THE NEW STOCK MARKET
Law, Economics, and Policy
Merritt B. Fox, Lawrence R. Glosten & Gabriel V. Rauterberg
Conclusion
This book has aimed to provide a systematic legal and economic analysis of a wide range of important issues in the modern equity market. As a conclusion, we hope to point to three major open questions about the structure of financial markets that stand in need of much greater empirical and theoretical exploration. We will quickly discuss how market structure differs across asset classes; the question of the optimal degree of pooling of informed and uninformed traders; and how technological innovations at exchanges might affect the character of liquidity provision. For each, our emphases will be two-fold: first, to demonstrate the continuing power of the informational paradigm to cast light on these issues, but second, to also use them as an occasion for appreciating its limits. The first open issue is to consider the similarities and differences between the equity market’s structure and the many other markets important to finance— including debt (whether corporate, municipal, or government bonds), options, commodities, futures, and over-the-counter derivatives—and whether and how those differences should make for distinct regulatory treatment of those other markets. To our knowledge, the legal literature has only begun to see the importance of thinking about the differences—often subtle and surprising—of
R e g u l a t i o n o f B r o k e r - D e a l e r s | 294
microstructure across asset classes.1 Yet, as we hope to illustrate, it is an issue of potential first-order importance for regulators. In keeping our focus on the adverse selection model of liquidity provision, a key axis for thinking about different financial markets is their distinctive information structure. By this, we mean the importance, if any, of the role played by private information and information asymmetries in that market. Consider, as a quick case study, a market with a very different informational structure than corporate equities, the corporate bond market. The structure of trading corporate bonds today is markedly different from the structure of the equity market. In fact, the corporate bond market more closely resembles the equities market of 1990 or even 1980. Recall that today’s equities market is dominated by electronic limit order books open to the provision of liquidity by any market participant. In contrast, the debt market resembles a pure dealer market, where every purchase order is filled by a sale from a dealer and every sell order is filled by a purchase from a dealer. There are no national corporate bond exchanges where marketable orders execute against nonmarketable limit orders and electronic trading platforms have made surprisingly limited inroads. In terms of law, the regulatory structure governing the market is minimal. Public offerings of corporate debt securities must generally be registered under the federal securities laws. Institutions that act as dealers in those securities must register with the SEC as broker-dealers, and transaction data is publicly disclosed.2 In terms of secondary market structure, however, there is no mandated transparency for quotations in corporate bonds, no order-handling rules requiring dealers to post customer limit orders, and no trade-through rule proscribing transactions at prices inferior to superior quotations available elsewhere. In fact, as late as 2001 there was no consolidated data on transaction prices. Furthermore, as we have seen, a major preoccupation of federal securities law for equities is insider trading, discussed extensively in chapters 5 and 6. Rule 10b-5 may apply to bond transactions, but current SEC activity suggests that there are few, if any, insider trading prosecutions based on individuals’ or institutions’ trading of debt instruments. Against the background of these regulatory differences between equity and debt markets, we note that spreads in the corporate bond markets are orders of magnitude higher than for NMS stocks. Does this fact call for any regulatory response? If so, what kind? Is is time for a Reg. NMS for corporate bonds or for more vigorous prosecution of insider trading? We suggest that the answers to these questions are helpfully informed by the informational perspective, and that they should be shaped by the answer to a more basic question: What role, if any, does private information play in the trading of corporate bonds?
