1.1 The measure at issue

This dispute concerns a set of regulatory measures imposing a 25% ‘charge’ on imported automobile parts used in the manufacture of motor vehicles in China. The charge is due if the imported auto parts have the character of a ‘complete vehicle’, something that is determined by the Chinese authorities based on criteria prescribed under three instruments enacted by the Chinese government.Footnote ¹

The criteria for such a determination are expressed in terms of particular combinations or configurations of imported auto parts or the value of imported parts used in the production of a particular vehicle model. The use in the production of a vehicle model of specified combinations of ‘major parts’ or ‘assemblies’Footnote ² that are imported requires characterization of all parts imported for use in that vehicle model as ‘complete vehicles’.

Various combinations of assemblies will meet the criteria, for example: a vehicle body (including cabin) assembly and an engine assembly, or five or more assemblies other than the vehicle body (including cabin) and engine assemblies. The use, in a specific vehicle model, of imported parts with a total price that accounts for at least 60% of the total price of the complete vehicle also requires characterization of all imported parts for use in that vehicle model as complete vehicles. Imports of Completely Knocked Down (CKD) and Semi-Knocked Down (SKD) kitsFootnote ³ are also characterized as imports of complete vehicles.

In other words, the measures at issue set up a regulatory regime and impose a charge with respect to imported auto parts used in the production or assembly of certain models of motor vehicles that are sold in the Chinese domestic market. They prescribe thresholds regarding the type or value of imported auto parts used to assemble specific vehicle models. If the imported parts used in a given vehicle model meet or exceed these thresholds, then all of the imported parts used to assemble that model are characterized as complete vehicles and are subject to the charge under the measures at issue.

The charge is imposed following assembly of the vehicles. The measures set out a number of procedural and administrative steps designed to determine whether the charge applies and ensure tracking and reporting of the imported auto parts, along with payment of the charge, in respect of the relevant auto parts. The amount of the charge imposed corresponds to the tariff rate applicable to complete vehicles, that is 25%. Imported auto parts that are not ‘characterized as complete vehicles’ under the measures at issue are subject to duties at the tariff rates for auto parts in China's Schedule of Concessions, that is 10% on average.

The system is based on the principle of ‘self-evaluation’ to be performed by the automobile manufacturer prior to beginning production of a new vehicle model that will incorporate imported parts and that will be sold in China. In case of a positive self-evaluation, the manufacture will apply to the Ministry of Commerce for an import license for the parts to be used in the vehicle model. The automobile manufacturer must provide a duty bond,Footnote ⁴ the amount of which is calculated on the basis of the projected quantity and value of the auto parts that will be imported each month. In practice, the amount corresponds to the applicable tariff rate for auto parts (10% on average) applied to the projected monthly importations of auto parts. The automobile manufacturer may then begin to import parts to be used in the production of its new vehicle model. When the auto parts characterized as complete vehicles for use in a registered vehicle model are then actually imported into China, they are subject to the duty bond as well as to ongoing tracking and reporting requirements imposed on the manufacturer importing those parts. The parts are not, however, subject to any ongoing physical confinement or any other restriction by customs authorities on their use in the internal market.

Article 13 of China's Decree 125 states that auto manufacturers importing auto parts characterized as complete vehicles must declare their importation of such auto parts and pay duties to the district customs office. Article 13 thus refers to the obligation not only to declare the importation of auto parts, but also to pay duties. However, actual payment of the duty/charge under the measures does not occur until the Verification Centre has completed the verification that the imported parts actually possessed the essential character of ‘complete vehicles’. This occurs after the manufacturers have finished the assembly of auto parts into complete vehicles.Footnote ⁵ It is this ‘duty’ that is to be paid by the manufacturers following verification that constitutes the ‘charge’ under the measure which is being challenged by the complainants.

1.2 The complainants' claims and China's defense

The United States, Canada, and the European Communities (the ‘complainants’) claimed that the charge imposed under the measures was an ‘internal charge’ that was inconsistent with Article III:2 of the GATT 1994 because it applied to imported auto parts, while no similar charge applied to like domestic parts. In addition, they argued that the additional administrative requirements on automobile manufacturers that use imported auto parts in excess of specified thresholds violated China's obligations under Article III:4 of the GATT 1994. According to the complainants, these requirements provided an incentive to prefer domestic auto parts. By using domestic parts, the manufacturers would be able to avoid all of the reporting and verification requirements.

In the alternative, if the measures were considered to impose customs duties, as argued by China, the complainants claimed that such a ‘duty’ (25%) was in excess of the relevant tariff bindings in China's Schedule of Concessions for automobile parts (10%) and was therefore inconsistent with Article II:1(a) and (b) of the GATT 1994.

The complainants also argued that the charge imposed by the measures equally applied to imports of certain unassembled or partially assembled motor vehicles, that is completely CKD and SKD kits.Footnote ⁶ The complainants claimed that China's tariff treatment of such kits under the measures was inconsistent with Article II:1(a) and (b) of the GATT 1994, as it exceeded the 10% import duty rate set for imported parts. Two of the complainants, the United States and Canada, claimed that the measures were also inconsistent with paragraph 93 of China's Accession Working Party Report, in which China promised to set at 10% the level of the duties due in respect of the importation of such kits, in case it would create a separate tariff line for kits.Footnote ⁷

China responded that the charge under the measures was an ordinary customs duty, within the meaning of Article II:1(b), and not an internal measure subject to Article III. China argued that the measures were not inconsistent with Article II of the GATT 1994 because they give effect to a proper interpretation of ‘motor vehicles’ in China's Schedule of Concessions. China explained that the measures ensured ‘substance over form’ in the administration of China's national customs law in that they allow customs authorities to classify, as complete motor vehicles, groups of auto parts and components that have the essential character of a complete vehicle, irrespective of how an importer chooses to structure the importation of these parts. China contended that the measures thereby prevent the ‘circumvention’ of China's tariff headings for motor vehicles. China argued that the treatment of the CKD or SKD kits was in line with its tariff binding for complete vehicles and thus consistent with GATT Article II. In China's view, it was consistent with China's obligation under its Schedule and Article II:1(b) of the GATT 1994 to treat CKD and SKD kit imports as complete vehicles in light of the principle that a complete set of parts ‘presented unassembled or disassembled’ (e.g., in a kit) is classified as the complete article, not as parts of that article.

