Gauntlet Risk Analysis
Simple Summary
- Gauntlet recommends that DEX stable pools increase the target balanced DEX concentration by 10% at each bucket level
- There are several risks Gauntlet would like to highlight for the community, that will need to be proactively managed
- If GHO < $1 and lending pool liquidity for assets used to mint GHO is low (i.e. utilization for those assets is high), then GHO repeg can slow or be difficult, especially for the set of GHO minters that are looking to completely close their position.
- Usage of emode minting GHO to restore the peg if GHO > $1 could lead to premature GHO supply explosion, especially when DEX size is not sufficient.
- Loss of liquidity in the GHO PSM, or devaluation of the assets in the GHO PSM (as seen during the USDC depeg) can lead to a significant depeg in GHO.
Gauntlet Analysis
From a risk perspective, Gauntlet recommends that DEX stable pools increase the target balanced DEX concentration by 10% at each bucket level. The initial analysis notes that usually, 15-30% of stablecoin supply is usually in DEX. However, GHO has many novel components. The fixed borrow fee, the interplay between the GHO markets and the regular lending markets (given that Aave v3 Ethereum will be the first facilitator), and the fact that GHO is a completely new market, increase the risk of extreme GHO depeg and GHO manipulation without sufficient DEX liquidity.
The community should decide at what speed is appropriate to scale GHO and determine the necessary bucket sizes and borrow rates.
We would also like to call to attention several key risk factors - particularly with regards to GHO liquidity. The community should be aware of these risks and discuss, if necessary, how to mitigate these risks.
- Aave v3 Ethereum as the facilitator.
- The interplay between GHO minting and the lending markets. Collateralized assets used to mint GHO will flow to the lending pool. Suppose GHO is trading < $1, then GHO minters can buy GHO externally to repay on the Aave v3 lending markets, helping to repeg. For those repayers that are looking to close their position as well and redeem their collateral through the facilitator, if there is insufficient liquidity on the lending markets for this collateral (i.e. the collateral has high utilization), then they may not be able to repay the GHO. An example of this is when a GHO minter needs to flashloan other assets to purchase GHO externally and needs their collateral to close out the flash loan. This could hurt the repegging.
- Because of this interplay, fixed and static fee (that requires governance to change) could lead to unmitigated GHO minting → dump when GHO is abnormally < $1. If the peg restoration mechanism is failing, then sophisticated traders can mint up to the remaining bucket capacity at the fixed rate and market-sell the minted GHO to further depress GHO prices. Until the redemption mechanism recovers and liquidity is restored we can expect GHO price to continue to fall (as seen with USDC). The trader can then buy back the GHO at a much lower price to close their position.
- Usage of emode minting GHO to restore the peg if GHO > $1 could lead to a premature GHO supply explosion, especially if DEX liquidity has not caught up. As mentioned in the initial analysis by Aave Companies, the maximum supply of GHO can be increased to bring GHO back down to par. The proposed emode implementation, at the current LT of 97.5%, enables less capitalized users at ~30x leverage to affect peg restoration greatly. The end result is that the pool will be rebalanced at the same TVL, but the global GHO supply will have expanded, decreasing the DEX GHO / GHO supply ratio. Without a PSM, GHO at this increased supply may be more susceptible to price manipulation.
- Usage of idle proposed PSM liquidity. As the initial review has indicated, a majority of DAI is minted through the Maker PSM. Recently, Maker governance has voted to onboard $100M USDC into Yearn. As such, idle liquidity in the proposed PSM for GHO will need to be managed carefully and prudently, given that yield strategies have associated inherent risks (smart contract, market, etc). Loss of liquidity in the PSM, or devaluation of the assets in the PSM (as seen during the USDC depeg) can lead to a significant depeg in GHO.
The above is assuming GHO is only released on mainnet Ethereum. Proposed GHO deployment on L2 could introduce further risks, which Gauntlet is continuing to study. We note that without more visibility into the PSM, we do not have full confidence about the pegging mechanism and consequent stability. As such, it is difficult/too early to provide more parameter suggestions for risk levers without the details on the overall mechanism.