Speech by SEC Staff:
Keynote Address to the 2009 SIFMA Market Structure Conference
James A. Brigagliano1
Co-Acting Director, Division of Trading and Markets
U.S. Securities and Exchange Commission
New York, New York
May 20, 2009
Thank you for inviting me here today to kick off this conference on securities market structure. I need to begin by reminding you that the views I express today are my own and not necessarily those of the Commission, the individual Commissioners, or my colleagues on the Commission staff.
One of the important lessons of the financial crisis of recent months is that the structure of a financial market matters, and that it matters most in times of economic stress. So this morning I want to begin by spending a few minutes looking back and assessing how the US securities market structure has performed during the financial crisis. My primary focus, though, will be to look forward and highlight a few market structure issues that I believe warrant the attention of industry and regulators in the coming year.
This conference is focused on market structure. I think we all can agree that the reason for existence of a market is to bring buyers and sellers together and enable them to trade as efficiently as possible. And a pretty good definition of "structure" is the manner in which the parts of a complex entity are organized and work together to achieve the purpose of that entity. So how have the individual parts of the complex markets for US-listed equities and options worked together during the financial crisis? Did they achieve their purpose of enabling buyers and sellers to trade efficiently through extraordinarily difficult economic conditions?
I am pleased that the structures of the US-listed securities markets held together and performed well when the chips were down. The individual parts of the securities markets, of course, are all of the intermediaries that play essential roles in their operation, from brokers to dealers to market makers to ATSs to exchanges to clearing entities. In the face of the sharp spikes in volume and volatility of the last year, the markets for US-listed securities have remained open and transparent. They continued to operate in a fair and orderly manner and perform their vital price discovery function. Buyers and sellers could see current prices and expect to execute their trades promptly at the prices they saw on their screens. Just as importantly, buyers and sellers could have confidence that their trades would be completed - that they would be processed, cleared, settled, and performed - in a reliable manner. This positive performance of the US-listed securities markets is a credit to the many individuals throughout the industry who are responsible for the highly complex systems and procedures necessary to operate modern markets. And we should never forget that "operability" is a key component of market confidence. One hears a lot of talk about shoring up market confidence, and I believe we all agree it is vital. Well, we can take some comfort that the first building block of such confidence is a market structure foundation that did not waver though the ground beneath it shook.
The markets for some other, non-listed financial products, of course, did not perform as well over the last year, in significant part because of a lack of transparency. In this respect, the financial crisis offers yet another opportunity to re-learn a lesson that seems easy to forget with time. In good economic times, market structures may not seem to matter as much. It is in tough times, when buyers and sellers need them most, that markets need to be well organized and properly regulated to fulfill their reason for existence. When it comes to managing financial risks, there truly is no substitute for financial products traded in transparent markets capable of generating price discovery that properly incorporates the risk of those products. Reliable markets with good prices enable all interested parties - from traders to clearing entities to regulators - to appreciate financial risks and act accordingly to protect themselves and the broader financial system.
Of course, one of the hallmarks of an efficient market structure is that it must fit the particular economic and technology environment in which it operates. Yet economic conditions and technologies are continually changing. Indeed, the securities markets have thrived as competitive forces have led entrepreneurial industry participants to innovate with new technologies and new trading tools. So the challenge for the industry and regulators is to monitor these changes and update a market's structure when needed. I want to spend the rest of my time this morning focusing on a few market structure issues that I believe would be a helpful part of the dialogue between the industry and regulators in the coming year.
The first issue is sponsored access to exchanges. As you know, the Commission staff has been gathering facts and reviewing issues relating to sponsored access for some time. Broadly speaking, sponsored access can be defined as an exchange member providing its customers with electronic access to the exchange using the member's identifier. Under such an arrangement, the customer is provided a dedicated line or port to the exchange's execution system, so that its orders do not first pass through the member's systems before reaching the exchange. As electronic trading has become so prevalent in recent years, this type of access to exchange execution systems, which often is unfiltered, has increased significantly. The appeal of the arrangement, of course, is that it helps preserve anonymity and enables the fastest possible trading. And for some of the most sophisticated, high-volume firms in today's trading environment, every microsecond counts.
