Remarks From the Financial Policy Joint Conference on Market Fragmentation, Fragility and Fees

Regulating in a New Paradigm

It's terrific to be here today. When Jonathan first brought up the idea for this conference, I was thrilled. After all, I'm a self-professed geek when it comes to economic research and analysis-although I admit I can't actually do all the calculations. So I'm especially pleased that we have this opportunity to bring regulators, industry representatives and academics together to discuss the significant challenges to the markets and, more specifically, market structure.

For 75 years now, FINRA has been overseeing the U.S. securities markets, using a combination of innovative technology and boots on the ground to protect investors and keep markets fair. Like the technology behind our regulatory programs, academic research is also an invaluable tool for us. We depend on it to help us better understand the growing complexities of our capital markets, and identify gaps in our regulatory system so we can work to close them.

I don't need to tell you that better analytics and research equal better outcomes. As a regulator, our goal is a well-run, well-regulated market that benefits everyone. So there is an incentive for us all to work together to move research forward and develop solutions to continuously improve the markets for investors. By bringing together experts from academia, government, industry and the legal practice, we can have candid, informative discussions about current issues relating to the capital markets.

With this dialogue, we discover and better understand economic factors driving these issues, and adapt our regulatory approaches accordingly.

From our perspective, there is more we can do, particularly in developing a disciplined and structured way to evaluate the issues and inform regulatory decision-making. Today, we collaborate with academia in a number of ways. For example, we rely on our Economic Advisory Committee to advise us on academic research and macroeconomic developments that may impact our regulatory programs. Committee members-some of whom I'm happy to see are here with us today-present original research and discuss economic developments to better inform FINRA initiatives. And in other instances, we partner to conduct research. For example, Chester Spatt and colleagues worked with our Transparency Services staff to better understand the impact of transparency on the behavior of 144A and publicly traded, securitized products such as agency mortgage-backed securities.

It's our hope that this conference will generate productive conversations about how we can collaborate more effectively. This dialogue is critical, because really, what we're facing is a new regulatory paradigm-one informed by vast, expanding amounts of data and increasingly sophisticated analytics. If there's one lesson that all of us have learned, it's that today's markets are all about using data to its full potential-whether you're a securities firm, a regulator or an academic. So in that vein, let's start with some success stories and then explore areas where there is a particular need for more research and discussion.

As a regulator, it's my strong conviction that we must bring to bear the power of math and its many forms. Big data and well-constructed analytical methods are essential to more efficiently and effectively monitoring the markets and preventing investor harm.

At FINRA, we are committed to continuously evaluating the impact of our regulatory programs and responding appropriately when research suggests potential problems in the market. It's these cases that call for a disciplined and structured approach, and where regulators, the securities industry and academia can leverage each other's expertise. Take, for example, the approach we have used to expand transparency in the fixed income market.

In 2002, when FINRA launched TRACE-the Trade Reporting and Compliance Engine for fixed income products-we partnered with the academic community to study what effect the systematic collection and reporting of trades would have on the corporate bond market. The initial studies were designed to independently assess the effect on liquidity and to provide a foundation for FINRA and our Industry Advisory Committee to determine the best approach to phase in transparency.

From this research, we learned that expanding transparency to other fixed income segments didn't affect liquidity. Armed with that knowledge, we subsequently expanded TRACE to include the entire corporate bond market. Additional studies, including one conducted by the SEC, indicated that investors realized considerable savings in the form of reduced execution costs. Based on the benefits we saw, we were able to move forward with the expansion of TRACE to include agency debentures such as Fannie Mae and Freddie Mac bonds.

Similarly, when we expanded TRACE to cover asset- and mortgage-backed securities in 2011, we again partnered with the academic community. Given the more complex and bespoke nature of securitized products, we initially committed to collect and study trading information on these instruments in order to propose a sensible transparency regime. We enlisted the assistance of four independent academic teams, whose studies and insights helped inform a policy of staging dissemination to first cover To Be Announced securities followed by specified pool transactions in agency mortgage-backed securities. In early 2015, we'll begin disseminating asset-backed securities built on consumer credit.

The academic studies have helped FINRA in a number of ways. First, they provide a truly independent and objective window into the fixed income products and their trading activity. Second, they dissect the trading activity in ways that would be hard for us to replicate internally, and third, they provide a diverse set of perspectives that illuminate various aspects of the marketplace.

