Jean Monnet Center at NYU School of Law



Previous|Title|Next

III. Economic Analysis of Judicial Tests

In this Chapter, we analyze the current jurisprudence of the AB&P regarding ETMs from an economic viewpoint. In the first two Sections, we discuss some factors and assumptions which will be necessary in analyzing judicial tests from an economic perspective. Section A discusses the definition of efficiency which is used in the first three Sections of this Chapter, namely the potential Pareto efficiency (PPE), or otherwise known as the Kaldor-Hicks efficiency, which is commonly used in the field of law and economics. Section B analyzes what kinds of costs have to be paid when invoking ETMs and how the costs could be allocated among countries. This Section shows that ETM-imposing countries are prone to prefer measures which impose a heavy burden on international trade. In Section C, we evaluate the "means and ends" test and the LTRA test, which are judicial tests which the AB&P have invoked, comparing these tests with the comparative cost-benefit analysis, which is the best "trade-off device"152 in light of the Kalder-Hicks efficiency. In Section D, we analyze the LTRA and "means and ends" tests in light of problems PPE has.

A. Potential Pareto Efficiency (PPE)

In the field of economics, Pareto efficiency is normally utilized as the standard for efficiency. Pareto efficiency is defined as the situation where "goods cannot be reallocated to make someone better off without making someone else worse off."153 For example, suppose that there is a case in which Country A invokes an ETM by which it can gain benefits of $10 million, whereas the ETM inflicts $1 million damages on Country B's export to Country A. In this hypothetical example, whether this state of affairs (the situation prior to the invocation of the ETM) is Pareto optimal depends on the amounts of the transaction costs when Country A negotiates with Country B, asking it to put up with the trade harms caused by the invocation of the ETM. We assume that, except for Country B, no other countries are affected by the ETM. If the amounts of the transaction costs are less than $9 million, both Country A and Country B are better off when Country A compensates Country B by the amount exceeding $1 million. In this situation, Country B is better off since the amount of the compensation exceeds the trade harms derived from the imposition of the ETM. That is, if the transaction cost is less than $9 million, the invocation of the ETM would be Pareto superior to the situation prior to the invocation of the ETM, given that voluntary bargaining between Country A and B would be held.154

One objection to the concept of Pareto efficiency is that the conditions for Pareto superiority are very difficult to be satisfied in the real world due to imperfect informational problems and the fact that most of the transactions affect third parties.155 The Kaldor-Hicks efficiency is more "practical" in the sense that it ignores whether a gainer from an action actually compensates a loser from the action.156 When benefits derived from an action exceed the costs, doing the action is "efficient" in the sense of PPE. That is, PPE focuses on the ability of gainers to compensate losers regardless of whether they actually do so; that is why PPE is also called potential Pareto efficiency(PPE). In other words, PPE is a kind of Pareto efficiency under the assumption that transaction costs are zero.157 With regard to the above-mentioned hypothetical example, the invocation of the ETM is considered to be Pareto optimal since Country B's damages ($1 million) are outweighed by Country B's benefits ($10 million) derived from the ETM.

Throughout the first three Sections of this Chapter, PPE will be utilized as the standard of efficiency since it makes analysis much simpler by allowing us to ignore the transaction costs for bargaining.

B. Costs and Benefits of ETMs

There are three factors which would be affected by the invocation of ETMs: harms to international trade (HT), benefits derived from the amelioration of environmental problems (EB), and administrative costs (AC). When EB outweigh the total of HT and AC, the ETMs actually improve world overall welfare in the sense of PPE (EB>HT+AC). The optimal ETMs are those which improve net benefits (EB-HT-AC) the most among available ETMs.

HT, EB, and AC are allocated among states as following:

- With regard to HT, when a country invokes ETMs, the benefits from international trade usually decrease.158 According to a classical international trade theory, trade measures always decrease measure-imposing countries' trade benefits. Nevertheless, as standard public choice theory suggests, countries sometimes impose trade restrictions so as to protect their industries. In this case, it can be said that the measure-imposing country has a distributional preference which is preferable for the industries. On the other hand, the trade benefits of exporting countries can either improve or deteriorate, depending on the demand elasticity of targeted products in importing countries and the qualities of the trade restriction. For example, if the measures in question are quantitative quotas and the demands of the imported products concerned are inelastic in the measure-invoking countries, the trade benefits in the exporting countries would increase.

- EB are enjoyed only by measure-imposing countries if the measures aim at protecting the domestic environment (this is the definition of the "domestic" environment). On the other hand, if the objects of measures are the global environment, all states enjoy EB derived from the measures since the global environment is one of the public goods and the measure-invoking countries cannot exclude "free riders".

- All of AC of trade measures are usually borne by the measure-imposing countries.

We assume that ETM-invoking countries attempt to maximize their welfare. Therefore, they try to employ measures which improve their welfare the most. However, the measures which maximize ETM-invoking countries' welfare are not necessarily the measures which maximize overall world welfare because measure-imposing countries do not have to bear all the costs derived from the invocation of the measure. There is divergence between ETM-invoking countries' and world interests.

