John Taylor has an analysis of the American Recovery and Reinvestment Act which shows that ARRA didn’t actually lead to a substantial increase in government purchases and therefore failed to do a huge amount to boost GDP. He concludes:
More generally, the results from the 2000s experience raise considerable doubts about the efficacy of temporary discretionary countercyclical fiscal policy in practice. In this regard the experience with the stimulus packages of the 2000s adds more weight to the position reached more than 30 years ago by Lucas and Sargent (1978) and Gramlich (1978, 1979).
Paul Krugman and Noah Smith both rightly critique the last sentence here. Per Smith, “Lucas, Sargent, etc. thought that government purchases wouldn’t raise GDP” which is the reverse of Taylor’s conclusion. ARRA didn’t boost GDP because a temporary boost in government purchases won’t boost GDP, and ARRA didn’t boost GDP because it didn’t lead to an increase in government purchases are incompatible positions. They’re just both criticisms of ARRA and therefore perhaps emotionally satisfying to people who dislike Barack Obama.
That said, I really do think Keynesians need to pay more to Taylor’s first quoted sentence here. It seems clear that, in practice, we cannot and should not count on discretionary fiscal stimulus to be the centerpiece of our stabilization strategy in the event of short-term nominal interest rates hitting zero. The case that engineering a dramatic, rapid temporary run-up in government purchases would be a good way of responding to a downturn seems to me to be fairly solid. It’s much less clear to me that such a thing is doable in a Madisonian political system that features a large level of partisan disagreement about the appropriate size and scope of the government. We need to work on improved “automatic stabilizers,” we ought to be targeting a higher level of inflation so that hitting zero is less likely, and we need to explore the unorthodox monetary tools.