Kredi derecelendirme üzerine bir model belirleme

Turkey has an institution-based finance system. In this system all types of businesses acquire the necessary funds mostly from the banks. That is why banks are the most important financial institutions in Turkey. In this regard the most important risk that the banks face is credit risk, which can be defined as the probability of default by the customer and subsequent loss incurred by the bank as a result of the default. For this reason it is of utmost importance by the banks to manage their credit risks efficiently. One of the ways to accomplish this objective is to evaluate the credibility of the businesses that apply for a loan or a line of credit. This evaluation process is called credit rating. In this thesis a multivariate statistical model is developed to accomplish the credit ratings of the businesses that are the potential loan customers of a bank. A unique data set is formed from the loan applications of a bank. Businesses whose loan applications are accepted and businesses whose loan applications are rejected are included in this data set. These businesses form the cases of the model. Independent variables are financial data (financial ratios and percentage changes). The best model for credit rating is found to be the logistics regression model (logit model). The most important variables that distinguish accepted and rejected businesses are found to be the net profit margin and debt ratio. As net profit margin increases, probability of acceptance increases. As the debt ratio increases probability of rejection increases. Key words: Credit rating, banking, loans, multivariate statistical methods