The Definition of ‘Covered Entities’ under the GPA: The General Scope and Escape Conditions

Abstract

The WTO Government Procurement Agreement (GPA) does not legally define what entities should be covered by the Agreement. However, its member Parties list their ‘covered entities’ in a series of schedules. The list approach has complicated accession negotiations and discourages Parties from providing a ‘wider’ range of entity coverage. Moreover, the list approach raises some tensions and a lack of legal certainty, especially concerning those that are not strictly ‘government entities’, such as State-owned enterprises (SOEs). This problem is exacerbated in the case of modern SOEs in developing countries, many of which can bear both public and private features. Given these conditions, the author proposes a definition of ‘covered entities’ to facilitate accession negotiations and the future expansion of the GPA. The proposal is based on a comparative study of the GPA and the EU public procurement regulations. It develops a framework by which all publicly controlled entities are presumably covered by the GPA. Nevertheless, Parties can rebut GPA obligations by proving that an entity competes with other commercial entities under normal market conditions.

1. Introduction

Since the World Trade Organization (WTO) was founded, it has regulated most trade discrimination through multilateral initiatives. However, the WTO's principal agreements exclude government procurement from non-discrimination rules. This is due to specific exemptions in GATT Article III: 8(a) and GATS Article XVII, both concerning national treatment obligations. These exclusions suggest the need for a specific agreement to govern government procurement regulations.

The (WTO Government Procurement Agreement (GPA) 1994 (which came into force in 1996) was an independent plurilateral agreement, not a complementary Code after the GATT. The achievement of the greatest extension of entity coverage and the addition of signatories from developing countries was a ‘built-in’ mandate of the GPA.¹ During the Uruguay Round (1986–1994), the coverage negotiation avoided the contentious issue of ‘what is government procurement?’. Instead, the original Parties determined the ground they wanted to cover, and the Informal Working Group categorized each Party's applicable entities into groups A, B, C, and D respectively in the GPA Appendix I. A revised GPA was adopted on 30 March 2012. It added over 200 new contracting entities to both the central and sub-central levels.² It has been provisionally estimated that the renegotiation gains about 80–100 billion USD in market access annually.

The revised GPA is the most critical international government procurement regulation under the WTO. While it has many strong points, the GPA does not legally define the entities that should be covered by the Agreement. Notably, the evolution of the GPA has produced a list approach whereby Parties list their ‘covered entities’ in a series of schedules. Article II of the revised GPA states that the agreement applies to covered procurement made by ‘covered entities’. However, the GPA text does not define those ‘covered entities’.³ According to Appendix I of the GPA, those ‘covered entities’ are generally listed as ‘central government entities’ (Annex 1), ‘sub-central government entities’ (Annex 2), and ‘other entities’ (Annex 3). ‘Other entities’ are a miscellaneous category, referring to entities whose procurement policies are not substantially controlled by, dependent on, or influenced by central, regional, or local government.

Article II.2 follows the fashion of the GATT and GATS, but it is not as clear as the GATT and GATS. Compared with the GATT and GATS, which handle trade measures and practices in goods and services, respectively, the GPA is virtually sui generis in handling trade of a government nature. Whether an entity is ‘governmental’ or not is more essential in the GPA context than in the GATT and GATS. Consequently, whether an action is being taken by a ‘governmental entity’ or a ‘private person’ is, at first instance, relevant to the application of the GPA. Therefore, the language in Article II.2 of the revised GPA is less clear and therefore less effective than in the GATT. Furthermore, different threshold⁴ and tendering procedures⁵ apply to whether an entity is classified as central, sub-central, or other government entities. Therefore, it is very necessary to make clear precisely what kind of government entity should be subject to the GPA obligations.

Specifically, the divergence in understanding the ‘covered entities’ lies in Annex 3: ‘other entities’. It includes ‘entities which bear governmental features but apart from the entities which are governmental bodies according to the constitutional laws of each Party’.⁶ Most entities in the GPA Parties' annexes are State-owned enterprises (SOEs), utilities enterprises, and their equivalents. Unlike understanding of the ‘central and sub-central government entities’ among WTO members, SOEs and utilities have different relationships with their governments. Consequently, due to the diversity of those ‘other entities’ in WTO members' national economies, there is no consensus on which ‘other entities’ should be covered under the GPA.

The significance of the concept for the development and future expansion of the GPA has been acknowledged. Additionally, the definition of ‘covered entities’ remains problematic, not just within the GPA context but also in defining terms such as ‘government entity’, ‘public body’, and ‘State-owned enterprise’. These definitions are frequently at the heart of disputes among WTO members. Importantly, the WTO Appellate Report indicates that these unclear definitions continue to be a source of contention among members.⁷

