1 1 THE UNITED STATES SECURITIES AND EXCHANGE COMMISSION 2 3 4 JOINT CFTC-SEC ADVISORY 5 COMMITTEE ON EMERGING 6 REGULATORY ISSUES 7 8 9 10 11 12 Monday, May 24, 2010 13 9:04 a.m. 14 15 16 17 Securities and Exchange Commission 18 100 F Street, N.E., Room L-002 19 Washington, D.C. 20 21 22 23 Diversified Reporting Services, Inc. 24 (202) 467-9200 25 2 1 C O N T E N T S 2 3 PAGE 4 Welcome and Opening Remarks 5 A. Mary L. Schapiro, Chairman, SEC 3 6 B. Gary Gensler, Chairman, CFTC 4 7 Introduction of Committee Members 6 8 Introduction of Staff 14 9 Summary of Events of May 6, 2010 18 10 Joint CFTC-SEC Staff Report 11 A. Staff Presentation of Preliminary Findings 21 12 B. Discussion 61 13 Additional Data Review and Next Steps 14 A. Staff Presentation 88 15 B. Discussion 101 16 V. Discussion and Approval of Bylaws 113 17 VI. Conclusion 120 18 19 20 21 22 23 24 25 3 1 P R O C E E D I N G S 2 CO-CHAIR SCHAPIRO: Good morning. Chairman Gensler 3 and I are pleased to welcome you and to call to order this 4 first meeting of the Joint CFTC-SEC Advisory Committee on 5 Emerging Regulatory Issues. 6 Today is May 24, 2010, and this meeting is being 7 held in accordance with the Government in the Sunshine Act. 8 All members of the Committee except one are present 9 either in person or telephonically. I believe Drs. Engle, 10 O'Hara, and Stiglitz are on the telephone, and Dr. Ruder has 11 an unavoidable conflict, although he may be able to join us 12 later. 13 Before we begin, let me make a few logistical 14 points. First, Gary and I are co-chairing the Joint 15 Committee meeting and as co-chairs we'll share responsibility 16 for leading our discussions today. So if you'd like to be 17 recognized, please just turn your tent card on end, and we'll 18 try to recognize you as soon as possible. 19 Also, for those of you who are joining us by phone, 20 if you could please make sure you have your phone on mute 21 when you're not speaking, that would be enormously helpful, 22 as we are being Webcast and interference, I think, will make 23 it difficult for many listeners and watchers. 24 And now, I want to turn it over to Chairman 25 Gensler. 4 1 CO-CHAIR GENSLER: Good morning. I would like to 2 start by thanking Chairman Schapiro for all of her efforts on 3 behalf of the investing public, as well as the staff of the 4 SEC for hosting this first meeting of our Joint Advisory 5 Committee on Emerging Regulatory Issues. 6 I also want to thank my fellow CFTC Commissioners, 7 Mike Dunn, Jill Sommers, Bart Chilton, and Scott O'Malia, for 8 all of their support in establishing this Joint Committee and 9 for all of their work on behalf of the American public. I 10 know that each of us look forward to receiving the advice of 11 this expert panel. 12 It was during the presidential transition actually 13 that Chair Schapiro and I first discussed possibly setting up 14 a Joint Advisory Committee. Little did we think it would 15 take a year and a half, a joint harmonization report, an act 16 of Congress and, yes, a 1,000 point drop in the Dow before 17 getting it done, but I'm pleased that we're here finally for 18 our first meeting of the Committee. 19 And an Advisory Committee designed to view emerging 20 risk in our financial markets is long overdue. While I am 21 not suggesting if we had had this Committee up five or ten 22 years ago we wouldn't have had the financial crisis, I do 23 believe that there were emerging risks that demanded 24 thoughtful analysis, whether they were the use of new 25 products like credit default swaps or developments in the 5 1 securitization markets or, more recently, rapid 2 advancements in technology and significant changes in market 3 structure. We can be certain that we will continue to see 4 innovations in the markets that will require thorough review. 5 I think also that the CFTC and SEC will benefit 6 greatly from hearing from an Advisory Committee together. In 7 fact, this is the first time our two agencies have ever 8 shared an advisory committee and that we have been two 9 separate agencies since the 1930s. We both exist to protect 10 the investing public and promote transparent, fair, and 11 orderly markets. 12 Our two agencies' oversight spans across future 13 securities and, hopefully shortly, the over-the-counter 14 derivatives market places. It's essential that we work 15 cooperatively to regulate these markets, and this Committee 16 can assist us in achieving that goal. 17 The CFTC and SEC will benefit from having outside 18 experts thinking about emerging risk to markets, advising us 19 on them. This morning's meeting and more generally the Joint 20 Committee is about looking around corners, looking into the 21 future about where we need to take regulation. 22 And markets do change. Market structures, 23 practices, technology, all change, and they change rapidly. 24 There's constant innovation. I hope the Committee of outside 25 investors can help us insure that regulation stays abreast of 6 1 these changes. 2 Our panel is comprised of a diverse and 3 accomplished group. I'm pleased that we have three former 4 Commission Chairmen. We also have three distinguished and 5 award winning professors of economics and finance, and a 6 former director of one of the largest asset management firms 7 in the world and head of an independent regulator. 8 The first task, of course, is we are looking at May 9 6th. Those events have significant implications for 10 investing public and the American businesses. It's essential 11 we continue our review and with the contributing 12 circumstances of price volatility, more work must be done to 13 accomplish this goal. 14 The Committee has had an opportunity to examine the 15 staff findings released. I wish to thank the staffs of both 16 the SEC and CFTC, an enormous amount of work around the May 17 6th report, and though there's much to do, the staff has 18 examined liquidity dislocations, linkages between markets and 19 the role of electronic and algorithmic trading, and today we 20 look forward to hearing from the staffs and having the 21 panelists ask questions. 22 With that, I thank you, and I turn it back to Mary. 23 CO-CHAIR SCHAPIRO: Thanks, Gary. 24 We're going to tag-team on the introduction of the 25 Committee members, but let me also acknowledge the presence 7 1 of my fellow Commissioners and thank them for their support 2 really during the last year and almost a half that I've been 3 here, but most particularly since the events of May 6th that 4 have required us to move very quickly on a number of fronts: 5 Commissioner Kathleen Casey, Elisse Walter, and 6 Troy Paredes, and I assume that Commissioner Aguilar will be 7 joining us at some point. 8 Gary, do you want to start the introductions? 9 CO-CHAIR GENSLER: If I can just find where I'm 10 supposed to turn to to do that. Ah, here we go. 11 We're pleased to have eight distinguished members 12 of the Committee, and the Chair and I extend our appreciation 13 for them taking the time. Now, if I may introduce, Chairman 14 Born, if I may call her that, is a retired partner of Arnold 15 & Porter, where she was the head of the firm's derivatives 16 practice representing domestic and international clients and 17 legislative, litigation, regulatory, and transactional 18 matters involving derivatives transactions in financial 19 markets from '96 until 1999. When we first met actually, 20 Brooksley held the position that I am now honored to hold, 21 the Chair of the CFTC. While at the CFTC, Ms. Born served as 22 a member of the President's Working Group on Financial 23 Markets and Technical Committee of IOSCO, and I believe 24 currently is also on the Financial Crisis Inquiry Committee. 25 Do we tag-team or do I keep going? 8 1 CO-CHAIR SCHAPIRO: You keep going. 2 CO-CHAIR GENSLER: I keep going. There, I learned. 3 Rick Ketchum to my left is Chairman and Chief 4 Executive Officer of FINRA. Before assuming his current 5 responsibilities, he was CEO of NYSE Regulation from 2006 to 6 2009. He served as the first chief regulatory officer of the 7 New York Stock Exchange. Mr. Ketchum was previously General 8 Counsel of the Corporate and Investment Bank at Citi Group. 9 He also spent 12 years in NASD and the NASDAQ Stock Market, 10 where he served as President of both organizations, and prior 11 to that he spent 14 years at the SEC, has a wealth of 12 regulatory and market knowledge. So we thank him for joining 13 us. 14 CO-CHAIR SCHAPIRO: To my right is Jack Brennan, 15 the Chairman of the Financial Accounting Foundation, Board of 16 Trustees, which is the independent private sector 17 organization responsible for oversight of the FASB and GASB. 18 He also is the Chairman Emeritus and senior advisor of the 19 Vanguard Group, which I believe is now the largest asset 20 manager -- mutual fund company. Sorry. 21 Mr. Brennan previously served as Vanguard's 22 Chairman and CEO and, among many other industry activities, 23 was past Chairman of the Investment Company Institute and is 24 a Governor of FINRA. 25 Professor Robert Engle, who is joining us by phone, 9 1 is the Michael Armellino Professor of Finance at New York 2 University's Leonard N. Stern School of Business. In 2003, 3 Professor Engle was honored with the Nobel Prize in economic 4 sciences for his work in methods of analyzing economic time 5 series with time varying volatility. 6 To Jack's right is Dean Susan Phillips. Susan has 7 just retired as a Dean of the George Washington University 8 School of Business. Her areas of specialization include 9 monetary policy, regulation and supervision of financial 10 institutions, derivatives, financial management, and the 11 economic theories of regulation. 12 In 1981, Dean Phillips was appointed to membership 13 on the CFTC, where I had the distinct honor and pleasure of 14 working for her, and she became its Chairman in 1983. 15 Upon leaving the CFTC in 1987, she returned to 16 academia at the University of Iowa, but then was called back 17 to Washington to serve on the Board of Governors of the 18 Federal Reserve System, and she has authored dozens of 19 scholarly publications. 20 CO-CHAIR GENSLER: The next member of the Joint 21 Committee joining us by phone is Professor Maureen O'Hara. 22 Professor O'Hara is the Robert Purcell Professor of Finance 23 at Cornell University's Johnson Graduate School of 24 Management. Professor O'Hara's research focuses on issues of 25 market micro structure, and she is the author of numerous 10 1 journal articles, as well as the book Market Micro Structure 2 Theory, which is going to be very helpful in what we're 3 looking at here. 4 Her most recent research looks at the role of 5 uncertainty in affecting the liquidity and valuation of 6 securities. 7 Next, also joining us by phone is Professor Joseph 8 Stiglitz. Professor Stiglitz is a Professor of Finance and 9 Business at Columbia University and chair of the University's 10 Committee on Global Thought. He is also the co-founder and 11 Executive Director of the Initiative for Policy Dialogue at 12 Columbia. 13 In 2001, he was awarded the Nobel Prize in 14 Economics for his analysis of markets with asymmetric 15 information. He was a lead author of a 1995 report of the 16 Intergovernmental Panel on Climate Change, which shared the 17 2007 Nobel Peace Prize, quite an accomplishment. 18 Professor Stiglitz was a member of the Council of 19 Economic Advisors from 1993 to '95, and served as its chair 20 from 1995 to '97. He has also served as Chief Economist, 21 Senior Vice President in the World Bank; has chaired both the 22 Commission of Measurement of Economic Performance and Social 23 Progress and the Commission of Experts on Reform of the 24 International Financial and Monetary System. 25 CO-CHAIR SCHAPIRO: And finally, last but certainly 11 1 not least is David Ruder, who has an unavoidable conflict, 2 out of the country today and is unable to be with us, but 3 David is the William W. Gurley Memorial Professor of Law 4 Emeritus at Northwestern University School of Law, where he 5 previously served as Dean. 6 From 1987 to 1989, Dr. Ruder served as Chairman of 7 the Securities and Exchange Commission, and he is currently 8 Chairman of the Mutual Fund Directors Forum. 9 We are, indeed, fortunate that this Committee is 10 made up of some of the brightest minds in the country, and we 11 are deeply in their debt that they have agreed to help 12 provide the SEC and the CFTC with their views and 13 perspectives on the emerging regulatory issues that affect 14 both agencies. Chairman Gensler and I both look forward to 15 benefitting from your insights and wisdom. 16 The establishment of this Joint Committee was one 17 of the 20 recommendations included in the agencies' 18 harmonization report issued last year. The mandate of this 19 Committee is broad. It's charged with identifying emerging 20 regulatory risks relating to both the SEC and the CFTC; 21 assessing and quantifying the impacts of such risks, 22 including their implications for investors and market 23 participants; and furthering the agencies' efforts on 24 regulatory harmonization. 25 The events of May 6th certainly highlight the need 12 1 to identify risks and engage in meaningful regulatory 2 cooperation. As you well know, during a 20-minute period on 3 that afternoon, the U.S. financial markets failed to execute 4 their essential price discovery function, experiencing the 5 decline and recovery that was unprecedented in its speed and 6 scope. That period of gyrating prices both directly harmed 7 investors, who traded based on flawed price discovery signals, 8 and undermined investors' faith in the integrity and the 9 fairness of our markets. 10 The events of May 6th serve to remind us once again 11 that our financial markets are inexorably linked. Events 12 that occur on markets regulated by one agency inevitably 13 affect the markets regulated by the other. Given this, 14 Chairman Gensler and I agreed that the first item on this 15 Joint Committee's agenda should be to conduct a review of 16 the market events of May 6th so that we can be sure we fully 17 understand how market structures, disparate trading 18 conventions, different rules across markets and other factors 19 contributed to the events of that day, and so that we can 20 begin to craft appropriate regulatory responses. 21 Over the last 18 days, the staffs of both agencies 22 have been hard at work conducting a comprehensive 23 investigation. Last Tuesday, CFTC and SEC staff released a 24 joint report detailing the preliminary findings of our 25 investigation. 13 1 In addition to working with the CFTC on the joint 2 report, at the SEC we have worked with the securities 3 exchanges to fashion measures to help protect against a May 4 6th recurrence by imposing a limit on the extent to which 5 certain individual stock prices can move before trading in 6 that stock is paused. These proposed circuit breakers have 7 been published for comment for ten days. 8 At this stage we continue to focus on the events 9 that may have triggered the unusual volatility, but 10 regardless of the triggering cause, we believe that the 11 initial volatility was magnified by a variety of factors. 12 Areas I expect we will probe further with the Committee's 13 input include the possible linkages between the steep decline 14 in the prices of stock index products and simultaneous and 15 subsequent waves of selling in individual securities; a 16 generalized severe mismatch in liquidity possibly exacerbated 17 by the withdrawal of liquidity by electronic market makers; 18 the possibility that this liquidity mismatch may have been 19 aggravated by disparate trading conventions among various 20 exchanges; the possible involvement of stub quotes; the use 21 of market orders and stop loss orders that might have 22 contributed to market instability; and the impact on exchange 23 traded funds which suffered a disproportionate number of 24 broken trades. 25 Even before May 6th, the Commission had launched 14 1 initiatives to strengthen the U.S. securities markets and to 2 protect investors. We had proposed rules that would prohibit 3 flash orders, increase the transparency of dark pools of 4 liquidity, prohibit broker-dealers from providing unfiltered 5 access to exchanges, and proposals to create a large trader 6 reporting system. 7 Earlier this year we also issued a concept release 8 on market structure that solicited public comments on the 9 impact of different trading strategies, including high 10 frequency trading on our markets and investors. These issues 11 will also be at the center of a market structure roundtable 12 which we're holding in a couple of seeks. 13 And on Wednesday of this week, the Commission will 14 consider a proposal to create a consolidated order tracking 15 system to allow effective cross-market surveillance. 16 Let me thank the Committee again for joining us 17 today, and I look forward along with you to the staff's 18 presentation and ensuing discussions. So now I'll turn the 19 meeting over to the staff so that the CFTC and the SEC, who I 20 just have to say have been working tirelessly over the last 21 18 days to present their preliminary findings to the Joint 22 Committee. 23 Let me introduce Robert Cook, the Director of the 24 SEC's Division of Trading and Markets, to introduce the 25 presenters. 15 1 Robert. 2 MR. COOK: Thank you, Chairman Schapiro. 3 Good morning. Before we begin our presentation, on 4 behalf of all the staff, I'd like to welcome the members of 5 the Joint Advisory Committee and to express our deep 6 appreciation to you for agreeing to offer your time, your 7 insights, your expertise to the Commission staff as we work 8 to address the issues associated with the market events of 9 May 6th and, beyond that, to further strengthen and harmonize 10 the work of our two agencies. We look forward to working 11 with you. 12 I'd also like to express our appreciation to our 13 Chairmen and to the Commissioners from both agencies and to 14 thank them for the invaluable leadership and support they 15 provided us over the last few weeks. We look forward to 16 continuing to work to strengthen our markets under your 17 steady guidance. 18 I'm joined today by some colleagues from the SEC 19 staff: David Shillman, Associate Director for the Office of 20 Market Supervision in the Division of Trading Markets; and 21 from our Division of Risk Strategy and Financial Innovation, 22 Co-Deputy Directors Jonathan Sokobin and Gregg Berman. 23 I'd like to take this opportunity to thank the many 24 staff at the SEC both here in Washington and the many 25 regional offices who have been working very hard over the 16 1 last few weeks and who will be working very hard over the 2 coming weeks on this project. There are far too many names 3 to mention, but their contributions deserve our collective 4 appreciation and acknowledgement. 5 Let me turn it over to my colleague, Steve Sherrod, 6 to introduce his staff. 7 MR. SHERROD: Thanks, Robert. 8 I want to join Robert in thanking the Advisory 9 Committee and my Commissioners and Chairman Gensler -- 10 particularly at last Friday's surveillance briefing where 11 they asked many probing questions that we continue to 12 investigate. 13 Joining me today are Andre Kirilenko from our 14 Office of Chief Economist, on my left, and Richard Shilts, 15 the Director of our Division of Market Oversight. 16 As Robert mentioned, we have had a tremendous staff 17 effort to bring all hands on deck to work on this issue over 18 the last two weeks. 