We are all seeing that premiums are getting out of control in recent weeks, steadily widening the spread between oracle and LP price. Based on the current MIR reward distribution mechanics, we can anticipate that the spread is going to get even wider if no appropriate changes are made in the MIR reward distribution system.
Understanding the MIR Protocol’s system:
The price pegging system is based on the correlation between premiums and the long/short APR you’d get for LP and sLP positions. If the premiums are getting higher, the APR for shorting the mAsset should rise accordingly, encouraging the user to open short position, hence bringing down the LP price to the oracle price.
The APR number is based on the varying factors of MIR reward distribution mechanics, the LP liquidity value (TVL in UST) and the shorted token amount vs. LP price. The MIR reward distribution mechanics is a mathematical formula, written in the documents, which basically relies on the weighting of the mAsset relative to other mAssets (higher weight → higher MIR distribution for the mAsset relative to other mAssets), and also varies if within that allocation, the MIR distribution for LP relative to sLP changes, which is subject to the premium.
Based on the current settings of the formula, the LP-sLP reward distribution is skewed towards LP in 60:40 - max. 40% allocation if premium is larger than 6.25%, if it’s below it and towards 0% premium, than the allocation would even rise to 100% for LP. It means that even at 15%+ premium, you’d get higher rewards allocation for longing the mAsset relative to shorting it. Obviously, we need to change the LP vs. sLP reward calculation if we’re to encourage shorting the LP price back down to the oracle price. Else the protocol will fall apart.
Other than that, the liquidity provided to an mAsset (and also the amount of shorts and the LP price) also alters the long/short APR calculation - higher liquidity value causes lower long APR, higher number of shorts causes lower short APR. Increasing the weight index for mAssets with high liquidity would rebalance this progress, but given the liquidity distribution of all mAssets throughout the protocol, there’s no need just yet.
Changing LP/sLP reward distribution formula
We can change the reward calculation for LP to Rk=(1-2,5rs) … and for sLP to Rk=(2,5rs) … Simply by multiplying the rs (short ratio) by 2,5 (based on the table at /Staking Tokens (LP & sLP) - mirror), the reward distribution rapidly changes in our favor, at 2% premium distributing symmetrically, at 6.25% premium it’s skewing towards 100% allocation for sLP relative to Lp. This is an easy, yet effective method for the recalibration. The best option would be though, if the formula would distribute symmetrically from 0% premium as a mean, and exponentially increasing/decreasing the long/short APR based on the absolute distance from the mean in either positive, or negative price premium. In case of expanding the short ratio table to negative premiums as well, Rk=(1-|2,5rs|)… for LP and Rk=|2,5rs|… for sLP gives a good guideline.
Given the significant change in long/short APR based on the above, it would be enough to bring back the price peg. In case one would fasten the process, the best option IMO would be to lower the collateral ratio by 10% among the most depegged mAssets. In this way a little credit impulse would help starting the move in the good direction.
…Hope not miscalculated something, long night.