[mrchicity]

Why's it a waste? Low latency traders facilitate risk transfer in thousands of instruments at razor thin margins using automation. They keep prices efficient through arbitrage or predictive modeling. Even during extreme market stress I can trade SPY in my brokerage account within a penny of its true value. If the S&P 500 index futures or the S&P 500 stocks move, someone will update the price of SPY instantaneously.

You want them competing with each other. To win they have to make the tightest price before a competitor, which either lowers their margins or requires them to bring new information to the market sooner.

If you accept that markets require intermediaries to bridge liquidity gaps in time, place, and product, then high speed traders are far less wasteful than the firms they replaced. Consider this: In 2000 Goldman Sachs bought Spear, Leeds & Kellogg (SLK), a large NYSE dealer. SLK employed 2500 people and earned over a billion dollars annually. That's one firm on one exchange. Each market had thousands of men in jackets yelling at each other, making a much bigger spread on transactions, and giving less accurate pricing.

Today there are a couple thousand people involved in low latency trading across all markets/firms and the entire industry makes a few billion dollars a year. Teams with a handful of quant researchers make markets in every listed stock globally.

[prklmn]

How much of that is simply splitting pennies and front running slower traders? Is that truely creating value? Seems like they’re siphoning value to me.

[mrchicity]

Not sure what you mean by splitting pennies, earning the bid/offer spread? That's common but the person who paid the spread to them got to trade instantly instead of waiting for another natural buyer or seller. The spread is payment for providing a service, just like an insurance company earns the spread between what you pay in premiums and their expected losses on claims. High speed market makers compete with each other so the spreads they end up earning are tiny.

The majority of high speed traders are trading their own account and don't have customer orders to front run. If you prefer a loose definition of front running to mean something like "reacting to changes in the market faster than others" then yes, but I don't see anything problematic with using public data to make your prices more accurate faster.

If you believe markets operate well without intermediaries, there are block crossing services where institutional investors can try to match up with one another. I'm sure institutions would prefer trading that way, but it turns out finding someone trying to do the exact opposite trade at exactly the same time is very difficult--volume transacted on these systems is small.