August 23, 2018

Response to the SEC's Proposed Regulation on the Best Interest for Retail Securities Broker-Dealers

Key materials
Rule details

Regulation Best Interest
Agency: Securities and Exchange Commission
Proposed: May 9, 2018
Comment Period Closes: August 7, 2018
Submitted: August 23, 2018
Document No. 2018-08582
RIN: 3235-AM35

Thank you for the opportunity to comment on the Securities and Exchange Commission’s (SEC’s) proposed Regulation Best Interest for retail securities brokers who provide their clients with incidental investment advice. The Mercatus Center at George Mason University is dedicated to bridging the gap between academic ideas and real-world problems and to advancing knowledge about the likely consequences of proposed regulation for private markets. Accordingly, this comment represents the views of no particular party or interest group. It is designed to help the SEC assess whether Regulation Best Interest is economically justified, as required by statute. Owing to the empirical difficulties with traditional cost-benefit analysis, I recommend that the SEC withdraw Regulation Best Interest until it has developed and tested a Coasean transaction cost theory of the market for investment advice and used it to determine that the rule is likely to reduce the parties’ costs of transacting.

Introduction

Regulation Best Interest (RBI) addresses the conflicts of interest between securities brokers and their retail clients that commonly bedevil principal-agent relations. It recommends adding a best-interest standard to the Financial Industry Regulatory Authority’s (FINRA’s) well-established suitability rule for retail brokers when they provide their clients with investment advice incidental to their main function of executing the client’s trades. Specifically, “when making a recommendation of any securities transaction or investment strategy involving securities to a retail customer,” RBI would require brokers “to act in the best interest of the retail customer at the time the recommendation is made without placing [its] financial or other interest ahead of the interest of the retail customer.”

Among other things, RBI would impose legal duties on brokers to exercise reasonable care, to fully disclose all material conflicts of interest, and to establish written policies and procedures reasonably designed to identify and disclose or entirely eliminate these conflicts. Breach would risk SEC enforcement actions carrying significant penalties and civil liability to aggrieved parties for securities fraud.

In addition to the SEC’s charge to protect investors, the four main securities statutes now require the SEC to show that all new regulatory rules “promote efficiency, competition, and capital formation.” Federal case law and its own internal pronouncements have interpreted this to require a thorough economic analysis of proposed major rules, including traditional cost-benefit analysis (CBA), and to quantify costs and benefits as best it can.

Clinton-era Executive Order (EO) 12866 explicitly requires executive agencies to perform CBA of proposed regulations to “maximize net benefits.” Although independent agencies such as the SEC may be exempt from its mandate, the order outlines best practices for any economic analysis of federal regulation.

EO 12866 also provides the following seemingly sensible test for federal regulation, in brief: “The private sector and private markets are the best engine for economic growth. . . . Federal agencies should promulgate only such regulations as . . . are made necessary by compelling public need, such as material failures of private markets to protect . . . the well-being of the American people” (emphasis added). It also requires federal agencies to “base [their] decisions on the best reasonably obtainable scientific, technical, economic, and other information concerning the need for, and consequences of, the intended regulation.”

In its “Economic Analysis” section, RBI identifies a number of empirical questions relevant to assessing the costs and benefits a best interest standard might generate, but also repeatedly states that they are unanswerable because the data necessary to quantify costs and benefits under various assumptions are unavailable. This finding highlights the inherent difficulties of doing reliable CBA in this setting.

This comment identifies an alternative approach to assessing proposed regulation based on Nobel laureate Ronald Coase’s path-breaking work The Problem of Social Cost. At the time Coase was writing, mainstream economists and policy commentators generally believed that government regulation is justified when markets fail to efficiently allocate resources owing to so-called “externalities”—situations in which one party takes an action that imposes costs or bestows benefits on another party but fails to account for those costs or benefits in choosing his activity level. As a result, the acting party may do too much or too little of the activity, leading to inefficient resource allocation.

Coase turned this belief on its head. He showed that any threat of inefficient resource allocation creates an opportunity for market participants to profit by internalizing the externality through private transactions. At some point, however, the costs of transacting overwhelm any associated benefits, and some “inefficiency” may persist. If transaction costs were zero, the parties would naturally negotiate for the socially optimal activity level because that would maximize their joint profits. There would be no need for government regulation.

Transaction costs are real costs, just like production costs, and they are never zero. For any given set of regulatory rules, the parties will negotiate what they privately perceive as efficient (and profit-maximizing) resource allocation. The outcome can be seen as an equilibrium in the sense that neither party has any incentive to make further adjustments given the transaction costs they face.

