Credit Scoring and the Sample Selection Bias

For creating or adjusting credit scoring rules, usually only the accepted applicant’sdata and default information are available. The missing information for therejected applicants and the sorting mechanism of the preceding scoring can leadto a sample selection bias. In other words, mostly inferior classification resultsare achieved if these new rules are applied to the whole population of applicants.Methods for coping with this problem are known by the term “reject inference.”These techniques attempt to get additional data for the rejected applicants or tryto infer the missing information. We apply some of these reject inference methodsas well as two extensions to a simulated and a real data set in order to testthe adequacy of different approaches. The suggested extensions are an improvementin comparison to the known techniques. Furthermore, the size of the sampleselection effect and its influencing factors are examined.