Determinants of Withdrawals of Anti‐Dumping Complaints in the EU

Abstract

The study proposes a model explaining what determines the emergence (or re‐emergence) of collusion between complainants and defendants during anti‐dumping (AD) investigations. Based on theoretical and empirical evidence, we assume that collusion results in withdrawals of complaints and is thus observable. The probability of collusion can be explained by the variables from four areas: domestic political economy, international strategic trade policy, international industry‐level bargaining, and industry and product characteristics. The model is verified with probit regressions for the EU AD cases, having good explanatory power.

Key Words:

• Anti‐dumping
• withdrawals
• collusion
• EU

Jel Codes:

• D43
• D72
• F13

Acknowledgments

The author is grateful to Jørgen Ulff‐Møller Nielsen for his valuable comments.

Notes

1. Department of Commerce and International Trade Commission.

2. Bureaucrats may get ‘captured’ by the industries which they regulate because they are assessed on the basis of the performance of these industries, e.g. import‐ competing industries because of trade balance considerations. Bureaucrats do not receive rents from protection so they do not maximise these rents optimising the level of protection but rather maximise protection. Finally, bureaucrats are naturally those who operationalise the policy (whereas politicians just set the goals); thus, bureaucrats can create complex and non‐transparent policy instruments which serve their own goals best.

3. This can be derived from the general motives for entering into collusion (see Lipczynski and Wilson, 2001, p. 62)

4. There is a case study supporting this assumption. Messerlin (1990) found a number of ‘twin’ AD and antitrust cases investigated by the European Commission between 1980 and 1987 concerning cartels of chemical enterprises which were first collaborating in antidumping cases. A similar phenomenon could be noticed in the US (Pierce, 2000).

5. The NAFTA, the Uruguay Round GATT negotiations, and the issue of MFN for China, all in 1993 and 1994.

6. Protection did not increase with output but geographical concentration, unionization rates, and the number of firms did affect protection levels. In general, the study showed that politicians placed ca. 15% more weight on campaign contributions than on consumer surplus.

7. As Zanardi (2004b) showed, the role of new, ‘non‐traditional’ users has increased rapidly, measured with the numbers of initiated investigations in the recent years.

8. In general, both economic analysts (such as e.g. the Economist Intelligence Unit in its Country Monitor) and scholars in economics use most often GDP as the way to measure foreign market size. For the policy‐makers, the risk of unemployment augmentation – due to decline in exports and, thus, in aggregate demand – may be the ultimate motivation.

9. ‘Relative’ means that GDP per capita of an exporting country was divided by the EU GDP per capita and unit value of exports was divided by the unit value calculated for the intra‐EU trade for a product to which each AD case referred.

10. NACE Rev. 1, abbreviation stemming from the French label ‘Nomenclature statistique des Activités économiques dans la Communauté Européenne’

11. ‘M’ added as a prefix to the name of the variable indicates that the median ratio was taken, and ‘A’ indicates the mean.

12. ‘T’ prefix before the name of the variable indicates that the total was employed.

13. Note that ‘high’ levels of pseudo‐R‐squared are much lower than ‘high’ levels of R‐squared (Whitehead).