NBER WORKING PAPER SERIES
THE IMPACT OF TRADE ON INTRA-INDUSTRY REALLOCATIONS
AND AGGREGATE INDUSTRY PRODUCTIVITY
Marc J. Melitz
Working Paper 8881
http://www.nber.org/papers/w8881
NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachusetts Avenue
Cambridge, MA 02138
April 2002
The author thanks, without implicating, Alan Deardor, Jim Levinsohn, and Elhanan Helpman for helpful
comments and discussions. Funding from the Alfred P. Sloan Foundation is gratefully acknowledged. The
views expressed herein are those of the author and not necessarily those of the National Bureau of Economic
Research.
© 2002 by Marc J. Melitz. All rights reserved. Short sections of text, not to exceed two paragraphs, may
be quoted without explicit permission provided that full credit, including © notice, is given to the source.
The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity
Marc J. Melitz
NBER Working Paper No. 8881
April 2002
JEL No. F1
ABSTRACT
This paper builds a dynamic industry model with heterogeneous firms that explains why
international trade induces reallocations of resources among firms in an industry. The paper shows how
the exposure to trade will induce only the more productive firms to enter the export market (while some
less productive firms continue to produce only for the domestic market) and will simultaneously force
the least productive firms to exit. It then shows how further increases in the industry's exposure to trade
lead to additional inter-firm reallocations towards more productive firms. These phenomena have been
empirically documented but can not be explained by current general equilibrium trade models, because
they rely on a representative firm framework. The paper also shows how the aggregate industry
productivity growth generated by the reallocations contributes to a welfare gain, thus highlighting a
benefit from trade that has not been examined theoretically before. The paper adapts Hopenhayn's (1992a)
dynamic industry model to monopolistic competition in a general equilibrium setting. In so doing, the
paper provides an extension of Krugman's (1980) trade model that incorporates firm level productivity
differences. Firms with different productivity levels coexist in an industry because each firm faces initial
uncertainty concerning its productivity before making an irreversible investment to enter the industry.
Entry into the export market is also costly, but the firm's decision to export occurs after it gains
knowledge of its productivity.
Mark J. Melitz
Department of Economics
Littauer Center
Harvard University
Cambridge, MA 02138
and NBER
mmelitz@harvard.edu
1 In troduction
Recent empirical research using longitudinal plant or Þrm-level data in several countries has ov er-
whelmingly substan tiated the existence of large and persistent productivit y differences among es-
tablishments in the same narrowly deÞned industries. Foster, Haltiwanger and Krizan (1998)
summarize this research by concluding that “... within sector differences dwarf betw een sector
differences in behavior.” In related work, Haltiwanger (1997, Table 1) reports that 4-digit industry
effectsexplainlessthan10percentoftheoverallvariationinthegrowthratesofoutput,em-
ployment, capital stocks, and productivity across establishments in the U.S. from 1977 to 1987.
Complementing this evidence on the extent of within sector heterogeneit y, other studies hav e sho w n
that the bulk of resource reallocations across Þrms remains internal to the speciÞc sector. Davis
and Haltiwanger (1999) summarize this evidence for the U.S. and report that less than 1 in 10 job
reallocations reßect employment shifts acro ss sectors. Levinsohn (1999) reports similar numbers for
most industries in Chile following wide-reaching trade liberalization. Evidence reported in Roberts
and T ybout (1996) and Davis and Haltiwanger (1999) conÞrms that these patterns are not speciÞc
to the U.S. and that substantial within sector reallocations between heterogeneous Þrms are also
prevalent in developing countries.
If these large intra-industry reallocations w ere unrelated to the heterogeneous characteristics of
Þrms, then their separate existence would not necessarily make them important determining factors
of industry performance. On the other hand, if the reallocations are related to Þrm characteristics,
then the nature of the link between the tw o signiÞcantly affects several important aspects of industry
performance. Although the analysis of this link between Þrm characteristics and industry evolution
is an ongoing research program, enough evidence has b een collected to demonstrate its existence and
relevance for industry performance. The main Þrm characteristic found to be empirically linked to
in tra-industry reallocations is Þrm productivit y.
1
The strongest evidence of this link pertains to Þrm
entry and exit decisions. Productivity differences between entering and exiting Þrms signiÞcan tly
cont ribute to aggregate industry productivit y c hanges over time. Additionally, a large number of
studies have documented a strong correlation bet ween Þrm exit and lo w productivit y (Þrm age
is also correlated with exit: younger Þrms have disproportionately high failure rates). Finally,
some studies have also found evidence that reallocations unrelated to entry and exit contribute to
1
Firm age and capital vintage are other importan t explanatory characteristics that hav e been highlighted in some
studies, although their impact may be limited to their effect on productivity.
1
industry productivity growth b y redistributing market shares among incumbent Þrms.
