Research on factors other than the investigated imported goods causing damage to the domestic industry in trade defense cases

According to the provisions of the relevant WTO Agreements on Trade Remedies (TAR), the application of TAR measures is only carried out when there is dumping/subsidy/sudden increase in imports, damage to the domestic industry, and a causal relationship between the investigated imported goods and the damage. Therefore, the analysis of damage and causal relationship are the main issues of a TAR investigation. Proving the causal relationship between the investigated imported goods and the damage to the domestic industry is based on the examination of all relevant evidence before the investigating authority. The investigating authority must also investigate other known factors that also cause damage to the domestic industry and the damage caused by those factors will not be counted in the impact caused by the investigated imported goods. Among other factors, factors (known) that may be taken into account in this case include: the quantity and price of non-dumping imports, decline in demand or changes in consumption patterns, trade-restrictive actions or competition between domestic and foreign producers, developments in technology, export capacity and productivity of the domestic industry. The list of factors other than imports is provided for guidance only.

1. Consider GATT/WTO precedents: “factors other than imports”

There is no legal requirement in the provisions of the existing SPS agreements that imports alone are the cause of injury. The provisions of the agreements recognize the possibility that factors other than imports cause injury to the domestic industry, but they do not provide a method or procedure for proving a causal link between the injury and the imports. However, Article 3.5 of the Anti-Dumping Agreement (ADA) and Article 15.5 of the Countervailing Duty Agreement (SCMA) require the investigating authority to consider all “known factors” other than imports that are causing injury to the domestic industry at the same time as the imports, and that the injury caused by these factors cannot be attributed to the imports. This is known as the “other factors” analysis. The ADA and SCMA have identified the following factors as determining the injury:

- The volume and price of non-dumping and non-subsidized imports

- Reduced domestic demand or changes in consumer tastes

- Trade restrictions and competition between domestic and foreign producers

- Technological developments

- Exports and productivity of the domestic industry.

The separation and analysis of injury from other factors is a common practice in many GATT/WTO disputes. Some of these fall under the ADA, while the majority of these fall under the SGA. The Appellate Body in the US – Wheat Gluten case, one of the first safeguard cases, observed:

It is reasonable to assume that the determination of injury caused by increased imports must follow a certain procedure for separating the injury caused by other factors. If the effects of other factors are not separated and distinguished from the effects of increased imports, it will be impossible to assess the injury caused by a key factor. As we have said, the final determination of the “causal link” between increased imports and serious injury can only be made after the impact of increased imports has been assessed in its entirety, and this assessment must be separated from the impact of other causes.

In US-Lamb, the Appellate Body, referring to Article 4.2(b) of the Safeguards Agreement (SGA), held that a finding based on a single factor, “increased imports” in the SGA, may be on a shaky basis, since the investigating authority assumed that other factors did not cause injury, and attributed the entire impact to imports. According to the Appellate Body, the “other factors” consideration provided for in Article 4.2(b) requires a reasonable and adequate explanation of the “nature and extent” of the effects of injury caused by “other known factors,” as distinguished from injury caused by increased imports.

The US – Hot-rolled Steel case, a dispute involving the application of antidumping measures, is perhaps the most important dispute on the “other factors” analysis. Japan, the petitioner, argued that the USITC had not adequately assessed the effects of “other factors” on the US industry, such as the increase in capacity of small mills and the subsequent expansion of US steel suppliers, the General Motors strike, the reduction in demand for hot-rolled steel from the tube and coil industry, and the selling price of non-dumping goods. Furthermore, Japan argued that the USITC had “failed to ensure that the injury caused by these other factors was not attributable to imports.” This was one of the earliest “other factors” cases brought at the WTO. During the GATT era, only a handful of disputes considered the role of “other factors,” and there was no requirement for such an analysis. For example, in the US – Atlantic Salmon case, a GATT Panel analyzed the role and scope of the “other factors” provision of the Tokyo Round AD and CVD Agreements. Norway, the complainant, argued that the USITC should have conducted a full investigation, not just made a general determination of all possible causes of significant injury to the domestic industry, and that the USITC should have isolated and excluded all effects of other factors that could cause injury from the effects of imports. The Panel noted that, other than assessing the impact of imports as provided for in Articles 3.1, 3.2 and 3.3 of the [Tokyo AD Code], the USITC is not required to determine the extent to which injury is caused by other factors. According to the Panel, there is no legal obligation for the investigating authority to analyze separately the injury caused by other factors and the injury caused by imports from Norway.

