The political economy of trade liberalization and environmental policy.

Fredriksson, Per G.

1. Introduction

The global trading system is currently undergoing another round of multilateral trade liberalization, and regional economic integration is occurring in several pans of the world. The trade-environment link has become an important pan of trade liberalization talks. It is commonplace for industry representatives to argue that less stringent environmental regulations give less developed countries (LDCs) a comparative advantage in pollution-intensive goods and that industries with high abatement costs will migrate to LDCs. Environmentalists argue that international differences in environmental laws and regulations may be reduced to their lowest common denominator through political pressure.(1) This is not a fear of trade liberalization itself but of the effects of lower trade barriers on the political determination of environmental regulation.

There is little empirical work explicitly studying the effect of openness on the stringency of environmental regulations. Husted and Logsdon (1997) report that the North American Free Trade Agreement (NAFTA) has induced Mexico to strengthen its environmental policies.(2) In simulations by Perroni and Wigle (1994), trade policy is found to have little impact on the quality of the environment, given the level of environmental regulation. However, they found that global increases of pollution taxes to the optimal levels would improve environmental quality significantly. It follows that, if environmental policy determination is influenced by changes in the trade policy regime, environmental quality will be affected.

This paper investigates the underlying forces determining environmental policy, in particular the nature of the political economy effects of trade liberalization on environmental tax policy. In the model used, the pollution tax policy choice of the government is influenced by industry and environmental lobby groups. The lobby groups offer the government prospective political contributions, the size of which corresponds to the pollution tax policy selected (see Bernheim and Whinston [1986] for the original model and Grossman and Helpman [1994], Fredriksson [1997, 1998], Schleich [1997], and Aidt [1998] for applications). In addition to contributions, the government values aggregate social welfare.

This paper contributes to the literature by predicting and explaining the political equilibrium pollution tax policy in tariff-protected sectors when pollution abatement is possible, thus extending Fredriksson (1997) and Aidt (1998). The politically determined tax is compared to the welfare-maximizing pollution tax. Whereas the latter tax is found to be inefficiently stringent because of the need to correct the economy's two distortions (pollution and trade protection) with only one policy instrument, the politically determined tax depends also on the relative political pressures in equilibrium.

Second, I show the effects of trade liberalization on the special interest groups' lobbying incentives on the pollution tax issue. Ceteris paribus, when output contracts in the polluting sector following trade liberalization, both lobby groups reduce their lobbying on the tax issue in the new equilibrium because the marginal return to a change in the tax decreases.(3) Thus, this paper contributes to the literature on "ecological dumping," which discusses the government's incentive to set environmental regulations such that marginal abatement costs deviate from the marginal damage from pollution. In a model of lobbying with sector-specific interest groups, Rauscher (1994) finds that export industry lobby groups have ambiguous lobbying incentives.(4)

A change in environmental policy has implications for environmental quality. I show that, if the pollution tax falls (increases) sufficiently as a result of trade liberalization, pollution increases (decreases) through the political channel discussed.(5) These are new effects of economic integration not previously discussed in the literature.

Fourth, I study the effect of trade liberalization on pollution tax revenues. Several OECD countries are currently considering implementing or raising pollution taxes in order to raise revenues (see Pearce 1991; Oates 1993; Barde and Owens 1996). I study the impact of freer trade on the amount of pollution tax revenues raised in political equilibrium. Tax revenues may fall due to the political economy effect on the pollution tax rate. This is consequently not a traditional Laffer curve effect. I also show that pollution may increase simultaneously as tax revenues decrease. Finally, the political economy effects on environmental policy may yield counterintuitive effects on trade patterns of lower trade protection.(6) Changes in tax policy accompanying trade reform may counteract or reinforce the basic effects expected from trade liberalization.

Hillman and Ursprung (1992, 1994), Leidy and Hoekman (1994), and Schleich (1997) discuss political economy links between trade regimes and environmental policy. Hillman and Ursprung examine how environmental concerns influence the political determination of trade policy decisions. Leidy and Hoekman discuss the impact of environmental instrument choice on trade barriers. They argue that industry lobby groups will favor inefficient pollution control because this leads to an increase in the level of protection.(7) Schleich discusses the choice between environmental and trade policy instruments in the presence of consumption and production externalities.

The paper is organized as follows. Section 2 outlines the model. Section 3 explains the game and characterizes the political equilibrium. Section 4 shows the welfare-maximizing tax rate, which is contrasted with the political equilibrium. It also examines the impacts of trade liberalization on lobbying, the pollution tax policy, environmental quality, pollution tax revenues, and trade patterns. Section 5 concludes.

2. The Model

A small, open, competitive economy has two sectors: one produces the nonpolluting numeraire good z, the other produces the polluting good x. The economy is populated by industrialists (I), environmentalists (E), and workers (W). The population is normalized to one. All individuals have labor income, industrialists have factor income from production of good x, and environmentalists derive disutility from the pollution associated with the local production of good x. The environmentalists have an interest in the physical quality of the local environment because they are affected by pollution themselves.(8) Industrialists' and workers' utilities are given by

[U.sub.I] = [c.sup.zI] + u ([c.sup.x]),

[U.sub.w] = [c.sup.zW] + u([c.sup.x]), (1)

and the environmentalists' utility is given by

[U.sub.E] = [c.sup.zk] + u([c.sup.x]) - [Theta]X, (2)

where [c.sup.zk] is consumption of the numeraire good z by an individual of type k [element of] I, W, or E and [c.sup.x] is the consumption of good x, with world market price equal to one and [p.sup.*], respectively.(9) u([c.sup.x]) is a strictly concave and differentiable subutility function.