19 Robert. 20 MR. COOK: So just to briefly review the agenda for 21 this morning, first Steve is going to provide an overview of 22 the general market conditions on May 6th to set the stage and 23 the context for discussing what happened in the futures and 24 securities markets that afternoon. 25 Next I, with the help of Gregg, Jonathan and David, 17 1 will summarize the preliminary findings from the securities 2 side of the markets, and then Steve will summarize the 3 preliminary findings from the futures side, and then we'll 4 have some time reserved to take your questions regarding 5 these preliminary findings. 6 And then in the second half of the morning, we will 7 talk about next steps and analysis both on the securities and 8 the futures and through joint collaboration. 9 Before we begin on the preliminary findings, I 10 think it might be worth making one caveat. The investigation 11 into these events is continuing, and the joint report that we 12 prepared for you was able to capture only our preliminary 13 findings within a relatively short period of time. 14 In addition, as we'll discuss a little bit more 15 later, the reconstruction of the relevant events is 16 complicated by the enormous amount of data involved, which is 17 being compiled on the securities side from multiple sources 18 that do not have a consistent framework for collecting, 19 formatting, and structuring data regarding quote and trade 20 activity, all of which is to say that we need to keep in mind 21 that the information we're presenting today is subject to 22 change as we validate and correct data that we have and 23 incorporate new data into our analysis. 24 So with that, let me turn it over to Steve to give 25 us some general context on the market. 18 1 MR. SHERROD: Thanks, Robert. 2 Throughout the day on Thursday, May 6th, many 3 financial news outlets were reporting on political and 4 economic events that were creating uncertainty in the 5 financial markets. As reports highlighted the perceived 6 increased risk of default of certain European countries, 7 there was downward pressure on the euro in global currency 8 markets. We particularly note the European Central Bank held 9 a press conference and did not address the possibility of 10 purchasing Greek government bonds. 11 This raised concern over defaults, as reflected in 12 an increase in the price of premiums on credit default swaps 13 to protect against such events. 14 Throughout the day we also observed a broad decline 15 in the U.S. equity market as all major broad based equity 16 indices and equity index futures contracts spent much of the 17 morning and early afternoon in negative territory. This 18 decline in equity markets coincided with a rise in expected 19 volatility as captured by the rise in the CBOE S&P 500 20 volatility index, usually referred to as the VIX. 21 Prior to May 6th, the VIX in 2010 had averaged 22 about 20, indicating a relatively low level of expected 23 market volatility as compared to the levels that prevailed in 24 2008 and 2009. On May 6th, however, the VIX began the day at 25 about 26 and reached levels of about 40 during the day. 19 1 Beyond the effect that economic news was having on 2 credit and equity markets, there was also evidence of a 3 flight to quality as investors sought to extract funds from 4 perceived risky investments and to place them into what are 5 viewed as safer investments, such as U.S. Treasuries and 6 gold. This is evidenced by the fall in the ten-year Treasury 7 yield from a level of 3.58 percent on May 5th to an intraday 8 low of 3.26 on May 6th, before settling at 3.41 for the day. 9 Gold prices, which normally rise in times of market 10 uncertainty, rose on the COMEX where the nearby gold futures 11 contract went from approximately $1180 per ounce at the 12 opening to $1210 per ounce at the 1:30 p.m. Eastern time 13 close. 14 In addition to the flight to quality, global 15 currency markets were indicating concern over the European 16 Union. Shortly after 1:00 p.m., the euro began a sharp 17 decline against both the U.S. dollar and the Japanese yen. 18 CO-CHAIR GENSLER: Just for those on the phone, 19 we're looking at Slide 6. If you might mention. 20 MR. SHERROD: Sure, sure. Moving on to Slide 7, 21 during the afternoon on May 6th, staff observed continued 22 pressure in the market, and equity prices continued to 23 decline, along with the increasing expectations in the market for 24 the VIX for volatility. 25 Between 2:00 and 2:30 p.m., there was an increase 20 1 in the number of liquidity replenishment points or LRPs 2 triggered in the NYSE's trading system. These LRPs are 3 intended to dampen volatility by triggering what is known as 4 a go slow period of trading where additional liquidity is 5 allowed to enter the market. My colleague will discuss LRPs 6 further in his testimony about findings in the securities 7 markets. 8 The existing market decline accelerated and implied 9 volatility sharply increased. By approximately 2:45 p.m., 10 over 200 securities had fallen 50 percent or more from their 11 2:00 p.m. levels. Between 2:45 and 2:47 p.m., the Dow, S&P 12 500, and NASDAQ 100 all reached daily lows. 13 During that same period, all 30 Dow components 14 reached their intraday minimum, representing a range from 4 15 to 36 percent below their opening levels. The Dow bottomed 16 at about 9872, the S&P at about 1065, and the NASDAQ at 17 about 1752. The E-mini S&P 500 index futures contract, the 18 June 2010 contract, bottomed at 1056. 19 There were also instances of securities exchanges 20 declaring self-help against another exchange. A declaration 21 of self-help freed the declaring exchange from their 22 obligation to route orders to the affected exchange, and 23 again, my colleague will discuss self-help further in his 24 presentation. 25 In the futures markets, a confluence of economic 21 1 events, signals from various other markets, and a market 2 increase in sell orders in comparison to buy orders in the 3 limit order book suggest that a significant dislocation of 4 liquidity in the E-mini S&P 500 futures contract may have 5 occurred. 6 After bottoming, equity and equity index futures 7 markets began to rebound. At 2:50 p.m., the Dow was trading 8 at 10,232, and the E-mini S&P 500 was trading at 1,096. The 9 E-mini futures contract climbed further. By 2:53 p.m., it 10 was at 1,118, and the Dow closed at 10,520, down more than 11 347 points, or 3.2 percent from the prior day's close. The 12 E-mini settled at 1,122, or down about 41.5 points from the 13 prior day's close. 14 And we'll turn to the next slide with Robert. 15 MR. COOK: Thanks, Steve. 16 We'll now zero in a little more closely on the 17 preliminary findings presented in the report with respect to 18 the securities markets. The events of the afternoon of May 19 6th with respect to the securities markets can usefully be 20 described in terms of two broad but related themes. The 21 first is a precipitous drop of the major securities indices, 22 as Steve just noted, of more than five percent, followed 23 immediately by a rapid recovery, which occurred consistently 24 across various broad indices and products. 25 The second are the even more extreme price 22 1 fluctuations, mostly losses, that occurred for certain 2 individual securities which were followed relatively promptly 3 by reversions to the price levels consistent with the broader 4 market. 5 Moving on to Slide 10, for those of you who can't 6 see the presentation, and zeroing in a little bit more on 7 the graph showing how the indices moved during that 8 day and to just kind of go through the key hour period where 9 we're going to be focusing most of our attention this 10 morning. 11 So up until about 2:30, there was a significant but 12 not extraordinary down day that was consistent with many of 13 the economic factors that Steve just reviewed, but then the 14 market decline began to steepen. At approximately 2:40, 15 prices declined with extraordinary velocity. 16 By the way, we use 2:40 p.m. a lot in this 17 presentation because it's a useful benchmark for the market 18 prices just before the steep market decline. So you'll often 19 hear us referring to 2:40 p.m. 20 So by 2:42, the Dow is down 3.9 percent. It then 21 suddenly dropped 573 points, which was an additional 5.49 22 percent, in just five minutes of trading, and you can see 23 that spike downward there. By 2:47, it was down 9.16 percent 24 from the previous day's close. 25 The market then suddenly and dramatically reversed 23 1 itself, recovering 543 points in about a minute and a half. 2 By 3:00 p.m., the Dow was down 4.26 percent, and it ended the 3 day down 3.2 percent. And as you can see from this slide, 4 there are similar patterns occurring with respect to the S&P 5 500 index. 6 So while the overall market had a significant down 7 day, the closing numbers don't really tell the full story of 8 these dramatic moves down and then up during the 20 minutes 9 of trading in the mid-afternoon. Understanding the trading 10 activities that contributed to these plunge and recovery, as 11 well as the market structures that permitted it, are 12 obviously crucial areas for further analysis. 13 In this regard, one of the key areas we're focusing 14 on is the linkages between the cash and the futures market. 15 As indicated in the graph, the precipitous decline in stocks 16 and their subsequent recovery correlated very closely with a 17 drop in recovering the value of the E-mini S&P 500 future 18 contract. In a sense, this basically reflects basic market 19 dynamics since much of the price discovery in the broader 20 stock market can occur in the futures markets and traders who 21 believe that one market is overpriced might sell in the other 22 market and/or buy in the other market if they believe the 23 other market is underpriced. 24 So given that the E-mini price fell by more than 25 five percent in five minutes and then quickly rebounded after 24 1 the CME stop logic functionality was triggered, which Steve 2 is going to get into more later, it really isn't surprising 3 that the broader stock market indices showed similarly fast 4 and similarly large declines and recoveries, but of course, 5 the fact that the two moved together doesn't tell us exactly 6 what triggered the price movements. 7 Steve will be speaking later about some of the 8 potential sources of selling pressure in the E-mini contract 9 that came during that vulnerable time. I'll be talking later 10 about some of the potential sources of selling in the 11 securities markets. But I think the key point here is the 12 linkages between the two markets and understanding how one 13 affected the other because of the potential impact that can 14 occur when a trigger occurs in either markets, the follow-on 15 effects for the other market. 16 Moving to Slide 11, this just gives you some of the 17 numbers more precisely for the percentage daily lows relative 18 to the prior day's close. As you can see, each index or 19 product reached its daily low between 2:45 and 2:47 and then 20 recovered over the course of the day to a level higher than 21 their 2:40 p.m. values. 22 Note that the daily decline for the Dow was less 23 than ten percent, which is the first trigger for the existing 24 market-wide circuit breakers. So even if the Dow had gone 25 below ten percent during this time, because it was after 2:30 25 1 p.m., that would not have triggered the existing circuit 2 breakers. The Dow would have had to go down 20 percent after 3 2:30 p.m. in order to trigger the first halt under existing 4 market-wide circuit breakers. 5 We go to Slide 12. This slide helps demonstrate 6 the impact of the market declines on the individual 7 securities. It's a scatter plot representing the lowest 8 transacted price for a given stock on May 6th. Each dot 9 represents one stock, and the low return is calculated from 10 the previous day's close. 11 Most securities, about 86 percent, stayed within 12 ten percent of their 2:40 p.m. price. You can see a red dot 13 for the low in the S&P 500 just above the minus ten percent 14 line. Approximately 14 percent of securities reached lows 15 that exceeded ten percent of their 2:40 p.m. price. 16 You'll see a significant number of stocks had a low 17 transaction for the day well before the market instability. 18 These low transactions increased as you moved to the right 19 over the course of the day, but at about 2:45, there's a wave 20 of low trades as seen in the large blue mass at the bottom. 21 This cluster reflects a significant number of transactions 22 that occurred at prices close to zero, which may have 23 resulted from executions against stub quotes. 24 We'll be talking quite a bit about stub quotes 25 today, and it's discussed in the report. Stub quotes refer 26 1 to quotes that are posted by a market maker in order to 2 satisfy the requirement to make a two-sided market, but 3 because the market maker is unwilling or unable to provide 4 liquidity at the time, it puts up quotes at such a low or 5 high prices, such as a penny or $100,000, that they are not 6 really intended to be executed. 7 We have highlighted in this chart two stocks, A and 8 B, which we'll examine a bit more closely in a few minutes. 9 Stock A suffered a significant decline, and Stock B traded 10 at levels close to 100 percent loss. 11 If we move to Slide 13, this table helps describe 12 the lows during the "hot" period from 2:40 to 3:00 p.m. You'll 13 see approximately seven million trades occurred during this 14 time for almost two billion shares. Note that almost 70 15 percent of trades were executed for a loss relative to the 16 2:40 p.m. price. The bulk of the trades, almost 98 percent 17 of the loss, of the loss trades, or 69 percent of all trades, 18 were executed within ten percent of their 2:40 p.m. price. 19 Thus, the largest overall losses occurred within the ten 20 percent range, which is not too surprising, given that the 21 indices as a whole bottomed out just above the ten percent 22 loss level. 23 But this does highlight the importance of 24 addressing the issues that caused the precipitous plunge and 25 recovery of the broader indices. Even if we do not have 27 1 other securities whose value tumbled down to zero, as 2 reflected in the histogram you just saw or the scatter plot 3 you just saw, the losses to the securities track the index 4 were still enormous. 5 We're now on Slide 14. This scatter plot 6 represents the highest transacted price for a given stock for 7 the day, where that price had highs above zero relative to 8 the prior night's close. Note that there are far fewer 9 positives here than there were negatives in the scatter plot 10 we saw moments ago in the left-hand side of the chart, but 11 you can see a mass starting at around 2:43 p.m. 12 In addition, a small number of stocks transacted at 13 even higher prices. We've identified five stocks that 14 transacted at $100,000 per share. Again, these are likely 15 the result of executions against stub quotes. 16 Unlike in the scatter plot for the daily lows in 17 which there are several extreme lows occurring throughout the 18 day prior to 2:40, here there do not appear to have been any 19 extreme highs prior to about 2:44 p.m. 20 I would like to just sort of zoom in even closer on 21 a couple of selected securities so you can get a sense for 22 what was happening at the time, and Gregg is going to walk us 23 through two selected stocks, and we're now on to Slide 16 in 24 your materials. 25 MR. BERMAN: Thanks, Robert. 28 1 I'll draw everybody's attention to Slide 16, which 2 shows 90 seconds of trading for a particular large cap stock. 3 The first thing to notice is what's not on this graph, 4 namely, the S&P low or the Dow Jones low. This is occurring 5 at about 2:47, which is a full two minutes after the lows of 6 the major market indices. The graph is constructed to show 7 the range of trading and how that trading fits into the bid 8 and asks that were occurring during this 90 second period. 9 The red thick line at the bottom of the trading 10 range represents over each second the absolute worst national 11 best bid. So during a second period we should see a range of 12 bids and this one represents the worst. Hence we should not 13 expect to see trades below that. 14 Similarly, on the up side we have the green line, 15 which represents the absolute best offer during a second 16 period across all exchanges, and we should not expect to see 17 trades above that. The first thing to notice is that as the 18 stock began to decline, we do see some trades that spike 19 above the green line. Whether or not those are artifacts of 20 timing or those are actual trade-throughs remains to be seen 21 and requires further investigation. 22 The more interesting part is on the decline, which 23 started at about 2:47 and continues for about 15 seconds 24 before hitting the lows. What we see is that the trades did 25 not break through or trade through the lowest bids, but that 29 1 the rapid decline in price truly represents a dry up of the 2 liquidity and the collapsing of the bid prices themselves. 3 We had trades that went from the bid to the offer 4 and back and forth within a one second period of time. The 5 offer side collapsed about where the lows are, and you can 6 see that in the green dot, the green line collapsing in the 7 middle, and we also see some trade-throughs or potential 8 trade-throughs at that time. 9 As quickly as the stock fell, the bids immediately 10 climb after hitting the low. The trades follow up the red 11 line. The bid-ask spread narrows, and by the end of the 90 12 second period we actually have a recovery that is greater 13 than the initial loss to begin with. So all of this happens 14 in a minute and a half, and part of the next step of our 15 research is to understand better how the bids themselves 16 collapsed which we are not able to see on this particular 17 chart. 18 Now, this represents an example of a stock that had 19 a very large decline, but did not fall all the way to zero. 20 In the next example, on Slide 17, we see a dramatically 21 different picture. This stock starts off with a tight 22 bid-ask at just before 2:48. Again, this is after the market 23 has declined and while the broad market itself is on the 24 rebound. And for reasons which we still need to explore, the 25 bid side completely collapses in a five second period. 30 1 During that five second period, we see trades at 2 $20. This was the stock that was priced $40 five seconds 3 previously. We have within the same second trades going at 4 $7, $5 and $32 at the offer side. So all of this is within a 5 one second period of time. 6 Finally, by the time you get to 2:47 and 51 7 seconds, we see that the bid side has collapsed to zero and 8 that there are actually trades being executed at those 9 prices. Now, this, again, is important because this does not 10 mean that see bids that are going through or trades that are 11 going through the bid-ask. There is a matching mechanism by 12 which someone had to trade in and was agreeing to sell for a 13 penny. Somebody had to trade in and was agreeing to buy for 14 a penny. 