May 20, 2022
by Stephen T. Gannon, Lisa Weingarten Richards, Maggie Sklar

 

Terra/Luna - Valuable Lessons About What Stablecoins Are and Aren’t

After the UST token failed to maintain its $1.00 peg, commenters have debated the reasons for and consequences of this crash. The loudest voices have called for greater regulation not only of stablecoins, but of the entire digital asset industry. As a counterpoint, this note attempts to “pull the camera back,” review the facts behind the UST debacle, compare the UST structure to that other reserved “payment stablecoins,” analyze whether the flaws in the UST structure have exhibited connectivity to the rest of the financial system, and suggest some voluntary tools that the industry might consider to mitigate risks going forward. It concludes that the events leading up to and underpinning the UST failure do not support calls for extensive and compulsory industry-wide regulation, and that the most effective and fast-acting remedies will be voluntary adoption of new risk tools and disclosure standards by the industry.

Stablecoins have become extremely popular, and for good reason. They are often used by digital assets holders to move in and out of digital asset markets. Some stablecoins allow holders to stake them to earn staking rewards.[1] More universally, they can be used for immediate payments at a fraction of the cost of traditional wire transfers or credit card payments. Stablecoins can be used across borders to greatly speed up international cross-border payments and remittances. Because they are so efficient and available without the need for a traditional banking relationship, they can also help increase financial inclusion for underserved communities.

Despite the utility of stablecoins, the recent crash of the Terra Luna ecosystem (a blockchain platform and protocol created by Terraform Labs), which included UST, an algorithmic digital asset that Terra Luna referred to as a stablecoin, as well as the associated digital asset, LUNA, will likely be remembered on a short list of rapid and harsh financial catastrophes. But the lesson to take away from this crash is not the lesson that Treasury Secretary Janet Yellen and some others in government and the media have tried to provide—this is not a lesson about why all stablecoins and digital assets are risky and need to be heavily regulated. In fact, when looking at how genuine payment stablecoins (i.e., those that are tied to 1:1 cash and secure cash-equivalent assets) have been behaving amidst the recent downturn in the markets, the lesson is the opposite: despite external market fundamentals (such as rising interest rates, supply shortages, and the war in Ukraine) which have been moving all U.S. markets, genuine payment stablecoins performed as expected and remained essentially at their dollar peg. 

What is a Stablecoin and What was UST?

In analyzing this question, we must first clearly define the term “stablecoin.” A stablecoin is “a digital currency pegged to a ‘stable’ reserve asset like the U.S. dollar or gold. Stablecoins are designed to reduce volatility relative to unpegged cryptocurrencies like Bitcoin.”[2] With this definition in mind, one can evaluate the algorithmic asset UST to see if it was truly a stablecoin.[3]

Terra Luna’s website, both before the crash, and to date states as follows:

Stablecoins are only valuable to users if they maintain their price peg. The Terra protocol uses the basic market forces of supply and demand to maintain the price of Terra. When the demand for Terra is high and the supply is limited, the price of Terra increases. When the demand for Terra is low and the supply is too large, the price of Terra decreases. The protocol ensures the supply and demand of Terra is always balanced, leading to a stable price.[4]

UST maintained its peg through a redemption mechanism. In theory, one UST could always be redeemed for $1.00 worth of LUNA, and vice versa. This allowed the UST “stablecoin” to hold a $1 price in regularly functioning markets. If the price of UST fell below a dollar, arbitrageurs would buy it, burn it and mint $1.00 worth of LUNA, then sell that LUNA to earn a profit. This would reduce the supply of UST and raise its price back to $1. The reverse would happen if the UST price rose above $1. Arbitrageurs would buy $1.00 worth of LUNA, burn it and mint one UST, and sell that UST at a profit.[5] That would increase the supply of UST, thus moderating its price back to approximately $1. However, it was entirely reliant on the demand for LUNA, which had no traditional reserves. Accordingly, if there was no demand for LUNA, there would be no demand for UST.[6] Leading up to the crash, Terra had driven the popularity of the UST coin by offering 20% staking rewards for UST which was staked on the related Anchor Protocol.[7] This helped increase the demand for UST which at its peak was valued at over $18 billion in circulating supply, with over $15 billion in “Total Value Locked” on the Anchor protocol.[8]

When the price of UST and LUNA started to fall, Terraform Labs took steps (which turned out to be inadequate in the event) to protect the UST price. For example, it had previously purchased nearly $3 billion worth of Bitcoin that were available to provide liquidity and defend the $1 peg in the event of a market dislocation.[9] According to various sources, it appears that Terraform Labs sold much of this Bitcoin in attempt to raise the price of UST.