Production of good x is given by X, and [Theta] is per-unit damage that is a function of per-unit abatement A, that is, [Theta] = [Theta](A). The disutility suffered by an environmentalist equals total damage. Abatement requires labor alone and has decreasing returns, that is, [[Theta].sub.A] [less than] 0, [[Theta].sub.AA] [greater than] 0.(10) The government is restricted to one environmental policy tool, a pollution tax t [element of] T, T [subset] R, levied on pollution from production. Whereas the numeraire sector enjoys free trade, imports of good x have an ad valorem tariff denoted by [Tau] [greater than] 0, and thus the domestic price in the latter sector equals [p.sup.*](1 + [Tau]).(11) The ad valorem tariff is assumed to be determined by multilateral trade negotiations in which this small country has negligible bargaining power, and therefore the tariff rate is taken as given by all domestic agents.

Industrialists' and workers' consumption of x is given by the individual demand curves, which are the inverse of [u.sub.[c.sup.x]]. The aggregate demand curve for good x is given by D([p.sup.*](1 + [Tau])) and the aggregate consumer surplus derived from consumption of good x is

C([p.sup.*] (1 + [Tau])) = [summation over [for every]I, E, W] u [D([p.sup.*](1 + [Tau]))] - [p.sup.*] (1 + [Tau])D([p.sup.*](1 + [Tau])).

The remaining income is spent on good z.

Each individual has one unit of labor, and the total labor endowment equals l. The numeraire good z is produced by labor alone with constant returns to scale and an input - output coefficient equal to one. The economy's labor supply is sufficiently large for the supply of z to be positive, which implies that the equilibrium wage rate equals one. Disregarding labor costs, each producer faces a net price given by

p = [p.sup.*](1 + [Tau]) - t[Theta](A) - A. (3)

The inputs into production of good x are labor and an immobile sector-specific input.(12) The technology is constant returns to scale. With a wage rate equal to one and the world market price fixed, the specific factor reward [Pi] depends only on the producer's price, that is, [Pi](p). By Hotelling's lemma, the supply curve for good x is given by [[Pi].sub.p](p) = X(p), where [X.sub.p] [greater than] 0 and [X.sub.pp] [greater than] 0. Given t, the first-order condition of each firm's profit function with respect to abatement, omitting any firm-specific notation, yields

d[Pi]/dA = -X(t[[Theta].sub.A] + 1) = 0. (4)

Taking the total derivative of Equation 4 yields

dA/dt = [[Theta].sub.A]/t[[Theta].sub.AA] [greater than] 0. (5)

Assuming [[Theta].sub.AAA] is not too negative, [d.sup.2]A/d[t.sup.2] is unambiguously negative and A is a concave function of t, A = A(t). We define the tariff and tax revenue function as R(t) = t[Theta]X(p) + [Tau]M(t), where M(p) = [p.sup.*][D([p.sup.*](1 + [Tau])) - X(p)] is the net import demand function.(13) For analytic convenience, the revenues generated are assumed redistributed lump-sum and equal to all individuals.(14)

Environmentalists and industrialists are assumed to organize into lobby groups coordinating campaign contribution offers to the government, whereas workers have no stake in the policy outcome and thus do not organize.(15) There is no free-riding within the organized lobby groups, and the lobby groups do not cooperate. Let i denote lobby group type and [[Alpha].sup.i] the fraction of the population with membership in lobby i.(16) The lobby groups make campaign contributions denoted by [[Lambda].sup.i](t), i [element of] E, I, contingent on the pollution tax set by the government. Workers do not make campaign payments. The gross (of contributions) indirect utility functions of the environmental and industry lobby groups are

[[Omega].sup.E](t) [equivalent to] [[Alpha].sup.E] [-[Theta]X(p) + C([p.sup.*](1 + [Tau])) + R(t) + l], (6)

[[Omega].sup.I](t) [equivalent to] [Pi](p) + [[Alpha].sup.I][C([p.sup.*](1 + [Tau])) + R(t) + l], (7)

respectively, where [[Alpha].sup.i][C([p.sup.*](l + [Tau])) + R(t) + l] is the share of total consumer surplus, tax and tariff revenues, and labor income of lobby group i. The gross aggregate social welfare attained at policy t, ignoring political contributions, is given by

[[Omega].sup.A](t) [equivalent to] l + C([p.sup.*](1 + [Tau])) + [Pi](p) + R(t) - [[Alpha].sup.E][Theta]X(p). (8)

The incumbent government is assumed to maximize a weighted sum of aggregate campaign contributions and aggregate social welfare. Campaign contributions are used for campaign advertising, whereas a higher average social welfare increases the probability of re-election. However, in some societies, the contributions paid may take the form of illegal bribes or gifts.(17) Only interest groups make contributions, and hence the government's utility function is given by

[V.sup.G](t) = [summation over i [element of] E, I] [[Lambda].sup.i](t) + a[[Omega].sup.A](t),(9)

where a [greater than or equal to] 0 is the government's exogenously given weight on aggregate social welfare relative to revenues and political contributions.

