Two vulnerabilities come to mind: (1) it would reduce the cost of a 51% attack. It would become a 36% attack! Hehehe! Sure, there’s an argument that this would be good for tribe holders. It raises the secure price floor from around 2 times the PCV to around 3 times the PCV. But that “secure price floor” is only a floor in theory. We don’t know how well it will support the market at the moment because TRIBE isn’t fully liquid yet. And a lot of the tribe that can already be voted with (read: the tribe owned by the team and seed investors) cannot be sold. We’ll see what happens in the future. (2) The bank is owned by its debtors. All I can say about this is that it feels risky! 
On the other hand, some people might read “the bank is owned by its debtors” and think that it sounds like the epitomy of DeFi. I haven’t come to a conclusion about this. I’ve tried to come up with strategies that would abuse this power, but I haven’t been able to come up with any. I’ll wait for people that are better at breaking things than I am to give it a try.
I agree with you here. I think many of us (myself included) have a bit of a dogmatic resistance to the idea that FEI would be reserve-backed. In reality, the ecosystem would probably benefit from having a faucet (even if an indirect one) where people can sell FEI for ETH at $1. And what’s nice about your solution is that the size of that faucet will be dynamic, but capped at 20% of PCV in the case that the only use of the loans is arbitrage.