On factors affecting the change in the Gini coefficient of the credit scoring model

Abstract

This article delves into practical challenges associated with monitoring the performance of credit scorecard after its implementation, considering the natural degradation of models over time. The Gini coefficient is commonly employed to assess performance, yet this metric can be influenced by four distinct factors, of which only two indicate the necessity to rebuild the model. The relationship between the Gini coefficient and underlying factors is investigated, employing simple examples and giving appropriate managerial decisions. Moreover, in adverse cases, the positive or neutral dynamic of the Gini can conceal in itself the hidden factors that urge to rebuild the model. The methodology for isolating the impact of these factors on the Gini coefficient is developed. A real-world case study utilizing authentic data demonstrates the comprehensive nature of performance assessment.