The impact of the ESG factor on industrial performance
An analysis using machine learning techniques

M. Palynska, F. Medda, V. Caivano, G. Di Stefano, F. Scalese

Quaderno finanza sostenibile (paper) no. 4 - June 2024 [PDF]
 

Abstract
In recent years, the transition towards a more sustainable economic model has become increasingly important for both financial market participants and supervisory and regulatory authorities. In a context in which regulations are constantly evolving and the information ecosystem is progressively being refined, empirical analyses on ESG issues increasingly make use of artificial intelligence techniques. The aim of this study is to show how machine learning methods can contribute to understanding the relationship between ESG performance and corporate earnings performance. The study is based on data from over 850 European and US companies over the period 2007-2021. Its aim is to analyse the link between the ESG score (used as indicator of the company's sustainability profile) and EBIT (used as indicator of the earnings profile). The hypothesis tested in this study is that the three pillars E, S and G have a different impact on company earnings performance. The results indicate that the environmental score (pillar E) is more strongly associated with earnings performance than those relative to the other two pillars (S and G). Additionally, the work reveals differences in the results for European versus US companies, which may be ascribed to the different regulatory frameworks in the two jurisdictions. The research contributes to the literature on the use of machine learning techniques for the analysis of sustainable finance issues, showing how such methods can enhance research activity on these topics. However, the use of machine learning techniques cannot fully address the critical issues related to the quality of ESG metrics currently available. The ongoing definition of sustainability standards and metrics will facilitate the collection and analysis of structured data. Furthermore, the evolution of ESG rating regulations will increase transparency about the underlying methodologies.  
 

Authors
Marta Palinska - (during an internship at CONSOB at the time of the study)
Francesca Medda - (CONSOB at the time of the study)
Valeria Caivano - CONSOB, Research Department (v.caivano@consob.it)
Giovanna Di Stefano - CONSOB, Research Department (g.distefano@consob.it)
Francesco Scalese - CONSOB, Research Department (f.scalese@consob.it)

The opinions expressed, any errors and inaccuracies are solely attributable to the authors. In citing this work, therefore, it is not correct to attribute the arguments expressed therein to CONSOB.