Some [business cycle] theories hold that booms cause busts, because good times allow bad investments to build up in the financial system. According to these theories, the larger the boom, the larger the crash that follows.
Then there’s the so-called plucking model. Proposed by the legendary economist Milton Friedman, it holds that the economy is like a string on a musical instrument — recessions are negative events that pull the string down, and after that it bounces back. Just as a string snaps back faster if you pull it harder, this theory holds that the deeper the recession, the faster the recovery that follows. But you can only pluck the economy in one direction; bigger expansions don’t lead to bigger recessions.
Back in 1996, Roger Garrison published an excellent Austrian response to Friedman. But let me give a quick version here:
The data by which Noah Smith thinks Friedman's plucking model has been vindicated, are perfectly consistent with Mises' theory as to what causes recessions. To see why, let's just imagine a simple (and exaggerated) example.
Suppose the whole labor force is dedicated to making hardware. In a normal, sustainable period of growth, 10% of the workers make hammers, 40% make nails, 10% make screwdrivers, and 40% make screws. This is healthy, balanced growth, where the underlying capital structure is sustainable.
Now suppose because of central bank manipulation of interest rates, the price system is screwed up and all the workers start making hammers. This won't boost "total output." There will still be "full employment." It just means that the composition of output will be heavily skewed to hammers, and away from nails, screws, and screwdrivers.
This is obviously not a sustainable path. Soon enough, the supplies of nails and screws will run out. It does little good to be cranking out hammers, if there are no new nails coming online.
What I've just described is a metaphor for what happens during an unsustainable boom, in the Misesian sense. Now once the crisis hits, "total output" does indeed drop, as workers need to be reallocated back to a more sensible niche in the division of labor.
Now as a completely separate issue, it's clear that the increase in output measured from the trough is going to be directly related to how deep the trough is. In our example, the "crash" is going to be really bad, because the economy is going to suddenly run out of nails and screws completely. If instead the workers had only been slightly knocked out of balance, then the ensuing crash wouldn't be as bad.
But either way, my point is that the wild swings in "total output" will appear to follow a bust-boom pattern, rather than a boom-bust one. This is so, even though by construction, my story fits the Misesian pattern of an unsustainable production period being followed by an inevitable bust.
For more details, read Garrison's article, or my earlier "sushi article" that many people found illuminating.