Introduction

1The ongoing talks and negotiations on the third phase (post-2012) of the European Union Emission Trading System (EU-ETS) have recently made reference to the possible use of border tariff instruments to protect European companies subject to this environmental regulation. In fact, the European Union’s decision to unilaterally introduce a pollution permits market results in a loss of competitiveness for European firms both in domestic markets and, for their exports, in foreign markets.

2An indirect consequence of this loss of competitiveness in Europe is an increase in emissions in the countries that are not subject to similar environmental regulation and which export to the European Union. In such instances, there is “carbon leakage.” The definition given by the UNFCCC (United Nations Framework Convention on Climate Change) is an increase in emissions in the countries that have not ratified Annex B of the Kyoto Protocol following the implementation of reductions in the Annex B countries. Such reductions can result from a substitution of goods sold, but also, in the longer term, from offshoring. One of the solutions put forward is to then introduce border tax adjustments. In this context of heated international negotiations, the World Trade Organization and the United Nations Environment Program recently released a joint report in which they explicitly mention such border adjustments and the possibility of adopting them under certain conditions that remain to be defined. [5]

3The introduction of border tax adjustments boils down to putting in place import taxes and export subsidies. Border tax adjustments are an oft-used mechanism to protect the competitiveness of a country’s firms when a tax is applied to the local production base. In practice, border tax adjustments are equal to the tax on the local production base. In this paper we consider an environmental tax to correct a global pollution externality generated by industrial production. A tax makes it possible to correct the externality produced in the country in question but does not reduce the externality produced abroad. We therefore seek to ascertain whether tax adjustments in the conventional sense of the term can be used in this context.

4In economic terms, equality between the import tax and the export subsidy and the permit price corresponds to a situation where domestic and foreign companies have no market power. The reason is that, under pure and perfect competition, the import tax and the export subsidy are equal to the tax on emissions. A government that maximizes domestic well-being and takes into account global pollution caused by domestic and foreign emissions from industries in pure and perfect competition chooses an import tax and an export subsidy equal to the permit price. The emissions tax corresponds to marginal damage and is therefore a Pigovian tax. One of the objectives of this article is to understand how the realistic consideration of imperfectly competitive market structures affects the level of environmental taxes and border adjustments.

5In fact, recent works on border adjustments have not addressed the value of taxes and subsidies, but have taken these as exogenous and equal to the emissions tax or to the permit price. For example, Fisher and Fox (2009) analyze the efficiency of such instruments in reducing emissions and carbon leakage. Their model is consistent with the theoretical basis for this equivalence: firms are price takers. The authors compare border tax adjustments with other instruments such as local production subsidies. They consider environmental policy as exogenous and show that the best instrument to combat carbon leakage is the simultaneous use of an import tax and an export subsidy.

6However, it is worth noting that studies of the effects of the introduction of a market for pollution permits on competitiveness and carbon leakage have used models with imperfect competition. The reason for this is that the sectors covered under the EU-ETS are generally considered to be oligopolistic and characterized by firms’ substantial market power. Thus, Demailly and Quirion (2005), Ponssard and Walker (2008), and Smale et al. (2006) examine the introduction of border adjustments in a context of imperfect competition using calibrated models. This paper aims to provide a more explicit illustration of the mechanisms that influence the choices of these different instruments. Domestic and foreign market structures have complex effects both on the nature and on the levels of these various border adjustment instruments.

7This article is also an extension of the literature studying the relationship between global trade and environmental policy (Barnett 1980; Barrett 1994; Kennedy 1994; Ulph 1996; Straume 2006). However, in most cases these authors consider two-country models with one or two firms. They show that in a context of imperfect competition, trade liberalization can provide an incentive for governments to adopt strategic behavior. In fact, in the absence of conventional trade policy instruments, governments that seek to protect the competitiveness of firms may be tempted to relax their environmental policies. They therefore set an emissions tax below the Pigovian benchmark. This strategic behavior by governments is known as ecological dumping (or eco-dumping). Governments can also opt for a rigorous environmental policy. Resource allocation is not optimal in either case.

