A Blog by Jonathan Low

 

Dec 2, 2015

Evaluating Phone Usage to Assess Credit Worthiness

Data analysis suggests that how a phone user enters contacts (last name first indicates a better credit risk), whether they receive more texts than they send (another positive variable) or what time of day calls are made (more evening calls indicates price awareness) all correlate with whether a potential borrower is a good credit risk likely to repay a loan.

The concern is to what degree these behaviors remain indicative or whether financial institutions and those who provide data service to them dont want to spend the money to update their data, thus potentially penalizing customers - and the economy. JL

Elizabeth Dwoskin reports in the Wall Street Journal:

Smartphones apps generate data—texts, emails, GPS coordinates, social-media posts, retail receipts, and so on—indicating subtle patterns of behavior that correlate with repayment or default: even variables such as how frequently a user recharges the phone’s battery, how many incoming text messages they receive, how many miles they travel or how they enter contacts into their phone
A handful of Silicon Valley-backed startups are looking to revolutionize lending in the developing world, where banks are scarce and many would-be borrowers have no credit history.
Their strategy: Show me your smartphone, and my app will find out how creditworthy you are.
Smartphones can dramatically reduce the cost of lending, experts say, because the apps they run generate huge amounts of data—texts, emails, GPS coordinates, social-media posts, retail receipts, and so on—indicating thousands of subtle patterns of behavior that correlate with repayment or default.
Even obscure variables such as how frequently a user recharges the phone’s battery, how many incoming text messages they receive, how many miles they travel in a given day or how they enter contacts into their phone—the decision to add last name correlates with creditworthiness—can bear on a decision to extend credit.
In Kenya, Branch.co offers an Android app that lets users apply for a loan and get immediate approval and access to funds. The loans average $30, enough for a taxi driver to pay for gas or a fruit seller to stock up on produce. Branch charges between 6% and 12% interest—based on the borrower’s creditworthiness—and loans are usually repaid between three weeks and six months later.
Traditional microlending tends to be far more expensive—interest rates often exceed 25%—partly because lenders must visit borrowers in the field to assess their ability to repay. Banks have steered clear due to the high cost of building physical branches.
These app-based lending startups are backed by some of Silicon Valley’s biggest names. Branch, which was founded by microlending pioneer Matt Flannery , has received funding from Joe Lonsdale, co-founder of data miner Palantir Technologies. InVenture, based in Los Angeles, is headed by a former United Nations officer and funded by venture investors Chris Sacca and Zachary Bogue. Saida is funded by startup incubator Y Combinator. Omidyar Network—an investment firm and foundation established by eBay Inc. EBAY 0.28 % founder Pierre Omidyar—holds a stake in Lenddo, a lender that determines creditworthiness by analyzing social networks like Facebook. FB -0.30 %
By installing these apps on their smartphones, users grant them access to any information that may help assess the borrower’s creditworthiness—from the content of their texts and emails to the duration and volume of their calls.
InVenture’s algorithms, for instance, found that users who wait until after 10 p.m. to make calls—when rates are lower—are lower-risk borrowers. Somewhat counterintuitively, Branch found that users who are known gamblers—something the app would find out by scanning messages or payments to a gambling company—are more likely to repay a loan than nongamblers.
“You’re able to get in and really understand the daily life of these customers,” said InVenture CEO Shivani Siroya. Her company’s scoring formula, or algorithm, analyzes 10,000 so-called signals per customer.
These lending startups build on the popularity of mobile banking in many developing countries and the rapid rise of smartphone use. A Pew Research Center report from April shows that 34% of South Africans, 27% of Nigerians and 15% of Kenyans already own a smartphone.Customers of Branch and InVenture in Nairobi, Kenya, said they used the loans to pay for running or improving small businesses. Some had access to banks but felt the smartphone interest rates were better; others had been borrowing informally from neighbors at high interest rates.
The owner of a beauty and weight management center said small loans covered items such as skin cleansers when her bank account ran low.
Samuel Njuguna, a chef, said he bought plates, cutlery, and pots. “I’ve had to turn down a few business opportunities because of lack of equipment,” Mr. Njuguna said. Now, he says he is plowing most of the money back into his business.
“These are people that don’t have a credit score,” said Branch’s Mr. Flannery, whose earlier venture, Kiva.org, helped expand microlending. “Your digital trail can establish your financial track record.”
Lending startups like Branch could bring formal credit for the first time to between 325 million and 580 million people in emerging economies, according to a recent report by Omidyar Network.
While the smartphone lenders focus on emerging markets, their efforts to assess risk based on nontraditional data sources is part of a wider trend in Silicon Valley. Affirm, LendUp, ZestFinance and others use data from sources such as social media, online behavior and data brokers to determine the creditworthiness of tens of thousands of U.S. consumers who don’t have access to loans.
And competitors with deeper pockets are entering the field. Visa Inc. V 0.10 % has built mobile payment applications in Rwanda and is working with International Business Machines Corp. IBM -0.30 % to use records of retail transactions or remittances to create a surrogate credit score. Chinese e-commerce giant Alibaba Group Holding Ltd. BABA 1.27 % recently launched a credit-scoring program that uses the company’s own trove of transaction data to assess risk.
Privacy advocates have complained that borrowers might be denied a loan because of a Twitter TWTR 0.16 % post such as “my car has broken down.” U.S. companies have wide discretion to offer loans as long as they don’t sell credit scores or discriminate against minorities, women, or people with disabilities.
The Omidyar Network surveyed dozens of individuals in developing countries about the privacy trade-offs, and most said they had no problem sharing personal details in exchange for much-needed funds.
As a former official at the U.N. Population Fund, Ms. Siroya—InVenture’s CEO—has conducted more than 3,000 in-depth interviews with small businesses in developing countries. She said borrowers in these countries are far less risky than mainstream financial institutions think they are.
Soon, she said, she will have the data to prove it.
Customers of Branch and InVenture in Nairobi, Kenya, said they used the loans to pay for running or improving small businesses. Some had access to banks but felt the smartphone interest rates were better; others had been borrowing informally from neighbors at high interest rates.

0 comments:

Post a Comment