The purpose of present paper is to facilitate GPA expansion by proposing a definition of a ‘covered entity’ under the GPA to clarify understanding of ‘other entities’. This argument stands in contrast to the extant literature which focuses on the technical rules governing tendering procedures or the improvements of the revised GPA and its impact on international trade, transparency, and sustainable procurement.⁸ To date, research on GPA coverage has been mostly restricted to the scope of contracts, particularly service contracts.⁹ Several researchers have touched upon the entity coverage topic.¹⁰ Arrowsmith, Evenett, and Hoekman considered that limitation to market access and predominance of the positive list approach under the GPA discouraged developing countries' interests in joining the GPA, and the extensive derogation from the MFN principle had a severe impact on the negotiations on coverage as negotiation by the Parties results in Schedules has no common rules or patterns apply.¹¹ It is also observed that the lack of clarification of ‘covered entities’ makes it difficult for countries with large state sectors to consider accession to the revised GPA.¹² The works of Ping Wang and Skye Mathieson provide ideas on this issue from a particular perspective. Ping Wang suggests that the coverage of government procurement disciplines should consider the factor of government control and competition.¹³ Wang generally contends that when an SOE is publicly controlled or influenced to compete in the market, it should not be covered by the GPA. Mathieson suggests that a ‘factor-based’ definition of control should be formulated by integrating the relevant practices of the US, the UNCITRAL, the EU, China, and the GPA.¹⁴ This work provided a reference point for recognizing the existence of government influence or control. Regarding the competition factor, Mathieson suggests using the Herfindahl–Hirschman Index (HHI) to indicate whether the publicly controlled or influenced entity has competitiveness. The HHI is useful for indicating market concentration – namely, the distribution of competition opportunities. However, market concentration does not necessarily indicate whether a competitor in this market will or will not distort procurement decisions.

Both Wang and Mathieson provide a perspective on public control and competition, but further study is still necessary to define the two factors in order to distinguish the treatment of different SOEs, utility entities, and entities with special or exclusive rights as well as to determine how these two factors interact with each other. Furthermore, it is also particularly important to note the size of the procurement of those entities since this is relevant in deciding whether there will be a trade-distorting effect. This paper will propose that entities under public control are presumably subject to the discipline of the GPA, but these entities could rebut. If an entity can prove that it competes in the market like other commercial entities, it could be exempt from the GPA obligation.

In terms of methodology, the research presented here is grounded in a comparative analysis of two legal systems – the WTO/GPA and the EU Public Procurement Directives. The origins of the GPA can be traced back to the OECD's efforts on public procurement in the 1960s.¹⁵ During the 1970s, these efforts were significantly informed by the EU's regulations on public procurement.¹⁶ Subsequently, in 1977, the work influenced by the EU within the OECD was handed over to the GATT Secretariat, and it was titled the ‘Draft Integrated Text for Negotiations on Government Procurement’.¹⁷ Given their shared origins, the revised GPA was influenced by EU public procurement legislation.¹⁸ Consequently, the EU approach to similar matters could serve as a useful reference for addressing issues within the WTO/GPA.

The main body of the argument will be developed in the next three sections. Sections 2 and 3 answer the questions of why the current list approach should be abandoned and why a definition of ‘covered entities’ is desirable. Section 4 reviews the regulations of the GPA in the WTO legal regime and emphasizes the developments in some major recent FTAs, and sets the context for the discussion of the proposal in Section 5. Section 5 elaborates on the contours of the proposed definition from the aspects of political consensus, legal premise, legal structure, and conceptual framework, followed by policy advice at the end. Section 6 concludes.

2. Lessons Learned from the Negotiation History: Power-Driven Bargaining or Rule-Based Harmonization?

In general, government procurement trade negotiations can take three modalities: single-issue-based multilateral negotiations, multilateral multi-issue single undertakings, and single-issue-based plurilateral negotiations. Each is elucidated below.

2.1 Single-Issue-based Multilateral Negotiations

This modality is exemplified by Agreements within the WTO.¹⁹ Notably, the Tokyo Round Codes, negotiated between 1973 and 1979, were primarily subjected to multilateral disciplines, with the Code on Government Procurement being a notable exception.²⁰ However, attempts at single-issue-based multilateral negotiations on government procurement under the GATS, the WTO on government procurement, and the OECD initiative all encountered failure.²¹ The key takeaway from the failure on the OECD and the GATS forums is that any genuinely global procurement instrument should include specific provisions catering to the needs of developing countries, with the active participation of these countries in shaping such provisions.²²

2.2 Multilateral, Multi-Issue Single Undertaking

Negotiations in the Tokyo Round (1973–1979), the Uruguay Round (1986–1994), and the ongoing Doha Round (since 2001) follow this modality. The multilateral initiatives on government procurement, proposed in 1996 as part of the Singapore issues, were primarily driven by OECD countries, with the EU and the US playing leading roles. However, these negotiations reflected the interests of major players, and their success was influenced by their significant share in the trade and government procurement market. The Uruguay Round Agreement was perceived by most developing countries as not increasing economic opportunities but entailing high implementation costs and diverting resources from the more pressing market access concerns. The Doha Round was initiated after the Seattle debacle in 1999, driven by developed countries dissatisfied with the Uruguay Round Agreements on services and agriculture. Yet, many developing countries viewed negotiations on Singapore issues as unfavourable, believing that the new round did not promise a balanced ‘fair’ agreement.²³

The GPA negotiation similarly faced concerns from developing countries regarding costs and benefits. Only three developing countries signed the GPA, while others participated as observers.²⁴ Developing countries are concerned that their current share of the government procurement market is minimal, and the potential export gains from joining the agreement are expected to be marginal, at least in the foreseeable future.²⁵ Furthermore, gaining access to the GPA involves legal compliance, potential institutional reforms, and adjustments to legislation, incurring costs for national governments. Consequently, there is suspicion that developed countries stand to gain more from the entry of developing countries, while the latter may face burdens outweighing the benefits of GPA accession.²⁶ Therefore, many developing countries suspected that the GPA brought more costs than benefits, leading to a growing number of observers, albeit slow progress in accession negotiations.²⁷ Currently, all 35 observers are developing countries or economies in transition. The limited interest shown by developing countries in joining the GPA might be attributed to an imbalance in trade benefits.²⁸ Notably, major GPA players such as the EU, the US, and Japan saw little change in their share of domestic firms in total procurement covered by the GPA after signing, whereas smaller countries experienced a decline in domestic procurement share.²⁹