8This paper is structured as follows. The first section introduces the model. The second section examines the unilateral introduction of an environmental regulation and its consequences in terms of carbon leakage. An analysis of the optimal permit price will also be provided. The third section analyzes how the use of an import tax and an export subsidy affects domestic well-being with a given environmental policy. It is shown that a country that implements an environmental regulation has a clear incentive to introduce a tax on imports. In addition, when the government uses all instruments, it is shown that the optimal import tax is higher than or equal to the permit price and equal to marginal pollution damage. It is found that, when the number of firms in each market is high, the three instruments have the same value. The fourth section analyzes border tax adjustments according to two efficiency criteria, namely global well-being and the elimination of carbon leakage. A conclusion is provided in the final section.

Model

9Consider two geographical regions: the home region and the foreign region, denoted H and F respectively. Each region contains one sector of activity, which is homogenous between countries. Trade between the two regions is free and there are no initial trade barriers. The number of firms in country H (respectively, F) is n_H (respectively, n_F). Firms in both countries are symmetrical. Firms engage in Cournot competition. Let q_iL and x_iL denote the quantity produced by firm i and sold in country L∈{H,F}, respectively. Q_L (respectively, X_L) indicates the total quantity produced by firms in country L and sold in country L (respectively, exported). Production costs (excluding costs due to environmental regulation) are standardized at zero. [6] Demand functions are assumed to be linear and given by

10

11where P_H and P_F denote the prices of goods sold in countries H and L, respectively.

12We assume that a_F<a_H: the size of the market is smaller in region F. Three cases are of particular interest. When the number of firms is high in each country, the case in which firms have no market power can be studied. This case approximates to perfect competition. We also wish to study cases where the structure of the domestic market is very concentrated (n_H = 1), with a small number of foreign firms (n_F = 1) and with a large number of foreign firms. In fact, this corresponds to situations where the polluting industrial sector in question may be exposed to international competition to greater or lesser degree (for instance, electricity versus cement).

13Production technology emits pollution: one unit of goods produced generates one unit of CO₂. [7] To reduce emissions, the government of country L may wish to introduce a range of measures, such as an emissions tax or export subsidies, which will modify firms’ perceived marginal costs.

14Let C^x_L (respectively, C^l_L) be the marginal total cost of exports (respectively, local production) for a firm in country L. The problem facing firm i in country H for example can be written as

15

16The degree of market integration then depends on the value of the perceived marginal cost borne by firms. We can then deduce that when , at equilibrium, firms in country H export to region F; similarly, if , then firms in F export their goods. There is bilateral trade when both conditions are met simultaneously. There is unilateral trade when only one condition is met. There is autarky when firms have no incentive to export. Below we will focus on the more interesting case of bilateral trade.

17Emissions of pollution in one country stem from that country’s total production, sold locally and exported, E_L = Q_L + X_L. Thus, when no environmental regulation is in place (and therefore C^x_L = C^l_L = 0), pollution in each region is

18

19We consider an environmental damage function that depends on the sum of global emissions, λ(E_H + E_F), where λ is the marginal damage generated by global warming. If there is too much damage, it may be beneficial to sacrifice production of goods and, in doing so, the consumer surplus and firms’ profits derived from those goods. To ensure that the quantity produced is positive and that there is bilateral trade in all configurations studied below, we assume that λ < a_F.

20The government in country H seeks to maximize a well-being function given by

21

22where SC_H denotes the consumer surplus in this country, the profits of firms in this region, and RR_H tax receipts (positive in the case of a tax, or negative in the case of a subsidy) derived from the environmental regulation.

Unilateral emissions regulation

23When country H is in autarky and its firms are price takers, a government that maximizes well-being chooses a tax that is equal to marginal damage – this is the principle of a Pigovian tax. In the presence of imperfect competition and international trade, the unilateral introduction of a regulation gives rise to carbon leakage and the optimal tax differs from marginal damage since it is used with multiple objectives.

24We consider that, to emit CO₂, firms in region H must pay amount σ per emission unit. Since there is no uncertainty in this economy, there is equivalence between price and quantity instruments (Weitzman 1974). σ then represents, indifferently, either a tax or an emissions permit price. Environmental policy can be decided directly by a government or it can be exogenous. An exogenous permit price would reflect, for example, an international permit price imposed on the sector in question. An endogenous tax, on the other hand, would correspond to a situation in which the government devises a regulation for a particular sector.

