This study aims to analyze the relationships between the economic performance of Italian listed banks and their GRI disclosure (GRID), understood as the level of disclosure of their non-financial reports according to the GRI standards. The study selected 6 among the Italian listed banks with the highest capitalization as of 31/12/2020 and analyzed the relationships between their economic performance and their GRID by applying three models: Linear Regression, Support Vector Machines, and Decision Trees. The research highlighted the existence of positive relationships between the economic performance of banks – measured in terms of capitalization, size and leverage – and their GRID, while the relationship with profitability is negative. Unlike the analyzes that see disclosure as a factor capable of improving economic performance, this research starts from the assumption that the best economic performance favors a wider disclosure. Furthermore, the study applies machine learning which represents a non-traditional methodology, not yet fully exploited in the field of sustainability reporting.
JEL classification numbers: M21.
Keywords: Non-Financial reporting, GRI standards, Banking sector, Economic performance, Machine learning.