3. The Political Equilibrium

In stage one of a two-stage game, each lobby group i offers the government a contingent political contribution schedule taking the other lobby's strategy as given. A lobby i's strategy is a differentiable function [[Lambda].sup.i]: T [approaches] [R.sub.+]; it offers the government a nonnegative contribution for selecting a policy t. In stage two, the government maximizes its utility, given the strategies of the lobby groups. The government selects its optimal tax policy and receives the corresponding contribution from each lobby. The lobby groups receive payoffs represented by the function [[Omega].sup.i]: T [approaches] [R.sub.+]. The lobby groups keep the commitments made in the first stage.

A Nash equilibrium is a policy [t.sup.[convolution]] and a set of contribution schedules [{[[Lambda].sup.[i.sup.[convolution]]]}.sub.i], where each contribution schedule is feasible, that is, nonnegative and lower or equal to the lobby group's members' aggregate income; [t.sup.[convolution]] maximizes the government's welfare [V.sub.G](t) taking the contribution schedules as given; and given lobby j's schedule and the government's anticipated decision rule, no lobby i has a feasible dominant strategy.

Following Grossman and Helpman (1994), I characterize the political equilibrium using Lemma 2 of Bernheim and Whinston (1986), which shows all necessary and sufficient conditions for a subgame perfect Nash equilibrium. The two necessary conditions used to find the political equilibrium are(18)

[t.sup.[convolution]] maximizes [summation over i [element of] E, I] [[Lambda].sup.[i.sup.[convolution]]](t) + a[[Omega].sup.A](t) on T; (C1)

[t.sup.[convolution]] maximizes [[Omega].sup.j](t) - [[Lambda].sup.[j.sup.[convolution]]] (t) + [summation over i [element of] E, I] [[Lambda].sup.[i.sup.[convolution]]](t) + a[[Omega].sup.A](t) on T for j [element of] E, I. (C2)

Assume that the political contribution schedules are differentiable around the equilibrium point. Take the first-order conditions of C1 and C2 and substitute the expression resulting from the maximization of condition C1 into the first-order condition of C2. It follows that, around [t.sup.[convolution]],

[Mathematical Expression Omitted]. (10)

This equation implies that, at the equilibrium point, the political contribution schedules are locally truthful.(19) Substituting Equation 10 into the first-order condition of Equation 9 (i.e., using condition C1), we obtain

[Mathematical Expression Omitted], (11)

which is the characterization of the pollution tax that emerges in the political equilibrium when the political contribution schedules are differentiable.(20)

4. Results

In order to have a benchmark, I first find the tax rate set by a social welfare maximizing government. Note that the Pigouvian tax equals [[Alpha].sup.E] since this is the marginal social disutility from pollution.

PROPOSITION 1. The pollution tax set by a social planner satisfies

[t.sup.*] = [[Alpha].sup.E] + [Tau][p.sup.*][[Gamma].sup.[convolution]]/([[Gamma].sub.[convolution]] + [[Psi].sup.[convolution]])[[Theta].sup.[convolution]], (12)

where [[Gamma].sup.[convolution]] = ([X.sub.p][p.sub.t][t.sup.[convolution]])/X [less than] 0 represents the tax elasticity of output and [[Psi].sup.[convolution]] = ([[Theta].sub.A][A.sub.t][t.sup.[convolution]])/[[Theta].sup.[convolution]] [less than] 0 represents the tax elasticity of pollution intensity.

PROOF. The partial derivative of Equation 8 equals ([[Alpha].sup.E] - t) ([[Theta].sup.2][X.sub.p] - [[Theta].sub.A][A.sub.t]X) + [Theta][X.sub.p][Tau][p.sup.*] = 0, which yields Equation 12 after conversions to elasticities and rearrangements. QED.

The tax rate set by a social planner unambiguously exceeds the Pigouvian tax [[Alpha].sup.E] since the second term in Equation 12 is positive. This can be interpreted as an application of second-best theory. The economy has two distortions: a tariff and pollution from production. However, the pollution tax is the only available policy instrument. Since the tariff causes an excessive amount of polluting domestic production, the second-best pollution tax must exceed the Pigouvian tax rate in order to discourage production.(21)

I define political pressure (or pressure) as the intensity of a (lobby) group's political effort to change the policy in the political equilibrium. This is formally represented by Equation 10. We now solve for the political equilibrium pollution tax policy characterized by Equation 11.

