The position has an HF of 1.66 and does not represent an immediate risk to Aave v2.However, we recommend to 1. freeze CRV collateral and 2. set Curve LTV to 0, to prevent the account from continuing to add CRV and increasing concentration risk, due to the decrease in CRV liquidity over the past few months.
CRV token liquidity has been decreased 50% over the past couple of months, both on chain and globally. This can cause future risk, especially if the CRV used as collateral by this account continues to grow.
The account has been continuously adding CRV collateral as CRV price decreases, and increasing its stablecoin borrowing over the past couple months. Given the account is actively managed and frequently maintains its health, freezing CRV will incentivize the account to reduce its borrow or add other forms of collateral.
This is an interesting case to observe because it’s possible that this debt has no intention of ever being paid off, as in it may be a form of profit taking.
Considering that the wallet is tagged as belonging to the founder of Curve, this would be a way to take profit that:
Doesn’t harm CRV market value or optics of “founder selling”
Is more economically efficient (eg. eat slippage via debt LTV vs trying to market sell that amount)
Is tax optimized (or delayed)
Doesn’t sound like too bad of a profit taking strategy from his vantage point to be honest.
It’s definitely an edge-case in terms of intended product use case here, but I think this is a fair assumption of what might be going on here. I mean he did recently purchase a $60M mansion right?
As much as I love Curve as core pillar of defi and liquidity, I think this is a fair risk assessment considering the available liquidity in the market and risk of eating bad debt until CRV liquidity grows deeper.
Hey, we have been discussing this situation a bit with @yaron from RiskDAO
Have you also considered freezing CRV on Aave v3? There’s about a $21m isolated debt ceiling available on $33m CRV supply cap.
What’s more problematic here in our opinion is that this wallet has around $110m leverage on CRV across DeFi, outside Aave there’s another $20m borrowed on Fraxlend and $20m on Abracadabra. Both positions are on 75% LTV which is quite high for such a token and Fraxlend position has the weakest health factor compared to Abra and Aave.
Another idea would also be to gradually decrease LT at Aave v2 in order to better protect against bad debt when and if this user gets liquidated at high slippage. Especially because there’s additional debt on CRV outside Aave with worse collateralization metrics and I think Aave doesn’t want to be liquidating this user as the last one when most of the available liquidity is gone. Of course whoever liquidates the user first will also make the price worse…
I also don’t see many unutilized assets that could be used for repayment of debt here, most of the available assets are CRV tokens from vesting. So if the price continues to drop and freeze is enabled the user is forced to repay debt rather than make continued top-ups and make the problem worse afterwards. Best hope is the user can actually repay debt with external assets.
General leverage on CRV needs to be reduced one way or another, you can’t have a $110m leverage built on an asset that has over 10% price slippage when selling $5m on-chain. Could be that off-chain liquidity is underestimated here but I still think the ratio is not healthy.
We should also consider increasing the reserve factor for CRV on the Ethereum V2 market to 99.9%. This position has unhealthy characteristics, and the cost to hold this position is clearly underpriced.
I agree that we should move LTV to 0. Anyone who wants to use CRV as collateral can move to the V3 market.
I dont think freezing the CRV asset brings any benefit in this case. Even if we prevent this actor from borrowing more, we also prevent refillings, which are in any case legitimate and wouldnt help maintaining the position in an healthy state to begin with.
Setting the ltv to 0 in V2 is unfortunately not enforceable. The ltv 0 check can be circumvented using flashloans (this is not possible in V3).
I don’t see any reason to freeze the asset in V3, first the exposure is not as big to begin with, second the supply caps prevents from accumulating an excessive amount of CRV collateral, third there are many levers that we can change (lower the debt ceiling, further lower the supply cap, drop the ltv to 0) that unfortunately we dont have in V2 to prevent any risk.
The increase of the V2 reserve factor for CRV is an interesting approach, the protocol is currently sustaining a massive risk allowing this position and it’s not being paid nearly enough.
@gauntletnetwork a simulation of the potential bad debt risk from liquidating this position would be very helpful, do you have anything on that?
Could you clarify this? I assume you’re not saying that LTV requirements can be entirely circumvented in Aave v2. Is there something special about the “zero” value that produces unexpected behavior, such that a low but nonzero value is needed?
To clarify, this set of recommendations is only to freeze CRV and set CRV LTV to zero on Aave V2. This proposal will not impact Aave V3.
We agree that setting LTV to 0 is only a minimal improvement, however, it is directionally helpful. The benefits of freezing CRV on V2 include preventing the protocol from having more exposure to CRV. The amount of CRV concentrated on Aave, relative to the circulating supply of CRV, is already high. Given the limitations of V2 mechanisms, including the possibility of circumventing an LTV of 0, the only way to truly prevent more risk of this position is to prevent borrowing of all assets on V2. Freezing CRV on V2 is just a step towards preventing more exposure to the protocol from CRV.
While freezing the asset may prevent the situation from worsening, it doesn’t alleviate the potential risk to the protocol. We are running simulations and analyses to devise a more structured strategy moving forward, to optimize the following:
Gradually reducing the LT for CRV on V2 to reduce borrowing power and exposure.
Lowering the debt ceiling on V3 - this action will reduce the potential exposure throughout the protocol.
Raising the RF for CRV on V2 - this will have a very minor impact as borrows are currently disabled on V2, and the current total borrows stand at around 36M CRV. However, it will present a more appropriate reward for the protocol, given the risk the asset presents.
