So how do discount brokerages make money?
NET INTEREST
Suppose I were to give you $100, in return for your promise to give it back when I wanted it and pay me 0.27% annualized interest in the meanwhile. Suppose you invested this in a virtually riskless bond, perhaps a mortgage-backed security with government backing, offering 2.53% annualized interest. You’d earn $2.26 in net interest in a year.
Suppose I were to give you $200 billion dollars. Now I’m the American middle class and you’re Charles Schwab. You would earn something like $5.8 billion dollars in net interest income. This would entirely pay for your sideline business in running a brokerage. Stocks, bonds, mutual funds, branch offices, call centers, blah blah blah, it all exists to justify the only pricing page that matters, and all the verbiage on the pricing page is about how much you pay the customer.
This is an exaggeration, but not much of one. 57% of Schwab’s revenues are from net interest. The firm could literally give away every other service; discount the mutual fund fees to zero, do away with commissions, etc etc, and they would still be profitable.
[...]
Did anything useful come out of this research project?
Three things that I believe much more strongly than I did previously:
Commissions in discount brokerages will go to zero. Like most repricing, it will continue to be slow until it happens all at once. Interactive Brokers will likely still continue charging them, as it is least similar in character to the rest of the industry.
Roboadvisors are a bad business below scale. I have really liked the model as a user, but I am glad I’m an investor using a roboadvisor rather than an investor in a roboadvisor. The discount brokerage industry works because you can make decisions which are bad for users and lucrative for you (“Ahh you have entrusted me with cash money; let me park it safely and pay you 50 bps while keeping the next 150 bps for me, rather than parking it in the equally safe place any of my product managers could have implemented if they wanted to be fired instantly”); roboadvisors depend on making decisions which are good for users (“Let me minimize that cash drag for you”) and then being very explicit about costs which are anchored very low. (Again, hate to belabor a point, but Wealthfront charges 25 bps all-in (on top of the underlying ETFs) and every customer knows it; Schwab charges 18 bps for cash management alone and virtually no customer has ever even thought there could be a number there.)
Some technologists I am acquainted with should know one esoteric thing about equities markets. If you own a material amount of a publicly traded stock, and the supply of that stock is constrained, and some population of sophisticated people want to short it, maybe you should Charge More rather than letting your broker get that service from you for free. Speak to your trusted advisor, this can have fairly toothy consequences, etc etc.