This thread is mainly for discussion by those interested in arbitrage.
It seems to me that a number of arbitrages around the bitUSD discount/premium are difficult or not risk-free. Maybe others have better solutions. Good strategies around these will be important for liquidity around the peg.Discounts
If bitUSD is trading at a discount to fiat USD on an external exchange, but at the feed price on the internal exchange against BTS, an arbitrager can convert buy bitUSD with fiat USD on the external exchange, sell it on the internal exchange at a more favourable price, send BTS to the external exchange, sell for fiat USD and make a profit. If the converse were true, the reverse set of actions would ensure a profit. In this case, the arbitrage is fairly risk-free except for timings.
If bitUSD were at a discount on both exchanges, this is not possible. However, an arbitrager willing to wait for up to 30 days could buy bitUSD on either exchange, and in theory
sit with a sell order at the feed price on the internal exchange until it is hit at the feed price (*which it must be due to short expiries, as shown in another thread here https://bitsharestalk.org/index.php?topic=9861.msg128218#msg128218
), then sell the BTS again for fiat USD, making a profit. The technical problem with this is maintaining a sell order that moves with the feed price, which is costly and manual, and I'm not sure of the best solution here yet. So in practice, this does have risk of not getting hit when or where an arbitrager might want it to.Premiums
If bitUSD is at a premium to the peg on the internal (v BTS) or external (v fiat) exchange, but not on the other, the arbitrage is similar to that of a discount only in reverse. For example, if its at a premium on the external exchange, not the internal, buy bitUSD on the internal with BTS, sell on the external, buy back BTS. This gives a risk-free profit.
If bitUSD is at a premium on both exchanges (which it should be very quickly once the first set of arbitragers act as just described), then the trader can short bitUSD (and hedge their increased BTS exposure by either simultaneously selling BTS or going long bitUSD), in the hope of covering the short at a lower or zero premium. However there is no guarantee this premium will reduce if bitUSD demand remains as strong. The short may be forced to cover at the same premium (zero profit), then roll to try the same play again in the next 30 day period. One would think that supply and demand would balance at some point, in which case a profit should be realised when the premium narrows, but the timing is unknown, and not guaranteed. So in this case, the trader is betting on a movement in the spread (or really, a change in future bitUSD demand), rather than being able to implement a risk-free arbitrage.
The situation in this latter case can be made even more risky if there are very little BTS left outside of the collateral pool (if BTS is not much larger than 3x the market cap of bitAssets). In that case, its possible a shortage of bitUSD could be prolonged or even grow to an extreme if bitUSD demand is persistent, leading to a rising premium. Also, on this borderline where the BTS market cap is barely sufficient to cover the pool of bitAssets, any decline in the BTS price could force shorts to cover at an even more unfavourable price.
[Edit: As an after-thought, interest rates will have some impact on restoring discounts and premiums, question is how much? This could take up to 30 days for all shorts to be replaced (ignoring current grand-fathering). In the case of a premium, the floor is at 0% interest, and people may still want to own bitUSD for other reasons.]
Any improved ideas welcome.