3. Explaining Patterns of WTO Member Behavior at the WTO Dispute Settlement

Relevant data are missing for a subset of WTO defendants, including Thailand, Egypt, Guatemala, Argentina, Colombia, Chile, the Dominican Republic, Ukraine, and the Philippines. Collectively, these defendants account for over twenty cases, which are omitted from model due to lack of data on trade in intermediates.

I start the clock to calculate “time until compliance” from the AB report because the extent of a violation, and therefore the extent of a policy change that is required to bring compliance is only clarified for the defendants after a final ruling is delivered by the AB. As a robustness check, however, I run separate analyses in my supplementary appendix in which I calculate the “time until compliance” from the original panels’ report. The results are substantively similar.

Four additional details warrant attention. First, in cases where there were multiple complainants in a dispute, I distinguish each pair and count them as separate disputes. This decision is consistent with a number of works in the field (e.g., Horn et al. 1999; Bagwell et al. 2004) and was made so that I can observe the impact of additional WTO members in a dispute. Second, my sample does not include disputes that targeted horizontal measures, since these disputes did not target specific sectors. Although some horizontal measures may involve certain sectors more than others, I only examine responses to litigation if there is a sector whose import dependence can be measured. Third, a few disputes resulted in compliance around the same time there was a panel ruling. Even though the defendants in these disputes did not comply with a panel ruling per se, they demonstrated domestic policy change after a panel composition, in which the clarity of violations Ire most likely made before the circulation of panel reports. Finally, in certain disputes, defendants notified the WTO DSB of compliance, but I could not find the corresponding domestic legislation. These cases are coded “complied with” and I rely on WTO Members’ notification to the DSB and official WTO records to consider the extent of compliance.

The dataset is available at: https://​www.​oecd.​org/​sti/​ind/​measuring-trade-in-value-added.​htm. Accessed on 20 June 2020.

For the TiVA dataset, see http://​stats.​oecd.​org/​Index.​aspx?​DataSetCode=​TIVA2015_​C1#. Note that in cases where import data from TiVA was not available, UN COMTRADE database was used.

The data on FDI stocks were collected from the OECD where available and from EUROSTAT for the EU. For data on developing WTO members, I utilized UNCTAD. The closest year available was used if data were missing for certain years.

See OECD (1993), for an overview on HHI index and sector concentration.

Sector employment is used as a proxy for political importance by Hoffman and Kim, “The Political Economy of Compliance in WTO Disputes,” Unpublished Manuscript, 2009. Spilker (2012) and Sattler et al. (2014) consider agricultural sectors to be politically important.

As mentioned in the introduction, the analysis here draws upon my previous work. It is particularly influenced by my doctoral dissertation and a co-authored article I published by the Review of International Organizations titled ‘Internationalization of production and Compliance in WTO disputes’.

Results remain substantively similar when omitting outliers.

Replicatio files can be accessed through my website, www.​aydinbyildirim.​com.