We propose a Bayesian methodology that enables banks to improve their credit scoring models by imposing prior information. As prior information, we use coefficients from credit scoring models estimated on other data sets. Through simulations, we explore the default prediction power of three...