Credit Risk Modelling – Catering to the Unbanked Section

Credit Risk Modelling for the Unbanked
Today, it is estimated that around 2.5 billion people don’t have access to formal financial services while several million micro, small and medium enterprises cannot fulfill their financial needs. This can majorly be attributed to the traditional credit risk models which base their lendings heavily on parameters such as a credit report, salary history, documents to comply with identity verification and assets for collaterals etc, which are not always available with the lower income groups. Thus, Governments and businesses worldwide are devising different ways to satisfy the requirements of lower income households by making quality, affordable financial services approachable to them. One of the most optimum ways to bridge this information gap is by using Big Data. Big data can help companies profitably serve vast unbanked populations by fueling credit risk models, and can also help societies move towards full financial inclusion. New avenues of data and information draw us closer to this vision by enabling a more complete understanding of households’ financial needs. Here is an article by CRIF – a consumer, commercial and MFI credit bureau on the effective usage of alternate data for better credit risk modelling.

Existing business models for lower income households

Lack of documented credit information and inconsistent payment patterns of lower income households has always proven to be a challenge for lenders to perform reliable credit risk assessment and extend loans to them. Deprived of sufficient knowledge of formal loans, this section of the society often tend to lose out on the better financial schemes. With such a backdrop, lenders have adopted two credit risk models for lending a loan:

Tradtional Finance:

This is a simplistic model where loans are extended at a higher interest rate, heavier penalties and additional fees in order to compensate for the frequent losses incurred by defaulters.


These finances typically cover the microenterprises, women owned businesses and joint liability groups who are willing to apply for a short term loan and buy a new loan after each short tenure. This way, the default rate is very less and the risk is low too.

Alternative sources of information

The above models, although logical, are not very cost effective. They fail to cover the diverse needs of the economically active families on a continuous basis. Hence another path for lenders to perform Customer Due Diligence beyond performing a credit check is using modern data sources such as telephone company details, Utility bills, Retailers and Direct sales companies. Lenders should exercise data mining by identifying trusted data sources, securing their access to uninterrupted data and collect and convert these into critical customer insights. However, collecting data from each source for lenders can be hugely resource consuming. Also, considering the various regulatory policies in place that restrict the access and trade of data, a better solution is to strike partnerships with organisations experts in automated credit risk management.

Technology has allowed us to collect, aggregate, and analyze data in ways never before possible. Using mobile operator data (with customer consent) we can derive details like times of calls, information shared on SMS, location insights and Payments made using the phone. With retail data, we can get insights into the spending patterns and with utility bills, we can see if the user has a record of timely bill payments. There are tools developed specifically to bring together disparate data sets and analyse them to create new customer insights. These automated tools overcome the need of requiring a business expertise to understand optimal details of each sector and allocate weightage to data accumulated from diverse sources such as telecom, retail etc. Lenders need to invest in heavy duty data processing softwares for credit risk modelling using non-conventional sources of information. Lenders can effectively employ big data to create meaningful value for their organisation, more financial inclusion and a significant social contribution of sorts!

Twitter Facebook Instagram Linkdin