Enhancing Portfolio Performance: Incorporating Parameter Uncertainties into Zero-Beta

Purpose -This study examines a zero-beta portfolio strategy that accounts for the uncertainties in expected returns and betas, with the goal of improving investment performance by incorporating parameter uncertainty into the optimization process.Theoretical framework -The research is grounded in modern portfolio theory and robust optimization, drawing on the multifactor asset pricing model of Chen, Roll, and Ross (1986).It leverages the Kalman filter to estimate dynamic betas and their uncertainties, and incorporates analysts' forecasts to assess expected returns and their associated uncertainties.Design/methodology/approach -The study constructs two types of zero-beta portfolios: a long-short stochastic portfolio that maximizes the ratio of expected return to parameter uncertainty, and a long-short normal portfolio that focuses solely on maximizing expected return.Portfolio performance is evaluated using data from 2015 to 2022. Findings -The results indicate that the long-short stochastic portfolios outperform the normal portfolios on several performance metrics.Specifically, they exhibit higher realized returns, lower drawdowns, and a superior realized Sharpe ratio.In addition, the stochastic approach yields more accurate predictions with a significantly lower root mean square error.Practical & social implications of research -The findings provide insights for investors, fund managers, and practitioners seeking to improve portfolio stability and performance under uncertainty.However, reliance on analysts' estimates should be approached with caution, as deviations from expected values can still occur.Originality/value -This study contributes to the existing literature by empirically validating the benefits of incorporating parameter uncertainty into portfolio optimization.