In most developing markets, there are few ways to asses a person’s ability to pay future debts, or calculate what most would call ‘credit worthiness’. A typical credit history is compiled from various financial sources – beginning with a bank account, and typically extending to credit cards, loans, etc. But 48% of the world’s adults are considered ‘un-banked’; those without a bank account.
Businesses often feel it is risky to extend credit to the unbanked, but that’s over 2 billion people who are being underserved – almost half of the world’s population. This group of people may not have a bank account, but they still have incomes and spend money. In fact, according to the World Bank, the unbanked population is running 200 million small businesses and has an overall buying power worth USD 5 trillion. That’s too much money for businesses to ignore.
Communication service providers (CSPs) rely upon credit information for supporting their postpaid customers, but in many developing countries they often have little to none of the data they might traditionally use to make sound credit scoring and credit control decisions (for example, official proof of income and a credit history). One of the prime barriers is that financial institutions typically reside in cities that are closer to a higher concentration of people and capital found in urban centers. These institutions have shied away from engaging rural populations because of high transaction costs due to poor infrastructure, and a remote, widely-dispersed client base. This creates a dearth of financial as well as vital records, creating a significant impediment to assessing a person’s credit risk. In regions like Africa, which is home to the world’s fastest growing middle class, many products and services remain out of reach.
But there is an opportunity for lenders to chart another path. Instead of utilizing banking records to determine creditworthiness, businesses can now benefit from increased computing power and new sources of information and data, such as mobile-phone usage patterns, demographic data from social network profiles, geolocation data, social media relationships and others, to build better risk models. With these assets, and with scrupulous attention paid to privacy laws and customer consent and preferences, CSPs can make responsible credit control decisions in low-touch and low-cost ways.
New risk management approaches based on social media information can provide the data that service providers need to create credit profiles for subscribers with little or no formal credit history. CSPs have ignored this group of customers for years due to their lack of solid credit ratings, not wanting to extend access to products and services that may never be repaid. Excluded from the mainstream, these customers are left with few options. This exponentially expands a service provider’s customer base, without significant additional risk.
New data, new insights
The problem with traditional credit scoring tools is that they are overly dependent upon past financial data as a guide to the future, but for a huge percentage of subscribers this data simply doesn’t exist. We believe there is more to the picture – valuable pieces of information that are being overlooked by the business community. For many, an enhanced profile could be achieved through other online sources. With social scoring tools, service providers can take advantage of the explosion in data being generated from social media and other online sources.
What can social relationships and connections prove in terms of creditworthiness? Depending on how frequently a person posts online, social media can give us powerful insight into how and where a subscriber spends a good portion of their time or income. For instance, a new subscriber that lives in São Paulo but spends a lot of time vacationing in Florida may be a low credit risk, because his location hints that he has the means to spend money traveling abroad. Another social scoring capability is the ability to automatically correlate and analyze comments a user posts to Facebook or Twitter. A person’s friends and associates can also provide clues for creditworthiness and be a good indicator for predicting potential credit problems in the future.
I believe that applying algorithms on top of an applicant’s behavior-based data provides a more complete picture than just marking off a checklist of credit scoring requirements. After all, spending patterns that might make sense for a “soccer mom” in the United States of America might be deemed suspicious by an unemployed housekeeper in Nigeria. When looking for correlations in a wealth of data there may not be a single “right answer”.
This article was originally published on the WeDo Technologies Blog. It has been reproduced with their permission.