Thursday, April 5, 2007

Doubts in Woodford's Output Gap (amv)

Michael Woodford is one of the most brillant economists of our age. He is a real master in many respects: He is a creative theorist, a vivid writer (even though his Interest and Prices is indeed tough to read), a great 'modl'-builder (the reason why Interest and Prices is so tough to read), virtuosic with empirical results and excellent, when it comes to break down theory to concrete policy advice. It is not surprising that my colleague, fellow, friend and co-blogger, fg, is attracted to Woodford's rigour and mastery. I, however, still have some doubts in Woodford (but this may not be of any significance, since I'm a pighead and thus not easily persuadable).

Since this is not the place for a full-fledged controversy over Woodford's treatise, I want to single out one sinlge point: Woodford's concept of the output gap. This gap marks the difference between the actual growth of demand (which is identical to the actual growth of output, since the aggregate good market is presumed to be always balanced) and the 'natural rate of output' growth. The latter is a concept derived from the Real Business Cycle approach and describes the growth path given optimizing individuals adjusting prices and wages quickly to anticipated nominal distortions. The actual growth path, in contrast, is described by rigid wages and prices.

And here is my problem: The rigidity of wages and prices is not simply presumed (this was the old New Keynesian approach), but rather founded on the rational and optimizing behaviour of the representative household and firm. But if, for any given set of circumstances, it is optimal to set wages and prices rather rigidly, why is the outcome of this optimal behaviour inferior to a regime of flexible nominal entities? Is it not the general environment of any choice, which determines - together with the utility and cost functions - the optimality of choosing? If then rigidity is optimal, is it not exactly the environment of a particiular time and place, which determines this outcome? Is it thus legitimate to offer a counterfactual environment as a reference setting for optimal policy, that is, a setting of choice, which would determine the rational individual to act differently but which simply does not exist? Is it not just wrong that the difference between two states of the worlds, which cannot co-exist, should guide real world monetary policy? Does it make sense to consult a central bank to adjust its interest rate to a level which would be optimal in a world that cannot exist, if the central bank is supposed to act like that?

Is anyone willing to redeem me from my disbelief? I would be extremely grateful!

P.S.: A similiar critique can be found in Boianovsky and Trautwein.