15 Whether or not that was an intention or was that a 16 ramification of the way the systems and possibly stop orders 17 worked remains to be seen, but the market nevertheless stayed 18 for a full ten seconds or so at stub quotes of a penny or 19 more, and again, as rapidly as it collapsed, it automatically 20 rose again. After 2:48 we see a few trades that are still at 21 stub quotes that may not have caught up with the rising 22 bid-ask, and then by 2:48 and 5 seconds, we see that the 23 bid-ask spread comes back to pre-collapse levels, even though 24 you'll notice that there are actually no trades during that 25 period of time. So even though there is very, very little 31 1 activity for the last ten seconds or so of this slide, the 2 bid-ask spreads and the quotes have retained normal level. 3 MR. COOK: Thanks, Gregg. 4 I think it's important to emphasize that these are 5 just two examples of experiences of two particular stocks. 6 The first stock was a component of the Dow and the S&P 500. 7 The second stock that went down to penny trades is a 8 component of the Russell 1000 index. There are many 9 obviously different stories with respect to each different 10 stock, and I didn't want this to suggest that there's only 11 two patterns at play here. There are many different patterns 12 that we've observed, but it helps to show the decline in a 13 stock that declined significantly, but eventually recovered 14 before it hit bottom, and in Stock B a stock that went all 15 the way down to penny trades within a very short period of 16 time. 17 So now we'll talk a little bit about broken trades. 18 If you look at Slide 19, we have the table that we had shown 19 you before on the distribution of low trades relative to the 20 2:40 p.m. price. We've highlighted near the bottom here the 21 trades that were eventually broken pursuant to exchange 22 rules. So after the trading day on May 6th, the exchanges 23 determined to break all trades between 2:40 p.m. and 3:00 24 p.m. that executed at prices different, up or down, from the 25 2:40 -- sorry -- prices 60 percent or more away up or down 32 1 from the 2:40 p.m. price. 2 The exchanges do this pursuant to their clearly 3 erroneous execution rules that are designed to permit them to 4 cancel trades that in their determination were clearly 5 entered into in error. 6 Of the broken down trades, the bulk transacted at 7 prices 90 percent or more from their 2:40 p.m. price, as you 8 can see at the bottom row on the table. The overall market 9 value of the shares involved in the broken trades was, of 10 course, small due to the artificially low share prices 11 involved. But the shares involved in these trades would have 12 been valued at 212 million if we used their 2:40 p.m. 13 benchmark price. 14 If we move to Slide 20, in this figure, we show the 15 time distribution of the broken downward trades. This sample 16 includes all trades identified by the exchanges as broken. 17 So we show some small number of broken trades that occurred 18 outside of the 2:40 to 3:00 p.m. time frame. But most of the 19 broken trades transacted between 2:45 and 2:55. There were 20 about 11,400 or so broken trades between 2:45 and 2:50 and 21 about 4,700 broken trades between 2:50 and 2:55. 22 You remember from the scatter plot that we showed 23 earlier a significant number of these broken trades, about 43 24 percent, were at stub quote prices, here identified as less 25 than a nickel. Stub quote executions occurred in more than 33 1 200 securities, including large, medium, and small cap 2 issuers, but with a concentration in small cap issuers, which 3 would be consistent with the tendency of these stocks to have 4 less liquidity and thus a greater likelihood that sell 5 interest could overwhelm order books in times of heightened 6 volatility. 7 If we move to the next Slide 21, we received 8 information about short sale trades from the exchanges and 9 FINRA. They provided us with a flag for each broken trade 10 that was identified as a short sale, and here we've laid that 11 information on top of the same table. So for the critical 12 ten minutes, 2:45 to 2:55, short sales appear to represent an 13 important fraction of trades executed at prices below a 14 nickel. Shorts appear to represent about 70 percent of the 15 trades from 2:45 to 2:50 and 90 percent of the trades between 16 2:50 and 2:55, not all trades, but we're talking about about 17 the stub quote trades. 18 While this analysis is very preliminary and we have 19 not been able to verify the accuracy of the data we have 20 received, this phenomenon is something we will study closely. 21 We need to understand any role that short selling played more 22 generally in connection with the market decline. 23 We can speculate that these trade were the result 24 of short sale market orders, but they could not bid a bid 25 above a nickel. We can also speculate that the parties would 34 1 not have intended to sell short at these low levels. Note 2 that the short sale executions against stub quotes at or 3 below the national best bid would have been subject to the 4 alternative uptick rule that was adopted by the Commission 5 early this year, but won't be effective until November 10th 6 of 2010. 7 If we look at the next Slide 22, we can talk a 8 little bit about the types of securities that had broken 9 trades. This table is also based on data provided by the 10 exchanges and FINRA. All told, there were 326 securities 11 affected. Of these stocks, 56, or about 17 percent, were 12 primarily listed on the New York Stock Exchange; 225, or 69 13 percent, were primarily listed on Arca. No doubt, stocks had 14 broken trades. About 12 S&P 500 stocks had broken trades and 15 30 stocks in the Russell 2000 were broken. 16 Remember that this doesn't mean that there were no 17 serious losses in these other categories, just that the 18 trades in those other categories were within the 60 percent 19 threshold established for determining whether the trade would 20 be broken. 21 Importantly, about 227, or almost 70 percent of 22 stocks with broken trades, were ETFs. The 99 securities with 23 broken trades that were not ETFs include securities of a wide 24 range of issuers both large and small, and there may be a 25 range of explanations for their aberrant behavior. 35 1 But we'll spend a moment now to look a little bit 2 more closely at the experience of exchange traded funds. I'm 3 going to turn it over to Jonathan to walk us through these 4 slides. We are now on Slide 24. 5 MR. SOKOBIN: Thank you, Robert. 6 So 69 percent of the securities of the stocks that 7 experienced a broken trade were ETFs. A large number of ETFs 8 traded for a short period of time with very significant 9 intraday price swings. Twenty-five percent had temporary 10 price declines of more than 50 percent from their 2:00 p.m. 11 market price. 12 Slide 24 is the same scatter plot that you saw 13 earlier on Slide 12, but now only for the ETFs. There are a 14 couple of features here that are worth noting. We see little 15 evidence of the same pattern of increasing daily lows prior 16 to the 2:33 -- excuse me -- 2:43 to 2:45 period. Instead 17 there's a significant to the ETF sector that hits around 18 2:43, 2:45, and continues on all the way to 3:00 p.m. 19 Ninety-six percent of ETF shares that traded below 20 the 2:40 price traded at prices above or within the ten 21 percent level of their 2:40 p.m. price, but approximately 160 22 ETFs experienced lows for the day that were almost 100 23 percent lower than their May 5th close. 24 Although we don't show the chart, the analogous 25 chart for the trades of -- excuse me -- the high point of the 36 1 trades exclusively for the ETFs, we know that a number of 2 those highs were also ETFs with one ETF experiencing a daily 3 high of 275 percent above the May 5th close. 4 I'd ask you to turn to Slide 25. This slide, this 5 figure shows different categories of ETFs that had broken 6 trades and the proportion of ETFs in each category that 7 experienced broken trades. The 227 ETFs with broken trades 8 were in 838 ETFs. This chart only reflects ETFs and ETF 9 categories for which there were broken trades. 10 There is no evidence that there is a pattern in 11 ETFs, particularly around bond ETFs or ETFs that are focused 12 in areas that are unrelated to the broad market, and in fact, 13 there appears to be at least some weighting toward ETFs that 14 invest in large cap stocks. But we haven't investigated that 15 particularly closely yet, and it's one of the areas that we 16 tend to look in. 17 If you'll turn to Slide 26, this is a sample 18 trading period for an exchange traded fund. This is a large 19 fund that is designed to match the total stock market 20 exposure. This graph shows a period of 16 minutes, and I 21 would point out that you can see between 2:45 and 2:46 is the 22 point where the S&P 500 hit its low. That point, the spread 23 between the national best bid and national best offer remains 24 fairly tight, and as the stock experiences a lowering, a 25 downgrade in the price, that NBBO remains fairly tight. 37 1 Things look fairly normal until about between 2:45 and 2:46. 2 At that point the ETF appears to hit a shock where 3 the bid drops from about $50 to about $30 and then returns 4 quickly, and you can see by the mass of blue within the green 5 and red bars that there is significant trading within that 6 period. 7 The stock rebounds and recovers until a few, less 8 than a second later, and then all of a sudden another shock 9 hits, and the bid drops again. And at this point there is a 10 transaction at a low price of 15 cents below the national 11 best bid. And even as quickly again, as Gregg had said, as 12 quickly as the bid price drops it returns, and the spread 13 between the national best bid and national best offer return 14 to a spread that seems to be normal for this security. 15 MR. COOK: Thanks, Jonathan. 16 Again, that's just one sample ETF, and there are 17 many different patterns when we drill down on the other ETFs 18 that had significant aberrations in their trading. We're 19 going to be talking more later about what some of the 20 theories might be for why ETFs behave this way. 21 Continuing with our observations of what happened, 22 we wanted to talk a little bit about the liquidity issues, 23 and starting with a discussion about the LRPs, or liquidity 24 replenishment points, and the self-help issues that Steve 25 alluded to earlier. And I'm going to turn it over to Dave 38 1 Shillman to talk about this a little bit. 2 MR. SHILLMAN: Okay. Well, thanks, Robert. 3 I think I'd like, as Robert said, I'd like to talk 4 about two unique aspects of the U.S. equity markets that 5 we're examining to see whether or not they contributed to the 6 price declines we saw on May 6th. 7 As many of you know, you know, there are many 8 different places to trade stocks in the U.S. market. There 9 are about ten or so exchanges, many times that more 10 alternative trading systems, and many times more market 11 makers, a multitude of venues. But our regulatory framework 12 brings those multiple venues together through consolidated 13 market data, through private linkages supported by fair 14 access requirements and through a trade-through rule that 15 generally prohibits trades at a worse price and the best 16 displayed price. 17 And in the report, we look at whether certain 18 practices that have developed in this national market system 19 contributed to the events of May 6th and whether or not they 20 exacerbated the price declines we saw. 21 The first of these are the New York Stock Exchange 22 liquidity replenishment points, and today trading in the U.S. 23 markets generally is automated and very fast, but the New 24 York Stock Exchange has retained a volatility moderator for 25 its market known as the LRP where it will go slow when an 39 1 incoming order would result in an execution materially away, 2 generally one to three percent, from the last sale. During 3 this time, which can last from a second or less to several 4 minutes, the New York Stock Exchange seeks to attract 5 liquidity and then reopen in a way that moderates price 6 volatility on its market. 7 When the New York Stock Exchange does this, 8 however, the buy orders on the NYSE's book are unavailable to 9 sellers wishing to trade immediately, and one of the things 10 we're looking at is whether the unavailability of the buy 11 orders on the New York Stock Exchange -- the New York Stock 12 Exchange now overall is about 15 percent of the 13 market -- whether that unavailability to those that wanted to 14 execute immediately, you know, exacerbated the overall price 15 decline or, alternatively, we're looking at whether the 16 availability of a mechanism where a seller might get a better 17 price, actually attracted sell pressure to the New York Stock 18 Exchange to moderate volatility. So it's an open issue, but 19 one we've teed up and are going to probe more deeply. 20 The second aspect of the national market that we're 21 looking at is the self-help remedy, and basically, under Reg. 22 NMS a market can exercise self-help and not comply with the 23 trade-through rule if a market appears to be having systems 24 problems, and as you can see from Slide 28, during the key 25 period on May 6th, four markets, NASDAQ, NASDAQ OMX BX, BATS 40 1 and NSX, exercise self-help against Arca, which has about 13 2 percent of the overall market share, for a period that lasted 3 until towards the end of that half hour. 4 The facts remain in dispute as to 5 whether that self-help exercise was legitimate or not, but 6 the fact of the matter is during that period those who were 7 using the routers provided by those four exchanges did not 8 have access to the buy liquidity that may have been available 9 on Arca. 10 So, again, we're looking at the extent to which 11 self-help against Arca exacerbated the overall price decline, 12 but again, the unavailability of Arca really would have been 13 available for those who were using the routers of those 14 exchanges and anyone else who may have been exercising 15 self-help against Arca. 16 I think if you move to Slide 29, that gives a 17 graphical example of how the NYSE's LRPs increased in the 18 afternoon of May 6th, particularly during the critical 19 period, and the blue bar on the left is the average number of 20 stocks that had an LRP event lasting more than a second, and 21 that average is based on activity during the course of 2010. 22 On the right in the red bar are the numbers that 23 occurred on May 6th, and you can see that starting at around 24 1:00 the number of stocks impacted by LRPs started to 25 increase above the historical norms, with a significant 41 1 increase occurring around between 2:00 and 2:30, and then a 2 huge jump during the, you know, core hot period from 2:30 to 3 3:00, where over 1,000 securities on NYSE were impacted by 4 one or more LRPs, and then it gradually started to decline 5 over, of course, the rest of the day. 6 So, again, this is a phenomenon where we're going 7 to be looking at more deeply to determine whether or not it 8 had detrimental impact on the volatility on May 6th. 9 MR. COOK: Thanks, David. 10 Just a note that these bars show the number of 11 stocks impacted. They don't necessarily show how long they 12 were impacted. It could have happened for a second or 13 multiple seconds. 14 The other thing is that on the self-help 15 declarations that were shown on the previous slide, just to 16 note that those were in the equity markets. There are some 17 conflicting reports about potential self-help claims in the 18 options markets that we're going to continue to look into. 19 So continuing with our discussion of liquidity, if 20 we move to Slide 30 we see a portrayal of the volume activity 21 between 2:00 and 3:00 p.m., and you see the increase in 22 trading that appears as the broad market experiences a 23 disruption and then fails. 24 I would just note Jonathan is going to talk a 25 little bit more about liquidity, but there's an enormous 42 1 volume of trading on this day, 2.2 times the average daily 2 trading volume in the fourth quarter of last year. So May 3 6th had the second highest daily volume for New York Stock 4 Exchange listed stocks across all trading venues. The 5 markets processed more than ten billion shares in New York 6 Stock Exchange stocks alone, and May 6th had the highest 7 daily volume on record for all NASDAQ listed stocks across 8 all trading venues. 9 We'll discuss some of the consequences of this 10 again further, but just to again flag that this high volume 11 of trading result in billions of data elements, millions of 12 trades and thousands of securities, all executed in 13 milliseconds which all contributes to the complexity of 14 trying to recreate what happened here. 15 Jonathan, do you want to talk some more about 16 liquidity? 17 MR. SOKOBIN: Thank you, Robert. 18 If you'll turn to Slide 31, one of the important 19 questions that the staff has been asked to address is whether 20 electronic liquidity providers pulled back during the 21 relevant time frame. The activities of these electronic 22 liquidity providers are important because they've come to be 23 a dominant type of liquidity provider in equity markets. Some 24 estimates suggest that high frequency traders are 50 percent 25 of the total volume or higher. Individual firms can enter 43 1 large numbers of orders that can execute more than a million 2 trades a day. In order to get a better sense of the role 3 that these liquidity providers play, we asked the exchanges 4 to identify the top ten liquidity providers on that exchange. 5 For each of these top liquidity providers the exchanges 6 reported the total number of shares provided or taken by each 7 firm. 8 Liquidity provider for these purposes would provide 9 executable quotes. A liquidity taker would lift or hit a 10 resting bid by each firm. 11 What Slide 31 shows is the percentage of shares 12 traded. The blue line is the number of shares traded. The 13 green line is the percentage of that volume that the top ten 14 liquidity providers accounted for, that they were providing. 15 So that is the percentage of shares traded where the 16 liquidity provider was being hit, was not initiating the 17 trade. 18 The red line below it is for the same top ten, is 19 the percentage of shares traded at any given time for which 20 they were the liquidity taker. All right? So the difference 21 between the two is the extent to which these liquidity 22 providers were net providers of liquidity over the period 23 2:00 to 3:00 p.m. 24 What we see here is because the green line is above 25 the red line at all times, that the top ten liquidity 44 1 providers were, in fact, net liquidity providers across that 2 entire hour. And, in fact, sort of a simple test looking at 3 it with your eyes, not a formal statistical test, it seems to 4 be that these providers effectively provided the net same 5 number of shares throughout the hour. So even in the period 6 where volume spikes, these providers continue to provide 7 approximately the same proportion of share volume and take 8 approximately the same proportion of share volume that they 9 do throughout the hour. 10 But we have to be very cautious interpreting this 11 figure. It only measures liquidity provision with respect to 12 shares and not the prices at which the liquidity was 13 provided. For instance, if most of the liquidity provided 14 during this critical period was at stub quotes, we would not 15 easily conclude that the liquidity providers were dampening 16 the market volatility during the period. 