Dark Pools is an informative, quick read about high frequency trading, it’s origins, evolution, and current state. The text has a clear bias against HFT, however, it does bring a few interesting points to light.
It provides a great general overview of how HFT arose, from the origins of Datek to the ECNs like Island, Instinet, and Archipelago. The ECNs arose due to frustrations with the exchange floor specialists and brokers who could favor certain orders and were sometimes inefficient and slow. It was essentially a “boys club” that was unfair. Yet with the ECNs, what originally started as a quest to make markets more efficient eventually became an electronic variation of what existed in the past, an unfair playing field that only catered to those with the fastest speeds and closest server positions at the expense of the everyday, 401k and mom and pop investors.
Towards the end, Patterson touches on machine learning AI strategies that take in vast quantities of data and information from around the world to uncover actionable correlations. Using computing power to sift through fundamental information to make trades in the market makes the market more informative and efficient. Some HFT strategies take advantage of structural inefficiencies that exist in the market. While these are in the gray area, I believe these ultimately make the market better because these inefficiencies will eventually get discovered and fixed. Until then, if others make lots of $$ off these inefficiencies, then so be it, it’s a problem with the market structure.
I believe the problem is that these market structure inefficiencies aren’t discovered fast enough and will always exist within the system, simply because the regulators will always be a step behind. I do believe it’s a shame that many of the brightest people in our society spend their entire careers creating and tweaking algorithms to solve an artificial and ultimately, inconsequential problem of making trading decisions and executions fractions of a second faster. Yet, this goes back to my underlying belief that all of our actions are driven by incentives. Unfortunately, the monetary incentives in our society are such that this is the end result.
HFT as a whole though has declined significantly since the late 2000s and early 2010s. As competition ramped up, margins became smaller and as profits shrank, many firms closed shop. While HFT dominated daily exchange volumes, they’ve come down substantially. It seems like machine learning and AI strategies are the future: computers that process millions if not billions of fundamental data points and are able to learn on their own and make investment decisions. However, I believe there will always be a place for human traders in the market.
There’s a great book on market microstructure called “Trading and Exchanges” by Larry Harris. However, the first and only edition was published in 2002, it would be great if he released an updated version of his book. I’d also be interested to hear what George Soros thinks of HFT and how it fits into his Theory of Reflexivity in the markets.
JL Nexus 