Read Full Paper

Before providing individual investors with any investment advice or recommendations, financial consultants would usually conduct risk assessments for them in face-to-face meetings. Nevertheless, with the advancement of Fintech, coupled with the enforcement of social distancing due to the COVID-19 pandemic, there is a new trend of having contact-free advising, such as robo-advising, to ensure the sustainability of investments. Robo-advising is a platform consisting of interactive and intelligent components, through which customers receive personalized investment services online. However, there is no recognized standard for such robot-assisted profiling of an investor’s risk preferences. This study attempts to develop a framework on risk profiling through content analysis and provides answers in regards to what kinds of questions are relevant to clients’ risk profiling and how to assess their relevance.

This study applies inductive content analysis to understand existing risk assessment and identify patterns through the investigation and comparisons on representative questionnaires from banks and investment service providers. Then, five important risk factors to profile individual risk preference are developed through inductive reasoning, namely, setting of realistic investment objectives/goals, risk appetite of an investor, understanding of investment risk according to own practical experience and knowledge, investor behavior when suffering investment loss, and ability to take risks. This helps to set the boundary to build theory for robo-advising risk profiling.

To make the risk profiling involved in robo-advising even more successful, five crucial implementation recommendations are put forward in the research. First, the total number of questions should be limited to show that robo-assisted assessments are simpler and briefer than human assessments. Second, questions should be simple and straightforward, which can be answered with instant valid responses without assistance. Third, customers’ responses throughout the assessment should be consistent, and the robot should prompt the customer to re-input any inconsistent responses. Fourth, the risk scores/grades should be updated regularly, as the degree of risk one can take is likely to change from time to time, and such demonstration of risk scores enables investors to provide the most accurate information at the moment. Finally, the customers should not be asked to adjust their risk scores/grades within a short period of time, to avoid them timing the market by buying investment products with higher risk grades.

Currently, robo-advising faces the challenges of lack of personal customization and lack of confidence in the ability of algorithms to perform tasks. This study may contribute as a starting point for developing an organized and theoretical framework and protocol for systematic research on advising, as well as that for the development of methodological framework for risk assessment and personalized services, which are important in building trust with customers. In practice, managers may make reference to this research and gain a better understanding of which question types are useful for risk profiling and ways in which to develop risk profiling for robo-advising. In the future, similar research may be replicated in other regions or countries, and more comparisons on the risk profiling of robo-advising in different places may be conducted.