Back to Forum

Can a Robot Be a Fiduciary?


By Scott MacKillop | January 30, 2017

Articles By Us

A new regulatory model is needed for today's—and tomorrow's—robo-advisors.

Related Resources

Link to Online Article

Download PDF of Article

This article was orginally published in the February 2017 issue of Investment Advisor.

The question of whether a robo-advisor can be a fiduciary has become a hot topic. Recently, departing SEC Chairwoman Mary Jo White and outgoing DOL Secretary Thomas Perez, the Massachusetts securities regulator and two prestigious law firms have weighed in on the question. There is much at stake for the robos; there is little agreement on the answer.

A related question is whether robo-advisors are operating as unregistered mutual funds in violation of the Investment Company Act of 1940. The costs and burdens attendant to registration would cripple their business model. Again, there is little agreement on the answer.

Let’s examine these two questions and see if we can come up with some answers.

For purposes of our analysis, we define robo-advisors as firms that provide portfolio management services directly to consumers through the internet without the involvement of a human investment advisor. Wealthfront is an example. We are not talking about “hybrid” firms like Vanguard or Personal Capital that combine robo-technology with human advice. We are also not talking about firms like Jemstep that provide robo-technology to advisors for use in their firms. Our focus is on pure, online business-to-consumer robos.

Robo-advisers are here to stay and for good reason. But they don’t fit neatly into the existing regulatory structure and fall short of satisfying the fiduciary standards applicable to human advisers. A new regulatory category is needed.

In the Beginning

The fun started in May 2015 when the Securities and Exchange Commission’s Office of Investor Education and Advocacy issued with FINRA a joint Investor Alert discussing the “risks and limitations” of “automated investment tools.” These included the risk that the automated tools may rely on assumptions that are “incorrect” or “do not apply to your individual situation.” The alert also warned of robos’ questions for users that may be “over-generalized, ambiguous, misleading or designed to fit you into the tool’s predetermined options.” It cautioned that “an automated tool’s output may not be right for your financial needs or goals.” These warnings emboldened the anti-robo forces.

The following month, Melanie Fein, a Washington-based attorney with impressive credentials, issued a white paper entitled “Robo-Advisors: A Closer Look.” It questioned the fiduciary status of robos and argued that they were really unregistered mutual funds. She leveled other criticisms, too, including alleged conflicts of interest and questions about the perception that robos charge low fees. Fein’s white paper was prepared for Federated Investors, a non-robo.

Her broadside was followed in April 2016 by a policy statement issued by the Massachusetts Securities Division that openly questioned whether robos were fiduciaries. The division stated that robos “cannot fully satisfy their fiduciary obligations if they fail to perform the initial and ongoing due diligence necessary to act in the best interests of their clients.” The division was concerned that this failure “may render them unable to provide adequately personalized investment advice and make appropriate investment decisions.”

Feeling the heat, the robos issued a white paper of their own in October 2016 written by two attorneys from the global law firm of Morgan Lewis & Bockius, which represents the robo Betterment. The Morgan Lewis paper, “The Evolution of Advice: Digital Investment Advisers as Fiduciaries,” mounted a comprehensive defense of the robos, arguing both that they met applicable fiduciary standards and were not unregistered mutual funds.

Further contributing to this conversation were Mary Jo White and Thomas Perez, who made public comments supportive of the pro-robo forces. White said that robos offer retail investors “broader and more affordable access to our markets,” and that the SEC has “been considering how these so-called robo advisors […] meet their fiduciary and other obligations under the Advisers Act.” Perez was even more supportive, saying that robo Wealthfront has “a platform that enables them to lower fees, operate as a fiduciary and do well by doing good.” White and Perez also made general policy statements supporting the use of technology to expand the availability of advice.

Sorting It All Out

No one can argue with the policy of making advice more generally available to those who need it. Nor can one quibble with the other policy arguments made in the Morgan Lewis white paper. That is, robos give consumers more choice in how they access advice; they tend to have lower fees; and they use low-cost ETFs, which further keep costs down for consumers. These are all good things.

