I'd say its more correct to say orders from retail investors are valuable because:

1) They tend to be a relatively equal number of buys and sells. This is figurative manna from heaven for market makers

2) They don't mind crossing the spread, also a gift for market makers.

3) They can't really move the spread

and probably least importantly......

4) they aren't trading based on some sort of knowledge most of the time, ie they are typically just buying because they buy a bit each month or selling because they need money or selling to re-balance,

ie they aren't trading due to market signals, but rather due to their own personal situations.

Instead you are likely buying from a market maker. For pretty much any stock there are many market makers who are offering to both buy stock and sell it at the same time. The trick is they'll buy for, say, a penny less than they are willing to sell for. So if they can keep their buys/sells even they make a penny for every share they move. They're middle men.

Make sense so far?

OK, but there is a problem. Say that REALLY BIG HEDGE FUND has proprietary knowledge that some company is probably going to go up in value soon. So they go out and buy a lot of stock from a market maker. And then the stock goes up. Uh oh. The market maker sold a bunch of stock at PRICE and now instead of buying an equal amount at (PRICE - 1 penny) and making money the market maker has to buy an equal amount at (PRICE + the_stock_went_up_amt) and loses money.

Basically market makers would prefer not to trade with sophisticated investors making big trades with proprietary knowledge. When that happens and they lose it's called getting picked off. They want to trade with people like you and me who aren't trading because of any special knowledge but because we're just putting our regular $1,000 in our 401k for the month.

Retail brokerages like Robinhood are a good source of these types of trades.

(As with a lot of topics, there is a lot of nuance and detail underneath this relatively short description, but this is the gist of it.)

Retail flow is driven by things like needing money for purchases or having excess cash lying around.

Also retail flow is less execution cost sensitive.

That's what people mean when they say less sophisticated.

Order book data is valuable because it helps you figure out who everyone thinks is pretty, and retail investors make these decisions on much different factors than traders (I.e. Brand over financials)

https://en.wikipedia.org/wiki/Keynesian_beauty_contest

I buy the stock myself a few days later, but at $40 each for a grand total of $4,000, and I just made $1,000. You get your 100 shares, I make $1,000 plus fees. Now imagine doing this hundreds or thousands of times a day. Or I simply sell my own existing supply conveniently to my own buyer and I dictate the fees, which has value in itself. Sure I'm only making $5 per head per month with these low-end customers, but now I have 100,000 customers instead of 1,000.

A savvy investor wouldn't ask to buy these shares. He'd know that he's better off shorting himself. Unsavvy investors are all over the place and if you can get them to come to you, you can make a lot of money off them with minimal investment or risk.

On top of that you can offer them predatory loans for investing. This app offers loans, no idea on terms, but I can't imagine them being very nice. They put up collateral like their car or home, so you have no risk. They fuck up, or not, it doesn't matter, you're at least getting fees and probably screwing them in other ways. Oh and you're getting interest on that loan as well.

edit: Front Loading absolutely happens, it can be done in a way that's difficult or impossible to detect and has a lot of loopholes as well.

Reality is more like "I know an order for NVIDIA is going to hit at $50, but I see it for sale for 49.9998 all over the place, I can go scoop some up at 49.9998 and put for sale at 49.9999 and make some money.

But then real reality is a step more complicated about when it is legal to do this vs not (front running is not legal, but things that are similar can be), and dark pool vs not, and you are way past what I know ;)

In buying order flow, you are usually going in the opposite direction, selling out of you own inventory against a buy.

The firm isn't buying days later after a price craters. They are usually buying or selling out of their own inventory. Sometimes this is speculative trades, more often it is more mundane like saving transaction fees (and a higher certainty of transacting) from going to the market.

So let's calculate the fees. For a given stock, FCKD, the bid offer is $4.50, the lowest offer is $4.60. Transaction costs for this are $0.10 at the brokerage.

Investor A wants to buy a share, B wants to sell one, on the same brokerage. Okay let's see.

Brokerage does nothing (because the amount of shares they hold on the exchange for their clients doesn't change at all). So they don't pay anything for that either. No transaction fees, nothing.

A gets to buy the share, and has to pay the lowest offer + transaction costs for the share, or $4.70.

B wants to sell the share, and he gets to sell if for the highest bid, or $4.50, and in some cases another $0.10 in transaction costs.

So the brokerage "captures the spread", and of course the transaction costs. In this case they made $0.40 for doing ... nothing, no share was sold or bought.

This is the official story. Now in practice, this means that a rather large amount of value is "stored" at the brokerage. It is a long accepted practice in banking that only a percentage of the stored value is kept. So while you might think that as a result of this transaction the bank/brokerage makes $0.40 in profit, but it doesn't. Assuming 10% reserve ratio (which is on the high end, usually either 2 or 4%). It makes (100% - 10%) * 2 * $4.55 + $0.40, and pays that out to it's shareholders/managers/... (twice, because once in cash, one in a share)

Now you might think that's where it stops. Well, not quite. They then go to another bank, and they say "we are holding an asset worth $X on behalf of our customers, can we borrow against that ?", and the next bank say "sure ! if we can do the same". So bank A gets a 2 * $4.55 debt at bank B, and bank B gets a 2 * $4.55 debt at bank A. Since this also counts as reserves, they then only keep 10% of that money actually available, and pay 90% out to their shareholders.

So now the bank has paid out (100% - 10%) * 6 * $4.55 + $0.40 to it's shareholders.

Needless to say, keeping this whole situation stable if the market were to gasp drop even slightly is more than a little tricky. Hence, "too big to fail".

That's why you want to be a bank. This is the reserve currency system we currently work with.

>(100% - 10%) * 2 * $4.55 + $0.40, and pays that out to it's shareholders/managers

>(twice, because once in cash, one in a share)

>can we borrow against that ?", and the next bank say "sure ! if we can do the same

> Since this also counts as reserves

> pay 90% out to their shareholders.

> if the market were to gasp drop even slightly

> Assuming 10% reserve ratio (which is on the high end, usually either 2 or 4%)

Doesn't this also enable front-running? Regardless, this is useful info for HFT funds, right?

> Robinhood also earns money from rebates its gets for directing its order flow to broker dealers, though Bhatt insists “We do not sell data to anyone. We have never sold data to anyone. We just do not do that.” There have been misconceptions that Robinhood sells high-frequency traders its data to help them trade against the startup’s customers. But Bhatt says “The rules around this stuff as so tight. We’re not a social media company. If we even step slightly out of line with anything we all go to jail.”

Basically all they're claiming is that they're following the rules - i.e. if the rules allow selling "data" (or something that we won't call "data" but "order flow" or something) to HTFs or banks' dark pools, they're not saying they don't do it!

So they're selling some data (as opposed to running directly on an exchange, without any intermediaries). Or maybe not "selling" in the strict sense (I give you data, you give me money), but in an indirect way (I give you data, and you give me discount on your other services).

That's not what rebate means in this context. It is, in fact, actual money.

Robinhood also isn't just giving someone data. They're sending order flow to various market makers. Not data about the orders but the actual flow in need of execution.

This flow is a source of customers for market makers. Those customers get charged a fee. A % of that fee is rebated back to Robin Hood (or other retail brokers) for being originators of the customers.

https://en.wikipedia.org/wiki/Front_running