EDIT: updating this with more options from discussions in #governance
There is quite a lot to discuss overall, especially higher level things such as integrations and PCV diversification. We’ll be able to leverage the position of the protocol into some successful integrations and stabilizing the peg over the coming weeks.
For this discussion I’d like to focus on the short term situation regarding the launch and the peg:
The PCV and circulating FEI are massive
The only current use cases for FEI involve supplying it as liquidity on AMMs or trading it for ETH/TRIBE
There is a lot of sell pressure - secondary spot price of FEI is ~.90
Oracle price movements also affect the incentives and time to reweight
Direct incentives are working as expected. The buy and sell pressure are converging on the secondary market price of FEI between .88-.90c. This is keeping the reweights from happening because buying pushes the reweight into the future.
Overall I think it would be healthy to allow a lot of the built up sell pressure out so the reserves and price can stabilize. Here are some options arranged from shortest term to longest term:
Force reweights using the Guardian to allow people to sell. Will be a bit clunky because it would slingshot up and down, but this can be done immediately and communicated via twitter etc.
Increase the growth rate of incentives. This could end up in a situation where there is a lot of buy support and inflation but keeping the price low around .90. Can be taken straight to a DAO vote.
Setting a pre-defined cadence for reweights such as every few hours, or every time the price goes below some %. Requires a minimal contract change and DAO vote.
Using the reserves to stabilize the price more directly by selling ETH for FEI with some spread below the peg price. Requires more substantial contract changes but can likely be written, reviewed and voted on within 1-2 weeks. This creates a price floor on FEI and could be low like .95.
Capping the burn % to something like 2-5%. This leads to faster reweights and makes integrations easier.
I think it will be good to gather some more data on the mechanism especially after the staking rewards kick in tomorrow.
Lets open up the discussion and see how people feel about the above options.
I like either 2 or 3. Clearly waiting around for FEI to restore the peg for over 24 hours is not ideal, and it would be nicer to have the peg restored faster.
In terms of 2, I was suggesting in discord that to set the value r (rate of reward incentive change) to be tied to % off the peg, so the rate of reward change is faster if FEI is more off the peg. For example, r can be set to like 5x the base speed if FEI is 5% off the peg, and 20x the base speed if FEI is 20% off the peg. etc. instead of at a somewhat constant speed of 1. This will likely move the reward incentive faster than how it is now, and then the parity between burn rate and reward incentive would be reached faster, and allowing for the reweigh condition to be met faster.
in terms of 3, It’a actually an easy way to have a pre-defined cadence and set an expectation on how often a reweigh happens, maybe once or twice a day at the same hour. That makes the user experience less frustrating because there won’'t be as much of a guessing and uncertainty. The arbitrageurs can use this and probably move the price up between reweighs too. The downside is that this increases the complexity, and gives one more parameter to set manually, and there might be unnecessarily reweighs done daily if and when the protocol becomes more stable.
Maybe a combination of 2 or 3, so that r is set faster in some way, and then also there is a regular reweigh that happens as an expected “fail-safe”. I prefer a cadence of reweights to forced reweights because former is easier for the users to plan for.
Hi Joey - the suggested tactical short-term steps are understandable… I think it would be better to state the immediate goal for these steps. Is it to first and foremost restore (and maintain) the peg at 1…? Allowing people to sell is good… but that by itself won’t be the goal - rather a means to it.
I think reweights do need to happen more frequently but I’m also aware that ultimately the price would go below 1 immediately (regardless of it being a natural or forced reweight) as long as there is no incentive to hold $FEI. Demand for $FEI is key to absorb the selling pressure.
How to create demand for $FEI ? The traits that people look for in stablecoins are stability, liquidity and usability. I’m afraid that $FEI is currently lacking on all these fronts.
Usability → I struggle to see how $FEI could integrate with lending protocols. As a collateral, it’s very risky to use $FEI because you could encounter a severe burn penalty if the position needs to be liquidated. Maybe the liquidity mining programs will help provide a use case in the short term but that probably will not be enough though.
Stability-> I think we can all agree that regardless of the specificities of the current situation, the design of the protocol makes it so that $FEI is more likely to trade below peg than above peg. Above peg, it’s very straightforward for people to arbitrage. Below peg, it requires faith that at some point the peg will be restored sustainably. So people are going to ask themselves “would $FEI hold its peg after a reweight ?” and inevitably they’re going to answer “no” because there is no incentive to hold $FEI (see above points on usability). Therefore there is no arbitrage opportunity below peg as long as the burn penalty is higher th an the buy rewards. Buying $FEI under peg is not an arbitrage, it’s a speculative bet on the future value of $FEI in my opinion because the possibility of $FEI losing peg right after reweight is big.
