@article{bilas2026esg,
  title = {Do ESG Ratings Predict Credit Risk?  Evidence From Croatian Firms and A Machine Learning Perspective},
  author = {Vlatka BILAS and Tomislav RADOSand Lana FRKOVIC},
  year = 2026,
  url = {https://ibimapublishing.com/p-articles/47ENV/2026/4713926/},
  journal = {Communications of International Proceedings},
  volume = 2026 (5),
  doi = doi.org/10.5171/2026.4713926,
  abstract = {This study investigates whether ESG ratings provide useful information for credit risk assessment in an emerging EU member-state context, where sustainability reporting is rapidly expanding but still institutionally developing. The motivation for the study arises from the increasing regulatory and financial relevance of ESG disclosure in the European Union and from the practical need to understand whether ESG indicators can support creditworthiness evaluation. Although prior research often links stronger ESG performance with lower credit risk, less is known about whether this relationship is observable in smaller and less mature markets, particularly during the early phase of adjustment to EU sustainability reporting requirements. To address this gap, the paper analyses firm-level data from the Croatian Chamber of Economy for 2024–2025. ESG is examined both as an aggregate score and through its Environmental, Social, and Governance components, while credit risk is measured using HGK creditworthiness indicators. The methodology combines descriptive and stratified analysis, pooled and within-firm econometric models, ordered-response and transition analyses, and random forest prediction to assess out-of-sample performance. The findings indicate a weak and unstable relationship between ESG measures and HGK credit ratings. Changes in ESG scores do not systematically translate into changes in credit ratings, and ESG variables add only modest predictive value. Overall, ESG ratings in this context appear to reflect firm size, sector, and reporting capacity more than a strong standalone signal of credit risk.},
  keywords = {credit risk, ESG ratings, machine learning, Croatia},
  note = Article ID: 4713926
}