I’ve been to a lot of DSGE seminars. At Duke, at the Fed, heck we even have ’em in Oklahoma! One thing most have in common is that they use the Calvo rule to implement price stickiness.
The Calvo rule is a shortcut which assumes that a firm has a fixed probability of getting a chance to change its price in any given period. This probability is independent of how far their price is away from the optimum or how long it’s been since they changed prices.
Time after time I’d object, and time after time I’d be told that the Calvo rule was a benign modeling “trick” that made things less intractable with no real influence on the results.
This morning, I was pleased to discover the research of a young Chicago (Booth) researcher named Joseph Vavra. He’s doing a lot of very interesting work, but what really caught my attention was his piece, The Empirical Price Duration Distribution and Monetary Non-Neutrality.
Here’s the abstract:
Allowing for price adjustment probabilities that vary with the number of periods since an item last adjusted (duration-dependenceí) provides a significantly better fit of observed price spells in CPI and grocery store micro data than the Calvo model, even if the latter is extended to incorporate item-specific adjustment probabilities. Furthermore, extending the Calvo model to match both duration-dependence and cross-item heterogeneity, as observed in the micro data, leads to an increase of 100-230% in monetary non-neutrality, even with no strategic-complementarities. As much as half of this increase is driven by duration-dependent adjustment probabilities.
Nicely played, sir, kudos. Not so innocuous after all, that Calvo rule.
Given how our profession often works, I fear he may have trouble getting this published. But, since he has R and Rs at the QJE and Econometrica, I think he’ll be OK no matter what happens to this piece.