An ETM-invoking country does not have incentives to take into account the harms suffered in foreign countries in the following two ways:

First, that there is a domestic-market model in which the production of a product creates negative externalities and there are no administrative costs in implementing environmental regulations. If the producers do not have to bear the environmental harms as their private costs, the amount of the production exceeds the optimal. As a result, the society is better off by reducing the amount of the production. In this case, the marginal benefits derived from the decrease in the pollution exceed the marginal losses derived from the reduction in the use and production of the product. The government does not have an incentive to reduce the amount of the production beyond the point where the marginal losses derived from the reduction in the consumption and production exceed the marginal benefits of the reduction in the pollution. On the other hand, what happens in the context of international trade? Assume a case where State A tries to ameliorate an environmental problem which is caused by the importation of a product by restraining its importation. We assume three things: there is no "disguised purpose" for protecting its domestic industry in State A's ETM, State A has perfect information, and there are no administrative costs in implementing the ETM. Given that State A tries to maximize its private welfare, the ETM would restrict the importation of the product excessively from the viewpoint of world welfare maximization in the sense of PPE. That is, State A tries to reduce the amount of the production up to the point where the marginal benefits derived from the reduction of the pollution equal the marginal losses derived from the reduction of the consumption, which loss is suffered by State A's nationals. State A does not take into account the loss suffered by the exporters/producers in foreign countries. However, the optimal amount of the production from the viewpoint of potential Pareto efficiency would be the one in which the marginal benefits derived from the reduction of the pollution equal the total of the marginal losses derived from the reduction of the production, which loss is suffered by exporting countries, and the decrease in the consumption, which loss is suffered by the nationals of State A. In other words, countries tend to set "appropriate" (to maximize the ETM-invoking countries' private benefits) levels of health or environmental protection "inappropriately" from the viewpoint of world welfare maximization in the sense of PPE.

Second, even if we think of appropriate levels of health or environmental protection as given, countries tend to prefer ETMs which impose a heavy burden on international trade.159 In the following, suppose that there is a case in which Country D is considering whether it should invoke an ETM in order to protect the domestic environment. The "appropriate level of protection" which is determined by the government of Country D yield EB worth $40 million.160 There are three available measures which can achieve the appropriate level of protection as the Table 1 shows below.161 We assume that Country D has perfect information about costs of each option.

Measure

AC

HT

EB

World Welfare Improvement
(WW=EB-HT-AC)

A

$26.3m

$2.4m

$40m

$11.3m = ($40m-$2.4m-$26.3m)

B

$11.7m

$11.7m

$40m

$15.6m = ($40m-$11.7m-$11.7m)

C

$2.4m

$26.3m

$40m

$11.3m = ($40m-$26.3m-$2.4m)

Table 1: Three Available ETMs

As Table 1 shows, Measure B is the most welfare-enhancing from the viewpoint of PPE. The total of AC and HT are the least in Measure B, whereas EB are the same in all the measures.

Then, we examine the divergence between Country D's private welfare (PW) and world overall welfare (WW). Suppose that a third of the HT is allocated to Country D (PHT). Table 2 shows the divergence between D's (PW) and the world's welfare (WW). All of the AC are to be borne by Country D. Country D obtains all of the EB.

Measure

AC

PHT

EB

PW

WW

A

$26.3m

$0.8m = ( 2.4/3)

$40m

$12.9m

$11.3m

B

$11.7m

$3.9m = (11.7/3)

$40m

$24.4m

$15.6m

C

$2.4m

$8.8m = (26.3/3)

$40m

$28.8m

$11.3m

Table 2: Divergence between PW and WW

As Table 2 shows, Country D actually prefers Measure C because PW is maximized in Measure C even though the world welfare is maximized when it invokes Measure B. This result is not surprising because Country D has to bear all of the AC, while the HT are shared by other countries. Country D prefers measures in which the HT are large but the AC are small. This is how the behavior of ETM-imposing countries tends to derogate from optimal even when we think of appropriate levels of health or environmental protection as given.

In these ways, ETM-invoking countries try to externalize the costs of ETMs to outsiders. However, as Coase suggests, this incentive itself does not necessarily mean that the invocation of the ETMs will end up with sub-optimal results.162

In the above hypothetical example, the exporting countries have to bear additional trade harms of $9.7 million ($26.3m x 2 ÷3 - $11.7m x 2 ÷ 3), which is greater than additional Country D's benefits of $4.4 million ($28.8m - $24.4m) by taking Measure C instead of Measure B. Therefore, the exporting countries can be better off by bribing Country D any amount between $4.4million and $9.7million, suggesting Country D to invoke Measure B instead of Measure C.
However, this Coase-type bargaining between exporting and ETM-invoking countries does not necessarily occur because of the following reasons.

First, trade harms on exporters can be split among many countries. As a result, none of the exporting countries would have an incentive to bargain with ETM-invoking countries. For instance, in the above hypothetical example, suppose that the HT on foreign countries are divided over ten countries, each of which suffers the same amount of harm. That is, the additional amount of harms these countries suffer as a result of the invocation of Measure C instead of Measure B is $0.97 million which is much less than additional Country D's benefits of $4.4 million. Therefore, individual exporting countries do not have incentives to seek to bargain with Country D. To be sure, bargaining is still possible if the exporters can cooperate for the purpose of bargaining. However, if the number of involved countries is large, then their ability to come together for bargaining may be small since the difficulties of cooperating tend to rise with the number of countries involved.

Second, bargaining would not occur due to the existence of asymmetric informational problems. Assume in the above hypothetical example that all of the HT on foreign countries concentrate on one country, say Country E, and Country E can make only a single take-it-or-leave-it offer to Country D. Country E is risk-neutral. Country E is not exactly sure how much Country D gains additionally by invoking Measure C instead of Measure B. Suppose that Country E think that there are two possibilities: the additional gain by taking Measure C might be $2 million with a probability of 80%, or it might be $6 million with a probability of 20%. In other words, Country E expect that, if it offers $2 million to Country D, it would be accepted with a probability of 80% and be rejected with a probability of 20%. If its offer is rejected, Country E still has to suffer the loss of $9.7 million. Therefore, Country E's expected loss would be $3.54 million ($2 million x 0.8 + $9.7 million x 0.2) if it offers $2 million to Country D. On the other hand, if it offers $6 million, even though bargaining will occur for sure, Country E's expected loss is also $6 million, which is greater than its expected loss when it offers $2 million. Therefore, its best offer is $2 million. However, the actual additional gain by invoking Measure C for Country D is $4.4 million, which is greater than Country E's offer. Consequently, its offer will be rejected by Country D. As this example shows, bargaining would not occur since the parties to the bargaining would not have perfect information about the costs and benefits of available policy options.

