Introduction
As “software eats the world,”
the law must adapt legal frameworks that were designed for traditional businesses to new, technology-based business models. In the financial services sector, the emergence of robo-advisors—online services that use algorithms to generate investment recommendations for clients
—has raised questions regarding the regulation of digital advice. Regulators must grapple with whether entities that provide algorithmic investment recommendations can fulfill the fiduciary obligations
imposed on investment advisers under the Investment Advisers Act of 1940 (Advisers Act),
the primary federal statute governing investment advice.
This question grows in importance as robo-advisors become more popular. Industry professionals recognize that robo-advice technology will revolutionize how individuals receive investment advice.
In the past, the high cost of financial advice made such services inaccessible to all but the very wealthy.
By replacing human advisers with algorithms, robo-advisors are able to charge significantly less than traditional wealth management services, making them an appealing option for young investors and others with low account balances.
Since the first major services launched in 2010, the robo-advice market has grown quickly, accumulating nearly $45 billion in assets under management (AUM).
Experts expect the market to continue to skyrocket: A particularly aggressive projection predicts that robo-advisors will have $2.2 trillion in AUM by the year 2020.
Establishing a suitable regulatory scheme for robo-advisors is critical to their long-term viability. In monitoring these products, the U.S. Securities and Exchange Commission (SEC) must strike an optimal balance between protecting investors and allowing robo-advisors the latitude to innovate and develop. Statements from industry professionals and regulatory agencies and articles in the press have criticized the quality of robo-advice recommendations and have indicated skepticism that robo-advisors, as they currently exist, could ever meet the fiduciary standards of the Advisers Act.
This Note argues that such criticism does not accurately reflect the state of the law governing traditional investment advice and that robo-advisors are structurally capable of meeting the requirements of the Advisers Act.
Rather than concentrating on evaluating the quality of robo-advisor advice, regulators should instead focus on policing robo-advisor conflicts of interest.
This Note proceeds in three Parts. Part I provides background on the obligations investment advisers are held to under the Advisers Act. Part II considers robo-advisors, first introducing the product and business model, then analyzing whether robo-advisors are capable of meeting the duty of care standards of the Advisers Act, and finally overviewing conflict of interest issues in robo-advisors. Part III argues that, in regulating robo-advisors, the SEC should shift its focus away from the quality of robo-advisor recommendations and instead promulgate a rule that would make robo-advisor conflict of interest disclosures more transparent.
The Regulation of Investment Advisers
This Part provides an introduction to the laws that regulate investment advisers.
Section I.A provides historical background explaining how Advisers Act law evolved into its current fragmented state. Section I.B details how the landmark case SEC v. Capital Gains Research Bureau, Inc.
read an investment adviser fiduciary duty into the Advisers Act. Finally, section I.C details several of the specific obligations found within that fiduciary duty.
A. The Advisers Act: Consequences of a Rushed Enactment
This section begins by showing how the Advisers Act’s haphazard passage resulted in a statute of limited scope. It goes on to explain that, while the law has since developed to fill gaps left by the statutory text, the variety of mechanisms used has resulted in uneven law.
1. The Act’s Passage. — The Advisers Act is commonly acknowledged to be the weakest of the New Deal federal securities statutes.
This is, in part, due to its origins. In 1935, the Public Utility Holding Company Act (PUHCA) directed the SEC to conduct a study on investment companies and investment trusts.
In that study, the SEC detected potential investment adviser abuse; however, because its results were already years overdue by that time, the SEC did not delve further into the issue.
The PUHCA study’s cumulative findings prompted the Senate to introduce a bill that was eventually separated into two acts: the Investment Company Act of 1940,
which regulates companies that invest and trade in securities and companies that offer their own investment products to the public, and the Advisers Act.
Of the two statutes, the Advisers Act was the lesser priority.
In addition to being backed by less research,
the Advisers Act faced strong opposition from coalitions of investment advisers as Congress debated it.
This vocal opposition came after weeks of grueling negotiations over the Investment Company Act, and Congress was worn and eager to finalize the text.
To this end are anecdotes of the PUHCA study chief counsel telling Congress to “throw in the sponge” and “write a simple bill that . . . we can all agree on.”
These dynamics led to a statute that, on its face, appears extremely limited in scope.
As enacted, the Advisers Act covered only registration, disclosure, and fraud prevention; it imposed no further obligations on investment advisers and gave the SEC little enforcement power.
Later amendments have expanded and refined the SEC’s jurisdiction some,
but the statute remains structurally the same today as in 1940.
2. Filling (Some of) the Gaps Through Interpretation. — To bolster investment advice law and protect retail investor clients, courts and the SEC have taken a piecemeal approach to “fill[ing] the statute’s gaps.”
This approach has caused its own problems, however: Mechanisms used to develop Advisers Act law include, but are not limited to, federal court cases,
interpretive releases,
the SEC’s bully pulpit,
no-action letters,
and enforcement actions.
Because of this, there is no single repository of all requirements investment advisers must follow.
The law is scattered and standards are unclear.
The SEC’s heavy reliance specifically on enforcement actions (prosecutions in administrative court) to develop Advisers Act law further exacerbates the ambiguity in the law’s standards. Enforcement actions are difficult to interpret and apply for a number of reasons. For one, due to how the SEC allocates its resources, it generally takes action against only egregious violations.
This leaves the law thin in gray-area situations.
Next, enforcement actions against minor violations generally settle, meaning their results are never subjected to the SEC’s or a court’s independent critical analysis and do not contribute to the body of investment advice law.
Finally, enforcement actions are tied to very particular sets of facts but have less legal analysis than court opinions; this makes them more difficult to apply to other situations.
