This study develops a one-year probability of default prediction model for unsecured retail loan applicants for a particular top-tier Zimbabwean commercial bank. Banks need this model calculation for them to be internationally Basel II/III com- pliant. Binary logistic modelling was used on 10521 cases of which 2770 were defaulting and 7751 were non-defaulting. Retail credit risk parameters such as monthly income, marital status, time with bank, loan amount, other debts, gender, time with employer, loan period and home status were considered as critical predictors of probability of default. The model can be used for internal rating purposes using advanced internal ratings-based approach in ca
Download