[PASSED] Removing or Adjusting the Redeem Function

General Protocol
1

Situation:
Vesta does not have much market share in the largest collateral asset on Arbitrum, ETH. Abracadabra has $72m of ETH collateral locked, with $7.11m worth of MIM borrowed against it. In contrast, Vesta only has $2m worth of ETH collateral locked, with $0.7m worth of VST borrowed against it. It looks like ETH holders do not want to borrow VST. Tellingly, ETH is not the largest collateral asset on Vesta. That honor goes to GOHM, by a large margin. If Vesta can capture more ETH market share from Abracadabra, both the TVL and fee revenue of Vesta will grow exponentially.

Problem:
The reason why Abracadabra dominates in this space, is likely because they give Borrowers a much better deal. Abracadabra’s liquidation penalty is much lower at just 5%, and in the event of a MIM depeg, Borrowers can help to repeg MIM by buying cheap MIM and repaying their debt, profiting off the depeg in the process. In contrast, Vesta charges a much heftier liquidation penalty on ETH at 10%, while penalising Borrowers regardless of which direction VST depegs. If VST depegs upwards, the value of the Borrower’s loan has gone up, and it costs more to repay. If VST depegs downwards, Borrowers risk being forcefully liquidated at a 0.53% penalty through the redeem function. Borrowers lose no matter what. In view of this, there is no incentive for ETH collateral holders to borrow VST, as a MIM loan is more advantageous in every scenario.

Solution:
To make VST loans more attractive to ETH collateral holders, I propose to remove the redeem function altogether, as it is a constant threat hanging over the heads of borrowers. A borrower could be prudent and responsible, but still be liquidated at a 0.53% penalty in the event of a significant depeg, through no fault of his own. To avoid this, Borrowers have to keep a much higher collateral ratio then they would have liked, reducing the capital efficiency of their collateral.

It’s been proven that overcollaterallised stablecoins are good at keeping their peg, as borrowers are incentivised to buy the stablecoin and repay. A bad debt situation is not necessarily fatal; Abracadabra has 3.62% (https://bad-debt.riskdao.org/) of it’s portfolio in bad debts, but holds strong anyway. In Vesta’s current state less the redeem function, a significant depeg should not stay for long, except in a black swan event where a contract is exploited, or OHM price wicks down faster then it can be profitably liquidated. The latter case would not be fatal to Vesta if a more stable asset like ETH or BTC was the major source of VST loan, instead of GOHM which currently makes up almost 70% of all VST loan origination.

Alternatively, if redemption must be kept to protect Vesta from GOHM, allow the redeemed trove to keep the redemption fee, instead of distributing it as protocol fees. This allows the borrowers to at least participate in the gains of a depeg, similar to Vesta’s competitors, instead of punishing them for every depeg.

5 Likes
2

Personally, I don’t see the redemption mechanism as a protection from bad collateral/failed liquidations.
In fact, if VST is traded below $1 it is actually easier (more profitable) to liquidate, so allegedly the system is even more secure.

From business/community point of view, the down side is that if vst < $1 (for some time), then new borrowers are paying some extra penalty for the borrow (if they plan to sell it to frax or usdc).
Also while stable coin community/holders (e.g., LUSD) don’t complain much when LUSD > $1.03, they might complain/panic if it is sub 0.97.
A panic/run to the bank will eventually be mitigated by lenders who get to repay their debt at lower price. But could be bad PR.

So eventually it is mostly a trade-off between the UX of borrowers vs the UX of VST holders, rather than a risk trade-off.

1 Like
3

Good points. Also there’s not the same incentive for borrowers to repay their debt since their position doesn’t accrue interest like other CDP protocols.

Edit: With Mikey disclosing launching an AMO to help stabilize the peg, it could be possible to not need the redemption feature or it could at the very least not make redemptions profitable any longer. However let’s see it in practice first.

1 Like
4

Thank you for your thoughtful post @Arcrenciel. Although I would argue that gOHM isn’t likely to wick down to zero given its current redemption mechanism of inverse bonds, it is a fair concern. I’m fully in line with your reasoning to increase VST’s ETH backing.

