The Impact of Intra-Industry Trade on the Environmenteconweb.ucsd.edu/~carsonvs/papers/889.pdfLast...
Transcript of The Impact of Intra-Industry Trade on the Environmenteconweb.ucsd.edu/~carsonvs/papers/889.pdfLast...
Last Revision: November 4, 2000
The Impact of Intra-Industry Trade on the Environment
Abstract
Empirical evidence suggests that international trade ‡ows take increasingly the form of intra- ratherthan inter-industry trade. So far, however, the literature has not addressed the environmental impactof trade liberalization in the context of two-way trade. The purpose of this paper is to …ll this gap andto show that under oligopolistic competition, the environmental e¤ects of trade di¤er signi…cantly fromthose predicted under standard inter-industry trade models. We show that environmental deteriorationis determined largely by which country liberalizes trade and by the nature of pollution. Implications forFDI and anti-trust policies are also discussed.
JEL classi…cation : F12, F18, F23
Keywords : Environment; Intra-industry trade; Foreign direct investment; Trade liberalization.
We are grateful to Brian Copeland for detailed comments on an earlier draft. Comments from Kar-Yiu Wong, Keith Markusen and Larry Goulder as well as from participants at the ”World Trade and theWTO Conference”, University of Washington, and at the ”SCCIE Workshop on Emerging Trade Issuesat the WTO”, University of California at Santa Cruz, greatly improved the quality of the present paper.Funding from the UCSC Social Science Division Research Grant is gratefully acknowledged. Any errorsand omissions are of course our own.
K.C. FungUniversity of CaliforniaEconomics Department217A Social Sciences 1Santa Cruz, CA 95064USAEmail: [email protected]
Andréa M. MaechlerSwiss National BankBörsenstrasse 15P.O. BoxCH - 8022 ZurichSwitzerlandEmail: [email protected]
1 Introduction
Recent years have witnessed two interesting developments in trade theory, namely, the emergence
of intra-industry trade as a widespread phenomenon and a revitalization of the concern about
the environmental consequences of international trade.1 By providing a link between these
two strands of the literature, the present paper departs from the standard inter-industry trade
approaches used in the trade-environment debate and o¤ers new perspective to revisit this issue
in the context of intra-industry trade.
The concept of intra-industry trade emerged when it became apparent that much of the trade
growth experienced over the last …fty years is attributable to an increase in two-way trade within
the same industry. This is particularly true for the manufacturing sector, among which rank some
of the most pollution-intensive industries, including chemicals, metallurgy, mining, and paper
and pulp. Bernhofen (1999), for example, has tested and supported the statistical evidence that
a signi…cant share of international trade in certain highly polluting industries, such as the coal
market, takes the form of two-way trade.
For illustration, Table 1 presents indexes of intra-industry trade for selected industries in 1998.
While the …rst column presents intra-industry trade indexes for the U.S., the second column
presents those for Western Europe. These two sets of data allow to illustrate the extent to which
intra-industry trade is not a U.S.-speci…c phenomenon but rather a widespread characteristic of
international trade. An index close to one (such as 0.98 for machinery and transport equipment
in Western Europe) indicates that for this industry Western Europe imports almost as much as
it exports, while an index close to zero (such as 0.30 for clothing in the U.S.) indicates that for
1 For excellent surveys on the environment-trade nexus, see Anderson and Blackhurst (1992), Beghin and Potier(1997), Beghin, Roland-Holst and van der Mensbrugghe (1995) and Low (1992). See Bhagwati and Srinivasan (1997)for critical discussions on the international harmonization of environmental standards. Finally, Rauscher (1997)o¤ers an insightful chapter on the interlinkages between trade and the environment in an imperfect competitionframework. Although the main focus of his work considers how environmental policies can be used to achievetrade-related policy objectives, he introduces a monopolistic competition model to sketch some of the environmentalconsequences of intra-industry trade.
1
this industry, the U.S. is only an exporter or an importer.2 Overall, although the average intra-
industry trade indexes are high in both regions, it is higher in Western Europe than in the U.S.
(respectively, 0.96 and 0.85). While earlier empirical work focuses on industrial countries, more
recent studies suggest that this phenomenon is not limited to industrialized countries and can be
signi…cant in developing countries.3 In fact, a new consensus is emerging from recent experience
suggesting that trade liberalization has been a powerful vehicle for the growth of intra-industry
trade in a wide range of countries.4
The observation that an increasing share of trade expansion is taking the form of intra- rather
than inter-industry trade is of particular importance when examining the far-reaching environmen-
tal consequences of trade liberalization. For example, the literature recognizes that the adjustment
costs associated with shifting resources across industries (i.e., from import-competing industries
to export-competing industries) are generally much larger than those associated with switching
resources across di¤erent product lines within the same industry.5 Following the same argument,
trade liberalization is also likely to yield very di¤erent e¤ects on the environment depending on
whether the underlying structure of trade leads to inter- or intra-industry trade expansion. This
suggests that, in the presence of intra-industry factor movements, the interactions between inter-
national trade and the quality of the environment need to be analyzed in their own right, that is,
independently of what occurs in the presence of inter-industry factor movements.
So far, however, the question of international trade and the environment has not been addressed
in the context of intra-industry trade. Thus, the purpose of this paper is to …ll this gap in the
2 For a recent empirical study of intra-industry trade in pollution-intensive homogeneous products, see therecent work by Bernhofen (1999). This paper uses the Brander (1981) model of intra-industry trade in identicalcommodities to conduct an empirical investigation in the context of the petrochemical industry in Germany andthe United States.
3 For a comprehensive survey of the literature on intra-industry trade, see Greenaway and Milner (1986),Greenaway and Torstensson (1997) and Helpman (1987). Tharakan (1984) provides evidence on intra-industrytrade between OECD countries and the developing world.
4 For example, Hoekman and Djankov (1996) presents evidence for trade liberalization-induced intra-industrytrade expansion in Eastern Europe, Lee and Lee (1993) in Korea and Menon (1994) in Australia.
5 Krugman (1981), for example, shows that in the context of intra-industry trade it is possible for all factors ofproduction to gain, which can alleviate adjustment costs. See Menon and Dixon (1996, 1997) for a general discussionon factor market disruption under inter- versus intra-industry trade.
2
literature and to add elements of intra-industry trade to the current debate on trade and the
environment.
[Table 1]
Broadly, the literature on the environmental consequences of international trade distinguishes
between three e¤ects, namely, the scale e¤ect that always increases pollution (trade induces more
output, which means more pollution); the technique e¤ect that always decreases pollution (trade
makes cleaner technologies available, which reduces pollution) and the composition e¤ect (trade
alters the pattern of production).6 To isolate the link between intra-industry factor relocation
and environmental damage, we put aside both, the technique and the scale e¤ects and focus on
the third and largely contentious e¤ect, that is, the composition e¤ect of trade.7
The composition e¤ect of trade relies on the theory of comparative advantage to predict how
trade alters the composition of national output and thereby, the quality of the environment.8
For example, trade may deteriorate the quality of the environment if it causes, say, the relatively
capital-abundant countries to specialize in the (dirty) capital-intensive goods.9 Trade may also
deteriorate the quality of the environment if it induces the dirty industries to relocate away from
the industrialized countries imposing tougher environmental standards towards the lower-income
countries requiring lower environmental standards.10 According to these arguments, global trends
6 This decomposition is …rst introduced in Grossman and Krueger (1993, 1995), elaborated in Copeland andTaylor (1994, 1995a, 1995b) and then tested in Antweiler, Copeland and Taylor (1998). Note that the scale and thetechnique e¤ects can incorporate the income e¤ect, which is discussed in Copeland (1996), Copeland and Taylor(1999) and Grossman and Krueger (1993).
7 The technique e¤ect relies on a very di¤erent set of theoretical premises dealing with the role of technologicalprogress. The scale e¤ect requires a general equilibrium analysis, while our model, which follows the tradition ofthe new oligopoly literature, is set in a partial equilibrium framework. It is important to keep in mind, however,that whatever the environmental consequences of trade via the composition e¤ect are (which Antweiler, Copelandand Taylor (1998) …nds to be only slightly negative), the full impact of trade may be either positive or negative,depending on the relative strength of the (positive) technique and the (negative) scale e¤ects.
8 The theory of comparative advantage predicts that a country tends to export products that use intensivelythe factors of production with which it is relatively well endowed, and import those that use intensively its scarcefactors. With respect to the environment, the e¤ects of trade depend on whether a country becomes an exporteror an importer of pollution-intensive goods in the post-liberalization situation. In the absence of transboundarypollution, trade is good for the domestic environment in the former case, while it is bad in the latter case.
9 This hypothesis, which is refered to as the ”simple factor endowment” hypothesis, relies on factor-abundanceto explain the link between trade and the environment.
10 The literature refers to this hypothesis as to the ”pollution-haven” or ”dirty industry migration”.
3
in environmental degradation are explained by inter-industry factor relocation.
The present paper, however, o¤ers a very di¤erent approach to address the trade-environment
nexus. It focuses on intra-industry trade as the main culprit of environmental degradation. In this
context, the composition e¤ect of trade generates a very di¤erent kind of change in the pattern
of production, namely, one that alters an industry’s output mix across international oligopolies
rather than modifying a country’s commodity mix across industries. In particular, we …nd that
global trends in environmental deterioration are determined largely by two factors, namely: (i) by
which trading partner liberalizes trade, and (ii) by the nature of pollution (i.e., whether pollution
is local, transboundary or global). For example, our model predicts that the global environment
can bene…t from trade liberalization, provided that the most pollution-intensive country liberalizes
trade.
