1
Anna Jermak (12720771)
Preferential Rules of Origin – Requisite Protection
or Constraint on International Trade?
2
Abstract The paper discusses the real nature of preferential rules of origin and their effect on trade within Preferential Trade Areas as well as between members of Preferential Trade Agreements and third countries. It evaluates the restrictiveness of a framework within which preferential rules of origin operate, considering regime-wide rules of origin that introduce flexibility into that framework and the characteristics of preferential rules of origin in three groups of Preferential Trade Agreements. Arguably, due to the protectionist purposes for which preferential rules of origin are used, their stringency often greatly exceeds a minimum level necessary to prevent trade deflection. In effect, the utilisation rate of preferences is significantly lower than it would be had lenient rules of origin, or no rules of origin at all, been applied. Hidden protectionism is also the reason why desperately needed changes to the preferential rules of origin system, bringing in balance and simplicity, have been implemented slowly and to a limited scope.
Table of Contents
1. Introduction ... 3
2. Trade-distorting impact of preferential rules of origin ... 5
2.1. Preferential rules of origin as a trade barrier ... 5
2.1.1. Preferential RoO as a trade barrier between PTAs partners ... 5
2.1.2. Preferential RoO as a trade barrier between PTAs partners and third nations ... 9
2.2. Preferential rules of origin’s potential derogation from Art. XXIV GATT ... 11
3. Mitigating measures – regime-wide rules of origin allowing for the relaxation of the restrictiveness of product-specific rules of origin ... 12
3.1. Cumulation ... 13
3.2. Other measures... 16
4. The practical side of preferential rules of origin – PTA’s analysis ... 19
4.1. The EU PTAs ... 19
4.2. NAFTA ... 22
4.3. PTAs between developing countries ... 25
5. Proposed changes to the preferential rules of origin framework ... 27
6. Conclusion ... 31
3
1.
Introduction
Any discussion relating to Preferential Trade Agreements (PTAs) – reducing tariffs for certain goods to participating countries – cannot go without addressing the issue of the rules of origin (RoO). They aim to restrain the so-called 'free-riders' by preventing trade deflection, that is, the transshipment of goods by a non-qualifying country to the final PTA-member through a lower-tariff PTA-member. Preferential RoO provide, then, for conditions under which a product is deemed 'originating' in a country eligible for favourable treatment by the importing country. They are continuously present in virtually all PTAs around the globe, constituting an essential element of Free Trade Agreements (FTAs), where each member keeps its own external tariff, and Generalised Systems of Preferences (GSPs) schemes through which developed countries accord tariff preferences to developing countries on a non-reciprocal basis. Preferential RoO play, however, a much weaker role in fully implemented Customs Unions characterised by a common external tariff, where trade deflection is not an issue.
Despite their seemingly straightforward nature, preferential RoO entail various legal issues and dilemmas. First of all, neither the General Agreement on Tariffs and Trade (GATT) nor any other WTO Agreement specifies qualities preferential RoO ought to demonstrate in terms of their trade-restrictiveness. While the WTO Agreement on Rules of Origin requires that non-preferential RoO do not create “restrictive, distorting, or disruptive effects on international trade” and should not be used directly or indirectly as “an instrument to pursue trade policy objectives”,1 such requirements are non-existent in Annex II to the Agreement addressing preferential RoO.2 The Panel’s decision in the US-Textiles case – not appealed by India despite receiving harsh criticism – specifies, however, that assertions of non-preferential RoO having “restrictive” effects on international trade must be backed up by solid evidence and be proven to adversely affect more than just one WTO member’s trading.3 Still, irrespective of the efficacy of the law governing non-preferential RoO, a significant amount of freedom is given to PTAs members to determine the level of stringency of preferential RoO necessary – in their minds – to deter potential trade frauds.
1 WTO Agreement on Rules of Origin, art 2(c) and 9(d).
2 ibid Annex II: Common Declaration with Regard to Preferential Rules of Origin.
3 WTO, United States: Rules of Origin for Textiles and Apparel Products – Report of the Panel (20 June 2003) WT/DS243/R [6.158-6.160].
4
This brings us to another vital issue – preferential RoO may act against their own objective and be used for purposes other than mere prevention of trade deflection. Indeed, they are increasingly used as trade policy instruments prevailing over preferential trade.4 Due to their complexity and poor transparency, preferential RoO has become a serious non-tariff barrier to trade, leading not only to high administrative and operational costs but also to underutilisation of PTAs themselves. As the European Commission noticed back in 2003, the existing framework for determining, managing, and verifying preferential RoO is simply not well-adapted to the international economy.5
How restrictive is then restrictive enough for preferential RoO? Is there a line at all? Are the current form and complexity of RoO truly necessary to prevent trade transshipment, their trade-distorting impact being merely the side-effect? This paper argues that although a certain level of stringency of preferential RoO is indispensable, the protectionist purposes they are often used for go beyond their alleged objective and, indeed, amount to a constraint both on trade between PTAs partners and between PTAs members and third nations. At the same time, however, achieving preferential RoO restrictive enough to prevent trade deflection but lenient enough not to generate any negative trade effects in third countries appears impossible due to the PTA framework focused on promoting intra-PTA trade they operate in. That puts into question the rigid division between preferential RoO being either necessary or negatively impacting trade. It seems they are both, albeit the amount of harm they cause (and to whom they cause it) does depend on the level of their stringency. The aim of PTAs should, therefore, be to implement non-protectionist RoO preserving a balance between trade creation on the one hand and the prevention of trade deflection, on the other.
The structure of the paper is as follows: section 2 will present arguments demonstrating trade-distorting effects of preferential RoO, both on trade between PTAs partners and between a PTA-member and third nations. It will also examine preferential RoO's compliance with Art. XXIV GATT. Section 3 will focus on measures allowing for the relaxation of the restrictiveness of preferential RoO and their effectiveness. Section 4 will approach the paper's topic from a more practical perspective, evaluating three models of PTAs: EU, NAFTA, and
4 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin: A World Map and Trade Effects' (Paper prepared for the Seventh Annual Conference on Global Economic Analysis: Trade, Poverty, and the Environment, the World Bank 2004) 2.
5 Commission Green Paper COM (2003) 787 final of 18 December 2003 on the future of rules oforigin in Preferential Trade Agreements [2003].
5
one between developing countries. Section 5 will elaborate on changes that have been suggested to improve the preferential RoO framework. Finally, section 6 will conclude.
2. Trade-distorting impact of preferential RoO
Any considerations of the nature of PTAs involve a question of whether they act as stepping stones or rather as stumbling blocks for the achievement of multilateral trade liberalisation. Though such deliberations are not per se a subject of the current paper, the operation of preferential RoO is relevant to how the above question is answered, especially in light of the PTAs' constant proliferation. The current section will explore the hindering impact of preferential RoO on the liberalisation of trade between PTAs members, and between PTAs partners and third countries, by following Estevadeordal's and Suominen's division between production and administrative costs RoO impose.6 In sub-section 2.2., the compliance of preferential RoO with Art. XXIV GATT will be briefly looked into.
