As a Westerner, getting a credit card is only slightly more complicated than tying my shoes. My world is raining with opportunities to borrow money to go to school, open a store, consolidate loans, or buy an iPhone 6. For poor global citizens, though, the game is different. In fact, there isn’t a game. A small business loan in developing countries usually requires a bank account (which 80% of sub-saharan Africans do not have), carries exorbitant interest rates (50%+), and are otherwise unreachable or unaffordable to the majority. The reality is that it’s simply not profitable for lending agencies to accommodate these high-risk, small-scale borrowers.
The irony is that these are precisely the people whose lives can be most dramatically improved by a small amount of liquidity. Their hardship is enormous, and emerging from it is often complex or impossible due to systematic constraints (poverty traps). Enter the world of microfinance, where nonprofit agencies like Kiva and Grameen seek out and create loans for those in the greatest need. With a small amount of cash, these industrious borrowers seek to start small businesses and pull themselves out of poverty. Microfinance loans may not be the best ROI, but in terms of human impact, the dollars are well invested (catalytic philanthropy).
For the past month, I have worked with data from one such organization and novel player in the microfinance game, Kiva. Kiva is an organization that allows individual (mostly US) to lend money via the Internet to low-income entrepreneurs and students in over 80 countries. Borrowers register with Kiva's in country partners and international lenders can immediatly begin providing capital to the borrowers business.
The hard part of lending money is (surprise!) getting paid back. Not only is it difficult to tell a reliable borrower from a negligent one solely from their online application, but the process of veting this loans is a long and lengthy process. Kiva combats this by using in country partners to provide local referrals, and a range of other criteria, but still a substantial proportion of first-time loan applicants end up paying nothing back. This is an especially bad experience for lenders: it’s one thing if you loan money to a person who falls into hardship and only partially repays you, but it’s entirely different if someone takes your generous loan and runs away. These “fraudsters” are a threat to the faith of lenders in the institution as a whole.
So for my Capstsone project, I undertook the task of creating a model to predict which borrowers were likely to completely skip town with their loan money.