By restricting our analysis to the context of oligopolistic intra-industry trade, our model iso-
lates the source of trade to be determined solely by the strategic interactions of oligopolistic
polluting …rms and not by cost di¤erences and/or by economies of scale.11 In reality, of course,
the determinants of trade are manifold, including di¤erences in factor endowments, income lev-
els, technological capabilities and regulatory and institutional frameworks. But given the strong
empirical evidence in support of a substantial and ever-growing volume of intra-industry trade, it
remains useful to analyze the likely impact of international trade on the quality of the environment
in this limited context.
This paper is organized as follows. Section 2 presents our basic duopoly model of intra-
industry trade and derives the environmental consequences of trade liberalization under local,
transboundary and global pollution. Section 3 extends the basic model to verify the validity of
our results in the presence of multiple …rms. In Section 4, we examine how trade collusion alters
our basic results. In Section 4, we introduce foreign direct investment (FDI) to discuss the impact
11 Note, however, economies of scale would be a key determinant of environmental deterioration in the contextof monopolistic competition.
4
of trade liberalization on the quality of the environment when one of the trading partner engages
in FDI activities. Finally, concluding comments are provided in Section 5.
2 Basic Model
2.1 Framework
We develop a standard intra-industry trade model in the tradition of Brander (1981) and Brander
and Krugman (1983). There are two countries, a home country and a foreign country, each
holding one polluting …rm. Consumers consume goods which are produced both at home and
abroad. Because most intra-industry trade occurs between similar economies, it is safe to assume
that the two countries have similar tastes and endowments. The home (foreign) country produces
a pollution-intensive good which is sold in the home (foreign) market and which can also be
exported to the foreign (home) market. The pro…t function of, respectively, the home and the
foreign …rm is written as:
¼x = x1px1 + x2px2 ¡ ex(x1 + x2) ¡ txx2 (1)
¼y = y1py1 + y2py2 ¡ ey(y1 + y2) ¡ tyy1 (2)
where xi (i = 1; 2) represents a pollution-intensive good that is produced in the home country and
yi (i = 1; 2) represents a pollution-intensive good that is produced in the foreign country. The
subscripts indicate whether a good is sold in the home market (1) or in the foreign market (2).
For example, while the …rst term in equation (1) represents the domestic sales of the home
…rm, the second term represents its export sales. The third term represents the home pollution tax
ex; which the home country imposes on all home-produced pollution-intensive goods, regardless
of whether they are sold at home or abroad. Finally, the fourth term represents the foreign
import tari¤ tx; which the home country faces when exporting its pollution-intensive production
abroad. Similarly for the foreign pro…t function in equation (2), where the …rst term represents
its export sales to the home country, the second term represents its domestic sales to the foreign
5
market, the third term represents the foreign pollution tax ey imposed on the foreign-produced
pollution-intensive good and the fourth term represents the home import tari¤ ty imposed on
foreign exports.
To highlight the environmental e¤ects of trade liberalization, we assume zero production and
transportation costs and focus solely on pollution taxes and/or trade protection. While the as-
sumption of zero production and transportation costs greatly simpli…es the interpretation of the
results, it does not a¤ect its general conclusions.
We assume a symmetric linear demand structure and allow for di¤erentiated goods (that is,
we assume ¯1 = ¯2 6= °)12 :
pxi = ®1 ¡ ¯xi ¡ °yi
pyi = ®2 ¡ ¯yi ¡ °xi
where ®i (i = 1; 2) > 0, ¯ > 0, and ° > 0. Assuming that the …rms are Cournot-Nash players,
the …rst-order pro…t maximization conditions yield the following outputs:
xCN1 =
2¯(®1 ¡ ex) ¡ °(®2 ¡ ey ¡ ty)D
(3)
xCN2 =
2¯(®1 ¡ ex ¡ tx) ¡ °(®2 ¡ ey)D
(4)
yCN1 =
2¯(®2 ¡ ey ¡ ty) ¡ °(®1 ¡ ex)D
(5)
yCN2 =
2¯(®2 ¡ ey) ¡ °(®1 ¡ ex ¡ tx)D
(6)
where D =¡4¯2 ¡ °2
¢: Note that the concavity assumption of the aggregate utility function
Ui associated with the above demand functions implies that (¯2 ¡ °2) > 0. Furthermore, we
assume that the own-price e¤ects are greater than the cross-price e¤ects, that is, ¯ > °.
12 This demand structure is associated with the following aggregate utility function: Ui = wi + mi wherewi =
©®1xi + ®2yi ¡ 1
2
£¯(x2i + y
2i ) + 2°xiyi
¤ªrepresents the expenditure on xi and yi and mi the expenditures
on all other goods in country i.
6
2.2 Conditions for Intra-Industry Trade
The existence of intra-industry trade requires that xCN2 > 0 and yCN
1 > 0. Thus, the conditions
for intra-industry trade can be expressed as:
xCN2 > 0 iff
2¯°
>(®2 ¡ ey)
(®1 ¡ ex ¡ tx)
yCN1 > 0 iff
2¯°
>(®1 ¡ ex)
(®2 ¡ ey ¡ ty)
According to the above conditions, intra-industry trade is more likely to occur in the presence
of: (i) greater product di¤erential (high ¯=°); (ii) low import tari¤s, ti; and (iii) lower envi-
ronmental standards imposed on the exporting country i (high ®i ¡ ei), where i represents the
non-liberalizing country.13
2.3 Unilateral Trade Liberalization
Assume that the home (foreign) country implements a unilateral trade liberalization, such that
dty < 0 (dtx < 0): The impact of such a trade policy on the demand for pollution-intensive goods
is given in the following equations:
@xCN1
@ty=
°D
> 0 (7)
@yCN1
@ty= ¡2¯
D< 0 (8)
For example, by unilaterally liberalizing trade, the home country facilitates the access of foreign
pollution-intensive goods into the home market. Given our earlier assumption that the own-price
e¤ects exceed the cross-price e¤ects (i.e., ¯ > °), the increase in imports exceeds the fall in the
demand for the domestically produced good (i.e.,¯̄@yCN
1 =@ty¯̄>
¯̄@xCN
1 =@ty¯̄). Thus, on aggregate,
13 The results in the present paper are derived under the assumption that the pollution-intensive goods aredi¤erentiated goods (i.e., ¯1 = ¯2 6= °): Assuming homogenous goods (i.e., ¯1 = ¯ = °); however, would not altersigni…cantly our results, except that the conditions for intra-industry trade become stricter, requiring an even lowerimport tari¤ and/or an even higher environmental di¤erential between the two countries for our results to hold.
7
a unilateral trade always increases the domestic consumption of the pollution-intensive good and
reduces its domestic production.
Similarly for a foreign trade liberalization, which increases the demand for the imported good
by more than it decreases the demand for the domestically-produced polluting good, raising the
aggregate foreign consumption of the pollution-intensive good and reducing its foreign production.
@yCN2
@tx=
°D
> 0 (9)
@xCN2
@tx= ¡2¯
D< 0 (10)
In summary, a unilateral trade liberalization (dti < 0) raises the pollution-intensive output in
the non-liberalizing country and lowers it in the liberalizing country. Furthermore, the liberalizing
country increases its aggregate consumption of the pollution-intensive good.
Although these results set the stage for our analysis, we need to go one step further and to
model how the volume e¤ects of trade liberalization presented in equation (7)-(10) above translate
into the environmental impact of trade liberalization. This is what we do next.
2.4 Environmental Damage
In this section, we focus on how a unilateral trade liberalization a¤ects the environment in a world
where the quality of the environment is a by-product of the production of pollution-intensive
goods.14 To tackle this question, we assume that pollution harms consumers without a¤ecting
the production possibility frontier and that the source of pollution may lie outside of national
boundaries. We di¤erentiate between three types of pollution, namely: (i) local pollution; (ii)
transboundary pollution; and (iii) global pollution.15
14 Although most of the existing theoretical work focuses on production-generated pollution, Copeland andTaylor (1995, 1999) examine, respectively, pollution as a by-product of consumption and pollution as it a¤ectsproductivity.
15 Although there is an extensive literature on transboundary pollution (see Dean 1992 for a survey), there islittle work examining the environmental consequences of trade in an imperfect competitive framework and evenless in an intra-industry setting (see Rauscher 1997).
8
Although there exists no strict line delimiting local pollution from transboundary or even
global pollution, the impact of various types of production-generated pollutions on the environ-
mental quality may be very di¤erent in scope. When pollution is local, a country’s environmental
damage is determined entirely by its domestic polluting production. Examples of local pollution
abound and include urban air pollution, local soil erosion and local contamination of water, air
and land. When pollution is transboundary, environmental damage spills across national bound-
aries and the quality of a country’s environment is determined by the sum of two components,
namely, the pollution generated by its own domestic polluting production and a fraction of the
pollution generated during the polluting production abroad. Acid rain, for example, is often cited
as an illustration of transboundary pollution. The production of certain pollution-intensive goods
releases pollutants in the air, which are then carried by winds hundreds of miles away from their
production site before being deposited again in the atmosphere through rain, fog or snow. In
this situation, a country cares about the type and the quantity of pollution generated abroad.
