Conflicts of Interest Associated with the Use of Predictive Data Analytics by Broker-Dealers and Investment Advisers

October 10, 2023

Docket ID: Release Nos. 34-97990, IA-6353, File No. S7-12-23

To the Honorable Gary Gensler, Chair of the Securities and Exchange Commission, and the Honorable Commissioners Hester M. Peirce, Caroline A. Crenshaw, Mark T. Uyeda, and Jaime Lizárraga:

The Competitive Enterprise Institute (CEI) is a non-profit public interest organization committed to advancing the principles of free markets and limited government. CEI has a longstanding interest in applying these principles to the rulemaking process and has frequently commented on issues related to oversight of rulemaking and the regulatory process. One of CEI’s major objectives is reducing regulatory barriers that affect access to capital and investor choice. CEI has pursued this objective through policy
analysis, Congressional testimony, and litigation.

On behalf of CEI, we are pleased to provide comments to the Securities and Exchange Commission (SEC) on its proposed rule titled “Conflicts of Interest Associated with the Use of Predictive Data Analytics by Broker-Dealers and Investment Advisers.”

Background

The proposed rule seeks to introduce new regulations under the Securities Exchange Act of 1934 (Exchange Act) and the Investment Advisers Act of 1940 (Advisers Act). The primary objective of the proposed rule is to address and potentially neutralize certain conflicts of interest that may arise from broker-dealers’ or investment advisers’ interactions with investors when using modern technologies–including artificial intelligence (AI)–that are designed to optimize, predict, guide, forecast, or direct investment-related behaviors or outcomes. The SEC contends in this proposal that information provided to retail investors from the adoption and utilization of these emerging technologies, referred to as “predictive data analytics” (PDA), has led to increased conflicts of interest. The SEC is therefore proposing amendments to existing rules under both the Exchange Act and the Advisers Act. These amendments would mandate that broker-dealers and investment advisers identify conflicts of interest with respect to these technologies, as well as explain how they will neutralize or eliminate them, in addition to requiring the creation and maintenance of written records to that effect.  

However, in reviewing the proposal, the SEC’s economic analysis falls far short of conclusively demonstrating the existence of a tangible problem that necessitates regulatory intervention. The analysis does not provide sufficient empirical evidence to support the assertion that current market dynamics, driven by the use of PDAs and related technologies, are causing widespread harm to investors. Moreover, the proposed solution, as outlined, lacks clarity on how it will directly lead to beneficial outcomes for investors and the broader citizenry.

Further, prominent legal experts contend that the rule as written may violate the First Amendment’s prohibition on government suppressing truthful speech. In an op-ed for the Wall Street Journal, former attorney general Bill Barr and former U.S. Representative Barbara Comstock make the case that the rule will prevent ordinary investors from receiving truthful information from their broker-dealers and investment advisers, leaving that information only in the hands of financial elites. Barr and Comstock write that the rule “also creates glaring constitutional issues by restricting what information firms can communicate to the public without adequate justification.” They explain: “As courts have recognized, there is no ‘securities law exception’ to the First Amendment. Under the U.S. Constitution, unless a communication is deceptive, the mere fact that it imparts to a customer information consistent with a speaker’s own interests can’t possibly justify these sweeping restraints.”

Given these issues related to the First Amendment, and given gaps in both the identification of the problem the rule is supposed to solve and in the analysis of the proposed solution’s potential impact, in addition to any serious consideration to alternatives beyond the proposed rule, it would be prudent for the SEC to reconsider this proposal. We respectfully suggest that the proposed rule be withdrawn until a more rational and evidence-based approach can be developed.

The SEC presents no evidence of a systemic problem.

In a forthcoming paper for the Competitive Enterprise Institute, one of us presents a framework explaining the evidence that should be required to justify regulation of new technologies, such as AI. This framework has obvious relevance to technologies such as PDAs, which often rely in part on AI. The governance framework is also similar to the SEC’s own economic analysis guidance, and it includes the following steps:

Step 1: Demonstrate a problem exists

Step 2: Define the desired outcome

Step 3: Enumerate alternative solutions

Step 4: Rank alternatives

Let us first consider step one. The first rule when regulating new technologies is to demonstrate a problem exists. In August of 2021, the SEC requested comments from the public regarding the digital engagement practices (DEPs) of investment advisers and broker-dealers. At the same time, the SEC used a “Feedback Flyer” to gather comments from individual retail investors regarding their use of online trading or investment platforms. This was designed to provide the SEC with a better understanding of retail investors’ experiences on these platforms. These SEC actions followed significant media attention and congressional hearings related to the run-up and subsequent crash of Gamestop stock prices, as well as the role digital apps like Robinhood played in the corresponding trading.