Rather than asking whether the overall benefits of a new rule will exceed the overall costs, in a Coasean framework the proper question, far more simply, is whether the proposed regulation is likely to reduce the parties’ costs of transacting. If so, the parties can be counted on to make efficient adjustments to the new rule based on their “knowledge of the particular circumstances of time and place.” This knowledge is fleeting, circumstantial, and inherently unavailable to outside observers, which is exactly why quantified CBA of proposed regulations is so difficult.

RBI falls far short of showing that the benefits of a best interest standard would likely exceed the costs. As a complement to CBA, the SEC should focus on showing that a change from the suitability rule to a best interest standard is “made necessary by compelling public need, such as [a] material failure[s] of private markets” (emphasis added). Under the circumstances RBI addresses, a thorough transaction cost analysis of broker-client relations provides the “best reasonably obtainable scientific, technical, economic, and other information” for assessing the “consequences of the intended regulation.”

The informational burden transaction cost analysis would place on regulators is relatively modest. It would require the SEC to identify the current transaction cost equilibrium—in the sense described above—and to show that the effect of the new rule would be to reduce the parties’ costs of transacting in the new equilibrium (with due consideration for one-time transition costs).

Antitrust scholars and courts have relied on transaction cost analysis for decades to administer competition policy under the antitrust statutes. Indeed, transaction cost analysis is widely regarded as having revolutionized antitrust law. Where federal courts once routinely condemned vertical arrangements such as exclusive retailer territories, exclusive product dealing, resale price maintenance, et cetera, they now embrace vertical relationships for reducing conflicts of interest between manufacturers and their independent retailers.

The SEC has failed to meet its statutory burden of showing that RBI promotes efficiency, competition, and capital formation. Its economic analysis falls far short of providing sufficient analytical or empirical justification for the best interest rule, and recommending its adoption is therefore premature. Rather than reducing the parties’ transaction costs, RBI would very likely increase them, limiting investors’ options as a result. This conclusion is all the more forceful owing to the enormous legal uncertainty the rule would create.

The SEC Should Withdraw RBI Pending a More Thorough Transaction Cost Analysis of Conflicts of Interest in Retail Securities Brokerage 

In a subsection titled “Broad Economic Considerations,” RBI recognizes the potential power of transaction cost analysis for assessing the likely consequences of proposed rules. Although it phrases the discussion in terms of the principal-agent costs that attend conflicts of interest, it explicitly recognizes that brokers and their clients have ongoing incentives to correct inefficiencies in their relationship as a way to profit, and that principal-agent costs impede their success.

This theoretical analysis lays an excellent foundation, but RBI fails to follow through with sufficient evidence to draw any conclusion regarding the superiority of a best interest standard over the baseline suitability rule. The proposed rule repeatedly states that brokers are conflicted when giving their clients investment advice. On one hand, under the suitability rule they might give advice that serves their clients’ best interest. On the other hand, they might compromise their clients’ best interest by adjusting their advice, more or less, to increase their own compensation.

RBI repeatedly states that it is unable to quantify either the costs or benefits of a best interest standard under various assumptions because it “lacks the necessary data.” Indeed, it flatly states that the SEC has no way to know the extent to which brokers currently serve their own interest at the expense of their retail clients. Yet it summarily concludes that the best interest standard will ameliorate broker-client conflicts. Given this empirical vacuum, any conclusion that a best interest standard would meet a CBA test or otherwise promote efficiency is pure conjecture. Its net effect could easily be to increase transaction costs.

Some investors dissatisfied with their brokers’ amended commission schedules and others whose brokers no longer offer advice will search for new providers, some brokers might recommend inefficiently rebalancing their client’s portfolio away from high-fee managed funds toward low-fee indexed funds, and owing to legal uncertainty some clients will file frivolous complaints for apparent violations of the rule. All of these actions involve added transaction costs.

RBI neglects various theoretical and empirical studies and other evidence that would help the SEC understand existing industry organization and the likely transaction costs the parties face.

Securities brokers’ primary function is to execute securities trades in the secondary market on their clients’ behalf rather than to give investment advice. They are normally paid a commission of pennies per share to do so. Full-service brokers charge premium commissions and also provide their clients with incidental advice. Why would an investor, as principal, want to overpay for securities trades?

