CMStatistics 2021: Start Registration
View Submission - CFE
Title: ESG factors and firms' credit risk Authors:  Vittoria Cerasi - Bicocca University (Italy) [presenting]
Matteo Manera - DEMS Bicocca University (Italy)
Laura Bonacorsi - DEMS Bicocca University (Italy)
Abstract: The link between the risk of default and Environmental, Social, Governance (ESG) factors is studied using supervised machine learning techniques on a cross-section of European listed companies. We focus on ESG factors instead of ESG scores, to avoid relying on unobserved models used by rating companies to construct these ratings. The advantage of this approach is also that our results can be applied to non-rated corporations. We proxy credit risk by using the Altman z-score, which is a linear combination of accounting ratios used to classify companies in different categories according to their risk of default. Our sample is a cross-section of 1251 European firms in the year 2019, and it includes 590 candidate variables. Due to the huge number of variables involved, we employ techniques of supervised machine learning, in particular, the Least Absolute Shrinkage and Selection Operator (LASSO), to select the relevant explanatory variables. Since our objective is to predict the sign of the selected variables on the risk of default, we use Lasso for inference methods. The preliminary results show that a selection of ESG factors in addition to the usual accounting ratios helps to explain a firm's probability of default. We also develop a model to explain short-term credit rationing augmented for ESG factors to interpret the results.