Thank you all for your insights and discussions on this matter. I’d like to share a few thoughts from the perspective of the Aave-Chan Initiative (ACI).
Firstly, it’s important to remember the core ethos of DeFi, which is neutrality. The protocol should function effectively regardless of whether a pool has one large position or a thousand smaller ones with similar liquidation ranges. The intention of users or what they do with their funds is not our primary concern. Users should be free to utilize the protocol as they see fit.
Secondly, we need to be cautious about implementing “solutions” that could potentially do more harm than good. The user in question currently has a healthy Health Factor. Artificially increasing the odds of liquidation by increasing the Reserve Factor (RF) or lowering the Liquidation Threshold (LT) may not be the most sensible approach.
Lastly, we should generally encourage the use of Aave V3 over V2. One of the reasons this large position grew in V2 was due to overly conservative CRV parameters in V3, which prevented a seamless migration. V3 has caps that allow for more granular management, which is a significant advantage.
In conclusion, while we should certainly monitor and manage risks, it’s crucial that we do so in a way that respects the principles of DeFi and the functionality of our protocol. In our opinion, there is not a significant need for governance to enforce any kind of action at this point.
We would like to clearly state that we will vote no to any rushed plan that just reacts to social media chatter.
Why is it an issue if a user has no intention of repaying the debt? The protocol is making money on that debt, so its great. We dont make money on repayments.
Seems like the issue here is just V2 being an overly simplimistic and optimistic view of how a shared money market in defi should work, and misses most of the latest additions to lending protocols that would combat things like this. Large positions like this can easily be handled via stuff like:
Liquidation fee auctions
Position Size -based LTV penalties
Even then, Aave can do a bunch of things like:
LTV and util curve management
Pool migration announcements
Why do we let so much liquidity sit on ancient mechanisms?
Aave is making $600 per day over that position. On the other hand if the user is liquidated Aave could lose millions. No one knows for sure how much the loss would be, but $10M+ is totally possible considering the debt is $63M.
So not a great deal…
Altering the protocol to explicitly attack one user who has done nothing to violate any of the rules or functions of the protocols seems questionable to me. The evidence and things that show suspect behavior are even worse.
We’ve crunched some numbers and are going to pull a literal Minority Report. They’re guilty of future crimes and will be punished today.
We found out that they use more protocols than Aave.
The address is rumored to be a Curve founder. PS: Curve recently launched a rival product to Aave.
They may be getting liquidity on tokens by borrowing against them… which of course goes against the principals of a money market.
This whole thing seems crazy to me. I could at least understand it if they were a hacker or something but we’re going after a user who hasn’t violated or exploited or abused anything. They literally just using the platform like normal and now that’s deemed a cardinal sin, with things like “they use protocols other than Aave” or “they engage in borrowing and lending on borrowing and lending platforms” being presented as evidence of their crimes or marks against their character.
If we’re going after a specific user, imo there has to be some real wrongdoing going on. It can’t be some Minority Report prediction backed up with claims that they aren’t a maxi for the protocol and that borrowing against your tokens, the entire point of Aave, is now like unethical or something.
You have no idea whether he intends to repay the loan. He may be ecstatic to lose 45% of his collateral’s value, as you noted. But on the other hand, he may believe CRV is massively undervalued, and hence would rather borrow against his CRV holdings than sell at these ridiculously low prices. We just don’t know. What we do know is his health factor at this point is totally fine.
@pray.eth I agree when you say that we have no idea whether he actually intends to repay his debt or not, but I completely disagree when they say that this is fine because the health factor is high enough.
The health factor doesn’t take in consideration the market liquidity. Actually, the health factor is just a number that is the result of the Value of the Collateral divided by the Value of the Debt and multiplied by the Liquidation Threshold, which is an arbitrary number.
The fact that health factor > 1 doesn’t mean that a position is safe, especially when we’re talking about positions this big. It is very disingenuous for the Aave-Chan Initiative (ACI) to say the opposite without considering that the user is using 50% of the asset circulating supply as collateral. I understand that they’re doing this because they don’t want to create panic in the market, but at the same time this is putting Aave holders and Aave lenders at risk.
And by the way, this is the exact reason why debt ceilings were introduced in V3. To say that this is fine is very disingenuous…
Not sure why you need to call others “braindead” or “simpletons,” this kind of language simply discourages debate. Be better.
Hasn’t this user been adding collateral to his position and/or repaying part of his loan (in USDT)? This doesn’t seem like the actions of someone looking to walk away from his debt obligations. On the contrary, it seems like someone trying to properly manage his loan to avoid liquidation.
One one hand, this could possibly cause the liquidation by not allowing more CRV to top up position
On the other hand, this allows risk to balloon
Perhaps the position will never reach liquidation point. CRV does have great demand drivers as any project or protocol seeking a peg/liquidity strategy should be incorporating a Curve strategy.
I think the comments about this being about ‘competition’ are misguided. There is a legitimate risk of bad debt, should liquidation occur. Hopefully it won’t and there’s probably no direct ill intent here, but wise to consider the possibility of both outcomes as this indirectly becomes a bet on CRV price.
This is a delicate situation as Marc points out that the users are free to do as they wish, this is defi after all, and it would be weird to target an individual. However, this is also using the product in a way that probably wasn’t intended, at least in this specific edge case where a user does have access to a large % of collateral in a not super liquid market.
Probably worth opening a conversation with Michael and seeing what his plans are.
We have started a discussion on other V2 assets on the thread here.
Given that CRV presents a distinct risk profile, it necessitates a unique approach and set of recommendations which we will present shortly.