Semiparametric Option-Implied Information and Median-Variance Approach: A Game-Changer in Integrating Sustainable Practices in Portfolio Optimization

Portfolio weights often exhibit instability when positioning is based on expected returns estimated using historical data, with the normal distribution assumption in many models proving irrelevant for swift investor decision-making.To address these shortcomings, this study introduces a methodology incorporating asset information through implied volatility and a median-variance approach.The latter, notable for its flexibility in application without relying on normal-distribution assumptions, guides the construction of portfolios through stock ranking determined by implied volatility.This research employs the semiparametric option model of Extended Generalized Leland (EGL) to estimate implied volatility for stock selection, focusing on the Dow Jones Industrial Average (DJIA) index.The preselection phase integrates risk premiums into the model, with the resulting subset of stocks subject to various strategies considering short-selling and zero-correlation constraints to accommodate broader investment strategies setting of environmental, social, and governance (ESG).The findings reveal that the information derived from the median-variance approach significantly enhances portfolio selection, leading to markedly higher Sharpe ratios, increased returns, and reduced volatility.By providing a robust, flexible, and data-driven approach to portfolio selection, our methodology not only offers investors a means to make informed decisions but also holds the potential to guide sustainable investments.