The Missing Inflation Puzzle: The Role of the Wage‐Price Pass‐Through

• Published date: 01 February 2022
• Author: SEBASTIAN HEISE,FATIH KARAHAN,AYŞEGÜL ŞAHIN
• Date: 01 February 2022
• DOI: http://doi.org/10.1111/jmcb.12896

DOI: 10.1111/jmcb.12896

SEBASTIAN HEISE

FATIH KARAHAN

AY ¸SEGÜL ¸SAHIN

The Missing Ination Puzzle: The Role of the

Wage-Price Pass-Through

Price ination in the U.S. has been slow to pick up in the last two decades.

We show that this missing ination can be traced to a growing disconnect

between unemployment and core goods ination. We exploitrich industry-

level data to show that weakening pass-through from wages to prices in the

goods-producing sector is an important source of the slow ination pickup.

We develop a theory where markups and pass-through depend on rms’

market shares and show that increased import competition and rising mar-

ket concentration reduce pass-through from wages to prices. Wend strong

empirical support for these predictions.

JEL codes: E24, E31

Keywords: ination dynamics, import competition, market concentration,

Phillips curve, pass-through

Price ination in the U.S. has remained sluggish and below

the Federal Reserve’sination target of 2% during the decade-long expansion follow-

ing the Great Recession. This behavior of price ination has been considered puzzling

by many policymakers and academics especially given that the unemployment rate

We are very grateful to DavidDam, Meghana Gaur, and Will Schirmer for superb research assistance.

We thank our discussant Slavik Viacheslav for many insightful comments and sharing his codes with

us. We also thank Susanto Basu, Yuriy Gorodnichenko, Giuseppe Moscarini, Emi Nakamura, Giorgio

Primiceri, and Ken West for helpful comments and suggestions. We thank the participants of the NBER

Summer Institute 2020 Monetary Economics Group, the Barcelona GSE Summer Forum 2021, and the

Annual Research Conference of the Central Banks of Ukraine and Poland for their comments. The views

expressed in this paper are those of the authors and do not necessarily reect the position of the Federal

Reserve Bank of New Yorkor the Federal Reserve System.

S H and F Kare at Federal Reserve Bank of New York (E-mail: sebas-

tian.heise@ny.frb.org , fatih.karahan@ny.frb.org ). A Sis at University of Texas Austin and

NBER (E-mail: aysegul.sahin@austin.utexas.edu ).

Received August 3, 2020; and accepted in revised form August 16, 2021.

Journal of Money, Credit and Banking, Supplement to Vol. 54, No. S1 (February 2022)

© 2021 The Ohio State University

8:MONEY,CREDIT AND BANKING

Fig 1. Evolution of Core CPI Ination during Economic Expansions.

N: BLS and authors’ calculations. The left panel plots the cumulative core CPI ination (all items less food and

energy, seasonally adjusted) against time, starting at the quarter of peak unemployment of a givenrecession. The right

panel plots the same data against the unemployment recovery gap dened in the main text.

stood at 3.5% at the end of 2019—its lowest level in almost half a century.1In this

paper, we revisit this growing disconnect between unemployment and ination and

show that declining pass-through from wages to prices in the goods-producing sector

has been an important source of the slow ination pickup. We attribute the decline

in pass-through to rising import competition and increased market concentration and

provide strong empirical support for these two channels using rich industry-leveldata.

We also show that the twoexplanations are consistent with the predictions of a theo-

retical framework in which markups and pass-through are a function of rms’ market

shares as in Atkeson and Burstein (2008).