17 Moreover, anecdotal evidence provided to us through 18 interviews with market participants suggest that, in fact, 19 some major liquidity providers ceased providing liquidity 20 during this period. All of this says that this graph is only 21 the beginning of the analysis that needs to be done and 22 points us in the direction of one of the lines of inquiry 23 that we'll be taking. 24 MR. COOK: Thanks, Jonathan. 25 Just to underscore that point, obviously these are 45 1 average numbers, and one of the things we're engaged in now 2 is looking with more granularity at the activities of 3 particular liquidity providers, to focus on whether certain 4 firms who were normally liquidity providers became liquidity 5 takers and at what volumes, and I think we'll see some 6 interesting behavior there that we will point out to you as 7 we get the analysis complete on that front. 8 So that is at a very high level some of the factual 9 findings on the securities side, and we'll now turn it over 10 to Steve and his colleagues to talk about the futures side. 11 MR. SHERROD: Okay. Thank you Good. 12 MR. ENGLE: Hello. Can I say hello, that I've just 13 arrived? This is Rob Engle. 14 CO-CHAIR SCHAPIRO: Thank you very much. We will 15 be starting on Slide 33 if you have the slide deck with you 16 that we e-mailed over the weekend. 17 MR. ENGLE: I do have it. Thank you. 18 CO-CHAIR SCHAPIRO: Great. 19 MR. ENGLE: I will be ready in a second. Thank 20 you. 21 MR. SHERROD: And, again, this is Steve Sherrod 22 from the CFTC. We're on Slide 33. 23 CFTC staff conducted a preliminary analysis, and I 24 want to emphasize like Robert did earlier this is our 25 preliminary analysis, and we reviewed the activity of the 46 1 futures markets to better understand the events that took 2 place on May 6, 2010. The objective was to collect, analyze 3 that preliminary evidence that we had that might be 4 associated with possible causes of the events that occurred 5 in futures markets, including, but not limited to, erroneous 6 activities, such as fat finger errors, cyber attacks and 7 significant system malfunctions. 8 Staff's preliminary review has not at this time 9 found evidence of erroneous activities, no evidence of cyber 10 attack, and likewise no evidence of significant system 11 malfunctions. Rather, our preliminary findings suggest that 12 a confluence of economic events, signals from various other 13 markets, and a market increase in sell orders culminated in a 14 significant liquidity dislocation in the E-mini S&P 500 15 futures contract. This liquidity dislocation was also 16 preceded by some reduction in activity of certain liquidity 17 providers. 18 I'll discuss our review of the role of liquidity 19 and the return to a balance in trading coinciding with 20 triggering of a pre-trade automated safety feature which the 21 CME calls Globex stop logic functionality. 22 Turning to Slide 34, our review focused on trading 23 he liquidity provision in the June 2010 E-mini S&P 500 24 futures contract. That single contract in the E-mini S&P 500 25 comprised 78.2 percent of the total trading volume in the 12 47 1 most actively traded broad based stock index futures 2 contracts on May 6, 2010. An apparent imbalance of orders on 3 the sell side, of course, resulted in falling prices. 4 At the bottom of the price decline, the depth of 5 the order book declined, the price incline induced buyers we 6 believe to enter the market, and the bottom of the price 7 decline coincided with that CME Globex stock logic event. 8 Thirty-five. 9 Consistent with the broad market trends on May 6, 10 2010, trading volume in the E-mini S&P 500 futures was about 11 2.6 times greater than the average daily trading volume over 12 the prior month. On May 6th, trading volume in the S&P 13 E-mini contract was the fifth highest daily volume over the 14 past five years. 15 Furthermore, the contract experienced a 16 significantly higher level of trading during certain 17 concentrated periods of the day. In addition to the overall 18 high daily volume, intraday period by period trading volumes 19 significantly exceeded the average trading volume for the 20 same intraday periods observed over the prior month, 21 especially between 2:00 p.m. and 3:30 p.m., with a spike in 22 volume between 2:40 and 2:49 p.m. 23 As noted earlier this high volume was accompanied 24 by high priced volatility. The daily price range in the 25 E-mini S&P 500 was 112-3/4 points. This represents the 48 1 second widest price range for the day over the past five 2 years. The other four of the top five widest price ranges 3 over these past five years occurred during the financial 4 crisis in autumn of 2008, including the single largest daily 5 price range, 115-1/2 points on October 28th, 2008. 6 Within the trading day, the widest price range 7 between high and low prices calculated over ten minute 8 intervals was 59-3/4 points. That occurred between 2:40 and 9 2:49 p.m., and that coincided with the spike in volume. 10 Turning to Slide 36, this is Figure 29 from the 11 report. The blue bars represent trading volume for ten 12 minute intervals on May 6th, and the red bars represent the 13 average trading volume for the prior month. The volume spike 14 clearly occurs in the 2:40 to 2:49 time period. 15 The tan vertical lines with bar stops, those 16 represent the price range in the ten-minute intervals and, 17 again, the graph shows the largest price range of 59-3/4 18 points during the same time interval as the price spike. 19 According to the CME, over 250 Globex executing firms were 20 active in routing E-mini S&P 500 futures contracts orders 21 into Globex during the hour beginning at 2:00 p.m. 22 Globex executing firm is an entity that is directly 23 connected into Globex. Non-Globex executing firms access 24 that trading platform through a Globex executing firm. 25 Also during this hour of 2:00 p.m., Globex 49 1 transactions in the E-mini S&P 500 futures were recorded for 2 6,939 buy accounts, 6,873 sell accounts, 7,669 buy user IDs, 3 and 7,564 sell user ID. 4 A buy account is a unique Globex account that 5 executed one or more buy orders, and a buy user ID is a 6 unique operator ID and also we refer to that as a Tag 50 ID. 7 That identifies the party who entered the order on behalf of 8 the account. A Tag 50 ID may be authorized to enter orders 9 on behalf of multiple accounts, and as well a single account 10 may have multiple authorized Tag 50 IDs. 11 So to reconstruct the activity is a bit of a 12 challenge, to say the least. At 2:40 though the E-mini S&P 13 500 was trading at 1113. Five minutes later at 2:45, the 14 E-mini had fallen another 57 points and bottomed at 1056. 15 In the second of 2:45 and 27 seconds, the E-mini S&P dropped 16 12-3/4 points over a period of a half a second, 500 17 milliseconds on the sell of 1100 contracts by multiple market 18 participants. 19 This sequence of trades caused the market to trade 20 down to an intraday low of 1056. Further, in the bid-ask 21 spread in the E-mini S&P 500 market widened to 6-1/2 points. 22 That's 26 ticks. A tick is a quarter point. 23 This triggered the Globex stop logic, and that sent 24 the E-mini into a reserve state at 2:45 and 28 seconds. The 25 reserve state held execution of any transactions for five 50 1 seconds. This hold allowed enough orders to flow into the 2 market so that the next executed trade would be within six 3 points of the last trade. At 2:45 and 33 seconds, the E-mini 4 exited its stop logic reserve state. Upon exiting the 5 reserve state, 1,753 contracts were traded at a price of 6 1056-3/4. The E-mini began to recover at that point, and in 7 the single minute of 2:45 p.m volume in the E-mini spiked at 8 78,412 contracts. 9 So turning to Slide 37, I want to talk a little bit 10 about the role of liquidity in the E-mini contract. The high 11 volumes of trading along with sharp price movements suggest 12 that liquidity at times actually may have dropped off. 13 Liquidity reflects the ease with which certain amounts of an 14 asset can be bought or sold without exerting a significant 15 effect on price. Higher market liquidity can be interpreted 16 as a greater collective willingness to execute orders at 17 given prices. 18 While the notion of market liquidity cannot be 19 directly observed, we can describe it, but we can't directly 20 observe it. So market liquidity has multiple dimensions that 21 are hard to capture by a single indicator. Staff at the CFTC 22 reviewed multiple indicators of liquidity, including but not 23 limited to the trading volume, the bid-offer spread and the 24 depth. 25 With respect to these three indicators, high 51 1 liquidity may manifest itself as high trading volume. High 2 volume may indicate the presence of a large number of buyers 3 and sellers willing to transact in significant quantities. A 4 narrow bid-offer spread may also be an indication of high 5 liquidity as it reflects the existence of at least some 6 buyers and sellers willing to transact at prices close to 7 recent transaction prices. 8 And finally, the depth of the order book at 9 successive quotes is an indication of high liquidity in that 10 it reflects the ability to execute trades of size without 11 having to bear large price concessions. 12 Turning to 38, so looking at that first indicator 13 of liquidity trading volume, it did flash some signs of 14 potentially high liquidity, and that is high trading volume. 15 In the June 2010 E-mini S&P 500 futures contract during the 16 period of 2:30 to 3:00, trading volume was about ten times 17 the average daily volume for the same intraday period over 18 the prior month. The high trading volume though was 19 accompanied by significant volatility in the volume. This 20 suggests to me a dislocation of market liquidity with high 21 volume fluctuations possibly occurring at the same time that 22 orders are executed deep into the limit order book, two 23 indicators, flash signs of low liquidity, but at different 24 times. 25 The bid-offer spread widened just prior to the stop 52 1 logic event, and that was at 2:45 and 28 seconds, and it 2 briefly became variable, and the imbalance in the order book 3 with less depth on the buy side was followed -- and that was 4 at 2:30 -- it was followed by the decrease in market depth 5 also at 2:45 at the time of the stop logic event. After 6 that, there was a return to a relative balance between bids 7 to buy and offers to sell. 8 Turning to Slide 39, it's Figure 30 in the report. 9 In this figure the red line represents the volume, and the 10 blue line represents price. The graph displays the level of 11 volume in ten second intervals between 2:30 and 3:00 p.m. 12 Between 2:30 and approximately 2:45, volume rose 13 significantly while the prices fell, and between 2:45 and 14 3:00, volume fell and prices rose. 15 So transaction volume ranged from several hundred 16 contracts per second to several thousand contracts per 17 second. 18 Turning to Slide 40, it's Figure 31 from the 19 report. 20 We also reviewed the behavior of the bid-offer 21 spreads for the best and fifth best quotes in the 2:00 p.m. 22 hour, and in particular, we focused on 2:43 to 2:48 in this 23 figure. The bid-offer spread is a liquidity indicator based 24 on the characteristics of a limit order book. Specifically, 25 the bid-offer spread is calculated as the difference between 53 1 the highest quoted price to buy and the lowest quoted price 2 to sell. 3 This price difference is a measure of the cost paid 4 by a buyer or seller who wishes to transact immediately, and 5 similarly, the second, third, fourth, and fifth best bid and 6 offer prices represent transaction costs to the buyer and 7 seller willing to buy at increasingly lower prices and sell 8 at increasingly higher prices. 9 So the red line here represents the spread between 10 the best bid and offer, and the blue line represents the 11 spread between the fifth best bid and the fifth best offer. 12 The green line is the price of the E-mini contract. 13 Until approximately 2:45 p.m., both spreads were at 14 their minimums, and that's typically what we would observe in 15 this market of one tick in the inside bid-offer spread. At 16 2:45 and 28 seconds, the best bid-offer spread widened to six 17 and a half points, and at this time the Globex stop logic 18 triggered a five second reserve state. 19 Following the reserve state, the first and fifth 20 best quote spreads increased to the period maximum of 21 approximately two and three-quarter points and eight and a 22 quarter points, respectively, and that's 11 ticks and 33 23 ticks. The left-hand scale here is in terms of ticks or 24 one-quarter of one point. 25 By 2:50 and 40 seconds, both spreads declined to 54 1 about one and nine ticks. That's about a quarter point or 2 two and a quarter points, respectively. 3 Turning to Slide 41, we reviewed the depth of the 4 market by examining the sum of quantities of orders resting 5 through the fifth best bid and offer in the limit order book. 6 At 2:30 p.m. and continuing until shortly after 2:40 p.m., 7 significant order imbalances existed between buy orders and 8 sell orders with significantly more orders in the limit order 9 book to sell than to buy. 10 In addition, in approximately 2:45 the depth of the 11 limit order book declined dramatically. But the limit order 12 book became approximately balanced, that is, orders to sell 13 became approximately equal to orders to buy, which is a 14 balanced state for a market to be in. While this relative 15 balance remained through 3:00 p.m., there was notably less 16 depth overall in the market. 17 Turning to Slide 42, I've mentioned the pre-trade 18 automated safety feature on Globex, the stop logic 19 functionality, and I'll review that in more detail now. This 20 functionality was originally developed to address both 21 markets and deferred contract months with low trading volume. 22 The stop logic functionality is designed to stop a cascade of 23 stop loss orders, which is essentially an event where one 24 stop loss order triggers a series of other stop loss orders. 25 Under the CME rules, this functionality is 55 1 initiated when the last transaction price would have 2 triggered a series of stop loss orders that, if they were to 3 be executed, would have resulted in a cascade in prices 4 outside a predetermined range called the no bust range. In 5 this case it's six points in either direction generally from 6 the last transaction price. 7 On May 6, 2010, that was the first day in 2010 on 8 which the Globex system activated the stop logic 9 functionality in any of the equity index futures markets. 10 The stop loss functionality was activated in the S&P E-mini 11 at 2:45 and 28 seconds. 12 In a bond being triggered, the E-mini S&P 500 was 13 sent into what I've referred to as a reserve state. That 14 reserve state is where the executions are held for five 15 seconds. This hold allows orders to flow into the 16 marketplace. If the system can execute within six points of 17 the last trade, it executes. If the system cannot execute 18 within that range, then the reserve state would have been 19 extended for an additional five seconds. 20 At 2:45 and 33 seconds, the E-mini S&P 500 did 21 execute at a price actually up three-quarters of a point and 22 exited the stop logic reserve state, and trading continued 23 throughout the remainder of the day. 24 I'll turn with Slide 43 to my colleague, Andrei, to 25 present some of our large trader analysis. 56 1 MR. KIRILENKO: So we're now on Slide 43, large 2 trader analysis. 3 What we've done is that we've looked, in addition 4 to looking to market-wide measures of liquidity, we have also 5 looked at account by account trading activity trying to 6 identify those particular patterns that we can observe in the 7 way accounts have traded. Specifically, we looked at the 8 activity in the ten largest accounts by net volume and by 9 gross volume in order to see whether there were significant 10 imbalances between large buyers and sellers in the market. 11 We split the critical half hour between 2:30 and 12 3:00 into two periods, between 2:30 and 2:45, right before 13 the stop logic functionality kicked in, and 2:46 to 3:00. 14 And the net volume was computed for each of the accounts in 15 the E-mini S&P June 2010 futures contract. 16 We noticed that during the period from 2:30 to 17 2:45, the volume of trading by the top ten net sellers 18 exceeded that of the net buyers by 20,660 contracts. During 19 the period from 2:46 to 3:00, the volume of trading by top 20 ten net sellers, exceeded that of net buyers by 18,364 21 contracts. 22 Thus, for both periods the trading of the largest 23 net sellers exceeded that of the largest net buyers in the 24 market. In this market, for every seller there is a buyer. So 25 what this statistic illustrates to you possibly is that the 57 1 largest net sellers are selling in the larger lots than the 2 larger net ten buyers are buying. It doesn't mean that there 3 are no buyers for the sellers; that the buyers are possibly 4 buying in smaller sizes, which is consistent with the price 5 concessions that they may be demanding. 6 Finally, we also noted that one trader out of the 7 top ten trading accounts only entered orders to sell, which 8 amounted to approximately nine percent of the volume of the 9 trading during the period. That trader entered the market at 10 around 2:32 p.m. and finished trading by around 2:51 p.m. 11 Thus, the trader sold on the way down and continued to sell 12 as the price level rose. 13 We now turn to page 44. We've also done a 14 preliminary analysis of liquidity by examining trading 15 activity of particular groups of traders in the market in 16 order to ascertain how these particular groups behaved during 17 the critical period between 2:30 and 3:00. 18 As we said before, liquidity is something that is 19 not observable, and liquidity provision has to be 20 appropriately defined as well. For the purposes of this 21 analysis, we used the particular methodology to select 22 significant liquidity providers. 23 What we've done is that we looked at the ten 24 largest long by gross, by volume accounts and ten largest 25 short gross volume accounts, and from those we selected 58 1 accounts that were in both categories, that is, they were 2 both the largest long and the largest short, but their net 3 position during the period was no more than 150 contracts 4 long or short. That is, they stayed relatively flat during 5 this period. They traded a lot, but they didn't accumulate a 6 net position on either side. 7 We defined this group of six trading accounts as 8 liquidity providers. In the E-mini S&P 500 Globex electronic 9 limit order market, there are no designated liquidity 10 providers. That is, no trader has an obligation to provide 11 bid and ask quotation on demand. 12 The six accounts that we classified as liquidity 13 providers participated in about 50 percent of the volume of 14 all transaction size between 2:30 and 2:34 p.m. For each 15 transaction there are two sides, the buy side and the sell 16 side, and so we've computed how many sides these liquidity 17 providers have taken. 18 The remaining 4,573 accounts of the total of 4,579 19 transacting between 2:30 and 3:00 p.m. were defined as 20 liquidity takers. We deliberately made our calculations 21 robust and biased against us so that we could be reasonably 22 sure of this preliminary review. 23 So turning now to Slide 45, it's showing the 24 cumulative volume of transaction size of liquidity providers. 