The basic point here is that ever since the 1970s we’ve been teaching a story in which an economy with excessive unemployment is one in which inflation should keep falling. Since inflation today is roughly where it was a dozen years ago, this story suggests that we’ve been on average at full employment over that period. But we know that this period was marked by weak expansions and a terrible slump, so that can’t be true. Something is wrong.
I’ve argued that the data are more consistent with a paleo-Keynesian Phillips curve in which unemployment determines the level, not the rate of change, of inflation — which could make sense given anchored expectations and downward wage rigidity. But that’s an educated guess at best, and somewhat post hoc.
The thing is, however, that this is not a new problem. There was nothing like a Phillips curve, let alone an accelerationist Phillips curve, during the Great Depression. Here’s wages:
Wages fell as the economy was slumping, but bounced back thereafter even though unemployment was high. Applying modern arguments to these data one could easily have concluded that the economy was near capacity in, say, 1939 — and in fact many economists argued at the time that most unemployment was structural, due to the mismatch between skills and the requirements of the modern economy, and could not be cured with more demand.
But then came massive fiscal stimulus in the form of rearmament and war — and it turned out that there was plenty of slack in the economy, and American workers were just fine.