He argues that the Recalculation story falls apart
when you ask why, say, a housing boom -- which requires shifting resources into housing -- doesn't produce the same kind of unemployment as a housing bust that shifts resources out of housing.
This is a good question. I would say that answer may be that the dynamics of booms and busts are asymmetric.
Note that the housing boom took place slowly, and it built up over a period of years. Several economists, Krugman included as I recall, were already talking about a housing bubble in 2004, even though in hindsight the biggest price excesses were still to come. So the housing boom must go back even further, to at least the late 1990s. In contrast, the collapse of house prices took less than two years (assuming they have bottomed out, which may be a brave assumption).
I think it is reasonable to generalize this notion that booms are longer and relatively gradual, while busts are sudden. If this generalization holds, then it ought to be harder for the economy to adjust to a bust than to a boom.
As an aside, I do not want to associate myself with the view that Krugman finds in Schumpeter, which is that the unemployment of a recession is a good thing. Instead, I am telling a descriptive story. Imagine the economy being run by an imperfect central planner (the market), which finds itself for a period of time unable to come up with productive uses for an unusually large number of workers.
This Recalculation story does not imply that government can do nothing or should do nothing about unemployment. On the contrary, it suggests that the market is faltering. However, the Recalculation story does say that it matters how the government tries to redirect resources. If it increases demand where demand already is sufficient, or if it increases demand in sectors where the needs are only temporary, it is not helping.
One advantage of something like a payroll tax cut (or an employment subsidy, which is the same thing but sounds better to people on the Left), is that it does not require government planners to make the right guesses about which firms or industries should expand. Another advantage is that a payroll tax cut can start right away, without having to wait for bureaucrats to go through the procedural hoops needed for spending. One disadvantage of throwing money at state and local governments is that the contraction in that sector has been miniscule compared with the contraction in the private sector, and in fact in the long run we may need some contraction in the state and local government sector.
The Recalculation story does not tell me that we should drop the concept of stimulus and balance the Federal Budget. However, it tells me that we should not use a crude, aggregative notion of macro to justify what looks to me like a permanent transfer of resources and power from the private sector to government.
UPDATE: Russ Roberts says more about asymmetry. Also, Alex Tabarrok emails me a link to a paper by Bloom, Floetotto, and Jaimovich that uses Dark Age macro to argue that variations in uncertainty are a cause of recessions. When uncertainty rises,
firms become cautious and postpone decisions. Productivity growth falls since reallocation across production units pauses when uncertainty is high. Once uncertainty falls back down activity quickly resumes as firms address their pent-up demand.
Again, too much Dark Age macro for my taste, but I think that uncertainty and hesitancy are an important issues to consider.