Many fintech innovators are utilizing new data and innovating new types of analysis to extend the frontier of credit to include excluded and underserved microenterprises, particularly those owned by women. These models provide credit to those who would otherwise be excluded or underserved by creating credit scores that leverage new sources of data in a more considered way. Many also leverage artificial intelligence (AI) and machine learning (ML) to allow MSEs to continuously improve their scores and increase the amount of credit they can access over time.
Conventional credit-scoring approaches disadvantage low-income people and informal businesses because they don’t produce financial data like transaction records, tax receipts, and bank statements. Without such inputs, traditional credit-scoring algorithms lack reliable and consistent digital data sets on which to make decisions, so marginalized users lack credit scores or have “thin files” (indicating that they have limited credit history), preventing them from integrating into the formal financial system.
The inclusive credit-scoring models featured here use a mix of traditional financial transaction data in conjunction with alternative data like e-commerce transactions, personal purchases, mobile wallet data, geo-location, bill payment histories, and social media usage. Such data trails serve as a replacement for traditional repayment data used to assess a person’s creditworthiness. CGAP’s research into transactional data for credit scoring found that it can accurately predict repayment behavior, much like a credit history, and that such approaches can power inclusion. The scoring methodologies for alternative data vary in complexity, ranging from basic scorecard models to complex AI-powered models. In many cases, providers incorporate AI and ML after several loan cycles while in others, AI and ML are used from the outset. Fintechs have developed this approach to data alongside a number of other strategies, which together extend the frontier of credit access. These strategies allow fintech startups to extend access to finance to excluded populations without credit histories while still managing risk. They include measures such as alternative data, cluster-based financing, and other approaches that are explored in two additional Focus Notes: “No Small Business” and “Empowering Small Giants”.
While alternative data can help overcome weak financial information (at least for small-ticket loans), it still depends on the existence of a broad and inclusive digital payments environment. In countries where digital identity is well established and where digital payments are ubiquitous, there is more plentiful alternative data to serve as inputs for inclusive creditscoring. Individuals in urban areas, for example, that make greater use of such digital services create richer data trails and can be assessed by alternative scoring mechanisms more thoroughly. Similarly, MSEs that order digitally and accept digital payments are more likely to be eligible for digital credit products. While open data regimes are relatively new, they also promise to increase the quantity and types of alternative data and can potentially drive further inclusion.
Tailoring pricing and loan terms to match enterprise cash flows and decrease risk of over-indebtedness
Fintech startups are tailoring loan size, tenure, and repayment frequency to match MSE cash flows and inventory turnover, thereby decreasing risk of over-indebtedness. They are moving away from long tenure loans, vanilla interest rates, and steep late fees to create loan structures that better match the rhythms of MSE revenues, costs, credit needs, and repayment preferences. These innovations decrease risk of over-indebtedness for entrepreneurs since they time loan disbursement with MSEs needs for liquidity, and repayments with revenue flows.
Fintech startups can afford to tailor loan terms thanks to the savings provided by their technology. Traditional financial service providers have to offer “one size fits all” loans because that simplicity offsets the high costs of acquiring, assessing, and servicing loans manually. By contrast, digital data and tech-enabled operations bring down costs of onboarding, scoring, and extending credit. Moreover, tech brings down the cost of tailoring by digitally segmenting clients, automatically adjusting loan terms, and more.
Many loan tailoring examples are found in FMCG (fast moving consumer goods) value chains where merchants gain access to digital, rotating working capital advances in the form of inventory credit or BNPL services. These products are not usually referred to as loans and can sometimes feature short interestfree repayment periods or bullet repayments, offered in-kind (not cash) as part of an FMCG product offering. They allow MSEs to take on additional inventory without the upfront cost. offered in-kind (not cash) as part of an FMCG product offering. They allow MSEs to take on additional inventory without the upfront cost.
Other examples are connected to e-commerce or digital payment models in which repayments are linked to earnings, allowing for variable structures. For example, by deducting repayments as a percentage of a merchant’s daily sales, so that repayments are seamless and painless.
Excerpt of: Inclusive Fintech: How Startups are Bridging the Finance Gap for Microenterprises (CGAP, 2024)