Note: This blog originally appeared on allaboutalpha.com
A paper written late last year by Adam Clark-Joseph may have lit a (longish) fuse, and may yet set off debate over transparency in the high-frequency trading world, and the role of regulators in protecting (or spilling) trade secrets.
Were I disposed to a melodramatic turn of phrase or were I writing for a blog [oh, wait …], I would suggest that there is the possibility of a ‘wikileaks’ type of debate here, and that circumstances may have cast Andrei Kirilenko of the Massachusetts Institute of Technology into an unlikely role: the Julian Assange of account-labeled, time-stamped trading data. But that would be going too far. How much too far, we may see soon enough.
Exploratory trading by A/HFTs
But we begin with Adam Clark-Joseph. In late 2012 Clark-Joseph, of Harvard, was at work on a paper on “exploratory trading” that focused especially on the e-mini S&P futures market. Clark-Joseph found that algorithmic high frequency traders in this market consistently lose money on their “smallest aggressive orders.” But the loss is for them a happy loss, a felix culpa.
They are happy to lose that money because these small orders are “exploratory trading,” giving these traders (giving their computers really) “valuable private information” about supply and demand: the information is private in that it is not available to anyone not engaged in the high-tech HFT arms spiral. Further, Clark-Joseph wrote that this was a good starting point to address “social welfare implications of high-frequency trading.”
Fascinating as it is, Clark-Joseph’s conclusion isn’t the point of this post. The response to the data of which he made use: that’s the point.
Academic papers often circulate for some time before they are ‘published’ in any formal sense. Thus, it is unsurprising that on Dec.14, 2012, a month before Clark-Joseph’s paper was published, lawyers for Skadden Arps, one of the most august of Wall Street’s law firms, knew of it and included it as one of the causes for concern that inspired them to write a letter to the Commodity Futures Trading Commission (CFTC).
What Section 8 Says
Skadden Arps represents the CME Group, and it was the CME’s global system that generated the data of which Clark-Joseph made use. In the Dec. 14 letter, Mark D. Young and Jerrold E. Salzman of Skadden Arps, said that their client routinely provides to the CFTC such material as the global e-Mini S&P message traffic Clark-Joseph reviewed. They refer to this as Section 8 data because it is section 8 of the CEA that bars the CFTC from publishing “data and information that would separately disclose the business transactions or market positions of any person and trade secrets or names of customers….”
It isn’t at all clear that the Clark-Joseph paper discloses any person’s trade secrets, market positions, etc. He made use of the data in aggregated form, as you can see for example from Table 2 of his paper, as posted above, designed to make the point that algorithmic high frequency traders’ losses on their small aggressive orders are “non-negligible,” and too large to render plausible the claim that the point of continuing to accept such losses is a matter of inventory management.
Indeed, if the CFTC itself had posted this paper in its own name it seems unlikely that there would have been much to object to in that action, since the above quoted section 8 also allows the CFTC to publish “from time to time the results of any such investigation and such general statistical information gathered therefrom as it deems of interest to the public” so long as it doesn’t violate certain provisos such as disclosing trade secrets, etc.
Thus, the Skadden Arps letter makes the point that the section 8 data is showing up in “non-Commission sponsored publications,” and it contends that this is what violates the law.
The lawyers quote a footnote in which Clarke-Joseph thanks seven sources for “their assistance with the empirical component of this paper.” One of those seven was Andrei Kirilenko, who at the time of the letter was leaving his position with the CFTC as its chief economist. At the end of the year he took a position at the Massachusetts Institute of Technology.
The footnoted gratitude to Kirilenko caused the lawyers’ ears to prick up because they had previously encountered an article “co-authored by… Kirilenko… that specifically disclaimed status as a Commission publication” that also in their view made use of confidential data. So they charge that the chief economist was both using Section 8 data for his own publications and providing such data to non-CFTC economists for purposes of their academic research. Or, if I may mirror their own caution, they say that this “appears” to be the case.
This matter has arisen for discussion again because the Skadden Arps document, itself originally marked confidential, has only quite recently become a public document through the operation of the Freedom of Information Act. Nanex has posted it on their website.
It makes you go … hmmm. I haven’t heard back from either Dr. Kirilenko or Atty. Young at Skadden Arps yet (if either of you happen to read this, give me a call, or bounce back my email with a few words.) But since this is a blog, I will share my reactions frankly:
1) Public discussion and understanding of the issues that Clarke-Joseph and others have raised is of great importance for the future of market infrastructure, and indeed of the economy as a whole;
2) If he is wrong in worrying about exploratory trading, he should be shown to be wrong with the use of facts and reason, he shouldn’t be shushed;
3) In the absence of any specific disclosure of particular market positions, strategies, customer lists or the like, I don’t see any conflict between market transparency and the preservation of actual confidences – the two goals are in harmony.
4) If Kirilenko played the role the lawyers at Skadden Arps suspect: good for him.