cunicula has presented a very interesting scenario and no one tried to prove it impossible except by trolling or personal attacks. I'll try to show that it's not so easy as cunicula claims it to be.
You see, the described scenario makes 51% attack very profitable (because the central bank profits off of it). This means that a third party will be very much motivated to claim this power to themselves, so the central bank will be constantly pressured to increase its hash power to compete with the third party, ideally to the point of next-to-zero returns from monetary policy. What do you think about that?
There are many third parties that can be involved. We the People are not always motivated by profit and can make a 51% attack very unprofitable for any attacker.
Final counter 3: Block IPs not trusted; all central blocks would then go into thin air - ONE update.
torDoesn't matter: If I see a 51% attack and set my client to only trust the mtgox IP for instance then you can fake coming from all the IPs you want using TOR, but that will never get your block on my client.
It's funny I am called the pig here when other posters are displaying such immense ignorance of basic TOR/IP/internet function.
Woah, you'll have to educate my sorry ignorant ass here. What, are you gonna ask mtgox to ignore blocks coming from tor aswell? Identifying a 51% attack must needs be based in properties of the blocks mined, not which IP they come from. Andresens' crude PoS, for example, linked in this thread would work on block properties.
TOR is an experimental network with limited capabilities. I doubt that Bitcoin will remain completely anonymous. A 51% attack would require tremendous resources that would be difficult at best to remain hidden and secret.