2.1. Preferential RoO as a trade barrier
2.1.1. Preferential RoO as a trade barrier between PTAs partners
The production costs of preferential RoO emerge from numerous technical criteria inflicted by the RoO regime and relate to methods used for determining the substantial transformation. The requirement of the substantial transformation must be satisfied in situations when two or more countries contributed to the production-process of a certain product. It seeks to determine in which country a particular good has been substantially transformed and grants the origin of that country in line with each PTA's RoO. There exist three main techniques for establishing the origin – change of tariff classification, value-added, and specific manufacturing process, each entailing potential trade difficulties by often being burdensome to comply with.7
6 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in FTAs in Europe and in the Americas: Issues and Implications for the EU-MERCOSUR Inter-Regional Association Agreement' (Chaire Mercosur de Sciences Po 2004) 6.
7 Bernard Hoekman, Stefano Inama, 'Harmonization of Rules of Origin: An Agenda for Plurilateral Cooperation?' (2018) 22 East Asian Economic Review 3.
6
The change of tariff classification (CTC) method confers origin when imported inputs within a product have been transformed to such an extent that the exported product's tariff classification under the Harmonised System (HS) has changed. The higher the level of classification at which the change is required, the more difficult it is to confer the origin. Despite its ostensible simplicity, the CTC technique has a major flaw – it is based on the HS which was not designed with the objective of granting origin in mind.8 Consequently, certain goods that may have been substantially transformed without involving the modification of the tariff classification would nevertheless fail to comply with the CTC requirement. Sectors where the intermediate and final products share the same heading, such as the machinery sector, have been particularly disadvantaged by such rigidity.9
Additionally, multiple available HS codes of importing countries for a single product place a burden on manufacturers to find the correct one, further discouraging them from employing their preferences.10 Exceptions defining situations where a change of classification will not grant origin, frequently leading to the elimination of regional sources of competition in PTAs, constitute another restriction of the CTC method.11 Under NAFTA, for instance, the transformation from any chapter to tomato ketchup in chapter 21 confers origin, except for transformation from tomato paste in chapter 20.12
The value-added method defines the minimum percentage of the value of a product that must be added in the exporting country or, alternatively, the maximum percentage of imported inputs in the value of a product. Unfortunately for producers, this method's clarity of specification is balanced out by its uncertainty and complexity of application. As changes in exchange rates, production costs, and raw materials prices are unavoidable, they largely affect producers' observance of even seemingly simple RoO based on the value-added.13
Moreover, the fact they are inconsistently applied across different RoO regimes constitutes another burden on firms seeking to benefit from multiple PTAs.
8 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in FTAs in Europe and in the Americas…' (n 6) 6. 9 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in the World Trading System' (Paper prepared for the
seminar on Regional Trade Agreements & the WTO, Centre William Rappard WTO 2003) 32. 10 Bernard Hoekman, Stefano Inama, 'Harmonization of Rules of Origin…' (n 7) 5.
11 James Mathis, Regional Trade Agreements in the GATT/WTO: Article XXIV and the Internal Trade Requirement (TMC Asser Press 2001) 111.
12 David Palmeter, The WTO as a Legal System: Essays on International Trade Law and Policy (Cameron May 2003) 149-150.
13 WTO Secretariat, 'Rules of Origin Regimes in Regional Trade Agreements' (WT/REG/W/45, WTO Committee on Regional Trade Agreements 2002) 5.
7
Verification operations requiring advanced accounting systems impose additional financial as well as trade burdens on firms, further limiting the utilization rate of respective PTAs.14 Furthermore, the value-added rule gives dangerously great leeway to the customs of the importing country to subjective compliance evaluations.15
The specific manufacturing process (SMP) method defines manufacturing or processing operations relating to a product that are sufficient or insufficient to confer origin. Albeit, that it provides for much certainty, it is deemed susceptible to obsolescence and thus in need of recurrent modification.16 Documentary requirements of the SMP test, such as a duty to keep an inventory of production processes, can also be problematic, costly, and onerous to satisfy, constituting a barrier to trade for PTAs producers.17 Ironically, the producers are at the same time given an opportunity to abuse the SMP RoO and serve their own interests, as they themselves supply information about their products' manufacturing processes.
The administrative costs of preferential RoO arise from the procedures necessary to determine compliance with RoO. They relate to the costs exporters must incur to certify the origin of a product and the costs of the origin verification borne by the customs services in the importing country. When it comes to obtaining the certificate of origin, some certification mechanisms impose high costs on companies, additionally to expenses relating to potential delays in the certificate's acquirement and expensive accounting and inventory system requirements.18 Such bureaucratic and financial encumbrances discourage exporters and producers from pursuing preferential treatment under PTAs simply because once the costs of certification exceed the value of traded goods, it is no longer profitable. Customs authorities, on the other hand, are burdened by complex RoO in PTAs, which have particularly trade-restrictive effect in developing countries with weak resources and ineffective administration systems.19
In the long run, the above-mentioned production and administrative costs are often so great
14 ibid.
15 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in the World Trading System' (n 9).
16 Paul Brenton, 'Preferential Rules of Origin' in Jean-Pierre Chauffour and Jean-Christophe Maur (eds), Preferential Trade Agreement Policies for Development (the World Bank 2011) 164.
17 WTO Secretariat, 'Rules of Origin Regimes in Regional Trade Agreements' (n 13).
18 Paul Brenton, Miriam Manchin, 'Making EU Trade Agreements Work: The Role of Rules of Origin' (2003) 26 The World Economy 755.
8
that they neutralise preferential access under PTAs. Indeed, it has been estimated that for 149 countries in a period between 1999 and 2001, the compliance costs of preferential RoO diminished the trade-creation effects of PTAs by around two-thirds.20 Furthermore, it has been widely contended that as a consequence of escalation of production and administrative costs of restrictive RoO, trade in final goods between PTAs members has been reduced to much lower levels than it would have with loose RoO or no RoO at all.21 The reason for that has been the 'encouragement' provided by harsh RoO to final goods producers to import their inputs from third countries for a lower price and sell their final products at their home market, rather than to purchase inputs from their PTA partners at a higher price and sell their outputs there.22 The trade relationship between Turkey and the EU accurately exemplifies the harmful impact of excessively rigorous preferential RoO on trade levels. The World Bank has demonstrated that since these two partners are in a CU (negating the necessity for RoO on bilateral trade), they enjoy greater export volumes than they would have under an FTA, particularly an FTA with restrictive RoO.23
It can be consequently deduced that the more stringent preferential RoO are, the more severe impact they have on the intra-PTAs trade volume. Lenient RoO generally support commerce between PTA partners by making intra-PTA trade more profitable for them than trading with third countries. Stringent RoO, by contrast, tend to induce PTAs members to have commerce with their non-PTA partners, by raising preferential RoO’s fulfilment costs to unacceptably high levels. Therefore, at the point when compliance with practically complex and/or inflexible substantial transformation rules and the costs of acquiring the certificate of origin become more expensive for PTAs producers to satisfy than a mere payment of the MFN tariff, the producers lose any incentive to utilise their preferential treatment. The line where the attractiveness of preferential tariffs becomes insufficient to compensate for the costs of RoO is marked by preferential margins (the difference between the MFN tariff and the preferential tariff) – the higher the margins of preference, the more stringent RoO are
20 Jose Anson, Olivier Cadot, Jaime de Melo, Antoni Estevadeordal, Akiko Suwa-Eisenmann, Bolormaa Tumurchudur, 'Rules of Origin in North-South Preferential Trading Arrangements with an Application to NAFTA' (2005) 13 Review of International Economics 501.
