Context
A lot of context here: https://forum.alchemix.fi/public/d/437-discussion-aleth-rebacking/9
As a follow-up, It appears that all LPers will be able to be made as close to whole as possible through a combination of recovered alETH and ETH from the exploit, as well as CRV tokens from Curve DAO. The CRV tokens would be vested over the course of 1 year. However, Alchemix would have the option to pursue a veCRV whitelist and receive the full amount of AMO-allocated CRV up-front, so long as they were immediately max-locked in veCRV(precedent being set by JPEG'd)
_Note: CRV/ETH price = 0.0003 at time of writing. The calculations in this proposal assume that price, however volatility can affect the outcome. This volatility should be considered when regarding all numbers within this proposal.
Note: Coinbase has still made no indication that they wish to return the MEV profits from the exploit or provide Alchemix with remittance in any form. If these funds are later returned, they can be addressed in a separate proposal._
Note: L2 alETH is still doubly-underbacked due to the multichain exploit. The intent is to reback L2 alETH similar to the plan put forth, voted on, and executed according to AIP95b. That would be a separate proposal.
Accounting
veCRV: Alchemix has voting power equivalent to approximately 7.1m veCRV due to treasury positions in CVX, sdCRV, and SDT. https://debank.com/profile/0x3216d2a52f0094aa860ca090bc5c335de36e6273
Deficit: The mainnet AMO and thus mainnet alETH currently has a deficit of approximately 1100 alETH (Approximately 3.7m CRV at current prices.). This number assumes the AMO receives a pro-rata refund from the funds the exploiter returned, the stuck bridger surplus, and the arbitrage bot refund.
Proposed Solution - Propose veCRV Whitelist, Receive CRV Up Front, Max-Lock, TWAP out of sdCRV
The refund to the AMO is expected to be 3.7m CRV. This is less than the 4.3m sdCRV that Alchemix currently controls. For all affected LPers of all affected pools, Curve DAO is unable to offer the recompensation as liquid CRV and must instead offer it through a 1 year vest.
Instead of the 1 year vest, Alchemix would request/opt to receive the entire amount of CRV recompensation up front, with the stipulation that it is immediately max-locked in a mix of veCRV and sdCRV.
To facilitate locking in veCRV, Alchemix would apply for a veCRV whitelist, which already has precedent from JPEG’d (another victim of the Curve/Vyper exploit) [https://gov.curve.fi/t/vecrv-whitelist-request-for-jpegd-dao/9626]. For those unfamiliar, smart contracts cannot vote with veCRV unless they are whitelisted. If accepted, the whitelist would not be forced to rely on sdCRV, vlCVX, etc and instead would only need to use them where more beneficial to the DAO (such as maximizing the current veSDT position). Thus, the proportion locked in sdCRV would be used to obtain the most efficient use of the veSDT owned by the DAO, and the remainder would be locked in veCRV.
Next, Alchemix would begin to TWAP out of its current sdCRV position. The SDT position is locked and would be used to boost the remaining sdCRV position after alETH is rebacked. Given Alchemix has a significant share of sdCRV, selling the position may require seeking out OTC deals or selling in chunks and waiting for price to recover. However, sdCRV is historically the most tightly pegged CRV derivative and should recover quickly especially with the influx of newly locked CRV. The sdCRV would be sold for ETH, which would be used to reback alETH. Alchemix will seek to sell as much sdCRV as possible so long as the sdCRV/ETH ratio is greater than the CRV/ETH ratio at which Curve DAO allocates the refund. When the price is less than that ratio, Alchemix will TWAP out of the position such that the maximum time to complete the sale is 1 year.
Risks
The risk is that the CRV price could drop relative to ETH, which could still result in a deficit that would need to be paid for by the DAO treasury.
Benefits
So long as the sdCRV position can be sold in less than 1 year, this represents a faster path to rebacking than any other option (potentially as short as a few weeks). While the position is being sold, Alchemix will have up to 50% more veCRV voting power, which can be used to generate AMO profits. Those profits could also be used to reback alETH, if the TWAP takes longer than intended.
Bribing
Depending on the size of the AMO and the bribe multiplier, it may make sense to use the CRV or sdCRV to bribe the alETH liquidity pool. This would only be worthwhile if the expected return to the AMO was expected to be greater than the value of the bribe. This would be a potential discretionary option available to the multisig when conditions allow, but would not make sense as a default option.
alETH/ETH Refund
Note - this portion of the proposal is a pre-proposal. The actual proposal of which addresses receive what tokens in what quantities will need to be its own separate proposal, once Curve DAO has finished their analysis.
The Alchemix multisig currently holds the ETH and alETH returned by the exploiter. It also holds the ETH from the MEV bot that returned funds, as well as the alETH from the bridger clawback proposal. These funds will be returned to LPers proportional to what they lost. This will be done in the form of a claim contract where exploited LPers can claim alETH and ETH.
General Note - alETH Price / Arbitrage Opportunities
There are still many arbitrageurs, LPers, etc holding alETH that may be looking to get out as soon as the price is reasonable. The AMO should deliberately act slowly in the coming months, prioritizing maintaining size and revenue over alETH price, so long as the market is crabbing/going down. This is somewhat standard operation even with a fully backed alETH, where the supply would not be expected to expand (from a lot of new loans) during a bear market, as there are less rewards available to compensate LPers for the risks they bear from purchasing alETH from newly minted loans.
If the alETH price is low, it may also seem enticing for the protocol to reback alETH by using AMO profits to purchase discounted alETH off the market. This is essentially what the AMO will be doing anyway - profits will be used to buy ETH, which will be deposited to the AMO thus turning them into backing. Depositing ETH single-sided to the AMO will receive positive slippage similar to if the AMO were to purchase alETH then LP those funds. The only reason to outright purchase alETH would be if the AMO needs to withdraw liquidity single-sided and wishes to combine actions.
Transmuter
So long as alETH is unbacked and the market prices it based on being unbacked, the transmuter should not allow 1:1 redemptions and therefore should not be enabled. Re-enabling would be possible when alETH is rebacked or if deemed necessary to stabilize the market, by a future governance proposal.
Condensed Proposal
The steps and specifications for the proposal are as follows:
- Alchemix will seek a veCRV whitelist
- Alchemix will opt to receive the CRV refund up front, and will lock in a mixture of sdCRV and veCRV based on maximum veSDT efficiency.
- Alchemix will TWAP out of its sdCRV position for ETH to reback alETH.
a. ETH will be deposited to the AMO
b. Alchemix will sell as much sdCRV as possible when sdCRV/ETH exceeds the CRV/ETH ratio of the refund, otherwise will TWAP out of the position on a maximum timeline of 1 year.
c. At the discretion of the multisig, the sdCRV or CRV may be bribed so long as the expected return from the AMO is greater than $1 per $1 bribed.
d. Alchemix will seek OTC buyers for the sdCRV where possible to avoid market selling.
e. Alchemix will not sell more sdCRV than necessary to reback alETH unless approved in a future proposal.
- [Pre Proposal] Alchemix will use the data provided from Curve to refund LPers alETH and ETH proportional to what they lost and what was returned by the primary exploiter, the MEV bot, and from the stuck bridgers.
a. Official proposal would come when the quantities and eligible addresses are known.
b. The refund plan would need to be executed prior to re-establishing the alETH AMO.
c The AMO would be reestablished in the frxETH/alETH curve pool rather than establishing a new alETH/wETH pool.
Voting
A vote "for" indicates desire to proceed with the proposal outlined above.
A vote "against" indicates desire to resolve the alETH rebacking and LPer refund in a different manner than what is outline in this proposed.
