The argument of this article is strange. The "randomista" revolution essentially states that we know (or knew) very little about what actually work to foster d

The argument of this article is strange. The "randomista" revolution essentially states that we know (or knew) very little about what actually work to foster development and that RCTs are the best tool we have to obtain this knowledge. Macro-scale policies are important, but we have no tool to evaluate their effects appropriately. The article essentially says that this approach is wrong because we know that some macro-scale policies will have a much bigger impact than all those micro interventions. In this case, reducing barriers to immigration. But we do not know that, that's the whole point of doing RCTs in the first place. You can make an argument that we should spend less time focusing on micro scale interventions and more on macro-scale interventions because the former will never have a large impact and thus the best way to foster development is finding which macro intervention work. However saying that focusing on the micro scale is useless because we _already_ know about some macro intervention that work better than the micro ones is unfounded. The evidence on macro intervention is much, much weaker than the evidence from RCTs.