Third, even if exporting countries have perfect information regarding the costs and benefits of available ETM options, they would not try to bargain with Country D; rather they would choose to bring an action before the WTO Dispute Settlement Body (DSB) because the initial allocation of "property rights" is unclear. If one of the rules under the WTO clearly admits a right of Country D to invoke Measure C, exporting countries would try to bargain with Country D without bringing the case to the DSB. However, in most of the cases, it is unclear whether the GATT, the SPS Agreement, or the TBT Agreement allows Country D to impose Measure C in order to protect its domestic environment. In fact, it might be also possible that the AB&P interprets these agreements as they guarantee exporting countries' right of not being disturbed by ETMs like Measure C. As a result, the exporting countries may bring the case to the WTO prior to attempting to bargain with Country D in order to clarify whether they have a right to be protected from ETMs like Measure C. If they win, they have to pay nothing to force Country D to discard Measure C. Even if they lose, then they can try to bargain with Country D. In sum, it is rational for exporting countries to bring a case to the WTO in order to clarify the allocation of "property rights" prior to bargaining with ETM-invoking countries insofar as adjudication costs for the exporting countries are small.163

Because of these reasons, exporting countries and ETM-invoking countries might not bargain, or not be able to bargain successfully. Therefore, we need "trade-off devices" by which we alter the behavior of ETM-imposing countries appropriately.

C. Analysis of the LTRA and "Means and Ends" Tests

PPE implies that comparative cost-benefit analysis (CCBA) is the best way to maximize the net benefits obtained by ETMs. That is, in conducting CCBA, we compare the costs and benefits of available policy options and select the one which improves the net benefit (EB-AC-TH) the most. If the AB&P have perfect information regarding the costs and benefits of available ETM options, they can accomplish the optimal use of the world's resource in the sense of PPE by striking down ETMs which are not the most efficient among available policy options. However, the AB&P have never invoked CCBA in analyzing ETMs. Instead, they have utilized the LTRA test and the "means and ends" test.

In the following, we compare the LTRA test and the "means and ends" test with CCBA.

AB&P has defined the LTRA as the test which examines whether ETM-invoking countries choose the least trade-restrictive measure among reasonably available ones. "Reasonably available" has two meanings: one is whether a less trade-restrictive measure is administratively available at reasonable cost; the other is whether a less trade-restrictive measure can effectively accomplish the purpose of the measure. The analysis of the "reasonably available" standard has to be done using inherently value-laden judgments; the strictness of the LTRA test may vary depending on how the "reasonably available" standard is applied. It should be noted that the LTRA test does not pay attention to the magnitude of EB. Regardless of the magnitude of EB, ETMs would be upheld as long as the ETM is the least trade-restrictive among reasonably available options. Furthermore, the LTRA test neither "compares" nor "sums up" any value at all. That is, in the process of LTRA scrutiny, AC, EB, and HT are examined independently without comparing them to each other or aggregating them.

As Trachtman points out, the LTRA test is "overbroad and underinclusive" in light of world welfare maximization in the sense of PPE.164 The optimal measure under CCBA could be struck down under the LTRA test if there exists a less trade-restrictive alternative than the optimal option.165 On the other hand, it is also possible that world welfare would deteriorate as a result of the invocation of the least trade restrictive measure if the EB obtained by the measure is extremely small and the HT is relatively huge. Furthermore, it is also pointed out that the least trade-restrictive test is imperfect in the sense hat it excludes policy options by thinking of "appropriate levels of protection" as given.166 That is, there would be measures which can improve the net benefits more than any options which can accomplish the "appropriate levels of protection".

The "means and ends" test focuses on the relative relationship between HT and EB. AC is excluded from the scrutiny. When HT that occurs as a result of the invocation of an ETM is disproportionately large in comparison with the magnitude of EB obtained by the ETM, the ETM is struck down. Like the standard of the "reasonably availability" in the LTRA test, scrutinizing the "proportionality" between EB and HT inherently requires value-laden judgments. We have to wait for future cases to know how the "proportionality" standard is applied.167 In the process of the "proportionality" analysis, an examiner has to "compare" HT and EB. This is a big difference between the LTRA and "means and ends" tests.168

It should be noted that there is a big difference between the LTRA test and CCBA on the one hand and the "means and ends " test on the other hand. In the process of the LTRA test and CCBA, multiple measures are examined. Meanwhile, in the scrutiny of the "means and ends" test, the examiner does not pay attention to the availability of other policy options; ETMs are upheld as long as HT is not disproportionately large in comparison with the magnitude of EB even if there is another measure which improves world welfare more than the measure in question. Therefore, the "means and ends" test is imperfect from the viewpoint of world welfare maximization. Furthermore, it can EB said that the "means and ends" test would be inferior to static cost-benefit analysis, which examines whether benefits from the invocation of a policy option exceeds the costs of the policy implementation in the sense of PPE, since administrative costs are excluded from the consideration of the "means and ends" test.

As we have seen, neither the LTRA test nor the "means and ends" test can guarantee the realization of the maximization of world welfare in the sense of PPE. However, this result does not necessarily mean that the use of the LTRA test and the "means and ends" test by the AB&P is inadequate. As Dunoff and Trachtman point out, there are some limitations in the use of the law-and-economics methodology.169 The LTRA test and the "means and ends" test could be a good solution to these limitations in a particular situation.

