TL;DR Lack of demand for ZUSD leads to significant redemptions, which leads to ever-higher collateral ratios (CRs). This has made the stability pool (SP) unattractive for potential liquidations and as a way to generate demand for ZUSD. I’m proposing a fix to Zero that would pay some part of origination fees to SP depositors so that holding ZUSD/DLLR is more attractive than holding other stablecoins. This will lead to higher demand and therefore price for DLLR, which will limit redemptions, return CRs to a reasonable range, and create more opportunities to profit from liquidations. This will also encourage more borrowing and free up existing BTC collateral for other uses within the ecosystem – collateral that has no productive value to Sovryn at present. Finally, we’re currently subsidizing the SP with SOV. That could be eliminated, which would take away some dilution of SOV in the market.
There are some knotty issues with this, and it’s likely not a silver bullet. It involves tradeoffs with revenue, so stakers need to weigh in on this. I’m happy to answer questions.
Zero is dead
- As of 2024/08/28, Zero has a total collateral ratio (CR) of 878%. That is incredibly capital inefficient. In addition, it forces users to expose a great deal of collateral to smart-contract risk as well as risk of RBTC relative to BTC.
- Only about 5% of the supply is in the stability pool even though the stability pool is a primary means for absorbing excess supply and is being subsidized by 5% APR liquid SOV.
- Approximately 442K ZUSD have been redeemed in the last month.
- Out of 16M ZUSD borrowed cumulatively, about ⅓ has been paid back and ⅓ has been redeemed.
Zero is experiencing a Catch-22
The stability pool should absorb excess supply through an active liquidation incentive. However, the stability pool is caught in a vicious circle:
- With all collateral ratios above 775%, there is no chance of gains through liquidation.
- With no chance of liquidation gains, there is little incentive to deposit in the stability pool and create demand.
- Since demand is too low, redemptions are common, which incentivizes high collateral ratios to protect against redemptions. And the cycle repeats.
Thus, the stability pool cannot serve as a significant source of demand or a protection against excess redemption. And it cannot resuscitate itself.
Key principles
Generally speaking, Zero borrowers mint ZUSD to sell—either to buy another asset or off-board to fiat. (Some is converted to DLLR for lending or for the RBTC/DLLR AMM pool.)
Since ZUSD/DLLR can always be swapped for USDT or another stablecoin, the gain from holding DLLR needs to be higher than the gain (interest) from holding another stablecoin to incentivize willing DLLR buyers.
The stability pool needs to be more attractive to stoke demand relative to a different stablecoin. The incentive can be a combination of expected return plus decreased risk of a decentralized, bitcoin-backed stablecoin (DBBS).
Assume for the moment that the discounted value of a DBBS is negligible. In that case, the total expected return must exceed the interest rate on USDT. This return has two possible sources—a payment for providing liquidity and liquidation gains. Since liquidation gains are non-existent, the only way to incentivize demand is through a payment to stability pool participants.
The current payment of 5% APR SOV is insufficient. It is not competitive with current stablecoin interest rates on Sovryn (around 11%), and it is inconvenient to receive this payment in SOV rather than the stablecoin.
Adjusting the origination fee is a good strategy to regulate supply, but it doesn’t address the demand side. Zero is designed to create demand through the stability pool.
Solutions
Liquity V2
Liquity is the original Ethereum protocol that was forked to create Zero. Liquity V2 has an ingenious solution to this problem. Borrowers pay an interest rate of their own choosing directly to the stability pool, incentivizing demand. Redemptions are no longer ordered by collateral ratio but by interest rate. Lowest interest rates are redeemed first. This creates a direct incentive to avoid redemptions. If the value of the stablecoin is close to the redemption level, borrowers can adjust their interest rate upward, which has the immediate effect of making the stability pool more attractive to stablecoin holders than the redemption path. This will adjust demand upward, raising the price and moving it away from the redemption level.
There are at least three down sides to this solution:
- This is an entirely different system from the original Liquity. Implementing this on Zero would require a major reworking of the smart-contract system or the launch of a new Zero V2. A V2 launch doesn’t solve the problems with the existing system or help those with existing lines of credit or DLLR/ZUSD in their portfolio.