China argued that, insofar as the measures were found to be inconsistent with either Article II or Article III, they were justified under Article XX(d) of the GATT 1994. China expressed the view that the measures were necessary to address circumvention of the obligation to pay 25% duties on complete vehicles through the shipment of auto parts at 10% when those parts were later assembled into a complete vehicle.

4.1 Border measure or internal charge – a fine line

The main legal issue to be resolved in this case was the question whether the charge imposed by China on imported auto parts was to be characterized as a customs duty (i.e., a border charge) on the importation of auto parts, or rather as an internal charge on imported auto parts (i.e., an internal measure). Under GATT, different rules apply to these different types of charges.

Ordinary customs duties, as well as all other duties and charges imposed on the importation of goods, are covered by the tariff commitments set forth in a Member's schedule of concessions. GATT Article II provides that no ordinary customs duties are to be levied ‘on the importation’ of products above the bound tariff rate. In addition, it provides that products shall be exempt from all other duties or charges of any kind ‘imposed on or in connection with importation’. GATT Article II is the basic market-access provision that determines the maximum price of the entry ticket to be paid in order to obtain market access in a given Member country. GATT Article XI is the corollary of Article II, as it prohibits any ‘restrictions or prohibitions other than duties, taxes or charges … on importation’. GATT Article XI thus confirms that duties and charges as circumscribed by GATT Article II are the only type of border measure allowed for under GATT.

In contrast, the only obligation that Members have in respect of ‘internal taxes or other internal charges of any kind’ is to ensure that imported products that have paid their entry ticket into the market are treated no differently from like domestic products. That is basically what GATT Article III stands for. It imposes a requirement that internal taxes, and more generally any measures affecting the internal sale, purchase, transportation, distribution or use of products, ‘should not be applied to imported or domestic products so as to afford protection to domestic production’. This is translated into two practical obligations: (1) in respect of internal taxes or charges, not to impose internal taxes or other internal charges of any kind in excess of those applied on like domestic products; and (2) in respect of any other internal measures, to ensure that imported products are treated no less favourably than like domestic products. GATT Article III thus acts as an ‘anti-circumvention’ provision, as it intends to prevent a Member from retreating from its market-access commitments by imposing internal charges or introducing internal measures more generally in a manner that distorts competitive opportunities to the detriment of imported products. The Panel in the case discussed in our report correctly summarized the role played by both provisions in the system in the following manner:

However, the distinction between import duties/border charges, on the one hand, and internal charges, on the other hand, is not always an easy one to make. Both GATT Article II on import duties/charges and GATT Article III on internal taxes and charges contain references to the possibility that additional charges may lawfully be imposed or collected at the border, even though their imposition would exceed the bound rate.

GATT Article II.2 provides that nothing in Article II on border charges shall prevent a Member from imposing at any time ‘on the importation’ of any product ‘a charge equivalent to an internal tax imposed consistently with’ the nondiscrimination obligation of GATT Article III.2 on like domestic products. Similarly, the Ad Note to GATT Article III states that any internal charge that applies to an imported product and to the like domestic product and is ‘collected or enforced in the case of the imported product at the time of importation, is nevertheless to be regarded as an internal tax or other internal charge’ and is thus subject to the provisions of GATT Article III. Both provisions thus warn against simplistic conclusions about the nature of a measure from the mere fact that certain payments are imposed at the border or at the time of importation.

In the case discussed in this report, the Appellate Body had to determine whether the charge imposed by China on imported auto parts was an internal charge covered by GATT Article III or an ordinary customs duty/border charge covered by GATT Article II. The approach followed by the Panel and the Appellate Body in answering this question is in line with the approach adopted in prior GATT/WTO case law relating to the different nature of border measures and internal measures.

In older GATT cases like Belgian Family Allowances Footnote ⁵² or EEC–Parts and Components,Footnote ⁵³ Panels have concluded that if the event that gives rise to the liability to make the payment is an internal event that takes places following importation, such as their assembly, and if the charge is thus imposed on ‘imported products’ – i.e., products that have been imported already – the charge is of an internal nature, even when this charge is collected at the border. Similarly, the WTO Panel on Argentina–Hides and Leather, found that the prepayment at the border of a Value-Added Tax (VAT) was an internal measure because the measure applied to ‘definitive import transactions, but only if the products imported were subsequently re-sold in the internal Argentinean market’.Footnote ⁵⁴ The Panel referred to Ad Note to GATT Article III to conclude that ‘the fact that [the payment] is collected at the time and point of importation does not preclude it from qualifying as an internal tax measure’.Footnote ⁵⁵

In contrast, if the charge is ‘on’ or ‘connected with’ importation – i.e., in respect of a product that is being imported – and the liability to pay the charge arises as a consequence of importation, it will be considered as a border charge covered by GATT Article II.Footnote ⁵⁶ As noted by the GATT Panel in EEC – Measures on Animal Feed Proteins, recalling the GATT negotiating history, ‘The Sub-Committee at the Havana Conference considered (Havana Reports pp. 62–63, paragraphs 42–43, E/CONF.2/C.3/A/W.30 page 2) that “certain charges … were import duties and not internal taxes because … (a) they are collected at the time of, and as a condition to, the entry of the goods into the importing country, and (b) they apply exclusively to imported products without being related in any way to similar charges collected internally on like domestic products”’.Footnote ⁵⁷

The same approach to border measures as measures that condition access to the market and impose a condition on entry into the market is found in the WTO case law in respect of GATT Article XI prohibiting restrictions ‘on importation’. The Panel on India–Autos considered the ordinary meaning of the phrase ‘restriction … on importation’ as being a restriction ‘with regard to’ or ‘in connection with’ the importation of the product.Footnote ⁵⁸

The Chinese case discussed in this report shows some resemblance to the GATT case on EEC–Parts and Components, relating to so-called ‘screwdriver’ operations used to circumvent antidumping duties on finished products. In that GATT case, the (then) EEC expressed the view that the antidumping duties it had imposed on finished products were being circumvented by the sale of parts and components of such products to importers related to or associated with the exporters found to have been dumping their products into the EC market. The EEC considered that ‘in order to prevent circumvention, it is necessary to provide for the collection of an antidumping duty on products thus assembled or produced’ in the event one could link the assembled products to the products covered by the antidumping duties.Footnote ⁵⁹ The measure imposed by the EEC to fight what it considered to be circumvention of the antidumping duties consisted of imposing the same antidumping duties on the finished products assembled in the EEC on the basis of the parts and components exported by exporters that were subject to antidumping duties on the finished products.