To start with, the SEC fails to consider that many do not view the events surrounding Gamestop and other so-called “meme stocks” as a problem but rather as a watershed moment that demonstrated the power of ordinary investors when collaborating with beneficial communications technologies, as well as a new method of providing valuable information to the market about firms’ intrinsic worth. As one of us has written: “Just as short sellers benefit a market by reducing information asymmetry, … so do the ‘new longs’ add information about the hidden value of certain companies. All of the companies targeted by the Redditors have engendered good will through their business history and brand names. The surge in stock price could communicate to financiers that the companies could have much value with a retooled product. The share price of GameStop in particular probably won’t stay near the high it reached, but the prices and valuation of these companies may remain higher for a long time [and] may stay higher than they were before the price swing, as the market may have rediscovered some of their intrinsic worth.” In the preamble to the proposed rule, the SEC voices the concerns of commenters who responded to its requests for information and who are concerned about the use of PDAs and similar technologies by broker-dealers and investment advisers. The agency has provided some evidence highlighting potential conflicts of interest when these technologies are used in ways that might prioritize a firm’s interests over those of its investors. The SEC’s examples include:

• Robo-adviser case: The SEC references a recent enforcement action involving allegations against an adviser that marketed its “no fee” robo-adviser portfolios as being determined through a “disciplined portfolio construction methodology.” These portfolios were allegedly preset to hold a certain percentage of assets in cash because the adviser’s affiliate was guaranteed a certain amount of revenue at these levels. The adviser allegedly did not disclose this conflict of interest, which could have led to investors holding higher than optimal amounts on cash reserves in their portfolios.

• Operational challenges with PDAs: The SEC points out that the reliance on complex and opaque PDA-like technologies can result in operational challenges. For instance, a failure to identify and address conflicts present in the PDA technology used to steer investors could result in a firm’s inability to recognize risks to investors. This could lead to inadequate compliance policies and procedures, potentially resulting in harm to investors.

While this anecdote and the accompanying general concern shed light on some potential issues, they are not sufficient evidence to demonstrate a systemic, ongoing market failure with respect to PDAs in the investment sector. Anecdotal evidence, though sometimes compelling, represents isolated incidents and might not be indicative of widespread malpractice or systemic issues. To conclusively demonstrate a systemic problem, there would need to be comprehensive empirical evidence showing a consistent pattern of malpractice across a significant portion of the industry. In the absence of such evidence, the case for broad regulatory changes is much weaker, and there is a risk regulation will have unintended consequences for the market and its participants.

The SEC also postulates that the rapid data processing and analytical capabilities of PDAs could lead to unanticipated conflicts of interest that affect a large number of investors and potentially cause market disruptions.  The preamble speculates: “A given firm might not fully bear the cost of the use of these technologies, and thus might not fully internalize the full cost of the use of these technologies. The costs imposed on entities external to the firm are called negative externalities, and regulatory intervention may be needed to address these costs.” Having explained what negative externalities are and theorized that some might be present, the SEC fails to demonstrate any. Mere speculation of what may or may not be does not demonstrate that that the SEC’s proposal is needed, only that it “may be needed,” to use the SEC’s own language.

The SEC cites only potential harms and largely relies on general observations and abstract theorizing. For instance, although the SEC points out areas where conflicts of interest could potentially arise from PDA-like technologies, there’s little proof that issues like erroneous data or modeling drift favor firms’ interests over those of investors. Concrete negative outcomes for investors should be identified, with evidence linking the problems to the use of PDAs. The SEC does reference feedback from various stakeholders, but these references are often speculative as well, highlighting opinions about potential problems rather than actual ones.

All told, the evidence provided by the SEC that a systemic problem exists is closer to armchair theorizing than a comprehensive empirical analysis. The reliance on speculation, general observations and anecdotal evidence, without a robust empirical foundation, provides unconvincing evidence of the existence of a widespread market failure. Based on the information provided, it can be concluded that the SEC does not have evidence to demonstrate a clear and systemic market failure related to the use of PDAs.

Assertions about excessive trading and behavioral market failure are unconvincing.

Beyond the more traditional principal-agent problems and negative externalities the SEC asserts might exist in investment markets relying on DEPs and other PDA-like technologies, the proposed rule is noteworthy because the SEC asserts several less traditional forms of market failure may exist. These include:

• Excessive trading: The SEC is concerned that technology that involves rapid data analytics could lead to unnecessary trading, allowing the firm to collect extra fees or payments from the additional trading activity, as well as reduced returns for investors.