25 The blue line represents liquidity providers as we define 59 1 them, and the red line represents all of the other trading 2 accounts during the one-half hour time period. 3 At approximately 2:35 p.m., the liquidity providers 4 began limiting their trading activity as measured by volume 5 of transaction size in comparison to liquidity takers. By 6 2:45:28, liquidity providers accounted for 46 percent of the 7 volume of all transaction size and by 3:00 p.m., the 8 liquidity providers account for 41 percent of the volume of 9 all transaction size. 10 The decline in the participation of liquidity 11 providers and executed volume of transactions can be 12 interpreted as a partial withdrawal of liquidity by the six 13 significant providers during the period of significant price 14 movement. 15 I'm now going to turn to Rick Shilts, who is going 16 to present the summary of findings from the CFTC side. 17 MR. SHILTS: Yes, and this would be Slide 46. 18 In summary, CFTC staff review of the futures 19 markets on May 6th showed that a number of economic events 20 and market developments led to a broad-based market desire 21 for investors to lessen their exposure to risky assets. This 22 translated into a downward movement in prices across 23 financial markets in conjunction with significant trading 24 volume. 25 At or on about 2:30 p.m., evidence suggests that a 60 1 liquidity dislocation may have occurred in the E-mini S&P 500 2 futures market. At this time, trading volume increased 3 significantly and became highly variable at the time that 4 prices began to plummet. 5 In addition, the electronic limit order book in the 6 E-mini S&P 500 futures market exhibited a significant 7 imbalance of sell orders and buy orders. In the backdrop of 8 declining prices, this imbalance appears to have contributed 9 to a sudden liquidity dislocation despite increased trading 10 volume. 11 At approximately 2:45 p.m., several sell orders 12 would have been executed deep into the limit order book which 13 coincided with significant loss of depth triggering the stop 14 loss functionality. Activation of the stop logic 15 functionality on May 6, 2010, initiated a five second pause 16 in trading in the E-mini S&P 500 futures contract. After the 17 five second pause, the limit order book became more balanced, 18 and the price of the E-mini S&P 500 futures contract 19 recovered. 20 I'll now turn it back to Robert. 21 MR. COOK: Thanks, Rick. 22 That concludes the initial portion of our 23 presentation as to the initial findings, and we'd be happy to 24 take questions before we move on to lines of further inquiry 25 and research. 61 1 Thank you. 2 CO-CHAIR SCHAPIRO: Thank you, Robert, and thank 3 you to all of you for an excellent presentation. 4 And I would open it up to Committee members for 5 reactions or questions. 6 MR. KETCHUM: Thank all the staff. It was 7 extremely well done and charts very revealing. 8 A few questions on the equity side just to try to 9 understand a little bit better. First, in particular, with 10 respect to the two stocks and one ETF vis-a-vis the 11 trading with regard to both bid-ask spreads and the number of 12 trades outside those bid-ask spreads, I think it would be 13 interesting. It appears from those three stocks that none of 14 the -- first, there are very few trades outside bid-ask 15 spreads on the down side, and none of them seem to track the 16 Arca self-help moments per Dave Shillman's thing. 17 It would be interesting to see across stocks how 18 many trades outside of the bid and offer, the NBBO track Arca 19 self-help moments, if indeed that self-help thing resulted in 20 avoiding liquidity in the Arca book. You would assume there 21 were trade-throughs. 22 The other one that I'd appreciate your maybe taking 23 a look at is the LRP information obviously is very 24 interesting to understand the results during those times. It 25 would be good to do, again, similarly to take a look at at 62 1 least in a few stocks where there was significant volatility 2 LRP trade-throughs and the breadth and dimension of the 3 trades that occurred outside of the New York indications 4 during the LRPs. 5 The other thing that occurs to me is that on the 6 liquidity provider chart on page 31, I think, Jonathan, your 7 point. If you're really looking at the logic and the 8 implication of different types of or the absence of market 9 maker obligations, to your point, what you're looking at is 10 whether trading occurred on a relatively slanted basis at or 11 around prices around there. So I think it would be good to 12 look not just at whether liquidity was provided, but how much 13 of that liquidity, say, was provided within 20 percent of the 14 trading price at the beginning of the period or, again, 15 looking at the slope from the liquidity provider standpoint. 16 Were they providing anything close, to use old terms, market 17 continuity during that time or were they sitting down at the 18 bottom and simply trading? And it would be great to sort of 19 get a feel for that kind of thing. 20 And the last suggestion would be, while some of 21 these charts do this very well, there are obviously moments 22 on a stock-by-stock basis, as well as the E-mini, and the 23 E-mini ones show this, I think, particularly well, where I'll 24 call flex trading moments, the moments when the thing tanks, 25 and the second or in between second when that pressure is in 63 1 there, sort of looking at liquidity providers and whether 2 they were taking or providing liquidity at those moments when 3 the stock turned or when the stock tanked also would be 4 helpful in sort of identifying what really was the reaction 5 of the major quote liquidity providers during the time. 6 MR. BRENNAN: I want to reiterate Rick's thanks for 7 the extraordinary work, and it looks like a career's worth of 8 work in a couple of weeks. It's really outstanding. 9 This may not surprise you. My eye was drawn to the 10 ETF side of things, and a couple of questions. One is soft, 11 I think, and I don't know whether you have answers to it 12 anecdotally, which is whether in the ETF creation process, 13 did you have any sense for whether it was an inability or an 14 unwillingness to participate in that market when the bids 15 fell away. Because this is a more manual process in many 16 ways. It's much closer to the old specialist process today. 17 And so that's a question that I think would be 18 interesting to know more about, and then a related question, 19 thinking about the E-mini and how important are -- and it 20 may not be important at all -- but how important are the 21 futures markets in the hedging in the creation exercise. Is 22 that part of this? 23 You know, in your summary points, two of the six 24 related to the ETF business, and it's interesting and it's 25 relatively new, still relatively manual. It would be 64 1 interesting to know answers to both of those either 2 statistically or probably more anecdotally. 3 Jonathan, you look like you may have those answers. 4 I don't know whether you're supposed to answer questions or 5 we're just supposed to ask them. 6 CO-CHAIR SCHAPIRO: Go right ahead. 7 MR. SOKOBIN: Thank you. 8 I guess where we don't have answers yet, I think 9 we'll pass, but where we do, we'd like to provide them. 10 We talked to some participants in the ETF market. 11 They told us by the end of the day things were sort of back 12 to normal, and that there was, in fact, no particular issues 13 with end of day pricing. But we did ask and did receive for 14 some fund families information on creation and redemptions. 15 We have only very limited information, but the information 16 that we have to date suggests that there was no particular 17 difference around May 6th for funds, in creation redemption, 18 for funds that had broken trades and ones that did not. 19 CO-CHAIR GENSLER: Can I just ask? Jack might know 20 this better, but, Jonathan, are creations and redemptions 21 done at the end of the day or are they done, you know, down 22 to this nanosecond sort of time frame? 23 MR. SOKOBIN: Our understanding is that creations 24 and redemptions are only done at the end of the day. 25 MR. COOK: I mean, just to follow up on that one 65 1 point, it's an excellent question that we intend to look into 2 further because even if you set aside the creation and 3 redemption, obviously the price of the ETF is being driven by 4 the trading throughout the day and the ability of market 5 makers in ETFs, whether formal market makers or people who 6 are substantial liquidity providers in ETFs, to hedge in the 7 underlying securities would obviously be something we should 8 be looking at when we're thinking about the effect of prices 9 of the ETFs themselves. 10 So that linkage between trading or the ability with 11 certainty to execute a hedge in the underlying stocks versus 12 buying or selling the ETF is something that we really need to 13 explore because that might be where some of the breakdown is 14 occurring in the pricing of the ETFs. 15 MR. BRENNAN: It does feel that way because it 16 encompasses many of the macro issues that you're looking at 17 in some ways. So that's great, Robert. Thank you. 18 MS. PHILLIPS: Thank you. 19 And I also want to thank all of the staff for this 20 very complete report. It's very impressive. You obviously 21 have been crunching lots of numbers and printing out lots of 22 graphs, and it's very helpful. 23 I have a few questions, and some of them I think 24 you'll be able to answer and some may lead to maybe some 25 additional analysis. 66 1 First of all, I'm not as familiar with this stub 2 quote stuff, and I guess first of all I'd like to know what's 3 the theory behind having stub quotes at all. 4 MR. COOK: It's in some ways a vestige of a time 5 when market makers were the central liquidity providers and 6 there was a more formalized role for market makers in more 7 concentrated markets. But now that we have dispersed 8 markets, many different trading venues, many of them don't 9 have obligations to even have market makers on some trading 10 venues. 11 However, there continues to exist in the rules this 12 concept of a market maker and being a market maker can 13 provide you with certain regulatory advantages, and so some 14 firms may try to opt into those advantages by claiming to be 15 market makers, and whether or not they're really providing an 16 effective market, which I think is a point Rick was also 17 raising, is the key question here, and whether we've now 18 observed a situation where in order to avail themselves of a 19 regulatory status they're providing quotes that actually 20 undermine the integrity of the markets rather than promote 21 them, I think, is a very important policy question we need to 22 be looking into. 23 CO-CHAIR SCHAPIRO: It really comes, doesn't it, 24 Robert, from the obligation to make a two-sided market and to 25 have quotes on both sides, and if you really actually don't 67 1 want to at the end of the day provide liquidity, you just 2 widen out that spread dramatically, and that's how you end up 3 with a penny to $100,000, for example, as potential quotes, 4 but they're meaningless in some ways. 5 MR. COOK: Right. You would expect that most of 6 the time you're putting up quotes that you don't really 7 expect to be executed because you aren't able to provide 8 liquidity or you're not willing to provide it, and normally 9 it doesn't matter because no one ever or the order book never 10 disappears before that quote would be hit. 11 But obviously, this was an instance where the order 12 book fell away and people went right down to the stub quote. 13 MS. PHILLIPS: Okay. If you didn't have those sort 14 of nominal quotes, stub quotes, there just wouldn't be the 15 trades; is that right? Trading would have stopped? 16 MR. COOK: Yeah, I believe that's right. There 17 wouldn't have been a quote to execute against. 18 MS. PHILLIPS: Yeah. 19 MR. COOK: So who's executing against them, it's 20 likely market orders coming in saying, you know, in order to 21 execute at whatever the market is, which is another issue 22 that we're going to be looking at, is whether the order types 23 sort of combine to contribute to this problem, but, yes, it 24 would have been -- if there was no quote there, there would 25 have been no execution. 68 1 MS. PHILLIPS: During the day, I guess I'm 2 wondering were there interday margin calls and did the money 3 move as it was supposed to and were there -- and even at the 4 end of the day for the margin calls, did everything settle 5 out as it was supposed to? 6 I guess I'm sort of wondering if you're following 7 the money where were the problems. 8 MR. SHERROD: Well, in the futures markets with 9 respect to the trading that took place on May 6th, at the 10 CME and ICE futures U.S., the clearing and settlement process 11 worked effectively and without incident, and the CME 12 collected and paid its clearing members in a timely manner, 13 and ICE Clear U.S. also performed well with its clearing 14 members. With the end of day mark to mark calculations there 15 weren't any particular difficulties with those end of day 16 payments. 17 MR. COOK: And then on the securities side, yes, we 18 followed closely working with the clearing agencies. There 19 weren't any significant issues there in terms of settlement 20 of these trades, and obviously at the customer level the 21 clearing firms, there were a number of margin calls that 22 needed to be made, but we're not aware of any kind of 23 significant issues coming out. Obviously, there are, you 24 know, case-by-case; individual customers may have had issues, 25 but there wasn't a sort of systemic type issue coming out of 69 1 that. 2 CO-CHAIR GENSLER: I would add it's not that we 3 haven't heard from some market participants as to whether any 4 of the clearing computers may have had some nanosecond 5 slowdowns and so forth, but in terms of the raw dollars is, I 6 think, what Mr. Sherrod was talking about. 7 MS. PHILLIPS: So none of the firms then got into 8 capital problems? I mean, because this is a lot of money 9 moving around, you know, and sizable prices. 10 MR. COOK: We're not aware of any firm that had, 11 you know, a capital problem to the point of, you know, being 12 able to or needing to report it solely as a result of this, 13 but there are a lot of firms out there, and I'm not sure 14 we've been able to track through each and every one. We've 15 been focusing, obviously, on the bigger ones and working our 16 way down. So we might come back to you and tell you that 17 there were some that got a little into the red depending on 18 their trading activity, but so far we haven't identified any 19 firms that would rise to the level of significance to raise 20 for you. 21 MS. PHILLIPS: Okay. And this is something I guess 22 I don't understand. If you hit -- and this is on the 23 securities side -- if you hit one of these speed bumps with 24 an individual stock and there's a slowdown in trading, how 25 does this then affect the indexes? And does that cause 70 1 problems throughout the markets? 2 MR. COOK: Well, the trading is still occurring in 3 the security, but it's a slowdown on a particular exchange in 4 the case of an LRP or it's a decision by one exchange not to 5 route to another in the case of self-help. So the trading is 6 actually still occurring during that time. Obviously the 7 trades that would be occurring would be in some cases were 8 at prices that were significantly lower, but also each index 9 has its own rules about how to calculate its index pricing 10 and often take into account things such as a trading halt or 11 other situations with respect to a particular security, and 12 that could all -- so you know, there are some ways in which 13 index providers deal with these kind of events. 14 But just back to the point, the trading didn't halt 15 in a security system. 16 CO-CHAIR SCHAPIRO: I see. 17 MR. COOK: It halted in some cases, in the 18 particular venue -- a pause would be a better way to put 19 it -- in the particular venue for a moment. 20 MR. BRENNAN: I just want to follow up on that 21 because in the report I wasn't clear on that either. So 22 could the same stock have been being executed in New York at 23 a different price than it was executed in one of the 24 self-help declaration venues? Simultaneous? I'm just not 25 familiar enough with it. 71 1 MR. SHILLMAN: Basically, when New York goes into 2 an LRP, it will pause for, you know, a second or two or up to 3 a few minutes. During that time the other markets will 4 continue to trade that security, and then New York will 5 reopen following the LRP at a print, and that print is then 6 subject to, you know, the trade-through rule, and that would 7 have to be brought into the national market system and the 8 prices established. 9 MR. BRENNAN: So there's no ability to arbit beyond 10 a second or something else exchange to exchange. 11 MR. SHILLMAN: Yeah, there isn't an ability to 12 arbit because basically New York has shut down in trading for 13 a few seconds or a minute and then comes back up. 14 MR. COOK: But I think is a crucial point and also 15 raises the further research question that Rick was 16 identifying, which is to what extent does that correlate with 17 actual price declines in a particular security, right? 18 Because there were a lot of broken trades in securities that 19 aren't listed on New York. So New York only trades New York 20 listed securities. So its LRPs only affected New York listed 21 securities. 22 Many, as you saw from the presentation, there are 23 many trades that were broken in other securities that could 24 not possibly have been affected by the LRP because those 25 securities never trade on New York. So we have to match up 72 1 both in terms of volume and timing the implications of an LRP 2 or a self-help declaration what it would actually mean for 3 liquidity in the other trading venues. 4 CO-CHAIR GENSLER: Robert, if you answer that 5 question also on self-helps is there an opportunity to trade 6 away from a market if NYSE Arca had these self-helps? 7 MR. COOK: Right. So if, for example, NASDAQ 8 declares self-help against Arca, what that means is that 9 NASDAQ will not route orders to Arca even if Arca has a 10 better priced order, but NASDAQ can still execute orders on 11 NASDAQ. It can still route orders to other exchanges. 12 Moreover, other routers of orders can still go to 13 Arca. It only affects NASDAQ's routing of orders to Arca. 14 Those who are on Arca can also choose to send their orders to 15 another venue as well. So with the self-help the question is 16 how much resting liquidity was there at the exchange, at Arca 17 in this case, that got trapped there or was it still 18 available effectively because people just didn't go through 19 NASDAQ. They went directly to Arca, which is quite common. 20 The more sophisticated traders wouldn't necessarily rely on 21 one exchange to route to another. If they thought there were 22 better prices at that other change, they'd go there directly. 23 I'll ask Dave if he wants to amplify a little more. 24 MR. SHILLMAN: You know, I think that's right, and 25 the key question I think you started to mention is whether or 73 1 not that liquidity in Arca and much liquidity today, 2 particularly on the exchanges, is made through electronic 3 market makers who are rapidly in and out of venues. You 4 know, if, in fact, Arca appeared to have system problems, 5 they'd then just move their quotes elsewhere. Did that 6 liquidity really stick on Arca? 7 MS. BORN: Could you expand a little bit also on 8 the broken trades? I guess I was a little taken aback. That 9 seemed like a lot of broken trades to me, and were they 10 mostly these ones at the stub quotes or is this 11 routinely -- and have there been complaints about it? 12 CO-CHAIR SCHAPIRO: I can answer the last part. 13 Yes. 14 (Laughter.) 