| Abstract [eng] |
While functioning in accordance with the new, business and efficiency-oriented operating model, credit unions develop and begin functioning outside the community. It is universally recognised in scientific literature that as credit unions expand their activities beyond a community, social relations with credit union members weaken and the credit unions lose their social control element, which help them to better assess and manage information asymmetry and credit risk. So far, the analysis of problems of credit risk assessment in the context of credit unions has been limited. The thesis offers a theoretical concept to create a model for credit risk assessment, which allows selecting model creation methods and data so that the make the created model meet the business needs of the creditor. In line with the proposed concept, analyses from three perspectives were performed. In performing analysis from a creditor's perspective, a survey (interview) of 56 credit unions, structural analysis of the sector of credit unions, and analysis of the business loan portfolio of credit unions were conducted. In performing analysis from a modelled segment's prospective, i.e. that of small and very small enterprises, a statistical data sample, which incorporated financial and non-financial data of the enterprises, was formed and analysed. In performing analysis from an external factors' prospective, the external information structure of credit unions and regulatory requirements were analysed. The conducted research allowed creating a statistical model for the assessment of the credit risk of enterprises meeting the business needs of credit unions. The model creation methods were selected on the basis of an assessment of the specificity of a creditor's activities, risk tolerance, and applicable regulatory requirements. The model creation sample was formed with regard to a creditor's target crediting segment. The variables of the model were selected following a comprehensive analysis of scientific a d professional literature and assessment of the possibilities of its practical application based on the study of the external information infrastructure of credit unions. The formed scorecard of the model meets the recommendations of the Basel Committee on Banking Supervision. The cut-off point has been determined by statistical methods, on the basis of the specifics of the portfolio of business loans of credit unions, crediting margin and the LGD of the portfolio. The discriminatory power of the model has been assessed statistically and compared with the latest results of scientific research. The thesis proposes a method for the integration of the model in the decision support system of credit unions. |