• Behavioral biases: The SEC notes that certain platforms use PDA technologies to target investors with game-like features such as points, rewards, leaderboards, push notifications or use social proof and peer influence to sway investor behavior. These practices might exploit psychological biases, leading to impulsive and potentially irrational investment decisions.

• Nudges: The SEC highlights that while DEPs can be beneficial, they can also subtly take advantage of human psychological factors. The rapid rate of investor interactions, the size of datasets, and the complexity of algorithms can influence investor behavior to their detriment and the benefit of the firms.

In support of these points, the SEC does cite studies showing that excessive trading negatively impacts investment returns. Frequent traders might exhibit lower annual returns than infrequent traders, in part due to overconfidence. The SEC demonstrates that some stock trading apps appear to employ strategies similar to those used in the gambling industry to encourage frequent repeat betting, obscure costs, or they might offer complex instruments that result in large payoffs only in rare cases. The SEC claims these strategies may benefit from survivorship bias, whereby exploiting customer biases is more profitable and therefore becomes the dominant strategy in the industry. While the SEC cites various studies, there are shortcomings to the evidence provided. While excessive trading is highlighted as a concern, it is not definitively shown that PDAs are the primary or sole cause of this behavior. Nor has the SEC defined the level of trading it considers to be “excessive.” Without this information it is difficult to know the extent to which a problem might exist, if it exists at all. With respect to behavioral biases and nudges, the best regulators can do is to identify investors’ “revealed preferences.” That is, regulators can infer what investors believe is in their own interests from their actions. Whether these actions might deviate from their “true preferences” is a matter of speculation only. It is impossible to know investors’ true preferences since these are known only to the investors themselves, and in some cases not even to them. Because this information will always be out of reach to regulators, there can be no definitive proof of a “behavioral market failure,” a situation where cognitive biases lead consumers or investors to act in a manner inconsistent with their own preferences. This is a classic example of the knowledge problem identified by F.A. Hayek, where government officials lack the information they need to improve social welfare.

Ironically, the SEC’s proposed rule itself might be an example of a behavioral government failure, as it might be perceived as a knee-jerk reaction to public outcry in light of the Gamestop controversy, rather than as a well-reasoned regulatory response to a systemic problem in the marketplace. The urgency to “do something” in the wake of high-profile events can sometimes lead regulatory bodies to act hastily, succumbing to “availability bias.” Confirmation bias, where actions are taken to affirm a pre-existing belief, or, social desirability bias, which occurs when actions are taken to appease public sentiment, can also result in regulations that are not thoroughly vetted or adequately backed by evidence.

The fact that investors are likely to lose money or earn lower returns when they invest in a biased manner may explain why a rationality postulate is useful for modelling and predicting outcomes of markets. The relevant survivorship bias here may well be among unbiasedinvestors, as biased investors exit the marketplace after losing money. This does not negate the need to crack down on instances of fraud or exploitation where they exist, but it does help explain why systemic investor irrationality in markets may be relatively rare compared to instances of systemic rationality.

Regardless, the SEC is unable to distinguish between instances of rational and irrational investing with current information. In light of these problems, there is a risk that introducing new regulations governing the use of PDAs and similar technologies among investment advisers and broker-dealers might misallocate resources, exacerbating inefficiencies in these markets, and potentially introducing new biases where none were previously present.

The SEC’s proposal creates burdens without quantifiable benefits. The SEC’s proposed rule creates several new requirements for investment advisers and broker dealers. The SEC proposes that firms adopt and implement written policies and procedures that include a description of material features of, and any conflicts of interest associated with the use of, covered technologies used in any investor interaction. The proposed rule also requires regular reviews of the adequacy of these policies and procedures, as well as a detailed written description of the process by which the firm determines whether and how to eliminate or neutralize the effect of any conflicts of interest. The SEC is also proposing that firms create and maintain records to this effect.

The evidence offered by the SEC does not offer a clear causal theory, backed by comprehensive data, that outlines how the proposed solution will directly address the largely speculative problems and lead to better outcomes for investors. To support the proposal, the SEC primarily relies on the over 2,300 public comments it received in response to its 2021 request for information and feedback flyer. These comments offered various perspectives on the purpose of DEPs, how investors interact with them, and reflections on potential regulatory actions. They also provided views on the benefits and risks related to firms’ use of DEPs and associated technologies.

While the SEC has gathered a significant number of public comments and feedback, the evidence provided primarily consists of opinions and perspectives from various stakeholders. While these comments are valuable, they do not provide a robust empirical foundation with which to support the proposed regulatory requirements.