25Domestic firms pay the tax for units that are both exported and sold on the local market. Such a firm’s program is then written as

26

27Firms’ production costs in country F are zero

28

29Since here we have C^x_H = σ and C^l_F = 0, and according to the conditions seen in the previous section, the permit price that enables firms in country H to export is such that

30

31At equilibrium, the quantities produced by each country and sold in the various markets, under condition (1), are

32

33A firm in region H has a marginal cost equal to the permit price for any unit produced, whether for local production or for export. Introducing a permit price therefore results in an increase in domestic firms’ marginal costs, causing them to reduce their production. Because of strategic substitutability, foreign firms increase their production. However, the net effect is a contraction in the quantity sold in each market. Since a unit of production is equal to a unit of emissions, total emissions decrease. As expected, the decrease in pollution is accompanied by decreases in the consumer surplus and in the profits of firms in country H.

34Carbon leakage. In addition to disadvantaging the countries that introduce pollution permit markets, the presence of international trade leads to a relocation of pollution from one region to another. Carbon leakage in region F, which is the difference between [the pollution emitted in that region in the case of a permit market in country H] and [the pollution emitted when there is no environmental regulation in place in H] is written as

35

36Carbon leakage appears only if there is international trade between the countries. When both countries export, carbon leakage is equal to . The variations in exports and local production in the two regions, when firms in country H export, are here each equal to . Carbon leakage increases with the number of firms in each region. The larger the number of firms in region H, the greater the effect on the individual quantities of firms in country F, and the greater the increase in total production and therefore pollution in country F. An increase in the number of firms in country F gives rise to two effects: a decrease in the effect of the increase in costs in country H on production choices in country F, as well as an increase in total production. The second effect outweighs the first.

37The permit price that maximizes well-being in country H is given by

38

39where RR_H = (Q^*_H + X^*_H)σ. The profits of firms in country H decrease with the increase in the permit price. Similarly, the consumer surplus decreases with σ. However, in the case where trade is unilateral, total emissions are lower than those under bilateral trade. When the permit price is such that firms in country H continue to export, the optimal price is

40

41Three effects influence the choice of optimal permit price σ^*:

• When environmental damage increases, the permit price increases. The price leads to a decrease in total emissions and, as a result, in environmental damage.
• The permit price decreases with the size of the domestic market. To understand this effect, consider the hypothetical situation in which country H is isolated from international trade (for example a_F = 0 and n_F = 0) and has no environmental concern (for example λ = 0). In this case, the government is then faced with an imperfectly competitive industry that produces a homogenous good sold to its consumers. In order to reduce the distortions associated with industry’s market power, the government is then prompted to subsidize industry. Returning to our framework, this incentive to subsidize tends to decrease the permit price so as to not provide an excessive incentive for industry to decrease its production.
• The permit price can depend positively or negatively on the size of the foreign market, i.e. . This derivative depends on the relative market structures of the two countries. When the number of domestic firms is higher, it is positive. To understand this result, consider another hypothetical situation in which country H does not produce for local consumption and exports all its production (for example a_H = 0); like above, let us assume that country H has no environmental concern. The government must choose the instrument that helps its industry’s exports to be more competitive. Its objective is to maximize domestic firms’ revenues since the cost of the tax for firms equates to the government’s tax revenues.
  This framework corresponds to the strategic trade model developed by Brander and Spencer (1985), with the important difference that there are several firms in imperfect competition in both countries. [8] When the government decides to subsidize its industry, there is both a price effect and a volume effect. If the number of domestic firms is higher than the number of foreign firms, the price effect prevails and it is in the government’s interests to introduce a tax. In the opposite case, the volume effect prevails and the government therefore introduces a subsidy. This effect leads the government of country H to tax (to subsidize, respectively) its industry when its industry is more (less) competitive than that of country F.

42Thus, introducing a pollution permits market in one country modifies trade and, as a result, the trade balance of that country, the result being a decrease in global pollution. This is borne solely by domestic firms. Border tariff instruments could therefore be envisaged.

Border tax adjustments

43We will first examine the effect of introducing import taxes and export subsidies for the country that conducts a given environmental policy.