21 Jose Anson, Olivier Cadot, Jaime de Melo, Antoni Estevadeordal, Akiko Suwa-Eisenmann, Bolormaa Tumurchudur, 'Assessing the Costs of Rules of Origin in North-South PTAs with an Application to NAFTA' (Discussion Paper 2476, CEPR 2003); Paul Brenton, 'Preferential Rules of Origin' (n 16) 174; James Mathis, Regional Trade Agreements in the GATT/WTO… (n 11) 110.
22 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in FTAs in Europe and in the Americas…' (n 6) 6. 23 World Bank, ‘Evaluation of the EU-Turkey Customs Union’ (Report No. 85830-TR, World Bank 2014)
9 necessary to neutralise them.
Excessively strict preferential RoO commonly present in PTAs are not accidental; their real aim has been to safeguard PTAs members' importing markets from competing exporters from their PTA partners.24 In fact, the same protectionist interests that push tariffs themselves, rather than the goal to prevent the transshipment of goods, are believed to drive them.25 Such a protectionist practice has been not only negating the alleged “boost to trade” objective of PTAs but has been also posing a danger to its entire regime by making the idea of PTAs practically redundant. Assuredly, the goal behind the creation of a PTA is at least questionable if overly stringent RoO adopted within it de facto deny preferential treatment to its members.
2.1.2. Preferential RoO as a trade barrier between PTAs members and third nations
It is not disputed that preferential RoO contribute to trade diversion also by hampering commerce between PTAs partners and the rest of the world (ROW), especially in intermediate goods.26 Indeed, RoO demanding unreasonably high amounts of regionally sourced inputs as a condition for manufacturers to benefit from preferential treatment have the effect of inducing PTA-based firms to source their input domestically or from their PTA partners, as opposed to third-countries sourcing, despite the cost reductions entailed by the latter. By means of such hidden protectionism, PTA-based input producers introduce their own intermediate goods to input markets of its PTA partners, which results in less efficient production and higher prices of final goods.27 The use of non-local inputs by PTAs final goods producers becomes ipso facto more and more problematic with increasing severity of RoO, leading to less PTA-ROW commerce and more intra-PTA trade, in comparison to when
24 Wayne Crook, Jenny Gordon, 'Rules of Origin: Can the Noodle Bowl of Trade Agreements Be Untangled?' (Productivity Commission Staff Research Note, Australian Government 2017) 4.
25 Dillon Laaker, 'Preferential Rules of Origin: Protection or Deflection?' (2019) < https://www.peio.me/wp-content/uploads/2020/01/PEIO13_paper_98.pdf> accessed 14 July 2020.
26 Kerry A. Chase, 'Protecting Free Trade: The Political Economy of Rules of Origin' (2008) 62 CUP < https://www.cambridge.org/core/journals/international-organization/article/protecting-free-trade-the-political-economy-of-rules-of-origin/30C6AFA9D1649ECF8B07CD5769D591AD> accessed 14 July 2020; Antoni Estevadeordal, Kati Suominen, 'Rules of Origin: A World Map and Trade Effects' (n 4) 37;
Kala Krishna, Anne Krueger, 'Implementing Free Trade Areas: Rules of Origin and Hidden Protection' (Working Paper No. 4983, NBER Working Paper Series 1995) 3.
27 Patricia Augier, Michael Gasiorek, Charles Lai Tong, Philippe Martin, Andrea Prat, 'The Impact of Rules of Origin on Trade Flows' (2005) 20 Economic Policy 569.
10 more relaxed or non-existent RoO were applied.28
Preferential RoO can, thus, either discourage intra-PTA trade for the benefit of third countries or – quite the opposite – positively influence the volume of trade within the PTA, with detriment for the ROW. What is crucial to fully grasp different ways in which RoO affect international commerce is the analysis of their various trade-restrictiveness levels that can be seen as a scale. As Chase explains, one end of the scale represents permissive RoO enabling the transshipment of goods; the middle point specifies RoO restrictive just enough to avert trade deflection; the other end, in turn, is where extremely rigorous RoO are used.29 Following Chase's logic, RoO so restrictive that PTAs members find compliance with them too expensive to even bother to make use of them are the equivalent of the second end of the scale, limiting intra-PTA trade. However, RoO less restrictive, yet still sufficient to prevent trade deflection (around the middle point on the scale) encourage PTA-based firms to source their non-originating inputs from inside the PTA and not third countries, increasing trade between the PTA members at the expense of the ROW. It appears that either way, trade somewhere is restricted and the protectionist interests of either individual members of a PTA or the PTA as a whole are served.
The most desired scenario – preferential RoO severe enough to fulfil their main objective of preventing trade deflection and, simultaneously, mild enough to create intra-PTA trade without undermining commerce with third countries – would involve the exact middle of the scale of RoO restrictiveness. Such precisely measured level of stringency would be possible – although not easy – to achieve in practice, as it would exclude the possibility of using RoO as a market access instrument. Interestingly, in case of a successful attempt to create a perfectly balanced and exactly as restrictive as necessary system of preferential RoO, only the problem of excessively rigorous RoO, but not that of trade diversion from the ROW to inside the PTA, would be solved. The reason for that would simply be the continuous incentive for PTA members to trade with each other, rather than with the ROW. RoO give, therefore, the impression of being incapable of performing their main function – averting the transshipment of goods – without involving negative trade implications.
28 Antoni Estevadeordal, Kati Suominen, Jeremy T. Harris, Matthew Shearer, 'Bridging Regional Trade Agreements in the Americas' (Special Report on Integration and Trade, IDB 2009) 31; Antoni Estevadeordal, Kati Suominen, 'Rules of Origin: A World Map and Trade Effects' (n 4) 38.
11
Having said that, the fact that preferential RoO function within the PTA framework, which Chase seems to have disregarded, cannot be omitted from our considerations. The main characteristic of PTAs is favourable treatment given to their participants, and their main objective – the stimulation of trade between their members. Preferential RoO are merely a tool to ensure that such preferential treatment is not abused by third parties, thus their role should not be exaggerated. Unless preferential RoO are overly stringent and go towards the prohibitive RoO on the scale (what admittedly happens quite often), the level of their leniency, or even their existence, does not affect the fact that PTAs members are provided with a strong incentive to have commerce between themselves, as opposed to with the ROW. After all, the prospect of intra-PTA trade growth is why PTAs are created in the first place. Accordingly, preferential RoO from the middle point on Chase's scale to its first end cannot be held responsible for the trade-distorting impact caused by preferences offered by PTAs. Only RoO having a higher level of stringency than necessary to prevent trade deflection (RoO exceeding the middle point and going towards the second end of the scale) can be blamed for detrimental trade effects they entail. The more restrictive – and thus the more protectionist – preferential RoO are, the more acute their impact becomes.
