|Researchers:||Andreas Hackethal, Benjamin Loos, Alessandro Previtero|
The current rise of robo-advisers provides a new, potentially more cost-effective approach for offering financial advice. Using unique data from a large German retail bank, we plan to investigate the effects of robo-advisers on three aspects of clients' portfolios: i) stock market participation and share of risky assets; ii) behavioral biases and portfolio efficiency; and iii) learning and spillover effects between robo-advised and non-robo-advised accounts. Preliminary evidence suggests that after joining a robo-advising service, clients increase financial risk-taking, hold more diversified portfolios with a larger fraction of index funds, exhibit lower home bias and increase their (buy) turnover. Our research has the potential to provide key insights into the trade-offs associated with using robo-advisers. Many households rely on financial advisors for investment guidance. Despite the widespread use of financial advice, academic research has raised concerns about the cost and quality of this service. In recent years, a technology-based form of advice, commonly referred to as robo-advice, has emerged as an alternative. By providing low-cost access to passive equity investments, robo-advisers could reduce the cost of participating in the stock market and, more generally, the costs of holding equity. Robo-advisers may also be able to reduce client biases by proposing cost-effective, passive and well-diversified investments and it leaves no room for the potential biases that human advisors can pass on to their clients. And robo-advisers may draw attention to the fees that their clients are paying on their other, non-robo-advised accounts. Contrary to this positive view, the lack of a human advisor could limit clients' level of trust in the financial markets and reduce participation and share of risky assets (Guiso, Sapienza and Zingales, 2008). Moreover, without a human advisor, clients could not overcome the anxiety associated with holding risky investments (Gennaioli, Shleifer and Vishny, 2015) or select unsuitable investment risk.
We have obtained data from a large German retail bank that started offering a robo-advice service. Three features make them particularly suitable for addressing our research questions: i) roughly half of the robo-advised clients had existing accounts at the bank, allowing us to investigate investment behaviors before and after joining; ii) we can observe the behaviors of bank clients that did not join (i.e. clients working with human advisors or making their own investment decisions); and iii) we have information on the marketing campaigns that quasi-randomly targeted existing bank clients to advertise the robo-adviser. In contrast, data from independent robo-advisers would not typically include information on clients' prior investments, the investments of those who don't sign up for the service, or clients' additional investments outside of their robo-advised accounts.