PROPOSITION 2. The equilibrium pollution tax policy satisfies

[t.sup.[convolution]] = [[Alpha].sup.E] (1 + a)([[Gamma].sup.[convolution]] + [[Psi].sup.[convolution]])[[Theta].sup.[convolution]] + ([[Alpha].sup.E] + [[Alpha].sup.I] + a)[Tau][p.sup.*][[Gamma].sup.[convolution]]/([[Alpha].sup.E] + [[Alpha].sup.I] + a)(1 + [[Gamma].sup.[convolution]] + [[Psi].sup.[convolution]])[[Theta].sup.[convolution] - (1 + a)[[Theta].sup.[convolution]]. (13)

PROOF. Take the partial derivatives of (6), (7), and (8) and substitute the

results into Equation 11. Using Equation 4 yields

(1 + a)[[[Alpha].sup.E]([[Theta].sup.2][X.sub.p] - [[Theta].sub.A][A.sub.t]X) - [[Theta].sub.X] + ([[Alpha].sup.E] + [[Alpha].sup.I] + a)[[Theta]X + [Theta][X.sub.p] ([Tau][p.sup.*] - [Theta]t) + t[[Theta].sub.A][A.sub.t]X] = 0, (13[prime])

which characterizes the political equilibrium tax policy. Rearrangements and conversions yield Equation 13. QED.

Even with the two lobby groups present, the government is concerned about the two distortions in the economy, as in Equation 12.(22) However, lobbying activities distort this equilibrium and [t.sup.[convolution]] [not equal to] [t.sup.*] is likely unless all individuals are organized in lobby groups, in which case [t.sup.[convolution]] = [t.sup.*] (see Grossman and Helpman [1994] for a discussion of this feature of the model). The first term in the numerator of Equation 13 represents the positive political pressure of the environmental lobby group on the tax. Ceteris paribus, this pressure is higher the greater the marginal effect on pollution of a change in the tax and is represented by ([[Gamma].sup.[convolution]] + [[Psi].sup.[convolution]]). The second term in the numerator represents the taxpayers' pressure for tariff revenues (where lobby group members have a weight of [1 + a]). The greater are [Tau] and [[Gamma].sup.[convolution]] (in absolute value), the greater the incentive to raise the tax rate and thus tariff revenues through increased imports. The first term in the denominator represents the pressure for pollution tax revenues. This is a positive (negative) pressure on the tax policy if tax revenues are increasing (decreasing) in the tax rate (represented by (1 + [[Gamma].sup.[convolution[greater than] [less than] [[Psi].sup.[convolution]]) [greater than] ([less than]) 0). The environmental and industry lobby group members favor higher revenues to the extent that these groups benefit (represented by lobby group size). The second term in the denominator is the industry lobby's pressure for a lower pollution tax. Since the equilibrium per-unit damage influences the effect on pollution, factor rewards, and tax revenues, all terms except the pressure for trade revenues are multiplied by [[Theta].sup.[convolution]](23) We now turn to the effect of tariff liberalization on the pollution tax.

PROPOSITION 3. Suppose that the polluting sector contracts following trade liberalization. In equilibrium, trade liberalization has three partial effects on the pollution tax: (i) It reduces the lobbying effort of the environmental lobby group; (ii) It reduces the lobbying effort of the industry lobby group; and (iii) It has a positive (negative) effect on the pressure for revenues if it increases (decreases) the tax elasticity of revenues.

PROOF. Taking the total derivative of Equation 13[prime] yields

[Mathematical Expression Omitted], (14)

where the denominator is required to be negative (see footnote 20). The first two terms in the numerator, (1 + a)[[Alpha].sup.E]([[Theta].sub.A][A.sub.t][X.sub.p] - [[Theta].sup.2][X.sub.pp]), represent the change in the lobbying by the environmental lobby group. Since their sum is negative, trade liberalization induces the environmental lobby group to reduce lobbying for a higher tax on the margin. The third term in the numerator, (1 + a)[Theta][X.sub.p], is positive and indicates that trade liberalization implies reduced industry lobby group pressure for a lower tax. The sum of the last four terms in the numerator, ([[Alpha].sup.E] + [[Alpha].sup.I] + a)([Theta][X.sub.pp](t[Theta] - [Tau][p.sup.*]) - 2[Theta][X.sub.p] - t[[[Theta].sub.A][A.sub.t][X.sub.p]), represents the change in the tax payers' pressure for tax revenues. If this sum is negative (positive), the tax payers increase (reduce) their pressure on the tax. QED.

The sum of the three partial effects in the numerator of Equation 14 is indeterminate. The intuition is the following. In equilibrium, the environmental lobby group's political pressure on the pollution tax falls with a lower import tariff because production falls ceteris paribus. This implies that, on the margin, a change in the pollution tax gives a smaller effect on pollution. The industry lobby group's political pressure falls with the tariff because, on the margin, the pollution tax affects fewer units of pollution when production falls. Thus, the impact of the tax on factor rewards is reduced.(24)

Figure 1 illustrates a special case of Proposition 3, where, for simplicity, we assume that the pressure for tax revenues is independent of the tariff rate.

The length of the left-hand (right-hand) side arrow represents the equilibrium political pressures from the industry (environmental) lobby group. Part (a) shows a case where the lobby groups' political pressures are equal (in absolute value) before trade liberalization. In part (b), the environmental lobby group's lobbying effort falls by more than the industry lobby group's lobbying effort, and [t.sup.[convolution]] falls as a result of trade liberalization. In part (c), the environmental lobby group's lobbying effort falls by less than the industry lobby group's lobbying effort, and [t.sup.[convolution]] increases.

