If you look at the mirror list, some assets have high premiums and some have low premiums. The ones that have high premiums are the ones that have high short yields. The ones that have low premiums are the ones that have high long yields and low short yields.
Let’s ask a question: how might high short-farm yields be related to artificial buying pressure that keeps premiums high? Answer: by constructing a delta-neutral position on the short farm, of course…
I agree that the short position is important because without shorting there is no way to bring prices down. But there always is a way to short an mAsset, and that is by minting it! Shorting by minting is different from short-farming. I think the first is desirable and the second is not. Here’s an illustrative example:
Person A with UST decides they’re going to put it into Anchor/Mirror. They use a delta-neutral strategy to enter this market: they convert 2/3ds their money into aUST, use the aUST to mint 1/3 their money worth of mAsset at %200 collateral ratio, and provide the mAsset along with the last 1/3 of their UST to the long farm LP.
Person A’s total yield is 2/3 * 20% + 2/3 * LFAPR.
Person B with UST decides they’re going to use a different delta-neutral strategy to enter this market. but instead of minting the mAsset, they short-farm it with 2/3rds their money, They then buy this mAsset right back to cover their short, and get 1/3rd their money back in 2 weeks as UST; using that 1/3 UST and 1/3 mAsset, they provide to the LP to earn LFAPR.
Person B’s total yield is 2/3 * 20% + 2/3 * LFAPR + 2/3 * SFAPR.
The difference between Person B’s yield and Person A’s yield is 2/3 * SFAPR, paid to them by Mirror protocol itself, which is taking a dead loss on the difference. What is the difference in what A and B did? Person A minted and Person B short-farmed the same initial aUST. The mAsset that person A created has been provided to the liquidity pool through the long-farm contract. The mAsset that person B created has been bought right back, then provided to the liquidity pool through the long-farm contract. And if you look more carefully, you see that in the event of liquidation, Person B and Person A end up with the same assets after liquidation: 2/3 total money in long-farm LP token. In the end, Mirror is taking the same collateral for the same net transaction but stupidly is paying a greater rate for what person B is doing, even those what person B is doing is worse for the protocol.
So currently, the sLP does nothing for the protocol, and yet we pay people the SFAPR to make and hold them. Mirror is basically paying people to borrow their aUST to do nothing with it. This is an -exploit-, which needs to be patched.
Now I understand that the point of the sLP existing is for people to be exposed to swap fees while not necessarily holding a long position. That would be a valid use case for an sLP-like object. I think sLP implementation needs to be changed so that the UST from shorting is not returned, but is used to bind the mAsset into the sLP.
tl, dr: Minting needs to be incentivized over short-farming; but the protocol currently has this backwards and so it bleeds value.