- One incentive to keep the total collateral ratio high is lost. Borrowers will still want to avoid liquidation (when their collateral ratio falls below 110%). But this eliminates competition to avoid being among the lowest CRs in the system. (This may not be a real problem.)
- It eliminates a key USP because it is no longer a zero-interest loan that can be held without cost forever.
- Liquidity V2 is only available under a Business Source License (BUSL).
A slightly more complex approach is to switch the subsidy to DLLR instead of SOV and raise the level above the interest rate paid to rUSDT lenders. It is likely that much of the same infrastructure can be used to accomplish this but simply change the token. However, this would also require Sovryn treasury to purchase DLLR rather than simply mint preallocated SOV.
Organic rewards
Another option does not require a subsidy but can be implemented as organic rewards generated by the system. Rather than allocating the entire origination fee to SOV stakers, the system would allocate a portion of the fee to stability pool participants. A fee level that generates a return greater than rUSDT interest would incentivize users to hold or purchase DLLR and deposit it in the stability pool. This would make holding DLLR more attractive than holding rUSDT, which would create demand for DLLR being sold to bridge out or to buy other assets.
Due to the randomness of originations and therefore the volatility of payments to the stability pool, a higher average rate may be necessary to incentivize holders in the stability pool.
Automating this system would require smart-contract development, but it would be much simpler than completely reworking the system to something like Liquity V2. As an alternative, a portion of the fees could be paid to the Sovryn treasury to fund a fixed-rate payment using the existing SOV incentive rails but with DLLR payments.
As the collateral ratios are restored to more normal levels, liquidation would become a source of return in the stability pool, which would limit the growth required in the origination fee.
Estimating the required fee
As a starting point, we will use the existing stability pool level and two recent months of origination volume (5/3/2024–7/3/2024).
Let
R = target APR in the stability pool
F = fee rate for stability pool
S = stability pool value = 215,173
B = volume borrowed = 16.641M - 15.991M = 650K
P = period of borrowing (in months) = 2
R = FB(12/P)/S or F = RPS/B/12
Suppose we set the target APR R = 0.20 (20%). Then
F = 0.202215173/650000/12 = 0.011 (1.1%)
For a target 20% APR, we would need to allocate a 1.1% origination fee to stability holders. This would come from the overall origination fee charged to borrowers (currently 5%).
Note that as the stability pool grew, the origination fee would have to grown to keep the APR constant.
Two secondary modifications
Redemption fee addition
If origination fees grow too slowly to properly incentivize the stability pool, redemptions may continue to occur. Another improvement is to add a redemption fee paid to stability pool depositors. This would provide a further reward for holding DLLR in the stability pool and eventually make it more lucrative than redeeming. It would incentivize depositing to the stability pool even in the case where borrowing has dried up due to too much redemptions.
In the period studied above, 537K ZUSD was redeemed. This would have sent 2685 ZUSD to the stability pool for a two-month gain of 1.25% or an APR of 7.5% on top of other fees earned.
If half of this had been deposited to the stability pool instead, redemptions would have been cut in half and the final stability pool size would have more than doubled. This would dilute the earnings in the stability pool by a factor >2.
Wall of liquidity
Another modification is to create a wall of liquidity—a pool of BTC that stands ready to purchase DLLR and deposit it in the stability pool with the additional incentive of receiving the base redemption fee. That is, DLLR purchasers in this pool would get a 0.5% reward for purchasing DLLR (selling BTC) at the nominal market price of BTC in USD. This would defend against redemptions of LoCs.
For the period studied above, the pool would have needed enough BTC to absorb 537K ZUSD. This would have transferred that amount into the stability pool and diluted future earnings by a factor >3. Apart from a locking period, anyone could immediately withdraw their ZUSD and redeem again but would risk missing gains due to being in the stability pool.
Concrete cases
Existing stablecoin holder
Based on the calculation above, depositing DLLR/ZUSD will yield 20% APR over time. (Note that this is paid out in chunks as loans are created.) The stablecoin holder can choose to hold USDT or DLLR. DLLR lend APR is 6.3%. rUSDT lend APR is 9.5%. The user chooses to convert as necessary to obtain DLLR and earn 20% APR by depositing into the stability pool.