295 | C o n c l u s i o n
On the one hand, if private information is pervasive, shaping the provision of liquidity in a manner similar to the equities market, then active insider trading prosecutions in the corporate bond market could play an important role in shaping market outcomes. Specifically, insofar as there is private information, then the cost of liquidity will reflect a component due to adverse selection and reducing adverse selection can reduce the cost of liquidity. If, on the other hand, the role of private information is minor, then there is little need for regulatory tools like the prohibition on insider trading. In that case, the large spreads that we observe may be the result of a substantial shortfall in the competitiveness of the bond dealer market, suggesting that a very different kind of regulatory intervention is called for, presumably focusing on market structure rules and quote and transaction transparency. At the broadest level, leading economic theories of the role played by the trading of corporate bonds suggests that their “informational insensitivity” is a crucial feature.3 The basic idea is that market participants often trade corporate bonds precisely because they value the fact that they—and their counterparty—are unconcerned with attempting to discern whether the actual value of the instrument differs from the current trading price. This is light-years from the world of equities and the pervasive concern with adverse selection that shapes that market. But is this true of all corporate bonds? Perhaps not. The bonds of highly distressed issuers are unlikely to be informationally insensitive or to be providing the same financial service as the “safe” corporate bonds that Gary Gorton and others have emphasized. Indeed, the debt of distressed issuers is traded by hedge funds and other institutions that seem to invest considerable resources to become highly informed as to their cash-flow prospects. It seems like such debt securities might be usefully subjected to an insider trading regime under the misappropriation theory. The details we leave to future empirical research and theory, but the lesson seems worth emphasizing. Naturally, the nature of an asset—its cash flow profile, governance rights, contractual details—matters to a market. And both the bonds of a blue-chip issuer and the bonds of a company considering bankruptcy share fundamental similarities in terms of maturity dates, term structure, and fixed repayment of principal and interest. Yet, the informational structure of the secondary market for those two assets varies considerably, and these differences—the insignificance of information asymmetries in the market for high-quality debt, and their pervasive role in the market for distressed debt—can matter considerably to whether a given regulatory policy is unnecessary or important. Before closing this discussion of corporate bond trading, it is worth investigating another, non-informational dimension on which debt and equity differs.
R e g u l a t i o n o f B r o k e r - D e a l e r s | 296
This will suggest that some market phenomena may not be explained by our informational paradigm and that other areas of investigation will no-doubt be needed. Specifically, it should be noted that the other costs of liquidity provision that we have discussed may play a role. While the typical listed firm has only one series of equity, it might have hundreds of series of bonds, each with its own maturity, coupon rate and contractual provisions. Managing an inventory of all the series of bonds in all listed issuers would be prohibitively costly, and require a huge amount of capital. As a result dealers carry a smaller inventory but even this is costly because many corporate bonds trade quite infrequently. They then rely on searching for the counterparty to a trade that a client wants in a particular bond. How costly is all of this? We do not know. We turn now to a second open question. A basic fact of the equity market, familiar from earlier chapters, is that a liquidity provider on an exchange cannot tell whether its counterparty is informed or uninformed. Indeed, this fact generates the fundamental dynamic in which liquidity providers lose money to the informed by trading with them and charge the uninformed a higher price than they would otherwise need to pay. Using the terms informally, stock exchanges thus reflect a kind of “pooling equilibrium”—a state of affairs in which traders of different “types” in terms of adverse selection risk all transact together, from the mistaken to the uninformed to those with various degrees of private information. Indeed, the stock market is routinely characterized as anonymous. However, if anonymity is taken to mean that liquidity providers simply have no idea of the type of their counterparty, then this is frequently false. In myriad ways, liquidity providers often are aware, probabilistically, of the type of their counterparty. The internalization of purchased order flow is a clear example. If, say, Citadel purchases the order flow from Bank of America’s retail customers then it can safely assume that the vast majority of incoming marketable order flow is uninformed. The internalizer does not know the precise identity of the person on the other side of the trade, but that is not really the characteristic of interest. What liquidity providers are interested in is a counterparty’s type for the purposes of adverse selection risk, and an internalizing broker-dealer certainly has a good idea about its counterparties’ type when it purchases all of a retail brokerage’s marketable order flow. Or consider NYSE’s retail liquidity program.4 The program empowers liquidity providers to quote non-displayed subpenny orders that will only interact with specifically designated orders from retail traders whose brokers vouch for their retail status. This offers liquidity providers a lower adverse selection risk and retail traders price improvement. Dark pools may also be run in a fashion that attempts to keep out informed traders.