2.2. Preferential RoO's potential derogation from Art. XXIV GATT
Art. XXIV GATT allows for the formation of customs unions and FTAs between WTO members.30 On the face of it, neither Art. XXIV:8(b) – specifying that in FTAs duties and “other regulations of commerce” must be eliminated on substantially all trade between its members in originating products – nor Art. XXIV:5(b), pursuant to which duties and “other regulations of commerce” in FTAs towards third countries cannot be raised from the pre-FTA levels, dictates a specific regime concerning RoO.31 Such a hypothesis appears to be supported by the Appellate Body's report in the Turkey-Textiles case (the only WTO case directly addressing Art. XXIV GATT), where the requirements for a PTA desiring to benefit from Art. XXIV did not cover RoO.32
30 General Agreement on Tariffs and Trade, art XXIV. 31 ibid art XXIV:8(b) and XXIV:5(b).
32 WTO, Turkey: Restrictions on Imports of Textile and Clothing Products – Report of the Appellate Body(22 October 1999) WT/DS34/AB/R [58].
12
Yet, fear that from a legal perspective preferential RoO contravene Art. XXIV has been constantly present among scholars.33 Arguably, “other regulations of commerce” mentioned both in paragraph 5 and 8 are an “evolving concept”34 and could be interpreted as to include RoO.35 In such a case, preferential RoO restricting trade between PTAs partners would be in breach of paragraph 8 (unless they covered such a minimum amount of trade that
substantially all trade would not be affected). Similarly, in line with the Panel’s non-appealed
findings in Turkey-Textiles,36 preferential RoO resulting in increased trade barriers towards the ROW from before the creation of a particular PTA would violate Art. XXIV:5. In the event of a wide acceptance of such observations, preferential RoO distorting international trade would be in breach of the GATT, and even a stronger case for the reformation of the existing preferential RoO regime would be made.
3. Mitigating measures – regime-wide RoO allowing for the
relaxation of the restrictiveness of product-specific RoO
As it has been demonstrated in the previous section, preferential RoO can negatively impact both intra-PTA and PTA-ROW trade, while being part of the PTA framework in itself generating trade diversion. It depends on their stringency whether one’s PTAs partners or rather third countries are prejudiced. Regime-wide RoO provide for a certain degree of flexibility in the application of product-specific RoO and help counteract their trade-restrictiveness. The following section will examine such measures and will reflect on their effectiveness and practical implications.
33 James Mathis, Regional Trade Agreements in the GATT/WTO… (n 11) 112, 118; Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in the World Trading System' (n 9) 9.
34 WTO, Turkey: Restrictions on Imports of Textile and Clothing Products – Report of the Panel (31 May 1999) WT/DS34/R [9.120].
35 WTO, 'Part III. Article XXIV: Territorial Application – Frontier Traffic – Customs Unions and Free Trade Areas' (Analytical Index of the GATT 1994) 803
<https://www.wto.org/english/res_e/publications_e/ai17_e/gatt1994_art24_gatt47.pdf> accessed 14 July 2020.
36 WTO, Turkey: Restrictions on Imports of Textile and Clothing Products – Report of the Panel (n 38) [9.122-9.123].
13
3.1. Cumulation
Thanks to cumulation, PTA-based producers are able to use non-originating materials from other PTA members or certain regional groups without compromising the originating status of the final product. Cumulation is thought of as an efficient means of preserving the main function of preferential RoO – the prevention of tariff fraud – while bringing down their costs.37 The statistics confirm such a standpoint, indicating cumulation's positive impact on overlapping RoO and indicating that international trade would be 25% to 70% lower without it.38 There exist three types of cumulation, each operating on a different scale and having different nature: bilateral, diagonal, and full cumulation.
Bilateral cumulation allows each of the two PTA partners to use materials originating in the other partner's territory as though they originated in their own, for the purposes of acquiring preferential treatment. Thus, in an FTA between country A and country B, country A (and vice versa) can count components originating in country B as originating in its own country and freely use them in its products.
Despite being the most common type of cumulation, promoting PTA-originating inputs and applied in all FTAs,39 bilateral cumulation has been contributing to the neutralisation of the trade-hampering impact of RoO only to a limited extent. Since it is only applicable between two PTA partners, it has a short reach and curtailed effectiveness. A PTA member will only find bilateral cumulation advantageous if the use of its PTA partner's materials involves profit for it. Being given a choice between using the high-cost input from a PTA partner entailing bilateral cumulation, and using the lowest-cost input from elsewhere, a producer is going to choose the latter, unless harsh RoO furnish him with an “incentive” to source locally or even to forgo his preferences. Provided that the most profitable option for a manufacturer is to use inputs from the ROW and that RoO enable him to do so, the level of stringency of RoO will stay undiminished by bilateral cumulation, the latter being rendered meaningless. Such a scenario pertains to preferential RoO located closer to the middle point of the scale of RoO restrictiveness, rather than to its second end – RoO relatively lenient but still excessively
37 Patricia Augier et al., 'The Impact of Rules of Origin on Trade Flows' (n 27) 601. 38 ibid.
39 Maria Donner Abreu, 'Preferential Rules of Origin in Regional Trade Agreements' (Staff Working Paper, WTO 2013) 30.
14
rigorous for satisfying their main objective. Bilateral cumulation is thus of use only when rather strict preferential RoO – but naturally not so strict that PTA producers are forced to export on the MFN tariff – are applied.
Diagonal cumulation involves at least three trading partners connected by PTAs with the same set of RoO. It permits one country to use materials originating in other countries in the area without giving up the originating status of the final good. Accordingly, if countries X, Y and Z are tied by FTAs with identical RoO, an exporter from country X can use originating inputs from country Y or Z in his product and count them as originating in country X itself.
This type of cumulation is characterised by more flexibility and further reach than bilateral cumulation in that it is open to more than two countries, as long as they share PTAs with identical RoO. Its promotion of trade liberalisation goes therefore beyond the borders of one PTA and favours a larger group of beneficiaries, alleviating overly severe RoO more effectively. Diagonal cumulation provides, moreover, more leeway to manufacturers willing to benefit from preferential treatment, by not limiting them to sourcing non-originating materials solely from the PTA partner to which they want to export to. It has been demonstrated that PTAs allowing for diagonal cumulation have trade levels over 30% higher than PTAs without it.40
Yet, there exist some pitfalls of diagonal cumulation that bilateral cumulation does not imply. As diagonal cumulation is much less popular – it is frequently used only in the EU agreements – it may not be as widely available as it should in order to adequately fulfil its role. The fact that identical tariffs and administrative provisions regulating RoO are required from the countries inside the diagonal cumulation area further undermines its potential. Certainly, the framework of an enormous “spaghetti bowl” of PTAs does not facilitate the introduction of the same FTA tariffs and RoO in respective agreements, making diagonal cumulation hard to achieve. However, a convincing explanation of such a tough requirement has been provided by Augier et al. – if participating PTAs partners were allowed to implement RoO of different stringency, the rigorous RoO of one PTA could be circumvented by shipment of goods through another PTA country with more lenient RoO.41 Unfortunately,
40 Patricia Augier, Michael Gasiorek, ‘The EU and the Southern Mediterranean: The Impact of Rules of Origin’ (2002) preliminary draft.
15
the requirement that the value-added in the last processing stage surpasses the highest customs value of any of the inputs used from countries within the area, which Brenton believes significantly decreases the usefulness of diagonal cumulation, cannot be so easily justified.42 One could rightly point out that a final product may largely differ from its inputs without exceeding their highest customs value. In effect, diagonal cumulation – in spite of being relatively effective – for one reason or another does not reach its full potential.