B. Capital Gains and the Investment Adviser Fiduciary Duty
The most significant development in Advisers Act law has been the creation of the investment adviser fiduciary duty. The Supreme Court read this duty into the Advisers Act in 1962 through its first case interpreting the statute, SEC v. Capital Gains Research Bureau, Inc.
The Capital Gains case involved a registered investment adviser engaging in the practice of “scalping.”
In short, the adviser purchased large blocks of stock it intended to recommend, recommended those stocks to clients until the stocks increased in value, and then sold its own blocks to profit from the higher stock prices.
The Court found that, although the practice of scalping does not constitute common law fraud, it violates Section 206 of the Advisers Act,
the Act’s antifraud provision. Section 206 states in relevant part that
[i]t shall be unlawful for any investment adviser by use of the mails or any means or instrumentality of interstate commerce, directly or indirectly—
(1) to employ any device, scheme, or artifice to defraud any client or prospective client;
(2) to engage in any transaction, practice, or course of business which operates as a fraud or deceit upon any client or prospective client . . . .
The Court justified its interpretation by reasoning that, in enacting the Advisers Act, Congress intended investment advisers’ relationships with clients to be closer than that of an ordinary arm’s-length transaction.
As Justice Goldberg wrote for the Court, “the Committee Reports indicate a desire to preserve ‘the personalized character of the services of investment advisers,’ and to eliminate conflicts of interest between the investment adviser and the clients as safeguards both to ‘unsophisticated investors’ and to ‘bona fide investment counsel.’”
In other words, Congress intended for unsophisticated investors to be able to place a high degree of trust in their investment advisers. Later in the opinion, Justice Goldberg finds that, given the legislative intent, the Advisers Act should be read “not technically and restrictively, but flexibly to effectuate its remedial purposes.”
To be consistent with this, the term “fraud” in Section 206 must be read more broadly than its common law meaning.
Investment advisers, like other fiduciaries, have an “affirmative duty of ‘utmost good faith and full and fair disclosure of all material facts,’”
and “to employ reasonable care to avoid misleading . . . clients.”
Activity that breaches this affirmative duty—such as scalping—is therefore fraud under Section 206 and unlawful.
Justice Goldberg never explicitly called this affirmative duty an “investment adviser fiduciary duty” anywhere in the Capital Gains opinion,
but subsequent Supreme Court decisions have read his reasoning in this way. In Santa Fe Industries v. Green, which dealt with a different securities statute, the Court stated in a footnote that Capital Gains recognized “that Congress intended the Investment Advisers Act to establish federal fiduciary standards for investment advisers.”
Later, in the Court’s second interpretation of the Advisers Act, Transamerica Mortgage Advisors, Inc. v. Lewis, the Court more strongly solidified the investment adviser fiduciary duty, writing “[a]s . . . previously recognized, § 206 establishes ‘federal fiduciary standards’ to govern the conduct of investment advisers.”
Since then, countless court cases and SEC enforcement actions have read a fiduciary duty into Section 206, and it is commonly accepted that this duty originated from Capital Gains.
C. Obligations Within the Investment Adviser Fiduciary Duty
Since Capital Gains, the SEC has used numerous lawmaking mechanisms to define the bounds of the investment adviser fiduciary duty.
Fiduciary duties encompass both duty of care and duty of loyalty obligations.
Section I.C.1 focuses on the duty of care by discussing how the SEC regulates investment adviser competence and quality; section I.C.2 focuses on the duty of loyalty by discussing how investment adviser conflicts of interest are regulated.
1. Quality and Competence (Duty of Care) Obligations. — The SEC has not used its notice-and-comment rulemaking authority to develop investment adviser duty of care obligations,
and there is limited law governing the quality and competency of investment adviser advice. There are no federal standards requiring investment advisers to have credentials of any sort
—the SEC’s position is that clients should evaluate for themselves whether an adviser has the competence to manage their assets effectively.
Further, while Section 206 provides for two obligations that do stem from the duty of care—suitability and best execution
—the SEC does not enforce either to a rigorous standard.
Suitability requires investment advisers to reasonably determine that the advice they give is appropriate to a client’s circumstances.
For guidelines on the suitability obligation, attorneys commonly look to a rule the SEC proposed in 1994 but later abandoned.
Lawyers look to this abandoned rule because the rule’s introductory text states that the rule would only “make explicit advisers’ suitability obligations under the Advisers Act.”
Under that rule, advisers would need to “make a reasonable inquiry into a client’s financial situation, investment experience, and investment objectives”;
in addition, advisers must update client information regularly, so that their advice can be adjusted to changing circumstances.
Empirically, however, the SEC has not enforced suitability to the full extent of this language. It brings enforcement actions on suitability grounds rarely and only in severe cases—such as when an adviser is taking out margin loans and purchasing speculative, high-risk stocks on the accounts of clients with conservative investment objectives.
The second duty of care obligation is best execution.
Put simply, when an investment adviser selects a broker-dealer to execute the transactions she recommends, she must seek to ensure that the client’s total costs are “the most favorable under the circumstances.”
When determining what is “most favorable,” investment advisers can consider transaction costs, execution capacity, financial solvency of a brokerage firm, and the value of any research.
As with suitability, the SEC also does not enforce best execution aggressively. It generally will not take action unless advisers are failing to best execute in order to benefit themselves.
2. Conflict of Interest (Duty of Loyalty) Obligations. — Investment adviser law is far more rigorous in its governance of the investment adviser duty of loyalty. This is consistent with the fact that the Capital Gains Court was addressing a duty of loyalty issue when it created the Advisers Act fiduciary duty.
In fact, Capital Gains stated that Congress enacted the Advisers Act with the intent to address “all conflicts of interest which might incline an investment adviser—consciously or unconsciously—to render advice which was not disinterested.”
