When Words Sweat: Identifying Signals For Loan Default in the Text of Loan Applications

Oded Netzer, Columbia University, USA
Alain Lemaire, Columbia University, USA
Michal Herzenstein, University of Delaware, USA
We automatically process the raw text in thousands of loan requests from an online crowdfunding platform, and find that borrowers, consciously or not, leave traces of their intentions, circumstances, and personality traits in that text. Moreover, the text is predictive of default up to three years after it was written.
[ to cite ]:
Oded Netzer, Alain Lemaire, and Michal Herzenstein (2017) ,"When Words Sweat: Identifying Signals For Loan Default in the Text of Loan Applications", in NA - Advances in Consumer Research Volume 45, eds. Ayelet Gneezy, Vladas Griskevicius, and Patti Williams, Duluth, MN : Association for Consumer Research, Pages: 53-56.