Full cumulation is an extension of diagonal cumulation and applies to three or more countries likewise. It allows countries linked by the same PTAs to not only use originating materials from their partners but also to use goods produced within the zone after intermediate processing of such materials took place. All the processing undertaken in the area is looked at as though it took place in the last country of production, and must comply with RoO's substantial transformation requirements.43 Therefore, if countries F, G, and H are tied by PTAs, an exporter from country G can use products already processed in countries F and H and treat them as originating in country G itself, provided that all the processing made by the countries taken together complies with preferential RoO.
Full cumulation is indisputably more liberal than diagonal cumulation in permitting the use of non-originating materials from the ROW. In that way, it helps eliminate the negative impact of medium-restrictive RoO (leaning towards the second end of our scale) on trade between PTAs members and third countries. By allowing the division of production processes between a large number of beneficiary countries, full cumulation simultaneously enhances economic integration and trade within the area.44 Developed countries with high-cost labour may find it especially tempting, as it provides them with an opportunity to outsource some production stages to lower-cost developing countries, without jeopardising the preferential status of their final products.45 Thanks to full cumulation, developed countries can thus gain cheap outsourcing, and developing countries – more well-paid labour opportunities and economic growth.
Prima facie, the above type of cumulation appears, then, to satisfy everyone: the preferential
42 Paul Brenton, 'Preferential Rules of Origin' (n 16) 167.
43 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in the World Trading System' (n 9) 6.
44 Arvid Lukauskas, Robert M. Stern, Gianni Zanini, Handbook of Trade Policy for Development (OUP Oxford 2013) 585.
16
beneficiaries, third countries, both more and less developed countries. Regrettably, because it is rarely used (it is mostly applied in selected EU agreements, some GSP schemes and between NAFTA members), its reach is not far and very little restrictive RoO are compensated for. Full cumulation's burdensome documentary requirements, often demanding more comprehensive information than merely the certificate of origin sufficient to comply with diagonal cumulation, additionally impede its potential positive effects. Therefore, although full cumulation looks promising on paper, it is not a “game-changer” for trade liberalisation in practice.
3.2. Other measures
The previous section has pointed out how different types of cumulation, despite weakening the impact of excessively severe product-specific RoO, do so to a limited extent – bilateral cumulation due to its narrow application; diagonal cumulation because of its infrequent use in trade agreements and its additional requirements; and full cumulation for the reason of its rareness. Yet, cumulation is not the only measure aiming to make up for the over-restrictiveness of preferential RoO that deserves attention. De minimis clauses, the absorption principle, duty drawback, outward-processing, self-certification, and the possibility of alternative means of satisfying the origin requirements must also be considered in order to assess the effectiveness of the regime-wide RoO system as a whole.
De minimis (tolerance) clauses are part of the vast majority of PTAs and allow for a specific maximum percentage of non-originating inputs which can be used in the production of a final good without undermining its origin. They are usually applied in conjunction with the CTC and SMP criteria for conferring origin but their effect is nullified when they are applied with the value-added technique.46 Tolerance clauses provide for a noteworthy degree of leniency and encourage PTA-ROW trade by permitting “origin-friendly” use of a certain amount of products from third countries. Consequently, the negative impact of preferential RoO inducing PTAs producers to source components from their PTAs partners, as opposed to the ROW, can be to some extent diminished. It must be remembered, though, that only an
46 Paul Brenton, Hiroshi Imagawa, 'Rules of Origin, Trade and Customs' in Luc De Wulf and Jose B. Sokol (eds), Customs Modernization Handbook (The World Bank 2005) 196.
17
inconsiderable amount of non-originating inputs percentage, usually ranging between 7% and 15%, is allowed. Hence, the difference made by de minimis clauses is quite modest.
The absorption (roll-up) principle states that materials comprising non-originating elements but granted originating status according to relevant RoO are treated as fully originating after their incorporation into a final good, despite not having 100% of originating elements. Given that import content is not considered during the assessment of the substantial transformation, the roll-up principle provides an incentive to each PTA member to trade with third countries as well as with its PTA partners (when there is no diagonal cumulation and inputs from some PTA members count as non-originating). As it is used in virtually all PTAs – with certain sector-specific exceptions – its application is not undermined. However, the contribution the absorption principle makes in practice to the counteraction of the trade-distorting impact of RoO appears far from meaningful. Since it is only triggered once an intermediate material acquires origin, the amount of non-originating inputs it allows, and thus the incentive it provides to PTA members to have commerce outside their regional zone, is largely limited.
Duty drawback exempts producers of a final good containing non-originating inputs from a tariff payment on these inputs (or provides that the tariff is paid back) once the final good acquires the originating status. By offering duties exemptions, it provides encouragement for PTAs manufacturers to obtain their inputs from third countries for the most attractive price. At the same time, duty drawback furnishes PTA-based producers with an incentive to export inputs to their preferential partners, by making it more financially advantageous for them than selling inputs in their domestic market. Unluckily, such dual trade-creating effect is neutralised by a 'no-drawback rule' – commonly applied in PTAs – which shifts the producers' interest in sourcing materials from the ROW back to the preferential area partners (if there is diagonal cumulation) or to their domestic markets (if there is no cumulation). The no-drawback rule leads, therefore, to trade diversion and can nourish the protectionist interests of respective PTA countries.47 In that light, benefits of the duty drawback seem outbalanced by the negative impact brought by much more popular no-drawback rule.
Outward-processing permits some production processes of products with a value-added rule to be carried out outside the PTA or cumulation area while maintaining the origin of a final
18
product. It is particularly often used by developed countries for activities involving intensive labour and not significantly contributing to the final value of goods.48 In that way, less developed countries from outside the PTA zone can be more involved in some manufacturing stages. Furthermore, PTAs producers receive a stimulus to trade within the preferential area and to satisfy even stricter RoO, by being given a cost-saving opportunity to fragment their production processes outside it. With the outward-processing, it simply becomes more profitable for PTAs members to keep their commerce within the preferential system than it would without it. Yet, due to stringent conditions and rarity of the outward-processing, it does not do much to hamper the trade-distorting effects of rigorous RoO.49
According to self-certification – used in most PTAs – the exporter's signature on the certificate is enough to serve as a proof that goods in question are originating, which is not the case in certification by an industry umbrella group or certification by the exporting country government. The role of customs authorities becomes, then, minimised, the administrative costs of the producer's compliance with RoO reduced, and trade flows – increased.50 However, due to the fact that the importing country has a burden of demonstrating the eventual lack of requisite origin of certain products claimed to be originating by the exporter, the latter is put in a fraud-encouraging environment. Self-certification can, therefore, help with excessive administrative costs of preferential RoO but also raises the risk of their circumvention.
Lastly, thanks to the possibility of alternative ways of satisfying origin requirements, mere compliance with one out of two or more methods evidencing substantial transformation suffices to grant origin. Manufacturers (especially small firms) are by such means given much flexibility in satisfying their respective preferential RoO, and their trade with their PTAs partners is facilitated. Unfortunately, the flexibility of such a solution is offset by contrasting rules which confer origin on the condition that more than one substantial transformation method has been complied with. Such a requirement, applied in NAFTA and EU agreements, undermines with its restrictiveness trade liberalising effects offered by the alternative ways of satisfying the origin establishment techniques.51
48 Maria Donner Abreu, 'Preferential Rules of Origin in Regional Trade Agreements' (n 39) 14. 49 ibid.
50 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin: A World Map and Trade Effects' (n 4) 42. 51 Paul Brenton, 'Preferential Rules of Origin' (n 16) 170.
