Recent history is rife with comparisons between machine computing power and human (organic) computing power. Thanks to advances in technology, that comparison has reached the financial advice field. Specifically, robo-advisers have been stepping into the role of financial decisions and portfolio allocations. The question is — should they?
According to a recent study, not so much. The peer-reviewed study concluded that human financial advisers outscored their robo-adviser counterparts when it came to rates of return on investment decisions. Score one for the humans.
Ever since advances in technology gave us everything from the lightbulb to supercomputers, a delicate tension has existed in the human versus machine battle. The 1968 film 2001: A Space Odyssey showed computer HAL being “unplugged” because he tried to take charge of the human crew. But in real life, machines are not necessarily on track to take over, but they are gaining on their human counterparts in head-to-head competition. Here are some examples.
- In 1997, IBM’s Deep Blue beat Garry Kasparov in a chess match, the first defeat of a reigning world chess champion by a computer under tournament conditions.
- Jeopardy champions Ken Jennings and Brad Rutter lost to Watson, another IBM computer, in Jeopardy games held in 2011.
- More recently, in 2019, a Carnegie Mellon computer, Libratus, beat five human poker champions — all at the same time.
WHERE ARE WE TODAY?
The current age of technology is dovetailing with an upswing in demand for professional financial advice because of changes in the U.S. retirement system. Billions of dollars, which used to live on balance sheets as corporate pensions, have been moved to the control of employees as tax-advantaged retirement savings programs. Everyday people are now in charge of their own retirement funds. And not everyone has the degree of knowledge held by corporate financial experts, nor should they. This is where financial advisers come in.
These advisers are responsible for encouraging and coaching their clients to force savings as well as to change spending habits. Individuals partnering with financial advisers are shown to contribute at least 10 percent to their employer-provided plan, which is higher than the default contribution rate of 3.4 percent made through automatic enrollment. Those outside of employer-related financial programs might also turn to qualified advisers if they have saved close to 20 percent of their annual income over time.
Once that investor has determined the need for a financial adviser, the question becomes whether that source is machine or human. The machine version (the robo-adviser) provides advice with the help of automated systems and algorithms. This information is based on portfolio allocation theory as well as levels of investor-reported risk tolerance. Recommendations can focus on portfolio mixes or portfolio rebalances. Robo-advisers frequently rely on ETFs as the primary investment vehicles for the accounts.
Statistica estimated an annual growth of 18 percent in the use of robo-advisers, a growth that is not terribly surprising. Cost is a factor; many investors believe that robo financial advice is less expensive than the human kind. Additionally, automation bias — assuming a machine has better deduction powers than a human despite a human’s observation otherwise — can play into the decision. The investor might assume that machines can perform a higher degree of calculations detailing higher rates of return on specific investments. There is a reason for this bias, given the superior performances of machines, such as the examples listed above.
But machine superiority is an investor assumption. The recently published Journal of Wealth Management study, “Who is Better at Investment Decisions: Man or Machine?” reveals much different results.
THE AGE-OLD COMPETITION
The study, conducted by the author of this article in conjunction with others at Georgia State University, set out to test several hypotheses, and turned out some very unexpected results: one involving inflexibility concerning robo-advice algorithm applications in same risk profiles. Other interesting results showed evidence to dispel potential automation bias in financial adviser selection and the value of human financial adviser advice in recent market conditions. The researchers set up an age-old competition, that of machine versus humans.
My previous research demonstrated that financial advisers, whether robo or human, typically offer the following.
- Portfolio construction: General and specific direction about how and where to deploy current and future wealth accumulation
- Investment product connection: Access to investment products
- Coaching: Varying degrees of consulting, or discipline, directed back to the investor
In this study, the researchers selected a top robo-adviser, based on data from the entire field. The human factor came from invitations to elected financial industry association board members, with Mark Kosanke of Concorde Financial, chosen on the basis of fastest response to the invitation. The robo and the human (Kosanke) were presented with nine scenarios, and told to select portfolio asset allocations for each of the nine. The scenarios were broken down by age (a 30-year-old, 50-year-old and 70-year-old) and investment amount ($100,000, $500,000 and $1 million). Other variables, such as medium risk tolerance, were included in the calculations, and specific asset classes presented. The advisers were asked to select specific allocations to various asset classes and/or funds.
THE QUALIFIED RESULTS
After comparing variables, results and metrics such as Sharpe and Sortino ratios, the researchers found that all of the returns from portfolios selected by the human financial adviser were higher than those of the robo-adviser. Unexpectedly, the human won every scenario since his chosen portfolios had a higher return on investment every time. And yes, even with adviser fees included. The human won 18-0, and on average overall, beat the machine by 1.47 percent, net of fees (the fees charged by the human were 1.05 percent, and the robo .30 percent).
While additional research will need to be conducted, the reason for the result boils down to flexibility; the robo-adviser has a one-size-fits-all portfolio allocation algorithm insensitive to many factors (such as age and amount invested in this case). This points out that robo-advisers don’t have in-depth knowledge of the marketplace or the macro-environment, nor do they have ontological awareness or just downright cleverness and flexibility. In a word (or two): Humans rock.
HUMANS, MACHINES AND MODERATION
The research conducted wasn’t to provide the notion that robo-advisers aren’t useful when it comes to investment decisions and portfolio allocations. Robo-advisers may provide very useful information when it comes to selecting the right assets for portfolio inclusion, based on an investor’s risk-tolerance profile. However, it is the human financial adviser that understands the market fluctuations and broader economic base. It is the human financial adviser who (hopefully) has the experience with previous economic swings. And, it is the wise human financial adviser who knows his or her client and understands the client’s goals and objectives.
In short, this was not necessarily a “human versus machine” competition, as much as it showed how humans could possibly use robo-advice in an effort to help support investment decisions. Decisions that are still made solidly by humans — so far.
John Harrison is executive director of ADISA.