Modelling Kemp-Vanek Admissibility: The Effects of Free Trade Areas on Non-Members

Robert Waschik, Profesor Adjunto de la Facultad de Derecho y Administración, Escuela de Economía y Finanzas — Melburne (Bundoora)


(solamente en inglés)

Since its inception, Article XXIV of the GATT has allowed Preferential Trade Agreement (PTAs) as legitimate exceptions to Article I (MFN) as long as:

  • PTA members remove barriers on substantially all trade between PTA members over a reasonable period of time, and
  • duties set by PTA members on non-member trade are not higher or more restrictive than prior to the formation of the PTA

This implies that, by their very nature, PTAs must be discriminatory towards non-members, since even if trade barriers against non-members are not raised, the formation of a PTA will make imports from without the PTA more expensive relative to imports from within the PTA. 

In evaluating whether PTAs are welfare-improving or welfare reducing, theoretical and applied models typically use the notions of trade creation and trade diversion described in Viner (1950).  While many studies have noted that trade creation and trade diversion are not well suited to the task of evaluating the welfare consequences of a PTA, their application typically focuses on effects on members of the PTA.  For a multilateral institution like the WTO, evaluation of a PTA should focus on its effects on non-members. 

McMillan (1993) argues that Art.XXIV focuses on the wrong measures of the effects of PTAs.  By concentrating on the level of duties set by PTA members on non-member trade, Art.XXIV targets a variable which only indirectly affects welfare through its effect on trade volumes.  McMillan proposes a superior test for the admissibility of a PTA, by examining its effects on trade volumes with non-members: A PTA would be consistent with the GATT/WTO if it did not result in any reduction in trade between the PTA and non-members.  This rule has a strong theoretical foundation beginning with the work of Kemp (1964), Vanek (1965), and Kemp and Wan (1976).  Such a rule would ensure that a PTA would result in no reduction in non-member trade, implying that the PTA could not leave non-members worse off.  Any Kemp-Vanek admissible PTA could not be a stumbling block on the path to global free trade. 

To show how to evaluate whether a PTA is Kemp-Vanek admissible, we use a Numerical General Equilibrium (NGE) model of global production, consumption and trade to simulate the effects of a PTA which is currently being negotiated:  A Free Trade Agreement (FTA) between Australia and China.  Beginning with a dataset depicting a global general equilibrium for the year 2001, we simulate this FTA by removing all import tariffs on goods traded between Australia and China.  Under such an FTA, both Australia and China experience welfare gains, but all non-members of the FTA are worse off.  Results show that there are a number of sectors (textiles, wearing apparel, leather products in Australia and cereal grains, meat, dairy products, sugar in China) where imports into Australia and China from outside the FTA fall by more than 10%.  We conclude that such an Australia-China FTA would not be Kemp-Vanek admissible, and suggest that these results could be used to inform debate during Panel deliberations at the WTO and to identify those countries or regions which would be adversely affected by an Australia-China FTA. 

We then proceed to show how to construct an Australia-China FTA which would be Kemp-Vanek admissible, by introducing into our NGE model an endogenous tariff on each commodity imported by each FTA member from each non-member, and an endogenous export tax on each commodity exported by each FTA member to each non-member. 

When we simulate the FTA by removing all import tariffs on goods traded between Australia and China, these endogenous trade taxes adjust to keep trade between each FTA member and non-member at its pre-FTA level.  Any welfare changes or changes in trade volumes within the FTA will be due only to the liberalization of trade within the FTA, not at the expense of any losses in trade or welfare by non-members.  Results show that this version of the FTA between Australia and China is Kemp-Vanek admissible.  FTA-member trade volumes with all other regions are held constant, and non-member welfare is virtually unchanged as a result of the FTA.  Both Australia and China gain from this Kemp-Vanek admissible FTA, and China's increase in welfare is even larger than under the FTA with no endogenous taxes.  No non-member country or region is worse off, and global welfare is higher under this Kemp-Vanek admissible FTA than the non-Kemp-Vanek admissible version of the FTA which did not have endogenous trade taxes to keep non-member trade with the FTA from falling. 

If results from the initial FTA without endogenous trade taxes can be used to identify those regions with legitimate concerns over an Australia-China FTA, then results from this Kemp-Vanek admissible version of an Australia-China FTA with endogenous trade taxes can be used to suggest possible remedies, in the form of reductions in Australian and Chinese tariffs on imports from non-members, since this experiment identifies the endogenous tariff to which each FTA member must lower tariffs to hold constant imports from non-members.

This construction of Kemp-Vanek admissible PTAs is also directly relevant to issues related to the deepening and expansion of PTAs, such as the recent expansion of the EU.  Such expansions of PTAs have created concern about the scope for preference erosion to harm members of the original smaller PTA who see their benefits to PTA membership being diluted as new members join a PTA, or as fellow PTA members grant concessions to non-members.  Concern about preference erosion has already impacted trade negotiations at the WTO's Doha Development Round.  Our application of Kemp-Vanek admissible PTAs can help to illuminate the debate over the significance of preference erosion. 

Waschik, Robert, (2009), The effects of free trade areas on non-members: Modelling Kemp-Vanek admissibility, Journal of Policy Modeling, 31, issue 5, p. 648-663.



Robert Waschik is a Senior Lecturer in the School of Economics and Finance at La Trobe University, Melbourne Australia.  His research on Preferential Trade Agreements has used Computable General Equilibrium models to investigate how Art.XXIV can affect PTA members and non-members, for a given structure of a PTA and in environments allowing for the endogenous formation of PTAs.

> Guía para descargar ficheros