Ease of Subversion
Governments throughout history have earned a tidy profit through currency debasement. In the modern era of fiat currencies this has been pushed to ludicrous levels. Depending on which numbers you go with, for example, the US Dollar has been reduced in value by as much as 98% since 1900, with the surplus cash going directly to the government and its cronies.
Cryptocurrency is a threat to this major cash cow because it is community controlled, so it is difficult for one actor to subvert it to their own purposes.
Proof of work has no cap on its difficulty. To subvert a network would require ongoing investment in computing power and the electricity to run it. This is not only expensive, it is expensive unpredictably and continuously. Furthermore, it would be relatively easy for someone to notice that a government was doing it due to the space and power requirements for the hardware.
Proof of stake on the other hand, once you've purchased 51% of the tokens, that's it, the network is yours forever, and there's nothing anyone can do about it other than a hard fork. This is still expensive, but it's an easily calculable, one-time expense. Furthermore, it doesn't take a big pile of hardware and power usage that's hard to hide. So obviously proof of stake is preferable for large actors with plans to subvert and exploit a cryptocurrency for their own purposes.
Merely taxing proof of work schemes won't guarantee their extinction, even if it weren't an easy tax to evade. Currency which cannot easily be debased by a third party is, economically, an extremely valuable thing. Applying a tax is likely to simply drive up the price of the tokens to compensate. A total ban with harsh penalties for anyone caught using such systems has a much higher likelihood of success.