44Let τ be a unit export tax on exports from F to H. Goods produced by firms in country F are either sold on the local market or exported. In the second case, firms bear the tax on their exports. Let s be the unit subsidy allocated by the government of region H to the firms that export to F. A firm from region H bears a marginal cost that is equal to the permit price for any unit produced for local production and a marginal cost that is equal to the permit price minus the export subsidy. First, we consider the pollution permit price to be exogenous.

45At equilibrium in a situation of bilateral trade, production is written as

46

47Total production of region H, Q_H + X_H, increases because of the protectionist measure and the other region produces less. Total emissions decrease in region F and increase in H.

48Let be the value of the import tax above which country F no longer exports. [9] It is not in the government’s interest to set a tax higher than because this instrument only affects exports and such a value eliminates them.

49How does a government determine the optimal border tax adjustments? The revenue of the government in region H derives from sales of permits and receipts of the import tax minus the subsidies distributed to exporters, or RR^*_H = (Q^*_H + X^*_H)σ + τ X^*_F – sX^*_H. Well-being is written

50

51We then obtain the following deconstruction.

52Lemma 1. When environmental damage is linear and environmental policy is exogenous, the two border instruments (export subsidy and import tax) are independent from one another and domestic well-being can be written

53

54where

55

56and

57

58The two trade policy instruments can therefore be isolated. The import tax only affects exporting firms in country F and local firms and consumers in region H. Similarly, the export subsidy only affects consumers in country F as well as profits realized locally by region F and profits realized on firms’ exports from H. As total pollution depends on total production, the linearity of our model (function of environmental damage and production costs) makes it possible to isolate production affected by the subsidy and production affected by the tax.

59What incentives does the government have to set a positive tax? The effect of an import tax on well-being is transmitted via four channels: the income from the tax, the consumer surplus, domestic firms’ profits in the domestic market and the level of emissions:

60

61An increase in the import tax increases the marginal cost of foreign competitors and leads to an increase in the price of the good in the domestic market. Thus, the impact on consumers is negative and their surplus decreases. However, domestic firms’ profits increase because their competitors have higher marginal costs. The import tax has the effect of reducing foreign emissions and therefore has a positive effect on environmental damage. The effect of the tax on tax receipts includes a price effect and a quantity effect. In fact, if the tax is high, an increase in the tax reduces receipts. However, at low values of the tax this effect is positive.

62It is worthwhile to study the derivative of well-being with respect to the import tax for low values of this tax:

63

64Thus, starting from a situation in which no border tax is in place, the government has a strictly positive incentive to introduce such a tax. This incentive remains when environmental damage is zero: gains realized in terms of firms’ profits and government revenue outweigh the loss of surplus borne by consumers. As a result, there are two types of incentives to implement a border tax: strategic and environmental.

65Proposition 1. A country that introduces emissions permits has a strict incentive to set an import tax. The optimal value of this tax T*, at an exogenous permit price, is equal to [10]

66

67In principle, the optimal tax can be higher or lower than marginal environmental damage; its level relative to the permit price, considered here as exogenous, is also ambiguous. Market structures are decisive in determining these relative levels.

68The optimal tax increases with environmental damage but also with the permit price. The higher the environmental damage, the higher the tax must be to reduce emissions and, in doing so, total environmental damage. The level of the tax must increase in order to offset an increase in the permit price and therefore firms’ lost profits.

69When the number of firms is high in each market, we find the conventional result whereby the tax is equal to the permit price. In the case whereby each market is monopolistic, the tax is equal to and is therefore greater than marginal damage (because ). When the number of foreign firms is high and when there is only one domestic firm, the import tax tends towards marginal damage.

70Let us turn to the impact of an export subsidy on the well-being of country H. The effect of an export subsidy on well-being is transmitted via two channels: the impact on domestic firms’ exports and the level of emissions.

71

72An increase in the subsidy does not necessarily give rise to an increase in export profits for domestic firms since it modifies the sale price of these exports. We also have

73

74The derivative of the well-being function for a low export subsidy value can be positive or negative. When the effect on environmental damage prevails (because marginal damage is high), the value is negative. In the opposite case, the value depends on the relative number of firms in each country. When the number of foreign firms is higher, the government has an incentive to introduce an export subsidy.

