개인신용 위험관리의 현황 분석

Abstract

I discuss the role of the amount of information as a determinant of the quality of credit risk management. Specifically, I attempt to evaluate differences in the performance of a credit scoring model: the benchmark is when all the necessary informations are available, while the alternative state is when positive informations are lacking. I explain the US case based on existing studies, which show that exclusion of positive informations significantly decreases the per- formance of a credit scoring model measured by Type I and II error. I show circumstantial evidences for the Korean case that the current state of credit reporting system, which mainly delivers negative informations, is likely to be exposed to the same problem. The result strongly suggests that policy makers should introduce measures to extend information coverage of the credit reporting system.