A Map of Markups: Why We Observe Mixed Behaviors of Markups

• Date: 01 June 2017
• DOI: http://doi.org/10.1111/jems.12193
• Published date: 01 June 2017
• Author: Seongman Moon,Seong‐Hoon Kim

A Map of Markups: Why We Observe Mixed Behaviors

of Markups

SEONG-HOON KIM

Department of Economics

YonseiUniversity - Wonju

1 Yonseidae-gil,Wonju-si, Kangwon-do, 26493, Korea

s.kim.econ@gmail.com

SEONGMAN MOON

Department of Economics and Research Institute of Big Data Business

Chonbuk National University

567 Baekje-daero, Deokjin-gu, Jeonju-si Jeollabuk-do 54896 Korea

nopasanada0501@gmail.com

This paper proposes an explanation for mixed evidence on the behaviors of markups. The key

mechanism consists of two complementary channels through which firms handle uninsurable

business losses. One channel is based on cost-compensating motive, by which firms raise prices to

reflect higher losses stochastically associated with higher output levels. The other channel is based

on loss-balancing motive, by which firms lower prices to countervail higher losses stochastically

associated with higher output levels. The relative responsiveness of the two channels to a shock

depends on each firm’s fundamental characteristics and leads to a sharp division of markup

cyclicality across sectors.

How markups move, in response to what, and why, is almost terra incognita

for macro. ( . . .) [W]e are a long way from having either a clear picture or

convincing theories, and this is clearly an area where research is urgently

needed. — Blanchard (2009, p. 220)

1. Introduction

Economists have studied the behavior of markups first because it illuminates the charac-

teristics of market structure and second because it clues to the nature of business cycles.

For example, theories based on procyclical competition at the industry level predict

markups to fall during booms and thus leave room for monetary expansion to have less

pressure on inflation. Theories that predict markups to rise during booms would have

different implications on the nature of business cycles and policy consequences. Data,

however, show mixed evidence on the behavior of markups. For example, markups in

Textiles (SIC code 22) and in Apparel (SIC code 23) behave differently over the U.S.

business cycle (Rotemberg and Woodford, 1991, table 8.b; p. 109). Basu and Fernald

(1997), Gopinath et al. (2011), and De Loecker and Warzynski (2012) have also docu-

mented sectoral and locational differences in markups. The aim of this paper is to add an

We thank Yongsung Chang and seminar participants at the 2016 Korean Econometric Society meeting for

helpful comments and data. We areresponsible for all errors. Corresponding Author: Seongman Moon

C2017 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume26, Number 2, Summer 2017, 529–553

530 Journal of Economics & Management Strategy

explanation for how and why the markups may behave differently across sectors thereby

leading to unclear and mixed cyclicality as observed in data.

Our theory of markups takes the presence of uninsurable business losses to the

fore of story. Asset markets are incomplete over goods market outcomes: General Motors

safeguards its business from losses due to various natural disasters but cannot prevent

losses incurred through normal business operation. The Real Greek, asmall restaurant in

London, holds similar insurance policies. The restaurant loss-proofs its business against

unprecedented flood but is exposed to unprecedented customer visits. The managers

will then take their uninsurable business losses into account when making pricing and

production decisions.

We model uninsurable business losses within a simple price-setting newsvendor

environment.1In a world where firms set prices and produce outputs before the real-

ization of their market demand, they will take account of the full distribution of market

outcomes contingent to their pricing and production decisions. Each of the contingent

outcomes bears either of two loss consequences: excess supply versus excess demand.

Firms take the costliness of these loss consequences into their pricing and production.

We present two complementary motives which firms would act on: cost-compensating

motive versus loss-balancing motive.

Cost-compensating motive means to reflect the expected cost of business losses

likely to incur when producing more. This motive leads firms to add the shadow cost

of production to the traditional marginal cost. One important implication is that, even

in competitive markets, the additional cost forms wedges between prices and the tradi-

tional marginal costs. Consequently, the price–quantity relation showing a firm’s will-

ingness to supply comprises the additional cost schedule ridden over the traditional

supply curve. In this paper, to avoid confusion, we refer this extended supply curve to

‘‘offer curve.’’ It slopes upward in the conventional price–quantity coordinate, reflecting

the principle that the higher cost a firm must bear, the higher price it wants to charge.Let

us call this generalized principle of marginal cost pricing ‘‘effective cost channel,’’ or, in

short, ‘‘cost channel.’’ Indeed, this channel of cost compensation has been known in the

various contexts: for example, Prescott (1975), Carlton (1979), Eden (1990), Rotemberg

and Summers (1990), and Greenwald and Stiglitz (1993).

Loss-balancing motive means to countervail higher business losses likely to incur

when producing more. This motive leads firms to lower prices when increasing produc-

tion. According to the cost channel stated above, firms will associate higher production

levels with higher prices to the extent that increased production implies a rise in the

odds of excess supply over excess demand. However, higher production levels lower

the odds of excess demand at the same time. Profit-maximizing firms will choose pro-

duction levels to balance the probabilities of being in the excess demand and excess

supply states, with the relative costliness of the two states that depend on the price of

output sold and the price of output not sold. Increasing the price of output sold makes

the excess demand state more appealing relative to the excess supply state, leading the

firms to lower production levels. This results in a negative price–quantity relationship.

Let us refer it to a ‘‘hedge curve,’’ and call the generalized principle of loss-balancing pric-

ing ‘‘hedging channel.’’ This hedging channel would naturally arise where firms access to

incomplete spanning of assets over their stochastic goods market outcomes. However,

1. As is named after the metaphorical example of a street-corner newsstand, a typical price-setting

newsvendor model assumes that firms choose prices and order stocks beforethe realization of market demand.

See Petruzzi and Dada (1999) for a review of the newsvendor literature.