Domestic political implications of global value chains: Explaining EU responses to litigation at the World Trade Organization

Abstract

This article shows how mobilized sectors’ degree of integration to global value chains can account for variation in the European Union’s responses to litigation at the World Trade Organization (WTO). The behavior of the EU demonstrates a puzzling picture in bringing about compliance following WTO disputes, sometimes meeting its trade partners’ demands in a timely manner, while resisting resolving the disputes in others. In order to account for this variance, I aim to answer the research question: under which conditions does the EU meet its trade partners’ demands in a timely manner, rather than resist compliance, in the WTO dispute settlement mechanism? Combining comparative analysis with case studies, I find that the EU’s behavior in bringing timely compliance to disputes can be explained by mobilized targeted sectors’ level of integration to global value chains (GVCs). As disputes target certain measures that tend to benefit-specific sectors, the level of integration of these targeted sectors shapes the EU’s preference toward pro or against liberalization. The results of the study corroborate the finding that trade policy is shaped by the interests of domestic economic actors and that those engaged in GVCs foster liberalization multilaterally.

Appendix: Domestic Political Implications of Global Value Chains: Explaining EU Responses to Litigation at the World Trade Organization

Additional Qualitative Comparative Analysis (QCA): The Impact of Veto Players on Compliance at the WTO DSM

In my primary analysis, I was not able to use the number veto players as a potential explanatory condition that might impact the EU’s record of compliance at the WTO DSM. The reason is that there is no systematic data available on veto players of the EU for each and every policy change enacted. The only comprehensive data source is the widely used Political Constraints (PolCon) database of Witold Henisz, which does not have values for the EU government.⁵ The PolCon database also does not include information on veto players whose consent is necessary to bring policy change for each and every piece of enacted legislation, which is the appropriate variable for my analysis.

In response, I tried to compile a dataset and outline the number of veto players whose consent was necessary to enact legislation in response to WTO litigation for the disputes lodged against the EU. I tracked the EU’s domestic legislative changes following every single WTO dispute but I could only locate 14 domestic policies that were passed in response to a WTO dispute. Considering that only works in the field who use veto players as a variable are the ones that conduct case studies, I am not surprised by this limitation. This means that I have veto players information for about 45 % of my sample, which is too low to conduct a reliable analysis. However, I decided to run the analysis in order to show the potential impact of veto players by arbitrarily coding the number of veto players. Considering that the EU government inherently has substantial checks and balances within, I characterized the average veto players’ capacity of the EU to be “relatively high” and then used this coding for the missing values. The results of this illustrative analysis are outlined below.

Analysis of Necessity and Sufficiency

As shown in Table 3, the analysis of necessity with the condition of “low veto players” reveals that targeted sectors’ mobilization and GVC integration remain the two most significant conditions for timely compliance. The examination demonstrates that neither veto players nor the combination of veto players with sectoral mobilization and GVC integration qualify to be considered necessary for the outcome of timely compliance. The analysis of sufficiency shown in Table 4 below paints a similar picture. It seems that sector mobilization and GVC integration are both necessary and sufficient conditions for timely compliance without veto players.

Table 3 Necessity of the conditions leading to the outcome of timely compliance

Table 4 Sufficiency of the conditions leading to the outcome of timely compliance

The first row in Table 4 shows that of the disputes that affected mobilized and GVC-integrated sectors, 9 out of 10 resulted in timely compliance. These results are substantively similar to the primary results obtained in the article, but there seems to be an important caveat revealed in this sufficiency analysis. We can see on the 2nd row that of the disputes that targeted mobilized GVC-integrated sectors and needed the consent of low numbers of veto players for policy change, all of them resulted in timely compliance. Even though such a combination of conditions is quite unlikely – only 5 disputes are under this configuration – it does reveal an important detail.

Moreover, Tables 5 and 6 below are the results of the QCA excluding the missing values on veto players, which means the analysis loses more than half of the original observations – I thought readers might be interested to see the results for illustrative purposes. The examination demonstrates low number veto players to actually have a relatively high consistency level as a necessary condition for timely compliance. Also, the first row of Table 6 shows that the combination of the three conditions, “high sector mobilization,” “high sector integration to GVCs,” and “low veto players” results in timely compliance for all the disputes that fall under that configuration.