Finally, when pollution is global, a country’s environmental damage is determined by the aggre-
gate pollution-intensive production, regardless of the country of production. Examples of global
pollution include the deterioration of the ecosystem, the depletion of the ozone layer and global
warming.
The environmental damage of, respectively, the home and the foreign country, can be written
as:
Edx = dx(x1 + x2) + ´ydy(y1 + y2)
Edy = dy(y1 + y2) + ´xdx(x1 + x2)
where di (i = x; y) represents the environmental damage associated with the pollution-intensive
goods produced, respectively, in the home (x) country and in the foreign (y) country. ´i represents
the pollution spillover parameter and illustrates the extent to which pollution generated in one
9
country damages the environment in the other country.16 At one extreme, ´i is zero and one
country’s pollution does not a¤ect the quality of the environment in the other country. This is the
case of local pollution. At the other extreme, ´i is equal to one and pollution produced in either
country a¤ects both countries equally. This is when pollution is global. Finally, if ´i lies somewhere
between zero and one, pollution is transboundary. In this case, no country is completely insulated
from the other country’s pollution and pollution incurred in one country spills over to the other
country. Since the pollution parameter is between zero and one, the damage that a country incurs
from its own pollution is greater than the damage it incurs from the pollution generated abroad.
To clarify our analysis, we assume that the home country is the ’clean’ country, in the sense
that it imposes a higher pollution tax than the foreign country, or, ex > ey. Furthermore,
we also assume that the stricter environmental standards induce the clean country to generates
less pollution than the foreign country or, dx < dy. These assumptions allow us to focus on a
particular scenario, without compromising the generality of our results. Henceforth, we will use
interchangeably the terms of home and clean country on the one hand, and of foreign and dirty
country on the other hand.
We can turn now to the e¤ects of a unilateral trade liberalization on the environment.
2.4.1 Local Pollution
When pollution is local, the environmental damage of, respectively, the home and the foreign
country, is given by:
Edlx = dx(x1 + x2)
Edly = dy(y1 + y2)
16 While there is an extensive literature on transboundary pollution, it is mostly concerned with how environmen-tal regulations a¤ect trade, rather than how trade a¤ects the quality of the environment (Siebert, 1977 and Baumoland Oates, 1988). See Copeland and Taylor (1993, 1995a, 1995b) for a discussion on the role of transboundarypollution in the debate over trade and the environment.
10
Clearly, under local pollution, there is a direct link between a country’s level of polluting pro-
duction and the quality of its environment: a country raises (lowers) the quality of its environment
if it decreases (increases) its pollution-intensive production. The e¤ects of a unilateral trade lib-
eralization by the clean country on the environmental damage of, respectively, the clean and the
dirty country, are given by:
@Edlx
@ty=
°dx
D> 0
@Edly
@ty= ¡2¯dy
D< 0
We have seen above that a unilateral trade liberalization raises the polluting output in the
non-liberalizing country and reduces it in the liberalizing country. Thus, a trade liberalization by
the clean country transfers the production of the pollution-intensive good away from the clean
country towards the dirty country. This lowers the environmental damage in the clean country
and raises it in the dirty country, or, put di¤erently, this raises the quality of the environment in
the clean country and deteriorates in the dirty country.
Similarly, if the dirty country liberalizes trade, the production of the pollution-intensive good is
shifted away from the dirty country towards the clean country. Thus, a unilateral trade liberaliza-
tion by the dirty country lowers the environmental damage in the dirty country (which produces
now less of the pollution-intensive good) and raises it in the clean country (which produces now
more of the pollution-intensive good). This can be seen in the following equations:
@Edlx
@tx= ¡2¯dx
D< 0
@Edly
@tx=
°dy
D> 0
These equations highlight one very interesting point, namely, that a unilateral trade liberal-
ization actually improves the quality of the environment in the country that liberalizes trade,
regardless of which country is liberalizing. This is a very di¤erent outcome than the one generally
11
predicted under standard inter-industry trade models, where comparative advantage generally
drives the economic consequences of trade liberalization. This result can be formalized in the
following terms.
Proposition 1 Assuming that pollution is local, a unilateral trade liberalization improves thequality of the environment in the liberalizing country and deteriorates it in the non-liberalizingcountry. This proposition holds regardless of whether the clean country or the dirty country isliberalizing trade in the pollution-intensive goods.
2.4.2 Transboundary Pollution
In the presence of transboundary pollution, the environmental damage of, respectively, the home
and the foreign country, can be written as:
Edtx = dx(x1 + x2) + ´ydy(y1 + y2) (11)
Edty = dy(y1 + y2) + ´xdx(x1 + x2) (12)
where 0 < ´i < 1 (for i = x; y): Because now pollution produces in one country spills over to
the other country, the e¤ects of a unilateral trade liberalization by the home country on the
environmental damage of, respectively, the home and the foreign country, are given by:
µ@Edt
x
@ty
¶CN
=1D
©°dx ¡ 2¯´ydy
ª(13)
µ@Edt
y
@ty
¶CN
=1D
f°´xdx ¡ 2¯dyg (14)
To simplify their interpretation, these equations can be re-written in the following conditional
expressions:
@Edtx
@ty> 0 iff
dx
´ydy>
2¯°
(15)
@Edty
@ty> 0 iff
dy
´xdx<
°2¯
(16)
For notational purposes, we refer to the left-hand side of the above conditional statements as
country i’s ”relative pollution di¤erential” e¤ect. This e¤ect is measured by the pollution borne
12
by country i due to its own pollution damage over the pollution it bears due to the pollution
spillover from country j. On the right-hand side, we de…ne 2¯=° as the ”product di¤erential”
e¤ect, which is always positive according to our earlier assumption regarding the magnitude of
the own- and cross-price e¤ects.
We have seen that a unilateral trade liberalization by the clean country shifts the pollution-
intensive production away from the clean country towards the dirty country. As long as the home
country does not bear a large fraction of the foreign pollution (i.e., as long as the pollution spillover
from abroad, ´y; is not too large), the clean country (which bene…ts from lower production and
hence, lower domestic pollution) improves the quality of its environment by unilaterally liberal-
izing trade. The quality of the environment in the foreign country, however, always deteriorates
following a unilateral trade liberalization by the home country. This can be seen from equation
(16), where the conditional statement never holds.
Assuming that the home relative pollution di¤erential exceeds the product di¤erential, we can
summarize the results as follows:
Proposition 2 Assuming that pollution is transboundary, the clean country can improve its en-vironment by unilaterally liberalizing trade. But the higher environmental quality in the cleancountry is always achieved at the expense of a deterioration in the environmental quality of thedirty country.
So far, we have considered the case where the clean country liberalizes trade. However, once
pollution is no longer local, the e¤ects of a trade liberalization on the quality of the environment
are sensitive to which country is liberalizing trade. Thus, we turn now to the case where the dirty
country is liberalizing trade.
@Edtx
@tx> 0 iff
´ydy
dx>
2¯°
@Edty
@tx> 0 iff
dy
´xdx>
2¯°
A foreign unilateral trade liberalization shifts the pollution-intensive production away from
the dirty country towards the clean country. Again, as long as the foreign country does not bear
13
a large fraction of the home pollution (i.e., as long as the pollution spillover parameter from
home, ´x; is su¢ciently small), the dirty country can improve the quality of its environment by
unilaterally liberalizing trade.
The interesting point here is that when the foreign country liberalizes trade, the clean country
does not necessarily su¤er from higher pollution in the post-liberalization equilibrium. Indeed,
the quality of its environment can improve, even though it ends up producing a greater quantity
of the polluting good. This is possible if its new level of clean production generates less pollution
than that borne previously under the foreign pollution spillover. For this to be true, dx must be
relatively low and ´ydy relatively large.
Again, provided that the foreign relative pollution di¤erential exceeds the product di¤erential,
we can summarize these results as follows:
Proposition 3 Assuming that pollution is transboundary, the dirty country can improve simul-taneously its own environment and that of the clean country by imposing a unilateral trade liber-alization in the pollution-intensive goods. This proposition holds when the production process inthe dirty (clean) country generates a su¢ciently large (small) amount of pollution spillover.
2.4.3 Global Pollution
When pollution is global, the environmental damage is determined by the aggregate production
of pollution-intensive good (regardless of whether it is produced at home or abroad) and it is
identical in both countries. This global environmental damage can be expressed as:
Edg = dx(x1 + x2) + dy(y1 + y2)
According to this expression, the e¤ects of, respectively, a home and a foreign trade liberaliza-
tion on the global environmental damage are given by:
@Edg
@ty> 0 iff
dx
dy>
2¯°
(17)
@Edg
@tx> 0 iff
dy
dx>
2¯°
(18)
14
As before, we refer to the right-hand side of these conditional statements as the product
di¤erential e¤ect. But in the presence of global pollution, we refer to the left-hand side as the
”absolute pollution di¤erential” (equation (17) refers to the home absolute pollution di¤erential,
while equation (18) refers to the foreign absolute pollution di¤erential).
