Fuzzy Portfolio Optimization Using Tracking Error

Abstract

 The study compares the conventional mean-variance model with the fuzzy meanvariance model for better portfolio construction. Trapezoidal fuzzy number for fuzzy mean, variance, and covariance is computed. The models are applied to BRICS (Brazil, Russia, India, China, and South Africa). For comparison of two alternative models, efficient frontiers are constructed, and tracking error by using the MSCI emerging market index as a benchmark index is measured. The construction is based on both equally weighted and optimized portfolios. The results show that the fuzzy mean-variance model is a better approach as compared to the simple M-V model for BRICS in efficient frontiers settings. Moreover, it suggests that the fuzzy M-V model can help more in devising an investment strategy based on the risk-return relationship.