15 MS. BORN: Oh, I'll bet. 16 MR. COOK: Yes, I think if you go back, if we were 17 to go back to the graphs, many of the quotes, many of the 18 broken trades were at the stub quotes, at the lowest, if you 19 define that as sort of in the 90 percent to 100 percent range 20 of loss. Broken trades happen all the time because of 21 pricing irregularities that can happen on exchanges, and the 22 process is designed to be one that protects parties from a 23 situation where a trade for whatever reason gets entered into 24 at prices that no one would deem reasonable given where the 25 rest of the market is at that time. 74 1 Obviously in this case there was a large number, 2 and the exchanges convened after the market closed to try to 3 determine where to draw the line, and then it was through 4 those discussions that they drew the line at the 60 percent 5 point. 6 There's a lot of arguments on either side of that 7 question. If you break trades at a lower point, you may be 8 discouraging liquidity providers from coming in and offering 9 buy side liquidity because their trades will just be broken. 10 On the other hand, the lower you set the threshold 11 if it was 70 or 80, you know, you're keeping on the books 12 trades that are really at prices well below the market, and I 13 think one of the things that Chairman Schapiro has asked the 14 exchanges to do is to come up with a process for determining 15 at what level trades would be broken if this sort of event 16 were to happen in the future, where the key is to have 17 predictability consistency and transparency around the 18 process so that when you're trading today you have some 19 assurance as to what levels you could enter into trades and 20 have them not broken versus having them broken and rather 21 than have it be the result of an ad hoc process. 22 I think just one thing to point out about this is 23 that the reason we're focusing on broken trades today and as 24 a result of this event was precisely your point, that there 25 are so many of them. Broken trades, the process has worked, 75 1 you know, quote, unquote, worked to some extent because it 2 really has only hit individual securities, and in some ways 3 that's what it was designed to do, was to get the one off 4 unique situation where there was something peculiar about the 5 particular stock that cause a price aberration. 6 What this event showed us was that there was a 7 shock of some kind to the system and it led to the indices 8 bouncing, and then some stocks just had, you know, as a 9 result of that presumably some very significant outliers in 10 terms of trading, and it was the volume of that that has 11 caused us all to focus on broken trades in the way that we 12 haven't had occasion to before. 13 MR. SHILLMAN: One thing I'll add is that the stock 14 by stock circuit breakers that were proposed by the exchanges 15 last week should greatly reduce, once they're implemented, 16 the number the number of broken trades. 17 MS. PHILLIPS: Were you -- I guess both 18 agencies -- reasonably satisfied that the market seemed fully 19 integrated across both the derivatives and the underlying 20 securities throughout this entire day or hour? 21 MR. SHERROD: I think that's exactly an area that 22 we want to work on carefully. Robert and I have talked about 23 our areas for further review, and we want to carefully and 24 thoughtfully analyze, do a careful side-by-side review of to 25 500 stocks in the S&P versus the S&P 500 E-mini contract, but 76 1 to this point we have not completed that. 2 MR. COOK: I would agree with that, and I would 3 just add that if you go back in your mind to the very first 4 slide where we had the plunge and then recovery and those 5 various indices -- 6 MS. PHILLIPS: Yes, right. 7 MR. COOK: -- which included both the futures and 8 the securities, in some respects you could say, yes, they 9 were very integrated, but that also highlights the regulatory 10 concern that a shock in one market or something that's 11 happening in one market can also trigger events in another 12 market or at least they move together, and it means all the 13 more -- it just highlights the need all the more for close 14 integration of the way we think about how the markets work 15 together, and that if there's something going on in the 16 securities side that could affect the futures and the futures 17 that could affect the securities, we need to coordinate our 18 regulatory approach. 19 And this is something that we're talking about, for 20 example, in terms of some of the stock-by-stock type circuit 21 breakers and any other types of regulatory initiatives that 22 we might undertake. 23 CO-CHAIR SCHAPIRO: Thank you. That's very 24 helpful. 25 Brooksley. 77 1 MS. BORN: Maybe I'll start with just a follow-up 2 on the self-help mechanism that NASDAQ activated. How long 3 does that continue, and what's the mechanism for its 4 discontinuation? 5 MR. SHILLMAN: Well, the way that's designed to 6 work, when one market detects that another market is not 7 responding within a second and does that more than once, then 8 it is entitled upon giving notice to that other market that 9 it's exercising self-help to route away from the market and 10 not go to it. And what's supposed to occur at that point is 11 there should be a conversation between the market exercising 12 self-help and the market that is purported to be slow to 13 figure out if there's a problem, and if so, if there's a 14 problem, that market is supposed to market its quote as slow 15 so no one has to go to it or, if there's not a problem or it 16 has been solved, the market that's exercising self-help will 17 then begin routing to it. 18 So it's a process that is designed to give an 19 immediate recourse to a marketplace that detects another 20 market as slow, but then there's a mechanism to basically, 21 you know, find out what the truth is, and as soon as that's 22 determined, then basically once that market comes up again 23 and stops having systems problems, routing will then occur. 24 MS. BORN: And what happened on May 6th? 25 MR. SHILLMAN: Well, on May 6th, again, you know, 78 1 there's some disagreement as to whether or not the self-help 2 exercise was legitimate or not. There are two sides to this 3 story, but you saw from the chart there were four markets 4 that exercised self-help against Arca during the relevant 5 period, and towards the end of that hour, right before three 6 and a little after three, they all began routing to Arca 7 again. 8 So whatever the issue was was resolved within a 9 period of, you know, ten to 15 minutes. 10 MR. COOK: It might be fair to say that self-help 11 is something that does occur from time to time. 12 MR. SHILLMAN: It's exercised with some regularity, 13 on average about once a week. 14 MS. BORN: Let me also ask about whether and how 15 carefully you all are looking at the market quotes, the stop 16 loss quotes, and the stub quotes as causative factors here, I 17 mean, without kind of, quotes, without limit on it 18 presumably the market would have not dropped and had 19 executions at such a low level and probably there wouldn't 20 have been the enormously high quotes, high executions without 21 the stub quotes, too. 22 Is there consideration being given to requiring 23 that orders have limits? And what are the issues involved 24 with that? 25 MR. COOK: Yes, we are taking a look at that. I 79 1 mean, I think it would be -- at least preliminarily we're not 2 thinking of them as causative factors in the sense of having 3 caused the overall broad index declines in a very dramatic 4 way, but they are certainly exacerbating factors in terms of 5 the performance of particular securities. 6 So, yes, one of the key things we need to think 7 about is to the extent there were market orders or stop loss 8 market orders which are essentially orders that are pending 9 until you hit the stop level and then they become market 10 orders, and I think Steve had referred to that as sort of 11 potential contributing to a cascade of orders that could push 12 the price down; that is something we're very focused on. 13 There are a number of competing regulatory issues here. 14 Market orders are widely used and normally don't create 15 problems. They give the investor the certainty of execution 16 so they know that when they put in a order to sell, 17 it will sell. But what they're assuming when they do that is 18 that it will sell at a price reasonably related to what they 19 think of as the current price and in an orderly market. And 20 so there are a range of different approaches we can think 21 about here. One is better education for investors about what 22 it means to have a market order, the consequences of that, 23 particularly if events occur like this again. 24 But there are also other types of approaches that 25 we can think about in terms of both encouraging greater use 80 1 of limit orders or other types of filters that might be put 2 in place by exchanges or other trading venues that would 3 collar orders around a certain price relative to the market. 4 Those are very significant changes that affect the 5 way people behave in their trading activities, and so I think 6 we want to approach that very carefully, but we clearly 7 need to look at it very closely. 8 CO-CHAIR SCHAPIRO: Could I follow up on that, 9 Robert? To the extent we were able or the markets proposed 10 eliminating the use of stub quotes, does that change how you 11 think about things like stop loss orders if stub quotes don't 12 exist in the marketplace? 13 MR. COOK: Well, I think it might. I mean, I think 14 all of these are interrelated just as the single stock 15 circuit breaker will change the relative significance of 16 certain other potential exacerbating factors. Yes, so if we 17 were to get -- I think part of the challenge here will be to 18 identify which are the changes that we could make most 19 effectively without causing disruption in the markets and 20 then once those happen, are there any other changes that 21 would have to happen? 22 And frankly, we still need to evaluate which are 23 the changes that would be appropriate to make in the first 24 place. I think it's very early in the analysis of this, but 25 yes if you did something with stub quotes that might well 81 1 affect the need to do something with market orders. 2 MR. SHILLMAN: And certainly the outrageous prices 3 we saw May 6th were very likely the result of stub quotes, 4 and as we talked to market participants, we found virtually 5 no defenders of stub quotes, and to the extent those are 6 eliminated one way or the other, it could certainly reduce 7 the need for action on the market order side for the 8 outrageous prices, but I think it may still be worth thinking 9 about whether investors should be aware they could get a bad 10 price, albeit not a stub quote, unless they had better 11 controls or some sort of limit price. 12 MS. BORN: I think Steve Sherrod wanted to add 13 something. 14 MR. SHERROD: I just wanted to talk about the 15 pre-trade automated safety features that exist on both CME 16 and ICE. Those types of features extend to both volume, to 17 have a volume restriction, to avoid a fat finger error, but 18 they also extend to price. So there's a no bust range or a 19 range of reasonability where orders must be entered within 20 that range, and even a market order that would be entered as 21 a market protected order to be executed at a price no more 22 than so many points from the last transaction or the 23 prevailing bid and offer prices. 24 So there are a range of these features already 25 built in on the futures side that keep the futures market. 82 1 It doesn't stop trades from executing rapidly and bringing 2 the price up or down, but it does stop the price from spiking 3 down to a penny or up to 100,000. 4 MR. COOK: I'd just note there are some securities 5 markets that have similar logic in their order processing 6 systems. It's just not uniform across all the markets. It's 7 an area where each market has been able to offer to its 8 participants a range of options. But because we have 9 multiple markets with multiple different trading practices, 10 some actually have implemented this type of logic. In fact, 11 some of those markets had lower instances of broken trades, 12 which may well be the result of that. 13 MS. BORN: I also just wanted to address the 14 futures markets in other broad-based securities indices. I 15 know that the E-mini, particularly the June 2010 contract, 16 was a very big percentage of the overall contracts in 17 broad-based securities index futures, but I wondered whether 18 or not an investigation of what went on in the other index 19 futures is ongoing, number one. 20 And also, whether there has been an analysis of 21 individual stock futures and how they behave compared to the 22 behavior of some of the aberrational individual securities 23 trading. 24 MR. SHERROD: We have begun looking into the other 25 markets. In particular, we've received a report and 83 1 discussed trading in the Russell 2000. It is ongoing. As 2 you mentioned, most of the trading, about 78 percent of all 3 the trading is in the E-mini, and we focused on that first, 4 but we are looking into the others. 5 With respect to the second question, my 6 recollection of single stock futures trading volume for the 7 day was about 12,000 contracts. So it wasn't that terribly 8 significant and it's lower in our priority list. 9 MS. BORN: Thank you. 10 CO-CHAIR SCHAPIRO: Are any of our phone 11 participants interested in comments or questions? 12 MR. ENGLE: Yes, I would like to ask a question. 13 This is Rob Engle. 14 CO-CHAIR SCHAPIRO: Please go right ahead. 15 MR. ENGLE: I was very interested in the graph on 16 Slide 45 that showed the -- one of the slides that showed a 17 reduction in liquidity supply. I wondered whether this is an 18 extraordinary amount of reduction or whether this happens 19 typically when the market has high volatility. 20 MR. KIRILENKO: Thank you very much, Professor. 21 This is Andrei Kirilenko from the CFTC. 22 I think this is another issue that we'd like to 23 investigate further. At this point we're able to analyze or 24 this very particular period of time but we'd like to go back 25 and see what happened before 2:30, what happened after 3:00, 84 1 what happened on other days and investigate further how 2 liquidity is typically provided and then contrast it with 3 what happened specifically on this day. 4 MR. ENGLE: Right. Because we would anticipate 5 that as volatility goes up, you would see reduction of 6 liquidity supply in many cases. 7 MR. KIRILENKO: We've, again, focused on the six 8 largest. It would also be interesting to drill down to what 9 are the -- who sort of stepped in and provided liquidity 10 during that time, how it was provided. The information, the 11 data that we have is very granular. We could look at whether 12 the quote was sitting there or whether it was picked off. We 13 could look further into this and see, sort of drill down into 14 more details on how the liquidity provision changes as the 15 broad market experiences periods of increased and volatile 16 volume and volatile prices. 17 MR. ENGLE: And is there a similar analysis that 18 you are doing for the securities? For the equities? 19 MR. SOKOBIN: Yes. We will be doing a similar 20 analysis. The slide that we showed on liquidity takers or 21 providers was just the first step in putting together the 22 order books and trying to understand where liquidity was 23 being provided and where it was being taken. 24 MR. ENGLE: Okay. Thank you. 25 CO-CHAIR SCHAPIRO: Jonathan, what was that slide 85 1 number just for Professor Engle's -- he joined us a little 2 bit late. 3 MR. SOKOBIN: That was Slide 31. 4 CO-CHAIR SCHAPIRO: Slide 31. Okay. 5 MS. O'HARA: Hi. This is Maureen O'Hara. I wonder 6 if I could ask a question. 7 CO-CHAIR SCHAPIRO: Of course. 8 MS. O'HARA: My question actually builds on 9 something you've just been speaking about, which is the 10 construction of the order book. I wondered if anyone has 11 looked yet at cancellations and how those contributed to the 12 hollowing out of the order book as the, you know, sort of 13 problems became apparent. That's sort of question number 14 one. 15 And then question number two is just a technical 16 question maybe you could help me with. I know that a lot of 17 the exchanges sell their own data feeds to, well, HF firms in 18 particular, but to other firms as well, and I wondered to 19 what extent are the proprietary data feeds of the 20 exchanges -- how much out of sync were they with the regular 21 consolidated data feed? 22 That is, I'm trying to understand whether, in fact, 23 the problems affected all of the information that everyone in 24 the market was getting or if it was affecting some people 25 differentially. 86 1 MR. BERMAN: Those are great questions, and they're 2 questions that we definitely would love to know the answer 3 to. On the first part about the cancellation of the orders 4 themselves, so far we have a full slide on this that we'll 5 explain later in the presentation. The data that we've 6 received has been on the top of the order book for each of 7 the individual exchanges. The cancellation information that 8 we received has been on broken trades, trades that were 9 actually canceled later. 10 We have not yet analyzed the depth of the order 11 book to see to the extent that orders were actually placed 12 and then they were later withdrawn, which we know is a common 13 technique in many types of proprietary trading strategies and 14 algorithmic strategies so that's an area of keen 15 investigation. 16 As we'll get into, the volume of data is enormous. 17 We're dealing with literally tens and tens of billions of 18 individual data elements to try to reconstruct some of those 19 order books. 20 We have also heard, to the second part of your 21 question, on the feeds themselves. The data that we receive 22 is quite precise. It has eight significant digits so it 23 goes down to the millisecond level, but what we have heard 24 directly from the exchanges and from those who have sent us 25 the information, just because something says it was traded or 87 1 a bid was placed at 3.567 milliseconds does not mean that it 2 came before the 3.568. 3 There are latencies in the different feeds. The 4 time that it takes an order to transmit from one to another, 5 clocks get out of sequence. So that's one of the areas that 6 we need to investigate. To try to minimize that, we've 7 grouped everything by the second. So if you recall some of 8 the charts where we see apparent trade-throughs, some of 9 those might be artifacts of saying, well, that actually 10 happened the second before and that just happened to be on 11 the one second boundary, but it is an area that we do need to 12 discover a lot more information on. 13 MR. KIRILENKO: On the E-mini side, Professor 14 O'Hara, we are fortunate that the trading occurs in one 15 trading venue and all orders are matched in one by matching 16 algorithm. So we are able to reconstruct the limit order 17 book, and we are very keenly looking at cancellations when 18 they happen, who submitted them, and how they were done, as 19 well as other more granular information on the limit order 20 book side. 21 MS. O'HARA: Well, thank you. 22 CO-CHAIR SCHAPIRO: Are there other questions? 