Racial disparities in the traditional small business credit market are a well-documented phenomenon in the U.S. With the recent development of more automated and anonymized lending technologies comes a natural question: do fintech lenders serve minority-owned businesses who are in the blind spots of traditional lenders? In our recent project, we investigate whether minority borrowers are more likely to use fintech lenders in the 2020 and 2021 Paycheck Protection Program (PPP). Moreover, we evaluate whether fintech lenders can lower racial barriers by serving relatively weaker – from a credit risk standpoint – minority borrowers. We also study racial disparities in previous lending relationships and how this relates to fintech usage.
As an example of why fintech lenders can benefit minority and non-minority borrowers to a different degree, think about a comparison between applying for the PPP loans through Square and the local bank. To apply through Square, everyone can create an account on squareup.com. Following several simple steps without face-to-face human contacts, a borrower can get the loan approved. In contrast, loan applications through banks still require human contact via the telephone. There could be racial discrimination, either intentionally or unintentionally, in the process of human contact. Moreover, as the application procedure through telephone has a capacity limitation, banks are very likely to first serve customers with previous lending relationships who tend to be non-minorities. These distinct features make fintech lenders more appealing to minority borrowers.
We restrict our sample to 98,000 restaurants across the U.S. where we find a link on yelp.com. Looking at restaurants has several advantages. First, it allows us to build a proxy for minority-owned businesses using the food type from Yelp.com to address the data limitation of missing race and ethnicity information for almost 80% of the PPP loans. Second, because all restaurants are eligible for the PPP, loan recipients in our sample do not have differences in terms of regulatory restrictions. Third, we can use the Yelp rating of the restaurant, as a proxy for operational performance, to indirectly detect racial barriers. Suppose we find evidence of a more negative minority-non-minority rating gap (the difference in Yelp ratings between minority and non-minority restaurants, racial gap for short) for fintech lenders, it means that fintech lenders are more inclusive and easier to use for minority borrowers.
To start with, we find evidence of a positive and significant association between minority ownership and fintech usage. Black-, Asian-, and Hispanic-owned restaurants are more likely to use fintech lenders by 9.17%, 8.44%, and 1.22% respectively, with the sample mean being 9%. Observed business characteristics, including employment size, franchise, and business type, account for 12% of the variation in the fintech usage for Black- and Asian-owned restaurants and 28% for Hispanic-owned ones. Across-city differences account for 32% of the variation for Black-owned and 16% for Asian-owned restaurants. For the Hispanic group, we observe limited racial disparities in fintech usage after controlling for a cross-city differences.
Then we investigate the difference in racial barriers for fintech and traditional lenders. It is hard to directly measure racial barriers. Instead, we look at the gap in terms of the Yelp ratings. We find that the racial gap is more negative for fintech lenders. This suggests that fintech lenders, compared to traditional lenders, are more valuable to minority borrowers. Exploring heterogeneity among lenders, we find that the four largest banks in our sample – JPMorgan, Bank of America, Wells Fargo, and U.S. Bank – do not have a large racial gap in ratings. In contrast, relatively smaller banks have pronounced racial gaps.
As to the lending relationship channel, we also find suggestive evidence. We find that minority-owned businesses are less likely to have previous lending relationships. Moreover, our evidence shows that borrowers without lending relationships are more likely to gain a PPP loan through fintech lenders. Interestingly, after controlling for lending relationships, the positive association between minority ownership and fintech usage only changes slightly. This suggests that lending relationships cannot fully explain why minority borrowers are more likely to use fintech lenders.
We further study the relative importance of the “matching value” channel and the “lending relationship” channel using a structural estimation approach. The structural estimation approach has the strengths of taking into account the interactions among different players in the matching game and no requirement of data on transferred payments. Consistent with the regression analysis, we find positive coefficients for both channels. The positive coefficient on the interaction between the fintech lender and minority borrower indicators means that fintech-minority matches generate higher levels of matching values than other matches. Similarly, the positive coefficient on the lending relationship dummy means the match between a borrower and a lender with prior relationships has higher matching values than matches without prior relationships. Moreover, we find that the “fintech-minority matching channel” is 0.69 times as important as the “lending relationship channel” in the 2020 PPP and 0.49 times as important in 2021.
Our paper contributes to the literature on racial disparities in the small business credit market in several ways. First, we provide novel evidence to the scant literature on this important topic. Despite that there is a large body of literature on the PPP program, few papers look at racial disparities. Two contemporaneous papers, Chernenko and Scharfstein (2021) and Howell et al. (2021), like our paper, link the PPP dataset with other data sources to create a firm-level minority measure, which allows for a richer set of findings. Using a different research design, our evidence is largely consistent with the findings in the contemporaneous papers and provides further support for the existence of racial disparities in the PPP.
Second, we look at racial disparities through the lens of the operational performance measure. We show a specific type of financial inclusion role of fintech lenders are more inclusive to the lower range of minority-owned businesses.
Third, we emphasize the channels through which fintech lenders can address the issue of racial disparities. Our paper highlights a borrower-side benefit of fintech lending for the minority group. A more automatic lending process lowers the level of racial barriers to approaching lenders. Other contemporaneous papers focus on the lender side and show that an anonymized lending procedure can reduce taste-based discrimination. In addition to the channel captured by higher matching values between fintech lenders and minority borrowers, we also show the channel on racial disparities in lending relationships and quantify the relative importance of different channels.
Celine Yue Fei is a postdoctoral research associate at the University of North Carolina at Chapel Hill
Keer Yang is a PhD student at the University of Minnesota
This post is adapted from their paper, “Fintech and Racial Barriers in Small Business Lending” available on SSRN.
The views expressed in this post are those of the authors and do not represent the views of the Global Financial Markets Center or Duke Law.