From advice to action : drivers of investor behavior in the age of robo advice

The dissertation deals with complex decision-making processes of investors in the context of the changing financial advisory landscape, especially considering the increasing importance of robo advice. It examines both personal and structural factors that influence investment behavior in the presence of algorithm-based advice and provides insights into the interactions between classical concepts of behavioral finance and digital advisory platforms. The basic idea is based on the approach that investment decisions are often subject to systematic investment errors, which are particularly common within the group of inexperienced investors, and that professional financial advice can help to avoid such errors. A total of four essays deepen the understanding of how algorithmic financial advice shapes investor behavior in this context. The first article lays the methodological foundation for the work by examining, among other things, whether simplified survey instruments can reliably reflect investor preferences. By combining surveys and laboratory experiments, it is shown that self-assessments of risk preference and patience based on easy-to-understand questions are superior to more complex methods. These results are particularly relevant for robo advisors, as their designs rely on standardized and simple questionnaires to automatically capture investor preferences. The second article examines the acceptance of advice in the context of robo advice. The focus is on the influence of the design of user interfaces and the length of exploratory questionnaires. The results show that emotionally appealing interfaces significantly reduce the discounting of the recommendation, i.e. promote acceptance. This underlines the importance of a well thought-out interface design to improve investment decisions. The third article analyzes the dynamics of trust and the acceptance of advice, particularly with regard to the role of human involvement in hybrid advisory processes. It is shown that the involvement of a human advisor increases investors' trust in the advisor, but at the same time reduces the acceptance of the advice. This is attributed to the so-called “betrayal aversion,” triggered by the individually perceived risk of unethical behavior on the part of the advisor. The results highlight a fundamental trade-off: while human advisors strengthen trust, they can promote skepticism towards the recommendations themselves. The fourth paper examines the influence of overconfidence on investment behavior in contexts where algorithm-based advice is available. The results show that investors who overestimate their own competence in terms of their general financial knowledge trade more than non-overestimating investors, even when algorithm-based advice is available. The study concludes that robo advice alone cannot prevent systematic investment errors and that there is a research gap regarding the impact of integrating features to prevent such investment errors in robo advice software. In summary, the results primarily show which robo-advice-specific factors influence the behavior of investors. The dissertation thus forms the basis for a possible approach to the challenge of increasing the acceptance of advice in the context of robo advice in order to help investors in their investment decisions. The results therefore also have practical implications for developers of robo advice software. This dissertation also contributes to behavioral finance by providing evidence-based insights into the drivers of investor behavior in the age of robo advice and paving the way for more effective advisory approaches.