This discussion thread intends to drive consensus on how we are going to reduce the Zero Protocol origination fee in the future.
Current Situation
The total supply of ZUSD has been hovering at 4~5M for over 7 months.
The number of credit lines and collateral has been slowly decreasing.
The ZUSD net insurance was about 202K for the past 4 weeks. It stands at about 4.5% growth in 4 weeks.
The redemption volume was around 128K for the past 4 weeks; It was around ~2.8% of the total supply.
Babelfish has around 1.67M stablecoins; only ~57% of TVL is DLLR.
My thoughts
We should focus on growing ZUSD supply.
The acceptable level of redemption should be under 5% of the total supply.
Babelfish is working fine on balancing the stablecoins.
it is acceptable that the price of DLLR is not above $1 all the time. As long as it is not seriously depeg for a long period of time, then we should let the market work its magic.
The marketing of Zero Protocol should emphasize the redemption risk.
Next move
We should reduce the origination fee as long as the monthly ZUSD net issuance is under 10% growth and monthly redemption is under 5% of the total supply.
I will suggest to reduce the origination fee floor from 8% to 5%.
others options
The fee schedule actually has throttle to deal with larger redemptions with higher fees, and to throttle borrowing directly after large redemption volumes. However, the decay parameter is tuned such that the fee changes by a factor of 0.99 per hour. if the decay parameter is configurable, we can increase the throttle behavior. so that to prevent over-borrowing or over-redemption. “this requires code change!!!”
Thank you for taking this initiative! I fully support it.
I have a technical question, how much developer time would be required to set up a variable origination fee dependant on current ZUSD issuance and redemption rates? It is not algorithmic by any means but could serve as a swift and automatic safeguard to let out redemption pressure without waiting months between SIP’s to change it.
IF ZUSD net issuance < 10% growth AND monthly redemption <5% of TS THAN OF = X%, OTHERWISE Y%
This would require a change in the smart contracts that run Zero. Any change in the smart contracts introduces new risks and resets the Lindy effect clock. The Lindy effect says that the longer we’ve gone without an exploit, the lower the probability of an exploit in the near future.
So bottom line is that we need an overwhelmingly compelling reason if we’re going to change the smart contracts, and I’m not sure this is that reason even if we’d prefer this behavior in the abstract. In the meantime, we can always change this based on a SIP. We could also approve a proclamation SIP that instructs developers to change the origination fee monthly or quarterly based on an algorithm we approve.
I cannot give you any development estimate simply because I am not the maintainer of the Zero Protocol codebase.
In the TroveManager contract, it manages the fee schedule for the redemption fee and borrowing fee (origination). However, the TroveManger does not track any monthly net issuance and monthly redemption.
However, there is another way to achieve what we want.
ps. I added another options section to the post above.
the contract has throttle on the fee schedule to deal with larger borrowing or larger redemption, but it is not so effective. if we can make the decay parameter configurable, then we will be able to adjust throttle behavior.
I appreciate your contribution to this. I do have some questions and concerns though.
I’m not convinced that we should focus on ZUSD net issuance. By itself it doesn’t really tell us much of anything about the potential for redemption. If demand for DLLR grows significantly, there’s no reason that ZUSD net issuance should be throttled above 10%.
Also, what is your thinking about the specific threshold of monthly redemptions under 5%? 5% of total supply strikes me as a pretty large redemption. I’m open to just letting the market sort it out, but in that case why should we care about any percentage in particular?
Maybe what we should be focusing on is maximizing total protocol fees. If a 37.5% drop in origination fee stimulates a doubling of originations, we come out ahead. Trying to find the sweet spot would be a purely market-driven approach.
Eventually, I would like the fee floor for origination and redemption reduced to 0.5%. And let the market run its course.
I agree that growing ZUSD net issuance may not be the best strategy right now. The reasons behind what I suggested are that it is simple, does not require a lot of work to implement and we can react to the market change.
5% is my acceptable level of redemption. I don’t have any specific reason behind it. We should care about the level of redemption because the system is still in the bootstrap period in my opinion.
Maximizing total protocol fees can be a good strategy. I tried to work on a machine learning model to predicate the Zero protocol but the performance was not good. If we want a purely market driven approach, maybe we should tune the fee decay parameter so that throttle larger redemptions automatically. With a slower decay parameter, the fee will stay higher for longer after large origination or redemption.
I would like to see the Zero Protocol market run smoothly without our intervention.
I’m totally on board with your goals, and it makes sense to think of Zero as still in a bootstrap period.