Gathering opinions and feedback is essential, but to conclusively demonstrate the need for the proposed solution, there would have to be comprehensive empirical evidence showing a consistent pattern of malpractice across a significant portion of the industry and supporting how the proposed rules would effectively mitigate these issues. It has been said many times, and bears repeating here, that the plural of anecdote is not data.

Moreover, the SEC should identify concrete metrics by which it will gauge whether the regulation is having its intended effects. Without concrete goals and tangible markers of progress, it is unlikely the SEC even has a concrete idea of what it wants the regulation to achieve. In absence of clear goals, the regulation is unlikely to improve welfare for citizens.

The SEC fails to adequately consider alternatives to the proposed rule.

The SEC, in its rulemaking process, considered several alternatives to the regulation being formally proposed. These include:

• Independent third-party analyses: Allow (or require) firms to use external experts to check if they’re following the rules correctly.
• Senior personnel involvement: Make sure that top people in the firm, especially those who know a lot about technology, are involved in setting up and checking these policies.
• Exclusion for large datasets: If a technology uses big sets of data and the firm believes that this data doesn’t favor the company’s interests, then that technology might be exempted from some checks.
• Limiting scope of rules: Apply the new rules and record-keeping changes only to technologies used by broker-dealers that don’t make specific investment suggestions.
• Regular testing of technologies: Firms should check their technologies regularly, perhaps once a year or at some other set interval.
• Standardized disclosure: Firms should give investors a standard set of information about any conflicts of interest related to these technologies.

While the SEC should be given credit for at least listing these six alternatives, there is no evidence the alternatives were seriously considered as there are notable gaps in the agency’s evaluation. For one thing, the SEC failed to estimate the costs of any of the alternatives.

By contrast, the SEC did estimate the costs of the proposed rule. For simple covered technology firms, the SEC estimates the initial cost of the rule at $180 million with $90 million in annual costs thereafter ($11,150 and $5,575 per firm, respectively). Meanwhile, for complex covered technology firms, initial costs are estimated at $281 million with $140 million in annual costs thereafter ($156,100 and $78,050 on a per-firm basis).

However, without a clear understanding of the costs and effectiveness associated with each alternative, it is challenging to make any informed decision about which option is most efficient. It is fairly clear in some cases that the alternatives listed are more or less burdensome than the proposed rule. That said, the SEC did not estimate the benefits of any of the alternatives, including the proposed rule itself. This omission is significant because, without a clear understanding of the potential benefits, it is impossible to weigh the pros and cons of each alternative to identify the most efficient solution.

Given the lack of comprehensive cost-benefit analysis for each alternative, there is no concrete evidence that the SEC’s chosen proposal is the most efficient option. In fact, the absence of meaningful benefit estimates might indicate that the proposed rule, and the alternatives listed, may not yield any meaningful benefits, suggesting inefficiency in the SEC’s proposal. Without a complete analysis, there is little reason to believe that the SEC’s chosen approach is the best solution to the identified problems.

Conclusion

In light of the Gamestop controversy and the subsequent outcry from some parts of the financial sector, the SEC’s proposed rule regarding the use of PDAs by broker-dealers and investment advisers appears to be a reactionary measure rather than a well-thought out regulatory response. The urgency to “do something” has seemingly driven the SEC to act hastily, potentially succumbing to its own form of behavioral bias.

While the intention to protect investors and maintain market integrity is commendable, the evidence provided by the SEC, largely anecdotal and generalized, falls far short of conclusively demonstrating a widespread or systemic market failure in the industry. The cited studies and concerns do not establish a clear causal link between the use of PDAs and concrete negative outcomes for investors, leaving significant gaps in the rationale for the proposed regulations. Furthermore, while the SEC identified several alternatives to the proposed rule, it failed to provide a comprehensive cost-benefit analysis for each. This is unfortunate since the SEC had been making some progress on improving the quality of its regulatory impact analysis in recent years. If this proposal is any indicator, that progress may have stalled. In conclusion, the SEC’s proposed rule, driven by external pressures and a perceived need to act, seems to be characterized by irrationality as well as inefficiency. Regulatory actions, especially those costing hundreds of millions of dollars and having the potential to slow down technological innovations in markets, should be grounded in comprehensive research and analysis to ensure they address underlying issues effectively and do not introduce new, unforeseen problems into the market ecosystem. The SEC has failed to provide the minimal evidence needed in support of a new rulemaking. Due to these flaws in the rule and its potential First Amendment violations, this proposal should be withdrawn.

Sincerely,

James Broughel, PhD

Senior Fellow, Competitive Enterprise Institute

John Berlau

Senior Fellow and Director of Finance Policy, Competitive Enterprise Institute