In US – Hot-rolled Steel, Japan also argued that the explicit requirement that “[the investigating authority] must assess any known factors other than dumped imports” under the ADA is inconsistent with recognizing the existence of the possibility that other factors are causing injury to the domestic industry, and constitutes “a material change in the Agreement that renders the determinations in US – Atlantic Salmon invalid.” In considering this issue, the WTO Panel noted that it is the responsibility of the investigating authority to “consider and ensure that these other factors do not undermine the causal relationship that appears to exist between the dumped imports and the material injury to the domestic industry based on the consideration of the quantity and impact of the imports as provided for in Articles 3.2 and 3.4 of the ADA”. In general, the GATT Panel’s finding was that it was not necessary to quantify the injury. On appeal, the Appellate Body promptly rejected this view of the Panel and held that the requirement to clearly apportion and separate the impact of the other factors is essential to concluding whether the imports have actually caused material injury to the domestic industry.

The effects of other factors referred to in Article 3.5… apply only in cases where dumped imports and other factors are causing injury to the domestic industry at the same time. In order for the Investigating Authority to … ensure that the effects of the other factors mentioned are “attributable” to the imports, the Investigating Authority must make a full assessment of the effects of injury caused by these factors. Obviously, this analysis must include a disaggregation and separation of the effects of injury caused by other factors other than the imports.

According to the Appellate Body, if the effects of the dumped imports and other factors are mixed and cannot be distinguished, there is no way to be certain whether the effects of injury attributed to the dumped imports are in fact caused by the other factors. This issue was reiterated by the Appellate Body in the US – Line Pipe case as follows:

The other factors provision referred to in Article 4.2(b) serves two purposes. First, it is intended to serve the case where the injury is caused by more than one factor at the same time, to prevent the investigating authority from concluding that a “causal link” exists between the increase in imports and the risk of injury being caused by factors other than the increase in imports. Second, it is a tool to ensure that only an appropriate amount of the total injury is attributed to the increase in imports.In principle, the Appellate Body’s ruling on the “other factors” provision provides a useful approach as it aims to establish the injury caused by imports. However, the practical challenge lies in the implementation of this provision. It may be useful to consider how this approach might be applied in a real-world trade remedy dispute. Hot-rolled steel rebar was one of the products investigated in the US-Steel dispute. The defendants in this case (exporters facing safeguard tariffs) argued that the injury to the US domestic industry was caused by increases in input costs or cost of goods sold (COGS). The USITC acknowledged the existence of increased COGS but emphasized that import competition prevented the domestic industry from passing on these increased costs to end customers. In other words, import competition may be the reason why domestic producers are unable to recover their costs, but how can an investigating agency prove that the injury caused by increased COGS is not attributable to imports? Another example is that a change in the exchange rate may be the cause of injury to the domestic industry, but at the same time this factor may also be the cause of dumping or increased imports. In such a case, there is an interaction of many economic factors, which cannot be separated separately for the purpose of assessing “other factors”.

In the above discussion, the assessment of the impact of “other factors” also raises certain questions. Do all investigating authorities have to quantify and exclude the role of other known factors? Or, does this simply mean that the investigating authority should provide a reasonable and adequate explanation of the injury caused by other known factors? The two approaches to this issue are separation and separation. There is a view that the investigating authority cannot distinguish the nature and extent of the injury effects except by using arithmetic or economic calculations. On the other hand, if the requirement is to provide explanations without specific economic evidence, the analysis will be a purely qualitative one – a procedural requirement. It seems that in some recent cases the Appellate Body has found it unreasonable to impose a standard that is not clearly understood and that not many WTO members have the capacity or resources to meet.

More interestingly, in the famous US – Line Pipe safeguard dispute, the Appellate Body, in interpreting the relevant provision of the SGA, interpreted the “other factors” analysis as a purely qualitative one. The Appellate Body stated:

[To] satisfy the requirement of Article 4.2(b), last sentence, the investigating authority must clearly demonstrate, through sufficient and reasonable explanations, that the injury caused by factors other than imports is not attributable to increased imports. This explanation must be clear and unambiguous. It must not be metaphorical, ambiguous, or suggestive. It must be stated directly in clear terms.

