trade, growth, and the environment - university of hawaiinoy/362texts/copeland-environment.pdf ·...

7

Journal of Economic LiteratureVol. XLII (March 2004) pp. 7–71

Trade, Growth, and the Environment1

BRIAN R. COPELAND and M. SCOTT TAYLOR

1 Copeland: University of British Columbia. Taylor:University of Wisconsin. We are grateful for helpful com-ments from John McMillan, Gene Grossman, and the ref-erees. Copeland acknowledges funding from the SocialSciences and Humanities Research Council of Canada.

1. Introduction

For the last ten years environmentalistsand the trade policy community haveengaged in a heated debate over the envi-ronmental consequences of liberalizedtrade. The debate was originally fueled bynegotiations over the North American FreeTrade Agreement and the Uruguay Roundof GATT negotiations, both of whichoccurred at a time when concerns over glob-al warming, species extinction, and industri-al pollution were rising. Recently it has beenintensified by the creation of the WorldTrade Organization (WTO) and proposalsfor future rounds of trade negotiations.

The debate has often been unproductivebecause the parties differ greatly in theirtrust of market forces and typically value theenvironment differently. It has been ham-pered by the lack of a common language andalso suffered from little recourse to eco-nomic theory and empirical evidence. This isperhaps not surprising, because much of thework in this area is still quite new.

The purpose of this essay is to set out whatwe currently know about the environmental

consequences of economic growth and inter-national trade. We critically review both theory and empirical work to answer threebasic questions. What do we know about therelationship between international trade,economic growth, and the environment?How can this evidence help us evaluateongoing policy debates in this area? Wheredo we go from here?

To answer these questions, we discussboth the empirical and theoretical literaturewith the aid of a relatively simple generalequilibrium model where government policyand private sector behavior interact to deter-mine the equilibrium level of pollution. Themodel is developed in section 2 of the paperand then employed in various guisesthroughout. Our use of a model to organizeour review reflects an overarching theme ofour essay: economic theory needs to play amuch larger role in guiding empirical inves-tigation, suggesting alternative hypotheses,and disciplining inferences. The vast major-ity of empirical work in this field has littleconnection to explicit theory, and we arguethat this has left important policy debates,which hinge on the relative importance ofvarious theoretical magnitudes, badlyinformed.

The economic literature on these issuescame in two waves, with an initial surge ofactivity in the 1970s and a resurgence ofinterest stimulated by the policy debates

8 Journal of Economic Literature, Vol. XLI (March 2004)

2 See, for example, William Baumol (1971), Ingo Walter(1973), James Markusen (1975), Rudiger Pethig (1976),and Horst Siebert (1977).

3 Exceptions include Walter’s (1973) empirical paperand Pethig (1976) on the pattern of trade.

4 Other recent surveys that focus more on policyinclude Michael Rauscher (2001) and Alistair Ulph (1997).Hakan Nordstrom and Scott Vaughan (1999) provide agood comprehensive review of the trade and environmentliterature.

of the past decade.2 Much of the earlier lit-erature was normative, with a focus on issuessuch as gains from trade and optimal tradeor environmental policies.3 A large compo-nent of recent work also focuses on policyanalysis, but its most significant feature is itsconcern with positive issues: generating andattempting to test hypotheses about howtrade or growth affects environmental out-comes. We view these latter issues as funda-mental to resolving current policy debates,and so most of our essay will focus on thisaspect of the literature.4

We begin our analysis in section 2 byexamining the link between incomes percapita and environmental quality. Interest inthis link arose from the pioneering work byGene Grossman and Alan Krueger (1993)on NAFTA, which subsequently led to aburgeoning literature on what has come tobe known as the environmental Kuznetscurve (EKC). The environmental Kuznetscurve hypothesizes an inverse-U-shapedrelationship between a country’s per-capitaincome and its level of environmental qual-ity: increased incomes are associated withan increase in pollution in poor countries,but a decline in pollution in rich countries.This literature is important because many inthe trade-policy community have arguedthat trade and growth may actually be goodfor the environment. If environmental qual-ity is a normal good, increases in incomebrought about by trade or growth will bothincrease the demand for environmentalquality and increase the ability of govern-ments to afford costly investments in envi-ronmental protection.

Our review of both the theoretical andempirical work on the EKC leads us to beskeptical about the existence of a simple andpredictable relationship between pollutionand per-capita income. To investigate theEKC we employ a simple pollution demand-and-supply system linking pollution levels tonational characteristics (incomes, factorendowments, and technologies) and tradingopportunities (comparative advantage andcurrent trade restrictions). Much of the workon the EKC reduces this set of possibleexplanatory factors to essentially just one—incomes per capita—although it is not clearwhy we would want to impose these restric-tions on empirical estimation ex ante. Thisconcern receives some support from empiri-cal work that finds that the shape of the esti-mated relationship between pollution andincome is sometimes sensitive to functionalform, the sample of countries or cities used,and the time period chosen.

Despite these concerns, the EKC litera-ture has made two lasting and significantcontributions. First, it raised importantempirical questions about how trade andgrowth affect the environment, and launcheda significant research agenda. Second, it hasprovided quite convincing evidence thatthere is an income effect that raisesenvironmental quality. Moreover, there arestrong indications that this income effectworks because increases in the stringency ofenvironmental regulation accompany higherper-capita incomes. Therefore an analysis ofthe effects of trade and growth on the envi-ronment cannot proceed without taking intoaccount endogenous policy responses.

Unfortunately, because most of the litera-ture has relied on simple reduced-form esti-mation, there is very little work that isolatesthis income effect from other factors, such asthe scale effects of growth and capital abun-dance. Consequently we still know very littleabout when this income effect is strongenough to offset forces, such as capital accu-mulation or export-led expansion of pollut-ing industries, that are associated with

Copeland and Taylor: Trade, Growth and the Environment 9

increased demands on the environment.Some evidence on this score exists, but it islimited and specific to certain commonlymeasured pollutants (SO2 for example).

While the link between income growthand the environment is important, trade mayalter environmental outcomes through avariety of other channels. Trade may encour-age a relocation of polluting industries fromcountries with strict environmental policy tothose with less stringent policy. These shiftsmay in turn increase global pollution or theymay have a chilling effect on environmentalpolicy, as countries will be reluctant to tight-en environmental regulations because ofconcerns over international competitiveness.

To examine these additional concerns weevaluate the environmental impact of tradeliberalization in section 3, starting in a smallopen economy setting and then moving to aworld with many small economies and worldprice determination. Our analysis within thesmall open economy setting links the envi-ronmental impact of trade liberalization tothe choice of policy instruments, the flexibil-ity of policy and, most importantly, a nation’scomparative advantage. At this point we findit crucial to distinguish between two differ-ent hypotheses linking pollution regulationto comparative advantage. The first is that atightening up of pollution regulation will, atthe margin, have an effect on plant locationdecisions and trade flows. We call this a pol-lution haven effect. This hypothesis hasstrong theoretical support.

The second hypothesis is that a reductionin trade barriers will lead to a shifting of pollution-intensive industry from countrieswith stringent regulations to countries withweaker regulations. We call this the pollutionhaven hypothesis. The theoretical supportfor the pollution haven hypothesis is, in con-trast, quite weak, because trade theory sug-gests that many other factors, in addition topollution regulation, affect trade flows. Ifthese other factors are sufficiently strong,then it is quite possible for there to exist apollution haven effect, but have the pollution

haven hypothesis fail. This distinction has alarge impact on our discussion of policyissues.

Until quite recently, there was a consensusin the empirical literature that differences inthe stringency of environmental regulationhave little or no effect on trade and invest-ment flows (Adam Jaffe et al. 1995). Recentwork suggests this conclusion was prema-ture. Most of the literature prior to 1997 thatinvestigated the stringency of environmentalregulation on trade and investment flowsused cross sectional data. These studies areunable to control for unobserved hetero-geneity across countries or regions and typi-cally treat pollution regulations as exogenous.If pollution policy is endogenous or there areimportant omitted factors, then the estimat-ed results will be misleading. Several recentstudies have addressed these issues andfound evidence for the existence of a pollu-tion haven effect: the stringency of pollutionregulations does affect plant location andtrade flows. There remains little convincingevidence, however, to support the pollutionhaven hypothesis. Instead, the available evi-dence suggests that other factors are moreimportant in determining trade patterns thanare differences in pollution regulations.