A crucial lesson learned is the necessity of maintaining a balance between developed and developing countries throughout negotiations. An emerging consensus suggests that trade negotiations should address the specific concerns of developing countries, acknowledging implementation costs, and possible institutional adjustments, thereby easing political pressure associated with opening the government procurement market. The complex negotiations on government procurement underscore that the expansion of the GPA must be guided by the participation and incorporation of the special concerns of developing countries.

The process of GPA accession negotiations is heavily influenced by power dynamics rather than rational moral, legal, or economic criteria.³⁰ The shifting dynamics of power within the WTO are evident in the growing trading capacity of developing countries. This expanded capacity has notably increased their influence in WTO negotiations, and interests in participating in trade rule-making in the WTO, a trend further emphasized by the accession of China in 2001. Developing countries with limited trade power may feel compelled to join the GPA as a precondition for WTO membership, as seen in the case of Chinese Taipei.³¹ Conversely, developing countries with more substantial trade power, like China, are better positioned to resist pressure and negotiate based on their specific trade needs. Over the past 17 years, China has submitted six negotiation offers, demonstrating its commitment to bargaining on trade concerns, including the coverage of its SOEs.³²

The success of the power-driven approach was evident in the twentieth century when negotiations were primarily led by winning countries such as the US and the EU. However, in the twenty-first century, with the economic development of developing countries, there is a growing interest in participating in WTO trade rule-making. This shift in the power balance, including within the government procurement market, suggests that a rule-based approach would garner more consensus and support in future trade negotiations. A rule-based negotiation is not only a desire but also a necessity to achieve further trade liberalization among WTO members.

Moving on to the second point, the GPA negotiations coincided with a surge in international trade following the establishment of the WTO. This surge was propelled by unilateral trade liberalization efforts by many countries, including major emerging markets. The benefits of this international trade boom made many countries reluctant to tackle challenging issues such as government procurement. After the successful outcomes of the Tokyo Round and the Uruguay Round, governments had reduced interest and fewer incentives to invest political resources in trade talks. Without a rule-based approach that respects and incorporates common trade concerns, developing countries lack the political motivation to reform their domestic government procurement systems for GPA accession.

A crucial lesson from the failure of the OECD to negotiate a multilateral agreement on investment in 1995 is that reducing trade and investment barriers requires broad-based domestic support for open markets.³³ This lesson holds true for multilateral negotiations on government procurement, as observed in the Singapore Ministerial Meeting in 1996. The diverse economies of major developing countries make it challenging to include all state-owned enterprises indiscriminately. Without a rule for co-efficient negotiation regarding the inclusion/exemption of entities such as state-owned enterprises, the marginalization of developing countries' concerns in GPA rule-making or asymmetric pressures within the WTO may further dampen their interest in GPA accession.

2.3 Single-Issue Based Plurilateral Negotiations

Plurilateral negotiations centred around a single issue have gained prominence, driven by the constraints of multilateral modalities and the growing influence of emerging powers.³⁴ Maintaining its plurilateral nature since the Tokyo Round Code on Government Procurement in 1979, the GPA operates through a collection of preferential bilateral agreements among Parties.

While the GPA of 1994 witnessed a substantial increase in overall entity coverage, the plurilateral nature of the agreement meant that the opening of government procurement markets was not universally experienced. The impact of trade liberalization was primarily felt among developed countries, with the expansion of entity coverage being largely attributed to the bilateral efforts of the EU and the US.³⁵ Other Parties to the agreement did not witness a comparable level of government procurement market access expansion. Additionally, attempts to increase the number of signatories fell short, resulting in only 25 signatories at the time of the GPA's enforcement on 1 January 1 1996, a figure similar to that of the Tokyo Code on Government Procurement.³⁶

The negotiations surrounding entity coverage were marked by divergent interests. Parties like the European Union sought to broaden coverage, including utility sectors, while others such as Japan and Korea vehemently opposed such extensions. The absence of a shared definition of a ‘government entity’ fuelled persistent disagreements throughout the negotiations. This issue became more complex as the GPA invited developing countries, such as the BRICS nations with sizable state sectors, to join. The lack of a common definition of ‘other entities’ emerged as a significant obstacle in reaching an agreement among Parties with diverse economic structures. Despite an increase in signatories from 22 to 49 in the Revised GPA, negotiations on coverage are progressing only with difficulty.³⁷

The plurilateral nature reflects the disparity between the Parties. Among all the disparities, the entity coverage issue is one of the most fundamental. Although the list approach effectively pushes coverage negotiation into reaching a compromise, in a longer perspective, it has disadvantages.

