In banking practice, quantifying the probability of default is one of the most important elements of the lending decision, therefore it is also vital from a financial stability perspective. The aim of our research was to model the probability of default as precisely as possible in the case of micro, small and medium-sized enterprises. By linking the data from the Central Credit Information System (KHR) and companies’ financial statements, a database was created that covers all the SMEs with loan contract, thus we were able to examine credit risk based on a uniquely large group of enterprises. In our research, we created models that enabled us to produce estimates based on certain corporate features about the probability of default of micro, small and medium-sized enterprises. Our analysis revealed that modelling these size categories separately and managing non-linear effects in the case of several variables are especially important. In addition, the impact of the macroeconomic environment on credit risk also proved to be important in the fitting of our estimates.
Keywords: SME, credit risk, credit register, logit model, probability of default.
JEL codes: C25, G20, G21.