One explanation is that good advice and the execution of securities trades are complements. The better the advice a broker provides his client, the more trades the client will do and, appropriately, the higher the broker’s compensation. Bundling may be necessary because the transaction costs investors would have to incur to assess the quality of investment advice are prohibitive owing to moral hazard. Paying brokers for (costly) advice only if they are willing to incur the cost of execution is likely an informative signal to brokerage clients. What is more, the prospect of being terminated and losing a stream of premium commissions is an effective way to bond broker performance.

Economists Kenneth Garbade and William Silbur provide a transaction cost analysis of securities brokers’ fiduciary duty of “best execution” when trading on their clients’ behalf. A sufficient fraction of clients in the marketplace know their brokers have an incentive to act selfishly, and competition will therefore set the fee schedule to encourage efficient broker behavior. The article identifies the contractual arrangements that mitigate conflicts of interest by properly motivating brokers to trade on their clients’ behalf, including reliance on limit orders, bonded agents, and collective monitoring of agent’s performance. As a general matter, the authors show that the patterns of organization in securities brokerage are consistent with the hypothesis that brokers and their clients maximize the joint gains from trade net of transaction costs.

Evidence from experimental economics suggests that mandatory disclosure of conflicts of interest can actually hurt the principal. Daylian Cain, George Loewenstein, and Don Moore find that “disclosure can increase the bias in advice because it leads [agents] to feel morally licensed and strategically encouraged to exaggerate their advice even further.” This raises an important empirical question the SEC should address before finalizing RBI.

Hamid Mehran and René Stulz provide an extensive literature review of conflicts of interest in financial services, and conclude that

"[A]lthough conflicts of interest are omnipresent when contracting is costly and parties are imperfectly informed, there are important factors that mitigate their impact and, strikingly, it is possible for customers of financial institutions to benefit from the existence of such conflicts. The empirical literature reaches conclusions that differ across types of conflicts of interest but are overall more ambivalent and certainly more benign than the conclusions drawn by journalists and politicians from mostly anecdotal evidence.

. . . The existence of a conflict of interest within a financial institution does not mean that, in equilibrium, the customers of that institution will be harmed. [A] variety of mechanisms help control conflicts of interest and their impact. For instance, a financial institution’s concerns about its reputation might lead it to control conflicts of interest so that they have no material impact on its customers. Alternatively, a financial institution’s customers can rationally take into account how these conflicts affect the financial institution’s actions [and adjust accordingly]."

The US Supreme Court’s decision and reasoning in Leegin v. PSKS belies any suggestion that brokers’ compensation necessarily leads them to disserve their clients. For over 100 years before Leegin, federal courts had summarily condemned resale price maintenance (RPM) as illegal per se, thinking it facilitated monopoly. Leegin concluded that the economic scholarship is now “replete with procompetitive justifications for a manufacturer’s use of resale price maintenance.”

With RPM, a product manufacturer contractually obligates its independent retailers to charge no less than a stated price. Rather than compete on price, retailers compete by providing point-of-sale services valuable to consumers on which the retailer might otherwise skimp, contrary to the manufacturer’s interest. Among its procompetitive effects, RPM allows the manufacturer to avoid monitoring its retailers, which reduces transaction costs.

In the Court’s words, “It may be difficult and inefficient for a manufacturer to make and enforce a contract with a retailer specifying the different services the retailer must perform.” RPM avoids the manufacturer-retailer conflict of interest. The Court concludes that “the interests of manufacturers and consumers are aligned with respect to retailer profit margins.”

Section 22(d) of the Investment Company Act of 1940 allows mutual funds to use RPM to restrict the minimum price at which brokers can sell fund shares to retail clients through load fees disclosed in the fund’s prospectus. In a representative load fee arrangement, the client pays $100 to the broker but receives an investment in the fund of only $95. The brokerage firm receives the $5 front-end load, which it splits with the individual broker according to internal compensation policy. Just as in any other retail setting, the price the client pays includes compensation for the costs of marketing the good, but in this setting the compensation is itemized and disclosed to the client.

The load-fee form of broker compensation for selling fund shares is one of the practices RBI is intended to address, and yet RBI recognizes that this form of organization has evolved efficiently over time. Depending on the share class, the industry has adopted back-end loads and even made them contingent on certain events or actions by the parties, no doubt in an effort to reduce the inefficiencies from conflicts of interest. Evidence the SEC cites suggests that these adjustments have reduced transaction costs and better aligned the parties’ incentives.