The GATT Panel noted that the anti-circumvention duties are levied, according to the EEC regulation, ‘on products that are introduced into the commerce of the Community after having been assembled or produced in the Community’. The duties are thus imposed, as the EEC explained before the Panel, not on imported parts or materials, but on the finished products assembled or produced in the EEC. Their imposition is not conditional upon the importation of a product or at the time or point of importation.Footnote ⁶⁰ The GATT Panel found that the internal assembly of the parts into the finished product subject to the antidumping duty triggered the duty liability and thus that the measures imposed by the EEC were internal measures subject to GATT Article III.Footnote ⁶¹

In so doing, the GATT Panel rejected arguments by the EEC relating to the policy objective of the charges (to combat circumvention of antidumping duties) or their qualification under domestic law of the EEC as antidumping duties – i.e., border measures. Such arguments were not considered to be relevant to the determination of the nature of the charge, which was either a duty imposed on or in connection with importation within the meaning of Article II or an internal tax within the meaning of Article III.2.Footnote ⁶²

On the nature of the charge imposed by China, the Panel and Appellate Body reach a conclusion similar to that of the GATT Panel on EEC–Parts and Components. According to the Appellate Body, for a charge to constitute an ordinary customs duty in the sense of GATT Article II, ‘the obligation to pay it must accrue at the moment and by virtue of or, in the words of Article II:1(b), “on”, importation’.Footnote ⁶³ In contrast, the Appellate Body is of the view, like the Panel, that the adjectives ‘internal’ and ‘imported’ suggest that the charges falling within the scope of GATT Article III are charges that are imposed on goods that have already been ‘imported’, and that the obligation to pay them is triggered by an ‘internal’ factor, ‘something that takes place within the customs territory’Footnote ⁶⁴:

In general, this approach relating to the distinction between border measures and internal charges makes sense – border measures determine the amount that needs to be paid in order to be allowed into a market, while internal charges are triggered by the internal sale, offering for sale, purchase, use, or any other internal event that occurs subsequent to importation. In this case, it was the use of the imported auto parts in a certain combination with other auto parts of foreign origin that led to the imposition of an additional levy to be paid by the manufacturer of motor vehicles. Duty liability under the Chinese measure rests with the manufacturer of the vehicle, not the importer of the parts. If the car parts were ultimately not used to manufacture a car that China considered to be an imported vehicle based on the percentage of imported auto parts, the charge was not due, even if the manufacturer had originally informed the authorities to the contrary. It thus seems very clear that it is the use of the parts in a particular combination, once imported, that determines whether the charge is due. In our view, it is correct to conclude, as did the Panel and the Appellate Body, that this charge is not imposed ‘on the importation’ of the parts but is an internal charge. Arguably, China's measure could be qualified as a ‘reverse border adjustment’ – China does not impose an additional charge in order to impose the same internal charge levied on domestic products equally on imported products (the normal type of border adjustment measure provided for in Ad Note GATT Article III and GATT Article II.2); rather, China imposes the additional charge in order to adjust for a customs duty on cars that was not paid by the importer of the car parts considered as a complete vehicle. In other words, it seems that the internal charge is seen as a means of adjusting for the failure to comply with a border charge (i.e., the customs duty on cars). No GATT provision expressly provides for this type of adjustment and thus, arguably, such an adjustment is simply a GATT-inconsistent internal charge.

Nevertheless, it seems justified in a case like this one to wonder whether all this discussion about the right ‘qualification’ of the charge mattered much. Both the Panel and the Appellate Body base their analysis on the fact that Article II refers to charges ‘on importation’ of a product, while Article III relates to charges imposed on ‘imported’ products. In line with past GATT/WTO jurisprudence, the Panel considered that ‘a charge cannot be at the same time an “ordinary customs duty” under Article II:1(b) of the GATT 1994 and an “internal tax or other internal charge” under Article III:2 of the GATT’,Footnote ⁶⁶ and the Appellate Body silently seems to agree with this commonly held view.

However, as became clear in the case of India–Additional Duties, the language of GATT Article II.2 casts doubt on this commonly held view that there is such a strict separation. This case suggests that a charge that is imposed at the border and is equivalent to an internal tax may nevertheless be seen as an unjustified border charge that violates GATT Article II, as an internal tax that fails to meet the conditions of GATT Article III, or both.Footnote ⁶⁷ GATT Article II.2 provides that nothing in that Article relating to border measures shall prevent a Member from imposing ‘on the importation’ of any product a charge equivalent to an internal tax when this charge is imposed consistently with GATT Article III.2. This seems to provide a safe haven for a charge ‘on importation’ – i.e., a border charge – if it is equivalent to an internal charge. The nature of the charge would be that of a border charge covered by GATT Article II – why else would there be a need for this type of justification under Article II.2? If it were an internal charge, simply collected at the border, GATT Article II would not apply and there would be no need for this type of safe haven. And what is GATT Article II.2 actually saying – that a border charge (‘on importation’) can be justified as an internal charge so long as there exists a domestic equivalent for this charge and the charge complies with the nondiscrimination requirement of GATT Article III? This type of approach seems to argue against a strict and clean separation between border measures and internal charges based on certain essential characteristics of the measure.

Maybe a more pragmatic approach would be to say that measures that are imposed at the border should prima facie be considered to be border measures subject to GATT Article II. One should not assume that they are actually internal measures simply collected or enforced at the border, unless such charges are clearly linked to a domestic tax equivalent. Such was not the case here, and no argument to this effect was made by China. There obviously was no domestic counterpart to the 25% charge imposed on imported products. China clearly considered the charge to be an import duty, and the Chinese customs authorities were the ones collecting the charge.

