Credit risk management in banking, credit risk measurement models and a model application on Turkey

Credit risk arising from lending activities is one of the main risks for banks. Accurate measurement of credit risk is indispensable for effective credit risk management, and credit scoring and credit rating models are used extensively for this purpose. International Basel Accords also attach great importance to the measurement of credit risk. Banks' internal rating models can be used to measure credit risk within the scope of Basel standards when they meet the necessary conditions. Considering the number and volume of credit transactions, the use of these models has become indispensable. Credit scoring models developed in the literature and banking practice use various statistical methods. With regard to the difficulties in measuring credit risk, it is not possible to mention the existence of a perfect credit scoring model or an ideal statistical method. The aim of this thesis is to present a framework for credit risk management in the light of banking practices and national and international regulations, and to develop an applicable credit scoring model in the Turkish market. In this context, the available data of companies traded in Borsa Istanbul for the period of 2017-2020 were analyzed, and the factors affecting the credit risk of companies that announced credit restructuring with banks as a result of negative developments in the economy in the same period were investigated. A credit scoring model was developed using the logistic regression method, by defining the said firms as non-performing loans. Despite the constraints in the data set, the model was successful in distinguishing between performing loans and non-performing loans by using four financial ratios. It is expected that the model results will benefit the researches in the literature and the credit risk measurement models in the banking sector.