In 2016, several Internet companies in China announced their intentions to offer digital credit in rural areas. The development of digital credit aims to tackle limitations faced by traditional microfinance (high operational costs and in turn, high interests rates) by automating routine operation and exploiting existing digital records. The annual interest rate these Chinese companies promised is around 9%, significantly lower than the Asian median (21%) . Because the concept of digital credit is a relatively recent invention, there is very limited research evaluating its impacts on consumers. This pilot study aimed to work with an anonymous company in China to evaluate the impacts of their digital loans on credit-constrained households in rural areas.
Unfortunately, after only one month following its launch, the company’s digital credit program was suspended indefinitely; a new regulation from the China Banking Regulatory Commission greatly limited the scope of services that a financial technology company without a commercial banking license could provide.
Due to the indefinite suspension of the company’s digital credit product (which was also met with lackluster demand in the areas it was launched), the research team instead tested a household questionnaire with consumers of a separate husbandry loan product offered by the same company. In the rural county of Sihong in Jiangsu province, researchers surveyed eight households and conducted interviews with several village leaders. Although Sihong is one of the least developed counties in the province, there is not much extreme poverty in the area; households are generally able to make ends meet but lack opportunities to increase income.
The husbandry loan can only be used to buy animal feed. Households apply for a loan amount for each batch of their livestock (primarily chicken and ducks), and then use the credit to buy animal feed whenever they see fit. A key feature of this loan is a statistical model estimating the amount of feed the households need for each batch at the time of the withdrawal request. The system will automatically disburse the money within 30 minutes if the request is reasonable, but if the request deviates from the company's estimation, the system will suspend the transaction and assign an agent to verify the request.
The household survey collected information relevant to measuring a rural household’s well-being with a special emphasis on production activities and financial situation (the first of its kind in rural China). These eight households are selected from the research area, and not limited to husbandry loan recipients. They had diverse backgrounds; besides raising chickens, some families were planting wheat, others were planting mulberry and harvesting leaves to raise silkworms, and a few relatively well-off families were investing in aquaculture.
During Summer 2017, PhD student Zenan Wang visited separate chicken and duck farms in Shandong and Hebei in order to conduct qualitative interviews with husbandry loan recipients and observe the businesses. The chicken farms were fairly clean and advanced with environmental control systems. A typical farm room can house up to 13,000 chickens, and with the help of equipment, only two employees are needed to manage the farm. Annual profits (not excluding labor) can reach $15,000 USD for one room and according to the interviewed employees, the majority of chicken farms in China are similar in terms of size, cleanliness, and profitability. The loan recipients believed they are benefiting from the new opportunity of husbandry loans because they now have easier access to credit and are paying less interest.
Results and Policy Implications
Lackluster demand for loan product: Despite the near impossibilities of getting loans from national banks, most rural households in Sihong are able to borrow from local commercial banks at higher rates. While the company’s loan product offered a lower interest rate than the local commercial banks, it struggled to compete in other aspects. The online application for the loan product was daunting compared to the familiarity of the loan application with local banks. Additionally, the risk management model used to automate loan decisions was very conservative; loan recipients complained that the awarded loan-size was much smaller than what they requested and also smaller than what local banks would have offered.
Husbandry loan product offers attractive features: The households’ agricultural activities require a large amount of input investment at the beginning of the season, yet revenue is not collected until the harvest. Hence, a lack of credit constrains families’ abilities to expand their production. The recipients of the husbandry loan expressed their satisfaction with this loan, since it is very easy to withdraw funds and they are paying lower interest.
Important lessons learned from the design of this loan include:
• Withdrawal limit – The loan does not disburse all at once, but instead in smaller amounts at the recipients’ request. This withdrawal limit imposed by the company may serve as a commitment device for borrowers to better manage their farm. The convenience of receiving frequent small loans, if combined with smart contract design, could lead to an improvement for both borrowers and lenders.
• Fraud detection tools – Adding fraud detection tools can help reduce the risk the company faces, therefore enabling a more competitive interest rate. For borrowers, this will mean lower interest rates.
The suspension of the digital loan product hindered the research team’s ability to evaluate the impacts of the product on credit-constrained households in rural China. However, understanding the constraints that are keeping households from applying to loans can assist in the design of new policies aimed at improving loan take-up rates for the credit-constrained. Findings from this study also have the potential to effect discussions with Chinese commercial banks interested in designing a loan product with a commitment contract and small and frequent disbursements.
April - November, 2017
 Rosenberg et al., 2013
Photo Credit: Zenan Wang