It seems that the Panel could have reached the same conclusion of violation simply by accepting that China imposed a duty on imported parts that exceeded the tariff binding on auto parts as set forth in China's Schedule. The only relevant question before the Panel at that stage would be whether China was justified in imposing duties at the level reserved for complete vehicles on parts simply because they were later assembled into a ‘complete vehicle’ as defined by the Chinese authorities.

The answer to this question has to be ‘no’: all imported parts will one day be assembled into a complete vehicle and may thus become cars, but what is being imported are parts, not cars.Footnote ⁶⁸ Whether considered as an internal charge or as a customs duty, it is clear that China would be found to have violated its GATT obligations. Maybe the Panel could have done as the 1992 GATT Panel on the Canadian Liquor Board case did: it concluded that ‘the question whether the restrictions violated Article III.4 [on internal measures] or Article XI:1 [on border restrictions other than duties] of the General Agreement was … of no practical consequence in the present case’.Footnote ⁶⁹

4.2 Reviewing the meaning of domestic legislation – a matter of fact or law?

Article 17.6 DSU provides that an appeal ‘shall be limited to issues of law covered in the panel report and legal interpretations developed by the panel’. In other words, a challenge to a Panel's finding before the Appellate Body is limited to matters of law, and in principle it is not for the Appellate Body to review factual determinations made by the Panel.

The only indirect way of making the Appellate Body review factual determinations is through Article 11 DSU, which requires a Panel to make an ‘objective assessment of the matter’ before it. Article 11 DSU thus provides a way of challenging the Panel's analysis of the evidence or its general approach to establishing the facts. This appeal raised the question whether a Panel's interpretation of the meaning of municipal law is a legal interpretation by the Panel or a factual determination that, in the absence of an Article 11 DSU claim, cannot be reviewed by the Appellate Body.

In particular, the Panel had interpreted a provision in the challenged Chinese Decree 125 relating to the exemption for CKD and SKD kits to mean that such kits were subject to the charge under the Decree but not subject to the administrative requirements of this Decree. China appealed the findings of the Panel regarding CKD and SKD kits. It did not deny that CKD and SKD kits were subject to a 25% duty, but argued that this duty was not a duty imposed by the challenged Decree. China claimed that the Panel misread the Decree's exemption for CKD and SKD kits as being only a partial exemption. It thus requested the Appellate Body to examine the Decree and to review the Panel's erroneous reading of this Decree. The United States on the other hand argued that a Panel's ‘constructions of municipal law are factual determinations in WTO dispute settlement’ and that the Appellate Body may not review such findings de novo, but must accord them ‘the same deference as other types of factual findings made by Panels in WTO dispute settlement proceedings’.Footnote ⁷⁰

The Appellate Body, referring to its report on US–Section 211 Appropriations Act, rejected the US argument and considered that ‘When a panel examines the municipal law of a WTO Member for purposes of determining whether the Member has complied with its WTO obligations, that determination is a legal characterization by a panel, and is therefore subject to appellate review under Article 17.6 of the DSU.’Footnote ⁷¹ It then added the following qualification of this right to review a Panel's interpretation of a Member's municipal law, which we find particularly intriguing:

On the point of principle – whether a Panel's interpretation of the meaning of municipal law is a factual determination or rather a legal interpretation subject to Appellate Body review – the Appellate Body's ruling in this case is in line with earlier findings to this effect in India–Patents (US) and US–Section 211 Appropriations Act. Although we are of the view that this proposition is debatable,Footnote ⁷⁵ we wish to focus our commentary on the new twist given by the Appellate Body to its authority to review a Panel's interpretation of the meaning of municipal law.Footnote ⁷⁶

In the case discussed in this report, the Appellate Body qualified its ‘right to review’ a Panel's determination of the meaning of municipal law and stated that it would ‘not lightly interfere with a panel's finding’ when the Panel's assessment of municipal law goes ‘beyond the text of an instrument on its face’ and thus involves ‘factual elements’.

It is not clear to us what exactly the Appellate Body was trying to say in the bolded part of the above-quoted paragraph. It seems that the Appellate Body was suggesting that, if the determination of the Panel is based on the text of the municipal law without consideration of the application of the law or its interpretation by the domestic courts, then the Panel's reading of the meaning of the municipal law is subject to review by the Appellate Body. However, if the Panel relies on the application of the law or on domestic court rulings to determine the meaning of the law as such, then the Panel's determination rests on ‘factual elements’ that the Appellate Body would not ‘lightly interfere with’. Hence, the Appellate Body would view this interpretation as a factual matter, not subject to review, unless through Article 11 DSU.

The Appellate Body's reasoning appears to deny that, from the standpoint of international law, ‘municipal laws are merely facts’.Footnote ⁷⁷ That is why in the past the WTO Appellate Body has stated that the consistency of a Member's domestic laws must be assessed on the basis of the text of the law, its consistent application in practice, if applicable, and possibly the pronouncements of domestic courts on the meaning of such laws, as well as the opinions of legal experts and the writings of recognized scholars:

In Dominican Republic–Import and Sale of Cigarettes, the Appellate Body clearly rejected an attempt by the complainant to impose an analysis of its ‘as such’ claim based on the text of the legislation in isolation.Footnote ⁷⁹

This was also the approach followed by the WTO Panel on US–Section 301, which examined a section of US law, first to determine its meaning in order to subsequently assess whether the challenged section of US law was consistent with the obligations of the US under the WTO Agreements. This Panel correctly considered that it was not to:

The GATT Panel on US–Tobacco similarly considered that an examination of a Member's law required it to refrain from engaging in an independent interpretation of domestic laws and to treat the interpretation of such laws as questions of fact.Footnote ⁸¹

In other words, the Appellate Body's distinction between certain textual ‘legal’ elements and other nontextual ‘factual’ elements in the case discussed in this report is at odds with the way public international law, as well as established GATT/WTO case law, views the interpretation of municipal law. In addition, it seems that the Appellate Body forgot that in the case of India–Patents (US) – to which it expressly refers in support of its position that it can review a Panel's interpretation of domestic law – the Panel and Appellate Body did not simply examine the text of the Indian law in isolation. Rather, there was a considerable amount of evidence available in that case regarding the proper interpretation of the express terms of the Indian Patents Act, all of which was taken into consideration by the Panel and the Appellate Body in its review of the Panel's determination.

