial losses is actually less than that of banks. Even if the underlying assets are the same, tokenizing traditional deposits may increase the run risks faced by financial institutions. This is because tokenization makes deposits easier to withdraw quickly and on a large scale, and token holders may be more sensitive to market signals. Therefore, institutions providing tokenized deposits may need to hold more capital to address this increased risk, even if their asset bases are the same as traditional banks. Essentially, tokenized deposits introduce a similar asset-liability mismatch problem inherent in bank balance sheets, so they may also need to apply similar capital and solvency regimes.

Next, let's talk about the "blockchain" aspect of stablecoins:

In addition to financial risks, the use of tokenization and distributed ledgers also brings additional risk considerations related to technology, infrastructure, and operations. These non-financial risks have been emphasized in regulatory public consultations and proposals. Although advanced cryptographic technologies, permanent record-keeping, and traceable transactions can reduce certain security and compliance risks, these risks, widely referred to in traditional banking terms as "operational risks," face challenges when assessing the required capital. This is mainly due to the lack of sufficient historical operational loss data, and the entire assessment largely relies on specific technological choices. In a rapidly evolving and continuously upgrading infrastructure environment, the technology chosen by issuers may significantly impact the loss-absorbing capital they need to prepare.

In other words, we can reasonably assume that stablecoin issuers are less likely to lose users' funds compared to traditional banks because their operations are based on blockchain technology, which offers transparency, traceability, and digital native characteristics that traditional banks lack. However, for the same reasons, we can also reasonably assume that they are more likely to lose these funds.

Last year, the U.S. experienced a small-scale banking crisis, which prompted me to spend some time reflecting on the banking industry. I wrote:

The banking industry is, in fact, a way for people to unconsciously engage in long-term, risky bets; it pools risks, making everyone feel safer and more benefited. The reason you and I are willing to deposit money in banks is that we feel "the money in the bank" is very safe, and we can use it tomorrow to pay rent or buy a sandwich. Subsequently, banks use these deposits to issue 30-year fixed-rate mortgages. Homeowners can never borrow 30 years' worth of money directly from me because I might need the money tomorrow to buy a sandwich, but they can collectively borrow money from all of us because banks have diversified this liquidity risk among numerous depositors. Similarly, banks also lend to small businesses that might go bankrupt. These businesses can never borrow money directly from me because I need that money and do not want to risk losing it, but they can collectively borrow money from all of us because banks have diversified this credit risk among numerous depositors and borrowers.

However, the opacity of traditional banking allows it to utilize customer funds for riskier ventures more. Part of the reason for last year's regional banking crisis may be that this opacity is no longer as effective as it once was. Nowadays, more information is available online, rumors and panic can spread rapidly across the globe electronically, and there are greater expectations for market capitalization. As a regulator from the Federal Deposit Insurance Corporation said last year: "The rules haven't changed; the competition has just become fiercer."

The magic of traditional banking lies in the ability of banks to engage in a series of risky investments, pooling these investments together, and then issuing senior debt against these portfolios, which is what we commonly refer to as "dollars": 1 dollar in a bank account is a real 1 dollar, even if it may be backed by a pile of risky assets. However, stablecoins forgo this magic: while a dollar stablecoin is close enough to 1 dollar for most cryptocurrency uses, it has its own trading price. In good market conditions, its price might be 1.0002 dollars or 0.9998 dollars, but in poor market conditions, you know its price could drop to 0.85 dollars. This is essentially a banking operation without the guarantee that "1 dollar in the bank equals 1 dollar," and the real-time prices available 24/7 tell you how close it is to 1 dollar. This undoubtedly creates new regulatory issues and may also foreshadow some new situations for conventional banking in the future.