By unilaterally liberalizing trade, the clean country transfers the production of the pollution-
intensive good away from the clean country towards the dirty country. Since the rise in dirty
production exceeds the reduction in clean production, the unilateral trade liberalization by the
clean country always raises the global environmental damage. This is expressed in equation (18),
where the conditional statement never holds and hence, @Edg=@ty < 0.
This result may be reversed if the dirty country reduces its trade barriers, in which case a
unilateral trade liberalization shifts the pollution-intensive production away from the dirty country
towards the clean country, possibly raising the quality of the environment. While the aggregate
production is higher in the post-liberalization equilibrium (which tends to support higher global
pollution), the rise in the clean production exceeds the reduction in the dirty production. Thus,
with a su¢ciently large pollution di¤erential, the dirty country can still improve the quality of
the environment by unilaterally liberalizing trade.
Assuming that the foreign relative pollution di¤erential exceeds the product di¤erential, these
results can be summarized as follows.
Proposition 4 Assuming that pollution is global, a unilateral trade liberalization by the cleancountry deteriorates unambiguously the global environment. But a unilateral trade liberalizationby the dirty country can still improve the quality of the global environment, provided that thepollution di¤erential is large.
In this section, we show that when the driving forces of the environmental damage associ-
ated with trade expansion are fueled by oligopolistic competition, there is no justi…cation for the
traditional inter-industry factor-relocation hypotheses. Indeed, according to our model, the envi-
ronmental consequences of trade liberalization depend largely on two factors, namely, on which
country initiates trade liberalization and on the nature of pollution (as opposed to comparative
15
advantage). For example, under local pollution, the liberalizing country always bene…ts from trade
liberalization, while the non-liberalizing country always su¤ers from it, regardless of whether it
is the clean country or the dirty country that liberalizes trade. This result, however, depends
strongly on the nature of pollution. As pollution crosses borders, the global economy is better-o¤
when the dirty country liberalizes trade than when the clean country does so. These results sup-
port the inconclusive empirical evidence found in support of the inter-industry e¤ects of trade.17
2.5 Multilateral Trade Liberalization
So far, we have discussed the e¤ects of a unilateral trade liberalization on the quality of the
environment. Assume now that the home and the foreign countries implement a multilateral
trade liberalization, such that dtx = dty = d¿ < 0: Under this scenario, it is easy to show that a
trade liberalization always a¤ects negatively the quality of the environment, regardless of whether
pollution is local, transboundary or global. The main reason for this results is that a multilateral
trade liberalization raises the domestic production geared towards exports by more than it reduces
the domestic production geared towards domestic consumption. This is illustrated in the following
expressions:
@xMTL1
@¿+
@xMTL2
@¿=
° ¡ 2¯D
< 0
@yMTL1
@¿+
@yMTL2
@¿=
° ¡ 2¯D
< 0
Thus, a multilateral trade liberalization raises each country’s overall pollution-intensive pro-
duction, and hence, deteriorates each country’s environmental quality. This is shown in the
following equations:
@Edx
@¿=
1D
£(dx + ´ydy)(° ¡ 2¯)
¤< 0 (19)
17 For a recent survey on the empirical literature, see Rauscher (1997). An excellent collection of papers isprovided in Low (1992), where various aspects of this issue are discussed and reviewed.
16
@Edy
@¿=
1D
[(´xdx + dy)(° ¡ 2¯)] < 0 (20)
According to equations (19) and (20), a multilateral trade liberalization always deteriorates
the quality of the environment, regardless of wether pollution is local (´i = 0), transboundary
(0 < ´i < 1) or global (´i = 1). But the magnitude of this deterioration, however, depends
positively on the size of the pollution spillover parameter, ´i.
These results can be summarized in the following terms:
Proposition 5 A multilateral trade liberalization always deteriorates the quality of the environ-ment in both, the clean and the dirty country. This proposition holds regardless of whether pollutionis local, transboundary or global, although its detrimental environmental e¤ects are smallest whenpollution is local.
Note, however, that although a multilateral trade liberalization raises the aggregate level of
pollution, it is important to keep in mind that the literature has long argued that di¤erent countries
have di¤erent pollution assimilation capacities. Thus, as long as the environmental degradation
borne by each country does not exceed its absorptive capacity, the positive variety and pro-
competition e¤ects of freer intra-industry trade may well outweigh the negative environmental
e¤ect.18
So far, we make the assumption that the polluting …rms are quantity-setting oligopolists. How-
ever, we know from standard textbooks of industrial organization that this is not the only possible
mode of competition. To test the robustness of our previous results, we have also considered the
case where the polluting …rms are price-setting producers, rather than quantity-setting producers.
It is easy to show that under standard assumptions, the presence of a price-setting competition
does not alter the results derived under a quantity-setting competition and all of the propositions
derived in Section 2.4 carry over to the case of Bertrand-Nash competition.
18 Excellent discussions on countries’ di¤erent pollution absorptive capacities can be found in Bhagwati andSrinivasan (1997) and Low and Safadi (1992). Note that while the authors highlight countries’ variable capacitiesto assimilate pollution in defense against an international harmonization of environmental regulations, they bothrecognize that not all pollution is unacceptably bad for the environment.
17
3 Multiple Firms
In this section, we extend the previous analysis and examine how our earlier results change when
each country holds a number of …rms producing the pollution-intensive good, as opposed of having
a single polluting …rm. In particular, we allow for m (n) …rms to produce good x (y), where
(m;n) > 1: Under this new speci…cation, the pro…t function of, respectively, the home and the
foreign …rm needs to be modi…ed as follows:
¼ix = x1px1(X1; Y1) + x2px2(X2; Y2) ¡ ex(x1 + x2) ¡ txx2
¼iy = y1py1(X1; Y1) + y2py2(X2; Y2) ¡ ey(y1 + y2) ¡ tyy1
where Xi = xi + (m ¡ 1)vi and Yi = yi + (n ¡ 1)zi (for i = 1; 2): While both x and v are
produced in the home country, good x is produced by a single oligopolistic …rm and good v is
produced by the remaining (m ¡ 1) oligopolistic …rms in the domestic industry: Similarly for the
foreign country, where good y is produced by a single foreign …rm, while good z is produced by
the remaining (n ¡ 1) foreign …rms in the industry: If we assume now that all the …rms in a given
country produce a homogenous good, we can now drop the inter-…rm product di¤erential and focus
on the international product di¤erential, such that xi = vi and yi = zi: As before, the subscripts
indicate whether a polluting good is sold in the home market (1) or in the foreign market (2).
In the presence of multiple …rms, the optimal outputs require:
xCN1 =
(n + 1)¯(®1 ¡ ex) ¡ n°(®2 ¡ ey ¡ ty)E
xCN2 =
(n + 1)¯(®1 ¡ ex ¡ tx) ¡ n°(®2 ¡ ey)E
yCN1 =
(m + 1)¯(®2 ¡ ey ¡ ty) ¡ m°(®1 ¡ ex)E
yCN2 =
(m + 1)¯(®2 ¡ ey) ¡ m°(®1 ¡ ex ¡ tx)E
18
where E =£(m + 1)(n + 1)¯2 ¡ mn°2
¤, which is assumed to be positive.19
3.1 Conditions for Intra-Industry Trade
The existence of two-way trade requires that xCN2 > 0 and yCN
1 > 0. Thus, the conditions for
intra-industry trade can be expressed as:
x2 > 0 iffµ
n + 1n
¶¯°
>(®2 ¡ ey)
(®1 ¡ ex ¡ tx)
y1 > 0 iffµ
m + 1m
¶¯°
>(®1 ¡ ex)
(®2 ¡ ey ¡ ty)
Similar to the conditions derived under the basic case in Section 2, intra-industry trade is more
likely to occur in the presence of: (i) greater product di¤erential (high ¯=°); (ii) low import tari¤s,
ti; and (iii) lower environmental standards imposed on the exporting country i (high ®i¡ei). Note,
however, that since (q + 1) =q < 2 (for q = m;n) the conditions for intra-industry trade are more
binding in the m £ n case than under the basic case. Thus, when the polluting good is produced
by multiple …rms rather than by a single …rm in each country, at least one of the three conditions
stated above must be stronger for intra-industry trade to occur and hence, for our analysis to be
relevant.
3.2 Unilateral Trade Liberalization
The impacts of a trade liberalization on the demand for pollution-intensive goods are given by the
following equations:
@xCN1
@ty=
n°E
> 0 (21)
@yCN1
@ty= ¡(m + 1)¯
E< 0 (22)
19 When comparing the optimal outputs derived in the presence of a single …rm (equations (7)-(10) in Section2.3) with those presented here in the presence of multiple …rms, we note that the denominator in the former issmaller the denominator in the latter, or, E > D for (m;n) > 1.
19
@xCN2
@tx= ¡(n + 1)¯
E< 0 (23)
@yCN2
@tx=
m°E
> 0 (24)
Clearly, the presence of multiple …rms does not alter the proposition stated in Section 2.3.
It is interesting to point out, however, that it is not clear whether the e¤ects of a unilateral
liberalization on the environment will be larger or smaller (in absolute value) in the presence of
multiple …rms than in the presence of a single domestic …rm. This will depend on the respective
magnitude of two con‡icting e¤ects. While, on the one hand, the determinant in the single …rm
case is smaller than the determinant in the multiple …rm case (E > D), on the other hand, the
numerator increases with the number of …rms (m;n). This can be seen by comparing equations
(7)-(9) with equations (21)-(24).