This is an area that still needs much morework, in part because the literature has oftenblurred the distinction between the pollu-tion haven effect and the pollution havenhypothesis. This distinction is also importantto our answers to two key policy questions.In section 4 we first ask whether environ-mental policy should be constrained byinternational trade law to prevent countriesfrom using it as a substitute for trade policy.This issue lies behind much of the concernthat trade liberalization might lead to a “raceto the bottom” in standards as countriesweaken their environmental policy inresponse to the competitive pressures offreer trade. The empirical evidence on thepollution haven effect is relevant to thisquestion, because if pollution policy doesaffect trade and investment flows, as recent


5 One large omission from our review is any explicit dis-cussion of renewable or nonrenewable resource use andsustainability. For many in the developing world, the statusof fish stocks, aquifers, and forests are key environmentalindicators, but an analysis of trade’s impact on resource usewould take us too far afield. For recent work on theseissues, see the survey by Gardner Brown (2000), the seriesof case studies examining the impact of trade liberalizationon resource industries available from the United NationsEnvironment Programme at http://www.unep.ch/etu/pub-lications/index.htm, and the excellent book-length treat-ment by Ian Coxhead and Sisira Jayasuriya (2003).

evidence indicates, then it is possible thatweak pollution policy may be used as a loop-hole in trade agreements. The second keypolicy question is whether trade policyshould be used to achieve environmentalobjectives either at home or abroad. The dis-putes at the WTO arising from the U.S.import bans on tuna from Mexico to protectdolphins and on shrimp from various coun-tries to protect turtles are two prominentexamples of this type of issue. On this issue,evidence concerning the pollution havenhypothesis is relevant because much of theconcern over trade’s environmental effect—either at home or abroad—disappears if thehypothesis is false.

We do not provide unequivocal answers tothe questions we pose. Instead we try toreport on the current state of affairs andidentify the set of important but as yet unan-swered questions that we need to resolve tobetter understand the trade, growth, andenvironment link.5

2. Growth and the Environment

We start with the relationship betweenincome and the environment. This is a keyaspect of the debate. The main argument forfree trade is that it will raise nationalincomes; but if this is so, then it is importantto understand how higher incomes affectenvironmental quality.

The empirical literature on the relation-ship between environmental quality and per-capita income has proliferated over the pastdecade, following the seminal work ofGrossman and Krueger (1993). Using a

6 An earlier literature found some evidence to supporta hump-shaped relationship between the intensity of metaluse (consumption of a given metal per unit of GDP) andper-capita income. See Wilfred Malenbaum (1978). Note,however, that this does not imply a fall in resource use percapita after the hump is passed.

7 See Therese Cavlovic et al. (2000) and SusmitaDasgupta et al. (2002) for recent reviews of this work.

panel of data on air quality from 42 coun-tries, Grossman and Krueger found a hump-shaped relation between some measures ofair quality (such as SO2 concentrations) andper-capita income: pollution at first rises andthen falls with income per capita.6 ThomasSelden and Daqing Song (1994) found a sim-ilar pattern using data on sulfur dioxideemissions. For some other pollutants (suchas contaminated drinking water), Grossmanand Krueger (1995) and Nemat Shafik andS. Banyopadhyay (1992) found pollutiondeclines monotonically with income percapita; while for others (such as carbon emis-sions) pollution tends to rise with incomeper capita. The hump-shaped relation hascaptured most of the attention and for thisreason, this line of work is known as theenvironmental Kuznets curve (EKC) litera-ture, since it echoes Simon Kuznets’ (1955)finding of a hump-shaped relation betweeninequality and per-capita income. There arenow numerous papers that estimate an EKCfor various measures of environmental qual-ity, time periods, countries, etc.7

What is perhaps most striking about theEKC literature is the limited role that theo-ry has played in its development. This hascreated difficulties in interpretation, sincethe basic finding is consistent with manypossible explanations. For this reason, webegin by asking what theory has to say aboutthe relation between income and pollutionbefore moving on to the empirical work.

2.1. A Model

We need a model with three key featuresto help interpret the literature. We need atleast two goods that differ in pollution inten-sity to allow for the possibility of differencesin the composition of economic activity over

http://www.unep.ch/etu/publications/index.htmhttp://www.unep.ch/etu/publications/index.htm


8 The model is based on Brian Copeland and M. ScottTaylor (2003), which has its roots in Copeland and Taylor(1994). Martin McGuire (1982) and Rauscher (1997) usesimilar models.

9 Models with consumption-generated pollution havebeen somewhat neglected in the trade literature. For oneexample, see parts of Copeland and Taylor (1995b) andRauscher (1997).

10 Production externalities are discussed in Copelandand Taylor (1999) and Michael Bennaroch and HenryThille (2001).

time or across countries. We need at leasttwo primary factors of production to providea simple motive for international trade thatis independent of pollution regulation. Andfinally, we need endogenous pollution policyto examine how pollution may vary acrosscountries with different levels of per-capitaincome. In what follows, we opt for simplic-ity rather than generality and do not attemptto be exhaustive in our coverage.8

We adopt a static model and focus on pro-duction-generated pollution.9 Pollutionfrom a given firm harms consumers but doesnot affect the productivity of other produc-ers.10 There are two goods, X and Y, eachproduced with a constant returns to scaletechnology using two primary factors, capital(K) and labor (L). Denote the price of X byp, and treat Y as the numeraire. To capturedifferences in pollution intensity across sec-tors, we assume X generates pollution duringproduction, but Y does not pollute at all.

The production function for good Y issimply:

Y = H(Ky,Ly). (1)

where H is increasing, concave, and linearlyhomogeneous.

Production of good X generates pollutionemissions (Z). If firms do not undertakeabatement, we assume each unit of outputgenerates one unit of pollution, and that out-put of X is given by F(Kx,Lx), where F isincreasing, concave, and linearly homoge-neous. We can think of F as “potential out-put.” If abatement does occur, then for z £ F,output of X is given by:

x = za[F(Kx,Lx)]1-a, (2)

11 One can alternatively start with a joint productiontechnology and then (under some regularity conditions)invert it to obtain a production function that treats pollu-tion as an input. See Siebert et al. (1980) and Copelandand Taylor (2003).

12 Our functional form implicitly assumes that abate-ment has the same factor intensity as production or poten-tial output. This assumption is made for simplicity.

13 If the pollution tax is sufficiently low, firms will notabate at all, and a corner solution will result. At this pointz = x, and so e = 1. Referring to (3), this no-abatementsolution occurs if t £ ap.

where 0 < a < 1. Although pollution is a jointoutput, one can treat pollution (or environ-mental services) as an input for analyticalconvenience.11 A firm can reduce pollutionand maintain output constant by using moreprimary factors and adopting less-pollutingtechniques.12

If governments regulate pollution, weassume that firms face a price t for each unitof emissions that they release. This pricemay be implemented with either an emis-sions tax t or by a tradable emissions permitsystem, in which the government sets thetotal level of pollution Z, and the emissionsprice t is determined in the market.

Firms choose the emissions intensity thatminimizes their production costs. Let e = z/xdenote emissions per unit of output. TheCobb-Douglas form of the production func-tion implies that the share of emissioncharges in the value of output is a = tz/px,and hence at an interior solution, we have:13

3)

The emission intensity falls as pollutiontaxes rise; and it rises when the price of thepolluting good p rises because the opportu-nity cost of resources used in abatement ishigher.

To close the production side of the model,we require nonpositive profits in each indus-try, and full employment. These conditionscan be solved to obtain outputs as functionsof endowments, prices and policy:

(4)x x p K L

y y p K L

==

( )

( )

, , ,

, , ,

tt

ezx

pº = £at

1


14 It is concave in (K, L, and z) and convex in prices.Moreover, outputs and factor prices can be recovered withsimple differentiation: x = ¶G/¶p; r =¶G/¶K; w = ¶G/¶L.See Alan Woodland (1982) and Avinash Dixit and VictorNorman (1980) for the standard treatment of nationalincome functions, and Copeland (1994) for an applicationto environmental problems.

For a given pollution tax, it can be verifiedthat this model behaves much like the pro-duction side of the standard Heckscher-Ohlin model of international trade. Inparticular, an increase in the supply of capi-tal will increase the output of the capital-intensive industry X, and reduce the outputof Y. An increase in the supply of labor stim-ulates Y and contracts X.

We can summarize the production side ofthe model with a national income function.Because markets are competitive, the pri-vate sector maximizes the value of nationalincome for any given pollution level z. Thisallows us to write national income G as thesolution to an optimization problem:

(5)

where T is the feasible technology set. As iswell known, the national income functionsatisfies a number of useful properties.14

Most useful to us is the following:

(6)

The equilibrium price of a pollution permitis equal to the effect on national income ofan increase in allowable pollution; that is, ifmore pollution is allowed, national incomerises by the value of the marginal product ofemissions. If we instead think of the effectsof a reduction in emissions z, then the cost tothe economy is also ¶G/¶z. This is the gener-al equilibrium marginal abatement cost.Hence another interpretation of (6) is thatthe price of a unit of emissions is equal to themarginal abatement cost, which is a standardresult in environmental economics.

We assume there are N identical con-sumers in the economy. Each consumer

t = ¶¶Gz

.