3. The Drawbacks of the Current List Approach of GPA Entity Coverage

The current entity coverage approach derives some features from the OECD discussion on government procurement but has also developed its own features compared with the GPA 1994. On the one hand, the OECD forum decided that entity coverage would be based on a specific list of entities,³⁸ which states accepted as providing reciprocal levels of coverage between Parties. The Tokyo Round Code on Government Procurement, the GPA 1994, and the revised GPA all retain the list approach. Since the GPA 1994, the covered entities on each Party's list have been categorized into three lists under the revised GPA: Annex 1: central government entities; Annex 2: sub-central government entities; and Annex 3: other entities. On the other hand, the GPA has dropped the idea of a general definition as suggested by the OECD. The current entity coverage depends, to a substantial extent, on the political will for reciprocal trading concessions from other countries.

3.1 Limited Openness under Item-for-Item Reciprocity

A fundamental concept used in the GPA entity coverage negotiation is reciprocity. Loosely defined, reciprocity is the practice of making an action conditional upon an action by a counterpart.³⁹ Reciprocal negotiation ensures that the market access granted by one Party is frequently balanced against the market access granted simultaneously by its trading partner. Instead of uniformly opening all listed procurement opportunities to all the GPA signatories, as in the Tokyo Round Code on Government Procurement (1979), the GPA negotiations have been conducted bilaterally on the basis of stringent reciprocity at each purchasing entity level (‘central, sub-central, or other entities’).⁴⁰ The coverage of the Code was limited to goods procurement by central government entities that agreed through quid pro quo negotiations.⁴¹ Not all central government entities were generally covered. The Code was only applied to those government entities of which the Parties had ‘contributed’ to the Code, as listed in Annex I.⁴² However, statistics show that, due to the unclear coverage and high threshold value, the outcome was far from satisfactory.⁴³ In 1987, it was revised (namely, the Revised Tokyo Round Code on Government Procurement 1988).⁴⁴ However, the revised Code did not make any extension on entity coverage.⁴⁵

Due to discrepancies in government involvement in national economies between the Parties, the item-for-item reciprocity negotiations on the ‘other entities’ level are tortuous and controversial. One case in point occurred during China's accession negotiations. While acknowledging the significant improvements in the fifth revised offer from China, the EU and the US held that the gaps in the level of the sub-central government entities and SOEs were still significantly larger than their expectations.⁴⁶ In those situations, item-for-item reciprocity leads to numerous derogations in the Parties' coverage schedules.

In addition, even if one Party makes an extensive coverage offer with a general approach, that would not encourage other Parties to make an offer with the same degree of openness. For example, the EU has placed all the contracting authorities and public undertakings covered by the EU Utilities Directive in its Annex 3.⁴⁷ Currently, most of the general utility coverage is made by EU Member States, the EEA countries (such as Norway, Iceland, and Liechtenstein), and countries influenced/connected by/to that area (Moldova and Montenegro)⁴⁸ to align with the EU Public Procurement Directives. However, most other GPA members, such as Japan, Korea, Israel, Hong Kong, Canada, New Zealand, Australia, and the US, prefer to list limited coverage of the utilities entities.⁴⁹ The asymmetric situation in the utilities sectors makes the EU's expectation of a ‘more open’ market in utility sectors difficult to fulfil in the case of countries with large state sectors, such as, China, United Arab Emirates, Russia, and Vietnam.

Consequently, entity coverage is significantly more limited than could be achieved under a general most-favoured-nation obligation. Although the general coverage from the EU is considerably inclusive, under the item-for-item reciprocity approach in order to keep the same level of market openness, the EU made derogations to Parties on the ‘other entities’ level. For example, the EU derogates procurement by entities in the port, airport, and public transportation field from the US, and derogates procurement by entities in the electricity industry field from Japan, Korea, and Israel.⁵⁰ However, in the bilateral agreement between the EU and the US, the coverage is broader than under the GPA. One of the reasons why actual market access is much more limited than may appear at first glance is that, with a stringent reciprocity approach, there is no possibility of cross-sector exchanges or exchanges between different levels of entities, which contribute to a significant reduction of potential coverage.⁵¹

3.2 Extensive Derogations in the Entity Coverage Schedules

The consequence of the item-by-item negotiation is not only that coverage is significantly more limited than expected but, more importantly, that the GPA, in fact, harbours an intra-discriminatory trade regime among its signatories.

As noted above, the EU has included a long and complicated list of derogations in its coverage of ‘other entities’,⁵² and similar derogations also exist in other Parties' coverage schedules, such as those of Iceland and Canada. All these derogations are accompanied by a declaration that procurement by those entities will not be subjected to ‘covered procurement’ until the respective signatory has given comparable access to its procurement market.⁵³ As a result, a GPA Party can still discriminate against other Parties in cases where other Parties' government procurement does not fall within the Party's coverage schedules, even though the GPA enshrines the principle of national treatment and non-discrimination. As we will see, government procurement market access is not open to non-signatories of the GPA, and it is also partly closed to signatories. The potential trade distortion is therefore great, and trade barriers continue to flourish and distort trade between WTO members.⁵⁴

Another interesting point is that although those Parties made extensive derogations in the GPA, they provided more extensive coverage on government procurement in their bilateral preferential trade agreements. For example, the EU–Singapore FTA, in which Parties generally provided their coverage for central government entities, utilities and other entities, and public–private–partnerships and services. The EU gives Singapore access to nearly 200 central entities that it withholds from the GPA, and Singapore lists 54 entities in contrast to its GPA coverage of 23 entities. In the EU–Canada Free Trade Agreement implemented in October 2017, the EU extended its coverage, especially with respect to sub-central and other entities.⁵⁵ Similarly, in the EU–Japan Economic Partnership Agreement, which took effect on 30 January 30, 2019, the scope has been notably broadened to include the railway sector and sub-central government coverage.⁵⁶ Furthermore, the EU and the US have also issued proposals to expand the reciprocal level of market access to each other's government procurement market.⁵⁷

Therefore, although the GPA is the most crucial trade agreement on government procurement under the WTO, its influence may be dwarfed by the increasing use of bilateral free trade agreements (FTAs) to access other members’ government procurement markets.

