[RFC] Peg keeping mkUSD

[Sidn3yGottlieb]

Redemptions

There has been a growing number of redemptions over the past 48 hours as a result of mkUSD trending towards $0.995. While a 0.5% depeg isn’t alarming, it is something we are conscious needs to be taken care of so the protocol can keep growing.

Redemptions are a very important mechanism of Prisma as they essentially prevent mkUSD from going too far below $1. They allow redeemers to sell mkUSD for the collateral of a user with the lowest CR and pay a fee to the DAO. The redemption fee could be raised to 1% (or even more) which would prevent redemption from happening until mkUSD might reach $0.99 but there is little doubt not having peg keeping levers in the current RWA environment would lead to much worse depeg.

Peg Keeping Mechanisms

• The most obvious one suggested by Michael from Curve is another borrow rate raise as Prisma borrowing is below market (MakerDAO around 5%, AAVE are currently sitting in the 4-9% range, crvUSD dynamic rates are above 8%). Prisma’s interest rates cannot go above a hardcoded ceiling of 4%. This is something the DAO can consider though there is a chance it would likely drive away TVL from the protoocol

• More pools, currently the vast majority of mkUSD sits in a pool with crvUSD which has proven to be very good at keeping its peg but also currently slightly off peg. On Curve, borrow rates keep increasing until users close their positions restoring peg to crvUSD, a mechanism similar to redemption though less punitive. We’d like to introduce more pools paired with mkUSD like USDC in coming days so that mkUSD is less dependant on any given stable coin.

• smkUSD is a stability pool module we are working on that would tokenize stability pool deposits and their yields allowing for greater flexibility of stability pool deposit.

• Implementing Peg Stabilization Bonds: These are specialized time-locked liquidity provider tokens paid in StablecoinXXX, pairing mkUSD with StablecoinXXX. Offered in response to imbalances in mkUSD pools, these bonds not only help maintain the peg but also distribute PRISMA emissions and mkUSD interest. Their issuance, contingent on peg deviation, serves as an effective peg stabilization tool, especially when the lock duration and yields are appriately tuned to market conditions.

• Interest rebate
  In analyzing user behavior within our protocol, we’ve identified two main user categories: yield farmers and leverage seekers. Yield farmers are highly responsive to the interplay between interest rates and PRISMA prices, while leverage seekers are primarily concerned with securing the lowest possible borrowing costs. This differentiation in user priorities necessitates a nuanced approach to interest rate management. To address these varied needs, we propose the implementation of an mkUSD native yield mechanism for smkUSD token holders sourcing mkUSD from the fee receiver. This system would function as a dynamic incentive, adjusting the net cost of borrowing according to user type. For yield farmers, the system would offer a yield in mkUSD, effectively providing a rebate on any interest rate increases. For leverage seekers, it would ensure that the absolute cost of borrowing remains competitive, slightly below prevailing market rates. This approach ensures a balanced incentive structure, catering to the distinct requirements of both groups within our ecosystem.

In the meantime, be sure that your vault isn’t among the first on this list and consider moving to the stability pool if you are in a heavily imbalanced pool like the mkUSD/FRAXBp pool.

[JonisJon]

Thank you @Sidn3yGottlieb for posting this, its super important the protocol finds ways to help mkUSD get back to peg and slow redemptions (preferably make them a rare event).

I think while raising interest rates is the most basic tool prisma has, its not really getting to the root of the problem and going towards the max of 4% is likely to chase away significant TVL and start to make prisma uncompetitive in the market (e.g. higher rates + redemption risk starts to not make sense for most borrowers vs paying high rates somewhere else with redemption risk).

I think some other ideas here, like more pools, raising the redemption fee, and the interest rebate are worth trying and may help.

Some other ideas worth considering that came up in a lively discussion with the community in discord today:

(1). A dynamic mint cap, basically just don’t allow minting in the system whenever peg is below .995. @Sidn3yGottlieb mentioned this may not be feasible engineering wise and if that is the case I think some other sort of clamp down on the minting has the likelihood of being highly effective. E.g. lets at least lower all mint caps to their current levels so that redemptions are not always followed by minting pressure which pushes them back down. If we can temporarily pause or limit minting, combined with redemptions already happening, that should push mkusd back to peg. We can also then slowly raise the mint caps again once the peg is more stable (and raise them slowly enough to not cause the kind of sell pressure we have been seeing).

(2). Incentivizing lower CR’s within the system, e.g. if we have a subsection of “active debt” rewards that just pays any vault with a CR below 150% a higher rate of rewards it would incentivize people to take on more redemption risk and lower CR’s for everyone across the system. Redemptions are less of a problem when they are happening to CR’s close to 150% or lower, they are much more of an issue when they are happening to vaults closer to 200%.

