Download this Paper Open PDF in Browser

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

65 Pages Posted: 7 Nov 2016  

Oded Netzer

Columbia Business School - Marketing

Alain Lemaire

Columbia University, Columbia Business School, Marketing, Students

Michal Herzenstein

University of Delaware

Date Written: November 6, 2016

Abstract

The authors present empirical evidence that borrowers, consciously or not, leave traces of their intentions, circumstances, and personality traits in the text they write when applying for a loan. This textual information has a substantial and significant ability to predict whether borrowers will pay back the loan over and beyond the financial and demographic variables commonly used in models predicting default. The authors use text-mining and machine-learning tools to automatically process and analyze the raw text in over 18,000 loan requests from Prosper.com, an online crowdfunding platform. The authors find that loan requests written by defaulting borrowers are more likely to include words related to their family, mentions of god, short-term focused words, the borrower’s financial and general hardship, and pleading lenders for help. The authors further observe that defaulting loan requests are often written in a manner consistent with the writing style of extroverts and liars.

Keywords: loan default, text mining, consumer finance

Suggested Citation

Netzer, Oded and Lemaire, Alain and Herzenstein, Michal, When Words Sweat: Identifying Signals for Loan Default in the Text of Loan Applications (November 6, 2016). Columbia Business School Research Paper No. 16-83. Available at SSRN: https://ssrn.com/abstract=2865327

Oded Netzer (Contact Author)

Columbia Business School - Marketing ( email )

New York, NY 10027
United States

Alain Lemaire

Columbia University, Columbia Business School, Marketing, Students ( email )

New York, NY
United States

Michal Herzenstein

University of Delaware ( email )

Newark, DE 19711
United States

Paper statistics

Downloads
598
Rank
36,815
Abstract Views
2,117