Except the people were exploiting inefficiencies in markets (and still do today). HFT algorithms exploit inefficiencies in servers, ISPs, fiber optic cable, a competing algorithm's implementation, etc.
And the tall loud guys exploited inefficiencies in trading pits and human perception. Ask yourself why I explicitly described them as "tall loud guys in brightly colored jackets"? Those are not irrelevant details.
HFTs do the exact same thing that pit traders did. They just do it faster and cheaper.
And if you really think they don't provide a useful service, you can very easily NOT buy their services - just change one flag in FIX. The question to ask yourself is why everyone chooses to trade with them.
Those things can't be meaningfully separated from the market.
Exploiting inefficiencies in a competing algorithm is the same type of thing as exploiting inefficiencies in the mental model of the guy standing next to you. Taking advantage of better market proximity is the same type of thing as drinking less than the other guys in the pit.
If you presuppose that the purpose of trading is to make money in and of itself then sure. Of course the real (original?) purpose is more tightly coupled to reality and to real goods exchanging hands. There seems to be a fundamental difference between profiting from the exchange and profiting from the mechanism of exchange.
I'm not asking for how we ended up here. I get that faster = better and 1ms faster is still 1ms better. What I'm asking is what value it brings to the world — specifically the world outside of, historically, the pit, and today, outside of the computers executing trades.
Are goods more accurately priced? Is there more liquidity in the market? Is the market more stable? Is the market more efficient, or is it only the technical implementation of the market that's made more efficient by these? Do these benefits, if they're present, outweigh the cost of things like flash crashes caused by algorithms? If flash crashes hurt all of us, then shouldn't we all be benefited by the algorithms when they're doing well? Do we benefit and by how much?
This is why people want "nerdy HFTs" to be taxed in a special way. Maybe it's because they can't articulate the value or maybe it's because there is no value. I can't see the difference and I can't get anything other than evasive comparisons when I ask what the difference is.
Borrowing a point from tptacek, Vanguard has come out and said that they have lower costs due to HFT. Their customers save money over the previous providers of liquidity.
Awesome, I appreciate the link. It's still not totally clear to me how/why this helps Vanguard, or whether the pros outweigh the cons for the market as a whole, but this is in fact the type of evidence that I've been asking for. Thank you for sharing!
It helps Vanguard because they don't make money through active trading: they are literally the vanguard of the movement, now practically accepted as orthodoxy, that funds should be passive and diversified across the whole market.
If you buy and hold, HFT helps you, by reducing the cost of execution --- both by reducing the spread, which is a tax you pay any time you place a market order, and by literally reducing trading fees. In fact, even if you're an active trader, HFT usually helps you: the only people truly harmed by HFT are the ones who tried to make a living selling liquidity before, who are now being outcompeted.
