Modeling credit risk for small and medium-sized enterprises : evidence from Norway
MetadataShow full item record
- Master Thesis 
Primarily, this paper investigates the determining factors of default in the Norwegian small and medium-sized enterprises (SMEs). Using logit regressions on a database comprises over 280,000 Norwegian firms (with sales less than 500 million kroner, and employees less than 250 persons), three default prediction models are developed. These three models designed to predict default event in one, three, and five year from now based on the today’s available information. These models have out of sample prediction powers which are approximately 15% (on average) higher than the models which are available for Norwegian SMEs. A secondary objective of this paper is to examine the proposed models’ ability to decrease bank capital requirements based on the latest Basel Capital Accord’s guidelines for SMEs. Throughout breakeven analyses, for any combination of SMEs (as retail customers and corporates) in banks’ portfolios, all models show lower capital requirements than the one suggested by the Basel III. Furthermore, the Basel III suggested a one-year default probability model as the basis for capital requirements calculation under the Internal Rating Based (IRB) approach. By a simulation over a sample of randomly selected SMEs, capital requirements are calculated using probabilities resulted from the one-year model and mixture of the one, three, and five-year models (corresponding to maturities of intended loans). This simulation confirms that using one-year probabilities of default for longer maturities slightly underestimates the calculated capital requirements under IRB approach.