D. Evaluation of AB&P's Use of the LTRA and "Means and Ends" Tests

In the following Section, we examine whether the use of the LTRA test and the "means and ends" test could be supported in light of the distributional, ethical, and philosophical problems PPE has.

1. Critiques of PPE

There could be three main criticisms of the use of PPE concept in the context of international law: lack of attention to distributional problems among countries, the incommensurability of values, and problems with interstate comparison of the utility.

First, PPE assumes that there are no transaction costs in redistributing benefits regardless of whether winners actually compensate losers. PPE focuses on whether net benefits increase as a result of the invocation of a measure without regard to redistribtional results of the measure. For example, suppose that there is a measure which benefits the developed countries more than it harms the least developed countries. Assume that the redistribution of the benefits and costs are impossible because of high transaction costs. Under usual Pareto efficiency analysis, this measure cannot be considered to be efficient because the developed countries cannot be better off without making the undeveloped countries worse off. However, this measure is efficient under PPE because the benefits enjoyed by the developed countries exceed the costs suffered by the developing countries.170 PPE itself cannot philosophically, morally, or politically justify why harming others by conducting an action should be allowed when the benefits of the action are greater than the harms.

Second, in conducting CCBA or looking for a measure which accomplishes potential Pareto optimal, we have to "compare" and "amalgamate" various benefits and costs. However, as many have pointed out, it is a very difficult job to quantify all of the social values such as the environment, moral order, and security on a unified scale (for example, a monetized scale) in order to calculate the net benefits gained by a policy option.171 For example, in the context of the protection of endangered species, people value these species primarily for their existence per se, as opposed to values which could be enjoyed by the exploitation or use of the species. That is, people can obtain economic utilities from the very fact that the endangered species are preserved.172 Furthermore, some people also obtain "altruistic" utilities from the fact that endangered species are protected from cruelty or death. However, how can we quantify these kinds of values?173 As Sunstein argues, in real life, we evaluate these values in non-economic ways. As a result, describing these values in economic language would obscure the real nature of non-economic valuations which we are actually using.174 Furthermore, Dunoff maintains that the use of economic language in evaluating these "non-use values" means not only describing them inaccurately but also creating the risk of changing our understanding of these social values.175 In the context of ETMs, the problem of incommensurability of values is the most significant when we evaluate the benefits derived from environmental protection. While some of the environmental regulations are primarily based on use value, for example the protection of human health, the others refer to non-use values, for example, the protection of endangered species. In conducting CCBA, we have to quantify the benefits derived from ETMs so as to render these benefits and costs commensurable with those of trade harms and administrative costs. That is not an easy task.

Third, PPE entails the problem of interpersonal comparison of utilities. In the process of CCBA, an analyst compares benefits obtained by one person with costs suffered by another. However, there is no widely accepted method to evaluate the benefits/costs to another. In order to compare utilities between people, it is required not only that benefits or costs have to be quantified in a comparable scale but also that these quantifications have to accurately reflect the utilities of individuals. In the context of ETMs, this problem appears in the form of interstate comparison of utility. We are not sure how we can compare the benefits of an ETM-invoking country with the harms inflicted by exporting countries.

In the following, we analyze whether the use of the LTRA and "means and ends" tests are good methods for avoiding these problems that PPE has.

2. The LTRA Test

It should be noted that the AB&P have utilized the LTRA test on ETMs which aim at protecting the domestic environment, regardless of which agreement the disputes were brought under. Under the SPS Agreement, Salmon and Apple, the cases in which the LTRA test was applied, were all about the protection of their nationals' health from risks associated with food consumption. Under the GATT, in Asbestos, the Panel applied the LTRA test to a French decree designed to protect their nationals from health risks arising from the use of asbestos. In this regard, the AB findings in Gasoline, in which the purpose of the U.S. measure was to ameliorate air pollution caused by gasoline combustion in major population areas in the U.S., seems to be somehow odd in light of this general trend that the LTRA test is applied to ETMs for protecting the domestic environment since the "primary aimed at" test was applied in Gasoline. However, we should bear in mind that the chapeau analysis in Gasoline was very close to the LTRA test. The AB's approach in Gasoline presents a clear contrast to the AB's findings in Shrimp/Turtle, where the object of the U.S. measure was to protect the global environment and the AB employed a balancing approach in the chapeau scrutiny. Furthermore, we should be mindful of the general tendency that the chapeau tests seem to be more LTRA-like when measures in dispute are in pursuit of the realization of domestic interests. In Bovine Hide, the Panel found that the application of the Argentinean measure in issue, which was aimed at preventing tax evasion of income and value-added taxes, amounted to "unjustified discrimination" because an "alternative course of action" in which the Argentinean government would reimburse importers for the additional interest forgone would be available and administratively feasible.176 The prevention of tax evasion can be regarded as a measure designed to protect "domestic" interests even though it is irrelevant with environmental protection. In sum, regardless of whether under the independent paragraphs of GATT Article XX, under the chapeau of GATT XX, or under Article 5.6 of the SPS Agreement, the AB&P have invoked the LTRA test when the case is about measures designed to protect the domestic interests.

In order to evaluate whether the use of the LTRA test in the context of the domestic environmental protection is appropriate, we need to analyze the above-mentioned distributional, ethical, and philosophical weaknesses of PPE, which raise great concerns in the context of ETMs preserving the domestic environment.