3.3 Environmental Damage
Assuming that each country has multiple …rms producing the pollution-intensive good, the impact
of a unilateral trade liberalization by the home country on the quality of the environment can be
summarized in the following expressions:
µ@Edx
@ty
¶m£n
=mnE
½°dx ¡
µm + 1
m
¶¯´ydy
¾(25)
µ@Edy
@ty
¶m£n
=mnE
½°´xdx ¡
µm + 1
m
¶¯dy
¾(26)
The various scenarios under local, transboundary or global pollution can be represented by sub-
stituting ´i with the relevant pollution spillover parameter (´i = 0 for local pollution, 0 < ´i < 1
for transboundary pollution and ´i = 1 for global pollution): Similarly in the case of a foreign
trade liberalization,
µ@Edx
@tx
¶m£n
=mnE
½°´ydy ¡
µn + 1
n
¶¯dx
¾(27)
20
µ@Edy
@tx
¶m£n
=mnE
½°dy ¡
µn + 1
n
¶¯´xdy
¾(28)
While the presence of multiple …rms does not alter signi…cantly the general conclusions derived
in Section 2.4, both the direction and the magnitude of the environmental e¤ects of a trade
liberalization can be di¤erent in the m £ n case than in the basic case, depending on the speci…c
parameter values of the model. In particular, the overall e¤ect will depend on the relative size of
two e¤ects:
(i) First, the second term in the brackets of equations (25)-(28) is smaller in the m £ n case
than in the basic case, because (q + 1)=q < 2 (for q = m;n): This means that, in the presence of
transboundary pollution, a trade liberalization is more likely to be good for the environment in
the presence of a large number of …rms than in the presence of a single …rm in the liberalizing
country. Note that the direction of the environmental impacts of a trade liberalization depends
only on the number of …rms in the liberalizing country and not on the number of …rms in the
non-liberalizing country.20
(ii) Second, the multiplier in front of the brackets in equations (21)-(24) is larger in the m £n
case than in the basic case (mn=E > 1=D). According to this e¤ect, whatever the direction of
the environmental e¤ects of a trade liberalization, these e¤ects are magni…ed in the presence of
multiple …rms. Thus, the presence of multiple …rms magni…es the impact of a trade liberalization
on the quality of the environment: If a trade liberalization is good (bad) for the environment, its
impact on the environment will be better (worse) under the m£n case than under the basic case.
These results can be summarized in the following proposition:
Proposition 6 Generally, the presence of multiple polluting …rms does not alter the direction ofthe environmental e¤ects of trade liberalization, although it magni…es their impact. Nevertheless,if pollution is transboundary, there exists a range of parameter for which a trade liberalization ismore likely to be good for the environment when a large number of polluting …rms (as opposed toa single …rm) are located in the liberalizing country.
20 This is because following, say, a home trade liberalization, the foreign country produces a higher level of thepollution-intensive production. In equilibrium, the extent to which the foreign country increases its productiondepends on the number of …rms in the home country. This is shown in equations (21)-(24).
21
Note the interesting implication of this observation in the context of competition policy. Ac-
cording to our model, anti-trust policies can be pro-environment. This statement is particularly
true for the non-liberalizing country and in the case of global pollution.
3.4 Multilateral Trade Liberalization
In the presence of multiple …rms, the e¤ect of a multilateral trade liberalization on the environ-
mental damage of country i is given by:
@Edx
@¿=
mnE
½µ° ¡
µn + 1
n
¶¯¶
dx +µ
° ¡µ
m + 1m
¶¯¶
´ydy
¾< 0
@Edy
@¿=
mnE
½µ° ¡
µn + 1
n
¶¯¶
´xdx +µ
° ¡µ
m + 1m
¶¯¶
dy
¾< 0
where ´i; the degree of pollution spillover, can be substituted with the appropriate parameter
value to illustrate the environmental damage under local, transboundary, and global pollution.
As in Section 2.5, a multilateral trade liberalization always raises the overall level of polluting
production, which necessarily deteriorates the quality of the environment. The number of …rms,
however, magni…es the detrimental environmental e¤ect of trade liberalization.
Proposition 7 As in the basic case presented in Section 2.5, a multilateral trade liberalizationalways deteriorates the quality of the environment in both, the clean and the dirty country. Thenumber of oligopolistic …rms, however, magni…es the detrimental environmental e¤ects of a mul-tilateral trade liberalization. This proposition holds regardless of whether pollution is local, trans-boundary or global, although the degradation of the environment is smallest when pollution is local.
4 In the Presence of Tari¤-Jumping FDI
4.1 Framework
When tari¤ barriers are high, rather than producing the pollution-intensive good at home and
exporting it abroad, a …rm may choose to serve the foreign market by engaging in foreign direct
investment (FDI) and setting up production plants abroad. This is often referred to as ’tari¤-
jumping’ FDI because it allows …rms to jump high trade barriers abroad. In this section, we
present the case where the home …rm engages in FDI activities in the foreign country, while
22
the foreign …rm continues to produce the pollution-intensive good at home and to export part
of its production to the home market. We have chosen this scenario because it illustrates an
interesting counter-example to the common ”dirty-industry migration” hypothesis, according to
which trade liberalization deteriorates the quality of the environment. This result is driven by the
assumption that trade liberalization induces the pollution-intensive industries to relocate away
from the industrialized countries with higher environmental standards towards the lower-income
countries with lower environment standards and hence, causing a convergence towards the lowest
common environmental standards. In our model, however, we show that in the presence of tari¤-
jumping FDI activities, trade liberalization may not necessarily deteriorate the quality of the
environment.
To duplicate the ”dirty-industry migration” hypothesis, we assume that the …rm that engages
in FDI adopts the production technology of the host country. For example, when the clean
country engages in FDI activities in the dirty country, it adopts the environmental standards
imposed by the dirty country for its FDI production abroad. Since, in this model, the level of
environmental damage (di) is determined by the level of environmental standards (ei), the clean
…rm will generate a higher level of environmental damage with its FDI production abroad than
with its domestically-located production.
The pro…t function of, respectively, the home and the foreign …rms is given by:
¼FDIy = y1py1 + y2py2 ¡ ey(y1 + y2) ¡ tyy1 (29)
¼FDIx = x1px1 + x2px2 ¡ exx1 ¡ eyx2 ¡ Fx (30)
According to equation (30), the pro…ts of the FDI host country (i.e., the foreign country)
increase with total sales revenues (from both, the domestic and the export market) and decreases
with both the foreign pollution tax (ey) and the foreign tari¤ (ty). The pro…t function of the FDI
source country (i.e., the home country) di¤ers in two aspects. First, the home …rm pays the home
pollution tax (ex) on the polluting good it produces in the home country and the foreign pollution
23
tax (ey) on the polluting good it produces in the foreign country. Second, it incurs a lump-sum
set-up cost Fx to engage in FDI activities in the foreign country.
The …rst-order pro…t maximization conditions yield the following optimal supply functions:
xFDI1 =
2¯(®1 ¡ ex) ¡ °(®2 ¡ ey ¡ ty)D
= xCN1 (31)
xFDI2 =
2¯(®1 ¡ ey) ¡ °(®2 ¡ ey)D
(32)
yFDI1 =
2¯(®2 ¡ ey ¡ ty) ¡ °(®1 ¡ ex)D
= yCN1 (33)
yFDI2 =
2¯(®2 ¡ ey) ¡ °(®1 ¡ ey)D
(34)
where D = 4¯2 ¡ °2: As in the basic case presented in Section 2, the concavity assumption of
the aggregate utility function associated with the above demand functions implies that ¯2¡°2 > 0.
We continue to assume that the own-price e¤ects are greater than the cross-price e¤ects, that is,
¯ > °.
While the presence of home FDI activities in the foreign country does not alter the sales in
the home market (i.e., xFDI1 = xCN
1 and yFDI1 = yCN
1 ), it a¤ects those in the foreign market. In
particular, whether the foreign market sales are higher or lower under an FDI regime than under
the standard Cournot-Nash regime described in Section 2.1 depends crucially on the relationship
between the tax burden imposed by the foreign country (ey) and that imposed by the home country
(ex + tx). In particular,
xFDI2 R xCN
2 if (ex + tx) R ey (35)
yFDI2 R yCN
2 if (ex + tx) Q ey (36)
4.2 Condition for FDI
The home country chooses to engage in FDI activities in the foreign country if its pro…ts are
higher under an FDI regime than under an intra-industry trade regime. The relative pro…tability
24
between these two regimes is given by:
¼FDIx ¡ ¼CN
x = ¯£(xFDI
2 )2 ¡ (xCN2 )2
¤¡ Fx (37)
If this expression is positive, it is pro…table for the home …rm to invest abroad and hence, to
incur the initial set-up cost necessary to produce x2 in the foreign country. If (¼FDIx ¡ ¼CN
x ) = 0;
the home …rm is indi¤erent between producing x2 in the home country (and paying import tari¤s
to the foreign country) and incurring the initial set-up cost to produce x2 in the foreign country
(and evading the foreign country’s import tari¤s). Finally, (¼FDIx ¡ ¼CN
x ) < 0 corresponds to the
situation in which it is in the home …rm’s best interest to produce x2 at home and to export its
production abroad, rather than to invest and produce abroad.
