As one is able to score an individual based on his level of credit risk, one should be able to score a group of potential borrowers with similar characteristics. To achieve this end, one should acknowledge that the risk of a group comes from the diversity of its members. This research proposes a methodology for applying credit scoring to groups of borrowers with similar characteristics, relying on the composition of a particular group and on the distribution of the probability of default within the group. This methodology allows ranking different cohorts of the population by their risk level, and considers in the ranking the risk preferences of the decision maker.
De clic aquí para ver el Working Paper