G p K L z px y x y T K L z( ) max { : ( ) ( )}{x,y}

, , , , , ,= + Î

15 Homotheticity allows us to write the indirect utilityfunction as an increasing function of real income. It alsoensures that the relative demand for goods is unaffected byincome levels. This is a standard assumption in the inter-national trade literature and it allows us to focus on therole of environmental policy and factor supplies in explain-ing trade patterns.

maximizes utility, treating pollution as given.For simplicity, we assume preferences overconsumption goods are homothetic and theutility function is strongly separable withrespect to consumption goods and environ-mental quality.15 The indirect utility func-tion for a typical consumer is

(7)

where h is increasing and convex, I is percapita income (so I = G/N), b is a priceindex, and v is increasing and concave.Pollution is harmful to consumers and istreated as a pure public bad (all consumersexperience the same level of pollution).

The Demand for Pollution. In ourapproach, we treat pollution as an endoge-nously supplied factor of production. Thissuggests a natural way to think about thedeterminants of pollution is in terms of itsdemand and supply. Notice (6) can be inter-preted as the inverse demand for pollution.We illustrate this demand curve in figure 1.It slopes down because G is concave in z.More intuitively, we can exploit the structurewe imposed on technology to write pollutiondemand as

(8)

This is the same relation we would obtain byinverting (6) and using our assumptions ontechnology. From (8), we can see that thedemand for pollution slopes down for tworeasons: as t falls, firms pollute more bothbecause the emissions intensity e rises, andbecause the lower tax on pollution makes pro-duction of the dirty good more attractive (sothat output of X expands while Y contracts).

The Supply of Pollution. Pollution supplydepends on the policy regime. If there is no

z e p x p K L= ( ) ( , , , )/ t t

V p I z v I p h z( , , ) ( ( )) ( )= -/ b


t

Figure 1. Pollution Demand and Supply

Z°

t°

t

Gz

NMD

Z

regulation, then pollution supply is perfectlyelastic at t = 0. Pollution in this case isentirely demand driven. If there is an exoge-nous pollution tax to, then supply is a hori-zontal line. Shifts up or down in pollutiondemand raise or lower emissions.Alternatively, if there is a fixed overall pollu-tion quota in place (as in a tradable emissionpermit system), then the pollution supplycurve is vertical. Shifts in pollution demandraise or lower the price of emissions, buthave no effect on overall pollution.

In general, we expect pollution policy tobe endogenous; and in particular, we expectthat changes in per-capita income will leadto an increase in the demand for environ-mental quality, and (if governments areresponsive) a tightening up of pollution reg-ulations. The endogeneity of pollution policyplays a key role in both the theory andempirical literature.

There are two approaches to modeling thepolicy process. One is to simply assume abenevolent government chooses policy.Another is to adopt a political economyframework where the interaction of compet-

16 The use of political economy models in the trade andenvironment literature is still in the early stages. Examplesinclude Arye Hillman and Heinrich Ursprung (1994), PerFredriksson (1997, 1999), Sumeet Gulati (2001), CarolMcAusland (2003), Raucher (1997, ch. 5), and JoachimSchleich (1999).

17 See, however, our discussion of political economyelements in section 3 on policy implications.

ing interest groups determines policy.16 Wefollow the bulk of the literature on endoge-nous policy and adopt a representative agentframework in which the government pro-vides efficient policy.17

To determine the optimal pollution policy,the government chooses the pollution levelto maximize the utility of a representativeconsumer subject to production possibilitiesand private sector behavior. The govern-ment’s problem is:

(9)

Because we assume the economy is small inworld markets, the government treats thegoods price p as given. Hence dp/dz = 0 andthe first order condition from (9) becomes:

Max V I p z s t I G p K L z Nz

{ ( ( ) ) ( ) }/ , . . , , , /b =


18 R also varies endogenously with z; this is taken intoaccount when drawing the supply curve.

To simplify, recall that Gz = t, which is theprivate sector’s marginal valuation for a unitof pollution. As well, define R = I/b(p) as realincome of the representative consumer. Wecan then rewrite the first order condition as:

(10)

where MD(p,R,z) is a representative con-sumer’s marginal damage from pollution(the marginal rate of substitution betweenpollution and income). The optimal tax sim-ply implements the standard Samuelsonrule: the pollution tax is the sum of marginaldamages across all individuals.

If pollution policy is implemented effi-ciently, then (10) can be interpreted as thesupply of pollution. As shown in figure 1, thesupply curve slopes upwards becauseincreases in pollution tend to make environ-mental quality scarce relative to consump-tion. That is, a diminishing marginal rate ofsubstitution between consumption and envi-ronmental quality yields an upward slopingsupply curve. As well, exogenous increases inendowments or technology that increase realincome will shift the supply curve in:because environmental quality is a normalgood, marginal damage is increasing in realincome (MDR > 0).18

Market Equilibrium with Efficient Policy.The equilibrium level of pollution is deter-mined by the interaction between thederived demand for pollution and the aggre-gate marginal damage as captured in pollu-tion supply. Combining (6) and (10) yields:

(11)

Equation (11) determines the efficient levelof pollution zo, as illustrated in figure 1. Toimplement zo, the government can eitherintroduce a pollution tax to, or issue zo mar-

G p K L z N MD p R p K L z zz ( ) ( ( ) )., , , , , , , ,=

t = =N V V N MD p R zz I – / , ,[ ] ( )

V GN

VI z z+ = 0.ketable permits which would yield the equi-librium permit price to.

Scale, Technique, and CompositionEffects. It is useful to have a simple way tolink changes in the economy to environ-mental outcomes. An understanding ofthese links is critical for empirical workbecause we have to distinguish between theeffects of growth, trade, and other factors ifwe are to measure their importance.Grossman and Krueger (1993) used theconcepts of scale, composition and tech-nique effects as the basis of their discussion,and we proposed formal model-based defi-nitions in Copeland and Taylor (1994). Herewe employ these definitions to provide asimple decomposition.

Trade and growth both increase realincome, and therefore both increase theeconomy’s scale. To be more precise, weneed a measure of the scale of the economy;that is, we need an index of output. Thereare many ways to create such a quantityindex, but for simplicity, we will use thevalue of output at a given level of worldprices as our measure of the economy’s scale.Our measure of scale, S, is defined as

(12)

where pxo and py

o denote the level of worldprices prior to any shocks we analyze. Ifworld prices change, we continue to con-struct S using the old (base-period) worldprices. This is so that scale will not changesimply because of a change in valuation.

Given this definition of scale, and settingbase-period prices to unity, we now use (12)to write pollution as

(13)

where jx = pxox/S = x/S is the value share ofthe dirty good X in total output evaluated atbase-period prices. Hence pollution emis-sions, z, depend on the emissions intensity ofproduction, e, the importance of the dirtygood industry in the economy, j, and thescale of the economy, S.

z ex e Sx= = j

S p x p yxo

yo= +


Taking logs and totally differentiatingyields our decomposition:

(14)

where ẑ = dz/z, etc.The first term is the scale effect. It meas-

ures the increase in pollution that would begenerated if the economy were simply scaledup, holding constant the mix of goods pro-duced and production techniques. As anexample, if there were constant returns toscale and all of the endowments of the econ-omy grew by 10 percent, and if there wereno change in relative prices or emissionintensities, then we should expect to see a 10percent increase in pollution.

The second term is the composition effectas captured by the change in the share of thedirty good in national income. If we hold thescale of the economy and emissions intensi-ties constant, then an economy that devotesmore of its resources to producing the pol-luting good will pollute more.

Finally, we have the technique effect, cap-tured by the last term in (14). Holding allelse constant, a reduction in the emissionsintensity will reduce pollution.

Understanding the interaction betweenthese effects will play an important role indetermining how trade and growth affect theenvironment.

2.2 The Environmental Kuznets Curve

With our model and definitions in handwe now consider the literature on theEnvironmental Kuznets Curve. From thevantage point of our demand and supply sys-tem, a key difficulty immediately arises. TheEKC literature seeks to estimate a simplerelationship between per-capita income andpollution. But income and pollution are eachendogenous variables that are functions ofmore primitive determinants. Since differ-ent types of economic activity have differentpollution intensities, it would be surprisingto find a simple relationship between all pos-sible realizations of income and pollution.Our simple theory predicts instead that the

ˆ ˆ ˆ ˆz S e= + +j

19 Note that Gz = 0 since t = 0.

shape of the relationship between incomeand pollution should vary with the source ofincome growth.

To illustrate, we demonstrate how physicaland human capital accumulation yield differ-ent income-pollution paths. For simplicity,normalize the population so that N = 1, andsuppose there is no pollution regulation. Inthis case, the emission intensity is e = 1 andwe can specialize (8) to write pollution as

z = x(p,t,K,L) (15)

where t = 0. Income is given by

I = G(p,K,L,z) (16)

Suppose growth occurs via capital accu-mulation alone. Then differentiating (15)and (16), holding t = 0 and L constant,yields19

(17)

and

(18)

where exK > 0 is the elasticity of X outputwith respect to the endowment of capital, sr> 0 is the share of capital in national income,and ẑ = dz / z, etc. Equations (17) and (18)tell us that capital accumulation raises bothincome and pollution.

Combining them we obtain a reducedform relationship between pollution andincome:

(19)

(+)

With no pollution policy, there is a positive,monotonic relationship between pollutionand income if growth occurs via the factorused intensively in the dirty industry.

