Editor’s Note: Grameen Foundation is a global nonprofit organization,that helps the world’s poorest people—especially women—lift themselves out of poverty by providing appropriate financial services, life-changing information and unique business opportunities. With this in mind, Grameen Foundation collaborates with Qualcomm Wireless Reach on projects that aim to bring entrepreneurs out of poverty by harnessing the economic benefits of wireless connectivity.
After the Miami Heat won the 2012 NBA Championship, a Twitter exchange erupted between Mark Cuban, the outspoken owner of the Dallas Mavericks; and Skip Bayless, sports journalist, TV personality and ESPN commentator. This led to a heated episode of ESPN’s First Take that went viral. Cuban contested that Bayless and other sportswriters only spoke in generalities. Whether Bayless was speaking of Lebron James’ “biggest collapse of a superstar that we’ve ever witnessed” or praising the Miami Heat by saying “Miami wanted it more than Oklahoma,” Bayless’ comments, in Cuban’s view, were too vague for anyone to question. Unfortunately, the issue of using vague generalities to describe a situation reaches far beyond the basketball court.
Melinda Gates’s describes the way we, in the nonprofit sector, usually evaluate the success of our programs by analyzing data at the end of the project, if at all, instead of using real-time data throughout implementation—a practice she likens to “bowling in the dark.” Like Skip Bayless, we can make all sorts of subjective conclusions around the efficacy of our work if we are not expected to offer factual evidence to support our claims. However, unlike sportscasting, the consequences of not using data in our work can be considerably more harmful.
While many private sector companies have adopted quantitative, even experimental approaches to designing products and services, many development organizations continue to roll out programs based on “commonly held notions” about the needs and habits of the poor. Many development organizations continue to view the poor as a homogenous group, with uniform and static needs, and paternalistic notions of what the poor need tend to take the place of information on what most poor households actually use. The challenge before us is to develop industry practices around program design and management that rely more on empirical data than conventional wisdom.
In their book Poor Economics, Abhijit Banerjee and Esther Duflo observed that food subsidies are ubiquitous in the developing world because “government efforts to help the poor are largely based on the idea that the poor desperately need food.” But looking at data coming from India, they saw that despite “the country's rapid economic growth, per capita calorie consumption has declined.” Despite government subsidy, the poor are not eating more or better because of a confluence of poor information, the decline of heavy physical labor and a sense of skepticism around the prospects of radically improving their lives. Against such findings, they saw the raging debate about the existence of poverty traps and the value of foreign aid to be woefully disconnected from the way “poor people really live their lives.” Thus, they argued that policymakers need to “completely re-imagine the way they think about hunger,” and that governments need to “stop pouring money into failed programs and focus instead on finding new ways to truly improve the lives of the world's poorest.”
We found a similarly counterintuitive, albeit less severe, insight in our own work at Grameen Foundation of helping microfinance institutions adopt the mobile platform to disburse loans, collect payments and lower transactions costs. Our team in Asia recently completed a six-month pilot and found that borrowers, who regularly sent and received text messages—a proxy for their level of comfort with mobile technology—made significantly moretransaction errors than clients who did not use text messages to communicate with family and friends. The team explained that clients, who thought they knew what they were doing, were less likely to pay attention to instructions, and therefore, more likely to commit mistakes. This small, but unexpected finding is changing the way we are developing new product features, such as automated payments and assisted transactions, as we roll the program out to more branches.
Clearly, the task of using data to show social and economic change is much more complex than analyzing a basketball game, but if we want to reduce poverty, we have to similarly break down the tape, crunch the numbers, and make the adjustments.