(3). Another idea would be to implement a system wide LTV paramater over which vaults are safe from redemption. This wouldn’t stop redemptions, but it would make the UX better and also encourage redeemers to redeem the lowest CR regardless of collateral rather than just targetting wstETH even when its CRs are the highest (and this parameter could be adjusted by governance up or down depending on peg health as needed). For example if its set at 180% right now, it would help stem the flow of redemptions on wstETH but would still allow for the mechanism to work on other collaterals. This would make redemptions act more cross-collateral and create a better UX for redemptions in the system.

I think if we implemented (1) or (2) (but preferably both) that would go a long way towards helping to solve the redemption issue and get mkusd back to peg. The biggest thing IMO is we need to slow down mint pressure on the back of redemptions, a dynamic mint cap would be the best for that but barring that at least clamping down on mint caps across the board temporarily would achieve a similar result.

I think (3) would be great too, but more of an unknown exactly what the effect would be.

If we try things like that and its not working then increase the interest rate as well, but I think just clamping down on minting alone would be enough to get us back to a good place without raising the interest rate and chasing away TVL.

[narok]

All Liquity forks suffer from same problems. The balancing act between peg, interest rates and TVL is extraordinary complex. Furthermore, Redemption mechanism is a silent killer which in the name of peg destroys user experience and TVL by extension.

Prisma in its inception has been very attractive for users as it offered low cost borrowing with some of most sought-after collateral options. Alas as protocol matures, we will find ourselves having to choose between two undesirable choices:

• either we increase interest rates
• or suffer redemptions

Both ways lead toward decrease in TVL, which in turn will decrease token values, thus starting vicious cycles.

Given that Collateral-Backed stablecoins are a function of both supply (attractiveness of Vaults) and demand (Liquidity rewards), I want to argue for leaving the stabilization mechanism into the market’s hands.

Proposing a Paradigm Shift for Prisma

I propose we explore the option of redefining the Prisma protocol to be not a stablecoin issuer, but a liquidity unlocker for staked ETH.

This would be done by executing the following changes:

1. Remove Redemption mechanism
2. Leave mkUSD peg to free-float

By doing this we could explore a true Defi concept, to have a self-balancing, governance minimised protocol.

I believe this will make the protocol much more attractive to long-term holders, as interest costs would be lower from other competitors while they would not have to worry about potential redemptions.

On other hand it will leave the search for equilibrium to the natural crowd decision of the market, thus avoiding the constant oversight over peg.

Market Stabilizing Forces

What happens when peg falls (risk of over-peg is actually demand-driving by itself):

Vault Depositor:

• High incentive to repay:
  Lower cost of debt due to mkUSD < 1.00. This will prove very attractive to early users, offering them to make even potential capital gains on their loans.
• Low incentive to borrow:
  If mkUSD <$1 when they open the position, it effectively increases their MCR. Say that original MCR is 110%, Alice opens a position with $110 and takes out 100 mkUSD when mkUSD is at 98¢, Alice’s debt value is effectively $98. So effectively, Alice’s MCR is 110/98 = 112.2%.

Arbitrage actors:

• High incentive to buy discounted mkUSD on the market and liquidity mine till mkUSD approaches 1.00. Additionally it becomes also an attractive asset for speculation on the protocol itself. This could guide important volumes through our partner protocols.

For both sides of the play, mkUSD becomes more attractive as the peg slips further. From economics, there must be a point where these two forces meet and find a natural equilibrium. Obviously, the mkUSD can’t be valued at 1.00, but any lower it goes it takes the premium on Supply (Higher incentive to repay and higher cost to borrow) and offers extra returns for Liquidity miners. This would also allow certain cyclical rotation inside of vaults when peg falls, thus creating more fee revenue.

In conclusion:

Given these self-adjusting forces, I believe the notion of the peg as fixed 1.00 should be relaxed and that we shouldn’t worry if the peg starts to free float. Educating the community that it is a normal part of protocols ebbs and flows is crucial in order to drown FUD initially. Over time this shall be accepted as normal.

It is rather obvious that stablecoins are very carefully monitored and very susceptible to crowd-think. Providing a mechanism that is human-free might produce unease, but it is the way of Defi. We want applications and structures that self regulate. This would be a step in that direction.

[JonisJon]

Personally I don’t think moving away from the liquity model is the right way to go, then prisma loses all its differentiation. If its just another protocol with no redemptions and high interest rates its competing with things like crvUSD and aave etc which are already better at that type of model with a big headstart. I think its better to try some creative ideas and see if we can get the liquity model to stick, its a big differentiator compared to other CDP and lending models IMO.

[Sidn3yGottlieb]

Removing the redemption mechanism which is our most useful tool against depeg would certainly be a " I’m in the arena trying stuff. Some will work, some won’t. But always learning.".

@JonisJon your idea of incentivizing those willing to get redeemed is fun and potentially feasible but it wouldn’t help the peg, just let some people get rewarded for potentially taking a loss for others.

[JonisJon]

Agreed @Sidn3yGottlieb it doesn’t help the peg, but it does significantly help the UX of redemptions. I think if we did a combination of limiting the minting via mint caps and incentivizing the lower CR’s it would both help the peg and help most vaults able to lower their CR’s.