Therefore, it would seem proper for the Appellate Body to consider either that any determination by the Panel regarding the meaning of municipal law is a factual determination that is not subject to appeal (other than through Article 11 DSU), and thus to reverse its prior rulings in this respect; or to accept, as it seems to have done in the past, that the interpretation of municipal law in order to determine the law's consistency with the WTO Agreements is a legal interpretation that is subject to review by the Appellate Body, irrespective of the kind of evidence used by the Panel to determine the meaning of the law. A bit of both is not possible, and statements like the one made in this case do not clarify matters; quite to the contrary. The distinction that the Appellate Body seems to make between cases in which the Panel simply bases its understanding of the municipal law on the text of the law as such, without resort to the application of the law, and other cases in which ‘factual elements’ are incorporated in the analysis is, in our view, misguided.

Such statements, which seem intended only for the Appellate Body in a later case to refer to when it does not feel like reviewing a Panel's interpretation of a Member's municipal law, do not provide the system with ‘security’ or ‘predictability’. To the contrary, they create doubt and uncertainty as to the scope of the Appellate Body's review authority in important cases involving the interpretation of municipal law.

5.1 Incentives for abuse on both sides

It is clear from the above that the ‘legal bone’ in the case is relatively straightforward in the sense that the Panel's rulings and the Appellate body's decisions are very much in line with earlier rulings and decisionmaking and can therefore be considered as time-consistent. From a legal point of view this case can even be described as highly ‘predictable’ in the light of similarity with earlier dispute cases. The time consistency by the Panel and the Appellate body reduces uncertainty and enhances transparency in decisionmaking, which is important for trade.

From an economic point of view, time-consistency tends to rank second to the objective of welfare. The ultimate question to an economist is how policy affects the welfare of the parties involved and whether or not a policy is welfare-improving from a unilateral or from a multi-lateral perspective. However, in practice this is a much more difficult objective to verify than time-consistency of policy, but this should not prevent us from striving for better outcomes than a mere pragmatic approach.

Ideally, WTO rules should be in place to make certain that world welfare is maximized; this can be attained by drafting and applying rules to ensure that countries do not pursue their unilateral welfare-maximizing policies at the expense of other countries, which would be a ‘beggar-thy-neighbor’ policy where one country's gain would be another country's loss. From an economic point of view, what determines who is ‘wrong’ or who is ‘right’ in a trade-policy conflict is often the beggar-thy-neighbor policy content of the action undertaken. Policies where one country gains at the expense of another are unlikely to yield stability, which may hurt trade, and they are therefore undesirable. Extrapolated to this dispute case under scrutiny, some insight into who ‘gains’ and who ‘loses’ from the higher duty imposed on imported car parts could have helped the Panel to make an even more informed decision. From this perspective several important questions were not raised in the course of the dispute and consequently no factual evidence was gathered on them.

Economists feel particularly uncomfortable if no factual evidence is gathered to substantiate the allegations made on both sides of the fence, with China accusing foreign suppliers of ‘circumventing’ import duties (GATT Article XX), and importers accusing China of pursuing an ‘industrial policy’ aimed at favoring the domestic car industry (Article III:2 of the GATT).

Without sufficient factual evidence, the Panel cannot but remain rather speculative on certain accounts. For instance, the Panel asserted that the import duty of 25% on car parts, decided on in 2004 and taking effect in 2005, is an internal charge (Article III:2) favoring the import-competing Chinese industry of motor parts. But for this to be true, it would have been relevant to investigate whether such an import-competing industry existed and whether it was benefited during the time that the Chinese applied their elevated import duty on parts. In the absence of such an industry, car assemblers in China would not be in a position to source from an alternative domestic supplier and would be forced to pay the higher import duty on car parts, squeezing their profits. This would not be in their interest, and that would leave us wondering why China would introduce such a measure. Follow-up questions that come to mind are: Who are the car assemblers in China? To what extent are they foreign firms or domestic firms? In other words, ownership matters in order to determine who carries the burden of the duties. Most importantly also: Who buys the assembled cars? Is it domestic Chinese consumers or foreign ones, and to what extent is the import duty passed through to the consumer in the form of higher prices? That in turn will depend on the demand elasticity for assembled cars, which is often correlated with overcapacity in the industry.

But before we enter a more systematic treatment of some of the questions above, we can ask ourselves why China imposed a higher tariff on the imports of motor vehicles than on motor parts before 2004 in the first place. It is true that, in the past, many great nations have pursued similar tariff structures for the purpose of building their own manufacturing industries using the dual tariff structure as a tool of industrial policy.Footnote ⁸² A country that makes the imports of complete motor vehicles relatively more expensive than imports of parts provides an incentive in favor of the shipments of parts and invites new investment in assembly activities. Additional investment in car assembly can occur through domestic entry or foreign investors keen on ‘jumping’ the high import duty on complete motors. One underlying reason for China to pursue such a policy could be that assembly activities generate technology spillovers and know-how that China would like to internalize rather than simply import from abroad. Additionally, it may also reflect a concern for creating domestic Chinese employment, since assembly is usually a labor-intensive activity. While China has not been violating any rules by setting its dual tariff structure on motor vehicles versus car parts at the time of entry into the WTO in 2001, it is undeniable that this gave rise to certain incentives for foreign suppliers, namely to ship more parts instead of complete motor vehicles. Hence, we expect the beneficiaries of the dual tariff schedule imposed to be the foreign suppliers of parts and the losers to be the foreign importers of complete motor vehicles. But an adversely affected importer of complete motor vehicles, faced with a high import tariff of 25%, can try to circumvent the duty by taking his complete motor vehicle apart, shipping the parts to China, and paying 10% import duties, thus saving 15% in duties. Economically speaking, the incentive to circumvent China's dual tariff structure was definitely present before 2004, and there is nothing in the case analysis to suggest that this did not actually occur. Indeed, history tells us that circumvention of high import duties through a strategy of disassembling products into parts is a distinct possibility and a strategy pursued by importers of final goods adversely affected by import duties.