Alternatively, suppose growth occurs viahuman capital accumulation. Then we have:

(20)

(-)

ˆ ˆzs

IxLw

= e

ˆ ˆzs

IxKr

= e

ˆ ˆ .I s Kr=

ˆ ˆz KxK= e


20 In an interesting paper, Erwin Bulte and Daan vanSoest (2001) make a related point in the context of renew-able resource exploitation. Using a model of optimalresource use, they consider the relation between thesteady state resource stock and income, and point out thatthe shape of the relation depends on whether income risesvia an increase in the resource price or an exogenousincrease in non-resource-based income.

21 Some empirical support for this view can be found inWilliam Harbaugh, Arik Levinson, and David Wilson(2002) and other papers, as we discuss later.

where exL < 0 is the elasticity of X outputwith respect to the endowment of humancapital and sw > 0 is the share of human cap-ital in national income. Note exL < 0, followsfrom the Rybczinski theorem of internation-al trade: human capital accumulation stimu-lates the clean industry Y, which drawsresources out of the dirty industry X andlowers pollution. Hence when growth occursvia accumulation of the factor used inten-sively in the clean industry, there is a nega-tive monotonic relationship betweenpollution and income.

This simple example highlights how dif-ferent sources of growth will in general traceout different income and pollution paths.20

Theory suggests there may well be a stablerelation between pollution and various prim-itives such as technology and primary factorsof production, and between income andthese same variables. But unless all countriesgrow in exactly the same way, there is littlereason to expect that there will be a simplerelationship between pollution andincome.21 Accordingly, all theories that gen-erate an inverse-U-shaped environmentalKuznets curve must proceed by imposingmore structure than even our simple pollu-tion demand and supply model contains.

There are four main explanations for theEKC. Each explanation places restrictionson preferences and technology to makeequilibrium pollution a function of per capi-ta income, and to generate the desiredshape of the income-pollution relation. Weclassify these explanations by the key mech-anism driving their results. These are: (1)sources of growth; (2) income effects; (3)

22 Grossman and Krueger (1995) and others cite MosheSyrquin’s (1989) discussion of the structural transforma-tion in an economy during the development process (forexample, for agriculture to manufacturing to services) asmotivation for the view that predictable changes in thesources of growth during the development process couldaffect the pollution-income relationship.

23 This approach has been relatively neglected in theempirical literature until recently. Lewis Gale and JoseMendez (1998) include relative capital abundance in theirestimation of an EKC for sulfur dioxide, and TheodorePanayotou, Alix Peterson, and Jeffrey Sachs (2000) use ameasure of the capital stock in their study of CO2 emis-sions. Hermamala Hettige, Muthukumara Mani, andDavid Wheeler (2000) decompose industrial water pollu-tion into scale, composition and pollution intensity effects,and find that the manufacturing’s share of output follows ahump-shaped pattern, but that it is not strong enough toyield an EKC pattern for water pollution.

threshold effects; and (4) increasing returnsto abatement. Although all of these explana-tions describe forces that could interact, wewill isolate the key features of each in ourpresentation.

Sources of Growth. To obtain the risingand then falling portions of an EKC, even inthe absence of any environmental policy, wecould place restrictions on the growthprocess across all countries. This yields oneof the commonly mentioned explanationsfor the EKC, although it seems to lack a for-mal development in the theoretical litera-ture.22 Very simply, suppose policy is notvery responsive to income (i.e. a restrictionon pollution supply), but countries growprimarily via capital accumulation in theearly stages of development and by humancapital acquisition in later stages (i.e.restrictions on the time profile of demandshifters). Then pollution will rise and thenfall with growth in per-capita income, ascomposition effects driven by the factorgrowth drive the profile for pollution.Composition effects are key here, becausewe have assumed a zero policy responseeliminating technique effects, and the com-position effects of factor accumulationdominate scale effects in this model. Giventhese assumptions, changes in the sourcesof growth are reflected in the pattern ofpollution.23


24 Ramon Lopez (1994) provides an early treatment ofthis approach, as he shows how non-homotheticity in pref-erences between consumption and environmental qualitycan lead to an EKC. Kishore Gawande, Robert Berrens,and Alok Bohara (2001) provide an interesting variation onthe income effect approach—in their model, agents arefreely mobile and so income effects induce a sorting equi-librium in which higher income agents avoid pollutedareas.

The source of growth explanation for theEKC is important to our discussion for tworeasons. First, it demonstrates how the pol-lution consequences of growth depend onthe source of growth. Therefore, the analogydrawn by some in the environmental com-munity between the damaging effects ofeconomic development and those of liberal-ized trade is, at best, incomplete. Second,the sources of growth explanation demon-strates that a strong policy response toincome gains is not necessary for pollution tofall with growth. Hence the shape of theEKC need not be driven by income gainsmaking pollution policy more stringent.

Income Effects. An alternative widelycited explanation for the EKC is that itsshape reflects changes in the demand forenvironmental quality as income rises.24 Toillustrate this theory, suppose governmentsset policy efficiently, and consider the effectsof neutral technical progress. Let l be a shiftparameter representing technology, andagain normalize the population so that N=1.With neutral technical change, we can writeour GNP function as lG(p,K,L,Z).

Pollution is determined by:

(21)

and differentiating with respect to l andrearranging yields:

(22)

where D > 0, and eMD,R is the income elas-ticity of marginal damage. Neutral techno-logical progress shifts both the demand andsupply of pollution. Demand shifts because

dzd

MD R

l=

-1 e ,D

l lG p K L z MD p G p K L z p zz ( ) ( ( ) ( ) ), , , , , , , / ,= b

25 Similar results are obtained by considering theeffects of neutral factor accumulation on pollution. Lopez(1994) shows how the effects of factor accumulation on theenvironment depend on interaction between the elasticityof substitution between pollution and nonpollution inputsand the income elasticity of marginal damage. Copelandand Taylor (2003, ch.3) explore these issues in greaterdetail.

26 While the income effect theory is illustrated here interms of a benevolent government in a representativeagent economy, it can be modified to allow for politicaleconomy motives. For example, Lopez and SiddharthaMitra (2001) show how the presence of corruption willmove the turning point of the EKC to the right. Someempirical studies, such as Scott Barrett and KathrynGraddy (2000) include measures of political freedom as anextra shift variable in their EKC regressions; and find thatall else equal, increased freedom is associated with a clean-er environment. However, the theory outlined abovewould imply that the political freedom variables should beinteracted with income variables since one would expectpolitical freedom to influence the strength of the policy-induced technique effect.

the marginal product of pollution rises; sup-ply shifts because real income has grown.

Whether pollution rises or falls with realincome changes depends on the incomeelasticity of marginal damage.25 If the elas-ticity is less than one, then the supply shift isswamped by the demand shift and pollutionrises; if it is greater than one, the oppositeoccurs.26

Because the EKC has both an increasingand decreasing segment, this pure income-driven explanation requires a variableincome elasticity of marginal damage to gen-erate the required shape. As an example,suppose indirect utility is given by:

(23)

where x > 0 (and R is real income). The keycharacteristic of (23) is that the income elas-ticity of marginal damage is simply R/x.Using (22), we obtain an inverse-U relationbetween real income and pollution: pollu-tion rises with neutral growth if R > x andfalls with neutral growth if R < x.Environmental quality is a normal goodthroughout, but at low incomes, pollutionrises with growth because increased con-sumption is valued highly relative to envi-ronmental quality. As income rises, thewillingness to pay for environmental quality

V p I z c c R h z( ) exp( ) ( ), , /= - - -1 2 x


27 See Shafik (1994) and Douglas Holz-Eakin andSelden (1995). However, Richard Schmalensee, ThomasStoker, and Ruth Judson (1998) do find a within-samplepeak for carbon emissions per capita.

rises and increasingly large sacrifices in con-sumption are made to provide greater envi-ronmental benefits.

The income-effect explanation of theEKC follows from two assumptions: neutralgrowth and a particular assumption on pref-erences. Neutral growth restricts the magni-tude of shifts in pollution demand as growthproceeds; while the rising income elasticityof marginal damage ensures ever-strongertechnique effects. Composition effects playlittle or no role.

This explanation suggests that the rela-tionship between pollution and incomeshould vary across pollutants according totheir perceived damage. For example, wemight expect a very low x for directly life-threatening pollutants such as contaminateddrinking water. In this case, the EKC wouldbe (almost) monotonically decliningthroughout, as was found by Shafik (1994)and Grossman and Krueger (1995).Alternatively, x might be very high for pollu-tants whose harm is uncertain or delayed.Carbon emissions may fit this category, andindeed most studies have found that carbonemissions per capita tend to increase monot-onically with per-capita income.27

Threshold Effects. An alternative explana-tion for the EKC is based on thresholdeffects. Threshold effects can arise in eitherthe political process as in Larry Jones andRudolfo Manuelli (1995), or in abatementopportunities as in Andrew John andRowena Pecchino (1994), Selden and Song(1995), and Nancy Stokey (1998). Thresholdeffects lead to a very different relationshipbetween income and pollution in early ver-sus later stages of development. At low lev-els of economic activity, pollution may beunregulated entirely or regulation may havelittle impact on the profitability of abate-ment. Pollution therefore at first rises withgrowth. But after some threshold has been

breached, and policy is either implementedor starts to bind, these models predict pollu-tion declines with income—provided appro-priate assumptions are imposed on tastesand technology.