Similarly, in the US – DRAM v. Mexico – Olive Oil cases, the WTO Panel also maintained its position on the analysis of “other factors” without specifically examining the methods used to separate and separate the effects of “other factors”. In Mexico – Olive Oil, the Panel held that Economia, Mexico’s CVD investigating authority, had “carefully considered, disaggregated and separated the effects of each of the other factors from the effects of the subsidized goods, and reasonably concluded that these factors did not cause injury to Fortuny [a Mexican olive oil producer]. In other words, the Panel considered whether Economia had analyzed a known factor and supported its conclusion. During the Panel’s dispute resolution, the Panel asked Mexico to identify the paragraphs in its Preliminary Investigation Report and Final Investigation Report that addressed injury factors. Based on the information submitted by Mexico, the Panel upheld the Mexican investigation authority’s conclusion without considering whether the Mexican investigation authority’s analysis was properly quantified. The Panel’s conclusions support the view that once the investigation authority has identified and If the Panel and the Appellate Body list the causal factors, and provide a plausible explanation for their “other factors” analysis, they will not be willing to change their conclusions. The Panel in the US-DRAM case also made similar explanations.

It could be argued that the Appellate Body’s conclusions in the US-Line Pipe case have clarified the concept of injury due to “other factors” alone, and can be seen as a significant improvement over the current state of the art in this regard. Furthermore, the qualitative analysis of injury due to “other factors” advocated by the Panel in the Mexico-Olive Oil case is a testament to the practical difficulties in carrying out this type of analysis. In addition, an economic analysis by Chad Bown and Niall Meagher based on the information provided by the parties in this case found that the Mexican investigating authority failed to separate the effects of other factors on injury. This is a common practice in a number of trade remedy cases where the causal relationship for injury is “attributed” to imports without regard to the effects of other factors. In this context, an objective explanation – even a defective or imperfect explanation – can reduce the contestability of an investigation to an acceptable level. As the Appellate Body noted in US-Lamb, if the investigating authority takes full account of the nature and complexity of the data and acts on a reasonable interpretation of the data, the Panel is unlikely to change its findings. This view was reaffirmed in the EC – DRAM case when the Panel held that a “satisfactory explanation” that included “qualitative analysis” could still satisfy the requirements of clearly apportioning and separating the harm caused by “other factors”.

2. Analysis of the Impact of “Other Factors” and Their Shortcomings: The Application of the “Otherwise” Analysis

The preceding paragraphs have pointed out the shortcomings of temporal overlap or correlation analyses in establishing a causal relationship between imports and injury to the domestic industry. However, a series of WTO Panels and Appellate Body have emphasized the need to demonstrate “a genuine and substantial cause and effect relationship”. The key question here is how this requirement can be written in legal language, and applied in practice in accordance with Panel and Appellate Body rulings. The “otherwise” approach is a commonly used tool for establishing/determining causality in domestic investigations. Even in the context of trade remedies, the “otherwise” approach is a useful tool. For example, in the context of an antidumping/safeguard investigation, one might ask: “If there had been no dumped/subsidized imports, would injury have occurred?” This is a counterfactual question and requires an explanation of what would have happened to the domestic industry if there had been no increase in imports. This analysis does not ask directly about the cause of the injury to the domestic industry. In fact, the approach is based on the balance of probabilities: if it is highly probable that an event is the cause, it is treated as such. In the context of a trade remedy investigation, the “if-not” approach is useful in considering whether the increase in imports, or the dumping/subsidization of imports, is the most likely or most probable cause of injury. For the purposes of explanation, it is assumed that in such a case, the increase in imports, by volume, price, or other means, is the most likely cause of injury, and the purpose of the “if-if” test is to determine whether the assumption is true or false.

In the context of antidumping and safeguard investigations, the “if-if” test is often implemented through econometric models such as partial equilibrium models. Causality is typically examined by setting the level of imports as a constant or by allocating quotas while using actual industry data for all other variables, and comparing this output to actual output. These comparisons are intended to approximate the state of the domestic industry in the absence of imports. In US law, these tests are implemented through what is known as unitary analysis. This method compares the actual situation of the domestic industry with a model that simulates the situation of the industry in the face of import competition. COMPAS (the current model), which is commonly used in antidumping and subsidy investigations, uses three input indicators and six parameters (measures of elasticity) to determine the price and quantity effects of dumping. The COMPAS model is designed to calculate the impact of dumping at specified margins on domestic prices, domestic sales and total revenue. The main parameters used include: (i) the margin of dumping; (ii) the market share of the domestic industry; (iii) the market share of imports; (iv) the total elasticity of demand for the product under investigation; (v) the elasticity of supply (production) of the domestic industry; (vi) the elasticity of supply from fairly competitive imports; (vii) the elasticity of substitutability between imports and domestic goods; (viii) the elasticity of substitutability between fairly competitive imports and domestic goods; and (ix) the elasticity of substitutability between the subject imports and fairly competitive imports. Factors (iv) to (ix) are all based on assumptions and, therefore, are only approximate. The WTO rules on trade remedies do not prohibit the use of such advanced statistical/econometric models, and even WTO panels use them regularly, at least in arbitration cases.