First, as a general proposition, the distributional problem is significant when countries invoke ETMs in order to protect the domestic environment. The benefits of the measure concentrate on the measure-imposing countries by sacrificing the trade interests of foreign exporters. To be sure, the consumers in measure-imposing countries also suffer harms derived from the restriction on international trade. However, countries do not have an incentive to invoke ETMs unless benefits from environmental protection outweigh the total of harms on their consumers plus administrative costs, assuming the ETMs are "genuine"(that is, there is no protectionistic motivation in the ETMs). In other words, ETMs for protecting the domestic environment could be regarded as a kind of "beggar-thy-neighbor" policy; measure-imposing countries are better off by making other countries worse off, even though the net benefits to the measure-imposing countries might exceed the trade harms borne by other countries.177

However, we should also take into account national sovereignty rights to deal with domestic environmental problems. It is an ETM-imposing country, not other countries, which should be able to decide how the ETM-imposing country deals with its domestic environmental problems. The WTO seems to concur with this proposition. The AB&P have repeatedly underscored the point that WTO members are free to set their appropriate levels of health or environmental protection as they wish.178

In this regard, it could be said that the LTRA test would strike an exquisite balance between the sovereign right of a measure-imposing country to protect its domestic environment and a right of other countries to enjoy the benefits of free trade; the LTRA test does not intrude on ETM-imposing countries' discretion in determining its appropriate levels of environmental protection, whereas requiring them to minimize trade harms on other countries. The LTRA accomplishes the balancing of Members' rights by not balancing values. It is not coincidental that the AB&P emphasized the sovereign right of Members to decide their appropriate level of protection when they invoked the LTRA test under GATT Article XX (b) or Article 5.6 of the SPS Agreement. To be sure, distributional problems remain unsolved to some extent insofar as outsiders have to suffer trade harms, even though they are minimized. However, as Trachtman points out, compensation for redistribution could be indirect; the application of the LTRA test for all the ETMs for preserving the domestic environment implies that all Members would receive roughly equivalent payoffs. All Members are guaranteed the right to decide their appropriate levels of protection whereas they simultaneously have the right not to suffer trade harms which are more than minimized.179

Second, with regard to the problem of the incommensurability of values, it should be noted that the LTRA test does not "compare" any values at all: it evaluates the magnitude of trade harms independently from any other values; it does not pay attention to the magnitude of benefits derived from environmental protection at all180; furthermore, it considers administrative costs only to the extent of determining whether alternatives are economically feasible. The LTRA test has a strong moral advantage by neither comparing values nor quantifying the value of environmental protection. However, examiners of the LTRA test should bear in mind that there is a sloppy pitfall in the analysis of administrative feasibility. It is possible to argue that administrative feasibility should be determined in light of the magnitude of benefits derived from environmental protection.181 However, if we take into account the benefits of environmental protection in evaluating administrative feasibility, the LTRA test loses its valuable advantages: unnecessity of comparing environmental value with other values and of quantifying the magnitude of environmental benefits. In other words, the LTRA test becomes a mere subset of cost-benefit analysis.182 In this regard, we should be careful of how we interpret the AB findings in Beef with regard to the relationship between the LTRA test and the purpose of trade measures. In Beef, in the analysis of GATT Article XX (d), the AB noted that "[t]he more vital or important...interests or values" are pursued, the easier trade measures would be deemed to be "necessary" to achieve their ends.183 This AB's comment is subject to many interpretations. It might mean that the regulatory purpose should be taken into account in analyzing the administrative feasibility of alternatives. On the other hand, another possible interpretation is that the AB merely pointed out the general tendency that the more important the regulatory purpose is, the fewer available alternatives there are. Analysts of the LTRA test should avoid the former interpretation since this interpretation would erode the moral weight of the LTRA test. In this regard, it should be criticized that the AB examined the degree of importance of the French measure in question in Asbestos, while the analysis of the French regulatory purpose did not seem to play a significant role in examining the availability of alternatives.184

Third, as to the problem of interstate comparison of utilities, we should be mindful of the AB's approach when evaluating the magnitude of trade restrictions as a gestalt without distinguishing between trade harms suffered by foreign exporters/producers and those borne by ETM-imposing countries' consumers.185 By so doing, the AB&P avoid "comparing" utilities among countries.

In conclusion, the LTRA test is a good method for avoiding the distributional, philosophical, and moral problems that CCBA has. In particular, regarding distributional problems which are significant in the context of ETMs for protecting the domestic environment, the LTRA test significantly ameliorates the distributional disadvantages of the domestic ETMs, while guaranteeing the sovereign rights of ETM-imposing countries to decide their appropriate levels of health or environmental protection.

3. The "Means and Ends" Test

Shrimp/Turtle was the only case since the establishment of the WTO in which an ETM which aimed at protecting the global environment became the issue of the case.186 Unlike other environment-related cases brought to the WTO, the AB invoked the "means and ends" test in its analysis. In this case, the AB invoked the 93means and ends" test twice, in the analysis of Article XX (g) and in the chapeau analysis. First, the AB scrutinized whether there is "close and real" relationship between the purpose of sea turtle protection and the ends employed under Section 609.187 The AB found that the scope of Section 609 was not "disproportionately wide" in comparison with the purpose of sea turtle protection. Second, in the chapeau analysis, the AB found that requiring other countries to adopt essentially the same regulatory program as the U.S. without taking into account different conditions in other countries was not an adequate means to achieve the policy goal of sea turtle protection.188 Furthermore, the AB found that the fact that the U.S. did not allow the import of shrimps from non-certified countries even if the particular shrimps were caught using methods identical those employed in the U.S. could not be reconciled with the purpose of sea turtle protection.189

Unlike the LTRA test, the nature of the "means and ends" test is the balancing of values. The "means and ends" test juxtaposes trade harms and environmental benefits and compares them. In the context of ETMs for preserving the domestic environment, the AB&P have avoided "comparing" values. However, they did balance them in analyzing an ETM for protecting the global environment. The adoption of the "means and ends" test could be explained in light of how distributional, philosophical, and moral problems appear in the context of ETMs aimed at protecting the global environment.