I propose three changes to the protocol to increase VST’s ETH backing.

On the other topic of disabling redemption, I fully agree that it’s very much a non-optimal user experience. However, redemption does help to save the token from depegging downward. With redemption gone, a new mechanism should replace it to ensure that the token stays on peg.
I propose that we resolve this issue through:

  1. Establish a Peg Defending Fund - direct part of the revenue toward it and use it to buy VST when VST is off-peg.

My concern with point 1 is that Vesta unfortunately has not generated lots of revenue as the protocol only charges an origination fee. This leads to my point 2 recommendation:

  1. Be in line with the rest of the protocols and charges competitive interest fees.

Interest fee is effectively the biggest source of revenue for any lending protocol. Without charging interests, Vesta’s origination fee will unfortunately not sustain Vesta’s operations, let alone a Peg Defending Fund. In light of this, I propose that Vesta start charging a competitive interest rate on ETH at 0.4%, below Abracadabra’s 0.5%.

However, this only increases the cost of borrowing for users, and is defeating the original purpose of this discussion. To alleviate this and decrease costs on borrowers, I propose point 3:

  1. Enable the feature of depositing borrower’s ETH into Aave to generate yield.

The mechanism of depositing users’ deposits into other protocols to generate yield is something we have been working on internally. We call this project Mobilizing Collateral and it was previously mentioned in [RFC] GMX Vault Strategy. For something like GMX, where to deposit the asset is an easy task as there’s only one place to stake GMX, and that is GMX’s native protocol. For ETH, we have been exploring a stETH strategy, which has been put aside due to recent market developments.

Aave, being one of the blue-chip protocols, is arguably the one of the safest places to deposit Ether. While we are still working on a yield framework, Aave’s strong track record, sustainable yield, and instant stake/unstake all make it a perfect protocol to deposit Vesta’s ETH into.

In summary, I believe Vesta should disable redemption, charge an interest rate and forward the revenue generated to a fund that buys VST when it’s underpeg, and mobilize all ETH collateral to into Aave.

1 Like
6

Point 1: As you noted, the protocol doesn’t make much money, so it won’t have funds to meaningfully defend the peg. It’s still likely to be a decent source of profit for the protocol though.

Point 2: Most protocols that charge an upfront borrow/repay fee, do not actually charge an annual interest fee. Abracadara on Arbitrum is a notable exception; they still continue to charge no interest on ETH mainnet. However, i agree that as Abracadabra is our only significant competitor in this chain, we should only consider their practises. Thankfully, it doesn’t look like QiDAO is investing much effort and capital to compete on Arbitrum.

Anyway, in the grand scheme of things, a 0.4% interest rate is rather insignificant. I say go for it. For any borrower, i believe the main objective is to not have their collateral seized and sold, and a 0.4% interest rate of interest on top of the existing loan origination fee is a good trade for peace of mind, and the ability to borrow more.

Point 3:
The problem with ETH yield on AAVE, is that it’s close to non-existent. For most small users, the 0.15% APY they earn from AAVE would not be worth their gas fees to deposit and withdraw from AAVE, even with the low gas costs on Arbitrum. Even for a whale with a stack of 100 ETH, the interest they earn after a whole year, is a measly 0.15ETH. However, i do agree in principle that it doesn’t make sense for capital to be idle. This idea may need to be shelved until stETH unstaking becomes available, which should force the peg back into place.

1 Like
7

Except it is. Imagine if GOHM fails at scale right now. What would happen, is massive, permanent depeg of VST, as the GOHM debts turn bad. Who then suffers the fallout? Other vault depositors do. A depeg will enable large scale redemptions of ETH, BTC, GMX and DPX. Vesta is not an isolated system because of the redeem function. Depositors are putting themselves at risk of liquidation, if any single one of the collateral assets fail.

This is pretty much unique to Vesta. In other protocols like QiDao and Abracadabra, a failure of one of the other collateral assets is actually cause for celebration for borrowers in other vaults, because the resulting depeg means you get a discount on repaying your loan. In Vesta, you just get liquidated against your will with everyone else.

