To be more transparent, maybe IB should openly use the proceeds to create a portfolio that directly bets against those same positions that traders are overexposing themselves on, to show that if those accounts go down, IB will directly hold the counteracting market positions that will directly benefit from the customer's losses to protect the firm, hence a true hedge against their customer's positions. Sort of like what Goldman did with CDSs.
So if some guy is shorting a tonne of tail risk premium as part of their strategy on the S&P via a bunch of ES puts, then maybe the fees collected by IB from the pool of funds from over leveragers, should be used by IB to buy puts or put spreads (cheaper) directly betting against the S&P so the firm remains delta neutral against those customer's positions. They should enter the trades well after a customer enters theirs, to show they are not trading against their client's order flow but merely as a true hedge against tail risk.
Of course, customers positions are fluid so customers can close a position, IB has to close a hedge and execution costs can make this untenable. But sure, clearly IB won't do this and it is likely just something that goes into their top and bottom line, but seriously some people are way too brazen. E.g. the attitudes about 30% moves being 'impossible' just because historically nothing exceeded 30% on an index. I would think 'anything is possible' is usually the way to go here.
I think gap risk is the real fear. Some people here say 30% moves are 'not possible', but they are. For futures there are circuit breakers, so accounts not in compliance will be liquidated long before a 30% move. But their risk measure is based on the portfolio through a family of instruments including equities and equity options. Equity can sustain 30% gap moves easily and it can occur at random, not even during earnings. E.g. POT gap down after the uralkali news last year. So with a PM account and someone short the options too, they can probably get creamed in those cases from a gap move in the underlying of 30%.