Demand for BitAssets will BID UP BTS and thus create demand for shorting the BitAsset.
Or it could lead to a higher BitAsset premium rather than bidding up BTS price, and therefore not encourage BTS holders to short because they may fear that BTS price will continue to fall (leading to a further increase in the BitAsset premium). The hope is that the shorts will at some point say enough is enough and short the premium away and thus reduce BTS sell pressure, but in theory the premium could exist indefinitely and even grow since there is no mechanism to correct it back down to the price feed like there is for BitAsset discounts (the expired short covering mechanism). Also, even if they take this leap there is no guarantee that they will be able to buy back the necessary amount of BitAssets to cover at a reasonable price. So there could be a lot of BTS sell pressure added back at the time of covering negating the BTS sell pressure reduced when the short was matched.
However, if shorts knew that even if the price of BTS stayed the same over some fixed period of time that they would be able to profit over that period by shorting when the BitAsset was trading at a premium, they would be far more encouraged to short during high BitAsset demand thus supporting the process through which BitAsset demand translates to bidding up the BTS price and BitAsset premiums would be short lived (meaning a better peg). But the cost of this is that one cannot hold a BitAsset indefinitely; they would have an expiration time like shorts do (although there is no need for shorts and longs to have the same expiration period). And thus this would make BitAssets less desirable and less fungible. All fungibility would not have to be given up though. I'm imagining something like seasonal BitAssets where each BitAsset has four variants: Spring, Summer, Fall, Winter. For example, all BitUSD shorted into existence during the Winter season would be BitUSD-Winter and would expire two months into the Spring season. After the end of the Winter season no new BitUSD-Winter could be created until the next year. Even if the expiration period of shorts was raised from its current 1 month to 2 months (which might good idea for the sake of encouraging shorting), all BitUSD-Winter shorts would still expire prior to the expiration of the BitUSD-Winter longs. Two months into the Spring season, any existing expired or margin-called BitUSD-Winter cover orders would be matched with any outstanding BitUSD-Winter at the price feed at that time (actually the outstanding BitUSD-Winter would only be matched exactly at the price feed if there were no margin-called covers, otherwise the match price would be slightly more than the BTS/BitUSD price feed, to the benefit of the longs, to account for the margin-call covers offered at the 10% premium). The blockchain would automatically give BitUSD-Winter holders the appropriate amount of BTS (just like in black-swan liquidation) and the corresponding covered short owners would get back the remaining BTS collateral. So at any given time there would only be at most two BitAsset variants in circulation: the one for that season and maybe the one for the prior season.
Perhaps for convenience the blockchain could automatically (assuming the appropriate flag was set in the BitUSD-Winter balance record) place the BTS received for the liquidated BitUSD-Winter into a BitUSD-Spring buy order at a price relative to the price feed on behalf of the BitUSD-Winter owner. It could also move this order along to the BitUSD-Summer/BTS market if the order was not fully matched prior to the end of the Spring season, and then to the BitUSD-Fall/BTS market if the order was not fully matched prior to the end of the Summer season, and so on. And it would be really fantastic if the relative offset from the price feed for those expired seasonal BitAsset buy orders started out with some minimum (some negative percent offset) and grew monotonically with time to some maximum (some positive percent offset). For example, two months into the Spring season, any remaining BitUSD-Winter would be liquidated at the price feed and used in a BitUSD-Spring buy order at a price offset by -3% from the BTS/USD price feed, then this offset would grow to 1% from the price feed just prior to the end of the Spring season, then immediately after the end of the Spring season the order would still be at 1% from the price feed but in the BitUSD-Summer/BTS market, then 1 month into the Summer season the offset would have grown to +2%, and finally the offset would grow to its maximum of +3% from the price feed by the end of the Summer season. If all goes well, a BitUSD holder could leave their balance alone and hopefully have close to the same value of BitUSD whenever they check it in the future (and likely more if you account for BitAsset yield).