As is common in U.S. securities law, investment adviser conflicts are governed through a “disclosure-based” regime, rather than a “merit-based” one.
Investment advisers are permitted to have interests not in line with their clients’ interests (e.g., in the form of bonuses, commissions, or personal relationships), but if a conflict is material, they must disclose it fully and accurately.
The justification for this is that it could be in a client’s interest to use a conflicted investment adviser—perhaps the fact that her adviser receives outside commissions could lower a client’s fees—but for a client to make an educated choice, she must have full information about the conflicts.
To ensure full information, the SEC strictly enforces the disclosure requirement: There is no waiver for conflicted investment advisers who believe in good faith that, despite the conflict, they still put their clients’ interests first;
for advisers who take adequate internal precautions to address conflicts;
or for advisers who never acted upon a conflict.
The SEC has promulgated specific rules dictating how investment adviser conflicts should be disclosed. Investment advisers file a Form ADV when they register with the SEC.
They must update the form annually and more frequently if significant changes occur.
Part 1 of the Form ADV is primarily for SEC use and is formatted as a check-the-box, fill-in-the-blank form.
It asks questions regarding an adviser’s business, ownership, clients, employees, business practices, and disciplinary past.
Part 2 of the Form ADV is divided into a brochure (Part 2A) and brochure supplement (Part 2B).
The brochure has nineteen items.
In it, advisers must describe, in plain English,
much of the information they disclosed in Part 1.
Finally, the brochure supplement provides information about the professionals working with a client’s account.
Investment advisers must deliver both the brochure and brochure supplement to clients before or at the time investment advisers and clients begin their contractual relationship.
Afterward, advisers must update the documents and provide clients with a summary of material changes every year.
Including untrue statements in or omitting material facts from any of these documents breaches the Section 206 duty of good faith and violates Section 207 of the Advisers Act.
II. Issues of Regulating Robo-Advisors
As is evident from Part I, the statutory scheme governing investment advice was created with human advisers—how they are motivated and make decisions—in mind. The growing popularity of an automated alternative, the “robo-advisor,” presents a complication. Robo-advisors are powered by computers and algorithms.
This means not only that they lack the human judgment that traditional investors possess, but also that they cannot be illicitly induced. Part II of this Note analyzes the regulatory issues these products present and assesses how the SEC should address them.
Section II.A begins with an overview of the robo-advisor product and market. Section II.B addresses whether robo-advisors, as a technology, can exercise enough care to meet the investment adviser duty of care standard. Specifically, subsection II.B.1 presents the narrative that robo-advisors cannot meet this standard; subsection II.B.2 lays out an argument that they can. Lastly, section II.C looks at duty of loyalty issues through an analysis of how conflicts of interest affect robo-advisors.
A. An Introduction to Robo-Advisors
Before diving into the regulatory issues robo-advisors present, this section provides background on robo-advisors. Section II.A.1 begins with a more detailed explanation of the product, focusing on the characteristics that differentiate robo-advisors from traditional advisers. Section II.A.2 surveys the major players in the market and explains that money manager robo-advisors are increasingly dominating the market.
1. The Product. — Robo-advisors are automated services that provide investment advice through web or mobile platforms.
In contrast to traditional investment advisers, robo-advisors rely primarily on algorithms, rather than human judgment, to determine recommendations.
Clients fill out questionnaires with information such as age, household situation, income, savings, financial goals, and risk tolerance.
This information is put through a computer algorithm, which calculates an investment portfolio that is efficient and tailored to a client’s needs.
Because this model limits robo-advisors from making truly bespoke recommendations, robo-advisors primarily rely on passive indexing and diversification strategies
and utilize exchange-traded funds (ETFs) that track broad market benchmarks.
Without the costs associated with providing human advice, robo-advisors can charge significantly lower fees than their human counterparts.
As robo-advisors actively advertise,
their automated systems enable them to take advantage of certain strategies—in particular, threshold-based rebalancing and tax-loss harvesting. Rebalancing realigns how a portfolio is weighted to reduce drift from the original target allocation.
When an investment adviser begins working with a client, she determines the investment mix that would maximize returns given the client’s risk tolerance and allocates assets accordingly.
But as different asset classes perform differently over time, the client’s portfolio may “drift” from the predetermined allocations.
For example, if emerging market funds are performing especially well and generating high returns, a larger percentage of a client’s portfolio will become weighted in emerging market assets (and, consequently, a smaller percentage in other asset categories). The portfolio would eventually need to be readjusted to reflect the allocation percentages the investment adviser originally calculated to be ideal.
Because it is not feasible for human investment advisers to continuously monitor all of their clients’ portfolios, human advisers primarily rebalance at predetermined time intervals (“time-based rebalancing”).
Robo-advisors have the advantage of being able to program continuous monitoring into their algorithms.
They can therefore rebalance automatically when allocations hit certain percentages (“threshold-based rebalancing”).
This capability minimizes transaction costs and helps ensure that investment allocations continuously reflect client goals.
Tax-loss harvesting is the strategy of selling securities carrying losses at strategic points in time in order to realize capital losses that can offset other income.
After selling a security to realize said losses, an investment adviser will replace the sold security, often an ETF, with a similar asset to preserve the original asset allocation.
By using algorithms, robo-advisors can capture tax-loss harvesting opportunities more consistently than human advisers.
This is especially true because the so-called “wash-sale rule” prohibits taxpayers from claiming a loss on the sale of a security if they purchase a “substantially identical” security thirty days before or after the sale.
This rule makes manual tax-loss harvesting more difficult, but, through the use of technology, robo-advisors are able to identify replacement ETFs that are highly correlated, but not technically “substantially identical,” perhaps because the replacement fund tracks a different index.
