I would like to start by thanking DMO staff for their tireless work on this rule.  While the Risk Principles are short, that is not reflective of the work that has been done by staff to produce them.  This is the same DMO staff that worked on the much broader “Regulation AT”,[1] and I appreciate all of their work over many years.

Last June, I stated in my dissent to the Electronic Trading Risk Principles proposal[2] that I strongly support thoughtful and meaningful policy that addresses the ever-increasing use of automated systems in our markets.[3]  The proposal regarding Electronic Trading Risk Principles did not achieve this.  Far from utilizing over a decade of experiences that should have profoundly shaped how we address operational risks that are consistently unpredictable and have wide-ranging impacts, today’s final rule changes only a single word from the proposal aimed at codifying the status quo.  Accordingly, I respectfully dissent.

A little over ten years ago, on May 6, 2010, the Flash Crash shook our markets.[4]  The prices of many U.S.-based equity products, including stock index futures, experienced an extraordinarily rapid decline and recovery. In 2012, Knight Capital, a securities trading firm, suffered losses of more than $460 million due to a trading software coding error.[5]  Other volatility events related to automated trading have followed with increasing regularity.[6]  In September and October 2019, the Eurodollar futures market experienced a significant increase in messaging.[7]  According to reports, the volume of data generated by activity in Eurodollar futures increased tenfold.[8]  A lesson of these events is that under stressed market conditions, automated execution of a large sell order can trigger extreme price movements, and the interplay between automated execution programs and algorithmic trading strategies can quickly result in disorderly markets.[9] 

Recent events further amplify that in increasingly interconnected markets, which are informed by growing access to real-time data and information, we do not always know how and where the next market stress event will materialize.  This past April 20, the May contract for the West Texas Intermediate Light Sweet Crude Oil futures contract (the “WTI Contract”) on the New York Mercantile Exchange settled at a price of -$37.63 per barrel. The May Contract’s April 20 negative settlement price was the first time the WTI Contract traded at a negative price since being listed for trading 37 years ago. 

While the unusual fact that the price went significantly negative grabbed the headlines, the precipitousness of the price move was every bit as significant.  The price dropped more than $39 between 2:10 and 2:30 p.m. on April 20.  Overall, the price dropped $58.05 from the open of trading to its low on April 20, breaking its historical relationship with other petroleum-based contracts including the Brent Crude futures contract.  The WTI price moved more in 20 minutes than it does most years.  A contract that had never experienced a 10% move in a single day fell by more than 300% in a brief 20-minute period.  All of the contributing factors have yet to be accounted for, but one thing is certain—these were stressed market conditions.  An already oversupplied global crude oil market was hit with an unprecedented reduction in demand caused by the COVID-19 pandemic.[10]  Under stressed market conditions, automated trading has the potential to quickly make an already volatile situation even worse.

Technology glitches have continued to impact our markets.  Just yesterday, a large retail broker that was significantly impacted by the events of April 20 suffered a significant failure in data storage.[11]  Recent technology glitches overseas have hampered our international colleagues as well, handcuffing markets for extended periods of time without clear explanation.  In Japan this past September, the Tokyo Stock Exchange shut down for a day due to technical glitches in equities trading.[12]  Luckily, this glitch happened to coincide with all other Asian markets being closed and occurred the day after the first Presidential debate.  But this only emphasizes the outsized impact that a technical issue could have during volatile market conditions.  One can imagine what would have happened if the glitch had occurred the day before, during the leadup to the debate.[13] 

Just last month, Australia’s stock exchange lost an entire day of trading due to a software problem impacting trading of multiple securities in a single order.[14]  This discrete issue was enough to lead to inaccurate market data that necessitated shutting down the exchange for an entire trading day.[15] 

As we consider today’s final rule, there is a tendency to think that something is better than nothing, and that today’s risk principles—if nothing else—demonstrate the Commission’s belief that mitigating automated trading risk is important.  However, I continue to question whether these Risk Principles improve upon the status quo, or even do anything of marginal substance relative to the status quo.[16] 

The preamble seems to go to great lengths to make it clear that the Commission is not asking DCMs to do anything.  The preamble states at the very outset that the “Commission believes that DCMs are addressing most, if not all, of the electronic trading risks currently presented to their trading platforms.”[17]  The preamble presents each of the three Risk Principles as “new”, but then goes on to describe all of the actions already taken by DCMs that meet the principles.  If the appropriate structures are in place, and we have dutifully conducted our DCM rule enforcement reviews and have found neither deficiencies nor areas for improvement, then is the exercise before us today anything more than creating a box that will automatically be checked?  

The only potentially new aspect of these Risk Principles is that the preamble suggests different application in the future, as circumstances change.  As I said in regard to the proposal, the Commission seems to want it both ways:  we want to reassure DCMs that what they do now is enough, but at the same time the new risk principles potentially provide a blank check for the Commission to apply them differently in the future.[18] 

We do not know what the next external event to stress market conditions will be, but one likely possibility is climate change.  In establishing new rules for automated trading, I would have liked the Commission to have taken a more fulsome look at both the events of April 20, the COVID-19 pandemic more broadly, and the potential impacts of climate change on our automated markets.  The recently published Interim Staff Report on the events of April 20 provides a stark example of what can happen to automated markets under times of economic stress. 

