Juhudi Kilimo gives loan applicants psychometric tests to help make more inclusive credit decisions © Juhudi Kilimo
In an effort to build accurate risk profiles for smallholders that do not rely on conventional data, such as borrowing history, the developers of credit scoring systems are being forced to innovate.
To raise acceptance rates and cut processing times for farmer loans, Juhudi Kilimo, a provider of financial solutions to smallholders in Eastern Africa, is piloting technology from EFL Global, a private firm that uses psychometric testing to create risk profiles for borrowers across Africa, Asia, Europe and Latin America. The pilot – funded by the Mastercard Foundation – involves agents from six Juhudi branches in Kenya visiting loan applicants and inviting them to take tablet-based psychometric tests that EFL claims give lenders a picture of their character, including their self-control in relation to spending and approach to budgeting; based on this, EFL gives applicants a three-digit credit score. From its initial assessment of over 6,000 clients using the EFL tool, Juhudi found 6% of those in the lowest-scored quintile experienced 60-day arrears at least once in a typical 1-year loan, versus 1.5% in the highest-scored quintile.
Juhudi is now actively using the model in its credit decisions, accepting high-scoring clients that would previously have been rejected based on insufficient collateral, a lack of credit history or limited financial information. The financial provider is also allowing higher-scoring clients to access a loan valued at up to 100% of their collateral, versus 67% without the EFL data.
Nevertheless, while the EFL tool has predictive power, Juhudi’s partnership relationship manager, Elvin Walela, emphasises that the tool should still be used alongside traditional credit assessment methods. Visits by loan officers to applicants’ farms remain key to understanding a farmer’s likely income and creditworthiness, states Walela. In addition, some applicants have found the test difficult or complain that it takes too long and is intrusive – issues that Juhudi is now seeking solutions for with plans, for example, to trial an SMS-based test.
Banks driving demand
Following enquiries by banks who saw opportunities to expand their revenue by lending to farmers, but wanted to do it safely, the Thomson Reuters Foundation’s for-profit Bankable Farmer initiative was developed. The initiative aims to build credit risk profiles for African smallholders and charge lenders – its target clients – to access them; in June 2017, the first such client was secured – a Kenya-based commercial bank with a presence across Eastern Africa.
The Kenyan bank will help the foundation calibrate its credit-scoring model ensuring the selection and weighting of data meets lenders’ needs, says Saidah Nash Carter, Thomson Reuters’ head of innovation for Africa. The idea is to create a credit score so rich in data that banks are comfortable lending to farmers without collateral. The project is currently working with 60 data sets to build the scoring system but ‘the sky is the limit’ in terms of what could add value, says Nash Carter. For example, ‘smart’ technology to weather stations, water pumps and even tractors might produce useful data.
Faster lending decisions
The time-sensitivity of agriculture means the speed with which loans are disbursed is critical to farmers’ productivity. But, with traditional credit assessment methods, it can take 2-6 weeks for a Ugandan farmer to receive a credit decision, says Tamsin Scurfield, Opportunity International’s business development and partnerships manager for Uganda. However, use of an accurate credit score card could shorten this to seconds, explains Scurfield, while improving banks’ risk appetite as more informed lending decisions help defaults to fall.
Opportunity International’s credit scoring system, developed in partnership with local microfinance partner Opportunity Bank, will be launched for farmers in October 2017. The score card follows a ‘token’ model under which farmers earn points based on 17 equally weighted variables of which seven are related to farmers’ businesses, for example land size, crops grown and access to irrigation. The remainder focus on farmers as individuals, their households, the structure of the farmer groups they belong to, and their credit history where available.
So far, the initiative has profiled over 6,000 farmers, 580 of whom achieved the score card’s top ‘A’ rating. Farmers with very low scores will be linked to Opportunity International’s transformation centres to help improve business performance and hence, their creditworthiness. However, Scurfield acknowledges that complex variables outside farmers’ control, such as weather or pests, mean the score will not be 100% accurate. The score card is likely to be used alongside Opportunity Bank’s current credit decision-making processes, being tweaked over time as lending data reveals how well it predicts risk. In the future, Opportunity International aims to build a ‘statistical’ score card using client history data and algorithms to process data and create the credit score.
Integration of services
These innovative credit scoring schemes can help influence banks' lending decisions to take into account more systematically the myriad factors that can undermine a farmers' capacity and incentive to repay and so may be helping to overcome long-standing obstacles to farmer bankability. However to fully mitigate such risks, banks still need to rely not only on credit scoring, but also on creative structuring of financial products through integration of other services such as agri-insurance, extension, and offtake arrangements.