3.3 Lack of Legal Certainty

Furthermore, in practice, entity coverage negotiations and their implementation present extremely complicated difficulties due to the kaleidoscope of reciprocal arrangements between the Parties. Korea–Measures Affecting Government Procurement is a case in point.⁵⁸ The lack of legal certainty caused by the complexity of entity coverage schedules is likely to impair the commercial predictability (of whether a particular procurement is covered) under the GPA and set barriers to delisting a no-longer-qualified entity, which is not consistent with the original intention of the GPA.

Pursuant to the provision of ‘Modification and Rectification to the Coverage’, Parties can withdraw entities from Appendix I if the ‘government control or government influence’ is eliminated. Although the provision explicitly states that the privatization of the covered entities is a justification for withdrawal from the GPA obligation, withdrawal does not happen automatically because the Parties affected by the withdrawal proposal have the right of ‘objection’.

However, due to the absence of an expressed consensus on the standard for ‘effective elimination of government control or influence’, there are no specific criteria serving as the grounds for objection. As a result, modification could take an excessively prolonged period of time due to ‘objections’. For example, Japan's proposal to delist the East Japan Railway Company, the Central Japan Railway Company, and the West Japan Railway Company from Japan's Annex 3 took 13 years to become effective.⁵⁹

The prolonged process is not only unfair to the modifying Party but also has a ‘chilling effect’. The threat of a long and tortuous delisting process inhibits or discourages acceding Parties from providing a ‘more inclusive’ entity coverage offer, which is inconsistent with the built-in mandate to ‘achieve the greatest possible extension of the GPA coverage’.⁶⁰

From the perspective of the affected Parties, if the objection fails to prevent modification, the only remedy for the affected Parties is to retaliate by withdrawing equivalent coverage. This remedy can be obtained through an arbitration procedure but by no means has a ‘deterrence effect’. If the Parties conspired to abuse the toothless modification procedure, the proposing Parties and the affected Parties could both withdraw entity coverage, which will diminish the GPA coverage. At its extreme, annex modification is likely to fall into a vicious circle.⁶¹ The outcome would definitely be detrimental to the expansion of GPA coverage, let alone further growth.

These findings have significant implications for the provision of a general definition of ‘covered entity’. To avoid the drawbacks of the current entity coverage approach, a general definition of ‘covered entity’ can provide guidance for the expansion of GPA entity coverage.

4. The Rules Governing SOEs in the WTO Legal Regime

As discussed in section 2, what kind of government entities should be put in annex 3–other entity is one of the big concerns in accession negotiation, especially what kind of State-owned enterprises should be covered as ‘government entities’ under the GPA. WTO rules provide some guidance to some extent, but the existing regulations of SOEs in WTO trade agreements are inadequate in the new context of international economic order.

4.1 Article XVII of the GATT

The International Trade Organization (ITO) Sub-Committee's report emphasize that the objective of GATT Article XVII, addressing State Trading Enterprises (STEs), is to curb adverse trade effects.⁶² It aims to ensure STEs operate based on commercial considerations and in a non-discriminatory manner, preserving the value of negotiated tariff concessions. The Panel's statement of the aims suggests that GATT did not intend to establish a workable definition of State-Owned Enterprises (SOEs).

Following the ITO's failure, GATT's treatment of SOEs progressed. Article XVII: 1(a) of GATT 1994 categorizes State trading enterprises into three types: ‘State enterprises’, ‘Enterprises granted special privileges by the State’ (e.g., subsidies), and ‘Enterprises granted exclusive privileges’ (e.g., a monopoly). The article is not technically clear and formal. Instead, there are notes and explanations attached to the article providing an indicative list.⁶³ A ‘Panel on Subsidy and State Trading’ explains that the word ‘enterprise’ refers to either an instrumentality of government which has the power to buy or sell, or to a non-governmental body with such power, and to which the government has granted exclusive or special privileges. The rationale for the categorization lies in the principle of ‘State responsibility’, presuming that State enterprises, given their governmental authorization and powers, act as instruments of the State.⁶⁴ The principle is elucidated in the Understanding on the Interpretation of Article XVII of GATT 1994, which offers a functional definition of a State trading enterprise. This definition encompasses governmental and non-governmental entities, entities granted with exclusive or special rights, and those that direct the course of trade through their purchasing or sales activities.⁶⁵

Despite the importance of these provisions, case law regarding them is scarce within the WTO adjudicatory body. The WTO acknowledges a significant flaw in the efficacy of the entire rule, stating: Throughout the history of Article XVII, a notable gap has been the absence of a clear definition for State trading enterprises and State trading activities. Numerous attempts were made to formulate such a definition, all of which proved unsuccessful. This absence posed a considerable challenge in enforcing the transparency obligation outlined in Article XVII. How can a notification be made when the substance of what needs to be reported remains unclear? Consequently, it is probable that numerous State trading enterprises from various countries remained unreported for extended periods. Adding to this already problematic situation, very few GATT contracting Parties adhered to the annual notification requirement, even in cases where no State trading enterprises existed.⁶⁶