There are at least two possible ways toensure regulation is ineffective in checkingpollution at low levels of income. The first isto assume an abatement production functionwhere the marginal product of abatement isbounded. In this case, there will exist a set ofrelatively low pollution taxes for which firmschoose the zero abatement option; conse-quently, even though taxes may rise withgrowth over some range, this has no affecton abatement, and pollution rises with eco-nomic activity. An abatement function of thistype was used in Copeland and Taylor (1994)and is implicit in Stokey (1998). The modelwe presented in section 1 contains thisattribute because we note from (3) that if t <ap, no abatement occurs. This means thatthere is no technique effect to offset thescale effect of growth when incomes are low.In models with only one good, this ensuresthat pollution must rise with growth at lowlevels of income. In multi-good models, thepollution tax can still play a role by alteringthe composition of output at low levels ofincome, but it becomes much more effectiveonce the abatement threshold is breached.

Alternatively we can assume a fixed cost toeither abatement or policy. Suppose there isa fixed cost CR of setting up a pollution reg-ulation system. When national income is low,the aggregate willingness to pay to reducepollution to its first best level may be lessthan the fixed regulatory cost CR, in whichcase it is not worth setting up a regulatorysystem. With no system in place, pollutionrises lock-step with output.

Both threshold theories need two furtherassumptions to generate an EKC. First weneed an assumption on the growth processto restrict composition effects. This is typi-cally done by adopting a one good frame-work or by limiting substitution possibilitiesby functional form assumptions. We will


28 See Copeland and Taylor (2003) for an explicit devel-opment of this model.

consider neutral growth. Second, we needan assumption on preferences to ensure thatonce abatement occurs, the response of pol-lution taxes is sufficiently income elastic. Inour framework, this requires an income elas-ticity of marginal damage in excess of one.

With these two assumptions in hand, con-sider the impact of growth via neutral tech-nological progress. In the policy thresholdmodel, the demand for pollution shifts out astechnology improves and income rises. Thenet benefits of reducing pollution increasewith income because we have assumed theelasticity of marginal damage with respect toincome is greater than one, and hence therewill be a critical income level at which it isworth setting up a regulatory system.Further income gains then lower pollution.This simple model predicts a discreteimprovement in environmental quality at thecritical point; however, by introducingadjustment costs, we could obtain a smoothresponse.28

In the abatement threshold model, pollu-tion taxes rise with growth and eventuallyfirms move off their corner solution.Abatement occurs and further increases inincome drive pollution downward. Hencewe obtain an inverse-U relation between pol-lution and income that is kinked at its peak.

Since threshold explanations also rely onincome effects, they bear a close familyresemblance to the income-effects explana-tion. Both explanations rely on a strong poli-cy response to income gains as developmentproceeds, but they differ in their explana-tions for the initial rising segment.Threshold effect explanations predict a peri-od of inactivity in pollution policy and/or pri-vate sector responses to policy whereas theincome effect explanations predict small butincreasingly tougher policies and higher pol-lution abatement costs over time.

Increasing Returns to Abatement. A finalexplanation for the EKC is increasing

29 See Copeland and Taylor (2003) for such a model.

returns in abatement. The argument is sim-ply that as the scale of abatement rises itsefficiency may increase. These efficienciesmake abatement more profitable and henceeven if pollution policy is stagnant andunchanging pollution can fall as more abate-ment is undertaken. James Andreoni andArik Levinson (2000) develop this idea with-in a one-good endowment model anddemonstrate how this process can lead to anEKC. This explanation carries with it aninteresting twist on scale and techniqueeffects because as the scale of output rises,even with constant pollution taxes, firmsswitch to cleaner techniques of production.The scale effect creates its own techniqueeffect even with no pollution policy responseto higher incomes. As such, this theoryshares a common feature with the sources ofgrowth explanation in that an EKC pollutionprofile is compatible with no change in pol-lution policy over the development path.

In Andreoni and Levinson’s endowmentmodel, issues of market structure arisingfrom the increasing returns technology donot arise, but one can relatively easily extendtheir increasing returns to abatement expla-nation to allow for perfectly competitivefirms by using either industry-wide learningby abating29 or by employing the methods ofMarkusen (1989) and introducing interme-diate goods.

The Role of International Markets. Each ofthe theories we discussed above could gen-erate an EKC with no international trade,but without trade it becomes more difficultfor higher income countries to shed dirtyproduction. Hence, it is useful to considermore explicitly how trade affects the EKC.

One key role for international trade is tooffer an alternative abatement mechanism.Access to world markets offers an easyabatement alternative—import the goodfrom abroad when higher pollution taxesmake it more expensive at home.Consequently, trade makes pollution


30 See Kenneth Arrow et al. (1995)

demand more elastic and pollution becomesmore responsive to changes in policy.

As well, international markets create linksbetween country pollution levels, and thishas important implications for the interpre-tation of the EKC. In the income effectsexplanation for the EKC, rich countries canreduce their pollution either by abatingmore or by using policy to encourage dirtyindustry to migrate to poorer countries. Ifthe former process is the main driving force,then all countries could follow a similar path.But if it is the latter, then even if an EKCexists for rich countries, the newly industri-alized countries may not replicate the expe-rience of the current rich.30

A natural concern is whether country-spe-cific explanations are consistent with theoverall cross-country evidence. One relative-ly uncharted branch of theoretical researchis investigating whether one-country (orsmall open-economy) explanations given forthe EKC add up. That is, is there a fallacy ofcomposition lurking in the background? Atpresent, we know of no research addressingthese concerns head on, but existing resultsin the literature suggest further work may beneeded. For example, the Copeland andTaylor (1994) pollution-haven model pre-dicts a very different relation betweengrowth and pollution in autarky than in freetrade. If the income elasticity of marginaldamage is one, the scale and techniqueeffects of growth exactly offset each other inautarky, so growth has no effect on pollution.In contrast, in free trade, with the same pref-erences and technology, growth in the Northraises both Northern and Southern pollu-tion; and growth in the South lowers bothNorthern and Southern pollution. Theseresults follow from composition effects cre-ated by differences in pollution regulationsacross countries. Further work along theselines would be useful and indeed a multi-country perspective would seem to be a

31 See the surveys cited previously, as well as DavidStern (1998).

32 See also Gary Koop and Lise Tole (1999) who find noevidence for any empirical relationship between deforesta-tion and per capita income. Recent work by William Brockand M. Scott Taylor (2003) shows that while all countrieswill exhibit an EKC type relationship, countries will differin their turning points and rates of environmentalimprovement whenever they differ in initial conditions,their rate of natural regeneration, etc.

necessity if the pollutant in question hasstrong transboundary or global effects.

The Empirical Evidence. As noted above,there has been a flurry of recent empiricalwork linking economic growth to environ-mental outcomes. Much of this work hasfocused on either confirming or denying theexistence of similar relationships across dif-ferent pollutants, considering additionalexplanatory variables, such as incomeinequality or political freedom, or investigat-ing the robustness of previous studies.31

A growing body of work has found that theestimated EKC is sensitive to the sampleused. Harbaugh, Levinson, and Wilson(2002) examine the robustness of estimatesof the EKC for SO2. This was the focus ofthe original work of Grossman and Krueger(1993) and is one of the most widely citedexamples of the existence of an environmen-tal Kuznets curve. They find that the shapeof the curve is sensitive to changes in thetime period chosen and the set of countriesincluded in the study. This is suggestive of amisspecification of the model, which is whatthe theory developed above suggests—dif-ferent sources of growth across countriesshould yield different reduced-form rela-tions between pollution and income. DavidStern and Michael Common (2001) use dataon sulfur emissions in 73 countries over 31years, and by comparing OECD and non-OECD subsamples conclude that the evi-dence does not support a common EKCacross countries. Similarly, John List andCraig Gallet (1999) study sulfur dioxide andnitrogen dioxide emissions in U.S. statesover 65 years and conclude that there is nota common EKC across states.32


33 Lurking in the background of this study is a compo-sition effect operating through changes in the fleet of cars.This composition effect is not investigated in the paper,although it may be responsible for the jump in lead pergallon of gasoline use at low income levels shown in figure2 of the paper.

Despite the proliferation of papers in thisarea, very little work has gone into evaluat-ing the various hypotheses offered for theEKC, or more generally in examining howthe interaction between different sources ofgrowth interact with income and othereffects to determine the relation betweengrowth and pollution. Unless we can clarifythe causal mechanisms involved, the workwill be of little use in helping us understandhow growth or trade affect the environment.

