I read through the Tether manipulation paper. IMO it made two convincing points:
#1 Someone has a habit of doing this:
- Issuing new USDT
- Within days, moving that USDT to BitFinex, Bittrex, and/or Poloniex
- Using that USDT to buy crypto (seemingly a portfolio of BTC & others). They especially like to buy crypto when the price is just below whole numbers.
- Moving the resulting crypto back to BitFinex
- Rarely or never selling the crypto for USDT again
The authors argue that this is Tether/BitFinex themselves, and I think that this is in fact the most likely explanation. But the authors didn't address the alternative possibility of this being a particularly ham-fisted whale who is a close partner of Tether.
#2 Due to end-of-month trading, Tether has probably always been trading with USD deposited with them (fractional-reserve), though at least until March 2017, USDT was probably not complete monopoly money, since they did go to the effort of achieving an end-of-month USD balance.
The authors also tried to argue some other points which I didn't find convincing.
I took the paper's data at face value. There were several points where I thought that they could be cherry-picking data, but it's too difficult to check this sort of thing. Cherry-picking / confirmation bias is especially easy to do with block-chain analysis. And I know for a fact that their method of grouping block-chain transactions is not robust in general, though it probably was sufficient for what they did here.
I've thought for a long time that USDT is almost certainly a scam, and this paper makes me think so even more. Though I was actually a little surprised that this provides evidence (via the end-of-month trading) that USDT ever had any real USD.
The paper estimates that if you removed the USDT issuance events which the paper's authors regard as most likely to be BTC price manipulation, the BTC price would be $4100 as of March 31. But that's based on a whole pile of assumptions; I wouldn't give it much credence. I think that the collapse of USDT will be mostly limited to the obvious direct effects (ie. some exchanges would have major troubles, there'd be many people stuck with worthless USDT, etc.), and there would not somehow be a natural "rollback" of any gains which monopoly-money USDT may have driven. Also, the paper makes clear that all major crypto was affected, often much more than BTC, so this isn't any sort of argument against BTC in particular.