After declining multiple requests to extend the comment period that expired on March 16, the Commodity Futures Trading Commission received approximately 50 written responses to its proposed Regulation Automated Trading. A wide range of responders, including trade organizations, vendors, market participants, public interest groups and individuals, wrote comments.

Although responders generally approved of the CFTC’s objective of ensuring robust controls, transparency measures and other safeguards to mitigate risks arising from algorithmic trading on designated contract markets, the majority of comment letters criticized the proposed rules as too prescriptive.

The most passionate specific objections were raised regarding proposed requirements mandating that certain persons engaged in algorithmic trading who directly access DCMs be registered as floor traders, and that all so-called AT Persons maintain source code in a distinct repository subject to inspection by CFTC and Department of Justice staff. (AT Persons, as defined by the CFTC’s proposed regulation, would include most current registrants, including commodity trading advisors and commodity pool operators, who engage in algorithmic trading on or subject to a DCM’s rules.) Those raising these concerns claimed that registration requirements were unnecessary and that it was untenable that source code could be made so easily available, thereby potentially exposing the lifeblood of a trading business to ruinous inadvertent disclosure to third parties, including competitors.

Regulation AT was proposed by the CFTC November 24, 2015. (Click here for a comprehensive overview of Regulation AT in the article, “CFTC’s Proposed New Algorithmic Trading Rules Augur Potential Increased Obligations and Costs, and a New Registration Requirement” in the November 29, 2015 edition of Between Bridges.)

The two longest comment letters in response to proposed Regulation AT were submitted by FIA and CME Group.

FIA argued, among other things, “that all electronic trading should be subject to pre-trade and other risk controls appropriate to the nature of the activity” and not just the algorithmic trading of the limited universe of persons captured by the CFTC’s proposed regulation. This does not mean, said FIA, that every market participant should be required to implement his or her own risk measures. However, all electronic orders should be subject to “appropriate controls,” said the industry organization. Futures commission merchants or DCMs could provide these controls, wrote FIA.

FIA urged the CFTC to consider breaking apart its rule-making into three components. One would address pre-trade and other risk controls that the industry organization argued could likely be implemented most quickly (although any requirements should be more principles-based). The other two components would address (1) policies and procedures related to the development, testing and deployment of algorithmic trading and (2) registration, if necessary. In general, FIA argued that the CFTC’s proposals were too prescriptive and should be more principles-based.

FIA also objected to Regulation AT’s proposed requirement that all AT Persons use enhanced self-trade prevention measures, claiming that the incidence of problematic self-matching today is insignificant.  

In addition, FIA said that Regulation AT’s proposal requiring annual reports to be prepared by AT Persons and clearing FCMs and filed with DCMs were very burdensome, and for FCMs, redundant of annual reports they already are required to prepare. Moreover, argued FIA, DCMs would be “unable to meaningfully process and analyze the wide variety of policies and procedures related to the development and compliance of Algorithmic Trading Systems” that would be addressed in such reports.

CME Group questioned the fundamental benefit of Regulation AT. According to CME Group,

we believe that much of the well-intentioned Proposal is unwarranted and could be unnecessary … In our view, the Proposal does not provide a clear justification for why additional federal regulation is necessary or appropriate and how the new rules would fulfill [the CFTC’s] objectives in a workable manner.

CME Group also said that, despite being promoted as principles-based, the proposed Regulation AT was very prescriptive, requiring “specific performance approximately 87 times.” In general, CME Group argued that current requirements and practices afforded markets sufficient protection, and additional requirements proposed by the CFTC would be very expensive to implement and maintain.

The Managed Futures Association argued that CTAs and CPOs should be excluded from the “focus and scope” of Regulation AT. As an alternative, MFA recommended that the CFTC

...direct the National Futures Association to promulgate CTA/CPO regulatory requirements on operational system risk controls … If necessary, the Commission should amend its Part 4 regulations to address CTA/CPO operational systems risk concerns related to algorithmic trading in a manner suitable for CTA and CPO businesses and operations.