The conditional expressions given in (35) and (36), together with the home pro…t di¤erential
function given in (37) indicate that home export sales and hence, home pro…ts, are always smaller
under an FDI regime than under an exports regime when ey > (ex + tx). In this situation, it
is never pro…table for the home …rm to engage in FDI activities. Indeed, the home …rm engages
in FDI only when ey < (ex + tx), that is, when a smaller tax burden is levied on the home
good produced abroad under FDI than on the home good produced in the home country and
subsequently exported to the foreign country. While ey < (ex + tx) is a necessary condition for the
home …rm to consider FDI activities, it is not a su¢cient one. This can be seen from equations
(37), where the …rm’s relative pro…tability requires not only to produce a higher level of output
under an FDI regime than under an exports regime but also to cover its initial set-up cost Fx
associated with FDI. Thus, we need an additional condition to determine under which conditions
it is pro…table for the home …rm to engage in FDI activities.
For given pollution taxes, there exists a critical tari¤ rate at which the home …rm is indi¤erent
between undertaking FDI activities in the foreign country and exporting x2 to the foreign country.
This critical tari¤ rate, tcx; can be determined by di¤erentiating the home pro…t di¤erential given
25
in equation (37) with respect to the foreign import tari¤ tx:
tcx = (®1 ¡ ex) ¡ °2¯
(®2 ¡ ey)
The home …rm engages in FDI if the tari¤ level exceeds this critical tari¤ rate. Thus, while
the su¢cient condition for the home country to engage in FDI activities in the foreign country
is given in equation (37), the necessary set of conditions for the home country to undertake FDI
requires ey to be smaller than (ex + tx) and tx to be greater or equal to tcx:
Proposition 8 The necessary set of conditions for the home country to undertake FDI requires:(i) that ey is smaller than (ex + tx); and (ii) that tx is greater or equal to tcx:
4.3 Unilateral Trade Liberalization
In the previous section, we have shown that when ey ¸ (ex+tx), it is never pro…table for the home
…rm to engage in FDI activities, in which case the home …rm exports x2 to the foreign country.
In the context of the present paper, this is not an interesting situation since a unilateral trade
liberalization by the foreign country (dtx < 0) does not alter the home country’s initial decision
to export and thus, yields the same results as those already derived in Section 2.3.
If, however, ey < (ex + tx), the home country may choose to either undertake FDI activities
or export, depending on whether tx ? tcx: In particular, when tx > tcx, it is more pro…table for the
home …rm to incur the costs associated with FDI and to produce x2 abroad. But when tx < tcx,
it becomes more pro…table for the home …rm to produce x2 at home and to export its production
to the foreign market. This highlights an interesting implication for trade policy in the context of
intra-industry trade.
Assuming that the environmental damage generated under an FDI regime is signi…cantly dif-
ferent from that generated under an exports regime, a unilateral trade liberalization by the foreign
country can a¤ect the quality of the environment by inducing the home country to switch away
from an FDI regime toward an exports regime. Clearly, the extent to which one country bene…ts
from trade liberalization depends on several factors, including which country is liberalizing, which
26
country is experiencing inward FDI, which regime is more harmful to the environment and which
type of pollution is occurring.
We turn now to the environmental damage when the home country engages in FDI activities in
the foreign country to circumvent high trade barriers and to bene…t from the lower environmental
standards enforced in the host country.
4.4 Environmental Damage Di¤erential
Assuming that the home country engages in FDI activities, the environmental damage borne by,
respectively, the home and the foreign country is given by:
EdFDIx = dxx1 + ´ydy [x2 + y1 + y2] (38)
EdFDIy = dy(y1 + y2 + x2) + ´xdxx1 (39)
Again, by substituting the appropriate value for ´i; the degree of pollution spillover, the above
expressions illustrate the environmental damage under local (´i = 0), transboundary (0 < ´i < 1),
and global (´i = 1) pollution.
According to equation(38), the home country su¤ers from two sources of pollution. While the
…rst source is caused by the polluting production taking place in the home country, the second
source is generated by pollution spillovers from abroad, which are caused by both, the foreign
production and the home FDI production. Similarly for the foreign country, which su¤ers from
the pollution caused by the polluting production taking place in the foreign country (including
both, the foreign production and the home FDI production), and from the pollution spillovers
generated by the home production in the home country.
We want to determine which regime supports a higher level of pollution and hence, to examine
the environmental consequence of a given trade policy. This is possible by comparing the envi-
ronmental damage produced under an FDI regime (given in equations (38) and (39)) with that
produced under an intra-industry trade regime (given in equations (11) and (12)). In particular,
27
we are interested in solving the expression¡EdFDI
i ¡ EdBNi
¢for i = x; y under, respectively, local,
transboundary and global pollution. If this expression is positive, the environmental damage is
higher under an FDI regime than under a Cournot-Nash exports regime and a trade liberalization,
which discourages the home country to engage in FDI activities, is good for the environment. If,
however, this expression is negative, the environmental damage is higher under a trade regime
than under an FDI regime and a trade liberalization is bad for the environment.
4.4.1 Local Pollution
Assuming that pollution is local, the environmental damage di¤erential for the foreign country
between an FDI regime and an exports regime is given by:
¡EdFDI
y ¡ EdCNy
¢l= xFDI
2 dy +¡yFDI2 ¡ yCN
2¢dy 7 0 (40)
We have already shown above that if the home country engages in FDI activities, it must be
true that ey < (ex + tx). According to equation (36), this implies that yFDI2 is smaller than yCN
2
and hence, that the sign of equation (40) is ambiguous. More precisely, it is possible to solve for
the condition under which the environmental damage of the foreign country is higher under an
FDI regime than under an intra-industry trade regime. This is given in equation (41) below.
(EdFDIy ¡ EdCN
y )l > 0 iff2¯°
>(®2 ¡ ey) + (ex + tx ¡ ey)
(®1 ¡ ey)(41)
From a trade policy perspective, equation (41) says that there exists a parameter range over
which the foreign country can improve its environment by liberalizing trade to a point where tx <
tcx. Indeed, by reducing its tari¤ level below the home critical level, the foreign country discourages
inward FDI and thereby, redirects the home pollution-intensive FDI production away from the
foreign country towards the home country. While this policy can improve the environmental
quality in the foreign country, it deteriorates it in the home country. The latter result is shown
28
in equation (42), where the environmental damage for the home country is always lower under an
FDI regime than under an intra-industry trade regime.
¡EdFDI
x ¡ EdCNx
¢l= ¡xCN
2 dx < 0 (42)
Proposition 9 There exists a parameter range under which a trade liberalization by the foreigncountry, which induces the clean country to switch from an FDI regime to an exports regime, im-proves the environmental quality in the dirty country. Such a regime shift, however, unambiguouslydeteriorates the environmental quality in the clean country.
4.4.2 Transboundary Pollution
When pollution is transboundary, the environmental damage di¤erential between an FDI regime
and an exports regime of, respectively, the home and the foreign country is given by:
¡EdFDI
x ¡ EdCNx
¢t= (yFDI
2 ¡ yCN2 )´ydy +
£xFDI
2 ´ydy ¡ xCN2 dx
¤7 0 (43)
¡EdFDI
y ¡ EdCNy
¢t= xFDI
2 dy ¡ xCN2 ´xdx +
¡yFDI2 ¡ yCN
2¢dy 7 0 (44)
Assuming that the home country engages initially in FDI activities, we know that the …rst
bracket in equation (43) is negative, while the second bracket is positive. Thus, the sign of
equation (43) depends on the relative size of these two con‡icting e¤ects. But it is easy to see
that in the presence of a relatively small pollution spillover from the foreign country (i.e., a small
´y); the second (positive) bracket becomes very large, making it possible for the home country
to improve the quality of its environment by liberalizing trade. In particular, the environmental
damage in the home country is always higher under an FDI regime than under a Cournot-Nash
trade regime under the following condition:
(EdFDIx ¡ EdCN
x )t > 0 iff
2¯°
>(®2 ¡ ey)(´ydy ¡ dx) ¡ (ey ¡ ex ¡ tx)´ydy
(®1 ¡ ey)(´ydy ¡ dx) ¡ (ey ¡ ex ¡ tx)dx(45)
If this equation holds, a trade liberalization is always good for the environmental quality of the
home country. Thus, according to equation (45), a regime shift can be good for the environment
29
in the clean country, provided that the degree of transboundary pollution is relatively large (i.e.,
provided that ´y is relatively large).
As far as the foreign country is concerned, equation (46) shows that there exists also a pa-
rameter range under which a trade liberalization by the foreign country, which induces the home
country to switch from FDI activities to exports, improves the quality of the foreign environment.
In particular, the condition necessary for a trade liberalization to be good for the environment in
the foreign country is given by:
(EdFDIy ¡ EdCN
y )t > 0 iff
2¯°
>(®2 ¡ ey)(dy ¡ ´xdx) ¡ (ey ¡ ex ¡ tx)dy
(®1 ¡ ey)(dy ¡ ´xdx) ¡ (ey ¡ ex ¡ tx)´xdx(46)
By construction, dy is always greater than ´xdx, which implies that the right-hand side of
equation (46) is always smaller than one. Since the left-hand side of equation (46) is always larger
than one, the environmental damage borne by the foreign country is always higher under an FDI
regime than under a trade regime.