The work of Cornell researcher Maureen O'Hara-who previously served on FINRA's Economic Advisory Committee-is another example of the kind of study that drives regulatory actions. Maureen, along with Chen Yao and Mao Ye, two colleagues from the University of Illinois, studied odd-lot bias in equity Trade and Quote data. Their 2011 study suggested that odd-lot trades, which had been excluded from the consolidated tape because of their size, play a new and relevant role in the market. When odd-lot trades represented a trivial fraction of market activity, their omission from the consolidated tape was of little consequence. But new market practices mean that these missing trades had become both numerous and important. And, while these trades were invisible on the consolidated tape, they were not invisible to all market participants. The study recommended the SEC change the reporting rules regarding odd-lot trades. So last October, FINRA, in conjunction with the UTP Operating Committee and the Consolidated Tape Authority, began disseminating odd lot transactions via FINRA's Trade Data Dissemination Service.

Let me briefly point to two other examples of research that have informed our regulatory programs. Charles Jones and Gideon Saar-who by the way are also on FINRA's Economic Advisory Committee-have produced research on how market participants' behavior and technological advancements impact market quality. And Robert Battalio's recent study on the relationship between maker-taker fees and limit-order execution quality has prompted us to take a closer look at best execution and priced orders. These are great examples of how research has informed our regulatory actions and reactions. I can assure you that we will continue to pay close attention to work from academia that points to potential problems in the marketplace.

But in this new paradigm, I think we need to also ask, "What other research do we need? What's on our wish list? What else can we look at to make the markets more secure and efficient?" Not to suggest that I have all the answers, but here are a few of the areas where I think the markets or regulators can benefit from further academic studies and analysis.

Let's start with market structure. This is one area that has received a lot of attention in recent months, particularly in response to talk about "rigged" markets. While I can say with strong confidence that the markets are not rigged, the current market structure does indeed have some problems that regulators must address. Some of those necessary changes are already underway. For example, in June, SEC Chair Mary Jo White recommended steps to further promote market stability and fairness, enhance market transparency and disclosures, and build more effective markets for smaller companies. In prepared remarks, Chair White called for a comprehensive review of market structure that tests our assumptions about long-standing rules and market practices, re-evaluates past decisions in light of current conditions, and explores market-based solutions to issues.

FINRA supports these recommendations, and, specifically on the issue of enhancing market transparency, we agree that more oversight is needed for dark pools. In that vein, we began issuing our first reports of Alternative Trading Systems volume on a stock-by-stock basis in July, increasing transparency for ATS, or "dark pool" trading.

We're taking other steps as well. Later this week FINRA's board will consider several proposals that are intended to increase the scope of the trading information FINRA receives, and make trading activity more transparent to market participants and investors. The proposals are designed to enhance FINRA's oversight of high frequency traders and algorithmic trading activity.

The rule proposals will go through FINRA's standard rulemaking and comment process. And in whatever form they are ultimately implemented, we will need to measure the effect the changes have on market dynamics and on our ability to surveill and regulate the markets. We have an opportunity to look at more areas of the market, like OTC fragmentation, and ask tough questions about whether the SEC and other regulatory programs are working as intended.

As we consider this new regulatory paradigm--where data and analysis better inform our regulatory approach--we also want to closely follow how any changes in our practices and policies fare over time. A good example of this type of review can be found in Australia. A year after implementing a series of reforms related to dark pools, the Australian Securities & Investments Commission studied and published findings on the impact of the reforms. These reforms were intended to address the impact of dark liquidity on price formation and fairness, to protect lit orders from being traded ahead of dark orders at the same price, and to encourage more trading on lit exchanges. The Australian regulators found that the reforms successfully addressed the issues of fairness, had no impact on the bid-offer spread and reduced the proportion of below-block-size dark trades.

Another area that's calling for further research is broker-dealer finance and capital requirements. Governor Tarullo of the Federal Reserve Board and Boston Fed Director Eric Rosengren have recently spoken about this issue-and it's an important conversation to have since there have been significant changes to bank holding companies from the standpoint of leverage. At FINRA, we've been very focused on this issue with respect to risk and potential impact to the markets, but there are some complicated questions.

Recent discussions have focused on one aspect of the broker-dealer liquidity picture-their use of repos and other short term wholesale funding. Clearly, broker-dealers' reliance on short-term repo funding, especially when provided by lenders with limited capacity to assess and manage collateral, was a key point of vulnerability in the 2008 liquidity crises.

However, some discussions have lumped together all broker-dealer repo activities-irrespective of the collateral financed and the tenor of the funding-while perhaps not giving enough weight to significant improvements to the tri-party settlement process undertaken since the crisis. Certainly there remain important vulnerabilities still to be addressed in these markets, including some which may have potential systemic effects. But adopting policies intended to curtail broker-dealer repo activity across the board may be unnecessary, and would come at a high cost measured in reduced liquidity across key markets.