A few recent studies are useful steps inthis direction. Hank Hilton and ArikLevinson (1988) examine the link betweenlead emissions and income per capita usinga panel of 48 countries over the twenty-yearperiod 1972–92. This study is importantbecause it finds strong evidence of aninverted U-shaped relationship betweenlead emissions and per-capita income, andthen factors the changes in pollution intotwo different components. The first is atechnique effect that produces an almostmonotonic negative relationship betweenlead content per gallon of gasoline andincome per capita. The second is a scaleeffect linking greater gasoline use to greaterincome.33 This study is the first to providedirect evidence on two distinct processes(scale and technique effects) that togetherresult in an EKC.

To interpret the empirical evidence asreflecting scale and technique effects, oneneeds to rule out other possibilities.Although the authors do not couch theiranalysis in this context, their analysis implic-itly presents the necessary evidence. First,they document a significant negative rela-tionship between the lead content of gaso-line and income per capita (post 1983). Thisrelationship shows up quite strongly in just asimple cross-country scatter plot of lead con-

34 To be precise we should note that since lead contentper gallon is an average, and cars differ in their use of lead-ed versus unleaded gas, the composition of the car fleet islikely to be changing as well. Therefore, the fall in averagelead content may reflect an income-induced change in theaverage age of the fleet (which would lower average leadcontent) plus a pure technique effect.

tent against income per capita. We havedepicted this in figure 2 below.

Since lead content is arguably pollutionper unit output, it is difficult to attribute thenegative relationship in this figure to muchother than income-driven policy differ-ences.34 Our interpretation is simply thatregulation is tighter in higher-income coun-tries and this is driving down lead content(or e, emissions per unit output, in ourframework).

Second, the authors find a hump-shapedEKC using data from the post-1983 period,but in earlier periods they find a monotoni-cally rising relationship between lead emis-sions and income. The declining portion ofthe EKC only appears in the data once thenegative health effects of lead become wellknown. The emergence of the declining por-tion in the income–pollution relationship isvery suggestive of a strong policy response tothe new information about lead. The factthat this appears late in the sample makes itdifficult to attribute the decline in lead toother factors that could be shifting thedemand for pollution. For example, if thedeclining portion of the EKC was due toincreasing returns to scale in abatement,then it should appear in both the pre andpost-1983 data. If it was due to shifts in thecomposition of output arising naturally alongthe development path, why would it onlyappear in the post-1983 data? While it ispossible to think of examples where theseother factors are at play, the scope for mis-taking a strong policy response for some-thing else is drastically reduced in this study.We are therefore led to conclude that thedeclining portion of the EKC post-1983reflects a strong induced policy responsethat more than offsets the scale effect.


Figure 2. Real GDP Per Capita, 1992

Lea

d pe

r ga

llon

(gra

ms)

0 10000 20000

0

3.5

Another paper that attempts to determinewhether an income effect is responsible forthe downward turn of an EKC is KishoreGawande et al. (2000). They estimate anEKC for hazardous waste sites in the UnitedStates, and find that it is hump-shaped,although only a small percentage of countiesare on the downward-sloping portion.Because it is very expensive to clean up haz-ardous waste sites, they argue that theincome effect would be reflected in net out-migration rates. They find evidence that thenumber of hazardous waste sites in a regionincreases the net out-migration rate but onlyafter a threshold of income is reached, whichis consistent with an income effect drivingthe downward portion of the EKC.Moreover, the income threshold they esti-mate is indistinguishable from the peak oftheir estimated EKC.

Gale and Mendez (1998) attempted toassess the importance of composition effectsin predicting cross-country differences inpollution levels. They re-examine one yearof sulfur dioxide data drawn from Grossmanand Krueger’s (1993) study. They regresspollution concentrations on factor endow-

ment data from a cross-section of countriestogether with income-based measuresdesigned to capture scale and techniqueeffects. Their results suggest a strong linkbetween capital abundance and pollutionconcentrations even after controlling forincomes per capita. Their purely cross-sec-tional analysis cannot differentiate, however,between location-specific attributes andscale effects. Nevertheless, their work isimportant because the strong link betweenfactor endowments and pollution suggests arole for factor composition to affect pollu-tion demand. That is, even after accountingfor cross-country differences in income lev-els that may determine pollution supply,other national characteristics matter to pol-lution outcomes.

Finally, two recent studies attempt toassess the relative importance of scale, tech-nique and composition effects in accountingfor changes in pollution. Selden, AnneForest, and James Lockhart (1999) compareemissions of six air pollutants in the UnitedStates in 1970 and 1990 and decompose theobserved changes in pollution into changesin scale, composition of economic activity


35 They actually consider more than one techniqueeffect. They measure changes in energy intensity and acompositional effect reflecting changes in sources of ener-gy as well as an “other technique effect.”

36 They also investigate compositional effects within themanufacturing sector by regressing per capita income onthe average pollution intensity in manufacturing, holdingthe pollution intensity in manufacturing subsectors con-stant. They find that the average pollution intensity firstfalls with income, and then levels out.

(due to changes in sectoral output shares),and changes in emissions per unit of out-put.35 Although this is simply a measure-ment exercise based on aggregate data for asingle country at two points in time, thestudy is nevertheless important because ittakes seriously the need to investigate therelative strength of the three effects. Theyfind that technique effects were an impor-tant factor in explaining the fall in emissions.Although composition effects were present,they were not strong enough to account forthe downturn in aggregate emissions duringthis period.

Hettige, Mani, and Wheeler (2000) usepanel data on industrial water pollutionfrom twelve countries and try to isolate com-position and technique effects, and explainhow they vary with income. They decom-pose pollution into the manufacturing pollu-tion intensity, the share of manufacturing inthe economy, and total output, and then sep-arately regress firm level pollution intensi-ties, the average pollution intensity inmanufacturing, and the manufacturingshare on per capita income.36 They find ahump-shaped relation between the share ofmanufactures and per capita income; how-ever, they find this composition effect issmall in magnitude relative to the impact ofscale effects. Conversely, they find a strongtechnique effect: the income elasticity of thepollution intensity is about –1. Overall, theyfind that industrial water pollution tends toinitially rise with income and then flattenout, with the strong technique effect beingresponsible for offsetting the scale effect ofgrowth.

Summary. The EKC literature is impor-tant in several respects: it brought theempirical study of aggregate pollution levelsinto the realm of economic analysis; itdebunked the commonly held view thatenvironmental quality must necessarilydecrease with economic growth; and it pro-vided highly suggestive evidence of a strongpolicy response to pollution at higherincome levels. The literature expanded rap-idly because of the ease of estimation andthe potential relevance of its findings.Studies replicating or extending the methodsof early contributors have played a usefulrole in providing a check on the originalwork, but further work along these lines haslimited usefulness. Investigators must nowmove beyond the methods that sparked theliterature to develop methods useful inrevealing the causal mechanisms underlyingthe relationship.

To proceed further more guidance fromtheory is surely needed. We would expectthat scale, technique and compositioneffects all play a role in determining the rela-tionship between growth and the environ-ment. This suggests the focus on reducedforms linking only per capita income to pol-lution is unlikely to be fruitful. If we are toask more detailed questions of the pollutiondata, we will need different methods. Wesuggest a step back from the EKC methodsto consider theories determining the equi-librium level of pollution as a function of arelatively few factors. An approach that triesto disentangle the scale, technique and com-position effects, and which allows these tovary across countries has much more sup-port from theory and is more likely to gener-ate an increased understanding of whatdrives the relationship between growth andthe environment.

3. Trade Liberalization and theEnvironment

We now turn to the impact of interna-tional trade on the environment. We draw


37 See, for example, Gene Grossman and ElhananHelpman’s (1991) book.

38 We have followed the literature here in examiningseparately the effects of growth on the environment andthe effects of trade on the environment. More generally,we expect that growth, trade, and environmental qualitywill all interact with each other in interesting ways.However, there is currently little work available exploringthese interactions in growth models where comparativeadvantage is determined by both environmental policy andfactor endowments.

the usual distinction between trade andgrowth: trade liberalization changes relativegoods prices by opening up the economy toincreased foreign competition, while growthincreases endowments or improves technol-ogy at given external prices.

While this distinction is clear, it may notalways be accurate. There is empirical evi-dence that trade liberalization also stimu-lates economic growth, and at a theoreticallevel, trade can alter the rate of growth if itspurs innovation or factor accumulation.37

In addition, trade may also pave the way forlabor and capital mobility and technologytransfer. Hence, trade can set in motionforces that shift the production frontier aswell. For clarity however we will maintainthe distinction drawn above.38

We first examine the effects of trade onthe environment in a small open economyfacing fixed world prices to emphasize threemajor points. First, the effect of trade liber-alization on the environment depends on acountry’s comparative advantage, which inturn depends on country characteristics.There is no reason to expect trade to havethe same effect on all countries. Second, theeffects of trade on the environment dependon whether environmental policy is rigid orinstead responsive to changes brought aboutby trade. When policy is rigid we will showthat outcomes depend on the type of envi-ronmental policy instruments used by regu-lators. Finally, the welfare effects of tradeliberalization are sensitive to both a country’scomparative advantage and the flexibility ofits policy regime.