We now explore the effects of trade liberalization on the lobbying incentives of the interest groups. I define political polarization as the difference in absolute value of the political pressures. This is illustrated by the distance between the end points of the arrows in Figure 1. Since the focus here is the behavior of the lobby groups, we maintain the assumption that the pressure for revenues is independent of the tariff.

PROPOSITION 4. Suppose that the polluting sector contracts following trade liberalization. In equilibrium, trade liberalization reduces the political polarization between the environmental lobby group and the industry lobby group on the pollution tax issue.

PROOF. This follows from the proof of Proposition 3. From Equation 14, since both lobby groups reduce the political pressure on the pollution tax, political polarization falls with a lower tariff. QED.

When the polluting sector contracts, trade liberalization reduces the level of political polarization on the pollution tax policy since both lobby groups' incentive to spend resources on obtaining a more favorable tax falls in equilibrium. This is illustrated in Figure 1 by the reduced distance between the end points of the arrows in parts (b) and (c) relative to part (a).

On the other hand, in a model where the polluting sector expands as a result of trade liberalization, it follows that the lobbying effort of both lobby groups and, consequently, the political polarization increases.(25)

We now show the impact of trade liberalization on pollution when the tax rate is endogenously determined. This can be illustrated with the help of an elasticity. Define [[Lambda].sup.[convolution]] = dt/d[Tau]([Tau]/t) as the equilibrium tariff elasticity of the pollution tax.

PROPOSITION 5. In equilibrium, trade liberalization increases pollution if [[Lambda].sup.[convolution]] is sufficiently positive.

PROOF. Pollution is given by [[Theta](A(t))X(p)]. Using the envelope theorem, differentiation yields

d[[Theta](A(t))X(p)]/d[Tau] = X[[Theta].sub.A][A.sub.t] dt/d[Tau] + [Theta][X.sub.p] ([p.sup.*] - [Theta] dt/d[Tau]), (15)

which, after rearrangement, can be shown to be positive iff

[[Lambda].sup.[convolution]] [greater than] -[Theta][X.sub.p][p.sup.*][Tau]/[t.sup.[convolution]](X[[Theta].sub.A][A.sub.t] - [[Theta].sup.2][X.sub.p]),

where both the numerator and the denominator are negative, and thus [[Lambda].sup.[convolution]] [greater than] 0. QED.

A lower tariff reduces output in the polluting sector, given the pollution tax. Proposition 5 shows, however, that if the equilibrium pollution tax falls sufficiently (i.e., if [[Lambda].sup.[convolution]] is sufficiently positive), pollution may increase. This occurs through two effects. First, pollution per unit of output increases. This is the first term on the right-hand side of Equation 15. Second, the fall in output is reduced and may even be reversed if [[Lambda].sup.[convolution]] is sufficiently positive. Output increases if the sum of the last two terms in Equation 15 is negative. This is a political route to increased pollution through trade liberalization.

It must be pointed out that if, on the other hand, [[Lambda].sup.[convolution]] [less than] 0, tariff reform has an extra "cleaning" effect in the polluting sector in addition to the effect of reduced output. Then per-unit pollution intensity falls, and output falls further than otherwise would be the case.

We now study the effects of trade liberalization on pollution tax revenues.

PROPOSITION 6. In equilibrium, (i) trade liberalization may reduce pollution tax revenues, and (ii) this may occur simultaneously as pollution increases.

PROOF. Pollution tax revenues equal [t[Theta](A(t))X(p)]. Using the envelope theorem, total differentiation yields

d[t[Theta](A)X(p)]/d[Tau] = ([Theta]X + t[[Theta].sub.A][A.sub.t]X) dt/d[Tau] + t[Theta][X.sub.p] ([p.sup.*] - [Theta] dt/d[Tau]), (16)

The effect on pollution is given by the three last terms. In equilibrium, pollution tax revenues are increasing in the tariff rate in two cases.

Case (a).

[Mathematical Expression Omitted], (17a)

where the numerator is negative and the denominator is positive.

Case (b).

[Mathematical Expression Omitted], (17b)

where both the numerator and the denominator are negative. QED.

In both cases (a) and (b), [[Lambda].sup.[convolution]] is indeterminate. A positive (negative) denominator in Equations 17a and 17b implies that the tax elasticity of pollution tax revenues is negative (positive), that is, we are on the right-hand (left-hand) side of the top of the revenue curve. The numerator is negative and represents the direct effect on pollution tax revenues from a change in output due to a tariff change. In case (a), a sufficiently large [[Lambda].sup.[convolution]] (i.e., [t.sup.[convolution]] falls sufficiently) implies a negative effect on pollution tax revenues of trade reform. This occurs because the pollution tax falls faster than pollution increases. If at the same time the sum of the last three terms of Equation 16 is negative, pollution increases simultaneously as pollution tax revenues fall. Case (b) requires a sufficiently small [[Lambda].sup.[convolution]] (i.e., [t.sup.[convolution]] increases sufficiently). Here the pollution tax increases at a lower rate than the decrease in pollution. It follows that, in this case, pollution does not increase simultaneously. Note that the effects are not traditional Laffer curve effects but are of a political economy nature.

We finally turn to an analysis of the effect on the pattern of trade.