23 (No response.) 24 CO-CHAIR SCHAPIRO: Okay. With that then, I guess 25 I'll turn it back to Robert and Steve to talk about 88 1 additional data requirements and the next steps we're going 2 to take. 3 MR. COOK: Thank you. 4 So we'll just discuss briefly some of the further 5 analysis, and we've touched on a lot of this already in the 6 Q&A. So this will be familiar territory, but why don't we 7 begin with Slide 48 on the securities side? 8 And I'll turn this over to Gregg to just sort of 9 describe what this slide is about. 10 MR. BERMAN: Thanks. 11 On Slide 48, what we've done is try to present a 12 bit of a framework for how we're going to think about both 13 the next steps and the analyses going forward. As you can 14 see from the questions today from all the members, there are 15 many different directions and many different lines of 16 inquiry. 17 So what we tried to do was put them together in a 18 framework so that we can look for similar lines of inquiry. 19 Where we start is at the beginning of the day when there were 20 clearly some external shocks in the market or signals. 21 Whether or not they were real or whether or not they were 22 perceived remains to be seen, but for certain it was a 23 jittery day in the stock market. 24 Different market participants behaved differently 25 under these circumstances. Here we list four: market 89 1 makers, retail investors, institutional investors, algorithm 2 of traders, each with their own decision making process and, 3 most importantly, each with their own time horizon for making 4 those decisions. 5 The bottom part of the chart shows that once those 6 decisions are made, they are made under different venues, and 7 here we list three different markets just to give an example. 8 Those markets represent different categories. So one can 9 think about the market as being different exchanges. You 10 have Exchange 1, Exchange 2, and Exchange 3, and one exchange 11 will be affected by what happens on the other, whether it be 12 through self-help or through LRPs. 13 You can also think about the markets as being 14 different venues for how to think about the trading 15 themselves, derivatives markets, futures markets, cash 16 markets, et cetera. 17 When you have rebalancing, what happens in the cash 18 markets happens in the futures markets; what happens in the 19 futures markets happens in the derivative markets, and it 20 becomes a bit of a circle. So there's a lot of intermarket 21 feedback. 22 Even the order types themselves have a bit of a 23 feedback as we've heard about today where if the market goes 24 down and you invoke a stop order, that can drive the market 25 down even further and you have a bit of a cascade. 90 1 So we're thinking about all of these problems 2 within the framework that's listed here, and the important 3 thing is when we start to think about how these market 4 changes and the propagation of market signals creates a 5 feedback loop that then exacerbates the situation even 6 further. 7 When we think about feedback, we think about folks 8 changing their decision based on information that then 9 changes their decision again and again, and so forth. 10 Because of the time scales involved, it is unlikely that 11 investors are making real time decisions or even that traders 12 are making real time decision by watching the prices. These 13 things happened at the second or the sub-second level. 14 So to the extent that decisions were being 15 triggered by changes in prices, this most likely happened at 16 the algorithmic side of things where things can happen much 17 faster, or they happened in a pre-programmed way through stop 18 loss orders, market orders, et cetera, that were triggered 19 automatically. But we're looking less at the instances where 20 folks have made their own trading decisions based on watching 21 a millisecond by millisecond feed. 22 MR. COOK: Thanks, Gregg. 23 So within that overall context, we've identified on 24 the next Slide 49, some general themes that pick up on some 25 of the discussion we've been having and areas that we want to 91 1 continue to explore at a very high level. 2 You know we in the first instance plan to examine 3 further the types of activities that might suddenly generate 4 enormous demand for liquidity to buy. We spoke earlier about 5 the linkages between the futures markets and lots of 6 discussion today about the E-mini on the S&P 500, and 7 linkages between those and the cash markets and the extent to 8 which changes in one may have driven the other, which is 9 particularly important given the price discovery interplays 10 that are going on. That's going to be a continuing area of 11 focus. 12 We'll also be looking at whether there are other 13 types of trading behavior that could have contributed in 14 varying degrees to the downward price pressure. For example, 15 to what extent were firms employing directional strategies 16 that were triggered by signals that attempt to exploit 17 short-term price movements? This is a type of trading 18 activity that the Commission has raised in connection with 19 its concept release on equity market structure and one that 20 we'd like to understand more in terms of the events of May 21 6th. 22 Short selling is also another area we'll be looking 23 at as we talked about before, and while the overall short 24 selling on May 6th for the day doesn't seem to account for a 25 disproportionate percentage of trading volume, we need to 92 1 examine further what was really going on in the 20 minute 2 period, and as you will recall from the slides on the stub 3 quotes, there were a lot of short selling trades that came up 4 at that level. 5 So these are some of the broader types of issues 6 we're going to be looking at to try to isolate what some of 7 the triggers, or that may or may not be the right word, but 8 what some of the key instigators were of the overall decline 9 in prices. 10 We'll also be looking much further into what 11 happened to liquidity, studying particularly the activities 12 of market participants who normally provide liquidity in the 13 marketplace. As we've said, based on anecdotal evidence, 14 there's some suggestion that a large number of trades 15 executed against stub quotes -- sorry -- based on anecdotal 16 evidence and the stub quote trades, it would appear that some 17 professional liquidity providers temporarily didn't 18 participate in the market on the buy side, and we need to 19 understand how these firms are acting. 20 They may have been acting appropriately under 21 current rules, but we need to look at the data and determine 22 whether they withdrew from the market, to what extent, and 23 why, and as we've also been discussing, we need to look at 24 issues of potential trap liquidity due to different trading 25 rules or conventions across different market venues as we've 93 1 been talking in terms of LRPs and the exercise of self-help, 2 whether that was appropriate under the circumstances or not. 3 There's also some emerging evidence that large 4 internalizers may have ceased providing executions. We've 5 been talking about the exchanges as the key execution venues, 6 but if you go back for a second to the previous slide when we 7 were looking at different market types, you should also be 8 thinking of large internalizers of orders as a way, another 9 type of market, a very significant market out there for the 10 execution of orders and their activities or their changes in 11 their behavior during this time could well have had a 12 significant impact on the way that orders, particularly 13 retail orders, were affected during this time. 14 A third area of further inquiry will be the role 15 that different order types play. I think we've talked a lot 16 about this already. 17 A fourth will be understanding the experience of 18 ETFs. We've talked about the need to look at the behavior of 19 market makers in ETFs and whether their ability to hedge in 20 the underlying securities may have affected their willingness 21 to trade the ETFs and at what prices. We're going to be 22 looking at the creation of redemption process and to see what 23 evidence we can find of any variation in creation and 24 redemption activity that might be related to these events. 25 We also need to examine further whether the use of 94 1 ETFs by institutional investors, who often use ETFs to 2 quickly acquire broad market exposure, whether that led to 3 any selling pressure on the ETFs as the market began to 4 decline, and of course, we'll also be looking at whether the 5 exercise of self-help against Arca was a factor in the 6 experience of ETFs because many ETFs are listed on the Arca 7 market. 8 And then finally, we'll we keeping our 9 investigation open to considering other contributing factors 10 that may be identified to us as we continue our research and 11 analysis. For example, there were some questions today about 12 the role of the kind of systems and how they work together, 13 and whether there was any latencies in the messaging traffic 14 that may have contributed to this in one way or another would 15 be something we want to understand both for purposes of 16 figuring out to what extent that contributed to what we 17 observe, but also for purposes of forming policy responses. 18 And just finally we'd like to spend a minute. 19 We've been alluding for a while to the issues of data and 20 just want to describe once again sort of what the context is 21 in which we're trying to gather data. 22 And I'm going to turn this over to Gregg to 23 describe. This is Slide 51 in your materials. 24 MR. BERMAN: So I thought we'd just take a minute 25 to trace the history of a trade and to get an idea of the 95 1 type of data that we need in order to reconstruct the events 2 of May 6th and the data that we've analyzed so far. 3 So the trades that are indicated in blue at the 4 center of the chart, these are the trades that do happen on a 5 millisecond by millisecond level. They come across the 6 ticker; they come across the consolidated tape, and we've had 7 access to them. They generally come from a single source of 8 information that is supported by the New York Stock Exchange 9 called TAQ: trades and quotes, and we've spent most of our 10 time analyzing the data in that database. 11 In addition, there are two other databases that are 12 part of that same suite of offerings. One shows the national 13 best offer and the national best bid, and we've also been 14 able to absorb that information, line it up with the trade 15 information, and produce the charts and the graphs and 16 analysis that we've showed so far. 17 Yet a third database gives you the best bid and 18 best offer on an exchange by exchange basis, and we've been 19 able to analyze some of that information, and part of that 20 information has been part of our analysis so far. So trades, 21 quotes, both at the national level and the exchange level, 22 all in three different databases, different formats, 23 different systems and require different ways of merging that 24 information. 25 What we have not analyzed yet is the order book 96 1 that sits behind each exchange's best bid and the exchange's 2 best offer. Those books are built by events that each 3 exchange keeps track of. Now, the exchanges themselves need 4 to do this in real time because they need to determine 5 whether or not there's a match and whether or not they have 6 the best bid or the best offer, and if they're going to 7 reroute or if they're going to accept that and do it 8 internally. 9 In order for us to reconstruct that, we almost have 10 to reconstruct our own mini version of a trading floor and 11 pretend to be an exchange for the day, taking in all the 12 different orders and the order book. So that is something 13 that we have started engaging and will take some time to 14 complete that. 15 One thing to note is that we only have the order 16 book for or the potential to see the order book for the 17 exchanges. Internalizers we would not have any information 18 on that, and those are shown in dotted lines, and we need to 19 do this both for the offer book as well as for the bid book. 20 The orders themselves and how they were processed, 21 whether or not that was a limit order, a market order, 22 contingent order, et cetera, that comes yet in a separate 23 feed from what's called audit trails as opposed to the order 24 book, and those are kept by the exchanges also in different 25 formats, in different databases, and get you back to the 97 1 introducing broker where the trade was originally introduced 2 and how it originated. 3 Since there are multiple exchanges, we will be 4 collecting and have received some of that information already 5 and build that into the analysis to see who is actually 6 selling what and why. But we'll get more of the why in terms 7 of the nature of whether it was a limit order than the who 8 because the who only brings us back to the broker. 9 If we actually want to go back to the original 10 decider of the trade, we then have to go to each broker and 11 request what are called blue sheets, and in this database, 12 and each broker will have their own forms of this, we will 13 find some of the identifiers for whether or not it was a 14 hedge fund or an asset manager or a retail client. 15 The data is not designed across the board to be 16 collected in a massive scale. So the names of the 17 participants will only be as good as the way that they are 18 known by the accounts to the brokers. So the same hedge fund 19 can show up under five different names, and we may have to 20 try to piece together some of this information. 21 What's not shown are some of the colocators, folks 22 who sponsored access who we would not have direct information 23 necessarily from the blue sheets to understand exactly the 24 orders and when they were placed. 25 MR. COOK: Thanks, Gregg. 98 1 So I think the overall point is that what's within 2 the dotted line on this chart is stuff that we can -- is data 3 that we can have ready access to at any given time, and as 4 you move farther away from that, you're talking about pulling 5 in data from various sources that may not be maintained in a 6 consistent format and that takes varying degrees of time to 7 get in house and then has to be validated and made consistent 8 in order to be able to be analyzed. 9 That's the world that we live in today. The SEC 10 has a meeting scheduled for Wednesday to propose rules that 11 would implement what we are calling a consolidated audit 12 trail, which would really be designed to address precisely 13 this point, to give us the ability and the exchanges in their 14 regulatory capacity, the ability to very quickly get 15 information about orders from the point of origination all 16 the way through their life cycle, through execution in a 17 standardized format. It's a major undertaking, as you can 18 imagine, given the complexities involved, but it would move 19 us substantially forward in being able to undertake the type 20 of analysis that we'd all like to undertake quickly in 21 connection with an event such as May 6th. 22 Thank you. 23 MR. SHERROD: I just want to echo some of the 24 difficulties Robert has we also have on the futures side, but 25 in many respects we have it much easier in that the E-mini 99 1 trades exclusively on the CME. So that data source is 2 confined to one exchange. 3 As Gregg mentioned though, we, too, have to do 4 inquiries to the clearing members and the carrying firms to 5 find out the names of the accounts. We have a large trader 6 reporting system where we receive the names of the accounts at 7 the end of the day, but for the intraday trade register, we 8 have to make reverse inquiries as does the SEC to find out 9 the accounts that are being executed through either the CME 10 platform or through the ICE platform. 11 I'd like to turn to Rick for the next steps on our 12 analysis. 13 MR. SHILTS: Yes, and this will be slide 53. 14 While the CFTC staff has been able to gain insights 15 into the events that occurred on May 6th, staff will continue 16 to review other information that it collects or will collect 17 as it relates to this market. Specifically, staff will 18 continue reviewing information from a special call on over 40 19 traders for their trading activity in the E-mini S&P 500 and 20 Russell 2000 futures contracts on May 6, 2010. 21 A special call is a CFTC directive to a trader 22 holding a reportable position to furnish any pertinent 23 information concerning the trader's positions, transactions or 24 activities. 25 Staff also will continue reviewing information from 100 1 a special call on swap dealers about their activity in 2 over-the-counter broad based security index derivative 3 markets on that day. 4 Staff also will continue its detailed review of 5 trader activity on May 6th through an examination of trade 6 register data. To date staff has reviewed over 25 gigabytes 7 of data in over 307,000 files related to the individual 8 trades that occurred that day, with more data expected. 9 In addition to these areas of analysis specific to 10 May 6th, staff at the CFTC continues to review electronic 11 trading and its effect on liquidity provision in the futures 12 market. Specifically, staff is focused on the practices of 13 high frequency trading and algorithmic trading. 14 We are also carefully reviewing pre-trade automated 15 safety features. 16 Now, turning to the next Slide 54, in addition, 17 both SEC staff and CFTC staff working together will review 18 correlated assets and equities including single stocks, 19 mutual funds, and ETFs, as well as options in the futures 20 market. The study would partly focus on examining 21 cross-market linkages by analyzing trading and stock index 22 products, such as equity futures, ETFs, equity index options, 23 and equity index OTC derivatives using, to the extent 24 practicable, market data, special call information and order 25 book data. 101 1 Now I'll return it to Robert. 2 MR. COOK: Thank you. 3 We'll be happy to take any discussional questions 4 or comments you might have. 5 MR. KETCHUM: I think the first basic comment is 6 that looking through on each of the points, you've really 7 anticipated a number of my questions before, and I think 8 you're looking at the right things. 9 Gregg, I would suggest from experience of having 10 tried to put things together in the fall of '08 that first 11 you have my deepest sympathy, and secondly, wherever you can 12 as you move through the blue sheet information and piecing 13 together the other pieces of the order audit trails, et 14 cetera, where you can focus on flex moments and very short 15 pictures of this rather than trying to recreate the whole 16 period at least as a first step I think you may dramatically 17 accelerate your time to be able to do some thinking about it 18 or for us to be doing some thinking about it. 19 So I would suggest you think about at least doing 20 it in stages because the data is truly overwhelming. 