The April 20 price plummet triggered both dynamic circuit breakers and velocity logic—exactly the type of risk controls discussed in the proposal that preceded the Electronic Trading Risk Principles proposal, commonly referred to as “Regulation AT.”  Regulation AT was formally withdrawn at the Chairman’s direction and without my support.  Further troubling, it was withdrawn before Commission staff had any meaningful opportunity to consider whether and how the risk controls in either Regulation AT or the Electronic Trading Risk Principles as proposed performed during trading around April 20. There was arguably no better test case, and yet we charged forward without looking back.  If the risk controls were effective, we should consider whether more specific risk controls along these lines should be part of the Electronic Trading Risk Principles, in order to be certain that all DCMs are prepared to maintain orderly trading during such a confluence of events.  If they are not, we should consider whether stronger risk controls are necessary.

I also think the Risk Principles would be improved if they were informed by a consideration of the possible impacts of climate change. The preamble states “The principles-based approach provides DCMs with flexibility to address risks to markets as they evolve, including any idiosyncratic events.”  Referring to events such as climate change as “idiosyncratic” downplays their impact and places regulators and DCMs in a purely reactive posture.  While we cannot know for certain what the next external event that causes stressed market conditions will be, that does not mean that we should remain idle until it hits.  As we will continue to experience unanticipated and unprecedented events that will impact our markets and the larger U.S. economy, I am concerned that a policy of simply checking a box will do nothing more than shield DCMs from public scrutiny and fault for the fallout. 

So often we hear that the markets have evolved from a technological and innovative standpoint at an exponential rate as compared to their regulators.  Rulemakings like this provide our greatest opportunity to proactively close that gap.  We need to be proactive.  Being proactive means studying the incidents of the past, like the Flash Crash, Knight Capital, and most recently April 20 so that we can recognize the precursors of events to come.  Instead of just reacting, we can predict, prepare for, and possibly prevent the next crisis event.

Again, while there is a temptation to advance this rule under the theory that something is better than nothing, in this case I do not think that the final rules add anything at all beyond the opportunity to take a victory lap.  In other words, the theme in this case is that nothing is better than something.  I believe that we can, and should, do better.  Therefore, I cannot support today’s final rule. 

[1] Regulation Automated Trading, Proposed Rule, 80 FR 78824 (Dec. 17, 2015); Supplemental Regulation AT NPRM, 81 FR 85334 (Nov. 25, 2016).

[2] Rostin Behnam, Commissioner, CFTC, Dissenting Statement of Commissioner Rostin Behnam Regarding Electronic Trading Risk Principles (June 25, 2020), https://www.cftc.gov/PressRoom/SpeechesTestimony/behnamstatement062520b. 

[3] The Commission’s Office of the Chief Economist has found that over 96 percent of all on-exchange futures trading occurred on DCMs’ electronic trading platforms.  Haynes, Richard & Roberts, John S., “Automated Trading in Futures Markets – Update #2” at 8 (Mar. 26, 2019), available at https://www.cftc.gov/sites/default/files/2019-04/ATS_2yr_Update_Final_2018_ada.pdf.

[4] See Findings Regarding the Market Events of May 6, 2010, Report of the Staffs of the CFTC and SEF to the Joint Advisory Committee on Emerging Regulatory Issues (Sept. 30, 2010), available at http://www.cftc.gov/ucm/groups/public/@otherif/documents/ifdocs/staff-findings050610.pdf.

[5] See SEC Press Release No. 2013-222, “SEC Charges Knight Capital With Violations of Market Access Rule” (Oct. 16, 2013), available at http://www.sec.gov/News/PressRelease/Detail/PressRelease/1370539879795.

[6] For a list of volatility events between 2014 and 2017, see the International Organization of Securities Commissions (IOSCO) March 2018 Consultant Report on Mechanisms Used by Trading Venues to Manage Extreme Volatility and Preserve Orderly Trading (IOSCO Report), at 3, available at https://www.iosco.org/library/pubdocs/pdf/IOSCOPD607.pdf.

[7] See Osipovich, Alexander, “Futures Exchange Reins in Runaway Trading Algorithms,” Wall Street Journal (Oct. 29, 2019), available at https://www.wsj.com/articles/futures-exchange-reins-in-runaway-trading-algorithms-11572377375.

[8] Id.

[9] Id. at 6.

[10] Interim Staff Report, Trading in NYMEX WTI Crude Oil Futures Contract Leading Up to, on, and around April 20, 2020 (Nov. 23, 2020), https://www.cftc.gov/PressRoom/PressReleases/8315-20.

[11] See Platt and Stafford, “Trading Outages Strike Again for US Retail Brokers,” Financial Times (Dec. 7, 2020), available at https://www.ft.com/content/cb99dc6f-a73e-41af-91fb-21a4aa606265.

[12] See Dooley, Ben, “Tokyo Stock Market Halts Trading for a Day, Citing Glitch,” The New York Times (Sep. 30, 2020), available at https://www.nytimes.com/2020/09/30/business/tokyo-stock-market-glitch.html.

[13] Id.

[14] See “Software Glitch Halts Trading on Australia’s Stock Exchange, to Reopen Tuesday,” Reuters (Nov. 15, 2020), available at https://www.reuters.com/article/us-asx-trading/software-glitch-halts-trading-on-australias-stock-exchange-to-reopen-tuesday-idUSKBN27W020.

[15] Id.

[16] See Behnam, supra note 2.

[17] Final Rule at 4.

[18] See Behnam, supra note 2.