Factors of ESG ratings assigned to commercial banks – the cultural and credit risk dimensions


  • Paweł Niedziółka SGH Warsaw School of Economics
  • Michał Bernardelli SGH Warsaw School of Economics
  • Zbigniew Korzeb SGH Warsaw School of Economics


commercial bank, ESG, rating, cross-cultural analysis


The purpose of the study was to examine the impact of cultural differences and credit ratings on the ESG (Environmental, Social, and Governance) scores assigned to commercial banks. The analysis was performed using ordered logistic regression. The Akaike information criterion (AIC) and the statistical significance of the explanatory variables were used as the method of comparing the models. Count R2 and adjusted count R2 were chosen as the measure of goodness of fit, but a non-diagonal element analysis of the contingency table was also performed. Based on the data of 330 banks from 50 countries the study proved that among all the clusters considered, the region with the highest ESG risk attributable to banks was the Arab countries, whereas regions with the lowest ESG risk were Western European and Nordic countries. Cultural variables found to influence ESG ratings were MAS (masculinity vs. femininity), PDI (power distance index), LTO (long-term orientation vs. short-term orientation), and UAI (uncertainty avoidance). In addition, a relationship between banks’ ESG and credit ratings was observed, i.e. an increase in the average credit rating reduces the chance of classifying a bank as high or medium ESG risk vs. ESG low risk. This is the first study focused on binding the cultural characteristics of a given country with the results of ESG risk assessments of banks registered in this jurisdiction as well as assessing the relation between ESG and credit risks. This kind of clustering with the use of an econometric model allows for capturing the interrelationships of many factors simultaneously with full interpretability of the results. An added advantage was the excellent accuracy of the model, even though the maximisation of the predictive power was not the key aspect of the research.