39 This approach has been frequently used in the tradeliterature. See, for example, Paul Samuelson (1954), andRudiger Dornbusch, Stanley Fischer, and Samuelson(1977).

40 For simplicity, we assume there are no trade barriersfor the numeraire good. This does not affect the qualitativeresults.

We then examine a two-region model toevaluate two of the major hypotheses in theliterature linking relative country character-istics to environmental outcomes: the pollu-tion haven hypothesis, and the factorendowments hypothesis. This then sets thestage for our review of the empirical work.

Trade Frictions. For modeling purposes,we need to be specific about the trade barri-ers that are being reduced. Some trade bar-riers (such as tariffs) generate revenue;others, such as distance, generate productiveactivities such as transportation to overcomethem; and yet others, such as bureaucraticdelays and regulations simply create tradingcosts. We don’t want to focus on the detailsof trade policy, but simply capture theeffects of increased opportunities to trade.To do so we assume there are some tradefrictions between countries, which we cap-ture by adopting an “iceberg” model of tradecosts.39

With iceberg costs, an importer who wantsto receive one unit of X from the foreigncountry has to ship 1+d units because d islost in transport. Trade therefore consumesreal resources, and the magnitude of tradefrictions increase as d rises.40

Trading costs drive a wedge between thedomestic and foreign price of a good. Asbefore let p denote the world price of X,then the domestic price of X for an import-ing country is:

(24)

Conversely, if Home exports X, then todeliver a unit of X to a foreign market (wherethe price is p), a home exporter must send1+d units, which are acquired locally at thedomestic price pe

d. Hence the domestic priceis lower than the foreign price:

p pmd = +( )1 d


41 More generally, we can distinguish between net andgross production frontiers, where the gross frontier repre-sents potential output, and the net frontier represents netoutput after resources are allocated to abatement. This dis-tinction is important if we want to capture techniqueeffects in the diagram. Details can be found in Copelandand Taylor (2003).

(25)

It is convenient for us to use “pd” to referto the domestic price, but the reader shouldkeep in mind that whether this price is aboveor below the world price depends on thecountry’s comparative advantage.

3.1 Rigid Policy

Fixed Emission Intensities. The effects oftrade liberalization on the environmentdepend on the environmental policy regime.We start with a simple case where govern-ment policy holds the emission intensity ofproduction fixed. This scenario is instructivebecause it simplifies the analysis by rulingout a technique effect, and may be a realisticapproximation of policy in many countries(at least in the short run) because much pol-lution regulation tends to target emissionsintensities, rather than overall emissions.This approach also allows for the special caseof no pollution regulation.

Consider a country importing the dirtygood X. The domestic price is initially abovethe world price, and as trade barriers fall, thedomestic relative price of X falls. As withgrowth, we can decompose the effects oftrade liberalization into scale, compositionand technique effects. This is illustrated infigure 3. The production frontier (for a givenemission intensity) is depicted in the top halfof the diagram,41 and pollution is measuredas a function of X in the bottom half.

Starting with producer price qo at point A,a trade liberalization reduces the domesticproducer price of X to q1. Production movesfrom point A to C, and pollution falls from zoto z2. If we measure the scale of output atworld prices p, then (hypothetical) move-

pp

ed =

+1 d

42 Note that trade liberalization generates a scale effect,even though it results in a movement along the productionfrontier, and not a shift of the frontier. In order to comparepollution across countries, we need to account for cross-country differences in the scale of production, whichmeans that we have to choose a quantity index to measurescale at common reference prices. As a result, any move-ment along a strictly concave production frontier will yieldboth composition and scale effects.

43 When both goods pollute, it is possible for the scaleeffect to dominate.

ments along the dashed line through AB(with slope p) keep the scale of the economyconstant. This allows us to decompose thechange in pollution into a composition effect(A to B) which lowers pollution from zo to z1,and a scale effect (B to C) which raises pol-lution from z1 to z2. There is no techniqueeffect in this example by assumption.

The scale effect is positive and tends toincrease pollution. Trade increases produc-tion efficiency (measured at world prices),and this leads to more output, and hencemore pollution.42 The composition effect isnegative, because protection is beingremoved from the polluting good, inducingproducers to shift towards the clean good. Inour simple model where only one good pol-lutes, the composition effect always domi-nates the scale effect, because tradeliberalization has an unambiguous effect onthe output of the polluting good.43 If theeconomy has a comparative advantage inclean goods, as in this example, trade is goodfor the environment.

If instead home exports X, then trade lib-eralization raises pd. Producers shift alongthe production frontier towards the dirtygood. This both increases the scale of pro-duction and shifts the composition of outputtowards the polluting good: both the scaleand composition effects reinforce each otherand lead to an increase in pollution.

In summary, with fixed emission intensi-ties, the composition effect is critical indetermining the effects of trade liberaliza-tion. Moreover, the sign of the compositioneffect is ultimately determined by a country’s


Figure 3. Fixed Emission Intensity

Y

C

B

Z1Z2

Z0

X

Z =e0x

q0

p

A

q1

comparative advantage. If a country has acomparative advantage in clean industries,then clean industries expand with trade; andconversely, if it has a comparative advantagein polluting industries, then dirty industriesexpand with trade.

Fixed Emission Permits. Now suppose thegovernment uses a marketable emission per-mit system to regulate pollution, and that itdoes not adjust the supply of permits inresponse to changes in the trade regime.Earlier, we noted the equivalence of permit

44 Rod Falvey (1988) obtains a similar non-equivalenceresult when comparing import tariff reform with importquota reform in models with multiple trade distortions.

and tax systems as a method of implement-ing the first best. But if we hold policyinstruments fixed in the face of shocks to theeconomy, this equivalence breaks down.44

If X is imported, then producers shifttowards the clean industry when trade is lib-eralized, as in the previous example. Thisreduces the derived demand for pollution.


45 Although full-blown marketable permit schemes arerelatively rare in practice, similar results would beobtained if the regulator enforces ambient air or waterquality standards. If trade liberalization increased thederived demand for pollution, a regulator enforcing a rigidair quality standard would respond by tightening up regu-lations, which would raise the shadow price of pollution.

46 The welfare effect of trade liberalization in the pres-ence of environmental distortions has been a major themeof the literature: see Baumol (1971), Pethig (1976), Siebert(1977), Kazumi Asako, (1979), and Copeland (1994).

47 If an inefficient instrument is used to control pollu-tion then there are three distortions to worry about: theinefficient choice of instrument; the inefficient level ofpollution; and the inefficiently low level of trade. Weassume the government uses efficient instruments; extend-ing our analysis to the inefficient instrument case is left tothe reader as an exercise.

For given emission intensities, this wouldreduce pollution. But with a fixed supply ofpollution permits, pollution will not change;instead the permit price falls. Producersswitch to dirtier production techniques. Thisnegative technique effect completely offsetsthe beneficial impact of the shift towardsproducing cleaner goods. Similarly, if X isinstead exported, trade liberalization leads toan (upward) adjustment in pollution permitprices, but has no effect on pollution.

An important implication of our analysis isthat with rigid pollution taxes or emissionsintensities, the environmental effects oftrade liberalization may be quite substantial.But if pollution quotas are in place, the envi-ronmental effects of trade liberalization maybe negligible.45

Welfare Effects of Trade Liberalization.The welfare analysis of trade liberalization inthe presence of environmental problemsdraws heavily on the theory of the secondbest (Richard Lipsey and Kelvin Lancaster1956).46 Prior to trade liberalization, thereare two types of distortions: trade barriersand inefficient pollution policy.47

Reductions in trade barriers can either alle-viate or exacerbate the problems caused byinefficient pollution policy. Consequently,standard gains from trade theorems do notapply.

To determine the welfare effects of tradeliberalization, consider the effects on the

48 If we instead model trade barriers as tariffs (t), thewelfare effects of trade liberalization can be written asdV/VI = tdM + (t - MD)dz. A reduction in tariffs will raiseimports, so the term tdM represents the gains from trade.As well, there is the effect on pollution noted above. Themagnitude of the gains from trade term differs accordingto the type of trade barrier in place; but the basic pointthat the welfare effects of trade liberalization depend onthe interaction between the trade and environmental dis-tortion is robust.

49 The standard “double dividend” literature (see LansBovenberg and Ruud de Mooij 1995; and Don Fullertonand Gilbert Metcalf 1997) considers the effects of pollu-tion regulation in the presence of distortionary taxes onlabor supply. That literature is also based on the interac-tion between two distortions.

utility of the representative consumer of asmall fall in the trade friction d. Replace pwith pd in (7), and differentiate to obtain:

(26)

Trade liberalization has two effects on wel-fare: the standard gains from increased tradeand the induced change in the environment.48

The standard gain from trade effect is pos-itive for both importers and exporters. IfHome imports X, then M > 0 and the domes-tic price of X falls with trade liberalization.As a result Mdpd < 0. If Home exports X,then M < 0 but the domestic price of X riseswith trade liberalization. Once again we findMdpd < 0.

