A0885
Title: Greener strategies, stronger results? Environmental sustainability and firm performance in Italy
Authors: Luca Patelli - University of Bergamo (Italy) [presenting]
Peter Cincinelli - University of Bergamo (Italy)
Daniele Toninelli - University of Bergamo (Italy)
Giovanna Zanotti - University of Bergamo (Italy)
Abstract: In the current context of environmental challenges, international institutions have established targets with the aim of mitigating the impact of climate change. The aim, focused on Italian firms, is to investigate whether sustainability, and more specifically the environmental dimension, affects the probability of observing firms in a state of financial distress. The probability of distress is estimated by considering financial and accounting variables, as well as variables derived from non-financial disclosures and good sustainability practices; variables available in external ESG databases are also taken into account. To predict the probability of distress, a statistical machine learning framework is relied on. These methods enable flexible handling of a large number of predictors of different types. In order to overtake the black-box nature of several machine learning algorithms, a set of explainable AI solutions is implemented. This approach allows assessing the variables' importance as well as to enhance the model's explainability, thereby facilitating a deeper and clearer comprehension of the prediction mechanisms and of the obtained results.