The “if not” approach has important applications in the area of ​​causality analysis. However, a major flaw with this approach is that it does not look for the cause of injury when multiple causes exist. Consider the example of an industry that is suffering injury. Suppose that the industry is operating under an outdated model, lacking skilled workers, high cost of debt, lacking an efficient investment model, and lacking investor confidence. However, the company still has a loyal customer base that is willing to buy at a competitive price. However, when imports begin to flood in, customers switch to buying imported goods. In a trade remedy investigation, the obvious question is: if – without the imports – would the industry have suffered injury? The domestic industry may seek antidumping protection, but the real cause of the industry’s poor performance is something else. Thus, the “if-else” approach does not, by its very nature, require an actual cause to occur. As Fumerton and Kress argue, it works well in simple examples of causality analysis, but is not useful in complex investigations involving multiple causes of injury and long investigation periods.

The “if-else” approach is not the same as the correlation approach, although they may appear similar. The underlying assumption of both approaches is that increased imports/dumping/subsidies are the primary cause of injury, but the correlation approach assumes that if two factors occur together, one can cause the other. The “if-else” approach does not make use of this assumption. It is more of an “all or nothing” approach, meaning: either the factor is the cause or it is not the cause.

This issue can be explained simply through a hypothetical example. Suppose we have to quantify the impact of several factors on the situation of a manufacturing industry. We can assume that the role of each factor is quantified through economic indicators in this case. Imports from the country under investigation are quantified as contributing 30% to the injury of the industry, and there are two additional causes of injury. Suppose that these causes each contribute 35% to the injury. The question is whether a trade remedy investigation can use the “if-not” method in this case?

In the above assumption, the other two factors, which together account for 70% of the injury, are the principal causes of the injury. It could be argued that even without the presence of imports, the domestic industry would still have suffered injury. In this case, imports would not be considered the cause of injury in the usual sense. In other words, the “if-else” approach fails to establish a causal link between imports and injury. One of the major disadvantages of this approach is that it can only account for a limited number of inputs. And, one might ask: Is 30% of the injury significant or insignificant? Prima facie, no. But could this factor have caused the injury on its own without the other factors? The answer depends on the case, but imports usually contribute to the injury to the domestic industry. How should an investigating authority deal with this information? It can be concluded that the “if-else” approach is useful only where there are only a limited number of known causes, but that it should be avoided where there are many. But, as noted above, where imports are the obvious cause of injury, it may be used.

The use of economic models is a core issue that is often debated in the context of determining injury and causality. A common view is that quantification should be as extensive as possible. As the WTO Panel in the US-Steel case put it, “[even] the simplest quantitative analysis can provide useful information … in determining the nature and extent of injury caused by factors other than imports to the domestic industry.” Regression methods are often used to assess the relationship between various economic variables. One particular regression method used in similar causality estimates is the Granger causality model. However, the Granger causality model requires a lot of input data. Econometric models such as COMPAS, used by the USITC, can also be useful in measuring the magnitude of the impact of imports on the domestic market. And, these models seem to be less well suited to distinguishing the relative impact of different factors that may have caused injury to the domestic industry at the same time. And, the conclusions of these econometric models are valid only when their assumptions are close to actual conditions. Moreover, the range of products investigated in AD and CVD cases is often very narrow, making it difficult to collect reliable econometric data. It is therefore not surprising that the COMPAS model analysis is often incorporated into the notice accompanying the USITC decision, but is not the deciding factor in most cases.

A challenge for most investigating authorities may be to consider the impact of “other factors” when there are multiple causes of injury to the domestic industry. The preceding paragraphs have mentioned “if-else” methods and econometric models such as COMPAS which are useful when dumped/increased imports are a clear cause of injury. However, these models have certain limitations when there are multiple causes of injury. Given these limitations, economists have used simultaneous equation models, such as the “Two-Stage least squares – 2SLS” model (a regression model) to estimate the supply and demand for the product under investigation. For example, Professor Thomas Prusa’s 2SLS model was used by the USITC in the 2000 cold-rolled steel investigation. However, these models require a large amount of reliable time series data and are therefore recommended for use only in certain cases that can satisfy these stringent input requirements.