Changes in pollution, however, can under-mine the benefits of trade liberalization. Tosee why, suppose emission intensities areconstant and regulation is lax (t < MD).Then if home exports X, pollution rises withtrade. Because the pollution tax is less thanmarginal damage, this increase in pollutionis welfare-reducing (t – MD)dz < 0. The netwelfare effect of liberalization is nowambiguous: the costs of increased pollutionhave to be compared with the benefits ofincreased goods consumption. If pollution issufficiently damaging, pollution costs willdominate.

On the other hand, if home imports X,trade liberalization may yield a double divi-dend by reducing pollution and generatingincreased consumption.49 With lax pollution

dVV

Mdp MD dzd

1

= - + -( )t


50 The main point here is that if pollution is regulatedwith quotas, trade liberalization cannot exacerbate the pol-lution distortion. A similar point was made earlier in theliterature on piecemeal trade policy reform, where Falvey(1988) showed that in an economy with multiple trade dis-tortions, alleviating one trade distortion will not exacerbateother trade distortions if import quotas are the instrumentof protection. See Copeland (1994) for further details ontrade policy reform in a world with many goods and pollu-tants. With pollution quotas in place, uniform tariff reduc-tions will increase welfare, but with pollution taxes (orfixed emission standards), then the welfare effects of tradeliberalization depends on whether trade protection isbiased towards clean or dirty goods. John Beghin, DavidRoland-Holst, and Dominique van der Mensbrugghe(1997) use a similar approach to investigate consumption-generated pollution, and Arja Turnen-Red and AlanWoodland (1998) consider multilateral reforms.

regulation (t < MD), the economy gainsfrom reduced pollution and the standardgains from trade.

The instruments used also play an impor-tant role in determining the welfare effectsof trade liberalization. If pollution regulationtakes the form of a binding aggregate pollu-tion quota, then trade must always raise wel-fare, even when marginal damage is highand pollution regulation is lax. Referring to(3.6), as long as the pollution quota is bind-ing, pollution does not change with trade lib-eralization, and hence we have (t – MD)dz =0. This leaves only the standard gains fromtrade, which as we have already shown mustbe positive.50

In summary, if pollution regulations areunresponsive, then the welfare effects oftrade liberalization depend on the pattern oftrade, the type of policy instrument used,and the existing stringency of pollution reg-ulation. If the number of pollution permits isheld fixed during trade liberalization, thenfreer trade has to raise welfare and has noenvironmental consequences. But if emis-sion intensities are unchanged with trade,then trade increases pollution in countrieswith a comparative advantage in dirty goods,and decreases it in countries with a compar-ative advantage in clean goods. And whenpollution policy does not fully internalizeexternalities, countries with a comparative

51 The analysis of the effects of trade liberalization onenvironmental quality has received less attention than theeffects on welfare. Lopez (1994), Rauscher (1997), andCopeland and Taylor (2003) consider these issues in moredetail.

advantage in dirty goods may lose fromtrade.

3.2 Flexible Policy

Now suppose government policy adjustsoptimally in response to changes in the traderegime. Pollution is determined by settingthe marginal benefit of polluting equal tomarginal damage as in (11), where the goodsprice “p” is interpreted as the domestic pricepd. When trade is liberalized, pd changes. Bydifferentiating (11), we obtain, after somemanipulation:

(27)

where D > 0, each of the elasticities is posi-tive, and the change in pd should be inter-preted as coming from a change in the tradefriction d. Trade liberalization yields both anincome effect (the first term) and substitu-tion effects.51

First, consider the income effect. Becausetrade liberalization raises real income (dI = - Mdpd > 0 when trade is liberalized) andenvironmental quality is a normal good, theincome effect will always tend to reduce pol-lution. The pollution supply curve shiftsback with liberalization and the strength ofthis income effect depends on the incomeelasticity of marginal damage, eMD,I.

The substitution effects of trade liberaliza-tion, however, move in opposite directionsfor dirty-good importers and exporters.There are two substitution effects, one inproduction and the other in consumption.On the production side, an increase in theprice of the dirty good stimulates productionof the dirty good, and this tends to increasethe demand for pollution. On the consump-tion side, an increase in the price of the dirty

dzp M

Idp dp

dMD I d G P MD P dz=

é

ëê

ù

ûú +

-é

ëêê

ù

ûúú

e e e, , , ,D D


52 See Copeland and Taylor (2003). The sign of the netsubstitution effect depends on preferences and technologyand hence can vary with assumptions.

53 One can show that if eMD,I £ 1, then pollution will risein a dirty-good exporter.

good raises the level of the price index forconsumption goods relative to the opportu-nity cost of environmental quality (given byt). Consumers would like to substitutetowards more environmental quality, and thepolicy maker responds by raising the pollu-tion tax (the pollution supply curve shiftsback for a dirty exporter). Therefore, the twosubstitution effects work against each other.One can show that the substitution effect inproduction dominates: the net substitutioneffect tends to increase pollution for a dirty-good exporter and reduce pollution for adirty-good importer.52

Putting the substitution and incomeeffects together, we can determine theimpact of trade liberalization on pollution.For a dirty-good importer, trade liberaliza-tion will reduce pollution. The increase inincome shifts back pollution supply, and thelower price of dirty goods leads to a reduc-tion in pollution demand. Both income andsubstitution effects combine to improveenvironmental quality.

On the other hand, for a dirty-goodexporter, pollution tends to rise via the sub-stitution effect (the demand for pollutionrises as the price of the dirty good rises) butfall via the income effect (the supply of pol-lution shifts back as real income rises). If theincome elasticity of marginal damage issmall,53 then pollution rises in a dirty-goodexporting country even though pollution pol-icy is fully optimal. Conversely, if the incomeelasticity of marginal damage is large, thenpollution falls.

If the income elasticity of marginal damageis increasing in real income, then we mightexpect pollution to rise in a low-income dirty-good exporter, but fall in a high-income dirty-good exporter (recall our example in (23)). Aswell, we would expect the policy response to

differ across pollutants, because both incomeand substitution effects will vary. In addition,the impact of trade liberalization would alsodiffer in settings where the policy processwas subject to threshold effects or if abate-ment exhibited increasing returns. We knowof no research, however, that investigates anyof these channels.

Welfare Effects of Trade Liberalizationwith Efficient Policy. When policy is setoptimally, then t = MD and the effect oftrade liberalization on welfare in (26)reduces to

We are left with the standard gains fromtrade. If pollution externalities are fullyinternalized, trade must always increase wel-fare. Trade may lead to an increase in pollu-tion if the economy has a comparativeadvantage in dirty goods, but this reflects anoptimal tradeoff between environmentalquality and consumption.

3.3 The Determinants of ComparativeAdvantage

It is apparent that composition effects playa key role in determining the effects of tradeon the environment. But composition effectsdepend on a country’s comparative advan-tage, and hence a major preoccupation ofthe literature has been an investigation ofwhich countries attract dirty industries whentrade is liberalized.

There are two major competing theories,although they are often not stated explicitly.The pollution haven hypothesis predicts thatcountries with relatively weak environmentalpolicy will specialize in dirty-industry pro-duction. In many versions of this hypothesis,countries with weak environmental policyare also low-income countries. An alterna-tive hypothesis is that environmental policyhas little or no effect on the trade pattern;instead, standard forces, such as differencesin factor endowments or technology, deter-mine trade. For example, under this view,

dVV

Mdpd

1

.= - ³ 0


54 We rule out other differences to highlight the inter-action between the pollution haven and factor endowmentmotives for trade.

capital-abundant countries tend to exportcapital-intensive goods, regardless of differ-ences in environmental policy. We will callthis the factor endowments hypothesis,although it can be interpreted more broadlyto encompass other motives for trade, suchas technology differences.

We can illustrate these competing theorieswithin our simple model by assuming thereare two regions in the world: “North” and“South.” We use an asterisk (“ ”) to denoteSouthern variables. North and South mayhave different factor endowments or pollu-tion policy, but we assume they are other-wise identical.54

The interaction between factor endow-ments and pollution policy in determiningthe pattern of trade can be illustrated usinga simple relative supply-and-demand analy-sis to determine autarky prices in each coun-try. To construct relative demand, note thatgiven our assumption that preferences overgoods are homothetic and separable fromenvironmental quality, the demand for X rel-ative to Y is independent of income and canbe written as RD(p), where RD’(p) < 0.Moreover, because preferences are identicalacross countries, the relative demand curveis the same in each country. This is illustrat-ed in figure 4 as the curve labeled RD.

Next, we need to determine the relativesupply curves for each country. Using (4)and exploiting constant returns to scale, wecan write relative supply as a function of K/Land prices:

(28)

This yields a standard upward-sloping rela-tive supply curve (increases in p increase thesupply of X relative to Y). Because North andSouth differ in factor endowments and pol-lution policy

trade, growth, and the environment - university of hawaiinoy/362texts/copeland-environment.pdf ·...

Documents