Not all Stablecoins are made equal.
A titanic shockwave recently shook the broader crypto world. It originated from heavy price movements on UST, the US dollar stablecoin built by Terra. This event urges us to remind cryptocurrency users that most Stablecoins differ by nature and that each design comes with its risks and costs. We believe that stablecoins must not rely on assets having systematic ties with them, and I will explain our stance in this article.
Families of Stablecoins
In the context of asset-pegged Stablecoins, we can form three distinct families.
- The (mostly) fiat-based ones: Like USDC,
- The soft-pegged, over-collateralized ones, like PAR or DAI,
- The algorithmic ones, like UST.
Fiat-based stablecoins are tokens issued by a central authority against collateral deposited in a bank account.
Such stablecoins require their environment to be friendly to be fully secured. Friendly in this context means non-negative interest rates and the presence of a network of banks willing to keep accounts open for companies doing business in cryptocurrency. As an entity ready to issue such a stablecoin, you store fiat money or cash equivalents into a bank, you get paid for it, and for each unit of currency stored in the bank account, you provide a token on the blockchain. Provided you can get paid for the operation; you have a profitable business. However, if the environment is such that storing the funds comes at a cost, you have to become creative to back your tokens (you need to channel revenue in your business), which therefore means having your token holders embrace the risk strategy, which can turn your business profitable. Fortunately, the US dollar has a friendly environment for such business.
Such Stablecoins have a central issuing authority and come with functions that allow the issuer to manage the tokens. A recent example of the public use of such a function is the event where Tether has frozen the USDT stolen by a particular attacker in a hack.
These Stablecoins are the backbone of the crypto industry and the reason for the growing trust in it. They come with the high cost of centralization for their users and partners, making them unsuitable for being a pure backbone for DeFi, but they are nonetheless crucial to our ecosystem’s health. In a sense, they are the middle ground that the community has accepted to live in a decentralized world still connected to the traditional economy.
Our industry has created stablecoins to provide an alternative to leaving the cryptocurrency world while wanting to escape the volatility of crypto native assets such as Bitcoin. What users wanted was not necessarily a backing guarantee but assets that they could believe would not diverge from fiat, and a fiat backing was the most straightforward way to reach this goal.
Soon, some up-and-coming players in the crypto community decided that the fiat-based approach’s tradeoffs would hurt the industry’s long-term decentralization goals and started providing alternatives that didn’t need to rely on banks and friendly environments. The algorithmic Stablecoins and the overcollateralized, CFD-like Stablecoins.
Decentralized and overcollateralized stablecoins
Decentralized overcollateralized stablecoins were born from the wish for more decoupling with any off-chain ecosystem. The mechanism behind these stablecoins came up with MakerDAO, and it is straightforward: every stablecoin is a form of debt to the protocol and must have sufficient collateral backing.
The predicate is that for a stablecoin to retain its value safely, it must have some other asset it can be redeemable for at any time. The sale of that asset must provide enough immediate returns to cover the price the token is trying to protect. Therefore the asset, which is then called the collateral, backs the stablecoin.
Unlike centralized solutions where an entity buys collateral and mints stablecoins, decentralized systems allow users to bring in collateral in exchange for the right to mint stablecoins. When users come back to the system, they bring back their stablecoins. The system burns the stablecoins and allows users to get back their collateral.
If both the collateral and the stablecoin are crypto assets, such a system sounds safe but doesn’t add much value unless the two are not correlated or can provide access to different utilities. Why would a user deposit a dollar stablecoin to mint another dollar stablecoin?
The answer of the protocol builders has been to accept volatile crypto assets as collateral (generally in addition to stablecoins) and create ecosystems to reward the use of the minted stablecoins, providing them with some utility. The ecosystems’ quality and rewards vary greatly and could be the subject of another blog post.
When accepting volatile collateral, the value is to allow the user to extract liquidity from their assets without selling them because they keep their exposure to the upside, but the risk is that in case of a price crash, someone must liquidate the collateral for the stablecoin to retain its value.
Decentralized overcollateralized stablecoins depend on their parent protocol’s risk management strategy, so they vary in risk levels. The risk management strategies of these coins tend to be about the debt ratios (how much more collateral do you need to provide to mint stablecoins) and liquidation incentives (how much bonus or discount do liquidators get when keeping the platform healthy), and total debt size (how much stablecoins can the platform mint using this collateral). These parameters are generally decided based on the expected liquidity and volatility of the asset and the stablecoin itself, alongside blockchain parameters like finality time (how long do transactions take to be considered final) and transaction cost, so liquidators can have a favorable environment to keep the platform reliable.
