Content area
Full text
The risk budget and risk appetite of investors are two key elements of professional investment advisory. Financial advisors typically complete a questionnaire to evaluate the client’s risk tolerance. In Germany, for example, financial advisors are legally bound to gather investor-specific information for the assessment of individual risk preferences. The collected data lead to a risk classification, which is associated with a specific equity exposure in the recommended portfolio. There is, however, no mandatory norm regarding the design of the inquiry or how the gathered information needs to be used to categorize clients. As this assessment of risk preferences is rather vague, human advisors obviously have some scope for the interpretation of their clients’ risk tolerance.
During the last five years, a new kind of business model has emerged: robo-advisory. The term robo-advisor is a composite of the words robot and advisor and describes online investment advisory, which is driven by algorithms and rational logic and excludes emotionally biased decisions. In this study, we analyze how robo-advisors evaluate the risk preferences of their users and how they translate the individual risk profile into a recommended portfolio. Our findings reveal significant differences between robo-advisors. There is a trade-off between the need to economize user time and development of an elaborate risk preference analysis. Robo-advisors therefore have the tendency to use fewer questions than the human advisors in our sample. In this context, it is surprising that robo-advisors sometimes ask questions that have no impact on risk categorization. Regarding the recommended equity exposure, robo-advisors do not deviate significantly from their human colleagues. It seems there is no or only little benefit from big-data analysis or risk assessment.
For our study, we generate a proprietary data sample consisting of 13 anonymized robo-advisors. We compare seven German robo-advisors with six market-leading ones from the United Kingdom and the United States. In the first step, the different web-based risk assessment questionnaires are analyzed. To gain insight into how the robo-advisors evaluate the risk tolerance of their clients, we first consider the number and type of questions used. Our algorithm then automatically tests all, or at least a wide range of, possible answer combinations and records the risk profile and the recommended equity exposure. This process allows us to deeply analyze the...