The Meltdown

In short, the ability of UST to maintain its stability was dependent on market forces and on the accuracy of the models on which UST was based to measure the impact of those forces as they changed. When those models turned out to be inaccurate in the face of unexpected market movements, the system itself failed. On Saturday May 7, there was a planned withdrawal by Terra of $ 150 million in UST from one pool in order to prepare for the launch of another pool.[10] However, this was followed by a large $350 million sale of UST (for USDC) that caused a significant imbalance with respect to the UST in the existing pool.[11] (Some have speculated that the UST sale was an “attack.”) The large sale led to selling pressure on UST, causing a downward spiral. Arbitrageurs redeemed UST for LUNA which they immediately sold, leading to a significant decrease in LUNA’s price, necessitating more LUNA being minted for each UST burned, thus creating a massive inflation in LUNA’s supply and driving its unit value down, ultimately to zero.

In response, the Luna Foundation Guard initially deployed nearly $1.5 billion of capital from its reserves in an effort to add liquidity to the ecosystem and protect the peg.[12] However, it was not enough. In fact, the deployment of Bitcoin from the reserves caused further selling pressure as the broader crypto market dropped along with UST. The situation continued to worsen, and LUNA Foundation Guard then sold virtually all of its remaining Bitcoin.[13] Once LUNA hit zero, most exchanges suspended further trading of LUNA, foreclosing any ability to recover.

The Instability of Algorithmic Stablecoins

In light of the UST debacle, it appears to be an open question as to whether algorithmically pegged coins should carry the moniker “stable.” Other algorithmic based projects, such as Iron Finance in June 2021 and Basis in late 2020, have also suffered significant disruptions. Unfortunately, because of its significant market cap, Terra’s losses were far more painful, with over $30 billion of value being destroyed in a single week. This significant market impact, in such a short period of time, appears to be the opposite of stability. Indeed, when algorithmic models have failed in other financial contexts, the consequences often are both severe and swift.

For example, more than 20 years ago, Long Term Capital Management (LTCM), a highly leveraged hedge fund, relied on mathematical models to closely predict the movement of bond spreads and trade on how the spreads narrowed, referred to as “convergence trading.”  Among LTCM’s advisors were Nobel Laureates Myron Scholes and Robert Merton. LTCM enjoyed very high returns for several years but was driven out of business by unexpected market risks arising out of the Asian and Russian debt crises of the late 1990s. To mitigate losses, LTCM was ultimately bailed out by a consortium of banks (making it distinct from the facts at hand). In short, no matter how complex and seemingly well-vetted the model, market forces have repeatedly caused unexpected and outsized risks to algorithmic models that may not be able to be mitigated within the constraints of the model.[14] Thus, it would appear more appropriate to refer to instruments such as UST not as “stablecoins,” but perhaps more accurately as algorithmically pegged coins (“APCs”).  Whatever terminology one chooses, however, implications of stability should remain outside the definition.

Reserve-Based Stablecoins

By contrast, other digital assets carrying the stablecoin label appear to have characteristics much closer to the commonly understood meaning of stability. Such stablecoins are backed by reserves that are a mix of cash and other liquid assets which match the value of the stablecoins issued on a 1:1 basis; that is, when a claim for redemption is made, cash or other liquid assets from the reserve can be deployed so that a stablecoin valued at a dollar can be redeemed for $1 in U.S. currency.

A stablecoin of this nature tends to hold its peg and is less subject to market fluctuations than those based on algorithms. They are not based on the ability of the algorithmic model to withstand stress but rather on the value and liquidity of the assets held in the reserve accounts. Of course, these values can fluctuate, but with proper reporting, the fluctuations also can be observed, thus enabling risks to be mitigated by, for example, substitution of one reserve asset for another. These are therefore inherently less risky than APCs.