2 Likes
8

I agree the huge dependency in gOHM is a systemic risk, but this is why borrowing vs other collaterals should be encouraged.

But my view about the technical depeg outcome is different than what you describe.

  1. A depeg of vst will not trigger liquidations.
  2. in case one of the vaults get rekt, a redeem mechanism might amplifies the run to the bank, as users would want to be the first to redeem. And the last users, who will try to redeem against rekt gOHM (in your example) will lose everything.
1 Like
9

Redemptions won’t be against gOHM though. It would be against ETH, BTC, GMX and DPX. Given that redemmers will likely market sell to realise an instant stablecoin profit, i forsee all the ETH, BTC, GMX and DPX vaults being wiped out completely. The gOHM vaults would be left with the ones holding bad debt, that can’t be redeeemed against.

The effect for ETH, BTC and GMX collateral owners probably will be a 50:50 coin flip. They would suffer slippage losses when they rebuy their assets. Whether they rebuy higher or lower, is really up to luck. Those with no liquidity on hand to rebuy are screwed.

For DPX though, they probably will be rebuying higher. This is because LP pools for these assets cannot absorb such a large sell without a significant dip. Meaning, their collateral is going to be seized and market sold at a fire sale discount…

10

We’ll encourage higher amount of ETH deposits through lowering ETH mint fee to 0.3% and deposit ETH into Aave to allow people to earn Aave-generated interest fee (which Vesta would take a 20% performance fee from). Going further on redemption, we can either disable redemption for all, or find a middle ground - only keep redemption for assets that which are not grade A, which are renBTC, gOHM, GMX and DPX.

Let me know what you think of this!

1 Like
11

Thank you for pressing on the redemption mechanism. However just a small correction on how the vaults will behave during a mass redemption event due gOHM collapse.

  1. gOHM vaults will get liquidated by its stability pool (SP). Collateral risk transfers over to SP stakers. After the SP is depleted, the remaining gOHM vaults will inherit liquidated vault positions. The constant liquidations will bring huge sell pressure to VST and cause it to depeg downwards.
  2. Redemptions will become profitable during the mass gOHM liquidations. (Here’s my correction) As users redeem other collateral, they are repaying the VST debt of the active vaults at the oracle price of the collateral. The vaults won’t get wiped out. Once the debt is repaid by redemptions, the remaining collateral in the vaults is safe from redemption.
1 Like
12

I’d like to not stray too far away from Liquity, but I can compromise. To encourage more ETH collateral, I’m okay with disabling redemptions against it.

1 Like
13

Sorry for the lack of reply here. Before we move things into the proposal stage, I think there’s still a couple options.

Option 1: disable redemption completely for all assets, and then put forth part of the treasury toward a peg recovery fund. We could start with a small amount such as 1m.

Option 2: disable redemption for ETH only, and put forth redemption pressure onto other assets (gOHM mostly)

@abmis @Arcrenciel @yaron would love to hear your thoughts on either of these options. I personally would opt for option 2 since option 1 would eat into the treasury.

1 Like
14

If we have $1m idling in treasury anyway, then Option 1 sounds better. Repegging is usually a profitable endeavor, unless the stablecoin has catastrophically failed, in which case the treasury funds would be needed to make good anyway.

If not, Option 2 is a great compromise.

1 Like
15

Until there is a variable interest rate for loans in vesta (that can be increased when vst < $1), and given that gOHM backed vst is already taking big part in the entire vst supply, I would personally feel more comfortable to start with disabling only ETH redemption, and see if it has a positive affect on the borrow demand.

2 Likes
16

While I agree with @Arcrenciel 's point I do believe Yaron’s suggestion makes more sense since it’s more iterative and also considers an interest rate model that lessens Vesta’s reliance on a arbitrage operation and more on an automated interest fee model.

On this note we will move forward with canceling redemption for grade A assets.

3 Likes
17

Sounds good to me.

For the variable interest/mint rate, may I suggest implementing a Proportional (P in PID) controller based on the oracle (Curve for now, hopefully chainlink at some point) price of VST.

2 Likes