My column for Bloomberg View makes it:
Last summer, Dylan Matthews reviewed the research on the stimulus for the Washington Post and dug up six studies that found a positive effect. Three of them were based on models that assume the stimulus worked. Three of them were supposedly empirical confirmations of this effect. These three all found that states (or counties) that got more stimulus money had stronger economic performances than places that received less.
But nobody denies that the federal government can shift the distribution of economic activity. If Congress were to give me $50 billion, I am sure car dealerships and liquor stores in my area would see an uptick in sales. That doesn’t mean the nation as a whole would come out ahead. (I am willing to go along with the experiment if Congress doubts this.)
Other research on the stimulus, meanwhile, has uncovered reasons for skepticism about its effect. . . .