Proposition 10 Under transboundary pollution, a trade liberalization by the foreign country,which induces the clean country to switch from an FDI regime to an exports regime, always im-proves the environmental quality in the dirty country. This regime shift can also improve theenvironmental quality in the clean country, provided that the foreign country generate a relativelylarge pollution spillover.
4.4.3 Global Pollution
When pollution is global, the environmental damage di¤erential between an FDI regime and an
exports regime for the global economy is given by:
¡EdFDI ¡ EdCN¢g
=£xFDI
2 dy ¡ xCN2 dx
¤+ (yFDI
2 ¡ yCN2 )dy 7 0
Again, assuming that the home country engages initially in FDI activities, we know that the
…rst bracket is positive, while the second bracket is negative, making the sign of the environmental
di¤erential damage ambiguous. Thus, under global pollution, there exists a parameter range under
30
which a trade liberalization by the foreign country, which induces the clean country to switch from
an FDI regime to an exports regime, improves the global environmental quality. In particular, the
necessary condition for the global environmental damage to be higher under an FDI regime than
under an intra-industry trade regime is given by the following equation:
(EdFDI ¡ EdCN)g > 0 iff
2¯°
>(®2 ¡ ey)(dy ¡ dx) ¡ (ey ¡ ex ¡ tx)dy
(®1 ¡ ey)(dy ¡ dx) ¡ (ey ¡ ex ¡ tx)dx(47)
In the context of our model, we have assumed that dy > dx. This makes the right-hand
side of equation (47) always smaller than one, while the left-hand side of equation (47) is always
greater than one. Thus, given our initial assumptions, the conditional expression given in equation
(47) always holds true and a trade liberalization by the foreign country, which induces the clean
country to switch from an FDI regime to an exports regime, unambiguously improves the global
environmental quality.
The intuition for this result is as follows. When the home country engages in FDI activities, it
produces not only more goods under an FDI regime than under an exports regime (i.e., xFDI2 >
xCN2 ) but its FDI production is dirty rather than clean. Thus, under an FDI regime, the production
of the home country is both quantitatively more abundant and qualitatively more pollution-
intensive. A trade liberalization, which induces a regime shift away from FDI activities towards
trade activities, is good for the environment.
These results can be summarized as follows:
Proposition 11 Under global pollution, a unilateral trade liberalization by the dirty country,which induces the clean country to switch from an FDI regime to an exports regime, improvesunambiguously the quality of the global environment.
We …nd that under certain conditions the results derived in the previous section can be reversed
if one of the …rms (say, the clean …rm) is engaging in FDI activities in the pre-liberalization
equilibrium. For example, assuming that pollution is global, we have seen in Section 2.4 that a
unilateral trade liberalization by the dirty country deteriorates the global environment, unless the
31
pollution di¤erential between the two countries is very large. If, however, the home country is
circumventing high tari¤ barriers and strict environmental standards by engaging in FDI activities
in the foreign country, a unilateral trade liberalization by the dirty country, which induces the
clean country to switch from an FDI regime to an exports regime, improves unambiguously the
quality of the global environment. While these contradictory results are driven by very di¤erent
forces, they both highlight the relevance of recognizing the underlying trade structure and of
examining the environmental damages of trade liberalization in their appropriate context.
4.5 Multilateral Trade Liberalization
So far, we have analyzed the environmental e¤ect of a unilateral trade liberalization when the
home …rm engages in FDI activities in the foreign country. Now, let us consider the impact
of a multilateral trade liberalization on the quality of the environment in the presence of tari¤-
jumping FDI. To do this in a straightforward way, we adjust slightly our framework in order to
introduce two-way FDI. Again, we assume that the …rm that engages in FDI adopts the production
technology of the host country. For example, when the clean (dirty) country engages in FDI
activities in the dirty (clean) country, it adopts the environmental standards imposed by the dirty
(clean) country for its FDI production abroad.
In the presence of two-way FDI, a multilateral trade liberalization (such that dtx = dty = d¿),
which induces both countries to switch from a FDI regime to an exports regime, a¤ects the quality
of the environment in country i (where i = x; y) in the following way:
¡EdFDI
x ¡ EdCNx
¢t= (2xFDI
1 ¡ xCN1 ¡ xCN
2 )dx + (2yFDI2 ¡ yCN
1 ¡ yCN2 )´ydy 7 0 (48)
¡EdFDI
y ¡ EdCNy
¢t= (2xFDI
1 ¡ xCN1 ¡ xCN
2 )´xdx + (2yFDI2 ¡ yCN
1 ¡ yCN2 )dy 7 0 (49)
where the e¤ects of a multilateral trade liberalization on the quality of the environment under
the di¤erent types of pollution are derived by substituting the appropriate value for ´i, the degree
of pollution spillover (namely, ´i = 0 for local pollution, 0 < ´i < 1 for transboundary pollution,
32
and ´i = 1 for global pollution).
Note that according to expression (48), the e¤ects of a multilateral trade liberalization on the
environmental quality of the home country are ambiguous, and therefore not necessarily bad for the
environment. This is an interesting result because it opens up the possibility (at least in theoretical
terms) for an environmental-friendly multilateral trade liberalization. This is illustrated in the
next conditional expression, where:
(EdFDIx ¡ EdCN
x )t > 0 iff
2¯°
>2(ex ¡ ey)(´ydy ¡ dx) + ¿(dx + ´ydy)
¿(dx + ´ydy)(50)
Thus, a multilateral trade liberalization can be good for the environment in the home country
if: trade tari¤s (¿) in the pre-liberalization equilibrium are high; transboundary pollution from
abroad (´y) is low (such that ´ydy ¡dx < 0); pollution-intensive products are highly di¤erentiated
(high 2¯=°); and, the environmental standards di¤erences between the trading countries are low
(ex ¡ ey > 0).
Similarly, we can show that the e¤ects of a multilateral trade liberalization on the environ-
mental quality of the foreign country are ambiguous also. This is illustrated in the following
expression.
(EdFDIy ¡ EdCN
y )t > 0 iff
2¯°
>2(ex ¡ ey)(dy ¡ ´xdx) + ¿(´xdx + dy)
¿(´xdx + dy)(51)
Again, a multilateral trade liberalization can be good for the environment in the foreign coun-
try if: trade tari¤s in the pre-liberalization equilibrium are high; if transboundary pollution from
abroad (´x) is high (such that dy ¡´xdx is small); pollution-intensive products are highly di¤eren-
tiated (high 2¯=°); and, if the environmental standards di¤erences between the trading countries
are low (ex ¡ ey is small).
33
This is an interesting result, for it predicts that a multilateral trade liberalization does not
necessarily deteriorate the quality of the environment. On the contrary, in the presence of intra-
industry trade (and under a reasonable set of assumptions), a multilateral trade liberalization can
be good for the environment in both, the dirty and the clean country. This result clearly contradicts
the often cited ”dirty-migration hypothesis” argument, according to which trade liberalization
is bad for the environment because it induces the …rms that operate in countries with higher
environmental standards to relocate in countries with lower environmental standards. Our model
suggests that, in the presence of intra-industry trade, this argument may not always hold and
multilateral trade can be good for the environment.
Proposition 12 A multilateral trade liberalization, which generates a shift away from FDI to-wards intra-industry trade in both countries, may be good for the environment in both the cleanand the dirty country and hence, for the global environment. This proposition is particularly truewhen the trade barriers are high, when pollution-intensive products are highly di¤erentiated, andwhen the environmental standards di¤erentials are low.
5 Conclusion
In this paper, we o¤er a new causal mechanism linking international trade to changes in the quality
of the environment. We argue that while trade expansion alters the composition of national output
and hence, its environmental quality, its driving force may be international oligopolistic rivalry
rather than comparative advantage. We do not claim to o¤er a complete resolution of the debate
over trade and the environment. Our contribution is to take a …rst step toward clarifying the role of
intra-industry trade in determining how trade liberalization a¤ects the quality of the environment.
First, our results show that when the driving forces that determine the environmental damage
associated with trade expansion are fueled by neither comparative advantage nor cost di¤erences
but rather by oligopolistic competition, there is no justi…cation for the traditional inter-industry
factor-relocation hypotheses. In particular, we …nd that environmental deterioration is largely
determined by two factors, namely, : (i) by which country liberalizes trade, and (ii) by the nature
of pollution (i.e., whether pollution is local, transboundary or global). This result o¤ers a potential
34
explanation for why the composition e¤ect of trade has only found inconclusive empirical evidence.
Second, this paper sheds light on the question as to whether there are still positive gains from
trade when environmental degradation matters. In the context of intra-industry trade, a unilateral
trade liberalization induces three e¤ects, namely, a product variety e¤ect, a pro-competition e¤ect
and an environmental e¤ect. While the …rst two e¤ects are positive, we show that the third e¤ect
is not necessarily negative. For example, when pollution is transboundary, the global environment
can bene…t from trade liberalization, provided that the most pollution-intensive country liberalizes
trade.