Table 5 Necessity of the conditions leading to the outcome of timely compliance (without the missing values)

Table 6 Sufficiency of conditions leading to the outcome of timely compliance (without missing values)

However, although the analyses above reveal some association of veto players with compliance, both the necessity and the sufficiency analyses reiterate the significance of mobilization and GVC integration. Even when including veto players, the results are substantively similar to the primary analysis outlined in the article and affirm that the EU brings forth timely compliance as a response to WTO disputes when the targeted sectors are mobilized and integrated to GVCs. Therefore, I refrain from making a conclusive statement about the impact of veto players on compliance across the board, but considering the data limitations, and the observed association I outlined above, the relationship between veto players and compliance would perhaps be better unpacked with alternative research designs.

Descriptive Appendix

Disputes Lodged Against the EU at the World Trade Organization Dispute Settlement Mechanism

Description of the dispute	Complainant(s)
DS7 – Trade Description of Scallops	Peru, Chile, Canada
DS9 – Duties on Imports of Cereal and Grains	Canada, USA
DS27 – Regime for the Importation, Sale and Distribution of Bananas	Guatemala, Honduras, Mexico, Ecuador, Panama, Colombia, USA
DS48 – Measures Concerning Meat and Meat Products (Hormones)	Canada, USA
DS69 – Measures Affecting Importation of Certain Poultry Products	Brazil
DS72 – Measures Affecting Butter Products	New Zealand
DS134 – Restrictions on Certain Import Duties on Rice	India, Thailand, Uruguay
DS141 – Anti-Dumping Duties on Imports of Cotton-Type Bed Linen	India
DS153 – Patent Protection for Pharmaceutical and Agricultural Chemical Products	Canada
DS154 – Measures Affecting Differential and Favourable Treatment of Coffee	Brazil
DS172 – Measures Relating to the Development of a Flight Management System	USA
DS219 – Anti-Dumping Duties on Malleable Cast Iron Tube or Pipe Fittings	Brazil
DS231 – Trade Description of Sardines	Peru
DS260 – Provisional Safeguard Measures on Imports of Certain Steel Products	USA
DS263 – Measures Affecting Imports of Wine	Argentina
DS283 – Export Subsidies on Sugar	Thailand, Australia
DS286 – Custom Classification of Frozen Boneless Chicken Cuts	Thailand, Brazil
DS290 – Protection of Trademarks and Geographical Indications for Agricultural Products and Foodstuffs	Australia, USA
DS292 – Measures Affecting the Approval and Marketing of Biotech Products	Canada, USA, Argentina
DS299 – Countervailing Duties on Dynamic Random Access Memory Chips	Korea
DS301 – Measures Affecting Trade in Commercial Vessels	Korea
DS313 – Anti-Dumping Duties on Certain Flat Rolled Iron or Non-Alloy Steel Products	India
DS316 – Measures Affecting Trade in Large Civil Aircraft	USA
DS326 – Definitive Safeguard Measures on Salmon	Chile, Norway
DS349 – Measures Affecting the Tariff Quota for Fresh or Chilled Garlic	Argentina
DS376 – Tariff Treatment of Certain Technological Products	Japan, Taiwan
DS385 – Expiry Reviews of Anti-Dumping and Countervailing Duties Imposed on Imports of PET	India
DS389 – Certain Measures Affecting Poultry Meat and Poultry Meat Products	USA
DS397 – Definitive Anti-Dumping Measures on Certain Iron or Steel Fasteners	China
DS405 – Anti-Dumping Measures on Certain Footwear	China

Cases that were Taken Out of the Sample

Dispute information	Reason to be excluded from the sample
DS246 – Conditions for the Granting of Tariff Preferences to Developing Countries	Horizontal measure
DS315 – Selected Customs Matters	Horizontal measure
DS408 – Seizure of Generic Drugs in Transit	Horizontal measure
DS104 – Measures Affecting the Exportation of Processed Cheese	Information unavailable
DS266 – Export Subsidies on Sugar	Information unavailable
DS337 – Anti-Dumping Measure on Farmed Salmon from Norway	Information unavailable
DS400 – Measures Prohibiting the Importation and Marketing of Seal Products	Information unavailable (outcome still pending)
DS115 – Measures Affecting the Grant of Copyright and Neighbouring Rights	Measure challenged against a single member state (Ireland)
DS135 – Measures Affecting Asbestos and Products Containing Asbestos	EU exonerated

