[variationvalue]

Because they can't use customer funds for these collateral requirements.

Note though that the collateral requirements only apply to firm-wide net buys. So if you're buying $100 of GME and another person on the platform is selling $100 of GME and those are the only two trades for GME that day, those trades would net out reducing the collateral requirement.

So on a day where there's a substantial net buy, money is basically being put up in twice, both by the customers and the broker.

[addingnumbers]

The question is why are there collateral requirements at all for a straight trade, though? In this example there is enough cash in the system to resolve everything without debt.

If you go to the bookstore and give them $105 for a rare $100 book +markup, and they say they can't get it because they don't have collateral, how would that make any sense? You gave them all the money they need to buy the book outright. Why collateral?

[variationvalue]

Ah this is a great question!

The answer mostly boils down to a policy decision to maximize safety of the overall system. You can imagine a situation where a malicious actor (not necessarily a broker servicing customers) who's a part of the DTCC system and decides to buy a highly volatile stock. Since the trade settles in two days, they don't really need to put up the money immediately, just need to be sure they have it in two days. Suppose that then this volatile stock plummets in value, this actor has a big incentive to not go through with the trade, since they're going to pay money for a stock that is now worth much less. This phenomenon is called counterparty risk. And the way to mitigate the risk is just to require collateral up front, with more collateral for more volatile stocks. Note too that the rules governing this risk mitigation are very distinct from the rules governing individual accounts (even different regulatory entities).

However, you could hypothetically carve out an exception for broker dealers since theoretically the customers have already put up the cash. But from a policy perspective you're now potentially betting the financial system on every broker implementing risk controls correctly. Why take that risk?

Ultimately the better solution is just to reduce the settlement cycle which reduces counterparty risk. Brokers want to go to T+0 (real-time) since it removes all collateral requirements on their end, but market makers like T+1 because then they can float more intraday (which is effectively passing the costs of mitigating counterparty risks back to the brokers).