21 Robert, on the point you made you made with respect 22 to latencies, one latency over and above the ones I think you 23 mentioned that I think would be of value to look at is the 24 latency of different lines exchanges and internalizers are 25 using to make their self-help decisions since some use common 102 1 lines, some use direct lines to different exchanges, and it 2 would be very interesting to understand whether the aberrant 3 use of self-help during that day had something to do with 4 latencies in lines as opposed to latencies in Arca. 5 I guess the last point I'd make, which is not to 6 suggest you recreate this in looking at other periods, but I 7 think we can overdo to some degree from a precision 8 standpoint between this particular experience which more 9 underlines the brittleness in the markets we need to address 10 than necessarily the one and only experience. 11 I think it would be valuable to at least on a broad 12 based way compare what happened on May 6th with what happened 13 in the fall of '08, which is the other time in which 14 unprecedented large numbers of trade cancellations occurred 15 in the equity markets, albeit in a time far more dire than 16 the news that you've described on the 6th. 17 And the last piece that would be helpful for you to 18 go back on particularly the extent this Committee is going to 19 consider things and as both agencies have indicated they plan 20 to consider the questions of system-wide circuit breakers as 21 well as any pauses in the market. There was one or two 22 events far earlier before system-wide circuit breakers were 23 expanded where those circuit breakers were hit. There was a 24 good deal written and analyzed vis-a-vis gravity impact and 25 concerns with respect to lower level circuit breakers, and it 103 1 would be great if the staff could go back and pull together 2 the analysis that was at least done at that time. 3 CO-CHAIR SCHAPIRO: Anyone on the phone have 4 anything further? 5 MR. ENGLE: Just one comment. This is Rob Engle. 6 When you look at the linkages between derivatives 7 markets, you're talking about options makers who are 8 presumably continuously hedging and some of this event could 9 actually spill into the options market in a way which we 10 haven't seen yet. But it would be interesting to understand 11 how those feedbacks work because they're probably working in 12 the same direction. 13 MR. COOK: Yes, we agree, and that's something we 14 intend to look at. When we think of the derivatives markets, 15 it won't just be the futures markets. It will also be the 16 options markets, and to the extent we can the OTC markets as 17 well. 18 CO-CHAIR GENSLER: I want to ask Professor Engle. 19 So earlier you made a very good suggestion about 20 looking at other periods of volatility, and Mr. Ketchum 21 suggested the fall of '08 or whatever volatile periods I 22 thought, but was it those periods that you were thinking 23 about, Professor Engle, to look back at the volatile periods 24 like in the crisis? 25 MR. ENGLE: Yes. 104 1 CO-CHAIR GENSLER: Or some other periods? 2 MR. ENGLE: No, I think it would be interesting to 3 see periods in the crisis, but also we've had shorter 4 intermittent periods of high volatility, of course, since 5 then that perhaps might more resemble this event, and so 6 maybe in less detail but more scope would be useful. 7 CO-CHAIR GENSLER: Great. So we also want to 8 recognize and set aside time for our respective 9 Commissioners. So I don't know which way I'm supposed to do 10 this, but I'll turn left first. We usually do it by 11 seniority, but Commissioner Dunn. 12 COMMISSIONER DUNN: Thank you, Mr. Chairman. 13 And thank the great panel we've got here and the 14 hard work that the staff has done on putting this together. 15 When I was acting chair of the Commission, I had indicated 16 that I thought we needed to have joint meetings with the SEC 17 and to look at harmonization of our regulations and to look 18 at risk that transcends both agencies, and I think this is an 19 important step. 20 I note that it took an act of Congress to be able 21 to do that, and this is probably in looking at the makeup of 22 this Commission and the start that we've got here, this may 23 go down as one of the more successful things that this 24 session of Congress has done. 25 I would also, Mr. and Madam Chairman, be remiss if 105 1 I didn't tell you what a great job you two are doing in 2 working together on harmonization. There's hardly a week 3 goes by that I don't hear from someone saying that, gee, what 4 great cooperation the Gary and Mary team are doing, and I 5 think it bodes well for both of our Commissions on doing 6 this. 7 I was doing some research on business process 8 reengineering getting ready to look at what Congress is going 9 to lay in our lap as we came up, and I came up with this 10 quote from one BPR group, and it said, "The problem is that 11 we're governing in the 21st Century with processes and 12 organizations designed in the 19th Century to work well in 13 the 20th Century." 14 And their results were they thought they needed a 15 new process and organization for governance in the 21st 16 Century, and I think this group is putting us well on that 17 way. Thank you again for your help. 18 CO-CHAIR GENSLER: Mike, thank you very much. 19 I normally call it the Mary and Gary show, but you 20 know. 21 But Commissioner Walter. I don't remember your 22 seniority. So I just want to make sure I have it in the 23 right order. 24 COMMISSIONER WALTER: It doesn't matter. 25 CO-CHAIR SCHAPIRO: Okay. All right. 106 1 COMMISSIONER WALTER: I would like to add my thanks 2 particularly to the staff for their tireless efforts and just 3 note in addition to our upcoming meeting on Wednesday I think 4 this shows the important of cross-market information being 5 available to regulators. It is very important to me that as 6 we go forward, for example, with respect to the OTC markets, 7 that regulators have full information regardless of 8 jurisdictional lines, and I think this exercise underscores 9 that. 10 So my thanks to everyone in the room, particularly 11 those of you who have volunteered in some way, shape or form 12 to work with us. 13 CO-CHAIR GENSLER: Commissioner Sommers. 14 COMMISSIONER SOMMERS: Thank you. 15 I also want to thank both Chairman Gensler and 16 Chairman Schapiro for pulling this Committee together. I 17 have found the dialogue today to be enormously helpful. 18 And thank you to the staff for all of your work on 19 this report, and I just want to say that I look forward to 20 your continuing analysis and the information that we're going 21 to get on your further steps. 22 Thank you. 23 CO-CHAIR GENSLER: Commissioner Aguilar. 24 COMMISSIONER AGUILAR: Thank you. 25 Good morning. First I want to apologize for 107 1 arriving late this morning, but months ago I had committed to 2 keynote the start of the 2010 Compliance Week conference this 3 morning. I certainly understand the importance of today's 4 meeting. 5 And I, too, would like to thank members of the 6 staff of both Commissions for their hard work, which is 7 evident, and I also want to thank the members of the Joint 8 Advisory Committee for their willingness to give their time 9 and expertise to these efforts. Thank you. 10 And certainly not least, I also want to thank our 11 respective Chairmen, the Mary and Gary show, and my fellow 12 Commissioners at the SEC and the CFTC for agreeing to form 13 this Committee. 14 The market disruption of May 6th and the inability 15 to promptly analyze and fully understand it has created a 16 sense of urgency and certainly deserve your and our 17 attention. It is essential that we respond to the events of 18 May 6th and address the issues they raise. Today's agenda 19 was a step forward in that discussion. 20 As we undertake this important work, however, we 21 must also remain focused on the broader issues related to the 22 financial markets. As Committee members, you have been 23 called together at a time of historical significance. It 24 should not be lost on anyone that the essential role of the 25 financial services sector is to facilitate the allocation of 108 1 capital to productive uses. The financial crisis revealed a 2 clear failure of markets and the financial sectors to serve 3 this role with widespread mispricing of assets, trillions in 4 losses, and damaging levels of unemployment and under 5 employment, together with an unprecedented concentration of 6 wealth at the top. 7 As this Committee does its important work and 8 prepared recommendations for the SEC and the CFTC, I ask that 9 you remember that our markets must serve the public interest 10 by facilitating the real economy and sustainable shared 11 prosperity. 12 Thank you. 13 CO-CHAIR GENSLER: Thank you, Commissioner Aguilar. 14 Commissioner Chilton. 15 COMMISSIONER CHILTON: Thanks. 16 There is an old saying in Washington. When in 17 doubt set up a task force or a committee, and in this case I 18 think it's really needed and not just on the May 6th stuff, 19 but for these emerging issues. so we can be more nimble and 20 quick and look around the corner. I think it's very helpful 21 to have this Committee. 22 I really hate it when in Washington we use 23 this -- and nobody has used this particular phrase 24 here -- but you know, this perfect storm stuff. I mean, I 25 think it sort of says we don't know what happened. Just all 109 1 of these different things, and in listening to staff, who I 2 know have worked really hard and done yeoperson's service, 3 I'm reminded -- remember Colombo? He would say, "And one 4 more thing," or, "if you could just explain this also," as he 5 was walking out the door, and in the meetings that we've had, 6 and I assume my fellow Commissioners in all of our 7 surveillance meetings, I keep feeling like there's one other 8 question I want to ask and one more thing. 9 But in listening to staff, you have those 10 questions, too. I mean, you're going to get more 11 information. You're going to look at the algo trading. 12 You're going to look at the flash trading. You're looking at 13 ETFs, and to me it's sort of a microcosm in part at least of 14 the entire economic calamity that we faced. You could make 15 arguments about why do we end up, you know, in 2008 and 2009 16 in this circumstance, legislation, regulation or lack 17 thereof. 18 But there are lots of questions that we still don't 19 know the answers to, whether or not it's with regard to the 20 OTC markets, you know, 600 trillion out there. Our futures 21 markets are only 5 trillion. It just seems to me we've got a 22 lot of questions, and one more thing out there. 23 So my take-away from this is it's great we're doing 24 this. I look forward to getting more information, but 25 overall we need financial regulatory reform. We need that 110 1 bill passed that's going to not have going to seek additional 2 information as part of as special committee, as part of a 3 special call or all of these inside the beltway phrases. We 4 need to constantly get this information, have the regulatory 5 tools and the funding in order to effectively implement them 6 and insure that we have efficient, effective markets that are 7 devoid of fraud, abuse and manipulation. And one more thing: 8 that we protect consumers above all. 9 So thanks very much. 10 CO-CHAIR GENSLER: Thank you, Commissioner Chilton. 11 Commissioner Paredes. 12 COMMISSIONER PAREDES: I just want to add my thanks 13 to everybody who has been working literally around the clock. 14 On the staff, this has been an incredible amount of work and 15 an incredible degree of effort on your part and an incredibly 16 impressive showing to this point. 17 We know there's still a lot of work to be done, and 18 I certainly look forward to the continued efforts in terms of 19 getting and analyzing the data and seeing where the data 20 ultimately takes us. 21 I also want to extend my thanks to the members of 22 the Advisory Committee for their sacrifices and all the 23 efforts that they'll be making in the coming months to help 24 us analyze and understand and figure out appropriate 25 responses. 111 1 And of course, to Mary and Gary for their 2 leadership. 3 CO-CHAIR GENSLER: Thank you, Commissioner Paredes. 4 Commissioner O'Malia. 5 COMMISSIONER O'MALIA: Chairman Schapiro, thank you 6 for hosting this inaugural event. I'm pleased that we have, 7 the Commission, appointed the distinguished group of experts 8 from the securities and future industry in order to advise 9 our Commissions on possible solutions to May 6th. I 10 appreciate their willingness to serve and hopefully we'll 11 have some solid recommendations that the two Commissions can 12 implement. 13 Based on first-hand information I've received in 14 the last several days, I encourage the committees to conduct 15 a more thorough investigation of the significant delays in 16 processing and clearing of trades caused by the overwhelming 17 volumes in the futures markets. 18 In a lot of this information I believe that the 19 statement on page 7 of the staff report which reads, 20 "Clearing and settlement processes worked efficiently, 21 effectively, and without incident significantly understates 22 the challenges created by the flash crash in the clearly and 23 settlement business." I've learned that thousands of trades 24 were unconfirmed long after trading stopped on May 6th. 25 On May 7th temporary shutdowns occurred in clearing 112 1 queues resulting in the backlogs of hundreds of thousands of 2 trade messages that lasted over the weekend. Without proper 3 confirmation of trades, FCMs and clearing members were unable 4 to accurately mark their books or set appropriate margin 5 levels with the highest degree of confidence. 6 I hope the Committee will, working with the staff, 7 will further investigate this matter and report back to the 8 Commissions regarding the impacts of the clearing and 9 settlement business as a result of the extreme trading 10 volume. 11 I also am interested to know if the futures 12 exchanges have implemented a crisis strategy in order to 13 respond to the next spike in trade volume. 14 I am particularly interested in final 15 recommendations from this panel regarding technology specific 16 issues that the newly established CFTC Technology Advisory 17 Committee can explore in the futures industry. 18 With respect to time constraints, I have provided 19 some written lists of concerns which I hope the Committee 20 will consider in their deliberations. 21 Finally, I'd like to thank the hard work of the 22 staff, both the SEC and CFTC and to produce a very thorough 23 analysis of the events on May 6th. 24 Thank you. 25 113 1 CO-CHAIR SCHAPIRO: Thank you all. 2 We have one Committee organizational item to 3 address before we recess the meeting. So you'll have to bear 4 with me. We actually have to approve our bylaws. 5 Briefly, the bylaws specify that the Joint 6 Committee will be jointly presided over by Chairman Gensler 7 and me or by our designated federal officers when directed to 8 do so. 9 In addition, the bylaws require that with respect 10 to voting a Committee member must be participating in a 11 meeting in person or by telephone or similar communication to 12 cast a vote. When a decision or recommendation of the Joint 13 Committee is required, the presiding officer will request a 14 motion for a vote. Any member may make a motion and vote on 15 that motion. No second is required. 16 Committee action based on a vote requires approval 17 of a simple majority of the votes cast at a meeting at which 18 a quorum is present. 19 I think a lawyer wrote all of that. 20 Before voting on this item, can I ask for a motion 21 to approve the bylaws? 22 PARTICIPANT: So moved. 23 PARTICIPANT: Second. 24 CO-CHAIR SCHAPIRO: Is there any discussion on the 25 bylaws? 114 1 (No response.) 2 CO-CHAIR SCHAPIRO: All in favor. 3 (Chorus of ayes.) 4 CO-CHAIR SCHAPIRO: Any opposed? 5 (No response.) 6 CO-CHAIR SCHAPIRO: Thank you. The bylaws have 7 been approved. 8 Let me again thank the Commissioners of both 9 agencies for their presence today at the Joint Committee 10 meeting. I also again want to thank the staffs of the SEC 11 and the CFTC for their excellent presentations and analysis 12 today. I know how many Saturdays and Sundays I've spent on 13 the phone with all of you, and you are true public servants, 14 and I am incredibly grateful to you. 15 You've done an incredible job analyzing data, 16 presenting preliminary findings in such a short period of 17 time, and we are very, very grateful. 18 I also want to thank the Advisory Committee members 19 for reviewing and digesting the joint report in a compressed 20 period of time and for offering today your insights and 21 recommendations on how we might go forward. 22 We will set up the next meeting of the Joint 23 Committee shortly. You've clearly given us some homework to 24 do and some additional data and analysis that will be useful 25 to you as you continue your work. 115 1 So with that, Chairman Gensler, I don't know if you 2 have any more comments? 3 CO-CHAIR GENSLER: I just again wanted to thank my 4 fellow Commissioners and the Commission of the SEC and 5 Chairman Schapiro for pulling this together and our Advisory 6 Committee members on short notice -- Susan Phillips reminded 7 me I called her on a Saturday evening to ask whether we could 8 announce her name on Monday. It was very, very generous of 9 you -- for pulling this together. 10 I can tell by the quality of the questions and the 11 dialogue today we as two Commissions are going to get a great 12 deal form this Advisory Committee. 13 And lastly, I want to thank the staffs again that 14 really, as Chairman Schapiro has said, has shown what public 15 service can be about. It is really great standards. It is a 16 high bar you have set though. 17 CO-CHAIR SCHAPIRO: So with that, the open meeting 18 of the Joint Committee will recess to allow for the Committee 19 members to have lunch and discuss administrative issues such 20 as the ethics rules and travel forms that you have to file 21 with us and public record requirements, and following that 22 administrative discussion the meeting of the Joint Committee 23 will adjourn for the day. Thank you all very much. 24 (Whereupon, at 11:54 a.m., the Joint Advisory 25 Committee meeting was adjourned.) 116 1 PROOFREADER'S CERTIFICATE 2 3 In the Matter of: JOINT CFTC-SEC ADVISORY COMMITTEE 4 ON EMERGING REGULATORY ISSUES 5 File Number: 265-26 6 Date: Monday, May 24, 2010 7 Location: Washington, D.C. 8 9 This is to certify that I, Don Jennings, (the 10 undersigned), do hereby swear and affirm that the attached 11 proceedings before the U.S. Securities and Exchange 12 Commission were held according to the record and that this is 13 the original, complete, true and accurate transcript that has 14 been compared to the reporting or recording accomplished at 15 the hearing. 16 17 18 19 ___________________________ _____________________________ 20 (Proofreader's Name) (Date) 21 22 23 24 25 117 1 REPORTER'S CERTIFICATE 2 3 4 I, Jon Hundley, reporter, hereby certify that the foregoing 5 transcript of 115 pages is a complete, true and accurate 6 transcript of the testimony indicated, held on May 24, 2010, 7 at Washington, D.C. in the matter of: Joint CFTC-SEC 8 Advisory Committee on Emerging Regulatory Issues. 9 10 11 I further certify that this proceeding was recorded by me, 12 and that the foregoing transcript has been prepared under my 13 direction. 14 15 16 17 Date: __________________________ 18 Official Reporter: __________________________ 19 Diversified Reporting Services, Inc. 20 21 22 23 24 25