November 24, 2015
Thank you. It’s a pleasure to be here this morning to consider the proposed rule on automated trading. I want to thank the Commission staff for the time they have devoted on this proposal. It is a timely topic.
As I have previously said, our markets have seen immense technological change over the last fifteen years.1 Futures trading used to involve “throngs of traders with jackets and badges using hand-gestures” to purchase futures and options.2 That trading structure has largely disappeared, with even CME closing the vast majority of its futures pits this summer. Meanwhile, algorithmic trading has substantially increased. Algo trading comprised less than 10% of futures volume at the turn of the millennium.3 Yet, “per CFTC staff’s estimates, for the most liquid U.S. futures contracts which account for over 75% of total trading volume, more than 90 percent of all trades make use of algorithms or some other form of automation.”4 Of course, these estimates are just that, estimates. We still do not have comprehensive, precise data on the percentage of trades created or entered by algorithms in many product classes. Clearly, further research and work remain for all stakeholders, from regulators, to industry participants, to academics and advocates of financial reform.
Yet, I do not believe this lack of information requires that regulators passively wait for this information to emerge. Simply waiting for that kind of data to materialize could allow problems to emerge in the interim that harm investors and the broader financial system. Given the current state of our economy and a global financial system still recovering from the 2008 financial crisis, that is a risk that I believe we must not take. Recent events have raised important questions about the impact and role of algorithmic trading in our markets. As I said earlier, this fall, “Even though the amount of algorithmic trading and definitions of these various terms are not crystal clear, what is clear is trades involving algorithms make up a substantial portion of our markets, and algorithms can and do malfunction at times, with negative effects on the markets. As a result, I believe we are obligated to consider if it is prudent to establish some regulations on algorithmic trading in our markets.”5
Today, we begin the process of potentially establishing those regulations. From what I have seen, I believe we now owe it to market participants, investors, and ordinary consumers to ensure that a reasonable level of regulation exists over this new trading technology. I have said such regulation should include requirements that entities utilizing algorithms to trade must use risk management strategies, be required to disclose information to regulators, and have people who understand the Commodity Exchange Act and our regulations involved in the creation and maintenance of their algorithms. I think this proposed regulation meets that standard and does so in a way that allows for innovation and continued development of this nascent technology.
Having said what I think lies at the core of this regulation, let me also be clear about what this regulation is not. The rule before us today should not substantially change how many firms utilize algorithms. If, as I hope, a firm already uses risk management strategies, has various protections against malfunctions in place, and retains the services of talented attorneys, this new regulation will not create significant new burdens for that firm. In effect, this rulemaking largely formalizes and mandates firms involved in algorithmic trading to engage in a variety of practices that they should already be doing for their own protection.
I expect that some observers will have issues with this regulation for not doing more to constrain the growth and use of algorithmic trading, and I expect there will be further debate. I do not regard this regulation as the final word on regulation of algorithmic trading. If there is clear evidence that more precise regulations are needed on this technology to protect investors or ward off systemic risk, I would support further regulatory action. And I am sure that, given the ferocious rate of change of this technology, we will need to update this regulation regularly to account for those changes. In many ways, this regulation is merely the first step in a process, it’s a starter home rather than a two-story. But we have to start somewhere, and starting with something that formalizes best practices and increases disclosure is an excellent place to start.
I have said numerous times that I support smart regulation, regulation that works. That goal is especially critical when it comes to regulation of such a nascent, significant, and widespread technology as algorithmic trading. I therefore hope we’ll get comments on this proposal from a wide swath of stakeholders, from industry experts, to end-users being impacted by this technology, to even ordinary investors and consumers concerned about the potential effects of algorithmic trading on commodity prices. I do not expect that everyone will have the same views on this subject or that there will be unanimity of opinion on any part of this rule. Even though I’ve only been in Washington for a year and a half, I’m experienced enough to know that people have different opinions on high-visibility issues like this one. However, I do encourage people to comment so that we can get a full and fair read of popular opinion on both this proposal and the topic in general. And if people have concrete evidence that algorithmic trading is distorting markets and needs to be curtailed, please submit it via a comment.
There are a few sections of this rule on which I think public comment would be particularly helpful. First, the proposal’s sixth and seventh questions ask about the nature of our proposed definition of algorithmic trading, including whether we should expand “the definition of Algorithmic Trading to encompass orders that are generated using algorithmic methods . . . but are then manually entered into a front-end system by a natural person….” The definition of algorithmic trading is at the heart of this proposal, and we need comments on this point. If there is evidence that a form of algorithmic trading poses systemic risks but is not captured by this definition, we should expand the definition to expand to cover that form of trading.
Second, section 1.83(a) of the proposal requires that persons engaged in algorithmic trading and registered as such with the Commission must prepare and submit an annual report to the Commission. These persons are required to include in their reports a description of the pre-trade risk controls in place, copies of policies crafted to comply with requirements regarding the testing and development of algorithmic trading systems and how their algorithmic trading systems comply with the Commodity Exchange Act and our regulations, and a certification by their chief executive officer or chief compliance officer that the information in the report is accurate and complete.
I think the current 1.83(a) does not ask registrants for enough information. Now, we don’t want to require each registered algorithmic trader to submit a tome of several thousand pages each year that lays out every arcane factoid about their trading systems. Such a requirement would bury our staff in paper and create significant expense for registrants. Yet, having already asked each registered algorithmic trader to submit an annual report, I believe we should ask for more information in the report. After all, at the point a company has to file an annual report, it should already be doing a comprehensive review of its policies. As a result, asking for one or two more pieces of information to be included in the annual report should not be a substantial additional cost to registrants. I therefore hope that commenters will let us know what additional information registrants should be required to submit in their annual reports. For instance, should we require registrants to submit information about how they train and monitor the staff responsible for handling algorithmic trading or about their order cancellation systems?
Finally, the proposal prohibits designated contract markets (DCMs) from paying market maker incentive program benefits for trades between accounts under common ownership. I think that’s a good change and worthy of being formalized in rule text. These programs serve a critical purpose of encouraging liquidity, but we don’t get increased liquidity by increasing the amount of trades a person does with herself.
However, I wonder whether this prohibition should not go further. Perhaps we should also prohibit DCMs from paying these program benefits for trades in which the benefits are, on a per trade basis, greater than the fees charged by the relevant DCM and affiliated derivatives clearing organization (DCO). Paying benefits for such trades seems tantamount to giving a subsidy to un-economic trades and thereby potentially risks distorting the overall market. I would therefore welcome comments about whether this section is adequate as is or whether we should also prohibit DCMs from giving benefits to such seemingly non-economic trades.
In closing, let me stress again that I want this rule to be both effective and workable. No one benefits from rules that work in the abstract but are confusing, impossible to implement as written, or full of gaps that prompt stakeholders to engage in widespread regulatory arbitrage. I believe this automated trading proposal is a commonsense effort at establishing reasonable regulation on a nascent technology, but if there are flaws with it, if it goes too far or not far enough, I want to know that now, before it is finalized. Thank you.
1 Keynote Address by Commissioner Sharon Y. Bowen before ISDA North America Conference, CFTC (Sep. 17, 2015), http://www.cftc.gov/PressRoom/SpeechesTestimony/opabowen-6.
Last Updated: November 24, 2015