Direct redemption is the simplest and most desirable guarantee for a stablecoin. Holders want to know that at any point in time, a stablecoin can be exchanged for its face value. A direct redemption mechanism is the simplest way to make this happen.
Many stablecoins such as FEI, FRAX, and LUSD use a redemption system as part of peg maintenance. FEI’s other peg maintenance mechanism, reweights, suffers from a number of drawbacks such as high implementation complexity, delayed response times to price shocks, and allowing some MEV through sandwiching. As such, Fei V2 will remove reweights in favor of direct redemptions.
Redemption will thus take on the full burden of peg maintenance. The primary way that redemptions are currently used is in risk-free atomic arbitrage. When FEI is trading below $0.99, arbitrageurs can buy FEI and redeem it for $0.99. However, standard DEX fees of 0.3% make this trade less efficient, and only risk-free when FEI is $0.987 or below. As such, keeping FEI above $0.99 requires raising the redemption value to at least $0.993.
Beyond peg maintenance, there are also utility and narrative benefits to raising the redemption floor. A stablecoin is more useful if it can be redeemed more closely to its face value. Raising this backstop would make FEI more competitive with other stablecoins that have similar redemption mechanisms.
To determine whether this fee range is reasonable we can look at fee ranges of other stablecoins that allow for redemption:
LUSD redemption fee: >= 0.5% (dynamic)
FRAX redemption fee: 0.55% (redemption allowed when FRAX < $0.9933)
Current FEI redemption fee: 1.0%
Proposed FEI redemption fee: 0.5%
These other protocols have been able to thrive in this fee range without issue.
Proposed Changes
Raise the reserve stabilizer redemption floor from $0.99 to $0.995.
Final Notes
Please share any thoughts you have on the above proposal.
I think its a good move. In a marketing perspective, its good to say $1 redemption with 0.5% fee instead of $0.995. Its easier to get the idea of how much you “pay” for redeeming.
For the future, I think its interesting to explore something as the design of Gyroscope of having some flexibility in the redemption price to avoid a bank run in case collateralization is low.
Good idea, direct redemption is far simpler than the current system, both technically and from a narrative standpoint.
Recently, I’ve been giving some thoughts on mechanisms to grow FEI, but I’m struggling with something. If the protocol never reweights, what makes FEI worth 1$ ? If redemption is 50bps under 1 USD, FEI calls itself a 1 USD stablecoin, but I think FEI is effectively a stablecoin worth .995$. It is only the narrative of being a 1 USD stablecoin that makes people trade it for 1 USD worth of ETH, or 1:1 with DAI/USDC/USDT for instance. Plus the fact that it has significant liquidity against ETH, so we know from ETH price fluctuations that FEI will eventually be worth more than 1$ sometimes. Or am I missing something ?
This is how I’m interpreting the two recent proposals by Storm:
Given the overwhelming support and pending integrations with Aave, Compound, and our activity on Curve, on top of the existing and thriving integrations, it is necessary to optimize the protocol for increased utility and velocity of FEI across DeFi.
increased caps and updated parameters of ETH, DAI, RAI, and DPI bonding curves
Add LUSD bonding curve
Bonding curves have been a very effective tool for PCV growth and diversification, and we should continue to leverage them moving forward while exploring additional use cases for yield generation with the acquired funds. Per the proposal, we are not incentivizing just making it available.
Guaranteed redemption (initially at $0.95) has proved to be one of the most effective peg stabilization tools. Peg stability is one of the main goals of the Fei Protocol and moving the redemption rate to $0.99 has proved effective, but still, dependant on periodic reweights for tightening of the peg. With the lowered buffer to $1.005 (reflected in the updated ETH bonding curve parameters) and increased redemption to $0.995 we are inviting the market forces (arbing) to keep the peg as close to $1 as possible, and are standing behind it with our evolving PCV strategy.
I think this is a great step forward and is aligned with the v2 path.
agreed. in the future it would be good to move toward some sort of dynamic fee system. LUSD’s system is pretty nice: each redemption increases the fee, and then the fee exponentially decays back down to 0.5% over time. you can see this visualized here. im not familiar with gyroscope’s system but will look into it
bonding curves address this. when FEI is trading above $1.01 (or whatever the bonding curve premium is) then arbitrageurs can buy FEI off of the bonding curve and immediately sell it to bring the price down toward the bonding curve price
one source of fluctuations is FEI’s pairings with volatile assets like ETH and DPI. another giant category of fluctuations is general supply and demand for FEI within each stablecoin usecase like yield generation. another source is using FEI to arb other tokens via FEI bonding curves and other AMM’s. due to these market forces its not possible to peg any digital asset at exactly a single value, instead all we can do it create mechanisms that keep it in as tight a range as possible
What is the harm in making the redemption even tighter - say $0.9975 vs $0.995? With the former, FEI gets yet another marketing narrative + an even tighter peg
the redemption mechanism depends on having an ETH-USD oracle, and the chainlink ETH-USD oracle’s rated precision is only 0.5%. using it in a way that demands higher precision could allow the protocol to be exploited
Entering last call on this proposal. Will run a snapshot vote on this proposal in a couple days unless people have significant objections or suggested revisions