References
Agarwal, Sumit, and Robert Hauswald. 2010. “Distance and Private Information in Lending.” Review of Financial Studies 23:2757–88.
Agarwal, Sumit, Shashwat Alok, Pulak Ghosh, and Sudip Gupta. 2019. “Financial Inclusion and Alternate Credit Scoring for the Millennials: Role of Big Data and Machine Learning in Fintech.” SSRN Scholarly Paper, Social Science Research Network, Rochester, NY.
Bank of England. 2020. “Open Data for SME Finance What We Proposed and What We Learned.” https://www.bankofengland.co.uk/-/media/boe/files/fintech/open-data-for-sme-finance.pdf?la=en&hash=FD4BC43BBD61EDEC5F8460C6BB7488EFDE647581
Barkley, Brett, and Mark Schweitzer. 2020. “The Rise of Fintech Lending to Small Businesses: Businesses’ Perspectives on Borrowing.” The Federal Reserve Bank of Cleveland, Cleveland, OH.
China Finance 40 Forum. 2020. “SME Digital Financing during the Covid-19.” China Finance 40 Forum Research Bulletin No. 551, June.
Chodorow-Reich, Gabriel. 2014. “The Employment Effects of Credit Market Disruptions: Firm-level Evidence from the 2008–9 Financial Crisis.” Quarterly Journal of Economics 129(1).
Cornelli, Giulio, Jon Frost, Leonardo Gambacorta, Raghavendra Rau, Robert Wardrop, and Tania Ziegler. 2020. “Fintech and Big Tech Credit: A New Database” BIS Working Paper No. 887, Bank for International Settlements, Basel, Switzerland.
Demirgüç-Kunt, Asli, and Leora Klapper. 2013. “Measuring Financial Inclusion: Explaining Variation in Use of Financial Services across and within Countries.” Brookings Papers on Economic Activity 2013:279–340.
Eriksson von Allmen, Ulric, Purva Khera, Sumiko Ogawa, and Ratna Sahay. 2020. “Digital Financial Inclusion in the Times of COVID-19.” IMF Blog, July 1.
Frost, Jon, Leonardo Gambacorta, Yi Huang, Hyun Song Shin, and Pablo Zbinden. 2019. “BigTech and the Changing Structure of Financial Intermediation.” BIS Working Paper No.779, Bank for International Settlements, Basel, Switzerland.
Hau, Harald, Yi Huang, Hongzhe Shan, and Zixia Sheng. 2018. “Fintech Credit, Financial Inclusion and Entrepreneurial Growth.” Unpublished.
Huang, Yiping, Longmei Zhang, Zhenhua Li, Han Qiu, Tao Sun, and Xue Wang. 2020. “Fintech Credit Risk Assessment for MSEs: Evidence from China.” IMF Working Paper 20/193, International Monetary Fund, Washington, DC.
International Finance Corporation. 2017. “SME Finance Gap: Assessment of The Shortfalls and Opportunities in Financing Micro, Small and Medium Enterprises in Emerging Markets.” International Finance Corporation, Washington, DC.
Jeffrey, Christopher. 2021. “Zhang Tao on the IMF’s fintech agenda, CBDCs and big tech oversight.” Central Banking, April 20. https://www.centralbanking.com/fintech/7823231/zhang-tao-on-the-imfs-fintech-agenda-cbdcs-and-big-tech-oversight
Leonardo Gambacorta, Yiping Huang, Zhenhua Li, Han Qiu and Shu Chen. 2020. “Data vs Collateral.” BIS Working Paper No. 881, Bank for International Settlements, Basel, Switzerland.
Leonovich, Paul. 2020. “Pandemic Preparedness for Financial Institutions.” IMF Special Series on COVID-19 International Monetary Fund, Washington, DC.
Metrick, Andrew. 2020. “Yale Program on Financial Stability: Government Interventions to Support SMEs during the COVID-19 Pandemic.” https://som.yale.edu/faculty-research-centers/centers-initiatives/program-on-financial-stability/covid-19-crisis
Vardoulakis, Alexandros P. 2020. “Designing a Main Street Lending Facility.” Finance and Economics Discussion Series 2020–052, Board of Governors of the Federal Reserve System, Washington, DC. https://doi.org/10.17016/FEDS.2020.052.
Digital banks are banks that do not have a physical interface and conduct all businesses virtually.
