How good are the credit scoring models of P2P lenders?
Peer-to-Peer lending platforms (P2P lenders) have been good for Small and Mid-sized Businesses (SMBs) and individuals. P2P lenders have reduced service costs and extended lending to those unable to obtain funding from banks, post Credit Crunch (2007/2008).
They have also offered relatively safe ways for investors to earn higher interest. But the track record of these alternative finance platforms is limited to a few years with easy monetary policies.
At the same time, regulatory bodies have not considered P2P lenders and other crowdfunding businesses a meaningful risk to the financial system stability, yet. But regulators have been monitoring the rapid growth of the new forms of finance these platforms provide.
The improvements P2P lenders bring may come at the price of a worse credit risk measurement, which can hamper lenders and endanger the stability of a financial system. This possibility has led to new research in the field.
In the technical report “Network-based scoring models”, submitted to the Journal of Banking and Finance earlier this year, Paolo Giudici and Branka Hadji Misheva of the University of Pavia, find that the predictive accuracy of P2P credit scoring methods could improve, using the information on borrowers’ networking.
Their findings show that including network information does improve predictive accuracy and, in addition, provides useful information on which borrowers are most central and, therefore, more contagious, from a systemic risk perspective.
Although the report considers regulators and investors of P2P lenders the main beneficiaries of the findings, we believe that the management teams of P2P lenders could benefit from the research too.
To download a copy of the presentation of the report, please click on this link:
Network models for credit risk management in P2P lending