First, when countries invoke ETMs in order to protect the global environment, distributional concerns are not as significant as when they impose ETMs in order to protect their domestic environment since countries other than the ETM-invoking countries also obtain benefits from the measure. As a result of global-environment-oriented measures, outsiders are not necessarily worse off. Hence, ETMs aimed at protecting the global environment could not be considered as a kind of "beggar-thy-neighbor" policy as domestic-environment-oriented measures are. In other words, the necessity of minimizing trade harms on outsiders is not so strong. In this regard, it should be noted that the "means and ends" test excludes administrative costs from its consideration. Administrative costs are exclusively borne by measure-imposing countries. On the other hand, trade harms and environmental benefits are shared by ETM-invoking countries and other countries. That is, the "means and ends" test compares the only part of the benefits and costs which are shared by the international community.

Second, with regard to the global commons, no countries can enjoy the sovereign right to decide their appropriate level of protection. Some denounce the invocation of ETMs for protecting the global environment because the imposition of ETMs means that measure-invoking countries force other countries to adopt their preferences concerning appropriate levels of protection for the global environment.190 If we assume this argument is true, the other side of the coin must be also true: prohibiting ETMs means that exporting countries force other countries to accept their preferences for low levels of protection for the global environment. As Lowenfeld observes:

[i]n a situation involving more than one country, if country A wishes to regulate and country B does not, it seems...equally fair to criticize B for attempting to impose its will beyond its borders as it is to criticize A for attempting to exercise its jurisdiction extraterritorially. Similarly, it seems equally fair to restrain states that cause global environmental harm as it is to restrain states that seek to use extra-territorial trade measures ensure that their own markets do not contribute to global environmental harm.191

As a result, in the context of global environment protection, the LTRA test does not have the philosophical, moral, or ethical weight it has in the context of domestic environment protection. As Dunoff maintains, applying the LTRA test to measures aimed at protecting the global environment may imply that trade values are elevated to a pre-eminent status in comparison with environmental values:192 trade harms have to be minimized, however, environmental harms do not have to be minimized. Therefore, balancing tests such as the "means and ends" test are more appropriate in the context of global environmental protection.

Third, by juxtaposing trade harms and environmental benefits, the "means and ends" test raises great concerns about the incommensurability of values and the interstate comparison of utilities. However, in comparison with cost-benefit analysis, these problems are less significant because the "means and ends" test is a balancing test with a moderate level of precision. By roughly juxtaposing values, the "means and ends" test creates room for the consideration of non-monetizable and non-commensurable values and avoids trying to produce a precise comparison.193

Because of these reasons, we can conclude that the AB's use of the "means and ends" test for ETMs aimed at protecting the global environment is appropriate.

4. Conclusion

As we have seen, the LTRA test is appropriate to use for ETMs designed to protect the domestic environment, and the "means and ends" test is suitable for analyzing global-environment-oriented trade measures. In this regard, the AB&P have appropriately differentiated the use of the LTRA test from that of the "means and ends" test. However, we need a qualification about these. This paper has examined the circumstances under which the invocation of the LTRA or "means and ends" test is appropriate as a general analytical framework for making a "trade-off" between environmental and trade interests. What this paper has not done in detail is to analyze how to apply these tests in actual cases. In particular, we have not discussed how to determine the standard of "reasonable availability" in the LTRA test and the "disproportionately wide" standard in the "means and ends" test except to discuss the point that the magnitude of the environmental benefits achieved by ETMs should not be taken into account when deciding the reasonable availability of alternatives in applying the LTRA test. The actual posture of these tests vis-à-vis ETMs would vary depending on how the AB&P set the thresholds of "reasonableness" and "proportionality" in real disputes194 even though the AB&P have not clarified these points yet.195 These points would be left to future studies and cases to be elaborated.

However, some comments should be made with respect to the scopes of the SPS Agreement, the TBT Agreement, and GATT Article XX. The TBT Agreement and the SPS Agreement should not be applied to the measures aimed at protecting the global environment since both of them adopt the LTRA test as their "trade-off device". According to Paragraph 1 of Annex A of the SPS Agreement, sanitary or phytosanitary measures are those designed to protect human or animal life or health within the territory of the Members. Therefore, it is reasonable to say that the scope of the SPS Agreement is limited to measures aimed at protecting the domestic environment. However, in the TBT Agreement, there are no provisions by which the scope of the TBT Agreement could be limited to ETMs aimed at protecting the domestic environment. Article 2.2 of the TBT Agreement provides that the legitimate objectives of technical regulations include "protection of human health or safety, animal or plant life or health, or the environment". This provision implies that the protection of the global environment could be one of the legitimate objectives of technical regulations. This broad scope of the TBT Agreement could be problematic in future cases.

On the other hand, GATT Article XX would be flexible enough. In the context of ETMs aimed at protecting the global environment, Article XX (g) should be utilized. Article XX (g) would be able to include all the ETMs aimed at the global environment under its very broad purview. Even though some of the ETMs for the domestic environment protection also fall under the scope of Article XX (g), the LTRA test could be incorporated into the chapeau test. This flexibility was exemplified by the AB findings in Gasoline.


