I especially like that short positions set their own initial collateral level. Doesn't that also mean that providing a grace period is unnecessary, since a short can decided how to collateralize their position and thereby determine the probability that their position gets called within a given period.
I wouldn't say it is unnecessary. Certainly being able to adjust the collateral ratio (via choosing initial collateral and adjusting the collateral of an existing short position through add margin and partial cover operations) helps the short owner adjust the likelihood of having their short position called. But the grace period gives certainty that the short position won't be called. For example, even in the situation where all BitUSD holders want to simultaneously exit, the owner of a short position within the grace period still doesn't need to worry about their short being called. That of course also assumes that the short position maintains sufficient margin to not be margin called and also assumes that settlement is not triggered either due to an undercollateralization event or through
the vote of the feed producers and majority of shorts. Although since the last trigger can only be done after 30 days, the short owner that could be affected by the settlement event would have had plenty of prior warning of that potential risk (with a 14 day grace period, they would have already been at least 16 days into the 30 day period at the time the short position was created).
I don't think shorts should have to hold onto extra funds to cover their position. Would it be possible for shorts to remove collateral from their position up to the margin call limit in order to place themselves at the head of regular call queue?
I agree with this. I believe that the short owner should be allowed to add and remove margin as they please as long as the collateral leftover after the operations are complete has a collateral ratio (at the current price feed) that is above the margin call ratio limit. This operation also makes partial cover more sane. There would no longer need to be silly tricks like breaking up your shorts into pieces to minimize the amount of spare BTS to keep outside of the locked collateral.
Additionally, why is there no yield now? I thought yield came from transaction fees as well as from the interest offered by shorts.
I suppose there could still be yield from transaction fees and overlap fees, but due to yield harvesting and also due to the fact that most of the yield would have likely come from short interest, I think it makes sense to just send the transaction and overlap fees to BTS shareholders instead and get rid of the idea of yield for BitAssets. The nonfungible bond market will likely be the best way of allowing longer term holders to get interest on their savings.
In this new BitAsset system, the people who can profit from eager shorts are those who are willing to buy the shorts offered below 99% of the price feed (if any exist), wait for X period of time (e.g. a few days) and then use the BitAsset they bought to redeem BTS at profit (assuming the price of BTS relative to the asset has not gone up too much in that period of time). They still need to take some speculation risk, so I don't think it makes sense to have the blockchain attempt to do it.