Thanks very much, @Llamaxyz team, for the proposal. Our data science team is continuing to analyze this proposal, but we just published an analysis on the forums on the risks of enabling LP tokens as collateral, more generally. We hope that this helps the community assess the risk/reward tradeoffs.
Forum post here.
Among other factors, the risk depends on the type of LP token listed, what the borrow asset would be, and the risk parameter configurations. In certain situations, there may be complex market risk implications even if there are conservative parameters applied (supply cap + low LT / LTV, etc.).
Gauntlet is still conducting analysis, but below are my initial preliminary thoughts on BPT for wstETH/wETH pool as an example of a collateral asset to be listed.
At first glance, this asset seems low risk because wstETH/wETH are correlated, and the use case will be to borrow wETH (so the borrow asset has high correlation to the LP token as well).
However, upon closer look, there are more complex risk dynamics that the community should consider:
- The user strategy is to deposit BPT, borrow ETH, stake their ETH, get wstETH, deposit wstETH into the Balancer pool, and repeat as much as possible to get leveraged yield.
- This may be highly lucrative to the user because not only do they get yield from staking ETH, but they also get BAL rewards, Aura rewards, swap fee rewards, etc.
- Thus, there will be high demand for borrowing ETH. This can increase ETH utilization in the V3 market.
- This presents several problems.
- First, if ETH interest rate curve isn’t high enough, then this won’t prevent high ETH utilization on V3 (the BPT strategy may be highly profitable, so the cost of borrowing ETH has to be very high until the strategy is no longer profitable), and high ETH utilization can interfere with atomic liquidations (for user positions that are using ETH as collateral). Also, there’s the standard issue of inhibiting users from withdrawing their ETH from the pool.
- Second, if wstETH is already listed on V3 at that time when BPT for wstETH/wETH is listed, then under high ETH utilization, the high ETH borrow rate can put stress on the existing wstETH/ETH yield farmers on Aave V3 (at that time) who are depositing wstETH as collateral and borrowing ETH.
- Third, if the utilization for ETH is high on V3 and the supply rate becomes attractive, this may incentivize users to withdraw ETH supply from Aave V2 and deposit it on V3 instead. The elasticity would depend partly on how the interest rate curves are configured. But if we imagine a scenario where a significant amount of ETH on V2 is withdrawn and supplied on V3 instead, ETH utilization on Aave V2 will spike. Now, the interest rate on ETH on V2 can put stress on existing stETH/ETH yield farmers on V2. You may argue that the additional supply that transitioned from V2 to V3 now alleviates the ETH utilization problem on V3, but that is not certain, and additional borrowing on V3 may fully utilize the additional supply depending on the interest rates. Additionally, you may argue that the high interest rate on V2 will attract more ETH supply on V2, but again, the elasticity is difficult to predict.
If Llama / the community proposes a list of the specific BPT tokens that are being considered, Gauntlet can help provide further analysis on the market risk side for those BPT tokens.