About the Event
A core challenge of the banking sector is securing sufficient data to accurately credit score agricultural clients, especially smaller sized farms. The lack of "financial footprints" of farmers can act as a real barrier to banks wanting to expand their lending to such clients. Fortunately, new technology aligned with big data analytics is providing a means of overcoming this challenge, enabling financial institutions to utilize "non-traditional" data for credit scoring farmers.
At our recent webinar, Michael Mbaka, Senior Innovations Specialist, FSD Kenya, spoke about the experience of FSD with regards to how banks and other financial institutions are starting to use alternative data to solve the credit scoring challenge. Michael identifies and details some of the alternative data sets and how they are used to generate credit scores for agricultural clients. He explains how these new credit scoring models work, and shares the promising preliminary results of organizations who have put the new designs into practice.
Featured Topic | Digital Finance in Agriculture
About the Presenter
Senior Innovation Specialist
Michael is involved in managing agricultural and risk related projects in FSD. Some of the initiatives include managing agriculture risks through an index based weather insurance pilot project and currently a national agriculture insurance macro programme; R&D to support financial institutions to develop innovative financial products as well as projects that targets managing other micro consumer risks.
Alternative Data for Financial Decisions: Lending Using Alternative Data
Key Messages from our Recent Webinar
AgriFin's recent webinar shared information about new credit scoring models built with non-traditional data which offer potential to help financial institutions (FIs) make better-informed decisions when lending to the agriculture sector. Michael Mbaka from Financial Sector Deepening Kenya (FSD Kenya) explained how currently, most the credit for individual smallholder farmers in developing countries comes from informal sources. While informal finance is often of critical importance to farmers, such informal lending mechanisms make it much harder for banks and NBFIs to start lending to such clients as there is usually no available records of loan disbursements and repayments, which FIs require for conducting credit scoring. Also, many farmers may have never held any form of a financial product from a financial institution, such as a bank account or savings account, and such information is often an integral part of an automated credit scoring process. This lack of formal financial historic data for farmers (sometimes referred to as their financial footprints) prevent banks from generating credit scores for them, further complicating the ability of financial institutions to lend to farmers.
However Michael noted that new approaches are being tested and piloted that offer the potential to gather new, non-traditional, sources of data which can overcome the challenge of farmers having no existing formal financial footprint. Michael noted two clear opportunities: i) the gathering and digitization of data that has historically been unavailable to banks; ii) the use of alternative data that can act as a proxy for creditworthiness.
Digitization of non-traditional data:
This is often the first logical step in trying to create financial records for farmers. Farmers may often have been utilizing the services of input suppliers, and village savings and credit groups. If banks can secure these records, they can digitize them and use the data to credit score the client. This can be supported by other very relevant data such as livestock valuations and vaccination reports all critical elements in understanding the effectiveness of the farmer's operations and the collateral they may have to secure against any lending.
However, innovations in big data analytics offer the potential to unlock new financing opportunities and help FIs reach more customers than they would by relying on traditional reporting alone, by providing alternative sources of data that may be used by FIs to conduct credit assessments. Digitizing the available nontraditional data is a logical first step. In some cases, institutions are collecting and using data from input suppliers and dealers and using this knowledge for character scoring. When an individual has been part of a group lending program, this data can be captured and provides deriving information from the group for peer scoring. Valuation of livestock performance and vaccination reports, collateralization of receivables, as well as other inputs are fundamental. As more institutions begin working in this area and reporting results, it will help standardize the process.
Michael noted the need for stronger convergence of two sectors--agriculture finance and mobile banking. Mobile service providers are expanding into more rural areas of developing nations. In some cases, they are bridging gaps between rural farmers and banking products and services by providing a low-cost and efficient platform for conducting transactions. These services and apps provide data that can be used to support automated credit scoring.
Big data analytics has opened an exciting opportunity for banks. Utilizing mobile telephone and mobile money data, and processing it through big data analytics, banks can start reaching more clients by generating credit scores using this alternative data. This is already happening in urban areas where small short term loans are provided to customers who are scored based on their mobile telephone usage and stats. At present, some programs are working to expand such credit risk assessment approaches to facilitate lending to farmers.
Banks interested in exploring alternative data applications can simultaneously develop the skills necessary to collect and process this information, while preparing to introduce new services, such as digital banking or apps. A first step would be for banks to work towards better understanding their rural agricultural clients and to keep more accurate records. For example, types) of crops grown and livestock raised and their performance. As this way of scoring credit is still in its infancy, FIs have both the challenge and flexibility of determining best practices.
Michael cautioned that these approaches are still in their beginning stages and that it is premature to know exactly how robust the effectiveness of these alternative data sets will be, but early results are showing promise in terms of expanding financial services and increasing inclusion. Further examination of statistical methodologies is needed alongside a better understanding of how to protect consumers and their privacy. Besides, as such services grow additional focus on how best to regulate these activities will be required.
Do you have experience applying "big data" into credit scoring models? Do you want to share opinions or ask questions of your peers about this topic? Then please join our Linked In Group by following this link.