19
Neither scholarly opinions and statistics, nor the above contemplations deny that regime-wide RoO add up some highly needed flexibility to preferential RoO, leading to a boost in trade flows. Assuredly, without measures such as cumulation, de minimis clauses, or self-certification, the preferential RoO system would be completely overwhelmed by the restrictiveness of many product-specific RoO. An assumption that regime-wide RoO fix all the flaws of the preferential RoO regime would be, however, overly optimistic and inaccurate. If regime-wide RoO were enough to compensate for the stringency of preferential RoO, current discussions on the negative trade effects of the latter would be non-existent. Yet, we keep hearing about the necessity of immediate reforms in that area. The reality is that many “mitigating” measures (such as full cumulation, duty drawback, and outward-processing) are rarely used in PTAs. Those commonly used, in turn, are often insufficient to combat unfavourable effects that harsh RoO entail. Measures like bilateral cumulation, tolerance clauses, and the roll-up principle exemplify it.
4. The practical side of preferential RoO – PTAs' analysis
Sections 2 and 3 have clarified the trade-distorting impact of severe preferential RoO and the insufficiency of regime-wide RoO to completely offset it. An investigation into the practical implications of these findings seems like a rational next step on the path of exploration of the preferential RoO’s true nature. Since each PTA is characterised by different nature of provisions preventing potential trade frauds, such provisions result in RoO of diverse levels of restrictiveness. The following section will investigate disparities between preferential RoO systems in different PTAs. More specifically, it will analyse the product-specific and regime-wide RoO, as well as the utilisation rate, of three groups of PTAs: the EU PTAs, NAFTA, and PTAs between developing countries.
4.1. The EU PTAs
The model of RoO in the EU PTAs can be described as highly complex and extremely detailed on a product-by-product basis. It incorporates the CTC at the heading level as the main method of evidencing substantial transformation, accompanied by the value-added and/
20
or the SMP techniques.52 For textiles and apparels, the SMP method has been predominantly used, harshly requiring clothing to go through particular tariff headings (from yarn to fabric to clothing) to become originating.53 The 2011 reform of RoO in the EU GSPs introduced more lenient conditions for the GSPs beneficiaries.54 It enabled the least developed countries' (LDCs) textiles' exports to qualify for the originating status after mere fabric to clothing transformation, resulting in a strong preference usage increase (from 40% to 100%).55 As far as the EU RoO are deemed medium-restrictive, the level of their stringency for textiles and apparels from non-LDCs is therefore relatively high, feeding the EU's protectionist interests and contributing to trade diversion at the expense of fabrics from the ROW.
Since the commencement of the Pan-European system in 1997, product-specific and regime-wide RoO across the EU's FTAs were harmonised, introducing diagonal cumulation in that area. The cumulation zone was significantly enlarged eight years later with the accession of Mediterranean states and became the Pan-European-Mediterranean system. Currently, apart from the EU members, it comprises the European Free Trade Association countries, signatories to the Barcelona Declaration, the Faroe Islands and participants in the EU's Stabilisation and Association Process.56 As a result of the 2011 reform, unidirectional diagonal cumulation with third countries having FTAs with the EU became open to GSP beneficiaries, with the exclusion of some agricultural products.57 Few EU agreements (inter alia with EFTA and South Africa) also provide for full cumulation.58 The rejection of the position of a 'spoke' in relation to countries in Eastern and Southern Europe by the EU – which the launch of the Pan-European system inevitably entailed – was a big step towards a less protectionist trade policy, further strengthened by the reform of RoO for GSP schemes. A noticeable improvement of low trade levels from before the 1997 reform by diagonal
52 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in FTAs in Europe and in the Americas…' (n 6) 8. 53 ibid 14.
54 United Nations Conference on Trade and Development, 'Generalized System of Preferences: Handbook on the Rules of Origin of the European Union' (New York and Geneva 2013) 1.
55 Stefano Inama, 'Ex Ore Tuo Te Iudico: The Value of the WTO Ministerial Decision on Preferential RoO for LDCs' (2015) 49(4) Journal of World Trade 591.
56 Council Decision 2013/94/EU of 26 March 2012 on the conclusion of the Regional Convention on pan-Euro-Mediterranean preferential rules of origin [2012] OJ L54/3, art 1(3).
57 Commission Regulation (EU) 1063/2010 of 18 November 2010 amending Regulation (EEC) No 2454/93 laying down provisions for the implementation of Council Regulation (EEC) No 2913/92 establishing the Community Customs Code [2010] OJ L307/1, art 86(7).
58 European Economic Area Agreement, Protocol 4 on Rules of Origin; Agreement on Trade, Development and Cooperation between the European Community and its Member States, of the one part, and the Republic of South Africa, of the other part, Protocol 1.
21
cumulation only proves the success of such a move.59 Nonetheless, the exclusion of textile products from the cumulation system for non-GSP beneficiaries involves potentially distorting impact on commerce.
De minimis rule is part of the EU RoO regimes and permits a higher percentage of non-originating inputs than many other PTAs worldwide (around 10%-15%).60 However, it generally does not apply to textiles and apparels, except for the specific maximum weight of non-originating materials it permits.61 The roll-up principle, on the other hand, has been consistently applied in all EU FTAs.62 So has the outward-processing, yet subject to numerous limitations, such as the preclusion of textiles and clothing.63 Although duty drawback has been normally excluded from the EU trade agreements, it has been allowed in PTAs linked to development purposes, such as the GSP schemes.64 Moreover, a phase-out period temporarily permitting duty drawback has been provided for in certain FTAs, further alleviating the restrictiveness of the EU's RoO regimes.65 When it comes to the certification costs, the two-step system – placing the administrative burden on the exporter and his country's government – has been applied.66 The lack of possibility of self-certification has been compensated for by a more lenient method for “approved exporters” making regular shipments and being authorised by the exporting country's customs services.67 Stricter RoO in the EU FTAs have been further mitigated by the option of satisfying them (for around 25% of products) in alternative ways.68 Regime-wide RoO have been, then, quite generously applied in the EU preferential system, especially towards developing countries and non-clothing products.
The hard-to-meet restrictiveness of the EU's RoO and their low utilisation rates for
59 Wayne Crook, Jenny Gordon, 'Rules of Origin: Can the Noodle Bowl of Trade Agreements Be Untangled?' (n 24) 8.
60 Regional Convention on Pan-Euro-Mediterranean Preferential Rules of Origin, art 5(2). 61 ibid.
62 ibid art 5(1); Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in the World Trading System' (n 9) 22. 63 Regional Convention on Pan-Euro-Mediterranean Preferential Rules of Origin, art 11;
Maria Donner Abreu, 'Preferential Rules of Origin in Regional Trade Agreements' (n 39) 14.
64 Commission Working Document 7341/10 of 10 March 2010 on the future of 'duty drawback' in the rules of origin of EU's Free Trade Agreements [2010].
65 Economic Partnership, Political Coordination and Cooperation Agreement between the European
Community and its Member States, of the one part, and the United Mexican States, of the other part, Annex III, art 14(7); Agreement establishing an association between the European Community and its Member States, of the one part, and the Republic of Chile, of the other part, art 14(6).
