CBK directive to create new credit-scoring platforms

The decision by the Central Bank of Kenya (CBK) to withdraw the approvals it granted to unregulated digital (mobile-based) and credit-only lenders as third party credit information providers to credit reference bureaus (CRBs), leaves a significant number of digital lending platforms in limbo in as far as credit scoring is concerned.

The Credit Reference Regulations of 2013 allowed CRBs to receive information from third-party sources, provided the CBK issues no-objection.

But the directive by the Central Bank now means CRBs will have to mask (or undisplay) prior information received from these parties and subsequently stop receiving new information. This creates new kinds of problems for digital lenders.

First, it creates a huge number of “new borrowers” (or otherwise referred to as thin files in credit scoring parlance). When a previous borrower’s data is deleted, it converts the borrower into a new one (for which there is no reference). These so-called thin files often present a big challenge to CRBs since there is no way to assign them any score in the absence of borrowing history. In short, CRBs are not able to report new users who have never taken a loan before.

Second, for as long as the CBK directive stays in force, digital lenders will now have to invest in internal credit scoring machines, which maybe costly for a significant number of them. The more established digital names, and there are two out of the over 50 players in the local digital lending space, have CRBs only as secondary validation points since they already make use of their internal scoring machines for primary reference.

However, the less established ones, and there are many of them, do not have internal scoring machines and have had to rely on CRBs for cross-referencing. However, I need to point out that they will still be able to ping the CRBs, subject to the usual condition of customer consent, whether electronically or written, but the data will be somehow not be up-to-date and not useful in effective credit deaccessioning.

Consequently, this directive has created some sort of referencing gap for a number of digital lenders. But it also creates an opportunity for alternative scoring credit scoring platforms. By connecting through application programing interface (APIs), alternative scorers can act as the primary reference point for the unregulated credit-only lenders, whether digital or otherwise. And they probably are much more sophisticated. They employ machine learning algorithms (Artificial Intelligence) and statistical modelling on the data available on customers’ cell phones, to predict their credit worthiness.

They also perform better when it comes to fraud mitigation, such as identity theft and data integrity issues that are prevalent in the digital credit space. As users generate real-time data, such as behaviour on such online platforms as social media as well as phone usage, the alternative platforms are equally able to generate real-time credit scores. Further, their powerful tools give them the capability to even score thin files; that is, one does not have to have taken a loan elsewhere or held a bank account for their credit score to be generated.

I have already seen a few platforms actively jostling for visibility on this front, including a recent entrant Weza.io. But most importantly, this need for alternative scoring also calls for CRBs to move away from the traditional scoring models.

@GeorgeBodo