The problem with this idea (aside from the gratuitous use of bitcoin, and the pricing out of people who can't pay) is that punishing people isn't the hard part of online moderation. For most simple cases, a system of warnings, temporary suspensions, and permabans mostly works. Yes, you've got to deal with sockpuppets, but my understanding is that they're usually fairly easy to recognise, and the extreme cases are usually rare and notorious figures. Nothing in this process is so complex that it requires a stake system to be fixed.

No, the problem is that determining bad behaviour (or at least, doing so in a way that is fair, just, and broadly accepted) is very difficult. There's no objective naughtiness meter that detects when someone's behaving in bad faith. There's no set of words that always, regardless of context indicate undue rudeness. There's no perfect demarcation between just asking questions, and harassment.

So basically any system is going to have to make a series of judgement calls, and those judgement calls are going to be specific to the context of the person (or people) making the judgement. And most importantly: other people will disagree with that judgement call! Not least the poor victim of your unfair banhammer (or the poor victim of the abusive spammer you choose to do nothing about).

Money's not going to fix that problem - if anything, it'll make it worse when people have more on the line.