The growth of mobile loans in the local market has seen tremendous leaps as new institutions dive into such businesses to reap the benefits of the extremely lucrative sector. In 2018 alone, Kenya has seen the launch of two high-profile microfinance loan apps, and this can only grow as banking institutions realize the potential attached to such services.
The dynamics of such apps, among e-commerce websites that offer the same services, are based on their ability to leverage users’ digital footprints to decide credit scores and creditworthiness. Any person connected to the internet leaves some form of personal information on the websites they visit, including emails and phone addresses, to mention a few – and lenders pick the trail from there.
It also appears that some lenders are determining if you qualify for a loan based on the type of a phone in your possession and the platform its running on. To put it differently, Android and iOS, in conjunction with other data forms, are being used to predict if you will default loans and predict consumer behaviour as traditional systems do. This is a channel that is being developed to include machine learning and AI technologies to replace traditional credit bureau approaches that are still in place – but have serious privacy concerns as once predicted by David Gerrold, where he argued that ‘having all that connectivity is going to destroy what’s left of everyone’s privacy.’
The above highlights are echoed by a research done by The National Bureau of Economic Research that analyzed thousands of purchases done on a Germany-based e-commerce platform that operate in the same manner as Kilimall and Jumia – but applies digital footprints to offer loan services to customers.
The research looked into a number of parameters that include data trails users leave on sites albeit passively such as the device used for purchases, how they arrived at the site (could be an email referral, ad clicks and so forth), time of purchase and the kind of email address used.
The analysis illustrated the importance of insights gained from a loan applicant’s phone platform (Android or iOS) as an average iOS device is priced significantly higher than its corresponding Android gadget. As such, the difference in default rates between iOS and Android users is equivalent to the difference between median credit scores and the 80th percentile of the same score.
Basically, iPhone users have a better credit score than 80% of the population whereas Android users are right there in the middle (median). Since a higher credit score means you are more likely to repay, then iPhone owners are a better bet to give a loan to. Your guess as to how such information is used is as good as mine.
Additional inferences from the study were as follows:
Customers who purchase stuff from an e-commerce site using a mobile phone are more likely to default a loan than those who place orders on PC.
Those who use outdated email clients, such as Yahoo and Hotmail are also likely to default.
Here is a funny one: if you input your email incorrectly during purchases, you are a potential defaulter.
If your email has your first or last name, like email@example.com, then you are 30 percent less likely to default than something like ‘princessminaj98.’
How you arrive at e-commerce platform is an important parameter too; if you are scouting for price comparisons from other sites, then you are more likely to default than a person who clicks on a targeted ad. Interestingly, this makes sense because people who make informed purchases go through a bunch of websites to ensure they aren’t being ripped off.
However, it is still unclear whether controversial metrics come into play when lenders determine the creditworthiness of an applicant. For instance, it is illegal to use race to bar a person from accessing credit services. But the research reveals that some people are wrongly disadvantaged from digital footprints. Such groups include those labeled as ‘risky’ even when they are not.
Also, many of you can afford iPhones, but choose to use an Android phone for personal reasons.
The study is summarized by admitting that there’s a correlation between credit scores and digital footprints, but not sufficient to be used singularly. For instance, some people qualify for loans, yet their online lives dictate otherwise. Thus, lenders need to combine traditional procedures and digital offerings to build accurate scores.