Risk Scorecard

ScoreCard Improvisation


  • To review scorecard for a car loan provider in UK
  • The client had a scorecard developed by another vendor but was dissatisfied with the performance
  • The client engaged us to review and update the scorecard with additional variables


  • Bootstrapping was used to overcome the limitation of a small sample.
  • Reject rates were taken as a surrogate for default rate.
  • Model Gini = 66.74* & Model KS = 53.85** indicating high quality scorecard


  • Details on scorecard development methodology not available.
  • Important variables such as Job Type, Residence Type and Residence Years were not considered while building the model.


  • Scores recalibrated with default data after sometime.

Download Case Study

    Share This Case Study