Alternative data ‘lends’ a hand

Alternative data ‘lends’ a hand

July 6, 2016    

Worried about how to improve your credit score? At least you have one. The harsh reality is that credit bureaus do not have data on the vast majority of the world’s population. According to the World Bank, credit bureaus cover less than a third (30%) of the adult population in the entire world, and in Sub Saharan Africa, they only cover 6% of the adult population.

What does this mean for credit lending? It means most adults in the world are excluded from access to formal credit. Why? There are a number of reasons that cause exclusion to credit, but most relate to the lack of credit bureau data used in traditional lending models. Traditional lending models rely on credit-bureau data for two reasons: to verify the identity of the applicant and to assess the risk associated with lending. If an applicant does not have a credit history, they cannot meet the minimum requirements of traditional lending models and are therefore excluded. This creates a catch-22 for ‘invisible’ customers (i.e. those without a credit history). To access credit they need to demonstrate they can manage credit. But in order to demonstrate they can manage credit, they need to have access to credit. Moral of the story, traditional lending models do not work for ‘invisible’ customers.

In the late 1970s, a new lending model gained momentum to overcome this challenge: microcredit. The microcredit model relied on community engagement to overcome the lack of information on potential new clients. There were two types of loans offered under this model: individual and group loans. Loan officers were responsible for knowing the community members and being able to assess the risk of lending to individuals and groups. In the case of individual loans, the loan officers were also responsible for collecting collateral. In the case of group loans, peer pressure was added to replace the need for collateral. While this model has shown promise for rural, low-income customers, it has also received a fair share of criticism and there seem to be more questions than answers on how much previously ‘invisible’ customers have benefited.

Asking new questions 

A new lending model is on the horizon. New data sources and analytical methodologies are allowing lenders to overcome their lack of information on ‘invisible’ customers and viably offer credit to more adults in this previously excluded segment of the market. Innovative credit lenders in this space are asking new questions to assess creditworthiness such as ‘What material is your roof made of thatched or metal?’, ‘What proportion of the contacts stored in your phone has both a first and last name?’ and ‘How long does it take you to fill out an application form?’ By analysing mobile phone data, social media data, psychometric data and geospatial data, companies like Lenndo, EFL, FirstAccess, Branch and Segovia are able to paint a holistic picture of the applicant. Combining these alternative data types with data analytics methodologies, ranging in complexity from regression analysis all the way to image recognition and machine learning, allows these companies to assess the applicant’s willingness and ability to repay a loan. Data and technology are enabling these players to confidently provide unsecured credit to a previously unserved market without ever requiring traditional indicators of creditworthiness.

While these models are still maturing in developed and emerging markets, some mobile money operators (MMOs) in the developing world are taking up similar alternative data-driven lending models. An interesting example of this approach has been rolled out by EcoCash in Zimbabwe. EcoCash offers microloans to their customers at no interest with a one-off 5% administration fee. Since the applicant is already a customer, EcoCash can use their SIM card registration to verify their identity. EcoCash uses several alternative data points in their model: transactional flows (value and volume) through the EcoCash wallet, transactional flows through the EcoCash Save Account as well as monthly expenditure on airtime and data. To qualify for the loan, consumers are required to save at least $5 USD a month for at least three months. EcoCash then matches the total amount saved and holds 25% of the funds in the savings account as collateral. The initial feedback from customers is positive and uptake is high.

Unintended consequences 

However, there are some potential downsides to these models. The ability to holistically understand customers and segment at an individual level through every text message, Facebook ‘Like’ and tweet introduces the potential of harmful selectivity. Providers can pick the most profitable customers only, or worse, unknowingly use data to discriminate against particular groups. Going forward it is important that we as an industry consider how to mitigate these unintended consequences to ensure we do not make things worse for customers. A critical lesson learned from the microfinance movement!

insight2impact is conducting a review of the use of alternative data and analytical methodologies in financial service provision. A Focus Note on emerging data analytics and data insights trends will be available mid-July.


This blog first appeared on the insight2impact website as part of their blog series on the use of alternative data in financial inclusion.

insight2impact (i2ifacility) was funded by Bill and Melinda Gates Foundation in partnership with Mastercard Foundation. The programme was established and driven by Cenfri and Finmark Trust.

Similar Articles
Risks, harms and opportunities in data-driven technology for financial inclusion
How can data and data-driven technologies impact financial inclusion efforts in Africa? ...
Building foundations for new behavioural scientists in Africa
Behavioural science studies how people behave and tries to answer why they behave in the way they do. This is do...
Financial inclusion innovation competition: DataHack4FI
DataHack4FI Season 3 report The DataHack for Financial I...