The dynamic changes in the business environment necessitate a modification of risk assessment, adjusting credit rating models and methods essential. In addition to economic performance, a company's profitability is also influenced by ecological and societal aspects. Therefore, the question arises whether purely financial factors still provide a solid foundation for credit rating. The consideration of non-financial, especially sustainable factors, can primarily be attributed to current regulatory developments. The financial system is ascribed a crucial role in promoting a more sustainable economy. Non-listed SMEs will play a significant role in this regard. SMEs are considered the backbone of the economy; however, the prevailing models for credit risk assessment are primarily tailored to the characteristics of large companies. Consequently, it is essential to develop specially designed credit rating models to optimally allocate resources while minimizing the risk of loan defaults as much as possible. To adequately reflect the current state of the literature, this master's thesis precisely examines which models and methods of credit rating are used for non-listed SMEs in the literature and what role non-financial, sustainable factors play in this context. To address these research questions adequately, a total of 18 empirical studies were identified as relevant and properly analysed using a systematic literature review. Academic research shows that the methods in the literature vary significantly: on the one hand, established models are applied and evaluated on specific datasets, and on the other hand, models are modified and expanded through the integration of additional factors. Furthermore, research also employs more advanced concepts and hybrid models to adequately capture the complexity of credit risk assessment. Nevertheless, logistic regression is the dominant method for determining the probability of default in current credit risk research for SMEs. The reasons for the popularity of logistic regression lie in its predictive ability, the good interpretability of the results, and its suitability for samples of different sizes. Alternative methods include linear discriminant analysis, support vector machines, multivariate regression analysis, the lasso-logistic model, the hazard model, and fuzzy-BWM. The integration of non-financial variables in the rating models shows that these variables, as predictors of corporate defaults, significantly improve the predictive capability and accuracy of the assessment models. Due to the partially limited availability of financial information for SMEs and the fact that financial metrics are often retrospective, non-financial variables, especially the age and size of the company, already constitute an integral part of credit rating models for SMEs. Regarding sustainability and ESG-factors, only one non-financial variable, the audit variable – as an indicator of corporate governance – stands out. Although the audit factor has been identified multiple times as a significant variable with governance relevance in terms of ESG-criteria, the result falls short of expectations. Despite the incorporation of sustainability topics into banking regulation and the increasingly noticeable impacts of climate change and social injustice, credit rating models for SMEs have not yet been adequately adopted for academic research.
Modelle und Methoden des Kreditratings in der Literatur und die Bedeutung von nicht-finanziellen, nachhaltigen Faktoren – mit Fokus auf nicht börsennotierte KMUs
Hamader, V. (Author). 2024
Student thesis: Master's Thesis