Using Mobile Data to Enable Financial Inclusion

Nineteen emerging experts from sub-Saharan Africa, working to advance financial inclusion at their national and regional levels, attended Rethinking Financial Inclusion (RFI) last month. An Executive Education course at the Harvard Kennedy School, RFI unites leaders in the space around data-driven and evidence-informed design of policies and programs to increase financial inclusion.

Through the lens of Smart Policy Design, the framework used by Harvard’s Evidence for Policy Design (EPoD) faculty to teach a problem-driven, collaborative approach to designing, testing, and refining solutions, they developed skills to systematically identify pressing problems in financial inclusion and use data to test innovative solutions. Chosen based on their experience and authority within their organizations and supported by the MasterCard Foundation, they arrived at Harvard with over 200 years of combined experience spanning the public and private sectors.

“Businesses try to target the poor, but we have a very superficial understanding of what motivates them,” said Dickson Osoo, Director of Projects and Program Management for Africa at MicroEnsure in Kenya. “RFI offers insight into those questions, particularly scarcity. That career-changing insight allows us to more effectively reach those customers.”

How can an organization do that effectively? Insights from an RFI model on alternative lending models, moderated by EPoD faculty co-director Rohini Pande with speakers Shivani Siroya, CEO of inVenture and Nicole Van Der Tuin, CEO and co-founder of First Access.

Imagine that two friends ask you to spot them $10. One of them is always punctual with her obligations; the other one still owes you $15 from two years ago. It’s easy to make a decision about who to lend to, right? But things get trickier when the person is a stranger. Faced with a request for capital from an unknown individual, most people wouldn’t feel comfortable assuming the risk.

In the world of financial services, unbanked users are a great unknown. Beyond the basics of identity verification, not having information about how they’ve behaved in the past makes it next to impossible to know how they’ll behave in future – and has historically reinforced their exclusion from the formal financial system.

History is changing, however: and the biggest driver of that change is often literally in our hands. Access to financial services hovers below 39% in much of the developing world, but 80% of the same population has access to a mobile phone. There are vast disparities by gender and geographic location, but those numbers inspire the question: How can we leverage the wealth of data produced by mobile phones to facilitate lending to the un- or underbanked, making it a winning proposition for banks and borrowers alike?

When Shivani Siroya was at the United Nations Population Fund (UNFPA) tracking how loans were used, her days were filled with time-intensive home visits, trips to borrowers’ businesses, and shadowing them at the market: all to try and understand how access to credit improved people’s lives. But the process was painstaking and difficult to scale. She realized that most of the data she spent a month gathering on a single person could be collected in seconds for thousands of users, using only their phones. Inventure, the company she went on to found, does just that – gets loan officers doing follow up out of the business of being “a walking, talking Quickbooks,” as she says with a laugh, and instead offers a direct customer relationship via a smartphone app.

For Nicole Van Der Tuin, CEO and co-founder of First Access, the spark came on the policy side. When it emerged that the Mumbai bombings of 2008 had been coordinated using unregistered prepaid phones, many developing countries moved aggressively to require personal identification for the purchase of SIM cards. This meant that individual users were being definitively tied to their data for the first time, creating a trail stretching over months and years that detailed exactly the sort of user history financial institutions need to analyze in order to determine risk.

What do texting patterns have to do with creditworthiness, you ask? A lot, according to our research, in both the developing and developed worlds. First Access in particular looks at four kinds of data: demographic, geographic, financial, and social. The first two help verify identity and provide context. Financial data provides a pattern of usage: and not only mobile money, as one might expect. Even regular top-up transactions provide incredible value, because they offer a window onto consumption patterns. And research has shown that borrowers with a broader social network are less likely to default, possibly because they have more support in place to help with repayment in the event of an emergency.

All of this requires buy-in from users, telecommunication companies, and financial services providers. But it’s a situation in which profits can rise for all involved. The borrower receives access to credit. In many cases, that investment in their business leads to increased mobile phone usage to deal with increased activities. And analyzing these borrowers relative to a microfinance institution’s existing portfolio can result in lower costs for the financial services institution. “Forty percent of applicants in Tanzania are so low-risk that they should be expedited for approval,” says Van Der Tuin. The cost of the decision-making process, however, is high enough to make many of these small loans needlessly unprofitable: unless you can find a way to remove the human Quickbooks from the equation.

Siroya is clear-eyed about the ultimate benefits of this approach: handing control over this data back to the customer. If they have access to their own information, they can take actions that will improve their credit score (either from private companies or the government bureaus), leading to fuller financial inclusion over the long term. “We have so much data,” she points out. “It’s our fault if we don’t understand the lives of our customers.”

So-called soft data can help create patterns that customers can use to legitimize themselves in the eyes of credit bureaus and larger financial institutions. But it can also have pernicious effects on vulnerable populations, particularly where digital or actual illiteracy are barriers. Course attendees asked about the regulatory structures in place for consumer protection. In India, the regulations imposed on organizations operating as formal lenders led Inventure to withdraw from the market. But in countries with interest rate caps, the pressure on financial services companies to keep loans affordable pushes them to find ways to streamline their client acquisition process – a need that First Access can help with.

RFI participants also pushed Siroya and Van Der Tuin to share their approaches to integrating the interests of multiple stakeholders in order to get these products off the ground: especially pertinent since the room contained many of the folks with whom their companies interact. An employee of SafariCom, for example, raised the issue of privacy concerns from telecommunications companies.

There was also the question of risk aversion from financial services providers, and users who might have legitimate concerns about how their data would be used. All of this had to be negotiated, both in real time as the companies were set up, and in the classroom during an engaged discussion on how to build something truly new in finance.

In the case of First Access, for example, we can see how varying interests were taken into account (and how it resulted in greater value for everyone concerned). Telecommmunications companies’ desire to safeguard their data was respected by operating on a server within their data center, such that no identifiable information leaves their facility. Users must opt into the system before their history can be reviewed in order to facilitate a decision. And analyzing financial service providers’ existing population in order to assess the risk of a prospective new borrower generates valuable, detailed information on the institution’s portfolio that can be used to improve service provision at lower cost.