A privacy-preserving robo-advisory system with the Black-Litterman portfolio model: A new framework and insights into investor behavior

Abstract

Recent financial sector changes, including strict privacy regulations, challenge robo-advisory companies with cybersecurity and data privacy. This study proposes a new framework integrating Homomorphic Encryption into the Black-Litterman portfolio model to safeguard robo-advisory investment strategies. The framework effectively balances privacy and accuracy while maintaining an acceptable level of privacy optimization error. Novel evaluation methods are also proposed to assess the trade-off between losses from privacy optimization and strategy leakage, from an economic viewpoint based on Expected Utility and Prospect Theory. It provides valuable insights into human behavior concerning privacy protection in portfolio management.