66 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in the World Trading System' (n 9) 24, 45. 67 ibid.
22
developing countries constituted essential reasons given by the European Commission for initiating alterations in the GSP schemes in 2011.69 Indeed, back in 1994, only 38% out of 79% of GSP beneficiaries eligible for preferential access to the EU market obtained it, paying a tariff lower than the MFN rate.70 In 2001, out of virtually all of Cambodia's products qualifying for preferential access to the EU, merely 36% actually entered it duty-free.71 Strict RoO – especially for clothing products – were to be blamed for such outrageous underutilisation, a well-known feature of the EU regimes.72 Since the late 1990s, however, all the changes discussed above have taken place, creating preferential RoO framework much more attainable for developing countries. Going back to the Cambodia example, Tanaka shows that its clothing export to the EU rose by 112% after 2011.73 In 2012, about 90% of imports eligible for preferences in the EU utilised them.74
Naturally, there are still issues that must be addressed by the EU preferential system, such as inequalities in the treatment of different PTAs partners and highly diverging utilisation rates between beneficiaries.Nonetheless, by means of the inclusion of developing countries in the preferential RoO framework and constant broadening of diagonal cumulation area, substantial progress towards less protectionist and more consistent RoO has been made by the EU. As will be seen in the next sub-section, the same cannot be said about NAFTA.
4.2. NAFTA
Specified on a product-by-product basis, NAFTA's RoO are predominated by the CTC on the heading and chapter levels, sometimes accompanied by the value-added or SMP.75 The change of chapter method – by its nature much stricter than a mere change of heading requirement largely present in the EU regime – has been one of the main reasons for the RoO
69 Commission Staff Working Document SEC (2012) 87 final of 27 January 2012 on trade as a driver of development [2012], sec 5.
70 Andre Sapir, 'The Political Economy of EC Regionalism' (1998) 42 European Economic Review 717. 71 Paul Brenton, 'Preferential Rules of Origin' (n 16) 171.
72 Paul Brenton, Hiroshi Imagawa, 'Rules of Origin, Trade and Customs' (n 46) 201.
73 Kiyoyasu Tanaka, 'The EU’s Reform in Rules of Origin and International Trade: Evidence from Cambodia' (770 Institute of Developing Economies Discussion Paper, IDE 2020).
74 Alexander Keck, Andreas Lendle, 'New Evidence on Preference Utilization' (Staff Working Paper, WTO 2012) 17-18.
23
regime in NAFTA being considered harsh and onerous.76 In that respect, NAFTA's RoO framework is, therefore, more restrictive than that of the EU. Quite a few similarities between the two systems are nevertheless noticeable: apart from extremely detailed and product-specific approach to RoO, both NAFTA and the EU have adopted rigorous RoO for textiles and apparels, requiring their progressive transformation from yarn to clothing (with the exception of LDCs in the EU PTAs).
Full cumulation is widely applied in NAFTA between its members, allowing the US, Canada, and Mexico to treat goods produced in the territory of other members like produced in its own territory, as long as relevant applicable CTC and value-added requirements are satisfied.77 Diagonal cumulation with third countries is, instead, precluded, despite it being a crucial feature of the EU PTAs. It is, however, de facto applied for automatic data processing equipment, in situations when the tariff applied by all three NAFTA members to third countries has identical value.78 The strict nature of the NAFTA-ROW cumulation regime visibly contrasts with full cumulation permitted within NAFTA, and with the generous system of diagonal cumulation applied by the EU.
De minimis rule is provided for in NAFTA on the condition that non-originating inputs do not exceed 7% of the transaction value of the good.79 Various non-originating materials (such as dairy, peanuts, and sugar) are excluded from its application.80 Not only is the allowance of 7% lower than in the EU agreements, but the number of exceptions is also much higher. In textiles, de minimis rule's reference to the total weight of the component, rather than its value, is, however, exactly what the EU PTAs constitute.81 Similarly to the EU, NAFTA also uses the roll-up principle, but restricts its application solely to the value-added calculations, and excludes from it motor vehicles.82 Duty drawback has been generally precluded both from the EU regimes and from NAFTA. The latter afforded some leniency by substituting drawback with a system providing for the refund of the lesser of the total amount of customs duties paid on imported goods and the exports of the good to another NAFTA party.83 The
76 Caroline Freund, 'Streamlining Rules of Origin in NAFTA' in C. Fred Bergsten and Monica de Bolle (eds), A Path Forward for NAFTA (Peterson Institute for International Economics 2017) 114, 117.
77 North American Free Trade Agreement (NAFTA) art 404.
78 Maria Donner Abreu, 'Preferential Rules of Origin in Regional Trade Agreements' (n 39) 11. 79 NAFTA, art 405(1).
80 ibid art 405(3)-(5). 81 ibid art 405(6). 82 ibid art 402(4). 83 ibid art 303.
24
outward-processing – conditionally permitted in the EU PTAs – is not part of the NAFTA RoO regime, similarly to alternative ways of satisfying RoO present in the EU framework.84 However, contrary to the EU that has not adopted self-certification, NAFTA has, providing more convenience to exporters.85 Overall, though, NAFTA's regime-wide RoO seem more restrictive than those of the EU, mostly due to their cumulation, de minis rule, and the roll-up principle provisions as well as the lack of both outward processing and the alternative ways of complying with substantial transformation criteria.
Within the framework of NAFTA, RoO are rightly placed among overriding causes of trade diversion and low utilisation rates of preferences (merely 64% in 2000).86 As Conconi et al. have found, restrictive RoO in NAFTA have reduced imports of parts from the ROW by 30% on average, by “encouraging” the US, Canada, and Mexico to source inputs from each other.87 It has been also contended that in spite of certain administrative amendments aiming at alleviating the NAFTA RoO regime, the rigidity of NAFTA's structure and the lack of flexibility from its members precluded the success of such measures.88 The US-Mexico-Canada Agreement (signed in 2018), despite updating the general RoO found in NAFTA, has not significantly diminished their stringency, either.
In effect, the restrictiveness of NAFTA is still at a much higher level than necessary to prevent trade deflection. RoO in sectors such as footwear, textiles, and apparel, in particular, are overly strict and prevent the use of available preferences by NAFTA's parties.89 Estevadeordal and Miller identify the very transition from the US-Canada FTA to NAFTA as marking the beginning of stringent RoO for these sectors.90 As a result, NAFTA parties –
84 Cheong Cho, 'Rules of Origin and Agricultural Trade Liberalisation in Major Free Trade Agreements' in Christopher Charles Findlay and Shujiro Urata (eds), Free Trade Agreements in the Asia Pacific (World Scientific Publishing 2013) 13.
85 NAFTA, art 501.
86 Jaime Serra, Valeriana Kallab, ‘Reflections on Regionalism: Report of the Study Group on International Trade’ (Carnegie Endowment for International Peace 1997).
87 Paola Conconi, Manuel Garcia-Santana, Laura Puccio, Roberto Venturini, 'From Final Goods to Inputs: The Protectionist Effect of Rules of Origin' (2018) 108(8) American Economic Review 2335.