Mutual fund companies that charge minimum loads on the sale of their shares apparently believe their investors are better off paying the load and receiving the services. In making this decision, the investment company’s profit or loss depends on how the load affects its costs of capital. If broker-provided services are worth more to retail investors than the load they pay, the investment company’s cost of capital will eventually fall. If not, it will rise.

Just as manufacturer and consumer interests are aligned with respect to retailer compensation, there is no reason to distrust the investment company to make the load-fee decision in a way other than to maximize the gains from trade for given costs of transacting. The gains are then shared in some way between the parties based, in part, on their outside options. All costs are internalized to the investment company.

The free rider problem also arises for full-service brokers who provide their clients with incidental investment advice in exchange for premium commissions. A client might split his wealth between brokerage accounts at two different firms, one with a full-service broker that provides advice and valuable services but charges premium commissions and the other with a discount broker that provides no services but charges low commissions. The client might take the full-service broker’s advice and use it, in part, to make investment decisions for his discount brokerage account, thereby reducing the total brokerage commissions he pays. This reduces full-service brokers’ ability to earn a return on their investment in providing incidental investment advice.

One way full-service brokers can prevent clients from free riding is by encouraging those clients to place their entire wealth with them. Unfortunately, the transaction costs to the broker of confirming this are extremely high. Brokers address this problem by giving clients with larger accounts (which is correlated with placing all financial assets with the broker) preferential access to investment advice. Surely free riding by retail clients is relevant to understanding the likely consequences of RBI, and it is an issue the SEC staff should be able to address empirically.

As RBI recognizes but fails to refute, the best interest rule could have consequences that harm investors rather than protecting them because, at least formally, it suggests that the client’s best interest is absolute and is not to be compromised even modestly to provide benefits to the broker. The implicit assumption seems to be that the parties are competing for a fixed pie, and that any value the broker captures must come at his client’s expense.

This stance fails to come to grips with economic reality. Investment advice generates a positive externality for clients that calls on the parties to balance delicate tradeoffs to expand the pie. Since the broker receives only a tiny fraction of any benefits he generates for his clients, the “problem,” if one exists at all, is that he will have too little incentive to make intelligent recommendations.

Taken to its logical extreme, the best interest rule could be interpreted to require full-service brokers to recommend a counterproductive “investment strategy” that gives their compensation no weight in the advisory calculus. By way of example, according to the letter of the rule they could be required to advise their clients to free ride on their recommendations. Doubtless FINRA and SEC administrative law judges and federal courts would eventually find ways to rationalize the best interest rule to avoid this and other inefficient outcomes, but having evolved for decades, the suitability rule suffers little or no risk of interpretive errors.

Two points are worth noting. First, there is nothing economically suspect about brokers charging their clients for advice if it increases clients’ net asset values. Second, in noisy securities markets, advice expected to do this ex ante can look like bad or tainted advice ex post. By establishing an objective basis for broker liability, however imperfect, the suitability rule guards against hindsight bias that could have a chilling effect on brokers’ willingness to make any recommendation at all. RBI fails to explain why the benefits of the best interest standard outweigh what could be enormous costs from legal uncertainty.

Absent clear evidence of self-dealing, federal courts are unlikely to hold a securities broker liable for selling his client a popular financial product, such as load fund shares that compete head-to-head in a competitive market against no-load shares, direct stock ownership, bonds, commodities, et cetera. The same is true of brokers who accept so-called revenue sharing payments from mutual fund advisers in exchange for marketing their proprietary funds, which they receive in lieu of lump-sum loads.

In re Morgan Stanley DW, Inc. (2003) supports this conclusion. There, the SEC claimed delayed revenue sharing payments that Morgan Stanley’s brokers received from fund manager Massachusetts Financial Service reflected a material conflict of interest that should have been disclosed to clients in more specific detail. Even though there were no serious investor complaints, the SEC found that Morgan Stanley had engaged in securities fraud. The SEC then ordered it to cease and desist; censured it; assessed it $50 million in disgorgement, prejudgment interest, and civil penalties; and required it to satisfy a list of 28 ongoing compliance plans. Yet in 9 of 11 civil suits that followed the SEC enforcement action, federal courts dismissed the case on the pleadings in favor of the defendant because they found the payments were immaterial to the average investor.