An incremental approach is observed, with specific issues addressed by the WTO adjudicatory body rather than a comprehensive guideline for all cases.⁶⁷ The case, US – Sonar Mapping, exemplifies the Panel's reliance on GATT provisions, particularly Article XVII: 2 (State trading enterprise) and Article III: 8(a) (non-discriminatory treatment), to determine if an acquisition qualifies as government procurement.⁶⁸ The Panel highlights the significance of ‘controlling influence’ in government procurement, examining factors such as payment source, use, ownership, and selection approval.⁶⁹ It refrains from defining ‘government entity’ but emphasizes the relevance of ‘government measures’ under the GATT, contrasting it with the GPA, which necessitates a clear definition of ‘government entity’ due to specific obligations on parts of government entities.⁷⁰ The GATT's focus on the nature of ‘government measures’ distinguishes it from the GPA, where the nature of the applying ‘body’ is crucial. This underscores the necessity of a definition of government entities for GPA application.

4.2 Article XXVIII of the GATS

The GATS does not explicitly mention State enterprises, State trading enterprises, or State-owned enterprises. The only concepts related to State-Owned Enterprises (SOEs) in the GATS are ‘monopolies and exclusive service suppliers’. In defining a ‘monopoly supplier of service’, Article XXVIII:(h) identifies three criteria indicating direct government involvement: government authorization, governmental establishment purpose, and being either a person (public or private) with member authorization or established by the member de jure/de facto in the relevant market. The definition of ‘exclusive service supplier’ in Article XXVIII:(h) also references ‘government authorization’ and ‘government purpose’. These factors aim to determine if the service provision aligns with non-discrimination obligations and market access commitments, not to define the purchasing entity.

The Panel's perspective on China–Electronic Payment Services supports this contextual interpretation, distinguishing a monopoly supplier as a sole supplier authorized or established formally or effectively by a Member, and an exclusive service supplier as one of a small number authorized or established by a Member, substantially preventing competition among them.⁷¹ This aligns with the interpretation that GATS does not intend to define the trading entity itself.

Apart from the lack of specific direct regulations on SOEs in the GATS, Article I: 3(b) of the GATS introduces an exemption from GATS scope for ‘services provided in the exercise of governmental authority’. According to Article I: 3(c) of the GATS, a ‘service supplied in the exercise of governmental authority’ is defined as a service provided by a supplier that operates neither on a commercial basis nor in competition with other service suppliers. Clearly, the core identifying factors for the exercise of governmental authority are the absence of a commercial basis and the non-competition condition. Building upon the insights from the adjudicatory report on China–Electronic Payment Services and Article I: 3 of the GATS,⁷² it is posited that the primary concern underlying the legal text is the ‘substantial impact on competition’. The causality between ‘government authorization and government purpose’ and ‘substantial impact on competition’ is apparent and indispensable.⁷³

WTO provisions and adjudicatory interpretations primarily focus on the discriminatory actions of the State rather than on those of the trading entity. The WTO framework, including GATT, GATS, and jurisprudence, has historically operated on the principle of ‘State responsibility’, treating trading entities as part of, or apparatuses of, the State in international trade. Those provisions were correspondingly made in 1993 when the SOEs underwent privatization as a worldwide fashion.⁷⁴ Thus, those trade rules match the SOEs at that time. However, after 2008, in the era of State capitalism, where SOEs operate independently in the market while still subject to government control, this approach faces criticism as it may not align with the diverse activities of modern SOEs.⁷⁵ Not all SOEs operate solely in accordance with public mandate policies. On the contrary, SOEs are more frequently seen undertaking solely commercial activities, or performing public policy functions while also engaging in commercial activities.⁷⁶

4.3 Article 1 of the SCM

The Agreement on Subsidies and Countervailing Measures (SCM) confronts the complex interplay between a public body and the concept of subsidies. Central to this is the interpretation of a ‘public body’ within Article 1 of the SCM, which defines a subsidy as a financial contribution by a government or any public body within a WTO member's territory. This definition is pivotal in determining whether subsidies issued by State supporting entities, such as State-owned banks or financial companies, fall under the SCM's regulatory purview. Notably, the negotiation history of the SCM demonstrates a deliberate introduction of ambiguity into the text, reflecting the reluctance of negotiators to commit to clear and certain meanings, thereby avoiding unwanted obligations.⁷⁷

This ambiguity has led to varied interpretations in WTO dispute settlements. For instance, in the US–Anti-Dumping and Countervailing Duties (China) case, the US Department of Commerce's stance, based on the ‘ownership approach’, suggested that government ownership or majority state shareholding in an entity sufficed to establish it as a ‘public body’.⁷⁸ However, the Appellate Body and Panels in various cases, including EC–Countervailing Measures on DRAM Chips and US–AD & CVD, have rejected the notion of using ownership as the sole criterion for recognizing a ‘public body’.⁷⁹

The Panels have developed a more nuanced approach, asserting that an entity must not only be government-controlled but must also exercise vested government authority or act as an immediate result of government control or influence to be considered a public body under Article 1 SCM. This approach necessitates two concurrent requirements: the entity must possess a core public feature (such as formal public legal status or vested government authority) and must exercise this authority in the case at issue.⁸⁰ The Appellate Body has taken a cautious approach in defining ‘public body’, avoiding an overextension of government boundaries. This is evident in the US–Countervailing Duty Investigation on DRAM case, where the Body posited that actions of State-owned corporate entities are prima facie private and not presumptively attributable to a Member under Article 1.1 of the SCM Agreement.⁸¹ This perspective marks a progression from GATT and GATS, emphasizing the commercial character of SOEs.

