Comparing Terra/Luna’s Failed Stablecoin Design to Other Stablecoins

Comparing Terra/Luna’s Failed Stablecoin Design to Other Stablecoins: Condensed

Jon Vokal

George Mason University

Stablecoins, a popular blockchain-based technology within the cryptocurrency industry, are tokens designed to be pegged to the value of some other asset, most commonly the United States Dollar. There are a variety of ways to design a stablecoin to achieve this peg. Some of these designs have failed, causing the token to “depeg” from the underlying asset. The most notable of these failures is Terra (UST), a USD pegged “algorithmic” stablecoin by Terraform Labs. Terra was pegged to USD not by reserves, but by a second token, called LUNA, meant to absorb volatility and stabilize the value of the coin. Terra depegged on May 9th, 2022, less than 2 years after its launch. It essentially hit zero on May 13th, and was shut down on May 27th. This paper will compare the designs of Terra to other popular stablecoins, such as US Dollar Coin (USDC), Tether USD (USDT) and MakerDAO/DAI (DAI) to assess why Terra-Luna failed while others have not.

USDC & USDT

USDC is a stablecoin distributed by the payments company Circle. The design and structure of USDC can be compared to a bank or warehouse that runs on the Ethereum Network. Circle backs every USDC in circulation 1:1 with a mixture of federal reserve notes and US treasuries (Circle, n.d.). Circle, unlike other stablecoins, is regulated as a money transmitter and undergoes regular audits that are filed with the SEC. Circle also publishes monthly attestation reports. As of their 11/10/2022 report, Circle held $11.1B in cash and $32.6B in treasures ($43.8B in total) to match the $43.6B USDC in circulation.

This model closely resembles that of a bank. The bank takes client assets into their reserves, and makes loans in order to generate interest on those reserves. A simplified banking balance sheet can be denoted as:

R + L = D + N + K

Where R = Reserves, L = Loans for assets, where liabilities are D = Deposits, N = Notes in Circulation, K = Equity Capital.

If a bank’s liabilities are greater than its assets, it cannot repay its depositors and is therefore insolvent. Circle’s transactions execute on the blockchains it functions on, rather than on a private or centralized system like a bank, or other transaction services such as Paypal. As long as this remains the case, and the blockchains USDC executes on remain functional, then it is unlikely USDC would fail. One risk to Circle, however, is that they are not a member of the banking system, and therefore they do not rely on the federal reserve to act as a lender of last resort.

Tether, or USDT, boasts a similar design as USDC. However, there is greater alarm pertaining to Tether’s solvency than USDC, even though Tether makes similar promises. Tether claims to back its assets 1:1, users can redeem their USDT for fiat currency at any time using tether.io, and Tether undergoes regular audits to show that their assets match their liabilities. The reason why USDT is under greater scrutiny is because Tether is not as transparent about the state of their reserves. Unlike USDC, Tether states all USDT is backed by cash or cash-like equivalents, but has not previously elaborated on what exactly those “cash-like equivalents” are. Assuming their auditors are not participating in or ignoring fraudulent behavior, Tether has taken steps to increase their transparency. As of December 2021, Tether has posted quarterly audits that break down the “cash-like equivalents” in their reserves. Tether’s September 2022 shows that they hold $39B in US treasury bills, $7B in money market funds, $6B in cash, $3B in reverse repo agreements, $3B in corporate bonds, $2.6B in other investments, and $6B in secured loans. The assets held total $65.7B. The total number of USDT in circulation is roughly $65.5B (Tether Limited, 2022). While there are still some questions about the truthfulness of their reporting versus Circle, Tether’s reporting as of 2021 does make it appear that the company is solvent. However, Tether does not file these audits with the SEC like Circle does, and is not regulated as a money transmitter. Fraudulent behavior has occurred with several cryptocurrency firms and there is always risk to holding assets with these firms, even if they claim they are solvent and undergo audits.

