Friday, April 30, 2010

Doing Calvo all wrong

On a number of instances, I have hyperventilated here about the misuse of the Calvo pricing hypothesis, which is an analytically convenient but empirically wrong hypothesis that a only fixed proportion of firms can change prices in any given period, and that price changes are not history or state dependent. There is a very limited set of circumstances where this is a reasonable assumption, and the following paper is by far not in that set.

Willem Van Zandweghe and Alexander Wolman study monetary policy in a Calvo world. The only thing that is endogenous is the money supply. This implies in particular that the share of firms adjusting prices is constant no matter what the money supply or its growth rate are. One would have thought that firms, when faced with a larger money growth rate and thus inflation, would increase the frequency at which they adapt prices. Also, money velocity is assumed to be constant, while we know very well that it is not.

The worse is that the paper is trying to determine what the optimal monetary policy is, and even worse, it is trying to give quantitative answers. There is no way any of the results can be credible.

11 comments:

Agent Continuum said...

No, this is a paper about the model, not about the economy. :-)

Also, they don't reference Barro-Gordon.

Kansan said...

This paper really questions the relevance of the research departments at the Federal Reserve Banks of Kansas City and Richmond. What is the use of this paper? And why are these Feds so proud of this paper to disseminate it as a working paper?

Anonymous said...

I don't agree with you. The Calvo approximation will be reasonable as long as the degree of price stickiness is close to invariant with respect to the variable being changed. In this case, they're studying optimal inflation. For "low" levels of inflation, the degree of price stickiness appears to be stable, see Gagnon (2009). For higher levels of inflation, price stickiness starts to vary with SS inflation. The breakpoint appears to be in the range of 10% annual inflation. Since they're looking at levels of inflation mostly well below this breakpoint, the Calvo assumption should be relatively innocuous.

Kansan said...

In other words, they got lucky by finding that the optimal inflation rate is below 10%. That is silly.

AlabamaFor You said...

Gagnon studies Mexico, a country with traditionally high inflation. Inflation below 10% is considered low there, so the fact that the frequency of price adjustment does not vary much below 10% is not too surprising. This tells nothing about the US.

Anonymous said...

Dhyne et al (2006) have same result for European countries.

AlabamaFor You said...

Looking at Dhyne et al. (2006), table 6 states that is all countries but the Netherlands (which has always had low inflation rates), prices changes do depend on the aggregate inflation rate. Your point is not valid.

rosserjb@jmu.edu said...

The idea that the Calvo pricing mechanism is endogenous is probably as likely as that the natural rate of unemployment is endogenous. It is likley that both vary somewhat endogenously over time due to various economic processes.

Anonymous said...

^^ In the ECB working paper, Table 8 shows that once you control for other factors, average inflation has no statistically significant effect on the frequency of price setting. Table 6 does not control for other factors.

AfricaForEver said...

Have people forgotten the Lucas Critique?

Anonymous said...

"The idea that the Calvo pricing mechanism is endogenous is probably as likely as that the natural rate of unemployment is endogenous. It is likley that both vary somewhat endogenously over time due to various economic processes."

This statement is similar to most of Rosser's work -- it says absolutely nothing.