2
A similar
reallocation process has also been studied at a higher level of aggregation: Basu and Fernald
(1997) Þnd that U.S. aggregate productivit y changes across the business cycle are partly driven
b y expenditure reallocations across sectors with differen t average productivity lev els. The inherent
inabilit y of representative Þrm industry models to explain the contribution of reallocations to
industry performance has prompted the development of a theoretical literature of industry dynamics
that emphasizes the role of Þrm level heterogeneit y. This literature, along with the previously
mentioned empirical evidence, is reviewed in Foster, Haltiwanger and Krizan (1998) and Tybout
(2002).
This paper adapts one of these recen t industry models with heterogeneous Þrms in order to an-
alyze the role of international trade as a catalyst for inter-Þrm reallocations within an industry. It
then describes how these reallocations affect both industry performance and welfare. The business
press often assumes the existenc e of this catalyst role of trade when describing how exposure to
trade has both enhanced the growth opportunities of some Þrms while simu ltaneously cont ribut-
ing to the down fall or “downsizing” of other Þrms in the same industry. Similarly, protection from
trade is reported to shelter inefficient Þrms. Rigorous empirical work has recently corroborated this
anecdotal evidence. Bernard and Jensen (1999a) (for the U.S.), Aw, Ch ung and Roberts (2000) (for
Taiw an), and Clerides, Lach and T ybout (1998) (for Colom bia, Mexico, and Morocco) all Þnd evi-
dence that the causation of the correlation between Þrm productivity and export status runs from
theformertothelatter: moreproductiveÞrms self-select into the export market. Aw, Chung and
Roberts (2000) also Þnd evidence suggesting that exposure to trade forces the least productive Þrms
to exit the industry (Þrms with higher productivity levels relativ e to the incumbent average exit
after the exposure to trade). Both of these selection effects (into the export market and out of the
industry) obviously reallocate market shares from less productive Þrms(whoexit)tomoreproduc-
tive ones (who export) and therefore contribute to industry productivity growth.
3
P av cnik (2002)
directly looks at the contribution of market share reallocations to sectoral productivity growth
follow ing trade liberalization in Chile. She Þnds that these reallocations signiÞcan tly contribute to
productivity growth in the tradable sectors. In a related study, Bernard and Jensen (1999b) Þnd
that within-sector mark et share reallocations towards more productive exporting plants accounts
2
The importance of this phenomenon varies across studies and is cyclically sensitive (see Foster, Haltiwanger and
Krizan (1998))
3
Forces other than trade also affect the reallocation of resources within an industry. Olley and Pakes (1996) Þnd
that deregulation in the U.S. telecommunications industry increased productivity predominantly through this channel
rather than through intra-Þrm productivity gains.
2
for 20% of U.S. manufacturing productivity gro wth.
By relying on a representative Þrm framew ork (at least at the level of the industry), general
equilibrium trade models have largely ignored these intra-industry reallocations and focused instead
on other consequences of trade, such as inter-industry reallocations or phenomena affecting all Þrms
in similar ways.
4
This paper attempts to Þll this gap by providing a general equilibrium model
with heterogeneous Þrms that explains how trade induces these selection effects and in tra-industry
reallocations. This model shows how exposure to trade will induce only the more productive Þrms
to enter the export markets (while some less productive Þrms continue to produce only for the
domestic market) and will sim u ltaneously force the least productive Þrms to exit. The paper then
shows how further increases in the industry’s exposure to trade (drive n either by trade liberalization
or the addition of new trading partners) lead to additional inter-Þrm reallocations towards more
productive Þrms. The model th u s explains how trade can generate industry productivity growth
without necessarily affecting in tra-Þrm efficiency. It also provides a theoretical foundation for
the recent empirical Þndings described above and rigorously shows how trade can contribute to
the Darwinian evolution of industries — forcing the least efficient Þrms to con tract or exit while
promoting the growth and success of the more efficient ones.
Another recen t paper by Bernard, Eaton, Jenson and Kortum (2000) also introduces Þrm-level
heterogeneity into a model of trade b y adapting a Ricardian model to Þrm-speciÞccomparative
advantage. Both papers predict the same basic kinds of trade-induced reallocations, although the
ch annels and motiv ations behind these reallocations vary. In addition, Bernard et al. (2000) sho w
how their model can be calibrated to provide a good ÞttoacombinationofmicroandmacroUS
data patterns. Ho wev er, subsequent work b y Brooks (2001) has sho wn that this feature is not
robust across countries. The current paper relies on the previously docume nted empirical micro
patterns and focuses on the theoretical explanations and motivations behind these patterns.
4
This last category includes models that assume a direct link between trade and Þrm lev el efficiency. In these
models, exposure to trade typically increases the efficiency level of all Þrms through a variety of channels: learning
effects, increased scale of production, increased innovation, higher quality or diversity of intermediate inputs, reduction
of agency problems between owners and managers. Clerides, Lach and Tybout (1998) and Bernard and Jensen (1999a)
speciÞcally test whether new exporting Þrmsbecomemoreefficient. Neither of these studies Þnds evidence supporting
this hypothesis. Tybout and Westbrook (1995) test and reject the hypothesis that increased productivity in Mexico’s
growing export industries was driven by increases in the scale of plant production.
3