You might be interested in this algorithm I derived for updating the origination fee to maximize profits. It probably has some similarities to your approach.
Thank you for the initiative. Zero is Sovryn’s most important product and it’s a shame that it’s not performing well.
The main problem is the high collateral rate required, as you probably know. If someone deposits their btc with 600% collateral, pays 8% origination fees and is then redeemed within a week, that is only one thing: non-functional.
From my point of view, the conclusion is the same with a 5% origination fee: if you are redeemed within a week with 600% collateral and pay a 5% fee for this service, the product is not functional.
Lower fees are nevertheless an important key to making redemptions less painful for users. But I think they should be a bit lower so that paying this fee several times doesn’t hurt so much. 2-3% comes to mind.
Nevertheless, the main focus must be on generating as few redemptions as possible (I would be more in favor of 1% of supply redeemed per month). How? At the moment, I can only see one way to achieve this: increasing the redemption fee to 2% and accepting a relaxed peg.
The protocol revenue would benefit from this.
The damage caused by the looser peg is unlikely to be greater than the damage caused to users (and especially new users) by the current redemptions.
As a convinced DLLR hodler, I don’t really care whether my dllr is currently trading at 0.98 or 0.99. As long as I can use it to generate yield in the pools and am able to exchange it for my btc in the zero protocol.
Please note: The zero users themselves do not pay a redemption fee for their loan, only arbitrage traders who redeem DLLR without their own zero loan pay this fee. If the arbitrage route via zero redemptions is more expensive than other routes (babelfish), that may reduce redemptions significantly. I think it’s worth a try.
Unfortunately, I don’t think a higher redemption fee is effective for this purpose. Here’s why. Let’s say we have a 2% redemption fee, which implies a peg floor of 98%. Excess ZUSD/DLLR will initially be dumped on the AMM until the AMM price of DLLR is at $0.98. The amount of DLLR required to move the peg from 100% to 98% is a fixed quantity that depends on AMM liquidity. The AMM will absorb that much DLLR with no redemptions. As soon as the peg drops below 98%, however, redemptions will resume at the usual rate to absorb the excess supply of DLLR from that point forward.
So the best that a high redemption fee can do is cause a temporary pause in redemptions until the new peg floor is hit. After that it does nothing.
Of course, in reality, users will avoid redemptions and will slow down borrowing in this situation. But they will do that earlier if the redemption fee is 0.5%.
The alternative way would be: If the peg is at 0.98, you can bring rbtc to Sovryn and sell them for a ~2% premium. 10,000$ worth of rbtc becomes ~10,200 DLLR. You could then exchange the stablecoins 1:1 for other stables via babelfish and take them out of the system. Then you could exchange the 10,200 USD stables for 10,200$ worth of btc on another platform. Of course, there are some costs here because of the bridges and the transactions. At 1%, this is certainly not profitable because $100 is easily spent on the fees. At 2% redemption fee, however, it could be more interesting and possibly cheaper to choose such a route instead of the zero route.
At least that would be my hope…Or is it a pipedream?
Any way, with regards to protocol revenue, a higher redemption fee would somewhat boost it.
Again, that would work short term. But it creates a supply-demand mismatch in Babelfish, and soon all the external stablecoins would be gone due to the outflow and you’d be back to looking for some other way to unload DLLR.
Balancing curves would slow that down some. But once the penalty reached more than 2%, the incentive would swing toward redemption rather than bridging out.
Having something that would work short term (we don’t know how long that is, really) is better than having nothing and just reduce fees. If we monitor the situation, bitocracy can step in before it gets out of hand. We have not battle testetd babelfish curves yet. We have not tested how a 2% redemption fee would play out.
You make good points but i would still be in favour of giving this a try. The last Zero SIP that put the origination fee to 8% was very succesful in terms of protocol revenue. Maybe this one turns out as good.
As a safety buffer, we could offer babelfish some adoption fund SOV for incentivizing bridging in of stablecoins, like we incentivize our AMM’s.
This would basically use the adoption fund for better Zero performance and should be absolutely worth it in the current market situation and a lot of eyes on Bitcoin Layer 2’s.
I bet that would be more productive than farming Dune Memes (granted, it’s one of the best films ever).
have you tried to implement the MVP yet?
Based on my understanding, we need to small change in the origination fee floor before we calculate the optimized fee. right?
Haven’t tried incremental changes like that, unless we want to look at the previous drops in the origination fee as incremental. We went from 13% to 8%, right? Might be worth looking at that.
I think spreading redemptions across all troves (lines of credits) is impractical. Because it will require a lot of state update. It wouldn’t not scale.