Staggered Pricing Models Face the Facts

by John Taylor, Hyun Shin, Frank Smets, Kazuo Ueda and Michael Woodford
Editorial Board

Introduction to a Special Issue of the International Journal of Central Banking

The International Journal of Central Banking from time to time will assemble and publish collections of papers on a common topic, especially when the editors see a critical mass of submissions in an active area where quick publication of the papers as a group will encourage discussion, create synergies, and increase the potential for further breakthroughs. We have decided to make this September 2006 issue of the IJCB the first such special issue, and we are planning more special issues for the future.

The topic for this issue is at the heart of monetary policymaking: empirical modeling of inflation dynamics. For more than 25 years, monetary economists have built models of inflation that combine some form of staggered price or wage setting with the rational expectations assumption, and in recent years, they have improved appreciably on these models, most importantly by increasingly emphasizing monopolistic competition and markups over marginal cost as empirical drivers of inflation. These improvements have in turn led to fundamental questions: How well do the newer models fit the data? Are they reliable enough to be used for practical policy evaluation? The questions are similar to those raised about earlier models of inflation — questions that gave impetus to the recent improvements in the models. But researchers now have more and better data — both micro and macro — as well as more specific models to compare the data with, allowing for more stringent testing.

In the leadoff paper, Silvia Fabiani and her colleagues at the central banks in Europe add significantly to the microeconomic database for testing and validating inflation models. Their detailed survey of firms' pricing practices in the euro area documents the time between price review and price change, the prevalence of markup pricing, and many other regularities across countries that help discriminate between models and thereby can identify more empirically accurate theories.

In the second paper, Jeffrey Fuhrer focuses more on macro data testing. He presents a simple stylized staggered pricing model and uses it to pinpoint areas of empirical deficiency, the most important being whether the newer models can explain the serial correlation or the inertia of inflation.

The following three papers then examine more detailed structural models of inflation dynamics. Fabio Milani shows that some of the empirical deficiencies pointed out by Fuhrer might be overcome by adding learning to the model, effectively altering the rational expectations assumption and thereby creating slower adjustment and more serial correlation of inflation, the interest rate, and real output. Gregory de Walque, Frank Smets, and Raf Wouters show that by adding firm-specific factors of production, the inflation inertia can be increased in a way that keeps the model tractable enough for policy evaluation and consistent with micro survey data on price setting. Argia Sbordone takes a limited-information approach to estimation and, by doing so, shows that sensibly sized parameter estimates for the staggered price- and wage-setting equations emerge and that these are consistent with the time-series data as represented in vector autoregressions.

In the concluding paper, John Roberts focuses on how the timeseries properties of the macro data that we use to test the models have changed. He examines how the regression coefficient of inflation on output, the variance of output, and the variance of inflation have all declined over time. He attributes these changes at least in part to monetary policy. But regardless of the attribution, such changes in the patterns of the data alter one's assessment about whether inflation models do or do not fit the data and are therefore an important issue for empirical work in this area.