PROPOSITION 7. Assume that the country imports the polluting good. In equilibrium, net imports of the polluting good may decrease with trade liberalization if the effect on the pollution tax is sufficiently positive.

PROOF. Using the envelope theorem, taking the total derivative of the net import demand function M(p), with respect to the tariff [Pi], yields

[Mathematical Expression Omitted], (18a)

which is positive if

[Mathematical Expression Omitted]. (18b)

QED.

Ceteris paribus, a lower tariff increases imports (demand increases, domestic output falls), which is represented by the first and second terms on the right-hand side of Equation 18a, respectively. From Equation 18b, this effect is reversed if the tax increases sufficiently in the tariff rate (and thus the tax falls with a lower tariff). This has a positive effect on output. If the domestic output increases by more than the domestic demand increases, imports consequently decrease.

The discussion in this paper has mainly focused on the case when the polluting good is produced by an import competing sector and thus freer trade implies a contraction of this sector ceteris paribus. Assume instead that the polluting good is the exported good (i.e., M(p) [less than] 0) and the form of protection is an export subsidy. This case is analogous to Proposition 7 in that exports increase if the tax falls sufficiently when the export subsidy is reduced. The effect on domestic output may outweigh the effect on domestic demand, and net exports increase.

In order to discuss the case when production in the polluting sector increases as a result of trade liberalization, assume that the country exports the polluting good and an export tax is levied.(26) It is easy to show that a reduction in the export tax increases lobby group polarization and, depending on the relative sizes of the increases in lobby group pressures, the pollution tax may rise or fall. The change in the pollution tax naturally determines the effects on pollution, pollution tax revenues, and trade patterns.

5. Conclusion

The paper presents a pressure group model with rival environmental and industry lobby groups that seek to influence the incumbent government's pollution tax policy in a tariff-protected sector. The theory developed explains and predicts the political equilibrium pollution tax policy, and this was contrasted to the tax set by a social planner, yielding a testable hypothesis of lobby group influence. Next, a number of effects of trade liberalization in an import-competing sector were identified. First, the political pressures on the pollution tax from the two lobbies fall as the tariff decreases, and the relative changes in lobbying efforts determine the final impact on the tax rate ceteris paribus. The pollution tax decreases (increases) with a lower tariff if the lobbying effort by the environmental lobby group decreases more (less) rapidly than the lobbying effort by the industry lobby group. I also found that, as a consequence, trade liberalization reduces the level of "political polarization" in the area of environmental policy, that is, the difference in the intensity of the lobby groups' equilibrium lobbying efforts.

Moreover, I showed that total pollution may increase with more liberal trade because the equilibrium pollution tax falls. Pollution tax revenues may also decrease when trade is liberalized. This may occur simultaneously as pollution increases. Finally, the expected effects on foreign trade may be reversed by the discussed changes in environmental policy. These political economy effects of trade liberalization are new contributions to the literature on trade and the environment.(27)

The model predicts that the change in lobbying activities on environmental issues following trade reform will depend on the resulting pattern of specialization. Countries with a comparative advantage in pollution intensive goods are predicted to experience an increase in political lobbying (in the sectors with increased output), and the opposite is true in countries shifting towards production in "clean" sectors following trade reform. Since "dirty" sectors tend to be capital intensive and "clean" sectors labor intensive, the model thus predicts that capital (labor) intensive countries will experience increased (reduced) levels of political conflict following trade reform. Another implication of this paper is that tariffs do not only distort consumption and production patterns but may increase the lobbying incentives in other areas of regulation, such as environmental policy, with a consequent waste of scarce resources.

I would like to thank Dicky Damania, Wilfred Ethier, Noel Gaston, Steven Lim, Jonathan Pincus, T. K. Rymes, two helpful referees, and seminar participants at the Universities of Adelaide, Sydney, Waikato, and Pennsylvania for insightful suggestions and comments. A travel grant from the University of Adelaide and partial funding from the Swedish International Development Cooperation Agency (Sida) is gratefully acknowledged. This paper was partially written while visiting the Department of Economics at the University of Pennsylvania. and I thank the Department for its hospitality. The opinions expressed are those of the author and not those of the World Bank or Sida. All remaining errors are my own.

1 The 1989 free trade agreement between Canada and the U.S. requires harmonization in technical regulations, standards, and test methods (Kaufmann, Pauly, and Sweitzer 1993). This applies to the licensing of pesticides. Canada has a more stringent licensing procedure than the U.S., and Canadian regulations do not require the consideration of economic benefits, which is mandatory in the U.S. (see Cropper et al. 1992).

2 For example, regulatory inspections of plants have increased from 1425 in 1990 to 13,993 in 1995 (Husted and Logsdon 1997).

3 On the other hand, if output expands (because the country has a comparative advantage in the polluting good), both lobby groups increase their lobbying.

4 Rauscher (1994). however, calls for a more explicit modeling of political decisions of environmental policy in open economies.

5 This is because when the tax rate falls (increases), pollution per unit of output increases (decreases). If, in addition, production in the polluting sector rises because the pollution tax falls sufficiently, the level of pollution unambiguously increases through trade liberalization. This is in contrast to Anderson (1992), who argues that tariff liberalization unambiguously reduces pollution.

6 In a static model, we expect output to fall in import competing and exporting sectors if they are protected by an import tariff and export subsidy, respectively.

