FV-80 | FinTech credit risk assessment: payment-history based measures vs. credit bureau scores

Prof. Y. Lengwiler, K. Rishabh 


Research Topic 
Financial Technology–FinTech– is revolutionizing the payments landscape with almost all major technology companies, such as Amazon, Apple, Facebook and Google, entering the electronic payments business. A by-product of electronic payments is that each transaction leaves a ‘digital footprint’ that contains relevant economic information about the parties involved. As technology has significantly reduced the cost of recording, aggregating and analysing the digital footprints, other payment-related information intensive activities are also going through transformation. One such highly information intensive activities is lending. 

We have entered into a research collaboration with Mswipe, an Indian FinTech company that processes electronic payments between customers, merchants and card issuers. Mswipe ventured into the lending business by offering its merchants an innovative form of credit: so called pay-per-sales (PPS) loans, where the loan amortization is not fixed but proportional to the borrowers‘ sales. Such loans can potentially be a win-win for the borrower and creditor. They give borrowers more flexibility as they allow repayment in good times and ease the budget constraint in bad times, which also reduces the default risk for the creditor. 

Compared to traditional banks, FinTech based payments companies gain an edge in offering such innovative PPS loans: First, they can costlessly acquire information about the borrowers digital footprints captured through past payment data. Second, they have higher seniority than other creditors as they can deduct the outstanding debt right from its source (i.e. the sales). 

Statement of the Problem 
A SECO study reports that two thirds of Swiss small and medium enterprises (SMEs) refrained to apply for bank credit, mostly because they felt discouraged by the high collateral requirements or the burdensome application process. SMEs have to face collateral requirement because they lack credible financial information and have non-existing or short credit histories. It is this reason that makes the application process1 also challenging for the SMEs. However, as SMEs use electronic point of sales systems, they provide payments service provider valuable data about their sales every time a card is swiped. Using the sales data thus generated, the payment service provider can partially assess a SME’s sales, bypassing the need for audited financial records or high collateral. Further, it can provide credit to SMEs that would not qualify for an institutional credit due to their low credit scores, since, now it observes the SME’s performance in real-time and understands the nature of its sales better than what is potentially captured by scores relying on past credit performances, if they exist. 

Our objective is to study: 

  1. How do FinTech based payment companies use their data to offer PPS loans to SMEs? 
  2. Is payment-history based credit assessment more informative than credit bureau scores? 
  3. If so, then which kind of data are more useful in predicting loan performance? 
  4. In which industries these alternative sources of data perform well? 

Importance, Usefulness, and Novelty of the Project 
This project will be first that would use actual transactions data to study merchant’s loan performance. Others studying2, 3 FinTech have only been able to use credit scores that FinTech themselves have developed using alternative data but never the underlying actual data. Most of these FinTech credit scores also take the credit bureau scores as inputs therefore they cannot study one as an alternative to the other. Our setting allows us to study this question cleanly. Our data is also very detailed. It covers each transaction that borrowing merchants made prior to and after taking the loan. Additionally, the project will be first studying the promising potential of PPS loans as new form of credit.  

1Institut für Finanzdienstleistungen Zug and Hochschule Luzern Wirtschaft, Studie zur Finanzierung der KMU in der Schweiz 2016, Staatssekretariat für Wirtschaft, SECO, 2017
2Frost, J., Gambacorta, L., Huang, Y., Shin, H. S., & Zbinden, P. (2019). BigTech and the changing structure of financial intermediation. BIS Working Papers, Number 779. April 2018 
3Jagtiani, J., & Lemieux, C. (2018). The roles of alternative data and machine learning in fintech lending: evidence from the LendingClub consumer platform. Federal Reserve Bank of Philadelphia, Working Papers, 18-13, March