2. Market Players. — The initial robo-advisors were independent, venture-backed start-up companies. Betterment became the first major player in the market when it launched in 2010; it has since raised over $200 million in equity funding
and has over $9 billion in AUM.
Wealthfront, the second largest independent robo-advisor, has raised about $130 million in equity
and has over $6 billion in AUM.
Other independent robo-advisors include Personal Capital, AssetBuilder, SigFig, and WiseBanyan.
The services these companies provide all vary some in their fee structures,
questionnaire structures,
amount of human support provided,
and investment strategies.
Seeing the success of independent robo-advisors, traditional money managers have more recently begun launching robo-advice services that work in tandem with the products and services they conventionally offer.
Charles Schwab and Vanguard were the two pioneers. Schwab entered the space in March 2015, when it launched Schwab Intelligent Portfolios (SIP).
SIP uses Schwab’s own ETFs and cash allocation programs and is free for Schwab clients.
Shortly after, in May 2015, Vanguard introduced its Personal Adviser Services platform, which charges thirty basis points and supplements its robo-advice with human advice through phone or video chat.
Benefitting from the size of its existing asset base, Vanguard’s service amassed $31 billion in AUM by the end of 2015
and almost $51 billion by the end of 2016,
far outstripping the size of all previously existing robo-advisors combined.
Other money managers have since followed suit: BlackRock acquired FutureAdvisor, a previously independent robo-advisor, in 2015;
Fidelity Investments and TD Ameritrade launched Fidelity Go and TD Ameritrade Essential Portfolios, respectively, in 2016;
Merrill Lynch released its service, Merrill Edge Guided Investing, in early 2017, and both Wells Fargo and Goldman Sachs have also announced plans to launch robo-advisors.
Money manager robo-advisors have given independent robo-advisors stiff competition.
Because names like Charles Schwab and Vanguard have the benefit of widespread brand recognition, they do not need to invest as significantly in marketing.
With this advantage, experts predict that money manager robo-services will completely overtake the market.
Supporting this are the facts that the growth rates of independent robo-advisors have been falling since mid-2015,
and traditional money manager robo-advisors are now the primary driver of robo-advisor asset growth.
Because money manager robo-advisors operate under a very different business model than independent robo-advisors,
this trend will have a significant impact on how the product develops.
B. Competence and Quality (Duty of Care) Issues
This section evaluates whether robo-advisors, as a technology, can meet the duty of care standards to which the law holds traditional, human investment advisers. Section II.B.1 summarizes arguments from the popular press, finance professionals, and regulatory agencies that robo-advisors are not structurally capable of fulfilling the investment adviser duty of care. Section II.B.2 then lays out an argument that, given the law on fiduciary duties and how the SEC has treated the investment adviser duty of care, robo-advisors are in fact capable of exercising enough care to meet fiduciary standards.
Before beginning, it is important to distinguish the analysis in this section from a more general discussion on how robo-advisors perform relative to human advisers. A common line of criticism is that robo-advisors perform worse than their human counterparts,
but there are also some compelling arguments that robo-advice is as good as, if not better than, human advice. For one, robo-advisors are able to rebalance and tax-loss harvest more efficiently than human advisers.
In addition, some traditional investment advice services are very basic, and robo-advisors can provide the same services at a fraction of the cost.
Lastly, robo-advisors are less impacted by emotional and cognitive biases than human advisers are.
This Note does not attempt to draw conclusions about these arguments, because an assessment of the relative performance of human versus algorithmic recommendations would be better suited for a finance and economics paper.
This section focuses only on the legal question of whether robo-advisors are capable of meeting the law’s duty of care requirements. For its purposes, comparing robo-advice with the quality of human advice, as well as analyzing the SEC’s past regulation of human advice, is relevant only insofar as such analysis indicates what the SEC should or should not permit with regard to robo-advisors.
1. The Narrative that Robo-Advisors Cannot Meet Duty of Care Standards. — There has been significant resistance to the regulatory acceptability of robo-advisors. Traditional financial advisers have put forth a number of articles arguing that the robo-advisor design is not capable of exercising enough care to meet the fiduciary standards of the Advisers Act.
For example, securities attorney Melanie Fein has written two oft-cited white papers arguing that robo-advisors do not meet a fiduciary standard of care.
The SEC and the Financial Industry Regulatory Authority (FINRA), the private regulatory body that oversees broker-dealers, have yet to take a strong position on robo-advisors, but the literature they have released has thus far been cautionary.
At the state level, the Massachusetts Securities Division has come out aggressively against robo-advisors, stating that “it is the position of the Division that fully automated robo-advisers, as currently structured, may be inherently unable to carry out the fiduciary obligations of a state-registered investment adviser.”
The arguments against robo-advisors’ regulatory acceptability center on three closely linked arguments. The first relates to the limitations of using questionnaires to extract customer information. Critics believe that using an electronic questionnaire to gather information about a client is not sufficient to satisfy the investment adviser duty of care.
An initial issue is that preset questions can miss vital information. The SEC and FINRA issued a joint Investor Alert on Automated Investment Advice, which stated that robo-advisor questions may be “over-generalized, ambiguous, misleading, or designed to fit [a client] into the tool’s predetermined options.”
Because of this, a robo-advisor’s recommendation may fail to account for factors such as an investor’s experience, time horizon, cash needs, and financial goals.
Next, questionnaires generally do not gather information on assets outside of a client’s account. This is problematic because, as the Massachusetts Securities Division puts it, “assets held outside of a client’s account directly impact the client’s total financial picture and, accordingly, the investment adviser’s ability to personalize advice and make appropriate investment decisions.”
Lastly, robo-advisors rely solely on the information gathered through a questionnaire; they do not confirm whether the information clients provide is accurate.