Using the same general principles to assess bank solvency, both Circle and Tether appear to be in sound financial standing. Additionally, both companies have developed technology that are interoperable with dozens of popular blockchain networks. Because these stablecoins are not “algorithmic” pegs to the United States Dollar, but rather issued on a basis of reserves, the underlying blockchain technology that executes the transactions is simple. Both USDC and USDT would not be expected to fail if they maintain their current standards and continue to back their issued tokens 1:1 with liquid assets. One reason that other firms may want to adopt different models for their stablecoins is that Circle and Tether are both centralized. Many proponents of cryptocurrency maintain that one of the most critical value propositions of the technology is that the systems are decentralized. LUNA and DAI use different designs for their stablecoins that both claim to be decentralized. This claim of decentralization can be easy to make, but proves difficult to properly implement.

Terra/LUNA

Terraform Labs launched its algorithmic stablecoin, Terra (UST), in September of 2020. Terra achieves its peg to the United States Dollar through miners absorbing demand volatility for UST. This is accomplished by miners staking Terra’s sister coin, LUNA, which in turn is used to mint UST.  From there, price fluctuations in UST can be resolved by the miners arbitraging the delta from the target price. For example, an arbitarguer can extract a profit when 1 TerraSDR = .9 SDR by trading TerraSDR for 1 SDR’s worth of LUNA, instead of the .9 SDR they would receive for it on the open market. When 1 TerraSDR = 1.1 SDR, they can trade 1 SDR worth of LUNA to the system to receive 1.1 SDR of UST (Terraform Labs et al., n.d.). This implies that demand volatility for UST is removed from the price by the minting and burning of LUNA supply in the system. This concept is similar to that of Maker/DAI, but there are issues with this design that actually make UST/LUNA function quite differently. First is the need for stable mining/staking demand and increasing UST demand, second is the introduction of leverage with LUNA as the only collateral.

1. Need for stable mining demand

Stable mining demand is required for the protocol to function properly, which Terraform Labs originally achieves through two methods, transaction fees and seigniorage paid by burning LUNA. The transaction fees vary from .1-1% based on transaction volume. Seigniorage is earned through LUNA burn. When demand for UST increases, the system mints new UST and receives LUNA in return. The system then burns LUNA, making mining power from LUNA staking more scarce, increasing the rewards paid by staking and transaction fees. Thus, the profit function earned for miners is as follows:

Where t is the work period.

In theory, the total rewards paid should increase as the economy on the network grows, with more transaction fees going out to miners and LUNA supply being burned. The inverse for both is true when the economy shrinks. This concept begins to illuminate a primary concern for this design: rising UST demand is the only way for profit from mining rewards to increase. Terraform Labs attempted to compensate for this by introducing “stability levers” that are adjusted to control for transaction fees and mining rewards to the amount of activity on the network. The equation for this is as follows:

Where f is transaction fees, b is burned supply, (1+g) is a growth constant, and R is mining rewards paid.

This implies that fees and supply burn increase inverse to mining rewards. If network activity decreases, fees and supply burn would increase. Terraform Labs claims this is a stabilizing mechanic that will allow for mining rewards to remain constant, even amidst heavy demand volatility on the network. They provide the following charts, which they claimed were subject to “extensive simulations and stress testing”, to demonstrate the concept:

Terraform Labs mention in their whitepaper that the target growth rate for staking rewards is 15%. This highly attractive rewards model was meant to keep mining demand stable. What this design overlooks is that even with low transaction volume, this model would still require continuous new liquidity flowing into the ecosystem to keep miner payouts stable. The second chart even implies a LUNA supply burn rate of over 100% to stabilize mining rewards during a long bear market, which isn’t sustainable.