As more stability pool deposits come in, the APR will be diluted and the flow into the pool will likely slow to zero. At a then-current stability pool balance of $215K, an inflow of $215K would reduce the 20% APR to 10%, assuming that borrowing continues at the same pace.
Note that with current liquidity a mere $100K in DLLR purchases would raise the price of DLLR by 8%.
Arbitrageur
An arbitrageur will monitor the price of DLLR in BTC terms or the price of DLLR relative to rUSDT/DOC. If existing stablecoin users are moving from rUSDT to DLLR, the price of DLLR will rise and will make redemption of DLLR for BTC with a 0.5% fee unattractive. Redemptions will slow or stop.
An arbitrageur may also be incentivized to purchase DLLR at a discount to rUSDT/DOC and deposit to the stability pool for higher APR, supporting the price of DLLR and making redemption unprofitable.
New borrower
A new borrower will consider the risk of liquidation and redemption in terms of the amount of collateral to deposit and the amount of DLLR to borrow against the available collateral. If the price of DLLR is on par with rUSDT/DOC and redemptions are slow, the borrower will feel safe in setting their CR above the minimum but below average. This will drop the average and incentivize other new borrowers to do the same and continue dropping the average CR.
Existing borrower
Existing borrowers have on average an 800% CR. This exposes their BTC to needless protocol risk if a lower CR is safe from redemptions. As DLLR demand increases to make redemption unprofitable, existing borrowers can either withdraw some BTC collateral and put it to work elsewhere in the protocol or borrow more DLLR and pay more origination fees to the protocol and stability pool depositors.
Origination fee algorithm
The ideal situation would be a dynamic, automatic origination fee for stability pool depositors. However, this could be adjusted manually by SIP or agreed-upon formula periodically. As DLLR flows into the stability pool, the origination fees to depositors would be diluted. The origination fee could be updated based on the formulas in the “Estimating the required fee” section.
Impact
As users grow more confident that they are protected from rampant redemptions, borrowing will increase and so will stability pool deposits. As stability pool deposits increase, the APR will be diluted. At the same time, increases in borrowing will increase the fees collected to compensate for the dilution.
As concerns about redemptions fade, borrowers will be incentivized to drop their CR. As CRs become lower, the potential for liquidation gains in the stability pool will rise so that the overall expected return for the stability pool isn’t so dependent on origination fees.
Given that the stability pool is very small now, it’s reasonable to think that the stability pool size will increase at a greater proportion than borrowing. If that happens, the margin for redemption protection will drop. In that circumstance we might have to increase the portion of the fee allocated to the stability pool.
If the overall origination fee is dropped while holding the stability pool portion fixed, that will incentivize more borrowing and increase the fee paid into the stability pool and boost the APR. That will incentivize an increase in the stability pool in the same proportion as the increase in borrowing, which should keep the APR stable and protect against redemptions.
If the origination fee is dropped and the stability pool portion is also dropped in proportion, the growth in borrowing will work against the drop in the stability pool fee rate. As a result, the stability pool would likely not grow in the same proportion as borrowing, which might eventually lead to redemptions unless the ecosystem grows to demand more DLLR organically before this happens.
If the stability pool grows to the point where the fees are diluted before they can be adjusted, the APR may fall below the APR of rUSDT. If that happens, the incentive to hold DLLR may not be sufficient to maintain the peg and redemptions will occur.
If users do not foresee a prospect of robust future borrowing activity, the expected future earnings in the stability pool will be discounted and the incentive to participate will be diminished. This is a key difference between this approach and Liquity V2. Whereas Liquity V2 requires existing borrowers to incentivize stability pool demand, this approach relies on future borrowers to do so.
Conclusion
Adjusting the origination fee has no impact on resurrecting the demand mechanism in Zero. Stability pool fees can address this. A number of implementation options exist with varying levels of complexity. The most promising from an incentive standpoint is to allocate a portion of origination fees to stability pool participants. This appears to have a moderate level of complexity in terms of implementation.
Just taking a stab, the calculation was that by adding 1.1% to the origination fee (or taking it out of the existing fee) and paying to the SP, we would achieve a 20% APR. But that’s only until the SP size began to increase, in which case we’d have to increase the origination fee allocation to the SP or hope that extra loans would raise the extra we needed as the pool grows.