In general, all of the approaches to determining causality and the impact of “other factors” that are currently used, both in the United States, the EU and elsewhere, have significant limitations. Most of the cases reviewed in this chapter rely on templates or checklists that only meet the minimum requirements. In fact, we can see the gaps in the approach to determining causality and injury used by most investigating authorities. It is therefore not surprising that Alan Sykes has stated that there is probably no country that can conduct a safeguard investigation without being challenged or without the conclusion of causality being upheld. Despite this fact, not all conclusions of injury and causality can be successfully challenged at the WTO, if the investigating authority has provided a reasonable and adequate explanation of the various causal factors. Under the usual standard in trade remedy investigations, especially anti-dumping, the Panel and the Appellate Body must uphold the findings of the investigating authority if they are based on the evidence in the record. The investigating authority, while using the traditional trend analysis method, can also use additional economic factor analysis models to distinguish and examine the role of other factors in as much detail as possible.

3. Conclusion

According to the WTO Appellate Body, the investigating authority must clearly demonstrate, through adequate and reasonable explanations, that the injury was caused by factors other than imports that are not attributable to imports – a requirement known as an “other factors” analysis.

While an “other factors” analysis has certain applications, the separation and separation of the effects of different factors causing injury is often more burdensome than anticipated. There is a widespread view that this analysis can only be done through quantitative analysis. However, this view remains controversial and the generally acceptable method for doing this analysis has not yet been established. WTO case law on this aspect is still evolving. Furthermore, the reliability of economic models will depend on the accuracy of the data and assumptions made. This paper has examined how certain computational models, such as 2SLS, have been applied in some cases by some investigating authorities, such as the USITC. However, the larger issue is that the use of complex econometric models has certain political consequences as well as inherent limitations. Even assuming that the econometric model is adequate and useful, the results of a regression may be difficult to understand for non-statisticians or policymakers or courts. Therefore, these requirements should not be applied to all WTO trade remedy investigations. Given the uncertainty about how to separate and distinguish the roles of other factors of injury, this paper argues that the strict “other factors” provision adopted by the Appellate Body is unnecessary and unreasonable.

The study also concludes that there has been a clear shift towards qualitative methods for the analysis of “other factors”, although the Appellate Body has not yet taken any affirmative position on these methods. Based on a review of some recent decisions such as US – DRAM, Mexico – Oil, Egypt – Rebar, Mexico – Pipe and Fittings, the study finds that the shift in the approach to injury and causation in WTO case law is away from sophisticated scientific or quantitative methods, and towards simpler, more widely accepted methods that respect the purpose of these trade remedies. In other words, the analysis of “other factors” is now widely considered to be an obligation to identify the various causes of injury and to demonstrate (the Investigating Authority’s) consideration of such causes through a comprehensive and explicit analysis. As long as these analyses are adequate and reasonable, panels are unlikely to challenge the Investigating Authority’s findings of injury and causality. In other words, if the findings of injury and causality meet the “minimum standard of reasonableness,” they should be considered acceptable.

The “other factors” analysis can only be satisfied by seeking all the causes of injury and fully assessing them, even if in a descriptive manner. A review of WTO disputes has shown that the rigorous economic or scientific “disaggregation” and “separation” of the effects of other factors advocated by the Appellate Body is not essential or feasible for all WTO members. With the exception of the United States, none of the respondent countries in WTO disputes have used econometric models to determine the causal relationship to injury to the domestic industry. In this context, a WTO member is free to choose the approach that suits its country.

This study also examines the applicability of the “if-not” method as an alternative measure for analyzing injury and causality. The “if-not” method is widely used by common law systems, and it can be argued that it is a good choice for analyzing injury in some specific cases. However, this method also has weaknesses if the case has multiple causes. Furthermore, even the “if-not” method may still require qualitative analysis to be properly applied in practice. The “if-not” method has been included in Unitary analysis by some investigating authorities with experience in trade defense cases, such as the United States. However, the prevailing view on this issue is that the “if-else” approach is only more useful when there are only a few causes of injury, or more precisely, when imports are the obvious cause of injury. Furthermore, the “if-else” approach may not be very appropriate, especially when imports are only one cause of injury, and not the most important cause.