However, we should not look the other way and give the robos a pass on legal requirements that apply to everyone else based simply on these policy arguments. Sure, everyone loves a disruptor, and the robos of today seem harmless enough. Jon Stein, CEO of Betterment, looks more like a choirboy than a portfolio manager; not a whiff of Wall Street.

But what about tomorrow? Robos are getting gobbled up by the familiar faces of the financial services industry faster than you can fill out an online risk tolerance questionnaire. If you held a robo conference three years from now, there wouldn’t be a hoodie in sight — just pinstripes. Any of the original robo firms that might be left will be under such pressure from the venture capitalists who backed them that they won’t have time to be concerned about the little guy.

And as attorney Fein pointed out, current robo practices aren’t nearly as praiseworthy as the robos’ squeaky-clean public image would suggest. Conflicts of interest exist. Fees aren’t always low. Investment options are often limited and may include proprietary products. Although today’s robo portfolios are relatively benign, focusing on buy-and-hold ETF strategies, tomorrow’s robo portfolios may, and probably will, include all manner of high-fee products and could even utilize hair-trigger market timing strategies.

For now, we should put aside the sweeping policy considerations and forget about how the robos look today. Whatever rules we settle on now we should be willing to live with for years to come.

So let’s take a quick look at some sticky legal issues that will help us answer the two important questions with which we began this discussion.

The Fiduciary Question

Investment advisors who are registered under the Investment Advisers Act of 1940 owe a fiduciary duty to their clients. This fiduciary obligation is not actually spelled out in the Advisers Act itself, but was breathed into the statute in 1963 by the U.S. Supreme Court in SEC vs. Capital Gains Research Bureau.

There are two parts of an advisor’s fiduciary obligation. The first is the duty of loyalty, which is the obligation of the advisor to act in the client’s benefit and to place the client’s interests ahead of the advisor’s. The second is the duty of care, which is the advisor’s obligation to act with the care, competence and diligence that normally would be exercised by a fiduciary in similar circumstances.

The key to satisfying these obligations is knowing the client.

Robos collect basic data about the client, such as name, age and address, and then run the client through what is essentially a short risk-tolerance questionnaire. Astronomer and computer scientist Clifford Stoll famously said, “Data is not information, information is not knowledge, knowledge is not understanding and understanding is not wisdom.” The robos collect data, but have very little knowledge about — and certainly no understanding of — the needs or interests of the client. How can you meet your fiduciary obligations to someone you don’t know?

I’ve written before about the challenges of using conventional risk tolerance questionnaires to gain an understanding of a client’s investment needs. On a stand-alone basis they are essentially worthless. Only in the hands of a skilled advisor do they become valuable tools.

It is simply not possible to understand the goals, experiences, preferences, risk profile and other relevant information about a client without talking to them. Anyone who has worked with clients knows this. The answers to a 10- or 15-item questionnaire provided in an online environment cannot even begin to scratch the surface. There may come a day when artificial intelligence has advanced to such a state that a human advisor does not need to be a party to that conversation, but that day is far in the future.

The Morgan Lewis paper acknowledges that fiduciary obligations require an advisor to have “a reasonable basis for its advice.” The Morgan Lewis lawyers do not directly argue that the meager data collected by the robos is sufficient to form a reasonable basis. Rather, they argue that the robos and their clients may agree to limit the scope of the fiduciary duties owed to the client and the amount of information the robo must collect. In effect, they are saying that by using the robos’ services, the client is agreeing to the limited nature of the interaction between the robo and client.

The implications of this argument are frightening. It says that an advisor can disclaim as much responsibility as the client will agree to. This is particularly distressing in light of the fact highlighted in Fein’s paper that at least one robo client agreement is 140 pages long!

