It is this:
Take the expectational Phillips Curve:
π = E(π) + β(u* - u)
where π is the inflation rate, E(π) is the market's expectation of inflation, u* is the natural rate of unemployment, and u is the unemployment rate.
This equation tells us that if expected inflation rises and the central bank does not want actual inflation to rise, it must take steps to drive the unemployment rate to:
u = u* + (E(π) - π)/β
Expansionary fiscal policy, coherent Austerians say, will drive up inflation expectations—will drive up E(π). A central bank that wants to hit its inflation target—and central banks do want to hit their inflation targets—will be forced to push up unemployment. The right policy to follow is one of fiscal austerity that will reduce expectations of inflation, and so allow for both full employment and price stability.
End of story.
This seems to me to be the analysis of somebody who is trapped in the 1970s, and probably trapped in Latin America as well--in someplace where expectations of inflation are genuinely unanchored, and respond rapidly and substantially to announcements of shifts in policy.
Now there are times and places in which market expectations of inflation are unanchored and do respond. Sometimes such a time and place even occurs in the North Atlantic: remember the inauguration of President Mitterand of France in the early 1980s?
But much more of the time in the North Atlantic confidence in central banks is high, inflation expectations are well-anchored, policy announcements and even short-run policy outcomes have little effect on E(π), and it requires a multi-year process of consistent upward surprises on inflation relative to expectations before inflation expectations stop being static and even become adaptive--that is, backward-looking so that this year's expected inflation is last year's actual inflation. And it requires an even greater phase change before expectations of inflation become materially forward-looking, and fiscal austerity becomes an effective employment policy via its role in calming expectations of future inflation.
We are not there now. There are no signs anywhere in the economy that we are going to be there anytime in the next generation.
So why make policy based on a model of the economy that was possibly and plausibly right for the start of the 1980s, but is clearly wrong now?