There’s several issues here.
First, even if we agree to go with a middle ground, then USDe should have been allowed to depeg for DeFi to .995 rather than be hard coded to 1. This is safer because there could be instances where the deepest pools of liquidity depeg more than that.
Second, I even if I agree with the framing that the goal is to separate out the case of temporary secondary price dislocation versus permanent impairment of collateral, this goal is not easily achievable in true tail scenarios. From the perspective of Binance, they can 1) use an oracle that looks at deeper liquidity pools that are not their own to determine if USDe should be liquidated as collat or 2) they can treat all of DeFi and other exchanges as invisible and exogenous to their own closed system. If they chose 2), then they are not able to determine what is a temporary dislocation versus a true impairment of collateral because it looks the same to them. So then how things played out is exactly the result. If they choose 1), then there are other tradeoffs. 1a) They are effectively loaning their own balance sheet out temporarily based on the trust they have toward the custodian, the oracle provider, and other exchanges. In extreme cases where let’s say USDe becomes hugely successful and much bigger, Binance would be betting the solvency of their exchange on trusted third parties. If they are wrong in their judgement, users hold the bag. At this point, you might as well just formalize a cross-industry clearinghouse thus reinventing the wheel from tradfi. That would at least be better than the current structure. 1b) In the time it takes MMs to cut over liquidity from USDe DeFi pools to CeFi, the liquidity on Curve/Uni/Fluid could deteriorate. In other words, local books guarantee higher solvency than expecting liquidity to be cut over from external books which is not as instant. Easy example is if two exchanges have local book dislocations and then wait to liquidate based on a DeFi oracle. MMs go to grab the same liquidity to service two exchanges. You can’t grab the same liquidity even though it initially looks like you can.
Third, at some point the tri-party custodian agreements will be tested. The custodian will have to decide whether an incoming margin call is real or fictitious. If they ignore a real margin call, Ethena is safe but the exchange and its users get rekt. If they fulfill a fictitious margin call, Ethena gets rekt and true impairment of collateral will happen.
Fourth, if Ethena has a special deal with exchanges for ADL protection, does that mean any firm with the same transparency, similar risk profile, and a tri-party custody agreement can get this privilege? If not, this would break exchange neutrality. If so, it puts undue burden on the exchanges to have case-by-case customized integrations to check up on any user who wants this. The issue here is that exchanges are taking on brokerage function. The exchange should be neutral while the brokerage doesn’t have to be. The brokerage risks their own balance sheet in isolation but the exchange, by risking their balance sheet, affects solvency for all users. This would also be a reinventing of the wheel from tradfi but long term would help avoid conflicts of interest.