That's why, from my perspective, this is an opportune area for balanced economic research. Specifically, it would make sense to look closely at:

• "run" risks that exist with short-term financing mismatches, and how those risks may differ where the securities held as collateral are highly liquid, such as treasuries, Fannies and Freddies;
• risks of market illiquidity that may be raised if short-term financing becomes so restrictive that investors pull away from one or more fixed income markets;
• the capacity of alternative providers to supply liquidity including during times of stress, the sources of their funding, and the attendant costs, along with the systemic implications of shifting funding sources; and
• developing measures that are "leading indicators" of liquidity issues that are simple to calculate-and while perhaps imprecise-can serve as the canary in the coal mine" to allow us and regulated firms to address a liquidity problem before it becomes critical.

We need to be careful that in addressing the problem of mismatched short-term wholesale funding, we're not inadvertently creating another.

I want to turn back to fixed income-but along the lines of market structure and trading issues-areas that are also ripe for analysis. Those of you who know me know that I have spent considerable amounts of time and energy thinking about how equity market structure can serve the interest of investors. But it strikes me as odd that we've spent enormous energy in equity markets to measure and save pennies or just basis points on execution quality, while in the fixed income market it's more a question of nickels, quarters and dollars. Our own TRACE data shows some noteworthy execution costs in small trades in corporate bonds. For many trades, the spread calculated on a matched-pair of retail-sized corporate bond trades with one leg going to a customer is moderate, often on average on the order of less than half a percent of the trade value. But the tail of the distribution of spreads is large and significant.

Transparency, as I noted earlier, is part of the solution. We have seen shrinkage in the spreads paid by retail customers in the bond market that has occurred with the publication and usage of TRACE data, but there is more to be done. In fact, later this week, FINRA's board will consider a proposal to require disclosure of pricing information for specified matched trades involving retail customers in corporate and agency debt securities.

Fixed income securities are inherently different from equities, and their market structure is different as well. So the question is how do we appropriately measure execution quality for fixed income trades? Given how few bonds of a particular name can trade on any day, what are appropriate measures of best execution, and where and when do they perform best?

Recent developments in the bond markets raise new questions about the impact of market structure and the role of transparency. We are seeing a trend in increasing bond market activity occurring in dark pools, ATSs that limit access to quotes and RFQ systems that are not registered as broker-dealers. What is the impact of increasing the complexity in the fixed income market? What are the strategies that lever the increased opacity inherent in these types of arrangements? Is there a role for pre-trade transparency to better protect investors?

My bet is that we could spend an entire conference on these issues, and, while we are at it, also consider the appropriate market quality measures during times of stress, to more clearly evaluate the impact of large bank holding company broker-dealers dramatically reducing the fixed income positions they hold on an overnight basis. Hopefully we will have some of that research by next year.

Another important lesson that FINRA has learned from analyzing large amounts of data is that trading is more and more about cross-product strategies. In our surveillance of patterns, we have found that about 44 percent of the manipulation-based alerts we generate involved conduct on two or more equity markets. And 43 percent of the alerts involved conduct by two or more market participants. Advances in technology and data have made strategies that combine equities, fixed income, derivatives and indices profitable. But with these advances also come more opportunities to manipulate markets, take advantage of customer information or otherwise abuse investors. Here again is an opportunity where the research community can help. We have seen examples of research that identify strategies and practices that can be associated with inside information or other market abuses. What are the right cross-market strategies that we should be focusing on in our surveillance? Regulators need to better understand trading strategies that are designed to harm investors and diminish market integrity so that we are in place to stop them cold.

We can take this discussion and raise it to another level. I spoke earlier about equity market structure and its importance. These strategies that cross multiple securities-or multiple financial instruments-may be possible precisely because of the differences in the market structures for equity, fixed income, futures, options, swaps and forex. These market structures may have developed independently and organically, but they exist in a more and more connected system. When the rules of trading differ across instruments and it is possible to arbitrage across those instruments, the rules must be aligned to ensure that investors are protected. So it is important that we understand better what cross-market strategies we should be focusing on in our surveillance.

Close

I've talked a lot about how regulators and academics have-and should continue to-work together to ensure regulation evolves smartly to make sure markets operate effectively. That's not to exclude those of you in the industry from the conversation-because of course you benefit from an environment where regulation is effective.

In many ways, the securities industry is an early adopter of the fruits of research. We've learned a lot from you. When it comes to regulatory research, there are more ways where we can collaborate. For example, in some cases, you control the data that researchers could use to address some of the issues I've laid out today. So, in the spirit of this conference, I close with a request to everyone in the room-a request that we work together to address issues of investor protection and market integrity.

An open and collaborative environment benefits everyone-both by avoiding serious regulatory mishaps and ideally creating an environment where we all can identify problems earlier, before there is any investor harm. Well-run, well-regulated markets benefit everyone and therefore there is an incentive for firms to be working with people who have the time and interest in trying to solve these difficult problems.