The Korea–Commercial Vessels case highlights the complexities in interpreting the notion of a ‘public body’. The European Community argued that the Korea Export-Import Bank (KEXIM) was a public body, citing its legal status, operational purpose, and access to state resources. Conversely, Korea contended that KEXIM, operating in a competitive market and fulfilling commercial functions, should not be considered a public body. The Panel ultimately stated that an entity is a public body if controlled by the government or other public bodies, rejecting the notion of an entity having a dual public–private nature based on its activities.⁸² The Panel's decision underscored the importance of legal certainty, rejecting the notion that an entity could oscillate between public and private statuses based on its market activities. However, the Panel did not elaborate on how government control should be determined, leaving a gap in the understanding of what constitutes a public body.

Certainly, neither of the two approaches presented in the European Community vs Korea case proves exhaustive enough to formulate a definition that encompasses all public bodies while ensuring the exclusion of entities incapable of implementing measures adversely affecting trade. However, when amalgamated, these two perspectives offer a more holistic understanding of a ‘public body’. The European Community's criterion contributes a formalist methodology for delineating the static features of a public body, while Korea's approach provides a functional perspective to identify the specific kind of public body subject to trade disciplines concerning non-discrimination and trade liberalization. Despite the Panel's rejection of the notion that the functional approach, in the context of the SCM Agreement, can justify any entity, it does have implications for defining a ‘government entity’ within the GPA.

4.4 The WTO Appellate Body's Approach to SOEs: Navigating the Intersection of Public Policies and Commercial Operation

The WTO Appellate Body has faced the intricate task of interpreting the conduct of SOEs within the complex framework of international trade law. This challenge is particularly evident in the context of balancing commercial considerations with public policy mandates. The following analysis delves into this nuanced area, exploring the intersection of competition policy with public sector companies and the evolving role of SOEs in international trade.

4.4.1 Competition Policy and Public Sector Companies: A Delicate Balancing Act

In the appeal hearing of the Canada–Wheat Exports and Grain Imports Board case, Australia and China supported the Panel's interpretation of ‘commercial consideration’, and China further pointed out that gaining market share is a better indicator of commercial considerations than replacement value. The European Commission (EC) concurred with the Panel‘s findings and further specified that interpreting the term ’commercial consideration‘ solely requires an assessment of the ’market behavior of State Trading Enterprises’, aligning it with standard private sector practices.⁸³’ The interpretation of ‘commercial considerations’ and ‘acting commercially’ in the context of State trading enterprises, and the Body's decision to avoid extending its analysis to encompass comprehensive competition-law-type obligations as seen in the Canada–Wheat Exports and Grain Imports case, highlights the Appellate Body's cautious approach and reflects a nuanced understanding of the unique nature of public sector companies.⁸⁴

Although the Panel and the Appellate Body were trying to prevent the introduction of competition concepts, the debate on the term ‘commercial considerations’ illustrates the role of competition policy as it applies to public sector companies. This is a recurring issue within the WTO. It extends beyond State trading enterprises, influencing the interpretation of a ‘public body’ under the SCM Agreement and of a ‘government entity’ under the GPA.⁸⁵ The Appellate Body's approach in these instances illustrates an effort to harmonize competition policy principles with the operational realities of public sector entities. To put it simply, the WTO Appellate Body's approach to the issue of SOEs reflects a sophisticated understanding and a delicate balance between public policy objectives and commercial operations. It is crucial in the evolving landscape of international trade, where the lines between public policy objectives and commercial operations of SOEs are increasingly blurred.

4.5 The Approaches of SOEs’ Regulations in the New Landscape

The rise of nationalist and protectionist sentiments, particularly in Western countries, has made it increasingly difficult to reach a consensus in multilateral trade negotiations. This is evident in the renegotiation of NAFTA into the USMCA, where the United States sought more favourable terms based on nationalist interests. Due to rising protectionist and nationalist policies, there is a growing scepticism towards multilateral trade agreements, and plurilateral agreements, such as the RCEP (Regional Comprehensive Economic Partnership) and the CPTPP (Comprehensive and Progressive Agreement for Trans-Pacific Partnership), and bilateral arrangements have gained prominence as viable alternatives to traditional multilateral trade agreements and significantly influenced the governance of modern State-Owned Enterprises (SOEs), marking a pivotal shift in international trade norms.