1. Introduction of leverage with LUNA as the only collateral

Terraform labs making LUNA the only collateral to its ecosystem was most likely the reason this design collapsed on itself. The only reason the LUNA token had value is that it was required for mining/UST stability and because of speculators purchasing the token. This became especially risky when Terraform Labs launched the Anchor Protocol, a platform built on Terra that allows investors to earn high yield on staked deposits and borrowing against their crypto holdings. Many products on the platform offered 20%+ APY for yield farming. Just like the mining mechanics, this design can only be functional if there is continuous new liquidity coming into the system to pay out the people that were there earlier. This led to UST’s market cap growth outpacing the market cap of LUNA in early 2021. The LUNA/UST market cap ratio peaked at over 6:1 in November of 2021, but the immense amount of yield paid out in UST from the Anchor Protocol led to this ratio being less than 1.5:1 the month before the collapse (Alden, 2022). Terraform labs would either need to keep injecting liquidity to pay these yields, or let the yields fall, reducing demand for UST.

Cyrus Younessi, head of risk at MakerDAO, pointed out many of the risks of Terra/LUNA’s design in 2018. “Terra is like the Maker model, if the only collateral you could post to create DAI is the MKR token itself… If Terra were to fall and break the peg, then it would depend on LUNA to save Terra. But, Luna would fall as investors would panic, and then Terra would continue to fall’ (Younessi et al, 2022). The “death spiral” that Younessi describes is exactly what occurred in May 2022 that led to the ecosystem's demise. UST depegging caused LUNA demand to decrease, and the panic set the collapse of the system in motion.

Unlike Maker/DAI, the entire ecosystem's collateral was built entirely on itself. There were no other tokens other than UST/LUNA that were holding up the system, and the promises made to investors and miners alike were unsustainable. If such an unwise loan with poor collateralization were to be made on Maker/DAI, they would be liquidated before any systemic risk is posed toward the system. Additionally, there are no 20%+ yield farms paid out in DAI. The primary profit opportunity is arbitrage, not high-yield staking. Terraform Labs identified this risk the month before the collapse, and launched the “Luna Foundation Guard” to hold over 80,000 Bitcoin (at the time over $3B) as a second layer to defend the peg. They bought no USD, USDC, or USDT. This entire position was liquidated trying to defend the peg. The Terra/LUNA ecosystem was not decentralized, it did not back its UST assets 1:1, and it was highly leveraged with insufficient collateral before its failure.

Maker/DAI

MakerDAO and DAI, like USDC, are ERC-20 tokens that are native to the Ethereum blockchain. The original design for DAI stabilizes as a “soft-peg” to the United States Dollar through a mechanism called a collateralized debt position, or a CDP for short. Originally, the only collateral one could put up for a CDP on the system was Ethereum. The ecosystem now allows for multi-collateral CDPs that accept BAT, USDC, and wBTC as additional options. These loans are over-collateralized, meaning that one must lock significantly more ETH in the CDP than there is DAI minted from the CDP. Maker is the governance token for the Maker/DAI system, and is not a part of the CDP system in order to avoid the possibility of a “death spiral” or other conflicts of interest. DAI is not an algorithmic stablecoin, as there are no equations trying to smooth the mining nor minting incentives to control supply on the platform.

DAI is pegged to the United States Dollar because there are profit opportunities from arbitrage as DAI rises above or falls below the value of $1. The cost of borrowing DAI becomes more expensive when the price falls below $1. This decreases the supply, causing the price to return to its target. Inversely, the cost of minting new DAI reduces when the price rises above $1, increasing the supply and causing the price to return to its target. The profit opportunity arises because those who hold a CDP can pay off their loan at a discount when DAI falls below $1, taking the necessary DAI out of circulation in order to cause the price to increase. When the price is above $1, ETH holders can create DAI for $1 but sell it for greater than $1, increasing the supply (Walters, 2022).