A well-described example of this occurred during the 1980s when Japan was heavily targeted by European antidumping duties on its imports of consumer electronics into Europe. The Japanese tried to avoid the import duties by ‘jumping’ the duties by setting up assembly plants in Europe so that they could start shipping to Europe parts on which they did not have to pay duties; the parts were then further assembled inside the EU into consumer electronics and sold on the European market. The EU tried to put a stop to this through the ‘screwdriver’ cases by levying import duties on the parts similar to import duties on final consumer electronics. Academically, the causality between antidumping duties and Japanese ‘antidumping’ jumping is well documentedFootnote ⁸³ but from a legal point of view the import duties were not WTO-consistent, since they discriminated foreign parts against domestic parts and were considered an internal charge.

Is this what has been going on in China? Has there been an increase in the imports of car parts? Has there been an inflow of FDI during the time of the dual tariff structure, accompanying the increase in imported parts, which could be a sign of circumvention? If the objective of the Chinese government was to create its own car assembly industry by using trade policy as an industrial policy, did they succeed in doing that? Did the Chinese car-manufacturing sector grow during that time?

For a WTO Panel, it should be relatively easy to get access to this kind of data, if only because it can ask the parties involved in a dispute to supply the relevant information. While the questions above are not directly related to the object of the dispute, for an economist ‘initial conditions’ are important. There can be little doubt that the imposition of the dual tariff structure, with 25% duties on imports of motor vehicles but only 10% on the imports of parts, created certain incentives and dynamics in the car industry in China that laid out the foundations or ‘initial conditions’ in the light of which this dispute case should be considered.

Without access to firms' private information, we can only turn to the available public information in an attempt to get a better understanding of some of the allegations made in the dispute case.Footnote ⁸⁴

We start in Figure 1 by examining the evolution of imports into China of complete motor vehicles as well as car parts to verify if the incentives that we believe to arise under China's trade policy are somehow reflected in the trade figures. Figure 1 shows that in 2002, right after China's entry into WTO, the imports of car parts equaled the imports of complete motor vehicles. But shortly after that, the imports of parts into China increasingly exceeded that of complete motor vehicles. This seems to suggest that the imposition of a dual tariff structure may indeed have created incentives for relatively more car parts being imported relative to complete vehicles. But after 2004, the gap between imports of parts and imports of complete motor vehicles narrowed, and they became equal to each other again in 2007. This seems to suggest that the measures after 2004 had taken effect. After an initial boost in the imports of car parts coinciding with an initially slower growth in imports of motor vehicles, imports in both types of activities converged again in 2007, possibly due to the circumvention policy on car parts that China pursued by that time, which again made imports of parts relatively less attractive.

Source: UN Comtrade, imports reported in dollars.

Figure 1. Imports of complete motor vehicles and car parts

In that same period, we also observe a substantial inflow of Foreign Direct Investment (FDI) in the car-manufacturing sector in China (in million USD). Table 1 shows that FDI especially increased before the years 2004 and 2005, but stabilized afterwards in the period that China imposed the circumvention duty that, under certain conditions, implied a similar duty treatment for parts as for complete motor vehicles. FDI inflows were highest in 2004 and 2005 but decreased again substantially in 2006. This seems to be consistent with the hypothesis that before 2005 there was a lot of tariff-jumping FDI taking place, and after the introduction of a high tariff on car parts, that kind of FDI inflow stopped.

Table 1. Inflow of foreign direct investment in Chinese car manufacturing

Source: Ministry of Commerce of PR China.

At first glance, these facts could be consistent with a story of duty circumvention by foreign suppliers of car parts. But in order to assert that circumvention has taken place, there is a missing piece of the puzzle – i.e., circumvention assumes that the same suppliers previously shipping complete motor vehicles are the same ones setting up assembly plants in China and afterwards switching to shipping parts.

In this dispute case, no evidence was collected to verify that the ownership of the importing firms was the same as that of the FDI firms, thus distinguishing the EU screwdriver cases mentioned earlier, where the same Japanese firms were involved in first shipping complete goods and afterwards setting up plants in Europe and shipping parts instead of motor vehicles. The facts observed – i.e., increase in the shipment of parts relative to complete motor vehicles and inward investment – could just reflect different firms entering the car-manufacturing industry attracted by the incentives provided by the dual tariff structure.

Clear evidence on circumvention would have to show that the same firms previously supplying the Chinese market with complete cars were engaging in inward FDI in China and altering their shipments to China into parts rather than complete cars in response to the dual Chinese tariff structure in the initial years after 2001.

An alternative interpretation of the facts is that China deliberately used its tariff policy as part of an industrial policy. China may have used trade policy for the purpose of building its own car industry. The tariff-jumping FDI triggered by the dual tariff structure may have been anticipated and intended in order for the technology embedded in the assembly of cars to spill over from foreign investors to domestic car assemblers once foreign firms had invested in China. Academically speaking, the effectiveness of such an industrial policy is questionable, since the empirical results on FDI spillovers are ambiguous and inconclusive. At best, horizontal spillovers of FDI amongst firms engaged in the same type of activity are found to be weak. Similar types of firms act more as rivals and try to prevent knowledge spillovers to horizontal competitors. The only robust result emerging from the spillover literature is the presence of vertical spillovers, i.e. between a foreign multinational and its indigenous suppliers,Footnote ⁸⁵ which could have given China a reason, in a second phase of its industrial policy, to pull up the import duty on car parts for the purpose of fostering local suppliers and to encourage vertical spillovers between the downstream FDI that had jumped the original tariff on complete vehicles and the local suppliers.