Liquidity → By design, the windows of liquidity at 1$ are very limited because of the exponential nature of the burn penalty. Volatility is enhanced by the direct incentives. I’m including here the thread of the CEO of Frax because I think it captures the issue quite well (no fud intended by mentioning another algo stablecoin, I’m a $FEI supporter) https://twitter.com/samkazemian/status/1378499341292236800?s=20
Regarding your suggestions:
1 & 3 - As said initially, more frequent reweights is probably a good thing to instill faith but it’s not a solution to the more fundamental problem : demand for $FEI
2- As long as the burn penalty > buy rewards, there is no real arbitrage opportunity so I don’t see why that would help with the peg
4- That could be useful. Essentially there are currently 2 prices for $FEI on FEI app : the buy price (oracle price + mint rewards) and the sell price (oracle price - burn penalty) and the asymmetric burn and rewards means that the protocol is capturing value (the spread) at the expense of buyers and sellers
Do not like the combination. Option 3 is not necessary. Because if there is no need to reweight, reweight shouldn’t happen, and certainly shouldn’t happen at pre-defined time interval.
But speeding up would be my preferable option. While setting a big parameter proportion to the distance is adding a parameter to the model, which adding complexity. If option 2 is prefered, i would just like the team manual tune the rate r a bit faster for now. Because the current situation should not happen at the normal operation.
By adding bit… the price shouldn’t settle at 0.9… with re-weight mechanism for an extended time (with more use cases around FEI). I would say probably the team could force-reweight(option 1) if current reweight takes too long to happen (another couple of hours and another couple of hours…which doesn’t help on our situation). And then observe what will happen, if it can release the sell-pressure.
At the meanwhile, adding use cases, integrations, etc.( you name it) are the gist to keep FEI stable for the long run.
Yeah think you hit nail on the head. Not sure how more frequent reweights help or any other similar suggestion helps without addressing why hold fei. Dont you actually get to the point that its rationale to buy for the incentives then dump on reweight to force off peg rinse repeat?
I think that option 3 (reweights every 3 hours) is reasonable to increase confidence in the protocol while there is no real use case for FEI, except holding it. By having a recurring reweight, people will start arbitraging and buying when the peg is below 1$. Pretty effective solution in my opinion.
(1) and (2) sends a signal to the market that you have strength and willingness to act quickly if needed, which I think would build trust. I like both of them, and would execute both. I think we want to avoid going to 100% loss of access within the first 48h.
(3) I think is healthy in itself long term, and the cadence can be adjusted during lower volatility periods if need be, but adds predictability, which arbritrageurs would love, and would help stimulate buying during low-peg. It could be announced as a 2 week special measure, considering genesis period, and re-approve/re-evaluate 2 weeks after.
(4) I think this is work worth doing for the long term, and that simply announcing that the work will commence, should bring trust and earn patience.
Some additional thoughts:
(option 0), some sort of public communication that is simple to understand, about all the things available in our toolkit to bring peg back up, and communicate that governance is taking place, and that the protocol stands behind FEI, and has massive resources, etc.
Option 3 is good and should be set to 60 minutes or even less.
I don’t see the other options benefits…
The important thing is that people that complains right now are the one that tried the short term gain and they are getting rekt by the protocol… mid to long term you don’t want these participants so if the protocol do his job then no reason to change this for short comers.
Again the plan from the beginning was to not preswap, buy tribe later, lp then stake
Ciao!
I like option 3, as long as reweights still require a minimum deviation from the peg.
That said, I think the larger issue is general demand for FEI. We still need more trading pairs, integrations w/ protocols, etc., so I don’t know that rushing to a hasty re-peg now is necessary.
But the problems mentioned by @federer remain. It’s hard to envision FEI as a collateral asset due to the burn incurred in liquidation, and we can’t rely solely on speculation for restoration from below peg.
I think the #1 priority right now should be liquidity for secondary, burn-free markets. This will create below-peg arbitrage opportunities, removing some reliance on speculation, and will provide exit opportunities for e.g. integrated protocols that need to liquidate positions collateralized by FEI.
Doesnt reweights every 3 hours just create an arb that you buy then immediately dump on reweight until all the fei is burned by the protocol trying to maintain peg on reweight? Could reduce confidence not increase it. Not sure more frequent rebasing is the solution
I agree !
I don’t know if for option 3 if it’s every 5 minutes what will be the impact on fei
If we can avoid bots to leverage the reweight it would be nice
I am in favor of Fei leveling out as quickly as possible.
This should be done in the simplest way possible in the shortest possible time, it will give people confidence in Fei Protocol.
After some time, you can change the technical solution for a better one. Now the main thing is to achieve 1 Fei = 1 USD
It’s better if the price mechanism brings back the peg more often. If that works, it would make more sense for arbitragers to buy and make the $ themselves, when they see a price lower than $1
(if you know the price is coming back in 10 mins anyway, you just buy that $1 for $0.98 yourself)
That being said the DI should prolly go and have it work like Maker. Fei was supposed to be undercollateralized but has way too much collateral right now. It’s easy to make it work like Maker
People’s behaviours will change if the price goes back to peg periodically. What is the most important thing to solve right now is FEI lack of adoption.
Without use case, there is no demand, and without demand, people dump.
By creating a recurring arbitrage opportunity via reweights, there is now a first use case for FEI: arbitraging. Arbers competing with each others will mitigate the peg issue really fast.
Can we get the exact formula for the reweight time calculation? This would surely help the decision process.
I favor solution 3 as the most pragmatic one. We should bring back the peg to 0.99 as soon as possible since time spent far from the peg damages fei perception as a stablecoin.