As an illustration, we can consider the graph appearing at Figure A. This graph compares the price movement of four popular fully-collateralized stablecoins,to the movement of UST during the second week in May. Figure B and Figure C show the movement of those collateralized coins over a longer time frame. Notably, each of these fully-collateralized stablecoins has voluntarily undergone independent certifications to make transparent the contents of their reserves. In these cases, the facts suggest rather strongly that transparency does appear to have a relationship to stability.

Response to the UST Meltdown – The Calls for Heavy Regulation and Its Challenges

Not surprisingly, in the immediate aftermath of LUNA's price hitting zero and many exchanges halting its trading, there were calls for broad and aggressive regulation not only of stablecoins but of the entire digital asset ecosystem. For example, Sen. Pat Toomey (R-PA) had already released a discussion draft of a bill proposing stablecoin regulation;[15] Treasury Secretary Janet Yellen noted to Congress the need for expanded regulation of stablecoins because of their “risks to financial stability”;[16] Sen. Cynthia Lummis (R-WY) announced that she will soon release a proposed bill that will contain provisions in connection with stablecoin regulation;[17] and numerous commentators used the opportunity to broadly condemn digital assets in general. While usually well intended, some of these calls significantly overshoot the mark by, first, confusing the differences between an APC and a reserved stablecoin and, second, by inflating the impact of the Terra Luna event beyond its causes and effects.

First, as noted above, the events surrounding the Terra Luna demise are neither unique nor do they amount to a crisis. As far as major market turmoil goes, banks escaped the UST meltdown largely unscathed. It did not pose a risk to the financial stability of the banking system or any related markets. The harm remained confined to the world of digital assets and there was no “contagion,” not even to other stablecoins (see Figure A). Second, the market is acting in ways that are familiar to participants. Redemptions appear to be more prevalent in stablecoins that are less transparent about their reserves than others that hold primarily cash and government securities. This is similar to the flow of funds from prime money market funds to government money market funds during the 2008 and 2020 crises. This tends to underscore the fact that stablecoins with high quality reserves are considered to be safe harbors in this environment. Third, and importantly, no redemptions have gone unpaid, and unlike the case with prime money market funds in 2008 and 2020, no government rescue plan has been necessary (or even considered) and no government funds or guarantees have been offered to back up or “bail out” stablecoin entities or investors.

Thus, the response to this event should fit the actual record. Terra was not a systemically important institution, and its failure did not have a knock-on impact to any of the significant pillars of the financial system.[18]  Hence, the prescription for mitigating this risk going forward should be proportional to the actual failure, fit the actual illness and its symptoms, and not address healthier parts of the ecosystem. In this case, the source of the illness was not stablecoins in general or the digital asset system as a whole. It was an algorithmic model which failed under stress. This is not the first time such a failure has caused financial losses and it will not be the last. However, attempting to prescribe intrusive regulation on other parts of the crypto eco-system as the only possible cure is akin to providing medical attention to the wrong body part.

Useful lessons can be derived from similar past failures which can be applied going forward to better inform investors and avoid similar risks. For example, one of the fundamental building blocks of any risk management system (RMS) is a robust model risk management (MRM) effort. Any MRM protocol includes basics such as the identification, monitoring, quantification, and governance of model risks, as well as the development of sensitivity tools to measure the impacts of changes to internal and external factors, including the use of “what if” analyses. Existing tools such as Bayesian statistics, machine learning, and artificial intelligence can be employed in generating scenarios to highlight a model’s potential weaknesses. Methods and means like these can be employed voluntarily in order to strengthen the risk environment in connection with APCs, to drive further transparency, and to restore confidence in such instruments and their utility.