The second argument relates to the fact that robo-advisors lack human perception.
The SEC and FINRA have flagged lack of human judgment and oversight as a potential issue for automated investment tools.
And critics of robo-advisors—from certified financial planners,
to Massachusetts Secretary of the Commonwealth William Galvin,
to Rutgers Law Professor Arthur Laby
—have implied that human connection and judgment are essential elements of the investment adviser fiduciary duty. They posit that only humans can connect with clients on a personal enough level to fully understand a client’s financial situation.
Robo-advisors are likely to miss the subtleties of a client’s situation that arise in conversation.
Third, and related to the lack of human perception, is the argument that robo-advisors cannot be fiduciaries because they are not equipped to address market failures.
Then-SEC Commissioner Kara Stein raised this position in a 2015 lecture, asking, “What does a fiduciary duty even look like or mean for a robo advisor? . . . Do investors using robo advisors appreciate that, for all their benefits, robo advisors will not be on the phone providing counsel if there is a market crash?”
While markets have been strong since robo-advisors first gained traction,
there is concern over how robo-advisors will function if and when there is an economic downturn. Critics argue that, in times of crisis, clients need a human adviser to talk them through decisions so that they do not take rash actions detrimental to their own long-term interests.
As a Wall Street Journal opinion article puts it, “An email or text message in the fall of 2008 would not have sufficed to keep millions of panicked savers from selling, with devastating consequences for their nest eggs.”
As an illustration, Betterment experienced this effect in the aftermath of “Brexit” when it halted all trading on its platform for about two and a half hours the morning after the U.K. vote to exit the European Union.
Betterment justified this decision as an effort to protect clients from making panicked decisions that would result in poor trade execution and higher transaction costs.
However, because it communicated the trading suspension poorly, many of its clients did not realize that transactions they put in that morning would not be executed until hours later, after the Betterment trading team deemed markets to have normalized.
The company received significant backlash as a result,
and the incident illustrates the limitations an automated system may have in times of crisis.
2. Robo-Advisors Can Meet Duty of Care Standards. — The arguments in subsection II.B.1, while appealing on a practical level, presume a higher and more rigid standard of care than exists in Advisers Act law. This subsection applies common law principles and interpretations of the Advisers Act to conclude that the investment adviser fiduciary duty of care is more lenient than robo-advisor critics recognize and that a well-designed robo-advisor meets the standard without issue. The discussion first establishes that fiduciary duties—and the investment adviser fiduciary duty specifically—are meant to be flexible; part of that flexibility includes having the option to adapt out certain investment adviser functions. Human advisers frequently do this, and robo-advisors should be able to as well. It next explains that, even if the investment adviser duty of care were more rigid, fiduciary duties can be modified when certain conditions are met. Robo-advisors meet these conditions and thus warrant a modification to the standard.
First, both common law and investment advice law hold that that the investment adviser fiduciary duty should be read flexibly, with the context of the investment adviser–client relationship in mind. Common law has long established that fiduciary duties are not a rigid package of obligations; the concept is flexible, and its bounds depend on the exact relationship between the fiduciary and the entrustor.
This flexibility exists because fiduciary duties have a practical purpose—to be a gap-filler in situations in which a fiduciary and an entrustor would not otherwise interact.
Consistent with this purpose, the extent of a fiduciary’s obligations under common law depends on the dynamics of a relationship,
meaning different types of fiduciaries have different levels of obligation.
Relative to investment advisers specifically, both the Senate and the House versions of the original Advisers Act bill recognized that the investment adviser–client relationship should be “personalized” and dependent on the circumstances of the agreement between the two parties.
Capital Gains solidified this idea into case law when the Court chose to read the Act “remedially”
to “preserve ‘the personalized character of the services of investment advisers.’”
Thus, the investment adviser fiduciary duty is more adaptable and the minimum requirements for investment advisers are lower than the arguments in subsection II.B.1 assume. Consistent with this, the SEC routinely permits advisers to adapt out common investment adviser functions so long as the terms between the parties are clear. For instance, advisers are permitted to prepare financial plans solely relating to a client’s investment circumstances at one point in time (and disclaim responsibility for updating the information on an ongoing basis); they can also provide advice on only one segment of a client’s portfolio (and disclaim responsibility for managing the client’s remaining assets).
Relative to robo-advisor regulation, no authority of law establishes that a comprehensive information-gathering process and human judgment are necessary elements of the investment adviser duty of care. Moreover, the lack of these elements is clear to clients when they choose to engage a robo-advice service. Robo-advisors should accordingly be able to adapt out these elements and still meet the investment adviser duty of care standard. There is some indication that the SEC is coming around to this position; during a 2016 speech, then-SEC Chair Mary Jo White said,
Providing financial advisory services electronically is different than the traditional adviser model, but in many respects our assessment of robo-advisors is no different than for a human-based investment adviser. Just like a conversation with a “real person” about a client’s financial goals, risk tolerances, and sophistication may be more or less robust, so too there is variation in the content and flexibility of information gathered by robo-advisors before advice is given.
Second, even if regulators were to accept that robo-advisors fall short of the investment adviser duty of care standard, robo-advisors would be ideal candidates for modifying the standard. Within industries in which fiduciary obligations are well defined, fiduciary duties can still be thought of as “default rules” to be modified based on the circumstances of a specific fiduciary–entrustor relationship.
Fiduciary law scholar Tamar Frankel has observed that under common law, the bounds of fiduciary obligations can be modified when five conditions exist: (1) The entrustor has independent will such that she can properly enter a contract; (2) when conflicts of interest exist, the entrustor has full information about the conflicts; (3) the fiduciary provides the entrustor with notice of the modification; (4) the substance of the modification is fair to the entrustor; and (5) the entrustor gives clear and specific consent to the modification.