152 See Trachtman, Cost-benefit Analysis, supra note 41.

153 Robert S. Pindyck and Daniel L. Rubinfeld, Microeconomics 588 (4th ed, 1997).

154 See Ronald Coase, The Problem of Social Cost, 3 Journal of Law and Economics 1 (1960).

155 See Richard A. Posner, Economic Analysis of Law 14 (5th ed 1998).

156 See J.R. Hicks, The Foundations of Welfare Analysis, 49 Econ. J. 696 (1939); Nicholas Kaldor, Welfare Propositions in Economics, 49 Econ. J. 549 (1939). See also Lucian Arye Bebchuk, The Pursuit of a Bigger Pie: Can Everyone Expect a Bigger Slice, 8 Hofstra L. Rev. 671, 671 n.2 (1980). Some works regarding "trade and environment" implicitly use PPE as the standard of efficiency. See e.g. Chang 95, supra note 21, at 2189 (stating that "the critical issue is whether the global economic benefits of the measure outweigh its global economic costs"); Daniel A. Farbar & Robert E. Hudec, Free Trade and the Regulatory State: A GATT's-Eye View of the Dormant Commerce Clause, 47 Vand. L. Rev. 1401, 1405 (1994). (noting that "in a community consisting of several smaller untis of government (a United States consisting of individual states, or a GATT consisting of individual nations), the ultimate question is whether the gain of the regulation for insiders outweighs the harm it causes to outsiders").

157 See Trachtman, Cost-benefit Analysis, supra note 41, at 53.

158 In this Chapter, we assume the trade-off relationship between trade and environment values by limiting the analysis of efficiency to static efficiency as many prior studies did. See id. at 33 (stating that "we are forced to choose the extent to which each value is to be implemented, to make tradeoffs among these values.") At the level of dynamic efficiency, it would be possible that the invocation of ETMs would result in the promotion of free trade since the threat by ETMs would give incentives to improving the cost performance of exporters. However, even if there exist ETMs which enhance both environment protection and free trade at the same time at the level of static efficiency, this fact itself does not negate the efficacy of the discussion in this Section since ETMs which ETM-imposing countries actually invoke tend to derogate from the optimal ones insofar as there is a divergence between measure-invoking countries' and world interests.

159 Here, we assume that trade-off relationship between administration resources and trade values, on the one hand, and environmental values, on the other hand. That is, environmental benefits can be obtained through using administrative resources and/or giving up trade benefits.

160 As PPE suggests, we assume here that all of benefits and costs are measurable and commensurable.

161 We assume here that the ETM-invoking countries produce "the reduction of environmental harm", using "trade harm" and "administration cost" as "production factors". If we think like this, thinking about EB as an "output" and about AC and HT as "inputs", we can draw the isoquant curve, the curve which shows the possible combinations of the inputs (AC and HT) that yield the same output (EB=40). In this example, the equation of the isoquant curve is AC2+HT2=402. Measures A, B, and C are points on the isoquant curve. B is the most efficient combination of "inputs" (AC and HT) to produce 40 "outputs" (EB). By doing such an analysis, we are assuming that "the law of diminishing returns" holds for AC and HT and the isoquant curve is concave up and the combination of AC and HT in which the sum of AC and HT is minimized is located in somewhere middle of the isoquant curve (AC=11.7 HT=11.7 in this example).

162 See Coase, supra note 154.

163 But cf. Joel P. Trachtman, The Domain of WTO Dispute Resolution, 40 Harv. Int'l L.J. 333, 354-5 (1999) (pointing out that, when the entitlement of a property is ex ante unclear, bargaining is likely to occur because person A who actually does not possess the entitlement may credibly threaten to take it and this threat provides the person B with an incentive to give A the information which clarifies the real entitlement of the property).

164 See Trachtman, Cost-benefit Analysis, supra note 41, at 72.

165 See Frank J. Garcia, The Global Market and Human Rights: Trading Away the Human Rights Principle, 25 Brook. J. Int'l L. 51 (1999) (pointing out that "the test evaluates measures favorably precisely insofar as their impact on trade is the least possible, despite the fact that more trade-impacting measures might be more effective in realizing the non-trade value. Not only does this trade-off mechanism fail to recognize the high priority which rights must hold in any policy determination, but in fact the necessity test turns this on its head, and privileges trade values over all other competing values").

166 See Trachtman, Cost-benefit Analysis, supra note 41, at 72

167 See McLaughlin supra note 13, at 882 (stating that "it is impossible to predict exactly what effect... the [means and ends test] may have on future trade/environment disputes. While it took a less doctrinal approach to the article XX exception,...it provided little else in the way of guidance to future panels. Much of the decision's language can be characterized as either overly broad or exceedingly narrow. For example, criteria such as substantial relationship that is every bit as substantial and `not disproportionately wide' are extremely broad and subject to discretionary interpretation").

168 It is normally said that the "means and ends" test is more moderate than the LTRA test. See e.g. Cho, supra note 80. Theoretically speaking, this is, however, not necessarily accurate because these two tests utilize very different factors; the factors which are utilized in LTRA test are the absolute values of HT, AC, and BE, whereas the analysis of "means and ends" test is based on the relative relationship between HT and BE. To be sure, "second" or "third" least trade restrictive measures could be sustained under the analysis of the "means and ends" test. However, it is also likely that a least trade-restrictive measure is struck down under the "means and ends" test.

169 See Dunoff & Trachtman, supra note 22, at 44-45.

170 See Trachtman, Cost-Benefit Analysis, supra note 41, at 53.

171 See e.g. Jane B. Baron & Jeffrey L. Dunoff, Against Market Rationality: Moral Critiques of Economic Analysis in Legal Theory, 17 Cardozo L. Rev. 431 (1996); Steven Kelman, Cost-Benefit Analysis--An Ethical Critique, Regulation, January 1981, 33; Mark Sagoff, Economic Theory and Environmental Law, 79 Mich. L. Rev. 1393 (1981); Cass R. Sunstein, Incommensurability and Valuation in Law, 92 Mich. L. Rev. 779 (1994).

