There are a few misunderstandings here.
ECAF is not a police or an enforcer. It's the judge. What does it judge? Disputes. What kind of disputes are under its jurisdiction? All kinds of disputes in the EOS blockchain which can't be resolved in any other way, because it's the default arbitrator.
The Constitution is a multiparty contract. That's the contract which defines the jurisdiction of ECAF. When ECAF makes a ruling for a dispute, the dispute has been something that was against the Constitution, for example, transferring tokens without the consent of the owner.
If we remove the default arbitrator from the Constitution, there will be disputes that won't get resolved. Users can try to get them resolved through traditional justice systems, but as most already know, it will be very difficult.
The arguments for and against a default arbitrator, IMHO, should be made from this perspective. Can we create a system which can resolve disputes in a protocol layer? If there are risks, are those risks so big that it's better to leave disputes not resolved?
We have already seen that there is much bigger demand for protocol layer dispute resolution that anybody could have imagined. Hundreds of open cases. If the protocol layer dispute resolution is removed, all those cases will be left unresolved.
I think that we really need a protocol layer dispute resolution. Not only for obvious problems, like theft, but for everything that we can't foresee. The more complicated system we have, the more there will be ways to create new kind of problems and disputes. If we don't have any kind of dispute resolution system, the end result tends to be bad.
And maybe even more important is to actually create a market economy. For free markets to function properly, we absolutely need private property rights and a dispute resolution system. Those both are formed by the current Constitution. V2 would ditch them out, so I'm not sure if we can say that we have free markets with it.
Your example how things could go wrong with ECAF on the protocol layer is not very convincing. Governments can't just say to it that it needs to shut down an app. ECAF can only make rulings on disputes. So first there needs to be a dispute. Let's assume there is. What kind of dispute would require to shut down the whole app? That would require pretty convincing arguments if it was to be accepted by BPs and the rest of the community. If ECAF was forced to create obviously false ruling, it would create a shitstorm immediately and very likely BPs would not enforce it. And even if it would somehow successfully to be sneaked in and enforced, without anybody noticing what's happening, the app could be brought back online on a different account. So nothing would be achieved, maybe the contrary, because now everyone would be aware of the app and it would get a lot more users and the next time it couldn't be shut down so easily.
Of course, there is always a way to replace ECAF it was to be corrupted or forced to follow outside orders. It requires a change in the Constitution. It will take longer than replacing BPs, but I pretty sure it would be easy to do. Because ECAF is used by everyone, everyone must care about it. This creates a huge incentive for everyone to make sure that it functions properly. And if it doesn't, everyone has an incentive to pressure it to fix itself, or to vote for another arbitration forum to replace ECAF on the Constitution.
RE: Proposal for an incremental Constitution and Dapp layer governance on EOS