Ahmed Mushfiq Mobarak of Yale University, who has experimented with encouraging Bangladeshi farmers to migrate to cities during the lean season, thinks it unfair to compare carefully tested projects to others where the cost-benefit numbers are “essentially made up”. Binning reliable, low-scoring projects for untested high-scoring ones would be foolish. But if the upshot is more scrutiny for promising projects, the exercise is useful. And almost anything would be better than spending money on projects because their backers can tell a good story, or because they are supported by powerful politicians.