This essentially turns the robo advisor into a self-help platform for do-it- yourself investors. Morgan Lewis argues that “goal-based” investors — those with very specific investment objectives like accumulating assets for retirement, college funding or saving for a vacation house — do not want or need to engage in a comprehensive financial planning process. For them, collecting less information is perfectly appropriate since the Advisers Act does not dictate the quantity of information that must be collected.

This is absolutely true. But what about the quality of the information? First of all, there is very little established science behind most risk tolerance questionnaires. Moreover, research has shown that a person’s risk tolerance changes over time and that people are notoriously bad at assessing their own tolerance for risk.

Then there is the problem of properly identifying a client’s goals. What if a client isn’t really sure about her goals or has a problem articulating her goals? What if a client has multiple goals? What if the client has conflicting goals? What if there’s a gap between an investor’s tolerance for risk and the amount of risk he needs to take to reach his goal? What if the investor simply doesn’t understand some of the 10 or 15 questions on the questionnaire? What if they think they understand the questions, but really don’t?

These are problems experienced advisors deal with frequently. How do robos deal with these problems?

Being a fiduciary comes along with important obligations and duties to the client. You cannot maintain that you are a fiduciary while disclaiming all the responsibilities that go along with that title. The weak link in the robos’ claim to fiduciary status is their lack of meaningful knowledge about the client and understanding of their needs.

The Unregistered Mutual Fund Question

Mutual funds pool the investments of a large number of investors and manage them collectively for the benefit of all. They must be registered under the ’40 Act.

The SEC has taken the position that an advisor who manages a large number of individual client accounts in a similar manner may also be subject to the registration requirements of the ’40 Act. This position recognizes that a single pooled account and a series of individual accounts managed in the same way are functionally the same.

The SEC established Rule 3a-4 under the ’40 Act to carve out a “safe harbor” for advisors who manage accounts in a parallel manner so they won’t be characterized as a de facto mutual fund. Robos maintain that they qualify for this safe harbor provided by Rule 3a-4.

There are a number of requirements that must be met in order to qualify for safe harbor protection under the rule. The robos arguably fall short on two of them.

The rule requires that “each client’s account is managed on the basis of the client’s individual financial situation and investment objectives, which should be gathered upon opening the account.” As discussed above, robos do not collect sufficient information to know a client’s “individual financial situation and investment objectives.”

Morgan Lewis suggests that an advisor and client can agree to limit the duties owed to the client under the fiduciary standard, but clients cannot waive the requirements of the rule. So Morgan Lewis argues instead that because robos provide risk tolerance questionnaires and online financial planning tools, and allow clients to customize their own portfolios, the clients “receive investment advice that is customized to their particular investment goals and needs.”

In reality, they receive no investment advice at all. What Rule 3a-4 calls for is personalized advice from a qualified professional. What robos provide is a box full of self-help tools.

Rule 3a-4 also requires that “some personnel of the investment adviser, who are knowledgeable about the account and its management, are reasonably available to the client for consultation.” This is a difficult requirement for robos to meet since their basic value proposition is that you can do it all online without having to talk to anyone.

Not to be deterred, Morgan Lewis argues that the “reasonably available” requirement is met because robos provide clients with “around-the-clock access to a great deal of interactive real-time information about their accounts.” This is unconvincing because it amounts to no more than the typical information available to clients through a custodian.

They also point out that robos “make a great deal of information about their investment philosophy and approach available through articles, blogs and social media posts.” This is thin stuff. Mutual funds also do this, and they still have to register under the ’40 Act.

Morgan Lewis valiantly continues, pointing out that robos “supplement their online offerings with telephone, email and chat features that allow clients to ask more specific questions about the management of their accounts in real time.” Certainly the fact that this all happens “in real time” makes it very compelling, but who does the client talk to when they initiate the “chat feature?” Probably not someone who is “knowledgeable about the account and its management” and definitely not someone who is familiar with the client’s “individual financial situation and investment objectives” since the robo never collected any of that information.

