The disciplining of preferential trade agreements (PTAs) by the WTO has been ‘relaxed’ recently as a result of the new transparency mechanism within which notified PTAs are being multilaterally reviewed. This is probably a blessing for a number of reasons including the success of the multilateral trading system in bringing tariffs down over the years (and the ensuing reduced trade diversion), the fact that modern PTAs deal with many non-trade issues as well (for which no WTO disciplines exist), and recent empirical literature suggesting overall positive welfare implications for those participating in similar schemes. This paper discusses these and other reasons to support the view that the WTO should rather focus on the multilateral agenda instead of diverting its attention towards disciplining PTAs. Here are a few key points:
 
?nbsp;    From Day One We Applied the Wrong Test

The GATT-test for consistency of PTAs with the multilateral rules aims at ensuring that PTAs will not be ?la carte: absent the substantially all trade-requirement, PTAs could be formed on one tariff line only. This could severely undermine MFN, the cornerstone of the GATT-edifice. So the GATT framers could not live with a GATT ?la carte, but could live with GATT-consistent PTAs which resulted in trade diversion, the evil Viner in his classic analysis warned us against. Indeed, especially in the ‘50s and the ‘60s when MFN rates were high (and thus, the potential margin of preference large) PTAs that would take the intra-PTA tariff rates to 0% could create substantial trade diversion if the PTA partners were relatively inefficient (un-competitive). The GATT would applaud while Viner’s worst fears would have been confirmed. One might legitimately ask the question whether the candle is indeed worth the flame?  Should we, in other words, be enforcing an economics-uninformed test in the name of avoiding PTAs ?la carte? How realistic is this risk anyway?
 
?nbsp;    What is the Counterfactual to Trade Diversion?

The classic Vinerian analysis would request us to calculate the trade created through the establishment of a PTA (since intra-PTA trade would be liberalized) and compare it to the trade diverted (since trade might be deflected from the worldwide most efficient source of supply to the intra-PTA most efficient source of supply). One of course, might cast significant doubt on the appropriateness of such measurement, since it assumes the counterfactual. What if countries refused to make the same MFN cuts if they were deprived of the possibility to go preferential? What if they refused to participate in the WTO altogether? This is not meant to put into question the classic Vinerian analysis. Indeed, Viner was interested in measuring the allocational impact of discriminatory integration; the point here concerns Realpolitik and suggests that there is no reason to believe that MFN cuts would be the appropriate counterfactual to preferences.
 
?nbsp;    Is Trade Diversion an Issue as it Was?

Recent empirical studies provide us with mixed evidence regarding the extent of trade diversion resulting from the formation of PTAs. We lack a comprehensive calculation of trade diversion for all PTAs (indeed one might wonder if one is feasible), but the on-going tariff liberalization of tariffs at MFN-level would strongly argue in favour of the thesis that the problem is not of the magnitude that it used to be.
 
Scholarship points to the (missing) incentives to agree on MFN tariff cuts following establishment of a PTA. Bhagwati (2002), Krishna (1998) and Limão (2006) all have contributed in making the point that, besides trade diversion created through the establishment of PTAs, members of PTAs behave as enemies of non-discriminatory trade liberalization, since they are unwilling to cut tariffs on MFN basis for fear of eroding the margin of preference that they have granted to their PTA-partners. They become thus, stumbling (as opposed to building) blocs opposing MFN trade liberalization, and frustrating the achievement of the basic WTO objective. The fear was probably legitimate at some point, but the question is how relevant is it today?
 
On the one hand studies like Karacaovali and Limão (2008) looking at the EU, and Limão (2006) looking at the US have provided empirical evidence that PTAs have behaved like stumbling blocs.  They ask the question whether MFN tariff cuts during the Uruguay Round are related to their preferential tariffs. The stumbling bloc-thesis would suggest that trading nations would have cut tariffs less in areas where they had preferential tariffs, and indeed this what these authors finds. Other studies however find the opposite. Estevadeordal, Freund and Ornelas (2008) examine the Latin experience with PTAs and find that Latin nations cut their MFN rates most in products where they had preferences in place. Baldwin and Seghezza (2010) use tariff data for 23 large trading nations and find that MFN cuts and preferences are complements not substitutes: preferences tend be zero or close to zero where nations have high MFN tariffs. Intuitively, one would associate the stumbling bloc thesis with large preferences in similar cases, but the authors show that this is not the case. The authors discard thus, the stumbling bloc- without supporting the building bloc thesis. Acharya et al. (2011) in similar vein, find that the impact of plurilateral PTAs on extra-PTA imports and exports is large and positive. If at all, recent empirical evidence hardly supports the uni-dimensional conclusion that PTAs are stumbling blocs per se.
 
