Non-compliance to Benford distribution as the portfolio default rate determinant in online retail lending

Article type: Research Article

Authors: Penikas, Henry

Affiliations: Research and Forecasting Department, Bank of Russia, Moscow, Russia | E-mail: penikas@gmail.com

Correspondence: [*] Corresponding author: Research and Forecasting Department, Bank of Russia, Moscow, Russia. E-mail: penikas@gmail.com.

Abstract: There is large online lending growth in volume world-wide. The credit risk concerns point to the fact that most of these loans might be used to redeem earlier borrowed funds. However, the true reasons for online borrowing and lending are unavailable. Benford law is one of the tools used by auditors to monitor how suspicious the transaction is. That is why I wish to study one of the publicly available lending portfolios. Our objective is to trace associativity of compliance to Benford law and reported default rates. I find that MAE is a more statistically significant determinant of the country portfolio default rate, than RMSE. Moreover, the least creditworthy portfolios seem to be those with the MAE around 52–56%, while the closest to Benford and the least adjacent distribution do not demonstrate that large default rates.

Keywords: Peer-to-peer lending, P2P, default rate, Mintos, maturity, LTV

DOI: 10.3233/MAS-221404

Journal: Model Assisted Statistics and Applications, vol. 18, no. 2, pp. 125-134, 2023