How Data is Changing Credit
By Michael Schlein, President & CEO, Accion
Michael Schlein, President & CEO, Accion
Around the world, two billion people lack access to formal financial services. There’s no single cause for this widespread and significant problem, nor is there a single way to address it. But there are new trends in data, analytics, and mobile phones that we can use to remove some of the impediments preventing people from accessing and using formal financial services–particularly credit.
In many cases, the principal barrier preventing applicants from borrowing is a fundamental lack of information about the customers themselves. Lenders reject potentially creditworthy applicants because they don’t have the detailed financial profiles that banks rely on in underwriting. These “thin-file” applicants–including first-time borrowers, legal migrants, and young adults–might be excellent customers, but lenders typically pass on them because they’re essentially invisible.
That invisibility results from how banks typically do business. In the U.S. and many other parts of the world, banks generally rely on only one piece of information to make credit decisions: a credit report generated by credit rating agencies. These reports are built on the assumption that past credit behavior predicts future credit risk.
Information can be much more valuable and predictive than credit scores
This assumption is problematic for a number of reasons. First, to be able to build a credit score, one needs access to credit to begin with; and to access credit one needs a credit score. Second, credit agencies in different countries use different systems to evaluate credit behavior, so a profile created in one country might not be recognized in another. When migrants move legally to a new country, their credit histories no matter how good they are stay behind.
A migrant who has already paid off a mortgage or a car is still “credit invisible” when they reach their new home.
This is a problem even in countries with complex, integrated formal financial systems: in the U.S., one in 10 Americans have no credit history, and an additional 19 million have outdated, “unscorable” histories. There are millions more “thin-file” customers in the U.K., Spain, and Canada, who face similar problems in addition to the billions of other financially excluded people who lack access to any formal financial services and are invisible to the sector.
This exclusion prevents entrepreneurs from launching businesses or parents from sending their children to school. It also costs banks quality customers.
But data can change that. In the last few years, new solutions have emerged that leverage the explosion of personal data created by the proliferation of mobile phones to better understand customers and predict their credit behavior. Combining a wealth of alternative data–including social media, internet usage, phone records, and bill payments–with robust analytics can help lenders understand what the applicants themselves are like. That information can be much more valuable and predictive than credit scores.
There are a number of innovative startups tapping into alternative data and using it as the basis for new lending models. In fact, some of them result from the very problems caused by banking’s overreliance on credit scores. When Aneesh Varma moved from the U.S. to the U.K., he couldn’t get a credit card despite having a strong financial history. His experience motivated him to co-found Aire, an alternative credit scoring startup that uses applicant-provided information and new sources of data to verify customers’ identity, profession, education, lifestyle, and financial knowledge. Aire uses this information to help “thin-file” customers and creditors work together.
Other companies demonstrate a great deal of creativity using alternative data to predict user behavior. Tiaxa tracks users’ phone usage and evaluates them on 70 different variables a day.
It analyzes whether users regularly top-up their air time, read and write text messages, travel frequently or use more than one cell tower, among many other behaviors. When a user runs out of airtime, Tiaxa uses this information to decide instantaneously whether to extend them a small amount of credit, a “nano-loan”, to talk on the phone, text, or surf the web. Though these “nano-loans” don’t represent much money, they can be essential to the user in the moment and the beginnings of a personal credit profile. The company is now starting to expand their service, using client data to make bigger working capital loans.
There’s also Konfio, an online lending platform that uses innovative credit algorithms and alternative data analysis to help small businesses in Mexico that lack access to credit obtain affordable working capital loans; and Tienda Pago in Peru, which works with distributors of large consumer goods manufacturers to provide “mom and pop” store owners with short-term working capital loans that they need to buy more inventory and increase sales through mobile platforms, all while building a formal credit history.
Alternative data’s growth is due in large part to mobile technology: although 2 billion people lack access to financial services, 85 percent of the world’s population–or roughly 4 billion people–have mobile phones that are almost constantly transmitting, processing, and receiving data. In fact, as mobile phones and the internet continue to spread throughout the developing world, there will be increasing amounts of high-quality, user-generated data at companies’ disposal. In 2011, 2 out 3 people in the developing world had smartphones by 2020, 4 out of 5 will. At the end of 2015, 3.5 billion people were using the internet–and 2 billion of them were in the developing world.
The financial sector is only just beginning to discover alternative data’s potential. As alternative data becomes more available, and as its analysis becomes more accepted, I hope that lenders discover other uses that go beyond “de-risking” new applicants, such as helping customers manage their financial health or using behavioral science and well-timed “nudges” that prompt them to save. It will take thorough data analytics and a great deal of experimentation, but the value for businesses and their clients–is too great to overlook.