The authors examine robo-advisor adopters and find that those who are underdiversified prior to adoption can increase their portfolio diversification after adoption because of a higher number of stocks and lower market-adjusted portfolio volatility while displaying higher returns. Investors who are highly diversified before adoption see no change in portfolio diversification and performance after adoption. Adopters see declines in such behavioral biases as disposition, trend chasing, and rank effect.
What Is the Investment Issue?
Most stock market investors tend to be underdiversified, which results in their having to bear idiosyncratic risks. Human financial advisers can help the investors diversify, but their services can be costly, and they are vulnerable to human behavioral biases and cognitive limitations.
Investors face a complex set of decisions with respect to portfolio rebalancing—choosing from a large number of securities and allocating wealth among the chosen stocks—which leads to suboptimal rules of thumb being used.
The authors look at robo-advisors—tools that delivers diversification advice to individual investors and that does not require the involvement of human advisers. They examine the adoption of the robo-advising tool and assess the impact on financial decision making.
How Did the Authors Conduct This Research?
The authors examine an automated portfolio optimizer introduced by a brokerage firm in India that focuses on Indian equities. This robo-advisor applies constrained mean–variance optimization. It enables rebalancing or adding stocks and simplifies trade execution by performing all the trades necessary to arrive at a target portfolio mix in one batch.
The data used in the study include information on individual brokerage client access to robo-advice, investor demographics, all daily trades, and month-end portfolio holdings. The time frame studied is 1 April 2015–27 January 2017. Investors in the study sample were already receiving financial advice from human advisers before being introduced to the robo-advisor to allow the authors to examine the adoption of robo-advising tools as opposed to examining financial advice in general.
In the first part of the analysis, users of the robo-advisor are compared with non-users in terms of demographic characteristics and trading performance. The second part of the analysis looks at the effects of using the robo-advisor on investors’ holdings (in terms of diversification and volatility), trading behavior (on a set of documented biases that earlier research attributes to individual investors), and trading performance (returns of the portfolio and of individual stock trades).
What Are the Findings and Implications for Investors and Investment Professionals?
The effects of using a robo-advising tool depend on investors’ starting levels of diversification.
Investors holding fewer than 10 stocks before using the robo-advisor increased the number of stocks in their portfolios, materially decreased their portfolio volatility, and improved their market-adjusted performance, on average. Investors holding more than 10 stocks pre-adoption decreased their number of stocks after using the optimizer, decreased their portfolio’s volatility to a smaller degree, and experienced no significant change in performance.
When examining the more popular behavioral biases that affect investors (e.g., the disposition effect for selling decisions, where investors have a tendency to realize gains more often than losses; trend chasing for buying decisions, where investors tend to buy stocks after a set of price increases; and rank effect, the tendency for investors to sell the best- and worst-performing stocks while neglecting stocks that display immediate performance), the authors find that those biases are substantially less pronounced when the robo-advisor is used, though the biases are not entirely eliminated.
In conclusion, the authors posit that financial institutions should target robo-advisors on undiversified investors and that future research should delve into the optimal design of robo-advising interventions tailored to the needs of different investor categories.