Not even Wall Street is safe from the machine learning megatrend. For many individuals looking to grow their wealth, hiring a human financial advisor can be expensive. The rise of the “robo-advisor” over the last decade has been an industry-wide response to this need for reasonably priced wealth management services. Robo-advisors apply machine learning to large datasets of investor preferences and investment outcomes to build customized portfolios. Once seen as a niche service, robo-advisors now manage over $224 billion globally, and are expected to grow to $1 trillion over the next four years.1 Unable to ignore this trend, financial services firms like Charles Schwab, who have historically benefited from the fees charged for their wealth management services, have started using these technologies to provide a more efficient and customized service for a lower cost.
In addition to increased competition, the machine learning megatrend has had a number of significant impacts on the wealth management industry, most importantly allowing Schwab to streamline its services. Traditionally, individuals looking for wealth management services would complete a form indicating investment goals, preferences, and risk tolerance. This form would then be reviewed with a financial advisor who would take time talking through the pros and cons of various investment options. Individuals would then meet periodically with their advisor to discuss any updates that may impact the investment strategy. This highly customized approach takes a significant amount of time, and is inherently subject to behavioral biases. Schwab’s introduction of the robo-advisor has allowed them to dramatically reduce the amount of time needed to review investor needs, while still applying and adapting a suitable investment strategy over time as the algorithm receives new information. This frees up Schwab’s human advisors to focus on situations that require real-person expertise, enhancing the overall quality of service offered by the company to its many clients.
In 2015, Schwab launched “Schwab Intelligent Portfolios”2 as its first machine learning application. This appeared to be the company’s short-term solution to make the investment process as cost-efficient as possible for individuals. However, if the history of algorithmic investing has taught us anything, it is that by relying on a “black box” to build, monitor, and adjust portfolios we may not fully understand how the data is driving decision-making and if those decisions will prove effective in future economic environments.3 It has only been three years since the launch of Schwab Intelligent Portfolios, and while it may have been successful thus far, it may fail to perform when the decade-long bull market comes to an end and the learnings from the recent period prove to be uninformative in the new market regime. In these scenarios, the ability of Schwab’s human advisors to see the “big picture,” educate their clients, and analyze the market beyond the algorithms will be a significant value-add.
With this medium-term view in mind, Schwab expanded its offering in 2017 to include “Schwab Intelligent Advisory,” a hybrid solution combining the benefits of its robo and human advisory services.4 This allows a human advisor to intervene when the client’s situation is too nuanced to fully rely on the automated service. Combining these two solutions also improves the feedback loop of the algorithms, avoiding uninformative updates when the market changes and inexperienced investors tend to panic. However, it still relies on the algorithms to make the investment process as efficient as possible, informing the human advisor and allowing them to focus their efforts on the most complex client issues.
The largest concern currently facing the robo-advisory business is an algorithm’s ability to prove successful through all market scenarios. By combining machine learning with human intuition, Schwab has the unique opportunity to improve its algorithms to serve as a stand-alone tool capable of managing suitable investments reliably throughout time. As human advisors utilize the algorithms to inform their advice, adjustments should be made to the automated questions asked, along with including automated follow-ups based on investor profiles to keep the feedback as informative as possible. While eliminating behavioral biases from advisors is important, it is also necessary to safeguard algorithms from potentially biases introduced by the client who may edit their investment profile over time in a reactionary way that does not effectively inform the investment decision-making process. With all these necessary improvements in mind, it is clear that there is still an important role for human advisors to play.
Another large issue with using machine learning algorithms to manage investment portfolios is their difficulty in considering all aspects of a client’s financial situation. The more complex the situation (typically positively correlated with overall wealth) the more difficult it is to fully rely on these algorithms. What is required to make these algorithms successful in all financial situations, and is it realistic to anticipate a world where human advisors are made obsolete?
(Word Count: 797)
- Duggan, Wayne, “9 Things to Know About Robo Advisors.” October 5, 2017, https://money.usnews.com/investing/investing-101/slideshows/9-things-to-know-about-robo-advisors?slide=10. Accessed November 2018.
- Charles Schwab, “Charles Schwab Launches Schwab Intelligent Portfolios.” March 9, 2015, https://pressroom.aboutschwab.com/press-release/corporate-and-financial-news/charles-schwab-launches-schwab-intelligent-portfolios. Accessed November 2018.
- Knowledge @ Wharton, “The Rise of the Robo-advisor: How Fintech Is Disrupting Retirement.” July 14, 2018, http://knowledge.wharton.upenn.edu/article/rise-robo-advisor-fintech-disrupting-retirement/. Accessed November 2018.
- https://pressroom.aboutschwab.com/press-release/schwab-investor-services-news/schwab-announces-schwab-intelligent-advisory. Accessed November 2018.