Calibration of the Cases: Characteristics of Sectors in Each Dispute Initiated Against the EU

Dispute Settlement No.	Timely Compliance	High GVC integration	High Mobilization	High export dependence
7	0.6	0.6	0.6	0.4
9	0.6	0.4	0.6	0.6
27	0.2	0.4	0.6	0.6
48	0.2	0.2	0.6	0.6
69	0.4	0.4	0.6	0.2
72	0.6	0.6	0.6	0.2
115	0.6	0.4	0.4	0.6
134	0.6	0.4	0.6	0.2
141	0.2	0.6	0.8	0.2
153	0.6	0.8	0.8	0.2
154	0.6	0.6	0.6	0.2
172	0.6	0.6	0.6	0.8
219	0.8	0.6	0.6	0.2
231	0.8	0.6	0.4	0.2
260	1	0.6	0.8	0.8
263	0.6	0.6	0.8	0.2
283	0.6	0.6	0.6	0.2
286	0.6	0.6	0.6	0.2
290	0.6	0.8	0.6	0.8
292	0.2	0.4	0.8	0.6
299	0.8	0.8	0.6	0.4
301	1	0.6	0.6	0.2
313	1	0.8	0.8	0.4
316	0	0.6	0.8	0.8
326	1	0.4	0.6	0.4
349	0.8	0.6	0.8	0.2
376	0.8	0.8	0.6	0.4
385	0.6	0.8	0.6	0.2
389	0.6	0.4	0.6	0.4
397	0.6	0.8	0.6	0.6
405	0.6	0.6	0.4	0.4

Details of the Coding Method: Qualitative Calibration

Outcome: Timely Compliance

The disputes that were resolved via a mutually agreed solution were considered in the set of “timely compliance.” For the empaneled disputes, if they are resolved within 15 months, they are considered to be in the set of “timely compliance,” which is the required time limit by the Dispute Settlement Understanding for the implementation of panel rulings.

• 0 Still in non-compliance – the demand has not been met.

• 0.2 Extremely untimely – e.g., 56 months.

• 0.4 Untimely meeting of trade partner’s demand, yet still relatively close to a timely limit – e.g., 19 months.

• 0.5 Maximum ambiguous point where I can’t decide whether or not the EU’s response was (un)timely.

  *(earlier than, and including, 15, is considered in the set of “timely compliance”)*

• 0.6 Considered timely compliance but still quite late – e.g., 14 months.

• 0.8 Extremely timely, significantly faster than the operationalized limit – e.g., 4 months.

• 1 Meeting a partner’s demand almost without a hesitation, perhaps within a month or so.

Conditions

High GVC Integration

It is a composite measure calculated by looking at 2 indicators. Firstly, the targeted sector’s outward FDI stocks are used, obtained from OECD Statistics. Secondly, the targeted sector’s import dependence is used – which refers to the amount of imports a sector uses for its output. Import dependence is calculated by using World Input Output Table Database (WIOD). Note that import dependence of sectors and FDI are highly (positively) correlated, yet in the instance where there was a contradiction, import dependence is taken as reference of GVC integration because FDI data are combined of EU member states’ FDI, and in the face of a national bias in EU-wide statistical measures, investment is not differentiated between intra-EU and extra-EU. Therefore, it can be argued that FDI data are overstated and thus import dependence is more reliable as an indicator of internationalization of production.

• 0 Completely not integrated – 0 % import dependent and 0 EUR in outward FDI stocks.

• 0.2 Demonstrates some level of integration – up to 2 % import dependent and holding up to 20 billion EUR in FDI stocks.

• 0.4 Moderately integrated but not sufficient enough to be “highly integrated” – up to 4.8 % import dependent and holding up to 49 billion EUR in FDI stocks.

• 0.5 Completely ambiguous to call EU’s sector integration to GVCs “high” or “low” – exactly 4.8 % import dependent and holding 49 billion EUR in FDI stocks – both are average values of import dependence and FDI.

• 0.6 Integrated sector, based on both FDI and import dependence, but still relatively low – between 4.8 and 6 % import dependent and holding between 49 and 59 billion EUR in FDI stocks.

