International Menu Costs and Price Dynamics

• Published date: 01 August 2017
• Date: 01 August 2017
• Author: Raphael Schoenle
• DOI: http://doi.org/10.1111/roie.12275

International Menu Costs and Price Dynamics

Raphael Schoenle*

Abstract

The pricing behavior of firms systematically differs across domestic and export markets. First, domestic pro-

ducer prices change approximately twice as often, the probability of synchronized price adjustment across

markets is 21% for upwards adjustments and 14% for downwards adjustments, the size of export price

changes is substantially larger and there are strong seasonality effects. Second, economic fundamentals can

only partially explain adjustment decisions and cross-market synchronization. Third, I attribute the remain-

ing unexplained part in adjustment decisions to differences in menu costs across countries. Model-implied

export menu costs are1.5% of steady state revenues and threetimes domestic market menu costs.

1. Introduction

While it is well known that firms price to market, we know very little about the

dynamics of price changes of the same product across countries. However, understand-

ing the timing of export price setting is key for understanding a number of issues in

international macroeconomics such as the global spread of inflation, fluctuations in

real exchange rates and optimal monetary policy in the open economy.

In this paper, I document several new facts about price-setting behavior in interna-

tional markets, using producer-price micro-data from the Bureau of Labor Statistics

(BLS). The novelty of my analysis is to compare price dynamics directly from the per-

spective of the producer who sets prices both in the domestic and export markets. In a

first step of analysis, I contrast price dynamics for the same products, defined at the

six-digit level, and for a set of matched firms selling in the domestic and export mar-

ket.

1

I show that the pricing behavior of firms systematically differs across markets in

terms of frequency, timing and size of price changes. In a second step, I show that eco-

nomic fundamentals such as inflation, productivity, demand, exchange rates and mar-

ket structure can only partially explain adjustment decisions and cross-market

synchronization. Finally, I present a dynamic menu-cost model of price setting and

attribute the unexplained part in adjustment decisions to differences in menu costs

across countries. I estimate that export menu costs have to be approximately three

times larger than domestic market menu costs to account for the empirical findings.

* Schoenle: Department of Economics, MS 021, Brandeis University, 415 South Street, Waltham, MA

02453-2728, USA. Tel: 11-781-736-2268; Fax: 11-781-736-2269; E-mail: schoenle@brandeis.edu. The

author would like to thank Alan Blinder and Marc Melitz for invaluable advice and encouragement.

Thanks are also due to Pinelopi Koujianou Goldberg, Saroj Bhattarai, Thomas Chaney, Gauti Eggerts-

son, Oleg Itskhoki, Nobu Kiyotaki, Kalina Manova, Emi Nakamura, Chris Sims, Mu-Jeung Yang and

seminar and workshop participants at Brandeis University, Erasmus University Rotterdam, ESSEC,

European School of Management and Technology, Federal Reserve Bank of Dallas, IMF Research

Department, Royal Holloway University of London, Princeton University, the Microeconomic Sources

of Real Exchange Rate Behavior Conference at Vanderbilt University and the AEA meetings in Denver

for helpful suggestions and comments. This research was conducted with restricted access to the Bureau

of Labor Statistics (BLS) data. The views expressed here are those of the author and do not necessarily

reflect the views of the BLS. The author also thanks the project coordinators, Ryan Ogden, and espe-

cially Kristen Reed, for substantial help and effort, as well as Greg Kelly and Rosi Ulicz.

V

C2016 John Wiley & Sons Ltd

Review of International Economics, 25(3), 578–606, 2017

DOI:10.1111/roie.12275

I start my empirical analysis by contrasting the frequency of price changes for the

same products in the domestic and export markets. I find that domestic prices change

2.1 to 2.7 times as often as export prices of the same product. This difference in adjust-

ment frequencies at the product level corresponds to a median difference in the dura-

tion of price spells of 5–10 months. My estimates based on the data for matched firms

imply a duration difference of 7.4 months. Using different methodologies to calculate

the frequency of export price changes, I find that differences between domestic and

export price-setting behavior are robust to the presence of missing values in the export

data.

Second, while the fact that export prices change less frequently than domestic prices

does not rule out that export prices adjust whenever domestic prices adjust, the data

show a striking absence of synchronization in export and domestic price adjustment.

The average monthly probability of upwards export price adjustment is 21% condi-

tional on complete domestic upwards price adjustment of the same product. At the

same time, the monthly probability of synchronized downwards price adjustment in

the export market is 14%. I find almost identical results when I additionally break

down the calculations by export destinations. There are two natural benchmarks for

which to compare these synchronization results: First, since US exports are denomi-

nated in US dollars in the data, the Law of One Price would imply perfect synchroni-

zation for the timing of price adjustment. A second benchmark is the fraction of other

products, defined at the six-digit level, which adjust in the domestic market within the

same four-digit category whenever a particular product fully adjusts in the domestic

market: the probability of synchronized domestic upwards (downwards) adjustment of

that particular product with other products in the same four-digit category is 48%

(50%).

In addition to computing probabilities of export price adjustment conditional on

complete adjustment of all goods in the same domestic product category, I regress

domestic and export price adjustment decisions of each good in a given product cate-

gory on the fraction of domestic price changes that have the same sign in that product

category. I find that the estimated coefficient on the fraction of same-signed domestic

adjustment is substantially larger in regressions when the dependent variable is the

individual domestic adjustment decision than when it is the export adjustment deci-

sion. Therefore, regression results also show that export price adjustment decisions are

not very synchronized with domestic price adjustment decisions. Because domestic

and export prices change at different frequencies and not in a synchronized way, my

findings suggest that we may have to carefully modify the standard modeling assump-

tion of identical Calvo parameters or even simultaneous adjustment of prices across

markets.

Third, I find that differences in price-setting behavior across markets also extend to

the differences in the size of price changes: conditional on adjustment, the median

absolute size of export price changes is 39% larger than the size of domestic price

changes of the same product. The median percentage difference in between export

and domestic absolute sizes of price changes is 1.75%. When I break down price

changes by destination, the median absolute size of export price changes is 30% larger

than the size of domestic price changes. The median percentage difference between

export and domestic absolute sizes of price changes is now 1.43%. Comparing the size

of price changes for matched firms, I find even more pronounced differences: the

median absolute size of export price changes, conditional on adjustment, is 2.5 times

as large as the size of domestic price changes, conditional on adjustment. The median

INTERNATIONAL MENU COSTS AND PRICE DYNAMICS 579

V

C2016 John Wiley & Sons Ltd