From Micro to Macro: A New Methodology to Discriminate Among Models

Abstract

We propose a new methodology to discriminate among models that uses both micro and macro moments to discipline model choice. The key insight lies in using macro moments conditional on micro moments to discipline the response of the main variable of interest following a key policy shock. Some of the micro moments may be sufficient statistics. In an application to discriminate among leading price-setting models, Calvo and menu cost, we show that both Calvo and menu cost models match key micro price moments. However, only Calvo replicates the irrelevance of kurtosis for monetary non-neutrality following a monetary policy shock. Our menu cost model can match the irrelevance of kurtosis, but at the cost of missing key micro price moments.JEL classification: E13, E31, E32

Investigating Nonneutrality in a State‐dependent Pricing Model With Firm‐level Productivity Shocks

“…Karadi and Reiff () and Hong et al. () also examine the ALL sufficient statistic in models that deviate from the specific assumptions of ALL. Karadi and Reiff's model has idiosyncratic shocks drawn from a mixture of normals distribution.…”

“…In ALL average inflation is zero, and individual firms' prices drift because of the nonstationary productivity process. Karadi and Reiff (2019) and Hong et al (2019) also examine the ALL sufficient statistic in models that deviate from the specific assumptions of ALL. Karadi and Reiff's model has idiosyncratic shocks drawn from a mixture of normals distribution.…”