In addition to the development and adoption of MRM tools for APCs, transparency, like sunlight, often can be the best source of renewed health. In this case, providing full knowledge of the type and quality of reserves for stablecoins likely would drive additional confidence into the system. Such disclosure can be composed of three, and possibly four, different elements: first, a disclosure of all reserve holdings at the beginning and the end of each month, including an analysis of the percentage of reserves composed of daily liquid assets and weekly liquid assets (i.e, those which can be readily converted to cash within one or five business days, respectively)”[19]; second, a disclosure of all changes to those holdings that have occurred intra-month (preventing any “switching” to more high-quality liquid assets near the end of the month); and third, the provision of an attestation from an independent accounting firm. Fourth, similar to the obligations taken on by senior executives under Sarbanes Oxley with respect to financial statements and controls (which are far more complicated than the composition of a stablecoin reserve), it may be healthy to consider certifications by an issuer’s senior management as to the accuracy of those monthly disclosures, subject to SOX-like penalties for false certifications.  

Such clear disclosures, both for APCs and for true stablecoins, will help address the risks inherent to APCs and will prevent harm to small as well as large-scale purchasers of stablecoins. These sorts of disclosures were lacking with regard to Terra Luna, and had they existed, may have had a beneficial impact on the market’s perception of risk. As such disclosures become more widespread in the stablecoin market, all stablecoins will be impacted, and customers will benefit. Given the similarities in stablecoin functionality, customers will be more likely to choose the one that is most fully disclosed.

Conclusion

Some commentators have noted that the need for end-to-end regulation of stablecoins is “manifest” or “urgent.” Despite the chaotic end of the UST coin, the opposite would appear to be true. Stablecoins based upon actual reserves have remained just that-- stable. Redemption requirements have been met. There has been no “contagion.” Indeed, the calls for immediate action appear to be based primarily on the old saw that “a crisis is a terrible thing to waste.” The Terra event is far from the demise of crypto. Rather, it is a painful lesson from which we will learn going forward. Remedies can be found in the understanding of the difference between algorithmically pegged coins and actual stablecoins, in the adoption of better model risk management, and in the insistence on robust transparency as to stablecoin reserves.[20]

Regulation is not required for any of these positive developments to come into effect unless the industry decides against voluntary efforts to put them in place. General calls for the banning of stablecoins or crypto display a misunderstanding of the digital asset ecosystem. Stablecoins are not so much assets but are part of, in the words of former Acting Comptroller of the Currency Brian Brooks, a “blockchain native payment system.”[21] They are not dissimilar to other payment systems that used to be more popular such as travelers’ checks, prepaid cards and wire-based money transmission devices.[22] The actual payment stablecoins themselves (those with quality reserves) have continued to operate as designed – stably -- throughout the entire Terra event. Changes can and should be made to improve disclosure and dampen the possibility of a reoccurrence. Indeed, such voluntary efforts should be encouraged by the regulatory community. However, to jump from voluntary improvements to RMS and reserves disclosures all the way to a heavy-handed licensing regime, seems both unnecessary and likely to cause delay and uncertainty, which could have a negative impact on voluntary measures as issuers could adopt a “wait and see” approach.

Much can be learned from the unfortunate destruction of the Terra Luna/UST project. However, in designing approaches responsive to that event, the industry, regulators, and legislators should ensure that the solutions fit the facts.  As to fully reserved and attested payment stablecoins, the factual record for significant regulatory intervention simply has not been made.

Figure A

Figure B

Figure C


[1] https://www.stakingrewards.com/stablecoins/; https://medium.com/illumination/how-to-stake-ust-stablecoin-to-earn-20-interest-2ffae899e5fa#:~:text=Create%20an%20account%20and%20wallet,post%20your%20UST%20into%20Earn.