One notable but overlooked feature of the arrangement was that the brokers who sold fund shares were paid only a portion of their compensation on sale. The remainder was back-end loaded in the form of trailing payments that continued for one year as long the clients retained their shares, terminating the moment they sold. If, as seems likely, some clients discover and sell unsuitable investments within a year of buying, this method of broker compensation mitigated any tendency brokers might have had to recommend ill-suited funds, somewhat akin to a guarantee. This had the effect of reducing clients’ up-front transaction costs of assessing their broker’s investment advice. Clearly, the market has continually developed compensation mechanisms that encourage brokers to provide suitable investment advice.

In January 2011, the SEC published its Study on Investment Advisers and Broker-Dealers (SEC Study) recommending that securities brokers be held to the same fiduciary standard as investment advisers. It relies heavily on a 2008 RAND Corporation report assessing retail investors’ understanding of various details regarding their investment accounts based on survey evidence. From this the SEC Study concludes that retail investors generally are unaware or confused “regarding the roles, titles, and legal obligations of investment advisers and broker-dealers, although [the report also finds] that investors generally were satisfied with their financial professionals.”

The most plausible economic inference from this observation is not that it reflects a problem to be solved but that most retail investors consider brokerage and advisory accounts to be sufficiently close substitutes and sufficiently unproblematic, and that it is not worth their time and effort to identify the finer differences between them.

To assume, as RBI does, that markets work well only if every consumer and producer cognitively calculates benefits and costs with full information is contrary to the vast weight of the evidence. As in any retail setting, for markets to work well it is sufficient that some investors—those who are on the margin between using a broker or an adviser—do their homework and allocate their funds accordingly.

Few consumers know or care whether their soft drinks are made from cane sugar or corn syrup, for example, but a select few are knowledgeable and sensitive to the difference. It is these marginal consumers’ decisions that guide relative prices and other terms of trade in the market for soft drinks. There is no reason to believe that the market for investment advice is fundamentally different, although it may suffer from unique informational problems. RBI recognizes this but fails to address the issue empirically.

Brokerage firm clients often maintain both a brokerage account and an advisory account. Many transfer funds between accounts from time to time, and brokerage firms normally make this easy and cheap to better compete for business. It is very likely that assets move fluidly between broker and adviser accounts as their relative merits vary. This is an empirical question, of course, but one the SEC is fully equipped to address. Rather than relying on pure conjecture, it should map out the contours of investor decision-making in response to outside events before approving the best interest standard. That is to say, it should build a testable theory and test it before proceeding.

There is no such thing as a conflict-free transaction. Pay a shop-floor worker by the hour and he has a tendency to loaf and pad his hours. Pay him by the piece and he will abuse the employer’s equipment to increase his output and compensation. The adviser-investor relationship is no different. Rather than charging a brokerage commission, advisers normally charge a periodic asset-based fee. For small account holders who plan to trade infrequently, advisory accounts are expensive compared to full-service brokerage accounts.

It is widely recognized that an adviser might charge a high fee to cover the costly effort it promises for actively managing the portfolio but refrain from making the effort and simply index it. What is more, investment advisers are authorized to use portfolio assets to pay brokerage commissions. This can include paying full-service brokers a high commission in partial exchange for investment advice or research—so-called soft dollars, an arrangement many consider fraught with conflicts of interest. In noisy securities markets, it is difficult for investors to hold advisers liable for these and other forms of misconduct even with their fiduciary status.

RBI asserts that the best interest rule should be mandatory, legally prohibiting the parties from contracting around it. And although it theoretically identifies narrow conditions in which a mandatory standard can be efficient, it provides no evidence that these conditions plague the suitability rule. If they do not, economic organization in the industry will inevitably evolve to reduce any ill effects while avoiding the appearance of explicit contracting, further increasing the costs of transacting.

Rather than summarily concluding that the industry’s current contractual arrangements are fraught with moral hazard to the detriment of retail investors, the SEC should begin by identifying (1) how brokers and their retail clients have adjusted to minimize any inefficiencies in the provision of investment advice and (2) what transaction costs prevent them from doing even better.

This is not to say private markets solve all problems or that government regulation is incapable of improving organizational efficiency. It is simply to say that RBI is justified only if, and to the extent, the SEC can show that it reduces the parties’ transaction costs. Sensible regulation must be premised on understanding why, and under what current circumstances, observed market practices reflect an equilibrium determined in part by the costs of transacting and how government regulation might make things better.

For the foregoing reasons, I recommend that the SEC withdraw consideration of RBI until it further studies the retail broker-client relationship using transaction cost analysis and thereby shows that RBI is likely to reduce the parties’ costs of transacting.