[PASSED] Adding Interest Rate to Help with VST Peg

General Protocol
1

In an effort to stabilize VST without impacting existing lenders via redemption, we propose phasing out redemption and introducing variable interest rate to the protocol.

Context

VST is currently de-pegged downward due to strong supply growth but lack of demand. While we are working on demand sinks such as the new staking program, it is much more direct to control the peg via controlling supply.

Currently, the redemption mechanism is the main method that the protocol has to throttle supply, bringing the token back to supply and demand equilibrium. However, this is quite possibly one of the worst user experiences across all CDP protocols as it takes borrowers by surprise - it is not up to the borrowers on when they are redeemed against.

Proposal

We propose that Vesta adds an interest rate on debt generated to help with throttling supply when needed. Charging an interest rate that is reactive to VST price would encourage some borrowers to pay back their VST debt, thus decreasing supply of VST.

Interest rate module v1

Vesta will optimize for the security and the peg of the stablecoin, therefore interest rate will be determined by three factors:

  1. Each collateral’s risk - which is determined individually according to each assets’ own risk profile
  2. The collateral’s backing ratio - how much the stablecoin is backed by a particular collateral relative to the total supply
  3. How underpeg the token is - If the token is underpeg, the interest rate across all collateral should increase. This will decrease circulating VST as higher interest rates encourage people to close vaults

Collateral backing ratio

VST is ideally backed by high quality collateral as defined by our risk team. Therefore we will compare the current backing of VST with the ideal collateral backing percentages for each collateral and charge an interest on the ones with higher-than-ideal backing ratio. In Vesta v2 we’ll potentially support negative interest rate - where the interest accrued would be streamed to collateral vaults with lower-than-ideal backing ratio, encouraging VST to be backed by high quality collateral.

The definition of ideal backing ratio will be determined by Risk DAO. It will consider a variety of elements in determining the ratios, including liquidity depth, history of the project, future plan of the project, centralization risk, etc. We are working with Risk DAO in publishing a framework for evaluating the collateral backing ratios soon.

Token peg

To help alleviate the issue of over-supply, the interest rate would be magnified depending on how under-peg the token is.

Where will the interest go?

Similar to all fees accrued to the system so far, the interest will go toward the ecosystem reserve. As mentioned above, v2 will potentially distribute interest rate to vaults with less-than-ideal backing ratio, facilitating a stablecoin that’s backed by healthier collaterals.

Implementation

The interest rate will be calculated manually, and published three days ahead of implementation. Collateral backing ratio part of the interest rate will be calculated and refreshed every week Monday 12 PM ET, and the token peg part of the interest rate will be calculated and refreshed every day 12 PM ET.

Timeline

The execution of disabling redemption and enabling interest rate will be implemented at the same time. As we hope to phase out redemption immediately, the features should be implemented upon passing of this proposal. A roughly timeline of two-week is to be expected.

2 Likes
2
  1. I agree that stopping redemptions is important and urgent. The abundance of redemptions killed a similar protocol in the past.
  2. At the same time improving the peg situation is also important. While it’s good for existing users already leveraged, it’s bad for new users & and thus adoption. It’s also terrible optics for the protocol.

It’s encouraging to see that the proposal is open to the notion of having a different interest rate per collateral. I think that’s key. I wanted to point out the unique nature of the GLP vault - it’s the only vault with a management fee (which is great UX, just my 2c, not complaining). The vault also pockets all esGMX. This week is the first week since the vault’s inception that Arbitrum GLP’s yield went below the minimal threshold so is thus paying esGMX. At this point in time Arbitrum GLP is paying 14.67% in WETH + 4.01% in esGMX for a total of 18.65% APR. Vesta is giving 11.83% APR, so is taking a 36.5% management fee (and yes in fairness the fee will go back to 20% on most weeks). All of that is to say that adding an interest rate on VST on top of that will be hard to stomach for many.

I think that adding a VST interest rate on GLP will be necessary, all I’m saying is that tuning it’s amount should take into account the fact that it’s the only vault with a management fee (that as we can see ranges from 20% all the way up to 36.5% this week). As far as the actual interest rate, one idea would be to set a target fee, between all 3 variables (management fee, esGMX, and VST interest). For example, Vesta gets 20% of all yield in total.

5 Likes
3

I’m not sure the GLP vault needs an interest rate at all given the (somewhat large) management fee. GMX shouldn’t need one, either, since Vesta gets more in admin fees than it would from any interest rate. Vesta should be putting a portion (or all) of those fees into incentives to bring more FRAX/USDC/DAI/USDT into the pools with VST. If we still want an interest rate, the management fee should be lowered accordingly.