7 The effects discussed in the present paper are of particular relevance to a small, open economy with a negligible impact on multilateral trade talks. Such a country must take the level of protection as given. Industry and environmental special interests, however, have a direct impact on environmental regulation. Hillman and Ursprung's (1992, 1994) and Leidy and Hoekman's (1994) research primarily applies to a large country with strong bargaining power in global trade negotiations.

8 The environmentalists may live in an area downstream on a river or in the direction of the wind leading from a polluting industrial plant or, alternatively, they may be altruists with a moral concern for the environment even though they are not personally harmed by pollution.

9 With quasi-linear preferences, corner solutions may arise. However, I assume interior solutions.

10 Subscripts denote partial derivatives.

11 In section 4, we discuss the case when the polluting good is exported and an export subsidy is used.

12 Many pollution-intensive industry sectors are resource-based according to United Nations Industrial Development Organization (UNIDO) (1982) (see Low 1992). These industries are immobile due to their main factor of production being a natural resource. The sectors are, for example, pulp and waste paper, petroleum products, organic chemicals, fertilizers, and nonferrous metals.

13 When the polluting good x is the exported commodity, the net import demand function is negative (see section 4).

14 In the Scandinavian countries, carbon taxes are channeled directly into general revenues (Organization for Economic Cooperation and Development [OECD] 1989). We recognize that this redistribution role does not represent a realistic description of all economies and that tariff revenues no longer are important for trade policy in most countries. As discussed below, this assumption does not drive our results, however.

15 See Olson (1965) for a discussion of lobby group formation.

16 Since the population is normalized to one, [[Alpha].sup.i] also represents the absolute number of members in lobby i.

17 Grossman and Helpman (1994) discuss some evidence of the corruptibility of politicians, and Potters and Sloof (1996) survey the empirical literature on the influence of interest groups. Pashigian (1985) shows how environmental policies are determined with political benefits in mind, and Cropper et al. (1992) find that lobbying has a significant impact on pesticide regulations in the U.S. Grier and Munger (1986, 1991), Endersby and Munger (1992), and Stratmann (1992) discuss the supply price of politicians' services and the relationship with lobby groups' strategies.

The game modeled here is a complete information game, and thus the campaign contribution schedules are known to all players (as in Grossman and Helpman [1994] and others). I have nothing to add to the model in this respect and recognize that the contributions may in some societies take the form of illegal bribes, and thus the informational assumption made does not correspond to all situations.

18 Condition C1 is taken from the definition of Nash equilibrium. It stipulates that the local government sets the pollution tax to maximize its own welfare, given the offered political contribution schedules. Condition C2 follows from the definition of the Nash equilibrium and states that the equilibrium tax maximizes the joint welfare of each lobby group j and the government, given the other lobby group's equilibrium contribution schedule. If this were not true, lobby j could alter its contribution schedule to induce the government to select the jointly optimal tax and capture most of the surplus from the change. Since the government would select the new tax and lobby j would benefit from the change, the original tax could not have been an equilibrium.

19 Intuitively, each contribution schedule is determined so that the marginal change in the contribution for a small change in policy matches the effect of the policy change on the lobby group's gross welfare, given that the contribution remains strictly positive. Thus, the curvatures of the contribution schedules accurately represent the lobby groups' preferences around the equilibrium. This implies that the contribution is the compensating variation (as long as it is positive). Dixit, Grossman, and Helpman (1997) show that, in a common agency model such as mine, a Pareto-efficient truthful equilibrium exists where the lobby groups use truthful strategies. They argue that such an equilibrium may be focal. Moreover, since there is complete information, truthful strategies is a way to obtain an efficient equilibrium in a noncooperative game.

20 For Equation 11 to be a maximum requires that the second-order condition satisfies [(1 + a)([[Alpha].sup.E][K.sub.1] + [K.sub.2]) + ([[Alpha].sup.E][K.sub.1] + [K.sub.2]) + ([[Alpha].sup.E] + [[Alpha].sup.I] + a)[K.sub.3]] [less than] 0, where [K.sub.1] = [Theta][[Theta].sub.A][A.sub.t](2[X.sub.p] + X) - [[Theta].sub.3][X.sub.pp] - X ([[Theta].sub.AA][[A.sub.t].sup.2] + [[Theta].sub.A][A.sub.u]) [less than] 0, [K.sub.2] = [[Theta].sup.2][X.sub.p] - [[Theta].s [[Theta].sub.A][A.sub.t] (2X + [X.sub.p] [[Tau][p.sup.*]) - 2[Theta][X.sub.p]([Theta] + t[[Theta].sub.A][A.sub.t]) + [[Theta].sup.2][X.sub.pp](t[Theta] - [Tau] [p.sup.*]) + tX([[Theta].sub.AA][[A.sub.t].sup.2] + [[Theta].sub.A][A.sub.u]) - t[[Theta].sub.A

21 The author is grateful to a referee for clarifying this interpretation.

22 Note that [t.sup.[convolution]] is unambiguously positive.

23 The impact of the lobby groups is illustrated by the differences between Equations 12 and 13.

24 Note that tariff revenues do not by themselves drive the results stated in Proposition 3. although these play a role for the final effect of trade liberalization on the pollution tax policy. Our focus is on lobby group behavior. In contrast to Fredriksson (1997). the availability of pollution abatement and the convexity of the supply function here play roles for the resulting effect of the domestic price on the lobby groups' behavior and thus on the tax policy.

