Russ Roberts writes:
Menzie Chinn invokes the CBO “estimates” to argue against those who say the stimulus didn’t work. Did the stimulus help turn the economy around and create jobs? I’m skeptical on logical grounds but I confess that I do not have strong empirical evidence on my side.
But those who defend the stimulus have no empirical support either…
…The CBO “estimates” are not an analysis of what the stimulus actually did but rather what some predicted it would do. They have done NO independent non-partisan analysis of what actually happened.
Yesterday, I had a headache. I took a couple tablets of aspirin. (Actually, it was ibuprofen, but the point remains.) My headache subsequently disappeared. I have no direct empirical evidence that the headache disappeared as a consequence of the aspirin, but I have a plethora of studies that suggest that aspirin (or ibuprofen in this case) can relieve headaches.
As the foregoing example suggests, it does seem to me there is empirical evidence. It’s just not the direct sort Professor Roberts desires. Admittedly, there is a dispute over the size of the multipliers — that’s why the CBO used ranges. For surveys of many studies of multipliers, see this Econbrowser compilation, especially this entry and this entry.