Robo-advisors meet all five criteria without issue: Their clients have independent will in choosing between services, and robo-advisors are required to disclose full information about all conflicts through their Form ADVs.
There is adequate notice: When clients seek robo-advice, they know that the resulting recommendations are based only on the information they entered into an online questionnaire and often made without the benefit of human judgment. The substance of the duty of care modification is also fair: Clients accept robo-advisors’ limitations in exchange for the lower prices robo-advisor services charge and for the above-discussed advantages they offer.
Finally, a client’s consent is specific because clients engage the service knowing the capabilities and limitations of robo-advisors. Thus, with each of Frankel’s criteria met, it follows that, even if a rigid investment adviser duty of care standard existed, the characteristics of robo-advisors would warrant modifying that standard.
C. Conflicts of Interest (Duty of Loyalty) Issues
The previous section established that robo-advisor duty of care issues are less pressing than critics contend. This section turns to the other core fiduciary duty, the duty of loyalty, which the SEC regulates more actively.
The discussion here begins by debunking the narrative that robo-advisors cannot be conflicted. It then illustrates how these conflicts can occur by overviewing the two types of conflicts most commonly disclosed in robo-advisor Form ADV brochures.
First, despite pushing a narrative that their algorithmic approach makes them more impartial than traditional human-based adviser services,
robo-advisors can and frequently do face conflicts of interest. Robo-advisors push the narrative that they are less prone to conflicts because robo-advice services, particularly those with no human support element, are able to avoid the “representative level” conflicts that result when human judgment is involved.
Their reasoning behind this narrative is that algorithms will never be tempted by commission incentives or by personal relationships to make recommendations that are not in a client’s best interest. This proposition has received some government support: In advocating for the Department of Labor fiduciary rule in 2015,
DOL Secretary Thomas Perez lauded robo-advisors as exemplary, low-cost investment advice services that act in the best interests of clients.
The notion that robo-advisors are unbiased is problematic, though, because it considers only employee–client conflicts (e.g., an individual employee receiving a kickback when recommending a certain bank’s products), and not firm–client ones (e.g., an entire firm receiving more compensation when recommending certain products over others).
By removing employee representative discretion from the advice they give, robo-advisors do eliminate the possibility that an individual employee’s incentives may conflict with clients’ interests. However, the possibility of firm–client conflicts persists.
Robo-advisor algorithms can be programmed to prioritize what is best for the firm, rather than what is best for a client.
Even when not done intentionally, the humans who design robo-advisor algorithms may be influenced by firm incentives,
and this could cause them to subconsciously bias algorithms to reflect firm–client conflicts.
Next, Form ADV brochure disclosures confirm the existence of robo-advisor firm–client conflicts. Robo-advisors most commonly disclose two types of conflicts: (1) utilizing an affiliated broker-dealer and (2) promoting affiliated services and products. The below subsections describe each in turn.
1. The Affiliated Broker-Dealer. — A number of robo-advisors have an affiliated broker-dealer, owned by the same parent, that executes all the transactions the robo-advisor recommends.
This practice is not unique to robo-advisors—almost twenty percent of all investment advisers have an affiliated broker-dealer, and the SEC permits the practice
—but it is nonetheless concerning to client interests. As a sample, Betterment’s brochure states that
clients must establish a brokerage relationship with our affiliated broker-dealer, Betterment Securities . . . . [C]lient authorizes and directs Betterment to place all trades in client’s account through Betterment Securities . . . . Clients should understand that the appointment of Betterment Securities as the sole broker for their accounts under this Wrap Fee Program may result in disadvantages to the client as a possible result of less favorable executions than may be available through the use of a different broker-dealer.
Thus, all client transactions are executed through Betterment’s affiliate, regardless of whether this is in a client’s best interests.
This benefits Betterment because its affiliated broker-dealer both collects a fee for the execution and also profits from the bid–ask spread.
The execution fee the affiliate collects is not an issue for Betterment clients; they, like most robo-advisor clients, are charged through a wrap-fee model, meaning they pay a flat, asset-based fee for all advisory, brokerage, and custody services.
This fee stays the same, regardless of the execution fee. The profits the affiliate makes from the bid–ask spread are concerning, on the other hand, because they can affect client returns. Betterment Securities has an incentive to quote less favorable prices than other broker-dealers so that it can profit from the spread,
and it is not deterred from doing so because, as the brochure establishes, Betterment clients cannot select a different broker-dealer without changing advisory services.
If Betterment clients pay more for assets than other investors in the market, their returns will be lower.
2. Promoting Affiliated Services and Products. — Robo-advisor services run by traditional money managers have additional troubling conflicts. These services generally charge clients a very low advisory fee. Instead of relying on these fees, the money manager profits when the robo-advisor recommends another one of the money manager’s proprietary products or services. These recommendations are problematic when they do not align with client interests.
As one example, Schwab advertises its robo-advisor, SIP, as a service with “$0 advisory fees, account service fees or commissions.”
However, when the service first launched, SIP’s brochure disclosed that a significant portion of every client’s portfolio—seven percent to thirty percent—would be allocated to cash and invested into Schwab’s retail banking service.
SIP allocated this portion to cash even when a smaller percentage would have been ideal for a client, likely because Schwab’s retail bank profits on the spread between the interest rate it pays on the deposit and the returns it makes investing the deposit.
This practice has significant negative repercussions for clients because they effectively miss out on all the returns they would have accrued had the assets allocated to cash been invested more efficiently. In fact, financial firm Raymond James estimated that SIP clients could be forgoing up to the equivalent of seventy-five basis points.
Schwab received significant backlash as a result of this disclosure
and has since changed SIP’s fee structure somewhat. SIP still does not charge an advisory fee, and it continues to promote Schwab’s ETFs and retail bank.
