“
This is nothing less than a whole new approach to economics. The randomistas don’t think in terms of models. They don’t believe humans are rational actors. Instead, they assume we are quixotic creatures, sometimes foolish and sometimes astute, and by turns afraid, altruistic, and self-centered. And this approach appears to yield considerably better results. So why did it take so long to figure this out? Well, several reasons. Doing randomized controlled trials in poverty-stricken countries is difficult, time consuming, and expensive. Often, local organizations are less than eager to cooperate, not least because they’re worried the findings will prove them ineffective. Take the case of microcredit. Development aid trends come and go, from “good governance” to “education” to the ill-fated “microcredit” at the start of this century. Microcredit’s reckoning came in the form of our old friend Esther Duflo, who set up a fatal RCT in Hyderabad, India, and demonstrated that, all the heartwarming anecdotes notwithstanding, there is no hard evidence that microcredit is effective at combating poverty and illness.13 Handing out cash works way better. As it happens, cash handouts may be the most extensively studied anti-poverty method around. RCTs across the globe have shown that over both the long and short term and on both a large and small scale, cash transfers are an extremely successful and efficient tool.14
”
”