It is not unimaginable that China, through the phasing of its trade policy, tried to create its own car industry.Footnote ⁸⁶ Introducing a high tariff wall on the imports of complete motor vehicles in a first stage favors domestic assembly of cars and stimulates foreign and domestic entry of car assemblers. Once the car assembly activity is well developed and foreign investors with their technology are in place, a higher tariff on the imports of car parts could constitute a second stage in China's industrial policy, where it aims to stimulate the domestic upstream production of car parts. Imposing a higher tariff on imports of parts favors ‘local sourcing’ to the benefit of the domestic import-competing car-parts industry but to the detriment of ‘foreign sourcing’ of parts, which brings us to the current dispute-settlement case under analysis. By pulling up the tariff on imported parts to 25%, imports of car parts will be adversely affected and car manufacturers inside China eager to save on costs are likely to switch suppliers and source from local suppliers (if present) that are not subject to the import duty.

The use of trade policy to build import-competing industries is a development strategy sometimes referred to as the ‘infant-industry’ argument for trade protection. However, by now we know that strategy is a fallacy. Baldwin showed that using trade policy is welfare-inferior to free trade.Footnote ⁸⁷ A better way to build a domestic manufacturing industry is through the capital market that should provide the necessary financing and loans to facilitate the construction of new industries. This is far less distortive and is welfare-superior to trade policy to achieve that purpose. Of course, for the Baldwin argument to hold, capital markets should function properly. In China, however, there is evidence to suggest that credit markets do not function properly and are far less efficient than in most developed countries,Footnote ⁸⁸ which may well explain China's incentive to revert to a relatively old way of constructing an import-substitution industry using trade policy.

The discussion above reveals that the incentive structure in this dispute case is ‘open ended’ in the sense that the rules in place undeniably give foreign suppliers an incentive to ‘circumvent’ the initial duty on complete vehicles (which is the Chinese stance in the dispute), although this remains unverified. At the same time, the Chinese government may also ‘abuse’ the higher tariff on car parts as a second phase in its industrial policy aimed at fostering its import-competing car-parts industry, which favors ‘local sourcing’ over foreign sourcing (which is the EU, US, and Canadian stance in the dispute) in the hope of benefiting from vertical spillovers coming especially from the foreign investors that previously jumped the high tariff walls on complete motor vehicles in the first phase of the industrial policy. This is a well-known development strategy used in the past by other countries, but it can be regarded as welfare-inferior to other domestic policies.

5.2 Four relevant questions related to the case

There are at least four relevant economic questions that were not addressed in the dispute case but that may have yielded a clearer understanding of the beggar-thy-neighbor policy content of China's trade policy.

Who bears the cost of the additional import duty?

A first relevant question would be whether the assembled motor vehicles in China were used for domestic sales to Chinese consumers or were meant for export destinations and consumers outside China. This is relevant in terms of who bears the ultimate extra cost of the surcharge on imported motor parts. If assembled cars are mainly for Chinese consumers, Chinese consumers are likely to pay a higher price for their cars as a result of the higher import duty on parts and Chinese trade policy would be carried domestically. If, however, the assembled cars are mainly exported, it is consumers in export markets that are likely to pay the additional price, but only if demand overseas is not too elastic, which is the second relevant economic question treated below. But let us first look at whether assembled cars are sold for exports or domestically.

Figure 2 below is derived from firm-level dataFootnote ⁸⁹ and gives us a peek at the importance of domestic sales versus exports of complete motor vehicles (2a) and car parts (2b). It appears that most complete motor vehicles are sold to domestic customers and a much smaller number is shipped for exports. While domestic sales remained relatively stable over time, exports continued to rise between 2004 and 2006. The only mention of exports in the dispute case under scrutiny is footnote 14,Footnote ⁹⁰ which states that exports are exempted from import duties, which may explain why they continued to rise. Figures 2a and 2b suggest that the import duties on complete car vehicles or car parts would mainly affect Chinese customers, given that domestic sales are more important than exports.

Figure 2a. Complete motor vehicles

Source: Oriana firm-level data.

Figure 2b. Car parts

How elastic is the final demand for cars?

A second relevant question is whether or not the final demand for assembled motor vehicles is elastic. The demand elasticity determines the negative slope of the demand curve and reflects how consumers move away from buying a motor vehicle when the prices go up. This elasticity is indicative of the extent to which an increase in costs resulting from an extra charge on inputs used in the production process can be passed through to the consumer. Put differently, the demand elasticity captures how sensitive prices for assembled motor vehicles are. Car manufacturers may pass the higher import duty on to consumers by charging a higher price, but if demand is very elastic, this would imply a large drop in the sales of cars, and therefore car producers may prefer to absorb the duty even though it implies a squeeze in their profits. To estimate the elasticity of demand directly is not possible for us, due to lack of data on quantities sold and prices charged, but we can look at car manufacturers' profits over the years to see what happens there. Figure 3 shows that profits of both car manufacturers and car-parts producers in China were on a slide before the introduction of the new tariff structure, with profits for manufacturers of complete motor vehicles much higher than those for car-parts manufacturers. But after 2005, when the new tariff structure went into effect, profits for both types of firms rose again, suggesting that the tariffs benefited both. This is indicative that the additional costs from the import tariffs were passed through at least in part to consumers, resulting in price increases that did not affect sales too much, leaving the car sector with higher profits than before.

Source: Oriana firm-level data.

Figure 3. Profits of motor vehicle assemblers in China

Is there an import-competing Chinese car-parts industry?

The third question is on the existence of an import-competing domestic Chinese motor-parts industry and whether it benefited from the higher tariff on car parts. If there is such an industry, to what extent did the import duty of 25% benefit them?

This is an important question, since the Panel's conviction of China on the basis of Article III:4 assumes there is. Violation of Art III:4 implies that a country uses trade policy to favor its local industry, since the underlying reasoning is that, when car assemblers in China are faced with more expensive foreign parts, they are likely to switch to domestically produced parts that have become relatively cheaper because the duty does not apply to them. The existence and size of an import-competing Chinese industry of parts is therefore important, since this would shed more light on how substantiated the claim is that the high import duty on foreign parts favors domestic parts suppliers.