297 | C o n c l u s i o n
We mention these three examples of the many possible to make clear that the current stock market also reflects a partial separating equilibrium in which the “type” of different traders is sometimes signaled, perhaps by the origin of their order flow (internalization), by disclosure by their broker (as with NYSE’s retail liquidity program), perhaps by their trading style, etc. As a result, the degree to which any secondary market reflects a pooling versus a separating equilibrium is to a significant extent a choice variable for regulators. It is thus a live and important policy question to ask: What is the optimal degree of pooling of the informed and uninformed? In principle, Congress could mandate that all equity trade occur on a single stock exchange. This would represent a significantly greater degree of pooling than the current multi-venue system reflects. Market forces themselves seem to be pushing in the opposite direction of greater separating. This is a profound issue, which we raise largely to illustrate how major architectural questions remain for microstructure theory to address and the value of the adverse selection perspective in thinking about them. A few quick things can be said now, however. In general, the issue of pooling versus separating reflects a basic trade-off in the extent to which the uninformed will subsidize the informed by sharing in the adverse-selection component of the spread when there is pooling. Thus, more pooling should generally mean the informed find their information gathering and trading activities more profitable. More separating means that liquidity providers are able to offer the uninformed cheaper liquidity, reflecting the diminished or nonexistent adverse selection risk, while the informed find their activities less profitable due to costlier liquidity. One background question then is whether there is enough information production in a given market. Here, too, in terms of finding the optimal degree of pooling, it will be important for regulators to appreciate that the degree of optimal pooling will likely differ across asset classes. Indeed, the degree of separation of the informed and uninformed not only varies in dizzying ways in equities across global jurisdictions, but it varies considerably across asset classes in any one jurisdiction. Even in those markets in which there are both the informed and uninformed, if the informational signals that a market generates are of less usefulness to the real economy than the equity market, then the degree of optimal pooling may be lower. Lastly, looking to the future of financial markets one cannot help but observe how technology is transforming their market structure. Consider just a few of these technological innovations in equity market structure. One innovation that has frequently appeared in the headlines has been “speed bumps”
R e g u l a t i o n o f B r o k e r - D e a l e r s | 298
at exchanges. This began with the exchange IEX, whose speed bump imposes a 350-microsecond delay on all incoming messages to the IEX order book and an equally timed delay on all information disseminated from the order book. After some vacillation, it seems that the Chicago Stock Exchange’s quite different speed bump will become operational. This speed bump is different because it is asymmetric, imposing a 350-microsecond delay on most incoming orders, but not on the ability of certain designated market makers to alter their limit orders. Or consider IEX’s discretionary peg order, which prevents non-displayed midpoint orders from executing during algorithmically defined periods of rapid change in quotations for a stock. IEX may seek to allow the discretionary peg functionality to apply to displayed orders. To all of these innovations, the adverse selection view has much to say. For the speed bumps, it suggests the need for solving for a complex equilibrium in which some speed bumps make liquidity provision more difficult by slowing down receipt of transaction data, while other speed bumps enable market participants to readjust limit orders in the face of new information, while slowing everyone else down. The discretionary peg order, intriguingly, may effectively prevent limit orders from executing precisely when the adverse selection environment is greatest, making IEX a “safer” environment for the posting of limit orders. We note this not because we are confident about the accuracy of these speculations, but because we are confident in the distinctive explanatory power of the informational paradigm to cast light on new questions and markets. We have focused on the U.S. stock market—perhaps the world’s best functioning market of any kind—as an avenue for studying how markets function— their dynamics, their mathematical properties, their social utility—in one of their purest and most elegant forms. Even there, much more remains to be understood. But we must leave all that to others, and to another day.
“In immensely readable fashion, The New Stock Market connects the fundamentals of market structure to new (and old) challenges: insider trading, market manipulation, high-frequency trading. A profoundly important look at how our stock markets have changed and the regulatory first principles necessary to keep them orderly and equitable as these changes continue.” DONALD LANGEVOORT, Georgetown University
“Integrating the perspectives of information economics and the law for understanding markets for trading equity, this book will be of considerable interest to students of markets and the law, as well as securities lawyers, investment bankers, analysts, economists, and regulators.” CHESTER SPATT, Carnegie Mellon Tepper School of Business and MIT Golub Center for Finance and Policy
Praise for
THE NEW STOCK MARKET “The New Stock Market achieves a difficult balance: it is accessible yet sophisticated. The mysterious new terms of market microstructure—‘high-frequency trader,’ ‘dark pool,’ ‘maker-taker’ rebates, and ‘internalization’—are all fluently explained, and this serves as a prelude to the authors’ careful weighing of the policy choices. Few books in this area have been this lucid and this rigorous at the same time.” JOHN C. COFFEE, Columbia University
“Equity capital markets are going through unprecedented change: new technology, new players, new venues, and new trading strategies. How can regulators respond to these developments without impeding market efficiency? These are the issues that Fox, Glosten, and Rauterberg analyze in their outstanding book, providing vital—and novel—insights and recommendations that should be welcomed by both regulators and investors. Highly recommended.” EDWARD F. GREENE, Cleary Gottlieb Steen & Hamilton
COLU M B I A U NI V E R S ITY PR ES S / N EW YORK cup.columbia.edu Printed in the U.S.A.