Digital footprints and financial inclusion

Can digital footprints lead to greater financial inclusion? This was the question that the researchers at the Consultative Group to Assist the Poor, a reputable research and policy center at the World Bank, set out to explore. After all, there are over two billion people in the world who have mobile phones but no bank accounts, with majority of them in low income segments. The latest CGAP report points out that Data Analytics - a lucrative business for consulting firms - can be put to good use towards financial inclusion. Little information is available on the unbanked poor, but the digital footprint they leave through their mobile phone usage is big data. Many meaningful and interesting insights are revealed through data mining and the mobile network operators are already sold on this. "As long as consumer interests are protected and privacy, security, and ethical use concerns are addressed, these data may become a useful way to reach unbanked poor people with a range of financial products", noted the researchers. The digital footprint is basically user details like voice calls and SMS pattern, value-added services, mobile internet usage and mobile money transactions. What this means is that mobile phone usage can provide a peek into lifestyle and behavior. Simple things like duration of calls, location, time of the day, topping up the accounts, can provide insights into as complex a phenomenon as debt repayment ability. This is huge, because even people with jobs and bank accounts often don have any credit history, but their mobile phone usage can come in handy. CGAP partnered with an MNO to test some hypotheses into how digital footprints could be used to deliver credit to the unbanked. One of the hypotheses was that frequent and consistent buying of cellular airtime demonstrated predictability in loan repaying capabilities. The other was that poor planning is manifested in an inactive prepaid account or one that consistently runs to zero balance before next airtime purchase. The researchers were unable to test their hypotheses and instead referred to the success of Cignifi. A Cambridge-based start-up, Cignifi built a credit scoring model in Brazil, using 50 variables from 2.3 million prepaid customers of an MNO and then back-testing this model against historical lending data from approximately 40,000 borrowers of that MNOs lending business. The Cignifi model proved to be an accurate predictor of default, as the scores positively correlated with default across the lending portfolio. Such credit scoring analytics are a long way off from replacing or even supplementing formal credit score reports that are based on past credit history. But, if it flies, imagine the impact on the socio-economic uplift of the poor who would have a decent shot at accessing finance. Since the mobile money market is in search of scale, the researchers recommend a "freemium" business model: offer basic transactions for free, build sufficient volumes of transactional data, and conduct analysis on that valuable data. For that to happen, more experimentation and more research is needed. More investment is required in the in-house analytics departments or partnering with analytics firms.