ESG-based index tracking with portfolio policy

We incorporate socially responsible investment goals in the index tracking portfolio problem. We propose a comprehensive decision-making framework to construct an ESG-based portfolio to track the market index while ensuring ESG performance at a specific confidence level. To solve the problem, we develop a hybrid method combining genetic algorithms and chance-constrained optimization. Moreover, we estimate a linear portfolio policy to determine the portfolio weights directly using stock-level characteristics. We provide the empirical application of the proposed method using the constituents of the S&P 500 index and quantify the trade-off between the tracking error and the ESG value of the portfolios. The results indicate that imposing ESG criteria at the highest confidence level increases the annualized mean absolute deviation (MAD) by more than 14.5% compared to non-ESG-based portfolios. However, the significance of this trade-off increases with a smaller portfolio size and a shorter investment horizon. Furthermore, the empirical analysis suggests that the size of the firms is more important than their value and momentum in portfolio rebalancing. The proposed system guides investors in their portfolio choice to set targets that best match their ethical as well as monetary goals.