• 0.8 Very high integration – over 6 % import dependent and holding over 59 billion EUR in FDI stocks.

• 1 Highest level of observed integration – 10 % import dependence and above, and holding over 100 billion EUR in FDI stocks.

High Mobilization

It was originally designed as a composite measure using both HHI and employment figures, but the problem of accessing HHI data further limited the calculation of the measure. I used HHI if and when available; otherwise, only use the percentage of employment (of the targeted sector) as a proxy.

• 0 Completely not mobilized – lower than 500 HHI score and less than 1 million persons employed.

• 0.2 Some level of mobilization – up to 700 HHI score and 1 million persons employed.

• 0.4 Relatively mobilized, yet not enough to be considered “highly mobilized” – up to 1000 HHI score and 1.3 million persons employed.

• 0.5 Maximum ambiguous point to call the sector’s mobilization “high” or “low” – exactly 1000 HHI score and 1.3 million employed.

• 0.6 Considered to have the minimum requirements to be in the set of “high mobilization” – HHI score between 1000 and 1500 and the number of employed between 1.3 and 2 million.

• 0.8 Highly mobilized – HHI score of more than 1500 and more than 2 million persons employed.

• 1 Highest possible levels of mobilization – HHI score of 2000 and employment of 9 million (the highest observed levels).

High Export Dependence

The measure is the export dependence of the targeted sector to the complainants’ market. Each targeted sector’s total value of exports to the complainant (for the year of the dispute) is collected from World Integrated Trade Solutions (WITS). The real values of exports to the complainant(s) are then calculated against the total exports of the sector, which gives the export shares of the targeted sector to the complainant(s). For the conjoined disputes with more than one complainant, export values are combined to give a reliable measure of export dependence. The 6-level calibration is calculated according to the average export dependence of the EU, which is 7.28 %. The minimum value of the EU’s sectoral export dependence is 0.1 percent and the highest is 24.9 percent.

• 0 No export dependence – no observable exports to the complainant.

• 0.2 Some export dependence, no significant threat of retaliation – up to 3 percent export dependent.

• 0.4 Relatively high export dependence, not enough to be considered to have “high” retaliatory capacity – between 3 and 7.2 percent export dependent.

• 0.5 Maximum ambiguous point: Not able to decide if the EU has “high” or “low” export dependence to the complainant – exactly 7.2 percent export dependent.

• 0.6 In the set of “high export dependence” but relatively close to being considered “not so high” – between 7.2 and 18 percent export dependent.

• 0.8 Very high export dependence, significantly higher than the 7.2 threshold, potentially able to press a significant threat of retaliation – between 15 and 24.9 percent export dependent.

• 1 The highest export dependence of the sector to the complainant’s market – 24.9 percent or more export dependent (the highest observed value).

Low Number of Veto Players

The measure considers the number of veto players in the EU government whose consent was necessary to bring domestic policy change in response to WTO litigation. A “low” amount of veto players refers to a policy change that required an executive action – e.g., commission regulation – and a high number of veto players refer to a policy that required a legislative action – e.g., Ordinary Legislative Proposal (OLP) of the EU parliament and of the Council. It is important to note that within the set of “high number of veto players” I consider legislations that required unanimity in the Council to be “very high” number of veto players.

• 0 No consent is required from any veto players of the EU – practically an empty classification.

• 0.2 Low number of veto players whose consent was necessary to enact policy change – i.e., EU Commission directives and regulations.

• 0.5 Maximum ambiguous point: Not able to decide if the number of veto players in the EU is “high” or “low” for the enacted legislation.

• 0.6 High number of veto players whose consent was necessary to bring policy change – i.e., Ordinary Legislative Proposal (OLP) with qualified majority voting (QMV) in the Council.

  *(I consider the EU to be composed of high number of veto players in the absence of any other information available about the policy that was modified or withdrawn)*

• 0.8–1 Very high number of veto players whose consent was necessary to bring policy change – i.e., unanimity in the Council as well as the consent of the parliament.

  *(The classification of “very high number of veto players” incorporates the highest number of veto players available in the EU – therefore policy change should be potentially the hardest. I thus consider disputes in this category to be fully in the set of “high number of veto players” and code 1 for their value in the set instead of 0.8)*