The personalized interaction that the SEC was looking for in Rule 3a-4 is missing. Robos aspire to use algorithms to weave data into personalized recommendations. Their technology is brilliant, but falls far short of replicating the interaction between a human advisor and a client.

A skilled advisor can take a conversation in an infinite number of directions depending on the information a client provides and the sense the advisor gets of the client’s knowledge, sophistication and emotional state. There are limited pathways through a risk tolerance questionnaire and no way to look a client in the eye online or hear the tone of her voice.

What to Do About This Mess

Why hasn’t the SEC taken any action against the robos for failure to meet their fiduciary duties under the Advisers Act or for failure to register their de facto mutual funds under the ’40 Act? Former SEC Chair Mary Jo White told us why when, speaking about robos during the agency’s Fintech Forum in November 2016, she said, “Regulators have an obligation to understand, monitor and, where appropriate, encourage such developments.”

White was absolutely right in taking a hands-off position. Robo technology has been — and will continue to be — a powerful positive force in our industry. It is transformative and will be a part of every financial advisor’s practice in the years to come. Shutting robo-advisors down because they don’t fit well into our existing regulatory structure would be a mistake. (And recall, we’re just talking about direct business-to-consumer digital advice platforms; see “How to Regulate Hybrid Robos” for suggestions on how to regulate the other robos.)

However, torturing existing law to fit them in would also be a mistake, and would have unintended negative consequences. If we allow robos to be considered fiduciaries, we will dilute the fiduciary standards that have served us so well, and we’ll be opening the gates to allow other less lovable players under the fiduciary umbrella.

By twisting the language and intent of Rule 3a-4 to allow robos to operate as they do, we will render the rule meaningless. Firms that want to run thousands of parallel accounts without knowing the needs, goals and objectives of their clients will have license to do so.

The robos are here to stay and will continue to make contributions that benefit both advisors and clients. Now that we have some familiarity with the characteristics of this new set of players, the regulators need to get to work modifying the existing regulatory structure to accommodate them. The robo foot simply doesn’t fit into the fiduciary glass slipper.

Perhaps a new regulatory category is the answer. Advisors assume a position of trust when they become fiduciaries. They enter into a relationship that may be limited in scope, but is complex in nature. Computers can follow logic, but can’t exercise judgment. At least for now, robo-advisors are unable to honor the duties that flow from a fiduciary relationship, so they should not fall under the same regulatory scheme applicable to fiduciary advisors.

Still, the robos have proven that technology can provide do-it-yourself investors with a valuable, low-cost way of accessing diversified, long-term portfolios. Instead of registering as advisors, what if they registered as product platforms that were available to those who did not need or want the services of a fiduciary advisor?

The regulations applicable to robo platforms could parallel those applicable to mutual funds. They would describe certain practices that are permitted and certain practices that are forbidden, while prescribing standardized disclosure to inform platform users.

The regulations could outline acceptable portfolio management practices. They might set restrictions on fees. They might set standards for the self-help tools offered by the platforms to ensure there’s some science underlying them. They might restrict the use of client data captured online so it’s not sold to other firms or used for cross-selling.

The disclosure requirements could be targeted to highlight conflicts of interest and prevent fraudulent or deceptive practices. They might describe the limitations of the robo self-help tools, much as the SEC/FINRA joint Investor Alert did. They might also require information about the qualifications of those building and managing the portfolios.

The key is developing a regulatory scheme that recognizes the unique nature of the robos. Technology platforms like the robo-advisors did not exist when the Advisers Act and the ’40 Act became law. The old regulatory schemes don’t work for these new players. Trying to fit robos into the same categories as either human advisors or mutual funds will not work and will only pervert the existing regulatory structures. New thinking is required.

The world needs the robos. They can provide a source of solid, low-cost portfolio management for those who cannot afford — or don’t want — to work with a human advisor. In a world where individuals are increasingly dependent on their own resources for financial security, it is worth creating a workable regulatory niche for the robos.

Read the paper here.