Monitoring the behaviour of credit card holders with graphical chain models

Consumer credit has become an enormous business in industrialized countries. Recently, finance agencies have started to develop new products aiming not only to widen their portfolio but also to keep active relationships with good clients already taken on file and to prevent bad clients from becoming a loss for the agency. As a result, models for the whole behaviour of the clients are necessary. They involve many related outcome variables, which altogether give a measure of whether the client revealed to be profitable or unprofitable. This paper aims to show the potential of graphical chain models in the described context.