88 Ticon Development Consulting, the Origin Institute, Trade Facilitation Services, 'Basics of Economic Policy no 21: Cross-Cumulation in Free Trade Agreements. Opportunities, Potential and Challenges' (study on behalf of the State Secretariat for Economic Affairs, Swiss Federal Department of Economic Affairs 2013) 21.
89 Olivier Cadot, Jaime de Melo, Antoni Estevadeordal, Akiko Suwa-Eisenmann, Bolormaa Tumurchudur, 'Assessing the Effect of NAFTA's Rules of Origin' (Research Unit Working Papers 0306, INRA 2002) 11, 26.
90 Antoni Estevadeordal, Eric Miller, 'Rules of Origin and the Pattern of Trade between US and Canada' (Inter-American Development Bank 2002).
25
albeit having access to preferential treatment that undoubtedly benefits them – cannot fully utilise it due to the over-severity of RoO. For the same reason, third nations' producers are also put in an underprivileged position.
4.3. PTAs between developing countries
The complexity of RoO in the EU PTAs and NAFTA stands in opposition to simple RoO found in trade agreements between developing countries. PTAs such as MERCOSUR (Southern Common Market), AFTA (ASEAN Free Trade Area), LAIA (the Latin American Integration Agreement) and COMESA (Common Market for Eastern and Southern Africa) contain very few – or no at all – product-specific RoO and tend to use general RoO across all items. The CTC on the heading level and the value-added criteria are the two most commonly used alternatives for the establishment of substantial transformation in agreements like LAIA, COMESA, and AFTA; combinations of the value-added method and SMP are also applied (e.g. in MERCOSUR).91 Some PTAs between developing countries provide for reductions of the minimum value-added percentage required for products originating in their least-developed members.92 ASEAN offers a 10% reduction to India, whereas MERCOSUR – 20% for Paraguay, Uruguay, and Argentina when exporting to Uruguay, on a temporary basis.93 By the above means, the qualification for preferential treatment is facilitated for the economically most vulnerable PTAs members.
All four PTAs mentioned above permit full cumulation among their members (with the exclusion of LAIA which only offers bilateral cumulation).94 When it comes to cumulation with third countries, only MERCOSUR allows for diagonal cumulation with Peru, Bolivia and the Andean Community (Bolivia, Columbia, Ecuador, Peru).95 In comparison to the EU RoO model providing for diagonal cumulation in the whole Pan-Euro-Med zone, diagonal cumulation provisions in developing countries' PTAs appear, therefore, very modest. The fact that many of them permit full – and not merely bilateral – cumulation between their members
91 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin in FTAs in Europe and in the Americas…' (n 6) 9. 92 Maria Donner Abreu, 'Preferential Rules of Origin in Regional Trade Agreements' (n 39) 7.
93 ibid.
94 ibid 10; Miriam Manchin, Annette O. Pelkmans-Balaoing, 'Rules of Origin and the Web of East Asian Free Trade Agreements' (World Bank Policy Research Working Paper 4273, the World Bank 2007).
26 should be appreciated nevertheless.
De minimis rule is not mentioned in any of the agreements in the context of preferential treatment.96 Similarly, the roll-up principle has been omitted from AFTA and LAIA but has been provided for in MERCOSUR (with the exception of the automotive sector) and COMESA.97 Duty drawback provisions have also been included in AFTA and MERCOSUR (again, except for automotive imports from Argentina and Brazil), in contrast to the EU's PTAs, NAFTA, and many PTAs in Africa and Latin America.98 Outward-processing and self-certification, however, are rarely available in PTAs between developing countries. Indeed, MERCOSUR and AFTA require certification conducted by a public body or an approved private umbrella entity, whereas LAIA and COMESA apply the rigorous two-step certification process, just like the Pan-Euro-Med area.99 Such harsh practice stands in opposition to the NAFTA regime that uses self-certification. Yet, unlike NAFTA, alternative ways of satisfying the requisite RoO are commonly applied in all four PTAs. Apart from duty drawback, the use of regime-wide RoO in PTAs between developing countries may not be as abundant as in some PTAs between developed countries. Still, the level of stringency of preferential RoO in agreements such as MERCOSUR or LAIA is significantly lower than in EU's PTAs and NAFTA. All in all, RoO in PTAs between developing countries are much more lenient, with or without the tolerance rule or self-certification.
The leniency and simplicity of PTAs become meaningless when their utilisation rate is disproportionately low. That seems to have been the case with AFTA, in which only about 10% of preferences on average were used in 2002 and only around 5% in 2007.100 As Manchin and Pelksman evidence, there exists a strong link between preferential margins of at least 25% in East Asian PTAs and the utilisation of their preferences.101 Estevadeordal and Suominen also indicate that reductions of preferential tariffs amount to higher utilisation rates.102 Such findings shed light on a crucial issue relating to preferential RoO – their
96 Antoni Estevadeordal, Jeremy Harris, Kati Suominen, 'Multilateralising Preferential Rules of Origin Around the World' (IDB-WP-137, Inter-American Development Bank 2009) 18.
97 Antoni Estevadeordal, Kati Suominen, 'Rules of Origin: A World Map and Trade Effects' (n 4) 18. 98 ibid 18, 20.
99 Antoni Estevadeordal et al., 'Multilateralising Preferential Rules of Origin Around the World' (n 96) 22. 100 Miriam Manchin, Annette O. Pelkmans-Balaoing, 'Rules of Origin and the Web of East Asian Free Trade
Agreements' (n 94) 13; Richard Baldwin, 'Managing the Noodle Bowl: The Fragility of East Asian Regionalism' (2008) 53 The Singapore Economic Review 449.
101 ibid 18.
27
leniency cannot always compensate for low preferential margins. In other words, even relaxed RoO that furnish PTAs producers with an incentive to make use of their preferential treatment will be insufficient if the difference between the MFN tariff and the preferential tariff is so low that it makes it unprofitable for the producers to comply with all the (even relatively minor) costs that RoO entail. Hence the visible differences in the utilisation rates of preferences between PTAs with higher preferential margins like LAIA (where the utilisation rate reached over 77% in 2010) and with lower ones like AFTA.103
Based on the above, it appears clear that despite the leniency of RoO in PTAs between developing countries, individual disparities in the levels of preferential and MFN tariffs largely influence their utilisation rates. In other words, the core of the problem has been not so much the stringency of preferential RoO as too low margins of preferences. In the EU's PTAs and NAFTA, instead, the overriding cause of preferences underutilisation, especially in certain sectors, have simply been restrictive RoO. Although the size of preferential margins is objectively pertinent to every PTA's utilisation, the foregoing comparison indicates how much more noticeable its significance is in PTAs where mild RoO are applied.
5.
Proposed changes to the preferential RoO framework
The system of preferential RoO – albeit varying from PTA to PTA – has been undisputedly failing in many ways by being overly strict and oppressive. Unsurprisingly, voices calling for reforms in that area started being heard, demanding the implementation of necessary changes into preferential RoO. The following section will elaborate on different proposals (including the author’s proposal) for how such an improved framework of preferential RoO could be accomplished, starting from the most moderate one and ending with the most radical ones.
The first proposal – arguably the most feasible one – encompasses the main findings of the paper. It is based on a premise that in order to succeed, the reform must achieve the right balance between precluding “free riders” from unfairly benefiting from preferential treatment on the one hand and not denying trade opportunities to PTAs members and third countries, on