The main vault we should be looking at in terms of interest is the GOHM vault. I really hope we’re not thinking about imposing an interest rate on already-opened loans, but only when loans are opened or adjusted. My two cents is that we ought to increase the mint cap on GOHM and simultaneously impose an interest rate on any new or any adjusted vaults. We can do the same for ETH, renBTC, and DPX although they’re not even 1/5th of the size of the GOHM vault. The mint fee should probably also be discarded.

We could also increase the cap on GLP to bring in even more streaming revenue which could then be directed towards incentives or outright market buying for VST.

5 Likes
4

We are looking into interest rate as a whole. Don’t worry, we are analyzing this scenario internally (what is better for the protocol as a whole and the users) . GLP & GMX vaults are unique (Yield-able vaults) which is why it has unique parameters. Once the interest rate is approved, and we come up with the parameters, we will be happy to discuss on this subject. Thank you for sharing your thoughts about it.

If we are implementing the interest rate, it will start at the moment of the implementation and will affect already opened vaults.

I personally want to apply interest rate on new vaults and those that are updated after the implementation. But, it wouldn’t make sense, let’s say, in an environment where a vault is already fulfilled.

The last thing we want to do is increasing VST cap on some vaults so we can have the interest rate applied. The VST cap is a security based on the token’s risk evaluation. (e.g: token failure, liquidity pool not deep enough, etc.) Not only for VST stability.

Overall interesting suggestion.
Thank you!

4 Likes
5

Would suggest the following.

Current vaults size and economics remain unchanged. If vault has capacity it can be capped as x date.

New vaults reflect new interest rate structure.

6

Here I propose a concrete example of how we’ll approach setting up interest rate. Generally, we’ll determine interest rate with three parameters:

  1. Token’s systematic risk
    Risk DAO has given each asset a rating, considering each token’s governance, oracle, redeemability and so on. Here are ratings of the assets currently:
  • ETH: A
  • gOHM: B - governance risk
  • renBTC: B - custodial risk
  • GMX: C - TWAP oracle and governance risk
  • DPX: B - governance risk
  • GLP: C - operational risk; might no be redeemable when in full utilization; underlying asset composition could change over time

  1. Rates across the competitions. It’s important to look across the landscape and observe whether an asset has a place to be utilized outside of Vesta. If so, what are they charge. In order to stay competitive, Vesta should charge a rate lower than competition.

  2. How under-peg VST is. Interest rate calculated from previous two rules should then be magnified by a factor when VST is under-peg to encourage system contraction via paying back their loans.

A quick note on backing ratio

While we have an ideal picture of what VST should be backed by, upon second thought, optimizing for backing ratio is a pursuit of making sure VST is secure. This goal is technically already covered via the first method. Comparing the two methods, it makes more sense to let the systemic risk of an asset determine its cost of capital instead of letting a non-controllable, subjective factor determine its cost of capital.

An example implementation

Determining the systematic risk: Assuming the incremental interest rate increase is 1% as asset security decreases - A: 0%, B: 1%, C: 2%. Assume that the token depeg magnifier is 1x - if VST is at 98¢, which is off-peg by 2%, we could add an interest rate of 2% across all vaults.

A framework for analyzing redeemable assets

GLP is a unique case where one of its most important properties - the utilization ratio - is actually gaugeable. I propose that GLP’s interest rate be calculated based upon the above metric plus the utilization ratio. More specifically, interest rate should increase as amount of VST minted from GLP vault approach amount of idle assets in GLP (amount of assets not being utilized), assuming all idle assets are redeemable.

7

Sers,

I would like to express my opposition to introducing the Interest Rates in Vesta.

I am very happy with the decision to remove the Redemption mechanism, as it negatively impacted user experience.

The introduction of Interest rates will similarly have a negative effect on our current users. It will inevitably reduce the attractiveness of our products and thus diminish our TVL. In my other post, I argued for adopting an expansion-oriented approach. I do believe that with two great products (currently non-existent in the market), GLP and gOHM vaults, we should seek to maximize our fees/profits and improve liquidity - even at the cost of depeg.

Depeg is temporary, as it is always brought back by the “payback discount”. For example - if you looped GLP 3 times in GLP Vault at 0.99 VST and VST drops to 0.97 VST, you will be able to pocket the difference by paying back circa 5% less of the collateral (that is 5% profit on a very stable asset). This is very attractive for current holders.

Instead of the Interest rate as a peg stabilization tool, I propose to raise caps and start bribing via Aura/Convex/Saddle/etc using our collected fees and/or VSTA. Another idea would be to try and acquire some of those Assets via Bonding or treasury swaps.

Additionally, I just want to add that Interest rate parametrization is quite a complex process even for vanilla assets. Not sure if that is the best use of the resources we have in Vesta team.

1 Like