The study, published in Finance Research Letters, focuses on determining whether ESG (environmental, social, and governance) information truly helps in early detection of companies experiencing financial problems. The findings differ between Spain and the UK due to the quality and homogeneity of the published data.
In the United Kingdom, where companies publish this information more comprehensively and uniformly, the inclusion of ESG indicators improves the accuracy of financial health predictions. However, in the analyzed Spanish companies, the lack of available data and its inconsistency across firms significantly limits the contribution of ESG information, even leading to worse predictions.
"More sustainability information does not necessarily mean better predictions. For it to truly help detect financial risks, that information must be rigorous, homogeneous, and comparable between companies," states Florentina Iulia Voicila, a researcher at the UCM's Department of Accounting and Finance.
The research utilized data from 77,140 Spanish companies and 10,085 British companies, spanning from 1996 to 2023 for Spain and from 1985 to 2023 for the UK. Companies were classified by their financial situation (secure, intermediate, or in difficulty) using the Altman Z-score, and traditional statistical models were compared with two machine learning models (AI) to assess the added value of ESG information.
Artificial intelligence models proved more effective, although their improvement in financial prediction thanks to ESG data was moderate. "Even in these more advanced models, the added value of ESG information depends on the quality of the available data," Voicila adds. These findings are useful for financial institutions, regulators, and companies themselves, and open avenues for new AI-supported analysis models.