There are, however, a variety of risks involved when trading firms have unfiltered access to the markets. These risks can affect many of the participants in a market structure, including the trader's broker, the exchanges, and the clearing entities. Ultimately, the risks can affect the integrity of the market structure itself. To address these risks, the Commission staff has worked with the SROs to develop a consistent approach to sponsored access. A few months ago, Nasdaq proposed to amend its rules to adopt such an approach.
I believe that the Nasdaq proposal has sparked a very useful public debate on sponsored access. Many commenters offered detailed discussions of the respective benefits and risks of various types of sponsored access arrangements. SIFMA's comment letter, for example, emphasized the importance of efficient market access to implement electronic trading strategies. The letter also laid out a "disaster scenario" in which a single trading firm's activity could lead to an enormous volume of failed trades that might overwhelm the financial capacity of many key market participants.
The Nasdaq proposal remains under Commission review so I cannot comment on the issues it raises. I do want to emphasize, though, the vital importance of dealing with these issues effectively and as soon as reasonably possible. I believe we have learned that dangerous financial risk conditions must not be ignored, even when they may appear difficult to resolve. Although it may be difficult or even impossible to foresee the particular event that triggers a systemic threat, the conditions that create a potential for such a threat often are foreseeable. With respect to sponsored access, the Nasdaq proposal and the views of commenters have greatly contributed to a better understanding of important issues among the industry and regulators. I believe that this improved understanding will pay off as sponsored access is addressed in the coming months.
I next want to focus on a group of issues relating to dark pools. In February of last year, SIFMA sponsored a symposium exclusively devoted to dark pools. Erik Sirri - then the Director of the Division of Trading and Markets - gave the keynote address. His remarks were intended to clarify some factual issues surrounding dark pools and to identify regulatory concerns that were worth watching. These included transparency, fragmentation, and fair access. Today's conference provides a good opportunity for an update on these concerns. I should emphasize, though, that the staff certainly has not reached any conclusions on whether a regulatory response is needed or, if so, what sort of response would be appropriate.
As Erik noted last year, dark pools can be defined in various ways. For our purposes today, I will just define them as electronic trading systems that do not display quotes in the consolidated public quote streams. The public quote streams are distributed by central processors that act on behalf of all the SROs in making market information widely available to the public. Over the last year, dark pools, which do not display public quotes, appear to have increased their percentage share of total trading volume in US-listed stocks, though that growth seems to have leveled off recently. Time will tell whether dark pool volume growth in fact has slowed permanently, or whether this is merely a temporary interlude due to unusual trading conditions since September 2008. I should note that the percentage of volume at other types of dark trading venues, such as OTC market makers, may have increased in recent months.
I would like to give you specific statistics on the trading volume of various dark pools, but there is very little reliable public information on dark pool trading activity. This lack of public information illustrates a transparency concern that warrants attention. Dark pools report their trades to the consolidated public trade streams, but their trades are identified merely as non-exchange, or OTC, trades. The public trade reports do not identify whether an OTC trade was reported by a dark pool and, if so, the identity of the dark pool.
Because of this gap in post-trade transparency, it can be difficult for the public to assess dark pool trading volume and evaluate which ones may have liquidity in particular stocks. Although many dark pools publicize volume statistics on their web sites, they do not use a uniform methodology and may effectively overstate their true executed volume because of double counting - counting both the buy and sell side of a single trade - or by including "touched" volume of orders routed elsewhere for execution with the "matched" volume that they actually execute. I am concerned that this state of affairs may not promote public confidence in the equity markets.
In deciding whether or how to address this matter, one issue to consider is whether there is any compelling reason for dark pools to object to improved post-trade transparency. While full pre-trade darkness is an important element of the business model of some dark pools, it does not appear that some form of improved post-trade transparency would be likely to interfere with their business models. Indeed, uniform and reliable trade reporting practices could help establish a fairer playing field because those dark pools that report their volume accurately would not be disadvantaged in comparison with any that might inflate their volume.