The Commodity Markets Council objected to the CFTC’s proposed definition of algorithmic trading that it implied could potentially impose many of the requirements of Regulation AT on commercial end users if they were deemed to engage in direct electronic access. This is because the definition captures “basic” order management functionality used by commercial firms “to implement trading decisions and parameters by which [a] specific order type is executed, which are determined by a natural person.” These functionalities, said CMC, include auto-spreaders to help commercial firms “create and trade synthetic calendar, inter-product and inter-exchange spreads;” iceberg or reserved quantity orders to help a firm bring “significant hidden liquidity” to a market by execution of a number of smaller orders; and trailing limits orders that constantly re-price certain orders at a fixed interval from the then prevailing market price. CMC expressly requested that the CFTC “categorically exclude” from the definition of algorithmic trading order management functionality (as opposed to order generation systems).

Trading Technologies also questioned the proposed definition of DEA, saying that it ignored many of the ways traders routed orders to DCMs, and an “FCM’s complete control over user set-up and risk management tools” in connection with many of these means. TT also argued that the utilization by the CFTC of its proposed rules regarding source code would “likely constitut[e] an unconstitutional taking of individuals’ property and is generally unnecessary to achieve the goal of the proposed regulations.”

Contrariwise, some commentators were generally supportive of proposed Regulation AT or thought it did not go far enough. According to Better Markets, for example,

The requirement for registration of Automated Trading firms is common sense and straight forward, and furthermore we are pleased to see the focus on several important safeguards that we have advocated in the past, such as the message throttles and the Commission access to algorithm source code.

Americans for Financial Reform similarly argued that “the self-regulatory approach taken [by the CFTC] falls far short of limits on the most dangerous and predatory practices made possible by automated trading technology.”

However, the Intercontinental Exchange cautioned that if not drafted and implemented correctly, Regulation AT could hurt and not help markets. A badly drafted Regulation AT, said ICE, “has the potential to disrupt currently effective risk management safeguards and may stymie the development of new and innovative methods of risk management.”

In a speech last week before the FIA’s annual International Futures Industry Conference in Boca Raton, Florida, CFTC Chairman Timothy Massad expressed his “hope” that the Commission would finalize Regulation AT by year-end. He promised that any final rule would respect and protect the confidentiality of source code “while at the same time ensuring that source code is preserved and is available to us when we need to reconstruct market events.”

My View: It was an error for the CFTC not to extend the comment period for Regulation AT. The proposed rule represents one of the most important and complex initiatives by the CFTC in recent years and it was wrong for the Commission not to provide additional time for more careful and detailed responses.

Although the CFTC’s objectives of mitigating market risk attributable to algorithmic trading and increasing transparency are laudable, the proposed rule mostly federalizes existing requirements of DCMs and the National Futures Association while adding many new prescriptive requirements that do not easily accord with existing market structure and practices (let alone where market structure and practices may evolve).

Like Charles Dickens’ observation in A Tale of Two Cities that “it was the best of times, it was the worst of times,” Regulation AT is both too overreaching and not reaching enough. It proposes too many onerous obligations on too many market participants solely because they are a registrant, but it fails to recognize that risk controls and other safeguards should apply to all electronic trading, regardless of whether the trader is registered or not.

Because the CFTC did not grant an extension for views, commentators could not prepare adequate cost and benefit analyses. However, as suggested by some commentators, the likely costs of Regulation AT are substantially greater than estimated by the CFTC.

If the CFTC believes there is a gap in existing regulations and wants to formalize and enhance existing industry safeguards around algorithmic trading systems, the better course would be for it solely to amend existing core principles for DCMs to require them to expressly impose appropriate requirements on members and non-members around such systems to protect their markets and ensure robust audit trails of orders. If the CFTC believes it needs to ensure better controls by gatekeepers of traders to DCMs (e.g., FCMs), it could consider principles-based amendments to existing risk-management rules.

In any case, the scope and complexity of Regulation AT, as proposed, is too great to be implemented as is. Although staff of the CFTC is to be commended for yeoman’s work in endeavoring to craft a simple but comprehensive set of rules to implement the Commission’s valid objectives, in the end, Regulation AT endeavors to accomplish too much and too little at the same time!