75Proposition 2. In certain cases a country that introduces emissions permits may have an incentive to introduce an export subsidy. The optimal value of this subsidy s*, at an exogenous permit price, is equal to

76

77The optimal subsidy increases with the permit price and decreases with environmental damage. Subsidizing exports amounts to lowering domestic firms’ marginal cost of production for exported goods. They already bear the permit price. Thus, the level of the optimal subsidy depends on the price level of the permits bought by companies. By lowering the marginal cost of production, the subsidy leads to an increase in domestic production and to a decrease in foreign production. However, the net effect on global production and therefore on pollution is positive.

78The export subsidy can depend positively or negatively on the size of the foreign market: . This effect depends on the relative market structure of each country. When the number of domestic firms is higher, this effect is negative. In certain cases the subsidy can increase the sum of firms’ profits net of the subsidies distributed by the government. When both market structures are competitive, the subsidy is equal to the permit price.

79We can conclude from this that under certain conditions, such as high environmental damage and a more concentrated foreign market structure than the domestic market structure, export subsidies equal to the permit price should not be introduced. In the general case, the relative levels of the permit price and the export subsidy are ambiguous.

80Another question can be addressed: can the introduction of border taxes modify environment policy in the long term? In fact, environmental regulation can be considered exogenous only over a certain period. Environmental and trade policy choices are not necessarily made at the same time. This leads us to study the effect of introducing import taxes and export subsidies on a country’s environmental regulation (τ > 0 and s > 0). Like in the previous section, we seek to determine the value of the permit price that maximizes well-being at given values of the import tax and the export subsidy. The optimal permit price is equal to

81

82where σ^* is the optimal price without border adjustments.

83Lemma 2. When the government introduces border adjustments, it has an incentive to tighten its environmental policy.

84The optimal permit price increases with the import tax and the export subsidy. The subsidy eases the global environmental constraint and the import tax reduces foreign emissions but increases domestic emissions. Thus, over time, the introduction of border adjustments requires the government to revise its national environmental policy.

85Now, consider a government that has these three instruments at its disposal. This case is interesting because it makes it possible to isolate the different effects. In fact, thanks to these three instruments, the government can more accurately correct the various distortions in pollution and competition. The optimality conditions of these three instruments are written as follows

86

87We then obtain:

88Proposition 3. When the government uses the three instruments simultaneously, the optimal import tax is higher than or equal to the permit price and equal to marginal pollution damage. At the optimum, the three instruments assume the same value when both markets are competitive (high n_H and n_F). At equilibrium, exports in country H are zero.

89It should be noted that the optimum corresponds to a situation in which foreign firms do not export (X_F = 0). This stems from the fact that domestic and foreign firms have the same production costs. It is then optimal for the government to ban domestic imports in order to reduce total emissions. When the government uses the three instruments simultaneously, the import tax is equal to marginal environmental damage and therefore to the Pigovian tax. The permit price is still lower than marginal environmental damage but tends towards it when the number of domestic firms increases. [11] In fact, the permit price is used both to correct the distortions in market power in country H and to reduce pollution. In this context of pure and perfect competition, as the permit price only corrects the externality, its price is equal to environmental damage. Likewise, the export subsidy corrects two distortions, market power in country F and pollution. If the number of foreign firms is high, the export subsidy then equates to the Pigovian tax.

90Let us now consider four extreme cases, in which the number of firms in each market can be either one or high. We set the values of the three instruments at the optimum.

91

92In the absence of market power, the three instruments assume the same value. We conclude from this that in a context of imperfect competition, the tax adjustment in the conventional sense of the term should not be used. We show in the Appendix that the export subsidy is never higher than the import tax. Subsidizing exports eases the environmental constraint faced by domestic firms. The relative levels of the subsidy and the permit price depend on marginal damage. High marginal damage requires a higher permit price than the subsidy to correct the externality. It should, however, be noted that a higher export subsidy than the permit price corresponds to a situation in which exporting firms are not subject to environmental regulation and yet are subsidized.

