Angus Deaton

Randomized control trials are a useful tool in development economics. But RCTs are not a gold standard; like other tools they face problems of inference, they cannot claim to have established causality and some of them raise serious ethical concerns.