Abstract
Based on a data set of nearly 43,000 personal loan applications from Germany, this paper empirically determines key factors of application fraud such as, for instance, the sales channel or the loan amount. This is done univariately as well as by employing a logistic regression, which is found to be a statistically significant approach for profiling loan application fraudsters. Besides in-sample ...
Abstract
Based on a data set of nearly 43,000 personal loan applications from Germany, this paper empirically determines key factors of application fraud such as, for instance, the sales channel or the loan amount. This is done univariately as well as by employing a logistic regression, which is found to be a statistically significant approach for profiling loan application fraudsters. Besides in-sample and out-of-sample verifications, we also prove the economic significance of our results by developing a fraud management framework taking the fraud rate, the average default cost due to fraud as well as the fraud screening costs into account.
Altmetric