Discussion

93In addition to the criterion of maximizing domestic well-being studied above, we consider two further criteria, namely global well-being and carbon leakage. The first criterion is the concern of supranational organizations. We seek to analyze the extent to which the import tax and the export subsidy are necessary from this viewpoint. We also wish to use a purely environmental criterion. Carbon leakage appears to be one of the biggest obstacles to unilateral environmental regulation. Moreover, when environmental policy is exogenous and the only available instruments are border tax adjustments, the emissions generated in country F and, more specifically, carbon leakage are a legitimate concern.

94Maximizing global well-being. Global well-being, W_H + W_F, is the sum of each country’s well-being function. Global well-being takes into account the profits of all firms, the consumer surplus in both markets, tax receipts, and environmental damage in both countries. Above we showed that the well-being functions of each country could be written as the sum of two functions, one determined by the import tax and the other by the export subsidy. Thus, when the government takes into account a global well-being criterion, both border instruments are still independent of each other when environmental policy is exogenous and the damage function linear.

95Eliminating carbon leakage. We seek to determine the import taxes and export subsidies that eliminate carbon leakage

96

Figure 1

Optimal Results According to the Criteria to be Obtained at Given Environmental Policy

Optimal Results According to the Criteria to be Obtained at Given Environmental Policy

97It should be noted that a number of solutions are conceivable when the import tax and the export subsidy are used simultaneously. However, we consider the levels of the taxes and the subsidies that respectively and independently cancel out the variation in the country’s emissions stemming from market H and market F.

98We present the overall results in Figure 1 and address the cases where the import taxes and export subsidies are used independently and simultaneously.

99Introducing an import tax increases domestic market power and reduces foreign CO₂ emissions. Symmetrically, the introduction of an export subsidy reduces market power in the foreign country and eases the domestic environmental constraint. We note that the optimal tax can be negative and therefore equate to an import subsidy for foreign firms when environmental damage is estimated to be low and domestic firms have substantial market power. When the number of firms in country F is high, the subsidy is low and can even be negative if environmental damage is very high.

100We note that the subsidy can be higher than the permit price when the market structures are uncompetitive. If both market structures are competitive, then no export subsidy should be introduced and an import tax may be recommendable if damage is high.

101Let us now turn to the results relating to the elimination of carbon leakage. The tax-subsidy pairing that eliminates carbon leakage both in the domestic market and the foreign market is such that the value of the tax is between one-half and one times the permit price, , while the subsidy is equal to the permit price. Thus, only production sold locally in country H is subject to the permits market and environmental regulation. The cost of the regulation is borne entirely by firms that do not engage in foreign trade.

102However, when the government uses only one border instrument, the value of this instrument is higher than the permit price. The tax that eliminates carbon leakage is between one and two times the permit price, . If there is a monopoly in region H, the import tax is equal to the permit price, τ = σ. The tax increases with the number of companies. The tax that eliminates carbon leakage is higher than the permit price. This also stems from the fact that the tax has no impact on imports when the permit price is borne by firms in country H both for exports and for production sold locally. Thus, to offset the effects generated in the two markets, the tax needs to be higher than the permit price. Conversely, only using the subsidy requires setting a value equivalent to double the permit price. This is explained symmetrically to the previous result. Eliminating carbon leakage with the sole use of the subsidy results in sacrificing domestic environmental policy.

Conclusion

103The unilateral implementation of environmental regulation in one country leads to a deterioration in the trade balance of that country and to a decrease in global pollution. The country has a clear incentive to introduce a tax on imports. However, it does not always have an incentive to introduce an export subsidy. Market structures play an important role in determining these incentives.

104Following the introduction of border adjustments, the government would want to renegotiate its environmental policy and make it more restrictive. When it has all the instruments at its disposal, the government has an incentive to use tax adjustments in the conventional sense of the term (that is, equality between import tax, export subsidy, and permit price) only in the absence of market power. In the presence of market power, the import tax should always be higher than the permit price and the subsidy, which raises the issue of whether such a measure would be politically acceptable. Our work has illustrated that market structures are important in determining the incentives and the effects of the introduction of border adjustments.

105Lastly, the use of border tax adjustments comes up against numerous hurdles in terms of the legitimacy of these instruments under WTO rules (Pauwelyn 2007). In addition to this legal obstacle, individual governments have an incentive to use these instruments to maximize national well-being and not global well-being.