172 See Richard B. Stewart, International Trade and Environment: Lessons from the Federal Experience, 49 Wash. & Lee L. Rev. 1329, 1340-41 (1992).

173 Some argue that these "non-use" values should not be taken into account in cost-benefit analysis. See e.g. Paul Milgrom, Is Sympathy an Economic Value? Philosophy, Economics, and the Contingent Valuation Method, in Contingent Valuation: A Critical Assessment 417 (Jerry. A. Hausman ed., 1993). However, economics which try to maximize the social welfare function can not provide a basis for excluding these "non-use" value from consideration insofar as people obtain utilities from these non-use values.

174 See Sunstein, supra note 171, at 854.

175 See Jeffery L. Dunoff, Rethinking International Trade, U. Pa. J. Int'l Econ, L. 347, 366 (1998).

176 See Bovine Hide, supra note 42, at para. 11.316-11.328.

177 See Chang 1995, supra note 21, at 2190 (stating that "Environmental trade measures [for protecting the domestic environment] may be suspect not because they have an asymmetric impact on domestic and foreign producers, but because our environmental standards yield asymmetric benefits for our nationals when applied at home and abroad").

178 See e.g. Asbestos Panel Report, supra note 28, at para. 8.179; Salmon AB Report, supra note 107, at para.199 (noting that "'the level of protection deemed appropriate by the Member establishing a sanitary...measure", is a prerogative of the Member concerned and not of a panel or of the Appellate Body"[emphasis in original]).

179 See Trachtman, Cost-Benefit Analysis, supra note 41, at 73.

180 See Farber & Hudec, supra note 156, at 1432 (stating that "in GATT,... tribunal encounter much stronger objections to decisions about the regulatory benefits of the measures in quesiton. GATT tribunals have few credentials to assess the success or social value of reglatory measures and lack any recognized political mandate to do so"[emphasis in original]).

181 See Virginia Dailey, Comment, Sustainable Development: Reevaluating the Trade vs. Turtles Conflict at the WTO, 9 J. Transnat'l L. & Pol'y 331, 376 (2000); Garcia, supra note 165, at 51.

182 See id. at 70 (pointing out that "if the reasonableness test amounts to a requirement that the least trade restrictive alternative not be so costly as to countervail the benefits of the regulatory measure, then it bears some resemblance to [cost-benefit analysis]").

183 See Beef AB Report, supra note 100, at para. 163,166.

184 See Asbestos AB Report, note 28, at para. 172-5.

185 In this regard, the magnitude of trade harms is evaluated based on the degree of the change of the competitive relationship between domestic and imported products, as opposed to the change of trade volume. See WTO Report of Appellate Body: Japan -Taxes on Alcohol Beverages, WT/DS8/AB/R. WT/DS10/AB/R, WT/DS11/AB/R (adopted Nov. 1, 1996) para. 109-110, available at <http://www.wto.org> (last visited Mar. 15, 2001) (noting that "Article III obliges Members of the WTO to provide equality of competitive conditions for imported products in relation to domestic products.... Article III protects expectations not of any particular trade volume but rather of the equal competitive relationship between imported and domestic products...").

186 It would be possible to say that the AB did not recognize strongly that sea turtle is one of the global environmental resources since the AB pointed out that "the sea turtles here at stake...are all known to occur in waters over which the United States exercises jurisdiction". See Shrimp/Turtle AB Report, supra note 15, at para. 133. However, it should be noted that the AB emphasized that the protection of sea turtles requires concerted and cooperative efforts among countries, citing several international documents such as Principle 12 of the Rio Declaration on Environment and Development, all of which underscore the importance of international cooperation for protecting the global environment. See id. at para. 168.

187 See id. at 135-142.

188 See id. at 161-4.

189 See id. at 165.

190 See e.g. Jagdish Bhagwati, Trade and the Environment: The False Conflict?, in Trade and the Environment: Law, Economics, and Policy 159,170 (Durwood Zaelke et al. eds., 1993) (arguing that the usage of trade measures implies that measure-imposing countries "force others into accepting [their]own idiosyncratic choice of ethical concerns").

191 Andreas F. Lowenfeld, Lawrence F. Ebb, International Business: Regulation and Protection (1964) (St. Paul: West Publishing Co.), 78 Harv. L. Rev. 1699, 1703- 04 (1965) (book review).

192 See Jeffrey L. Dunoff, Reconciling International Trade with Preservation of the Global Commons: Can We Prosper and Protect?, 49 Wash. & Lee L. Rev. 1407, 1449-50 (1992).

193 See Trachtman, Cost-Benefit Analysis, supra note 41, at 80.

194 In this regard, see supra note 167; Garcia, supra note 165, at 51.

195 As to the LTRA test, in Asbestos, the AB found that not only administrative difficulties but also the possibility of realizing the regulatory purpose pursued should be taken into account in determining the reasonable availability of an alternative. However, the AB did not clarify how difficult implementing the alternative has to be or how sure one must be that the alternative can achieve the regulatory purpose so as to consider the alternative to be reasonably available. See Asbestos AB Report, supra note 28, at para. 170-5. As for the "means and ends" test, in Shrimp/Turtle, the AB avoided elucidating the threshold of the proportionality standard by stating that "[the] relationship [between Section 609 and sea turtle protection]...is every bit as substantial as that which we found in United States - Gasoline between the EPA baseline establishment rules and the conservation of clear air in the United States94. See Shrimp/Turtle AB Report, supra note 15, at para. 141

Top|Previous|Title|Next

 

 


This site is part of the Academy of European Law online, a joint partnership of the Jean Monnet Center at NYU School of Law and the Academy of European Law at the European University Institute.
Questions or comments about this site?
Email Enfellows@exchange.law.nyu.edu