Creation of an inflation-beating investment portfolio based on ESG criteria managed by machine learning

Abstract

This paper explores the creation and management of an investment portfolio aimed at outperforming inflation, integrating environmental, social and governance (ESG) criteria, and using machine learning tools to manage the portfolio by improving investment. The main objective of the study is to design an investment strategy that is not only financially profitable, but also in line with sustainable practices. For the construction of the portfolio, various investment philosophies, markets, and assets in different sectors and geographies were analysed, prioritising those that meet the best performance and high ESG standards. The XGBoost machine learning model was then used for asset management, proving to be an effective tool in predicting future movements. The resulting portfolio was compared to the MSCI ACWI ESG Universal index as a benchmark, showing superior performance in terms of risk-adjusted returns and compliance with ESG criteria. In addition, various statistics and risk metrics were evaluated to ensure that the portfolio was not only cost-effective, but also consistent. The results of the study highlight the feasibility of combining responsible investments with advanced technologies to improve the profitability and efficiency of investment portfolios, offering a replicable and scalable model for investors seeking to contribute to sustainability without sacrificing profitability.