During 2014–2016, China’s banking regulator, the China Banking Regulatory Commission, issued 11 new privately-owned banking licenses, of which three were digital banks, that is, banks that do not have a physical interface with customers, with all services provided over the internet, particularly through mobile devices. These three digital banks are: MYbank of Ant Group, WeBank of Tencent Group, and XWBank of Xiaomi Group.
The short duration of MYbank loans is in sharp contrast with traditional banks. Nearly 57 percent of traditional bank loans for MSEs being longer than one year. Reflecting the short duration, these loans are often used as working capital for operational purposes rather than longer-term investment. In addition, using big data and machine learning models, MYbank dynamically provides pre-approved credit lines to a large pool of MSEs. These MSEs with credit lines then can borrow on the so-called 3–1-0 model, which promises user registration and application within 3 minutes, money transferred to an Alipay account within 1 second, and 0 human intervention (Huang et al. 2020).
This note does not compare digital banks with traditional banks for two main reasons: (1) most of digital banks’ borrowers have never obtained loans from traditional banks, so no comparable data are available; (2) digital bank borrowers that have obtained loans from traditional banks might experience difficulties in borrowing from traditional banks during the COVID-19 lockdown. Therefore, to ensure comparable samples, this note only compare MSEs that borrowed with those that did not borrow.
Lending during periods of financial constraint can boost sales and reduce layoffs—this has been documented in the literature (Chodorow-Reich 2014). For firms that received new opportunities during the pandemic (for example, online food delivery), greater lending support can help expand their business operations to meet the demand, thus increasing sales. For firms that lost opportunities during the pandemic (for example, offline barber shops), lending support can avoid closure of the business, massive layoffs, and suspension of regular equipment purchases and maintenance, all useful to sustain the scale of operations and sales both during the pandemic and when lockdown policies are relaxed.
MSEs that acquire credit access at the beginning of the month may drop out of pre-approved credit lines at the end of the month, given the daily reconsiderations of any MSEs for credit approval by MYbank (Hau et al .2018). This feature makes it difficult to select both treatment and control groups from firms with pre-approved credit lines. Therefore, the selected sample of MSEs that did not borrow during the pandemic could include those either in the pool with pre-approved credit lines or those not in the pool. However, this is unlikely to be a concern that could systematically bias the results as MSEs’ prior borrowing and sales activities were controlled for, as well as the sector and region.
This step is estimated using a probit model Φ1(Yi) = β0 + β1Pisectori + β2regioni + β3FirmChari where Y is the dummy for the firm having borrowed in 2019:Q4, Φ is the probit function and FirmChar includes past (during 2019:Q1-Q3) monthly average sales, loan balance, and the standard deviations of these variables.
The PSM estimation does not aim at finding the best statistical model for explaining the probability of borrowing, but to control, to the extent possible, for variables that could influence both borrowing and the outcome variable (sales growth).
This is also called the conditional independence condition of these firm characteristics for the probability of borrowing.
This assumption is verified by looking at firms’ actual borrowing activities during the pandemic. We find that, holding all other firm characteristics the same, firms with positive 2019:Q4 loan balances indeed had higher actual borrowing from MYbank during the pandemic.
MSEs in the treatment group are not immune to the shock of the pandemic and lockdown. The sales growth rates of the treatment group declined from 38 percent in March 2019 to 10 percent in March 2020.
Sales were not controlled since they are the dependent variable that can be affected by both the pandemic and lending activities (such as business sales during the pandemic). On credit lines, since the average historical monthly borrowing as well as the standard deviation of the borrowings in addition to other firm characteristics were controlled for, it is unlikely that sampling procedures resulted in systematic biases in firms’ credit lines.
It is not assumed that firms that did not borrow in 2019:Q4 did not have access to credit. Instead, 2019:Q4 is only used to represent the different propensities to access digital lending.
Regression results are available upon request, however, as confidence bands are not available, results should be seen as indicative for now.
Lending to small businesses using collateral is less feasible, especially during a lockdown. A collateralized loan would be expensive, since there could be added costs arising from collateral verification, documentation, registration, and monitoring—to some extent these need a physical presence. In addition, the cost of foreclosing the collateral may be higher than the loan exposure itself. As a result, the income from the loan may not be able to cover these collateral costs. Therefore, the traditional approach to using collateral to reduce information asymmetries is unlikely to be feasible for small business lending.