Of course trade diversion can result from instruments other than tariffs. It can result from say convergent environmental or public health policies across PTA partners. With respect to domestic instruments in general, nonetheless, there is no need for action: to the extent that a trade advantage has been conferred, it must be extended to all WTO Members automatically and unconditionally by virtue of Art. I GATT (MFN). PTAs in other words, cannot provide legal shelter for discriminatory domestic instruments since the latter were not meant to protect anyway, and hence, cannot be regarded as a restrictive regulation of commerce in the sense of the term embedded in Art. XXIV GATT.
 
Trade diversion can also result from say increased use of antidumping (AD) proceedings against non PTA partners, as the work of Prusa and Teh (2010) shows. Once again though, nothing much can be done about it: at a positive level, the only MFN obligation that WTO Members incur with respect to AD duties is to collect them on non-discriminatory basis; at a normative level, the burden associated with proving that under similar circumstances PTA partners privileged AD proceedings against a sub-set of the WTO Membership (namely, outsiders to their PTA) is quite high: except for conceptual issues, those carrying the burden of proof (that is, the Members asked to pay them) will have to also address issues such as opportunity cost of conducting another investigation, scarcity of administrative resources etc.

Concluding remarks

PTAs are formed for many, often idiosyncratic reasons. We cited some of the reasons and there are many more. Baldwin (1997), for example, tries to explore the validity of some of the rationales, and, more recently, Whalley (2008) attempts a similar endeavour. Some of the rationales advanced have even (persuasively) criticized. For example, the argument has time and again been advanced (and continues to do so) that countries have gone preferential because they were frustrated with the slow pace of multilateral tariff liberalization. I attach little value to this view. If true, then why did not they go for what Bhagwati has termed open regionalism, that is allow others to join their PTA assuming they had agreed on the tariff cuts decided? Other rationales hold more promise. Baldwin (2008), for example, develops a theory aiming to predict who goes preferential depending on the identity of the spoke and the hub that have already gone preferential. But we lack a dominant explanation that can serve as rationale across PTAs. This observation in and of itself would cast severe doubt on remedial action against PTAs since it is questionable whether the same remedy should apply to divergent situations.
 
The historic rationale for PTAs is of no much help either. Arguably, one reason for its inclusion is that the GATT negotiators were presented with a fait accompli. Two CU participated in the negotiation, the Syro-Lebanese customs union (Syria, Lebanon), and Benelux (Belgium, Netherlands, Luxembourg). Institutional arrangements probably had to be made in order to accommodate these contracting parties. Chase (2006) drawing from a series of archival records, begs to differ and points to a different direction: the author persuasively demonstrates that it was the US negotiators that designed this provision in order to accommodate a trade agreement that they had secretly reached with Canada. The US – Canada FTA did not see the light of the day then but only 40 years later. On the other hand, the view held by many that the inclusion of a provision on FTAs was there to accommodate the European integration process must be discarded. In Acheson’s (1969) record, Jean Monnet revealed his plans on European integration after the Havana Conference had taken place.
 
We are still struggling with the rationale but recent research paints a much rosier picture for PTAs than what was the case before. One contributing factor is the success of the multilateral trading system. MFN reduction of tariffs results in reduction of trade diversion created through PTAs. So we are now facing a problem less acute than before. Moreover, the content of PTAs has changed drastically over the recent years and moved to areas escaping the current WTO mandate. Finally, empirical evidence shows that PTAs can be welfare improving. And while all these changes were happening, the WTO continued to enforce an ill-informed and out-dated to constrain PTAs. Against this background, the shift towards a mere exercise in transparency (facilitated by the WTO Transparency Mechanism) should be welcome with relief. If at all, it removes the risk for false positives which can have important institutional (negative) external effects.