Price adjustments in a general model of state-dependent pricing

In this paper, we show that a simple model of smoothly state-dependent pricing generates a distribution of price adjustments similar to that observed in microeconomic data, both for low and high inflation. Our setup is based on one fundamental assumption: price adjustment is more likely when it is more valuable. The constant probability model (Calvo 1983) and the fixed and stochastic menu cost models (Golosov and Lucas 2007; Dotsey, King and Wolman 1999) are nested as special cases of our framework. All parameterizations of our model can be ranked according to a measure of state dependence. The fixed menu cost model has the highest possible degree of state dependence; the parameterization which best fits US microdata has low state dependence. The fixed menu cost model is inconsistent with the evidence both because it never generates small price adjustments, and because it implies a large fall in the standard deviation of price adjustments as trend inflation increases. Even though the state dependence of our preferred parameterization is almost as low as that of the Calvo model, it is well-behaved when we change the steady state inflation rate, matching the data at least as well as Golosov and Lucas' model.