[2] https://www.coinbase.com/learn/crypto-basics/what-is-a-stablecoin

[3] We note that while Tether is often considered the largest “stablecoin” by volume, it is not fully collateralized by high-quality liquid assets according to its website- https://tether.to/en/transparency/#reports. However, it does publish data about its reserves that is available for consideration.  Therefore, it does not fall under the category of fully collateralized stablecoins that we are discussing in this article. We also do note that although Tether very briefly (at about 3:00 am) dropped to 95 cents during the fallout of the Terra Luna collapse, it only stayed that low for approximately one hour and quickly regained its peg, which it has maintained to date. https://www.bloomberglinea.com/2022/02/17/were-not-revealing-our-secret-recipe-tethers-cto-says/; https://www.cnbc.com/2022/05/12/tether-usdt-stablecoin-drops-below-1-peg.html https://www.google.com/imgres?imgurl=https%3A%2F%2Fwww.ft.com%2F__origami%2Fservice%2Fimage%2Fv2%2Fimages%2Fraw%2Fhttps%253A%252F%252Fd6c748xw2pzm8.cloudfront.net%252Fprod%252Fdbe4e290-d206-11ec-ad39-7db7bbae9143-standard.png%3Fdpr%3D1&fit%3Dscale-down&quality%3Dhighest&source%3Dnext&width%3D700&imgrefurl=https%3A%2F%2Fwww.ft.com%2Fcontent%2F5887ef43-d43a-4608-a1ac-aacc99f076b9&tbnid=AT-C8Hf6qy9eGM&vet=1&docid=MN7t027TVdjnBM&w=700&h=500&source=sh%2Fx%2Fim

[4] https://docs.terra.money/docs/learn/protocol.html

[5] https://docs.terra.money/docs/learn/protocol.html

[6] https://paxos.com/2022/05/13/understanding-the-demise-of-ust-and-what-makes-a-stablecoin-stable/

[7] According to Kevin Zhou of Galois Capital (a Terra Luna critic), the money to finance these returns came in part from a large amount of Luna tokens held at Terraform Labs which would be sold to fund operations and to support the anchor protocol’s 20% yield reserve. (https://www.bloomberg.com/news/articles/2022-05-15/luna-crash-this-crypto-insider-warned-of-terra-ust-s-collapse)

[9] https://www.cnbc.com/2022/05/16/what-happened-to-the-bitcoin-reserve-behind-terras-ust-stablecoin.html

[10] https://twitter.com/stablekwon/status/1523315720012001282; https://www.coindesk.com/business/2022/05/08/ust-briefly-loses-peg-luna-drops-10/

[11] https://fortune.com/2022/05/13/terra-ust-stablecoin-crash-suspicious-potential-attack-george-soros/

[13] https://www.cnbc.com/2022/05/16/what-happened-to-the-bitcoin-reserve-behind-terras-ust-stablecoin.html. According to a May 16 post on Luna Foundation Guard’s Twitter page, on May 10, Luna Foundation Guard sold all but 313 of its Bitcoin. https://twitter.com/LFG_org/status/1526126703046582272.

[14] While algorithmic models are broadly useful, every such model can be subject to some form of tail risk. That underscores the importance of model risk management discussed infra. As John Maynard Keynes once said: “The market can remain irrational longer than you can remain solvent.”

[15] https://www.banking.senate.gov/newsroom/minority/toomey-announces-legislation-to-create-responsible-regulatory-framework-for-stablecoins

[16] https://www.cnbc.com/2022/05/13/regulators-anxious-about-stablecoins-like-tether-after-ust-collapse.html; https://blockworks.co/secretary-yellen-stablecoins-pose-significant-risk-to-financial-stability/

[17] https://www.law360.com/capitalmarkets/articles/1494379/sen-lummis-teases-bill-addressing-crypto-and-stablecoins

[18] Notably, the Federal Reserve’s May 9, 2022, Financial Stability Report (https://www.federalreserve.gov/publications/files/financial-stability-report-20220509.pdf) indicates that levels of High-Quality Liquid Assets (HQLAs) at U.S. banks are dramatically higher across the board and have increased significantly since 2020. At the same time, banks have extremely low levels or reliance on of short -term wholesale funding. By comparison, fully reserved stablecoins hold even more liquid assets and have little to no reliance on short term wholesale funding. The Terra event had no impact whatsoever on liquidity profiles either in the banking system or at fully reserved stablecoins.

[19] See generally SEC Investment Company Act Rule 2a-7.

[20] Some commentators have even suggested that a reserved stablecoin can lose its peg. While almost anything is possible, speculation should not be the basis of regulation. Such a scenario would be difficult to imagine in cases where reserves were held against all coins issued and were composed of cash, short term government securities and HQLAs, all of which are subject to a daily and weekly liquidity assessment.

[21] https://www.youtube.com/watch?v=fFbKOFQqxx8

[22] https://www.youtube.com/watch?v=fFbKOFQqxx8