FRB Staff Working Paper on High-Frequency Trading ("HFT") (with Lofchie Comment)
The Board of Governors of the Federal Reserve System ("FRB") published a Finance and Economics Discussion Series ("FEDS") working paper titled "Machines vs. Machines: High Frequency Trading and Hard Information," written by Yesol Huh.
The paper focuses on whether, and in what circumstances, high-frequency trading ("HFT") provides liquidity to the market, arguing that HFT liquidity-takers' use of machine-readable information causes an information asymmetry, which becomes more severe when markets are volatile. Using a statistical approach to measure HFT activity, the author argues that the greater the information asymmetry, the lesser the liquidity that is provided by HFTs to stocks that suffer the most from this information imbalance.
Additionally, the author discusses the potential implications for HFT regulations and for market-making activity in times of market stress.
Lofchie Comment: The statistical analysis in this article is complicated, but the conclusion is relatively straightforward: in some circumstances, high-frequency trading, or at least some types of high-frequency trading, may result in a reduction in the amount of liquidity in the market. This type of "negative" finding seems to raise a reflex reaction in today's environment: let's impose regulation. Since the world is not perfect, high-frequency trading programs do not furnish the mathematically perfect amount of liquidity in every market. The "problem" discussed in the paper, to the extent that it is a problem at all, is not deserving of urgent attention, nor does it rise to a level that calls for an attempt to regulate it away. Regulatory resources are limited and there are quite enough other material issues that demand attention. That said, this is a serious paper and several points made in it are worth particular attention. First, the battle between man and machine as to how trading takes place is not over. Second, it is far from clear that adopting regulations that put a speed limit on machines is the answer (and why would that do any good unless all machines acted at exactly the same speed)? Third, while HFT may withdraw liquidity from markets during volatile periods, human market-makers would do exactly the same thing (for those who remember, that was the case with the 1987 market crash). Regulators should attend to one aspect of the study in particular: its use of simulations. Historically, when the securities regulators have attempted to adopt rules governing market structure, they simply guessed at the effect of those changes (and their guesses don't appear to have been very accurate). The use of simulations could allow regulators to game-play the effects of different combinations of theoretical rules and perhaps give them a better chance of adopting rules that would produce anticipated and desired results.
See: Machines vs. Machines: High Frequency Trading and Hard Information.