Inferring Risk Aversion for Decentralized Investment Portfolios

A shift towards self-managed pension investments has allowed greater transparency, flexibility and control in the way individuals interact with their financial wealth. In contrast to traditional wealth management practices that rely on explicit assessments of individual risk aversion, platform-based investment management services can provide concise metrics that define individual risk aversion, but are computationally-intensive. Using a complete dataset obtained from the interaction of investors with investment management platforms, we provide a detailed insight into risk aversion by age, gender and reaction to investment performance history. We use a MapReduce model to efficiently gauge risk aversion levels in real-time to optimize individual glide paths and investment styles. The use of inferred assessments of risk aversion based on actual investor behavior is undermining the inefficient cohort-based approach to investment management. We anticipate they will eventually replace the need for subjective aversion assessments conducted by financial advisors.