Earlier this week, US Commodity Futures Trading Commission (CFTC) gathered several wise men to discuss how many angels could dance on the head of a pin. Actually, during the meeting of the regulator's Technology Advisory Committee, the discussion turned to a problem that was, amazingly, similar—how to define high-frequency trading.
Most were there to respond the CFTC commissioner Scott O'Malia's request to help the regulator develop a solid definition that extends beyond “I know it when I see it.”
I've always found it silly to attempt to create objective definitions for things that by definition are subjective, such as high-frequency trading and even best execution. Depending on your perspective, it is difficult to know what is “high-frequency” and what is “best.” The last-place finisher in a 100-meter dash might not be considered fast by Olympic standards, but I know he is certainly faster than I am.
Coming up with an objective definition for high-frequency trading will require selecting arbitrary thresholds. However, those thresholds need to be logical and defensible. In any high-frequency trading conversation the same terms always come up—latency, automation, proximity, aggressiveness and other related words.
As the CFTC's chief economist Andrei Kirilenko mentioned during the committee's discussion, measuring and collecting data on these terms would prove exceedingly complicated.
Kirilenko and his colleagues researched 15,000 trading accounts, examined each account's intraday and end-of-day trading inventory and found that only 16 accounts behaved in what they would term a “high-frequency” manner. In fact, they saw these accounts acting more as "appendages of the matching engine" by providing additional order-matching capabilities. Without them, says Kirilenko, there would be significantly more resting orders on the exchange's order book.
One suggestion, which I believe has merit, comes from Richard Gorelick, CEO of private trading firm RMG Advisors. Gorelick says regulators should avoid establishing arbitrary trading thresholds and look to regulate trades coming from automated trading systems. The data on which orders come from automated trading systems and who runs them is already being collected by exchanges and could be accessed easily by the CFTC if necessary.
No matter what the final definition is, the CFTC will need to make sure that the metrics it relies on are easily measured and collected. But knowing how regulators have worked in the past in developing definitions and standards, I'm not holding my breath.