25 See also the discussion below.

26 Grossman and Krueger (1993, p. 46) estimate that NAFTA will cause the U.S. and Canada to specialize in physical and human capital-intensive, pollution-intensive production. Mexico will specialize in relatively "clean" sectors (labor-intensive manufacturing and agriculture), a "composition effect."

27 The previously identified effects are the scale, composition, and technique effects (Grossman and Krueger 1993).

References

Aidt, T. S. 1998. Political internalization of economic externalities and environmental policy. Journal of Public Economics 69:1-16.

Anderson, K. 1992. The standard welfare economics of policies affecting trade and the environment. In The greening of world trade issues, edited by K. Anderson and R. Blackhurst. London: Harvester Wheatsheaf.

Barde, J.-P., and J. Owens. 1996. The evolution of eco-taxes. The OECD Observer 198:11-6.

Bernheim, B. D., and M. D. Whinston. 1986. Menu auctions, resource allocation, and economic influence. Quarterly Journal of Economics 101:1-31.

Cropper, M. L., W. N. Evans, S. J. Berardi, M. M. Ducla-Soares, and P. R. Portney. 1992. The determinants of pesticide regulations: A statistical analysis of EPA decision making. Journal of Political Economy 100:175-97.

Dixit, A., G. M. Grossman, and E. Helpman. 1997. Common agency and coordination: General theory and application to government policy making. Journal of Political Economy 105:752-69.

Endersby, J. W., and M. C. Munger. 1992. The impact of legislator attributes on union PAC campaign contributions. Journal of Labor Research 13:79-97.

Fredriksson, P. G. 1997. The political economy of pollution taxes in a small open economy. Journal of Environmental Economics and Management 33:44-58.

Fredriksson, P. G. 1998. Environmental policy choice: Pollution abatement subsidies. Resource and Energy Economics 20:51-63.

Grier, K. B., and M. C. Munger. 1986. The impact of legislator attributes on interest-group campaign contributions. Journal of Labor Research 7:349-61.

Grier, K. B., and M. C. Munger. 1991. Committee assignments, constituent preferences, and campaign contributions. Economic Inquiry 29:24-43.

Grossman, G. M., and E. Helpman. 1994. Protection for sale. American Economic Review 84:833-50.

Grossman, G. M., and A. Krueger. 1993. Environmental impacts of NAFTA. In The US-Mexico free trade agreement, edited by P. Garber. Cambridge, MA: MIT Press.

Hillman, A. L., and H. W. Ursprung. 1992. The influence of environmental concerns on the political determination of trade policy. In The greening of world trade issues, edited by K. Anderson and R. Blackhurst. London: Harvester Wheatsheaf.

Hillman, A. L., and H. W. Ursprung. 1994. Environmental protection and international trade policy. In The international dimension of environmental policy, edited by C. Carraro. Boston, MA: Kluwer.

Husted, B. W., and J. M. Logsdon. 1997. The impact of NAFTA on Mexico's environmental policy. Growth and Change 28:24-48.

Kaufmann, R. K., P. Pauly, and J. Sweitzer. 1993. The effects of NAFTA on the environment. Energy Journal 14:217-40.

Leidy, M. P., and B. M. Hoekman. 1994. 'Cleaning up' while cleaning up? Pollution abatement, interest groups and contingent trade policies. Public Choice 78:241-58.

Low, P. 1992. Trade measures and environmental quality: The implications for Mexico's exports. In International trade and the environment, edited by P. Low. World Bank Discussion Papers No. 159. Washington, DC: The World Bank, pp. 89-103.

Oates, W. E. 1993. Pollution charges as a source of public revenues. In Economic progress and environmental concerns, edited by H. Giersch. Berlin: Springer-Verlag.

OECD. 1989. Economic instruments for environmental protection. Paris: OECD Publications.

Olson, M. 1965. The logic of collective action. Cambridge, MA: Harvard University Press.

Pashigian, B. P. 1985. Environmental regulation: Whose self- interests are being protected? Economic Inquiry 23:551-84.

Pearce, D. W. 1991. The role of carbon taxes in adjusting to global warming. Economic Journal 101:938-48.

Perroni, C., and R. M. Wigle. 1994. International trade and environmental quality: How important are the linkages? Canadian Journal of Economics 27:551-67.

Potters, J., and R. Sloof. 1996. Interest groups: A survey of empirical models that try to assess their influence. European Journal of Political Economy 12:403-42.

Rauscher, M. 1994. On ecological dumping. Oxford Economic Papers 46:822-40.

Schleich, J. 1997. Environmental protection with policies for sale. IATRC Working Paper No. 97-2. Virginia Polytechnic Institute and State University, Blacksburg.

Stratmann, T. 1992. Are contributors rational? Untangling strategies of political action committees. Journal of Political Economy 100:647-64.

UNIDO. 1982. Changing patterns of trade in world industry: An empirical study on revealed comparative advantage. New York: United Nations.