Another common risk factor for overcollateralized stablecoins is known as price manipulation. Collateral with a manipulated price could allow minting beyond what is reasonable. The scenario goes this way:
- The price of some accepted collateral, say, ETH, is $2000, which would allow minting a certain quantity of stablecoins, maybe $1500.
- An attacker sees where the platform reads the price and manipulates this source to say that ETH is worth $4000 temporarily. Attackers generally do this via borrowing enough funds to create an immediate spike and repaying them immediately after the transaction. Such loans (called flash loans) are very cheap, or sometimes free, as the user borrows and repays the loan immediately, in a single transaction, making the attack affordable.
- The attacker immediately gets the ability to mint $3000 worth of stablecoins.
- The price comes back, and the stablecoin has now lost the quality of its backing.
The best way to prevent such manipulation is to rely on a trusted source that is expensive and ideally slow to manipulate, which is unfortunately too dull and safe for some protocol developers who can get exploited. We disagree with such a stance, and Mimo does use a trusted source for its price feeds.
PAR is an overcollateralized stablecoin, and Mimo has decided that its collaterals, debt ratios, debt limits, and liquidation incentives to be very conservative, as the environment for Euro stablecoins is still nascent.
This conservative approach is commonplace in the industry, but it has a cost in capital efficiency as users can not use all their assets to mint stablecoins because it would be unsafe. The research for more efficiency has led some players to build a new generation of stablecoins that could use capital more efficiently.
Algorithmic Stablecoins came from the idea that using the backing of an unrelated asset was not necessary for price management.
Lumping them into one family might not do them justice as they all have different algorithms, but they all fall under the promise of being decentralized, efficient, and reliable.
In the context of recent events, I will focus on the Terra USD stablecoin here: UST.
Algorithmic stablecoins rely on complex systems to maintain their pegs, and UST is no exception. The system is essentially composed of the coin and a market module. The market module can mint and burn both UST and Luna (Luna is the native asset of the Terra blockchain, meaning it’s the token you use to pay transactions to whoever maintains the chain alive). Their documentation clearly describes the expected mechanism for UST to maintain its peg, but I’ll outline it here to explain its risks.
The market module can mint both UST and Luna and will always provide 1 UST for 1USD worth of Luna, which should offer arbitrage opportunities to anyone outside the system. Does UST trade below peg? Bring some cheap UST and get $1 of Luna that you can now sell! Does UST trade above peg? Bring some Luna and get some expensive UST that you can now sell!
Are we seeing anything here? The assumption is that there will be constant market demand for the tokens. If the demand for both Luna and UST drops, your arbitrage opportunity becomes risky, as bringing UST will give you some Luna you can’t sell, and if UST stays de-pegged long enough for Luna to keep dropping, the arbitrage looks more profitable, so any arbitrage request will lead to minting even more Luna. Such a loop could destroy both UST and Luna.
Every non-fiat-backed stablecoin comes with some assumptions, but it is reasonable that for such a system to work, there should not be a systematic tie between the asset helping to sustain the peg and the one that needs pegging. Following this logic, Anchor sounds risky, but it was the most popular DeFi ecosystem, which could appear surprising.
The reason is Anchor. The platform offered an almost 20% annual yield on UST deposits. Anchor has designed the yield to come from loans’ interest rates, but it ended up being funded by a reserve. The same reserve supporting the users’ profit was supposed to help the stability of UST in the case of a crash.
Long story short:
- The crash happened.
- The reserve got emptied.
- Both Luna and UST crashed in a seemingly endless loop.
At Mimo, we believe that stablecoins should not rely on assets that have systematic ties to them. No matter whether they are algorithmic. Overcollateralized stablecoins can run the same risk (remember IRON and TITAN).
Our role at Mimo is to provide reliable stablecoins and educate users who aren’t familiar with the cryptocurrency ecosystem. We aren’t going for a fast-growing stablecoin that will die at the first attack as we intend to deliver consumer-grade products, but we also want to remain competitive in the DeFi ecosystem, so we have built our protocol to allow much flexibility without sacrificing safety. Our safety reserve provides help beyond what the debt ratios can help with, we can set our debt limits to very low for the riskier assets, and we use reliable sources to fetch the price of the collaterals we accept, limiting the possibility of manipulation.
We are going beyond this and will offer innovative ways for all of our users to enjoy a safe and rewarding experience when depositing collateral and using PAR.