The left panel of Figure 1 shows the evolution of the cumulative consumer price

index (CPI) in the U.S. starting from the business cycle trough for each of the past ve

economic expansions. As the gure shows, each recovery is associated with a lower

cumulative consumer price ination compared to the previous one, with cumulative

CPI increasing by 25% in the rst 20 quarters of the recovery following the 1981–82

recession, by 15% following the 1990–91 recession, and by less than 10% over 20

quarters in the last expansion. While the evolution of consumer price ination over

time is informative, the duration of expansions is an increasingly misleading indica-

tor of labor market recovery given the emergence of jobless recoveries. We therefore

propose a novel but simple measure of labor market recoverythat is consistent across

expansions—the unemployment recovery gap—and consider the evolution of price

ination with respect to this new measure. Our new metric computes the share of the

rise in the unemployment rate during the preceding recession that has been reversed

during the current expansion. An unemployment recovery gap of 100% thus implies

1. NY Times Upshot: Janet Yellenand the Case of the Missing Ination, June 14, 2017; Living Life

Near the ZLB, John C. Williams, Remarks at 2019 Annual Meeting of the Central Bank Research Asso-

ciation (CEBRA), New YorkCity, July 18, 2019.

SEBASTIAN HEISE, FATIH KARAHAN AND AY¸SEGüL ¸SAHIN :9

that the unemployment rate has declined back to its prerecession trough.2In the right

panel of Figure 1, we plot the evolution of cumulative CPI ination relative to this

new metric. The 1970s and the early 1980s clearly stand out as expansionary periods

with ination picking up rapidly after the end of the recessions. The three most re-

cent expansions exhibit a milder rise in cumulative ination. Thus, even taking into

account the emergence of jobless recoveries, the behavior of ination changed con-

siderably after 1990. This observation is consistent with the decline in the slope of

the price Phillips curve as documented in detail in Stock and Watson (2019) and the

references therein.

We examine ination in goods and in services and identify a notable change in

the behavior of core goods ination. While core goods prices rose by about 20% as

unemployment fell from recessionary peak to trough during the 1982–89 expansion,

core goods prices barely increased in the 2009–20 expansion. At the same time, core

services ination has picked up in all expansions. We analyze a counterfactual sce-

nario in which goods prices behave in the 2009–20 expansion in the same way as they

did during the 1982–89 expansion. Weshow that around 50% of the missing ination

relative to the 1980s expansion can be traced to goods prices despite core goods hav-

ing a weight of only 25% in the CPI. Interestingly, our empirical analysis shows that

the change in ination dynamics is not driven by a differential behavior of wages.

Wage ination has been similar in goods and in services in the last two expansions.

This result is consistent with Galí and Gambetti (2019) and Stock and Watson (2019)

who nd a more stable Phillips curve for wages. It is also consistent with Beraja,

Hurst, and Ospina (2019), who nd a strong relationship between employment and

wage growth at the state level during the Great Recession.

These ndings suggest that the changing pass-through from wage changes to price

changes is a promising explanation for the missing core goods ination. Toinvestigate

this possibility, we estimate the impulse responses to changes in wage ination on

both consumer and producer price ination following Jordà (2005). We uncover a

striking change in the relative behavior of wageand price ination: pass-through from

wages to prices was signicant and positive until the early 2000s, but then dropped

sharply and has been statistically indistinguishable from zero in the last two decades.

The Wald test for a structural break identies a structural break in 2004/Q3 for the

CPI and in 2002/Q4 for the PPI.

Our paper proposes two related explanations for the disappearance of wage-price

pass-through: rising import competition and increasing market concentration. A sub-

stantial literature has documented an increase in import competition in the U.S. man-

ufacturing sector since China’s WTO entry in the early 2000s (e.g., Autor, Dorn,

and Hanson 2013). At the same time, recent work has pointed out that market

2. There are alternative measures of labor market utilization that capture wage-growth better than the

unemployment rate such as the NEI developed by Hornstein, Kudlyak, and Lange (2014), the tightness

measure developed in Moscarini and Postel-Vinay(2017), and the aggregate hours gap in Faberman et al.

(2020). These alternative measures of labor market utilization typically start in 1994 due to CPS redesign.

We focus on the unemployment recovery gap measure due to its simplicity and availability of a longer

time series.