A1198
Title: ESG-based global minimum variance portfolios: Allocation weights, performance metrics and stochastic dominance
Authors: Alessandra Amendola - University of Salerno (Italy) [presenting]
Vincenzo Candila - University of Salerno (Italy)
Luigi Aldieri - University of Salerno (Italy)
Abstract: Over the past two decades, corporate Environmental, Social, and Governance (ESG) factors have received growing attention from both academic and investor perspectives. To date, empirical evidence on the actual impact of ESG integration in portfolio construction remains mixed, and its effectiveness continues to be debated in the literature. The study analyzes the composition and performance of Global Minimum Variance (GMV) portfolios constructed under different ESG configurations. Specifically, Fully Balanced (FB) GMV portfolios composed of Leader, Average, and Laggard ESG assets are compared with three configurations composed exclusively of Leader-only, Average-only, and Laggard-only ESG assets, respectively. To mitigate the impact of asset selection, 10,000 portfolios per year are simulated over the period 2015-2024, each constructed from a random subset of the Russell 3000 constituents. Overall, 100,000 portfolios are analyzed along three dimensions: (i) the weights of ESG categories within the FB-GMV portfolio, (ii) the comparison of Variance, returns, and Sharpe ratio across single-category portfolios, and (iii) the stochastic dominance of the Variance of single-category portfolios. The analysis of these dimensions is particularly relevant to ESG-oriented investors, who may adjust their portfolio choices accordingly. The results show that Leader-only portfolios deliver the best performance across all three dimensions.