Furthermore, our basic model indicates that a multilateral trade liberalization always deterio-
rates the quality of the environment. We show, however, that this proposition may be reversed in
the presence of tari¤-jumping FDI. This is an interesting result. The literature has long argued
that di¤erent countries have di¤erent pollution assimilation capacities. Implicitly, this argument
suggests that as long as the environmental degradation borne by each country does not exceed
its absorptive capacity, the positive variety and pro-competition e¤ects may well outweigh the
negative environmental e¤ect. Our results go one step further and suggest that, in the presence
of tari¤-jumping FDI, a multilateral trade liberalization may even have a positive impact on the
environment.
Our results also o¤er a new perspective on the role of competition policy in the context of
highly polluting industries. Our model suggests that the presence of a large number of …rms in
the non-liberalizing country magni…es signi…cantly the (negative) environmental impact of trade
liberalization. Thus, anti-trust policies may not always be good for the environment.
Finally, we …nd that market structure is an important factor when determining the environmen-
tal damage associated with trade liberalization. For example, we show that our basic results can
be reversed if polluting …rms are engaging in FDI activities in the pre-liberalization equilibrium.
Assume, for example, that high tari¤ barriers in the foreign country and/or strict environmental
standards in the home country induce the home polluting …rms to engage in FDI activities in the
35
foreign market. Then, even under global pollution, a unilateral trade liberalization by the dirty
country, which induces the clean country to switch from an FDI regime to an exports regime,
improves unambiguously the quality of the global environment.
It is important, however, to recognize the limitations of our analysis. Our analysis is limited
to the channel linking trade and the environment to the composition e¤ect via intra-industry
trade, ignoring other channels such as technological transfers and income e¤ects. Also, our model
su¤ers from extreme stylization. Nevertheless, the simple structure of our model can also be a
virtue since it lays bare the basic relationship driving our results and enhances our understanding
of the relationship between pollution and trade in the context of intra-industry trade. If any-
thing, this paper highlights how much remains to be done in advancing our understanding of the
many complex factors that in‡uence the environmental consequences of trade liberalization in an
increasingly integrated world economy.
36
ReferencesAnderson, Kym and Richard Blackhurst, eds., 1992, The Greening of World Trade Issues,New York: Harvester Wheatsheaf.
Antweiler, Werner, Brian R. Copeland and M. Scott Taylor, 1998, ”Is Free Trade Good forthe Environment?”, NBER Working Paper Series, WP 6707.
Baumol, Williams J. and Wallace E. Oates, (1988), The Theory of Environmental Policy,Cambridge: Cambridge University Press.
Beghin, John and Michel Potier, 1997, ”E¤ects of Trade Liberalization on the Environmentin the Manufacturing Sector”, The World Economy, 20(4):435-456, July.
Beghin, John, David Roland-Holst and Dominique van der Mensbrugghe, 1995, ”Trade Lib-eralization and the Environment in the Paci…c Basin: Coordinated Approaches to MexicanTrade and Environment Policy”, American Journal of Agricultural Economics, 77(3):778-786, August.
Bernhofen, D.M., 1999, ”Intra-Industry Trade and Strategic Interaction: Theory and Evi-dence”, Journal of International Economics, 47(1):223-244, February.
Bhagwati, Jagdish N. and Robert E. Hudec (eds.), 1997, Fair Trade & Harmonization:Prerequisites for Free Trade?, Volume 1: Economic Analysis, Cambridge, Massachusetts:The MIT Press.
Bhagwati, Jagdish N. and T.N. Srinivasan, 1997, ”Trade and the Environment: Does Envi-ronmental Diversity Detract form the Case for Free Trade?”, in Bhagwati, Jagdish N. andRobert E. Hudec (eds.), Fair Trade & Harmonization: Prerequisites for Free Trade?, Volume1: Economic Analysis, Cambridge, Massachusetts: The MIT Press.
Brander, James A., 1981, ”Intra-Industry Trade in Identical Commodities”, Journal of In-ternational Economics, 11(1):1-14, February.
Brander, James and Paul Krugman, 1983, ”A ’Reciprocal Dumping’ Model of InternationalTrade”, Journal of International Economics, 15(3/4):313-321, November.
Cole M.A., A.J. Rayner and J.M. Bates, 1998, ”Trade Liberalization and the Environment:The Case of the Uruguay Round”, The World Economy, 21(3):337-347, May.
Copeland, Brian R., 1996, ”Pollution Content Tari¤s, Environmental Rent Shifting, and theControl of Cross-Border Pollution”, Journal of International Economics, 40(3-4):459-476,May.
Copeland, Brian R. and M. Scott Taylor, 1999, ”Trade, Spatial Separation and the Envi-ronment”, Journal of International Economics, 47(1):137-168, February.
Copeland Brian R. and M. Scott Taylor, 1995a, ”Trade and the Environment: A PartialSynthesis”, American Journal of Agricultural Economics, 77(3):765-772, August.
Copeland Brian R. and M. Scott Taylor, 1995b, ”Trade and Transboundary Pollution”,American Economic Review, 85(4):716-737, September.
Copeland Brian R. and M. Scott Taylor, 1994, ”North-South Trade and the Environment”,Quarterly Journal of Economics, 109(3):754-787, August.
37
Dean, Judith M., 1992, ”Trade and the Environment: A Survey of the Literature”, in Low,Patrick (ed.), International Trade and the Environment, World Bank Discussion Papers No.159, Washington D.C., The World Bank.
Greenaway, David and Johan Torstensson, 1997, ”Back to the Future: Taking Stock onIntra-Industry Trade”, Review of World Economics, 133(2):249-269, June.
Greenaway, David and Chris Milner, 1986, The Economics of Intra-Industry Trade, Oxford:Basil Blackwell Ltd.
Grossman and Krueger, 1995, ”Economic Growth and the Environment”, Quarterly Journalof Economics, 110(2):353-377, May.
Grossman and Krueger, 1993, ”Environmental Impacts of a North American Free TradeAgreement”, in P. Garber (ed.), The U.S.-Mexico Free Trade Agreement, Cambridge, MA:MIT Press.
Hauer, Grant and Ford C. Runge, 1999, ”Trade-Environment Linkages in the Resolution ofTransboundary Externalities”, The World Economy, 22(1):25-39, January.
Helpman, Elhanan, 1987, ”Imperfect Competition and International Trade: Opening Re-marks”, European Economic Review, 31(112):77-81, February/March.
Hoekman, Bernard and Simeon Djankov, 1996, ”Intra-Industry Trade, Foreign Direct Invest-ment and the Reorientation of East European Exports”, CEPR Discussion Paper, No.1377,April.
Krugman, Paul, 1981, ”Intra-Industry Specialization and the Gains from Trade”, Journalof Political Economy, 89(5):959-973, October.
Krugman, Paul R. and Maurice Obstfeld, 1997, International Economics: Theory and Policy,New York: Addison-Wesley, 4th edition.
Lee, Hyun-Hoon and Young-Youn Lee, 1993, ”Intra-Industry Trade in Manufactures: TheCase of Korea”, Review of World Economics, 129(1):159-171, March.
Low, Patrick (ed.), 1992, International Trade and the Environment, World Bank DiscussionPapers No. 159, Washington D.C., The World Bank.
Low, Patrick and Raed Safadi, 1992, ”Trade Policy and Pollution”, in Low, Patrick (ed.), In-ternational Trade and the Environment, World Bank Discussion Papers No. 159, WashingtonD.C., The World Bank.
Menon, Jayant, 1994, ”Trade Liberalization and Intra-Industry Specialization: the Aus-tralian Experience”, Centre of Policy Studies Monash University, General paper No. G-107,June
Menon, Jayant and Peter B. Dixon, 1997, ”Measures of Intra-Industry Trade as Indicatorsof Factor Market Disruption”, The Economic Record, 73(222):233-247, September.
Menon, Jayant and Peter B. Dixon, 1996, ”How Important is Intra-Industry Trade in TradeGrowth?”, Open Economics Review, 7(2):161-175, April.
Rauscher, Michael, 1997, International Trade, Factor Movements, and the Environment,Oxford: Clarendon Press.
38
Siebert, Horst (1977), ”Environmental Quality and the Gains from Trade”, Kyklos,30(4):657-673.
Tharakan, P.K.M., 1984, ”Intra-Industry Trade Between the Industrial Countries and theDeveloping World”, European Economic Review, 26(2):213-226, November.
Thompson, Peter and Strohm Laura A., 1996, ”Trade and Environment Quality: A Reviewof the Evidence”, Journal of Environment & Development, 5(4):363-388, December.
World Trade Organization (2000), Annual Report 1999, Geneva: World Trade Organization.
39
Table 1:
Indexes for Intra-Industry Trade in Selected Industries, 1998
Western
US Europe
Agricultural Products 0.87 0.94
Food 0.90 0.97
Raw materials 0.78 0.81
Mining Products 0.70 0.78
Ores and other minerals 0.98 0.67
Fuels 0.57 0.74
Non-ferrous metals 0.89 0.90
Manufactures 0.86 0.98
Iron and steel 0.53 0.97
Chemicals 0.95 0.92
Other semi-manufactures 0.85 0.96
Machinery and transport equipment 0.91 0.98
Automotive products 0.79 0.96
O¢ce and telecom equipment 0.83 0.87
Other machinery and transport equipment 0.95 0.91
Textiles 0.78 0.97
Clothing 0.30 0.80
Other consumer goods 0.26 0.99
Total Merchandise Trade 0.85 0.96Source: WTO Annual Report, 1999
40