The nature of Best Execution responsibilities of market participants has been surprisingly complex in light of the evolution of trading technology and our regulatory system. My remarks highlight several facets: the NBBO, Regulation NMS and the trade-through rule, fees and rebates in trading and make-take pricing, rebate pricing tiers, proprietary data and pricing power by the exchanges, latency, and the geography of trading and post-trade opacity. These raise a variety of questions about the meaning of Best Execution. Perhaps the ambiguity in meaning reflects the Securities and Exchange Commission’s (SEC’s) desire to allow Best Execution responsibilities to evolve organically, along with technology and the trading process. At the same time, the amorphous nature of Best Execution makes it difficult for market participants to internalize their responsibilities and for regulators to enforce clear standards.
While Regulation NMS has facilitated the electronic evolution of equity markets and originally helped end the entrenched monopolistic specialist system that predominated equity trading, it is striking that NMS, including the trade-through rule and the system of make-take pricing used by many exchanges is at the core of the complexity of our trading system. In fact, NMS came about because of concerns by the SEC about whether broker-dealers were fulfilling their Best Execution duties, but the commission’s response was to impose responsibilities on the exchanges rather than broker-dealers by requiring a trade-through restriction through exchange linkages as a partial substitute for Best Execution, though only at the “tops” of the order book—adding significantly to the complexity of market structure. To the extent that there is an agency problem in order routing, the trade-through rule could have the potential to improve routing, as would, alternatively, greater focus on Best Execution. However, the trade-through restrictions, which focus upon facilitating small trades, also have the potential to increase execution costs for institutional orders, such as those of mutual funds.
In a presentation to the SEC’s Investor Advisory Committee (IAC) two years ago I highlighted the impact of rebate pricing tiers using large numbers of alternative tiers and somewhat customized pricing within our system of fees and rebates.[1] This clearly seems designed to price discriminate under reasonable assumptions about the form of exchange costs, i.e., the costs depend upon total activity, but do not depend upon the distribution across broker-dealers. The price tiering is significant in terms of Best Execution because it encourages small brokers to use larger ones for trading to obtain better rebates, arguably inconsistent with the requirement of the Exchange Act with respect to placing an undue burden on competition. Indeed, citing this problem specifically and my related paper,[2] three congressmen wrote to the SEC early last year suggesting a ban on pricing tiers.[3] The case for such a ban is even stronger than at the time of my presentation or the subsequent letter as the D.C. Circuit Court of Appeals later overturned the SEC’s effort to study and potentially ban rebates after a market pilot. Of course, if the SEC had banned rebates, then the tiering issue would have been moot.
Another key feature of current practice is that the pricing tier achieved by broker-dealers and others routing orders is only determined at the end of the month and is not transparent to asset managers, in contrast to the tiering based upon pre-determined pricing in airline frequent flyer programs. In that spirit and in the aftermath of my earlier presentation, the IAC itself called upon the SEC to require much stronger disclosure, but that has not progressed. This actually was a much weaker recommendation than those of the three congressmen. The unknown nature of the rebate rate leads to an agency conflict even for institutional investors and amplifies the agency conflict in order routing relative to the basic “make-take” structure.
I would reemphasize the earlier recommendations to the commission. In my view it should seriously consider banning price tiers for rebates and absent that should implement much stronger disclosure requirements. This is an important issue in a Best Execution context.
Another important issue involving Best Execution is whether different market participants need access to various proprietary datasets to fulfill their Best Execution requirements or whether such access should simply reflect the commercial preferences of various broker-dealers and asset managers? To the extent that particular data are required universally, is there any advantage to not including it in the SIP? The issue of data that would be required for Best Execution is complex, in part because of the market power of many of the providers of unique data, the range of proprietary datasets offered, the diverse business models for various intermediaries, and the locations of the various exchange data centers and trade reporting facilities.
Data are not uniformly available at all locations due to latency and the geographic structure of data availability. Yet much of the SEC’s regulatory framework, such as the meaning of the NBBO, the import of the trade-through rule, and the underlying meaning of Best Execution largely abstract from latency frictions and the geographic structure of trading. The diverse geographic points at which prices are measured and the latency to identify and access liquidity are fundamental. Even with faster technology and assumed continuation of Moore’s Law, the geographic structure of trading and the physical constraint associated with the speed of light suggest that frictions cannot disappear in the limit. Given that, and the importance of relative time in trading, latency is inherent and an important issue to address in the context of Best Execution.
With collaborators, I have recently been examining post-trade transparency, and the pattern of trading around the publication of trades. Our research highlights the difference in latency between off-exchange and exchange trading: We identify bursts of trading activity consistent with market participants learning about off-exchange trades from SIP broadcasts, but appearing to learn about on-exchange trades from faster proprietary feed broadcasts. The much slower response to off-exchange trading reflects greater post-trade opacity and latencies in reporting and the manner in which market participants utilize data. This highlights fundamental questions about the meaning of Best Execution in the presence of various latencies, including those associated with exchange vs. off-exchange trading. For example, what are the Best Execution responsibilities of various market participants in light of their potential knowledge of trading activity given the latencies?
The underpinnings of Best Execution, both conceptually, as highlighted by my remarks, and pragmatically from the perspective of the goals of the commission require further attention.
ENDNOTES
[1] Remarks of Chester S. Spatt, March 28, 2019, to the Investor Advisory Committee.
[2] Chester S. Spatt, “Is Equity Market Structure Anti-Competitive?” working paper, revised December 2020 and presented at the American Finance Association meetings, January 2021.
[3] Budd, Congressman T.; Congressman A. Mooney; and Congressman A. Wagner, Letter to SEC Chairman Jay Clayton, January 31, 2020.
This post comes to us from Professor Chester S. Spatt at Carnegie Mellon University’s David A. Tepper School of Business. It is based on his remarks to the Securities and Exchange Commission’s Investor Advisory Committee on June 10, 2021.