However, SIP now makes a “nominal calculation” and caps the compensation Schwab earns on affiliate services to the equivalent of thirty basis points.
If Schwab earns more than this amount, the excess is either refunded to the client or used to pay “account administrative expenses.”
This revised arrangement alleviates the conflict somewhat—SIP is not incentivized to favor Schwab products and services after Schwab reaches the cap. Until that point, however, SIP still has a vested interest in weighting client portfolios toward assets from which Schwab collects fees; without these outside fees, SIP is not sustainable.
This subsection uses SIP as an illustrative example, but conflicts relating to promoting affiliate products and services are common among money manager robo-advisors. Vanguard’s Personal Adviser Services charges clients thirty basis points
—a rate comparable to many independent robo-advisors
—and thus, unlike SIP, is not fully reliant on affiliate service profits. Vanguard nonetheless receives additional fees when its robo-advisor recommends its own funds and ETFs
and thus faces the same conflict. Similarly, Merrill Edge Guided Investing charges forty-five basis points,
while also still receiving additional fees for recommending affiliated products.
III. Recommendations for Regulating Robo-Advisors
Part II laid out the issues of robo-advisor regulation and explained that (1) robo-advisors’ duty of care issues are less concerning than critics claim; and (2) robo-advisors can and often do face conflicts of interest. The SEC has been in the process of learning about robo-advisors and determining how best to regulate them,
and this Part provides a recommendation for how the Commission should proceed. Section III.A argues that the SEC should shift its focus away from the quality of algorithmic advice and gives reasons why the regulatory focus should be on conflict of interest issues. Drawing upon this, section III.B proposes a two-part robo-advisor disclosure rule that would increase transparency.
A. Shifting the Focus to Conflicts of Interest
Section II.B.2 has established that, under both common law and investment advice law, the investment adviser duty of care standard should be interpreted flexibly. As such, the Advisers Act duty of care obligation is actually far more lenient than the dialogue criticizing robo-advisor quality assumes. This, coupled with the SEC’s general leniency in enforcing investment adviser duty of care obligations,
indicates that monitoring the quality of robo-advisor advice should not be a regulatory priority.
On the other hand, the Capital Gains Court created the investment adviser fiduciary duty expressly to address conflict of interest concerns, and investment advice law has always governed conflicts issues surrounding traditional advisers more rigorously.
This section establishes why robo-advisor conflicts are, for a number of reasons, even more concerning than the human adviser conflicts the SEC generally polices. Accordingly, the SEC should shift its focus to robo-advisor conflict of interest issues and enact a rule to monitor how conflicts are disclosed.
1. Programmed Bias. — First, robo-advisor conflicts have larger and more certain effects than human adviser conflicts. Individual employees of traditional investment advisers may be influenced differently by outside incentives;
some employees may be easily tempted by kickbacks and bonus incentives, while others are not. If a conflict biases a robo-advisor algorithm, however, that conflict will without a doubt impact all clients and their investment returns.
Thus, robo-advisor conflicts have a larger and more certain impact.
2. Unsophisticated Investors. — Second, robo-advisor clients are less sophisticated than most investment advice clients and will therefore have more difficulty understanding the consequences of conflicts. As a lower-cost service,
robo-advisors are marketed toward younger and less financially sophisticated investors.
The general functional purpose of the Form ADV disclosure is to equip customers with the information necessary to make educated decisions between services.
To facilitate this function, the SEC requires brochures to be written in plain English, “taking into consideration [the] clients’ level of financial sophistication.”
Given robo-advisor clients’ general lack of finance knowledge, robo-advisor conflict disclosures should be additionally transparent to be consistent with this requirement.
3. Market Trends. — Finally, properly regulating robo-advisor conflicts is additionally pressing because of broader trends in the robo-advisor market. First, the robo-advice market is rapidly growing.
In particular, robo-advisors have opened financial advice to millennials and, because young investors inherently trust technology and prefer their services to be delivered at a faster pace, they often actually find robo-advisors preferable to traditional advice.
As the demographic grows wealthier, robo-advisors are likely to continue to gain traction, and effective regulation will be increasingly necessary to ensure that clients are protected and maintain trust in the services. Second, the increased prevalence of robo-advisors operated by traditional money managers—which have greater potential to be conflicted than independent robo-advisors
—also increases the need for a comprehensive regulatory system for conflicts. Third, as artificial intelligence continues to develop, automated investment advice will only become more sophisticated.
The SEC should begin developing an effective means of regulation now.
B. Solutions for Addressing Robo-Advisor Conflicts
With the importance of addressing robo-advisor conflicts established, this section next recommends that the Commission promulgate a two-part rule that would make robo-advisor conflicts more transparent. On a broader level, the SEC should require robo-advisor firms to clearly specify when conflicts are intentionally programmed into their algorithms. Then, when robo-advisors purposely factor conflicting incentives into their algorithms, there should be a heightened disclosure requirement. More specifically, the SEC should require robo-advisors to provide clients with a “shadow commission” figure, which would quantify the effective amount conflicts of interest cost a client, each time the Form ADV brochure is delivered.
1. Delineating Intentional Conflicts. — The SEC should require that robo-advisors, in their disclosures, clearly delineate between conflicts that are programmed into their algorithms and conflicts that may affect the design of algorithms. Currently, the language in most brochures leaves this distinction unclear. For instance, this excerpt from Schwab’s January 2015 brochure implies that a conflict is programmed into SIP’s algorithm, but does not state so explicitly:
Because Schwab Bank earns income on the Sweep Allocation for each investment strategy, SWIA has a conflict of interest in setting the parameters for the Sweep Allocation. In most of the investment strategies, this results in a Sweep Allocation which is higher than the cash allocation would be in a similar strategy in a managed account program sponsored by a Schwab entity or third parties.