If the value of the ETH locked into a CDP were to drop, the user would have the option to pay off their debt at a premium to exit the CDP, or be liquidated of their position plus a 13% penalty. “Keepers” on the system run algorithms in order to spot risky CDPs to liquidate. MakerDAO claims that this system can withstand price volatility of almost any kind, suggesting that the system will remain functional as long as Ethereum is worth more than 0 and that the price of Ethereum doesn’t drop 50% within 5 minutes (Walters, 2022). As volatility picks up in the market, the collateral ratios for a CDP begin to increase. This is a clever defense mechanism of the system, whereas Terra/LUNA was more concerned about defending miner profits through bearish market conditions.

Maker/DAI is also the most decentralized stablecoin system of the four covered. First, there is no high-yield staking involved that requires MakerDAO to go searching for outside liquidity to try to collateralize the system. CDP holders are paid .5% APY on their loan, which transaction and liquidation fees cover with ease. Additionally, anyone that holds MKR tokens can participate in the governance system, making it a true decentralized autonomous organization. Even the process of liquidating bad CDPs on the platform is decentralized. By combining a decentralized system with proper incentives, Maker/DAI seems to have achieved what the other 3 designs have not: a properly collateralized, decentralized stablecoin.

Conclusion

Like any bank, exchange, or business, the primary defining factor for the success of DAI, USDC, and USDT is solvency. Technological design is a factor for success, but solvency is vital. The designs of these three tokens make sure that assets in the system match liabilities, where Terra/LUNA did not do proper diligence to ensure this. The promises made to miners, LUNA stakers, and yield farmers were far too unrealistic for the algorithmic design to be viable. It would not be outrageous to suggest that the design of Terra/LUNA and the actions of Terraform Labs were a ponzi scheme. While Circle and Tether are not decentralized systems, both firms demonstrate that they have taken the necessary steps to ensure that the currency in circulation matches that of the assets they hold in reserves. Maker/DAI ensures that the system is solvent through over-collateralization of the CDP loans, while also going great lengths to achieve a decentralized system. Assuming the design of USDC, USDT, and DAI remain the same, one could expect that these systems are unlikely to fail in the catastrophic manner that Terra/LUNA did.

Works Cited

Alden, L. (2022, May 15). Digital Alchemy: A Post-Mortem of the Crypto Crash. Lyn Alden. Retrieved November 25, 2022, from https://www.lynalden.com/digital-alchemy/

Circle. (n.d.). USD Coin (USDC) | Crypto that’s held to a higher standard. Circle. Retrieved November 22, 2022, from https://www.circle.com/en/usdc

Sandor, K., & Genç, E. (2022, May 20). The Fall of Terra: A Timeline of the Meteoric Rise and Crash of UST and LUNA. CoinDesk. Retrieved November 25, 2022, from https://www.coindesk.com/learn/the-fall-of-terra-a-timeline-of-the-meteoric-rise-and-crash-of-ust-and-luna/

Terraform Labs, Kereiakes, E., Kwon, D., Maggio, M. D., & Platias, N. (n.d.). Terra Money: Stability and Adoption. Terra Money: Stability and Adoption. Retrieved November 25, 2022, from https://assets.website-files.com/611153e7af981472d8da199c/618b02d13e938ae1f8ad1e45_Terra_White_paper.pdf

The Maker Team. (2017, December). The​​Dai​​Stablecoin​​System. MakerDAO. Retrieved November 26, 2022, from https://makerdao.com/whitepaper/DaiDec17WP.pdf

Tether Limited. (2022, November 18). Transparency. Tether. Retrieved November 23, 2022, from https://tether.to/en/transparency/

Tether Limited Inc. (2014). Tether White Paper. Tether: Fiat currencies on the Bitcoin blockchain. Retrieved November 23, 2022, from https://assets.ctfassets.net/vyse88cgwfbl/5UWgHMvz071t2Cq5yTw5vi/c9798ea8db99311bf90ebe0810938b01/TetherWhitePaper.pdf

Walters, S. (2022, July 7). What is The DAI? Complete Beginners Guide to Stablecoins. The Coin Bureau. Retrieved November 25, 2022, from https://www.coinbureau.com/education/what-is-dai-coin/