A Stablecoin Framework
mStable’s first Meta Vault, to be launched soon, will be stablecoin-based. Related literature is abundant but frameworks to assess their risk are missing.
mStable is a battle-tested project with a strong focus on security and an upcoming stablecoin-centric product release. Therefore, we created a Stablecoin Risk Framework to assess the depegging risk of the different possibilities.
Stablecoins have been the biggest narrative in DeFi for a long time and building a truly scalable and decentralized stablecoin remains a challenge. The UST crash has taught us that nobody is too big to fail. Designs are getting more complex over time and it is becoming increasingly hard to stay on top of them and assess their risk. But that also means that the market might be mispricing the risk.
At mStable we are building a generalizable vault architecture of which stablecoins are a central piece. Very soon, mStable is deploying the first Meta Vault, which will enable investors to get exposure to a diversified set of strategies with a single transaction. These strategies will be stablecoins pooled in Curve against 3CRV which are then staked in Convex to claim $CRV and $CVX rewards. We developed a framework to asses the risk of these pools which will allow us to make an informed, data-driven decision about the selection for our vaults. Eventually, this framework will become the guide we will use (and improve) for any subsequent new vault designs. First, a few disclaimers:
- We are only focusing our analysis on crypto-backed stablecoins. Fiat backed stablecoins risk comes from issuer risk and we don’t have the tools to rate any of that apart from what they are publicly disclosing.
- We are assuming that USDC, USDT, and DAI are relatively risk free. Any risk associated with these three will very likely be inherited by the rest.
- Nothing in this article is financial advice, it’s more of a research piece illustrating how we conceive and value stablecoin risk. Always do your own research and make your own investment decisions.
This framework was developed by deeply researching the different stablecoin designs and understanding where the potential pitfalls can come from for each of them. We assessed not only the risk of a stablecoin fully depegging but also the possibility of them temporarily slightly depegging, because while it’s not as harmful as an UST-to-zero event, it’s something that would eventually damage the yield. We then lined up the common points in all of the stablecoins and created a set of basic rules which we will probably be constantly iterating as the market develops.
mStable strongly believes in Open Source and Decentralization, so we are opening it up. We want to receive feedback from the community and enrich the framework with it. Now let’s jump to the interesting part.
General rules
For each stablecoin studied we chose to apply a set of 4 or 5 rules (depending on the design). Each rule generates a specific metric which is then weighted into a final score. Below, we lay out the explanations of the different rules, provide detail on how points are assigned, and then show the results for the analysed stables in a separate spreadsheet.
- Bad debt/ circulating supply (2 pts): stablecoins backed by collateralized debt positions (such as DAI or LUSD) can be exposed to incurring bad debt. This happens when the price of the collateral asset falls quicker than the liquidations. The protocol cannot cover their debt and a portion of the circulating supply remains unbacked.
- Hard asset backing (2 pts): many stablecoins are backed by other crypto assets. For these, we assessed the percentage of the redeemable supply that is backed by hard assets which we defined as ETH, BTC, and other stablecoins. We defined redeemable supply as the total circulating supply minus protocol owned liquidity and locked liquidity. Liquidity provided but not locked is still redeemable, as providers can pull it out.
- Fully redeemable (stables that base their pegging mechanism only on redeemable assets):
- Partially redeemable:
- Effective Peg Stability Module (PSM) (1 point): having a short term peg stability module is important to ensure that the stablecoin is always trading close to the peg value. An effective PSM is one that allows for 1 stable to be instantly redeemable for 1 USD worth of ETH, BTC, USDT, USDC, or DAI
- Lindy effect (1 point): relative to how long they have been on the market and how battle tested their design is. Adding points to those that survived bank runs and decreasing if they have suffered a hack and/or incurred in bad debt.
- Share of USDC+USDT+DAI liquidity in their Curve pools (1 point): Curve is the main on-chain source of liquidity for stablecoins and it acts as a backstop for preventing depegging events. The deeper the liquidity is, the less likely that it will depeg. We measure the deviation from the optimal share of the pool because it reflects the market’s risk assessment of that stablecoin. This can be derived from
1–1/numberOfAssetsInPool. For example, sUSD is paired against USDC, USDT, and DAI and not 3CRV. Therefore, it has an optimal liquidity of 75%. sUSD having 50% of USDC, DAI and USDT liquidity is equivalent to any asset paired to 3CRV having 33% of 3CRV liquidity
Results
As was mentioned before, USDC, USDT, and DAI are treated by this framework as risk free. At the same time, this framework will serve us to select the stablecoins that will be utilised to be paired against them. As a corollary, a score of 100% means that this stable is as safe as the base assets. A score higher than 80% means that this stablecoin is safe enough for our first Meta Vault so it will invest in these stable pools. For every on-chain backed stablecoin paired with 3CRV, we calculated the score (which can be verified here).
sUSD, mUSD, and OUSD scored (almost) perfectly, which is explained by the resilience shown in the past, their strong peg mechanisms and their deep liquidity.
FRAX has also shown great resilience but the imbalance of their pool costed them some points.
alUSD has a strong long term mechanism but rely on market conditions to get a strong PSM, which have costed it a few cents off-peg in the past. Additionally, their PSM is centralised, as it is up to the team to decide on when to deploy funds on their transmuter.
PUSd and LUSD suffer from being above peg. The former one has a centralised PSM that is not acting to lower the peg and the former one only has a downwards hard-peg at 1.
Finally, DOLA and MIM suffered bad debt and are now under-collateralised, which make them prone to depegging events.
Conclusion
On this first iteration, the risk framework returned that, of all on-chain governed stablecoins paired against DAI, USDC, and USDT, the stablecoins that could potentially be used by our first Meta Vault because they have a score higher than 80% are: sUSD, mUSD, OUSD and FRAX.
The results of the framework are dynamic: the shared spreadsheet will be updated every 30 days and, whenever there’s a significant change, we shall create a post on our forum informing it. The last criteria will be actively monitored, and a rebalance will be triggered when the share of liquidity drops to 25%, exiting the position with less than 1% of slippage.
Additionally, we will also consider utilising other centralised stablecoins such as BUSD, GUSD, or USDP because they have shown great resilience and do not add a significant amount of trust (considering that we are already utilising centralised stablecoins). Given that the framework cannot be applied to centralized stablecoins, a governance proposal needs to be created in which it’ll be discusses wether a certain centralised stablecoin is to be trusted.