Some brochures indicate that conflicting incentives exist, but are not intentionally programmed into the algorithm, by acknowledging “competing interests” while stating that the robo-advisor strictly abides by its investment methodology.
This, again, leaves ambiguity as to the effect of the “competing interests.” Still other brochures are explicitly equivocal: For example, FutureAdvisor, the robo-advisor owned by BlackRock, discloses that “[i]nvestment . . . in an affiliated product may mean that BlackRock, Inc. may receive directly or indirectly advisory fees and other compensation from the affiliated product.”
Problematically, each of these phrasings leaves unclear exactly how a robo-advisor’s conflicting interests affect investment decisions and client returns. The SEC could rectify this issue by requiring advisers to state whether conflicts are deliberately programmed into investment allocation algorithms.
A requirement that robo-advisors explicitly indicate whether conflicts are incorporated into their algorithms would be consistent with the intent of the brochure document. The brochure exists to educate investors
and, to facilitate this purpose, the SEC generally requires brochure disclosures to be as specific as possible.
Robo-advisors are capable of making disclosures about conflicts with greater specificity than traditional investment advisers, so they should be obligated to. Furthermore, the SEC’s instructions for brochures require advisers to “provide the client with sufficiently specific facts so that the client is able to understand the conflicts of interest . . . and can give informed consent . . . or reject them.”
If anything, robo-advisor clients, as less sophisticated investors, need more clarity to understand disclosures.
Currently, clients cannot determine from existing disclosures whether a conflict will certainly or only possibly affect recommendations.
Therefore, it would be entirely consistent with SEC policy to require that, when robo-advisors intentionally bias algorithms, they state so clearly in their brochures, without hedging language.
2. Addressing the Tiers. — Due to the different natures of unintentional and intentional robo-advisor conflicts, the SEC should establish different disclosure requirements for each type of conflict. Conflicting incentives that exist, but that are not intentionally programmed into algorithms, should be disclosed as traditional investment adviser conflicts are. These incentives could subconsciously influence algorithm programmers, but it would be impossible to determine if they actually do and, if so, to what extent. Because, as with human investment adviser conflicts,
there is ambiguity as to how unintentional bias impacts recommendations, the current regulatory scheme is appropriate.
There is, however, more certainty with conflicts that are deliberately programmed into algorithms. Increased transparency is therefore possible
and should be mandatory.
The SEC should implement a rule requiring robo-advisors to disclose to every client a “shadow commission” that indicates the impact of conflicts programmed into robo-advisor algorithms. To determine this figure, robo-advisors would first calculate for each investor the difference between what the client’s expected returns would be if the algorithm worked in the client’s best interest (i.e., if allocation decisions were not affected by conflicts) and what the client’s expected returns are in the actual algorithm (which is affected by conflicts).
Advisers would then need to convert this figure to a basis point equivalent, so that prospective clients are able to easily compare between advisers. This shadow commission figure should be provided to the client at the time the brochure is first delivered, and it should be updated each time the brochure is delivered thereafter.
Currently, Schwab’s brochure does quantify the effects of its conflicts to an extent, but the disclosure is inadequate:
While clients are not charged a Program fee for services . . . SWIA makes a nominal calculation that fully offsets in the amount of 0.30% the compensation its affiliates receive from ETF transactions in clients’ accounts. This includes advisory fees for managing Schwab ETFs™ and fees earned for providing services to third-party ETFs . . . if CSIA selects them . . . . If this affiliate compensation ever exceeds 0.30% of client assets, SWIA would refund the additional amount to client accounts . . . .
Here, SIP informs clients how much SIP is earning on each portfolio, but it does not state how much clients are forgoing as a result of the conflict.
If a client could be earning a higher return from another asset allocation, but is allocated to a less ideal asset so that the adviser can collect its 0.30% fee, the client should be made aware of the conflict’s entire impact. The “shadow commission” would accomplish this by accounting for the returns clients missed out on because of a robo-advisor conflict. It would also be simpler for unsophisticated investors to understand and would better enable investors to compare services and reach an informed decision.
A more draconian alterative would be to completely prohibit robo-advisors from intentionally biasing algorithms toward conflicting interests, but such a rule would not be feasible. A blanket prohibition would admittedly be even simpler for investors to understand and would protect investors most fully. It would, however, be inconsistent with the United States’ securities regulation regime, which favors disclosure over prohibition.
On a practical level, a prohibition would also harm the robo-advice market because it would cause money manager robo-advisors to exit the market to the detriment of investors. Money managers develop and offer robo-advisors to drive business to their other services and products
—if SIP were not permitted to favor Schwab’s ETFs and retail bank, the service would not exist.
These money manager robo-advisors are currently the strongest force fueling the growth of the robo-advisor market.
Their disappearance would stall the development of the technology, consequently limiting the investment advice options available to investors.
Conclusion
Robo-advisors are an innovative development with the potential to transform how Americans use investment advice. As services grow in size and popularity, regulators must find efficient and effective methods by which to regulate them. This Note argues that, despite skepticism from the popular press, investment advice professionals, and some government agencies, robo-advisors are structurally capable of meeting the Advisers Act’s duty of care standards. In regulating robo-advice, the SEC should shift its focus from advice quality to conflict of interest issues. Specifically, the Commission should impose a rule that requires robo-advisors to explicitly indicate when conflicting incentives are intentionally programmed into asset allocation algorithms. For these intentional conflicts, robo-advisors should be required to disclose a “shadow commission,” which would quantify for clients how much biased algorithms are costing them. Such a disclosure would give consumers access to more information so that they are in a better position to decide whether to reap the potential benefits of robo-advice.