We collected firm-level information on the group of firms in car manufacturing and the group of firms in China reporting ‘car parts’ as their main activity.Footnote ⁹¹ One of the variables of interest is firm-level investment in car assembly and parts. Figure 4 shows the evolution of fixed tangible assets of firms in China in the car sector, which is generally regarded as a good proxy for investment. While both complete motors and car parts attracted investment, the growth in investment is more spectacular in the car-parts industry between 2004 and 2006, the period where the circumvention duty of the Chinese authorities applied. The investment in machinery, land, buildings, etc., grew so fast that the car-parts industry by 2006 had reached a similar level of assets as that of the motor-vehicles sector.

Source: Oriana.

Figure 4. Investment in the Chinese car sector

This sparse evidence already suggests two important facts that substantiate the claim made by the WTO Panel that the local industry was favored by China's trade policy. It shows, first, that an import-competing industry for car parts existed at the time, and, second, that it seems to have grown substantially over time at a higher pace than the car-assembly industry during the period the circumvention duty applied.

Who owns the car manufacturing industry in China?

A fourth relevant question is about ownership. To the extent that the entire car-parts industry is Chinese, the effects of a change in import duties on car parts would fall on Chinese shareholders. However, if there is a substantial amount of foreign ownership in that sector, the effect of an import duty on car parts would also be shared by them.

Ownership information is typically very hard to come by. While our firm-level data set has some information on ownership, it is time-invariant and is far from perfect. From Table 2 we observe that, based on our micro data, the foreign ownership in both the complete-motor-vehicle sector and in the car-parts sector lies around 20%.

Table 2. Number of firms in the car sector in China

Notes: ‘Foreign’ includes: ‘Foreign Invested Stock Limited Company’, ‘Sino-Foreign Cooperative Venture’, ‘Sino-Foreign Joint Venture’, ‘Wholly Foreign Owned Enterprise’, ‘HK, Macau or Taiwan Joint Cooperative Venture’, ‘HK, Macau or Taiwan Joint Venture’, ‘Wholly HK, Macau or Taiwan Invested Stock Limited Company’, or ‘Wholly HK, Macau or Taiwan Owned Enterprise’. The rest of the firms are defined as ‘Domestic’.

Source: Oriana.

Furthermore, it can be noted that in the more downstream industry of complete motor vehicles, the number of firms is around 287, making this a relatively concentrated sector. Concentrated sectors generally are less subject to competition, and firms have more market power.

This is important for pass-through issues, as we know that more concentrated sectors are more likely to pass on adverse cost shocks to consumers. The total number of firms in the more upstream activity of the car parts is around 5,800 different producers, making this sector much less concentrated. Typically, competition in less concentrated sectors is stronger and market power lower. The pass-through to consumers is therefore likely to be more limited in car parts than in complete motor vehicles.

In terms of the nationality of the firms, we present in Table 3 the information available to us. The number of firms for which country of ownership could be retrieved is relatively small, but it gives us an idea of the proportions. Foreign ownership is relatively more common in the complete-motor-vehicle sector than in the more upstream car-parts industry. The largest group of foreign owners is the Asian countries (Japan, Korea, Hong Kong, Taiwan, and Singapore), followed by the EU, followed by North America (US and Canada). The presence of foreign firms in the Chinese car sector could reflect an outsourcing strategy where the purpose of foreign firms is to ship the goods they produce back to their country of origin. Or their presence could reflect ‘horizontal FDI’ for the purpose of serving the domestic Chinese market. In the first case, they would be far less affected by China's trade policy due to the exemption of exports (footnote 14 in the dispute case), i.e. goods produced for export markets are exempted from the high 25% duty on imports of car parts.

Table 3. Nationality of foreign firms

Notes: The sample used in this table includes firms that have information on their ownership AND have foreign owners. One firm can be owned by owners from different countries or regions.

Source: Oriana.

This can be verified in Figure 2, where exports continued to rise when domestic sales stabilized. In the second case, where FDI is of the ‘horizontal type’ for serving the domestic Chinese market, the import duties can adversely affect foreign firms' operations in China, depending on the pass-through to the Chinese customer.

Our gathering of factual evidence can only be regarded as indicative due to limitations in data, time, and manpower. Our main purpose was to raise questions rather than to pretend to answer them in full. But the many questions that the case leaves unaddressed suggest that the ‘economic bone’ is less straightforward to split than the legal one. If anything, the empirical facts that we gathered suggest that in the end the Panel most likely was right in ruling out the ‘circumvention’ hypothesis and embracing the ‘industrial policy’ hypothesis.

While the occurrence of ‘circumvention’ cannot be completely ruled out for reasons explained above, the facts pointing in the direction of China's use of trade policy as a tool of industrial policy are plentiful. The phasing of trade policy by China is almost a textbook case of an ‘import substitution policy’ pursued by other countries in the past. At the time of WTO entry, China first set a high tariff on the imports of motor vehicles to build manufacturing capacity in the upstream activity of car assembly, resulting in inward FDI and discouraging imports of complete vehicles. A couple of years down the road from WTO entry, China altered its trade policy by imposing a circumvention duty that raised the tariff on the imports of more upstream car parts, which are intermediates in the assembly of cars. This discouraged the imports of both car parts and complete vehicles and enhanced a spectacular increase in the investment in and capacity of the domestic Chinese car-parts industry. Furthermore, the exemption of exports from paying import duties on parts suggests that China combined its import-substitution policy with a well-known policy described by: ‘import protection as export promotion’.Footnote ⁹² In a model where import protection raises the market size for a representative domestic producer, domestic sales for the domestic firm increase This allows the firm to lower its average cost curve and makes it more competitive, resulting in more exports. Hence, the Krugman model pointed out that import protection may actually stimulate exports.

While too little factual evidence is currently presented to formally accuse China of pursuing these ‘beggar-thy-neighbor’ policies, all we can say is that some of the facts, such as the spectacular growth in the Chinese car-parts industry after the 25% duty on parts was introduced, are consistent with it.

From an economic point of view, it is interesting to note that ‘old’ policy recipes appear to survive despite the fact that we know that trade policies can only be second best compared to other domestic instruments. Building industries behind tariff walls is less effective than investing in properly functioning capital markets.Footnote ⁹³ Recent evidence also questions the ability to use import protection to raise exports when firms are heterogeneous.Footnote ⁹⁴