$1,000 * secure_threshold
to ensure premium redeems and liquidations remain profitableRecently, a user reported a vault that got slashed for 10% of its collateral, which made the vaults collateral ratio fell below the liquidation threshold, and got liquidated. The result was that the vault's collateral ratio dropped below 100%. Hence, the vault’s BTC was no longer secured, and liquidators were not incentivized to redeem iBTC to receive collateral because the collateral value would be less than the redeemed iBTC.
The vault was likely abandoned, which only had ~$130 in collateral and backed 0.003515 iBTC, so the damage is limited to a few USD. However, it highlights that we should change a few parameters to prevent such cases from happening again. Therefore, the proposal suggests to apply the following changes:
$1,000 * secure_threshold
or more specifically:Reducing the slashing amount will have two effects. First, the change will ensure that punishments for vaults will not cause the collateral ratio to decrease below 100%. This is true if a vault is close to the lowest liquidation threshold in the entire system, which is 105%, so the liquidated collateral would still contain a 3% premium for liquidators. Secondly, it drastically reduces the financial risks for vault operators in case their service is unresponsive or temporarily falls below the premium redeem threshold. This should also support the current attempt to onboard new vault operators and increase mint capacity.
Increasing the minimum collateral is required to ensure that premium redeems remain profitable. We have recently seen several days where fees for Bitcoin transactions were ~$20, so premium redeems with a 2% premium on amounts of less than $1,000 would have been unprofitable. The same issue, although to a lesser extent, holds for liquidations if we assume that someone has to (re-)mint iBTC before/after burning his iBTC in the liquidation process. In both cases, a Bitcoin transaction will be triggered, so the proceeds from the premium redeem or liquidation should cover at least the costs of such a transaction. The suggested amount should still be small enough to not jeopardize decentralization but also to ensure profitability for the actions mentioned above.
Numerical example
If the vault is completely offline and does not respond anymore, it will get slashed until the collateral ratio is below the liquidation threshold. Assuming a newly registered vault would have at least $1,300 collateral at at 130% collateral ratio and is unresponsive, then users would be able to drain the vault down to $1,050 or 105% collateral ratio, by getting reimbursed on their failed redeems. Below that level the vault would get liquidated. At this point, a liquidator would get ~$50 for the liquidation, which should cover the transaction fees on BTC (if we assume he has to (re-)mint iBTC) and some profit.
Currently the smallest vaults only provide ~$5 to liquidators so even a small increase in BTC fees make it unprofitable for liquidators to do this.
Given that the liquidated collateral of the vault does not fully cover the iBTC value anymore, it is suggested to use treasury funds to buy 0.003515 iBTC with INTR via HydraDX and burn the iBTC to receive 130.08 USDT, which shall reside in the DAO’s treasury. At a Bitcoin price of $41,418.36 (as of 04.12.2023), this would incur a loss for the treasury of about $15.51.
I do not agree... Even 10% premium is not a lot to cover all the hassle and transactions costs.
If the proposal is accepted, I could easily see many vaults being less careful and endangering the overall stability of iBTC. It would risk a depegging, and that would be terrible.
I do agree with asking for higher collateral -- if you want to play, play on kintsugi. Interlay is for the big boys (cough)
I'm in part agree.
Reducing the punishments will be better for vault and can be attractive for new vault owner. But I think 2% is too low to ensure vault engagement and bridged bitcoin security. Perhaps it could be a good option to set it to 5%.
In this case it's possible to set liquidation threshold to 110%. It ensure 5% min for liquidators, cause I think 3% could be dangerous for security, it's not so attractive for liquidators and if they don't do rapidly liquidation job bitcoin ratio could move to far to be guaranted by liquidation ratio. If liquidation threshold is set to 110%, we have to set Premium Redeem Threshold to 120%, don't care about secure threshold, it's could be change by the vault owner.
So in this case premium redeem and punishment will have the same cost => 5%, I think it's easier for vault owner to project what are the risk for owning vault.
I am completely in favor.
This makes a lot of sense.
Just want to double-check one thing:
We often-times talked about the volume of the premium redeem and how the premium should not be based on the full collateral of the vault but capped at the amount required to get back from "below premium redeem threshold" to "safe collateral threshold".
Is this plan cancelled with the updated numbers in here?
a vault that got liquidated because it did not process a redeem request
Please review wording here as the vault did not get liquidated, only slashed.
As for the meat of the proposal, I agree with the above. EIther the minimum collateral or the slashing % is too low. A 1500 USDT vault will have $1000 of bitcoin. A 2% premium will only net a redeemer $20, which could very well be lower than the bitcoin transaction fee at the time.
Edited
I'm generally in favour (despite operating a small vault that will be affected by this change), but I worry about this clarification by dom on Discord. Copying it in here for clarity:
- For existing vaults, you can continue to operate the vault
- When adding/withdrawing collateral, the chain checks that the vault either withdraws all or deposits such that it is above the minimum collateral
The ability to draw down collatoral in tranches when taking the vault offline would be very important to me (among other things to reduce risk when moving to a new vault). Any chance that this proposal can be amended to support gradual collateral withdrawal when below the threshold?
+1 Reduce the slashing
+1 Increasing the minimum collateral