How can you trust if someone will repay a loan? Look at their smartphone.

May 17, 2018 / Patrick D’Arcy
By replacing traditional credit scores with mobile phone data, entrepreneur Shivani Siroya has pioneered an innovative way to think about trust.
What makes you decide to trust someone? Maybe you have mutual friends or you know their family. Maybe you know other people who went to the same college or work at the same company. Or perhaps your trust is based on something less concrete, such as their friendliness, openness or a sense of stability they convey.
What if you can’t meet them in person because they live far away, and they’ve asked you for a loan? In the developed world, we use credit scores — calculated by analyzing a person’s banking, credit card and loan histories — to make a decision. But what if they don’t have that kind of history or even an address?
Entrepreneur Shivani Siroya, a TED Fellow, has pioneered a different way of figuring out whom you can trust (TED Talk: A smart loan for people with no credit history yet). Her Los Angeles-based company Tala collects over 10,000 data points from a single smartphone and uses this information to give people in the developing world a financial identity. Tala uses this identity, which is a numerical score, to make microloans between $10 and $500 to customers who own small businesses, such as food stalls and retail kiosks. Since 2014, Tala has provided more than $326 million in loans and boasts a remarkable repayment rate of over 90 percent. Because 69 percent of adults worldwide lack a credit score, this new approach could have enormous possibilities.
Siroya first started thinking about new ways to evaluate trust while she was working at the UN. Between 2006 and 2008, she interviewed more than 3,500 recipients of UN Population Fund microloans in Africa to determine the impact of this lending. As she says, “I started to realize all these people were getting credit, but we didn’t know how they were using the money, and we didn’t understand what their capacity for repayment was.”
One day when a friend visited Siroya, he asked her how she knew who could be trusted with a loan. “I was like, ‘Well, this woman is planning a wedding for three other kids in the community and she’s saving 30 percent of her monthly income for her son to take this computer class,’” she remembers. “And my friend was like, ‘That’s interesting. So you don’t have traditional data, but in a sense you’re actually using her daily life.’”
Something clicked for Siroya. She realized that if she could learn enough about a person’s daily life, she could determine how trustworthy they were and, by extension, how likely they were to repay a loan. But where could she find the data for people she didn’t know? She had a hunch that smartphones could provide a window into their everyday lives. To test the theory, in 2014 she raised an initial round of capital and made small loans to 10,000 individuals in Kenya; she and her team used this group to help figure out what mobile data points could serve as indicators of a person’s creditworthiness.
As it turns out, how someone uses their phone provides many insights into their capacity to repay a loan. For instance, the time of day someone makes most of their calls can be significant. Since nighttime calls cost more than daytime calls, night calls suggest they’re better off — and a better credit risk. Siroya and her team also found if an individual communicates with 58 contacts or more and makes phone calls that last for more than four minutes, these can indicate strong social networks, meaning they’re more likely to repay a loan. If more than 40 percent of a person’s contacts are listed by both first and last name in their phone, this attention to detail means they’re 16 times more likely to repay a loan than someone who doesn’t do this. Tala’s software crunches together these and thousands of other variables to establish your financial identity. “There’s not one feature that can be taken into isolation that can determine your credit score or your creditworthiness,” Siroya cautions.
After someone downloads the Tala app and requests a loan, it asks for permission to view key pieces of data on their smartphone. This data includes texts and calls, merchant transactions, app usage and personal identifiers. After assessing all this information — which takes just minutes — Tala decides whether a customer is creditworthy and, if so, immediately sends them money via a mobile wallet. The company doesn’t use gender, race, religion, ethnicity or national origin as factors in lending. In fact, its data team recently discovered a correlation between the number of vowels in a customer’s name and their ability to repay. On closer inspection, it realized this could lead to discrimination against certain tribes or groups, so the variable was deleted from its algorithm.
The more trusted people feel, she believes, the more trustworthy they will be. “If you’re someone who is not trusted, think of what kind of turmoil that does for you,” she says. “When someone who’s financially anxious has to interact with a financial system, they’re going to make incredibly reactive decisions. That doesn’t just translate into how they repay their loans but every other decision in their life — like how they take care of their families and how they run their business. They’re reacting from a place of insecurity.” As a result, Siroya has tried to design the app to make its users feel comfortable, empowered and trusted. “If you look at the credit card industry, they make money by people defaulting, or by people being late with their payments,” she says. “We think about how to flip that and incentivize positive behavior.”
Tala defines its relationship with customers to be one of “radical trust”: a mutually beneficial partnership in which the company takes the first risk — lending money — in order to put more power in the customer’s hands. “If you think about it, the entire global financial system is built on skepticism, not trust,” Siroya says. “We try to imagine what it would look like to start with believing in people — an assumption of trust rather than risk.”
Today, Tala is used by more than 1.3 million borrowers in Kenya, the Philippines, Tanzania, India and Mexico. It’s the second most popular app in the Philippines, fifth in Kenya, and sixth in Tanzania. Last month, Tala raised $65 million to fund its global expansion.
“We had a Kenyan customer recently say, ‘This is the Kenya we want,’” Siroya says. “He was talking about the idea behind our product and how that’s the kind of country he wants to live in — a place where people are trusted. It’s the kind of world that we all want. To be trusted can enable you to take chances and really improve your life.”


Leave a Reply