Consequently, one of the issues worth discussing with respect to dark pools is whether their post-trade transparency should be enhanced. Would this type of information be useful to investors and, if so, what is the best vehicle for disseminating the information? For example, would trade-by-trade disclosure by dark pools be most helpful, or would the public availability of more uniform, reliable summary statistics on trading volume be sufficient?
Another regulatory concern with respect to many dark pools relates to their transmittal of pre-trade messages about their available liquidity. Although dark pools do not publicly display quotes, many nevertheless are not entirely dark on a pre-trade basis. In an effort to attract order flow, many dark pools transmit messages to selected market participants notifying the recipient that the dark pool currently has an actionable order in a particular stock - that is, an order that currently is available for automated execution at the NBBO or better if responded to before it is executed or canceled. One of the regulatory issues that Erik Sirri noted last year was the potential for such messages to fall within the regulatory definition of a "quote." Most dark pools label their pre-trade messages as indications of interest, or "IOIs."
For our purposes this morning, I do not want to get bogged down in technical legal definitions, but instead focus on the policy implications of actionable order messages and whether they raise any market structure concerns. Given the speed and sophistication of today's order routing and trading systems, it appears that private actionable order messages are functionally and economically similar to public quotes. The systems are incredibly fast - both actionable order messages and quotes can be transmitted and responded to within a few milliseconds. Although a public quote always will have an explicit price, an actionable order message is implicitly executable at the NBBO or better.
Dark pools that use actionable order messages typically transmit them to networks of selected market participants. As a result, the widespread use of actionable order messages could create the potential for significant private markets to develop that exclude public investors. In particular, these private markets could wind up representing a significant volume of trading based on valuable information on actionable orders to which the public does not have fair access.
Another aspect of actionable order messages to consider is their potential effect on competition among trading centers and market fragmentation. Erik Sirri suggested last year that competitive forces seemed particularly apt to address the problem of fragmented dark pools. He noted that the users of dark pools seemed likely to pressure their operators to consolidate the pools to enable users to check liquidity more efficiently. Actionable order messages, however, are used by the operators of dark pools to alert users when they have liquidity and thereby may weaken this competitive force for consolidation. If actionable order messages enable many small pools to survive separately, market fragmentation could worsen.
The final issue I want to take up this morning is the potential effect of dark pools on the public price discovery function provided by the quoting markets. In particular, if dark pool volume were to increase substantially, could they impair the public price discovery function by diverting valuable marketable order flow away from the exchanges and ECNs that contribute to the public quote stream?
As an initial matter, it may be helpful to distinguish between two major types of dark pool trading mechanisms - block crossing systems and small order systems. Block crossing systems focus on arranging large trades between institutional investors. Their average trade size can be as high as 50,000 shares - much greater than the average trade size in the public markets of less than 300 shares. Block crossing systems thereby offer significant size discovery benefits to their participants, though they currently appear to execute a small percentage of dark pool trading volume.
Small order systems, in contrast, appear to execute the great majority of dark pool volume and to be the fastest growing type of dark pool. They also may have the greatest effect on price discovery in the public markets. In particular, they may attract a significant volume of very desirable small marketable order flow away from the public markets. This order flow could be considered the "cream" of the market if it comes mostly from longer-term investors who may be less interested and informed about very short-term price movements. This type of marketable order flow is desirable for those who supply liquidity through resting, non-marketable orders. To the extent desirable order flow is diverted from the public markets, it potentially could adversely affect the execution quality of those market participants who display their orders in the public markets. The concern here is that any practice that significantly detracts from the incentives to display liquidity in the public markets could decrease that liquidity and, in turn, harm price discovery and worsen short-term volatility.
That is the last of the issues I wanted to raise with you this morning. I know that the industry has a great deal of insight to contribute to regulators on all of these issues and look forward to hearing from you in the coming months. Thank you for listening.