Can robo-advisers be fiduciaries?

What does it mean to be a fiduciary? Putting the best interests of the client first through due loyalty and care, according to the Securities and Exchange Commission. Is a robo-adviser equipped to be loyal and, especially, “to care”?

The first aspect, loyalty, tends to be interpreted as not having conflicts of interest that might contaminate or even just cloud one’s ability to give advice that truly benefits the client ahead of the adviser’s own interests.

OK, automated programs probably get a passing grade on this one. They don’t have a personal stake to put above a client’s, as long as they are programmed with standardized criteria for determining the appropriate investments to recommend from a vetted menu. But what is appropriate?

That’s where “care” comes in. Caring depends on knowing the client.

BEYOND COMPETENCE

Blaine Aikin, chairman of fi360 Inc., brought this to readers’ attention in a Fiduciary Corner column for InvestmentNews last October. Sure, part of care is competence. In some cases, robos can perform this function — such as routine needs like regular rebalancing — even better than some human advisers, particularly those who don’t rebalance at all.

But knowing a client requires a discovery process that goes beyond a sterile set of questions — as few as a handful up to a dozen or so for most robos. Much of assessing a client’s needs comes from knowing the financial aspects of a person’s life beyond the pool of cash to be invested.

An expression from the 1980s comes to mind: GIGO. The acronym for “garbage in, garbage out” refers to the all-important factor of inputs. The computer can only spit out a calculation, a finding, or in the case of advice, a recommendation based on what it has been told.

Most clients have multiple financial needs competing for the same limited monetary resources available. Knowing how best to use them requires weighing each of the factors against the others.

Reporter Alessandra Malito wrote earlier this month about Betterment’s incorporating external accounts onto its clients’ dashboards to give them a fuller picture of their financial life, as other robo-advisers have done in varying degrees. A good first step, but syncing those inputs into investment recommendations made by the robo is still lacking.

As Aaron Klein, CEO of Riskalyze, said in the story, “When the first robo can look at someone wanting to invest $10,000 and say, “That’s a bad decision for you; instead of investing you should pay down student loans’ … that’s when we reach the point that they’re actually delivering advice.”

Robos are for the most part registered with the SEC as registered investment advisers and are required to act as fiduciaries. But as SEC commissioner Kara Stein said at the Investment Adviser Association Compliance Conference in Washington on March 10, the rules were written before the advent of online advisers.

NO DUTY?

“What would a fiduciary duty mean to a robo-adviser?” she asked. “Is there no fiduciary duty if it’s automated advice? How should the SEC be thinking about that and regulating that?”

And last week, the Financial Industry Regulatory Authority Inc., which regulates brokers held to a less-stringent suitability standard, put out guidance on requirements when offering robo-advice.

Essentially, responsibility for advice — even that which is incidental to a transaction — cannot be palmed off on an inanimate object. The broker must understand the algorithms that turn the inputs into a recommendation; this requires education and training on the tool at hand.

Notice we say “tool.” A computer program of any kind is a tool that — when complexity comes into the picture — requires human reasoning to round off the sharp edges. This is especially true when, in the case of robo-advisers, the GIGO effect is distinctly possible given the limited scope of current inputs.