He basically says that the fair value of dollar-tracking assets like BitUSD and NuBit is:
Fair Value = 1 USD
+ interest premium vs. USD
+ expected future appreciation (basically how far it is below parity)
- risk of total collapse
+ utility fudge factor (mostly affected by historical volatility, facility of transactions, exchange market depth)
For Nubits he then argues there is no interest premium and since it's trading at parity no expected appreciation, implicitly concluding that Nubit's utility factor must offset it's perceived risk.
For BitUSD he assumes interest premium and risk are comparable to Nubit which leads to this equation:
utility of Nubit = utility of BitUSD + expected appreciation of BitUSD
Because in the past BitUSD has been trading at a discount its expected appreciation is larger than zero concluding that its utility must be lower than Nubit's.
What is left when you remove Nubit from the discussion is that
BitUSD trades at a discount because it has a risk that is not offset by utility and there is no expectation that it might appreciate above parity.I basically agree with that summary (well I wrote it
) and would like to propose a fix:
Adjust a BitAsset's shorting threshold according to historical deviation from parity. For example if BitUSD has been trading at an average of 98 US¢ in the past few weeks allow shorting only at 102% of the feed price or above. I'm not suggesting that this is really urgent but if in a year's time with enough liquidity present the problem persists ...