Disagreement in Consumer Inflation Expectations

• Published date: 01 December 2023
• Author: TOMASZ ŁYZIAK,XUGUANG SIMON SHENG
• Date: 01 December 2023
• DOI: http://doi.org/10.1111/jmcb.12981

DOI: 10.1111/jmcb.12981

TOMASZ ŁYZIAK

XUGUANG SIMON SHENG

Disagreement in Consumer Ination Expectations

By carefully matching the data sets from the Michigan Surveyof Consumers

with the Survey of Professional Forecasters, we show that there exists sub-

stantial heterogeneity in the propensity of U.S. households to learn from

experts in forming ination expectations. Additional results for a group of

European economies broadly conrm this observation. We advance an ex-

tended version of the sticky-information model to analyze disagreement in

consumer ination expectations. Besides differences in consumers’ propen-

sities to learn, disagreement in our model arises from heterogeneity in con-

sumers’ fundamental ination and past expectations and experts’ different

views about future ination.

JEL codes:E32, E52

Keywords:consumers, disagreement, ination expectation, information

rigidity

D     public mat-

ters. Recent advances in macroeconomics have emphasized the role of disagreement

in signaling upcoming structural changes in the economy (Mankiw et al. 2004), and as

a proxy for uncertainty in driving business cycle uctuations (Bloom 2009). Yet, why

ordinary people disagree in their expectations, and how best to model this heterogene-

ity, remains an open question. We answer this question by matching household and

expert ination expectations and by building a theory of consumer expectation up-

dating.

Our theory has three key elements. First, consumers hold different beliefs about

price changes, gained from personal experiences on shopping and the previous

This paper represents the opinions of the authors and it is not meant to represent the position of Naro-

dowy Bank Polski. The authors are grateful to the participants of the Computational and Financial Econo-

metrics (CFE) Conference in Pisa, Italy, and the Georgetown Center for Economic Research (GCER)

Biennial Conference for their helpful comments. Wethank the editor, Pok-sang Lam, and two anonymous

referees for their useful suggestions. Longji Li provides outstanding research assistance. All remaining

errors are of the authors.

TŁ is the Economic Advisor, NarodowyBank Polski, and an Associate Professor,Institute

of Economics, Polish Academy of Sciences (E-mail: Tomasz.Lyziak@nbp.pl).X S S

is the Professor of Economics, American University (E-mail: sheng@american.edu).

Received May 16, 2019; and accepted in revised form March 22, 2022.

Journal of Money, Credit and Banking, Vol. 55, No. 8 (December 2023)

© 2022 The Ohio State University.

2216 :MONEY,CREDIT AND BANKING

ination rates experienced in their lifetime. Second, consumers obtain from experts

public information about the trends in future ination via newspapers and social me-

dia. Consumers are not constrained to rely on consensus expert forecasts, but are

allowed to learn from different individual expert forecasts instead. Third, households

can have different propensities to learn from experts.Consumers then combine public

and private information in forming their ination expectations.

The ingredients of our theory are motivated by the empirical ndings. Our pri-

mary database of household forecasts comes from the Michigan Survey of Consumers

(MSC) that contains both quantitative and qualitative ination expectations. We use

both forms of expectations to estimate the central tendency and dispersion among

consumers and, in particular, quantify the qualitative responses following the prob-

ability method. By carefully matching the database of consumer expectations with

that of experts from the U.S. Survey of Professional Forecasters, we nd that ina-

tion expectations between laymen and experts differ persistently from each other. It

is consistent with the results reported in the literature that households—in contrast

to experts—pay close attention to salient price changes, such as oil and food prices;

see, for example, Coibion and Gorodnichenko (2015b), Berge (2018), and Binder

(2018). By contrast, experts respond more to monetary policy and macro indicators.

We also observe substantially higher levels of disagreement among the public than

disagreement among professional forecasters that is reected in the opinions voiced

in media outlets.

Our model is closely related to the theoretical literature on expectations forma-

tion with information frictions. For instance, Mankiw and Reis (2002) propose the

sticky-information model that explains agents’ rational inattention in terms of lim-

ited resources and the cost of updating information sets. Carroll (2003) develops an

epidemiological model of expectations formation that can be viewed as providing

microfoundations for the Mankiw–Reis model. Our model differs from the sticky-

information model in an important aspect. Disagreement in Carroll (2003)’s model,

or in sticky-information model in general, arises only from different generations of

consumers using different information vintages and there is no disagreement within

a generation.1In contrast, our model generates disagreement within a generation due

to consumers’ exposure to different expert views about ination even under full in-

formation updating. Sims (2003), Woodford (2003), and Mackowiak and Wieder-

holt (2009) advocate the noisy information model that emphasizes the limited ability

of economic agents to process new information from noisy signals. In contrast to

the noisy-information model where agents always solve a signal extraction problem,

households in our model observe different views of experts and use these views as

direct inputs in forming their expectations. Importantly, households are allowed to

differ from each other in terms of their propensities to learn from experts.

Our paper builds on the burgeoning literature exploring cross-sectional distribu-

tion of forecasts. One strand of the literature examines the disagreement among

1.It should be noted that in another version of his study, Carroll (2006) mentions the possibility of

heterogenous propensities to learn.