Recent FTAs have increasingly incorporated provisions targeting SOEs within their ‘fair competition’ chapters. The European Union, for instance, has been at the forefront of advocating for ‘fair competition criteria’ to regulate the trade and investment activities of SOEs. This approach aims to mitigate the competitive distortions caused by the unique advantages that SOEs may possess, such as special powers or privileges. For example, the US–Australia FTA addresses State enterprises under its competition policy, mandating their operation in a manner that does not hinder trade.⁸⁶ Similarly, the US–Korea FTA includes provisions under competition-related matters, outlining specific powers that could obstruct free competition.⁸⁷ The EU–Singapore FTA, under its competition chapter, stipulates that entities with special or exclusive rights must not engage in anti-competitive practices.⁸⁸ The Singapore–Australia FTA further emphasizes ‘competitive neutrality’, ensuring no undue advantages are granted to any SOE.⁸⁹

4.5.1 Functional Approach: A Contemporary Trend in FTA Provisions

The United States has traditionally adopted an ownership-centric approach to define and regulate State-Owned Enterprises (SOEs), evident in agreements such as the US–Singapore FTA.⁹⁰ The EU, in agreements such as the EU–Vietnam Trade and Investment Agreement, follows a similar pattern by defining SOEs based on ownership, control over board appointments, or influence on strategic decisions.⁹¹ However, the ownership approach has faced criticism for its practical limitations.⁹² Scrutinizing ownership structures can be challenging for enterprises, leading to ambiguities in identifying SOEs and hindering foreign market entry and compliance strategies. Transparency issues persist, as governments can modify state ownership without immediate disclosure, potentially bypassing rules applicable to the government sector.⁹³

Contrasting earlier FTAs with those post-2008 reveals a growing preference for a functional approach in regulating SOEs. This approach eschews a rigid public–private dichotomy, focusing instead on the nature of an entity's impact on competition. Under this paradigm, the classification of an entity as public or private becomes secondary to its competitive behaviour.

The CPTPP and the USMCA have adopted a consistent definition of SOEs, which is also mirrored in the EU–Vietnam Trade and Investment Agreement.⁹⁴ In these agreements, SOEs are subject to trade disciplines specifically when they engage in commercial activities driven by commercial considerations. This dual requirement – engaging in commercial activities based on commercial considerations – serves as a threshold for applying trade disciplines to SOEs. It effectively confines the scope of regulation to scenarios where SOEs’ transactions exhibit anti-competitive behaviour. This approach marks a significant shift from traditional models of SOE regulation to a more functional and pragmatic approach in regulating SOEs, focusing on the nature and intent of the activities rather than merely the ownership structure of the enterprises.

4.5.2 Harvest Approach: A Progressive and Pragmatic Step

Following the discussion on the CPTPP, USMCA, and EU–Vietnam agreements, it is pertinent to examine the EU–China CAI's approach to SOEs. The CAI notably diverges from explicitly defining ‘SOEs’ within its text. Instead, Article 3bis introduces the concept of ‘Covered Entities’, encompassing a broader range of organizations, including state-controlled enterprises, private monopolies, and state-subsidized entities. This strategic avoidance of the specific term ‘SOE’ does not diminish the agreement's regulatory reach over such entities. Like the CPTPP and USMCA, the CAI subjects these covered entities to the principles of non-discriminatory treatment and commercial consideration in their commercial activities.

By not confining itself to the traditional concept of SOEs, the CAI expands its regulatory scope to encompass a wider array of entities that could potentially engage in anti-competitive actions, irrespective of their public or private legal status. This broadened scope is a significant departure from the more traditional approaches seen in other FTAs. The CAI's approach is reflective of what the author describes as a ‘harvest approach’, which deepens and refines the regulation of SOEs seen in recent FTAs. This approach is threefold:

1. 1. Legacy of the WTO: The CAI acknowledges the relevance of ownership and control criteria in determining a ‘public body’ but does not consider these factors as decisive. This stance aligns with, yet goes beyond, the WTO framework.

2. 2. Focus on Competition: The CAI emphasize the criterion of competition to delineate the scope of entities considered under its regulations, moving away from a strict ownership-based model.

3. 3. Functional Approach: By abandoning a formal definition of SOEs, the CAI adopts a functional approach, expanding discipline to a broader range of entities. This approach is particularly significant in the context where SOEs may differ culturally and economically. The CAI's approach captures entities based on their trade behaviour's competitive impact, offering the flexibility to incorporate a diverse array of entities from various cultural and economic backgrounds into international trade regulation.

By focusing on the functional impact of entities in the market, the CAI offers a flexible and inclusive framework that could potentially serve as a template for future international regulations, including the GPA entity coverage negotiation. This approach is particularly relevant in addressing the complexities of modern global trade, where the distinction between public and private entities is increasingly blurred, and the need for adaptable regulations is paramount.

5. A Proposal for a General Definition of the ‘Covered Entities’

At the time of the OECD discussions on government procurement codes, the EEC (EU as of 2008) suggested a general definition for use in its public procurement directives, which referred to ‘the State, the regional and local authorities, and other legal persons constituted under public law’.⁹⁵

Regrettably, the OECD Code on Government Procurement did not adopt this suggestion. With reference to a list of agreed entities, the following was added to Section I (2): ‘These entities are the national government agencies, or the entities entirely or substantially controlled by national governments in respect of their purchasing policies.’⁹⁶ The subsequent GPA negotiations take the list approach following the OECD Code.⁹⁷

The Tokyo Round Code in Article I (1) (c), which is similar to the language used by the OECD Code, states that ‘This Agreement applies to … procurement by the entities under the direct or substantial control of Parties and other designated entities, with respect to their procurement procedures and practices.’ Article I of the GPA 1994 states, ‘This Agreement applies to … any procurement by entities covered by this Agreement, as specified in Appendix I.’ The revised GPA Article II states that ‘covered procurement means procurement for governmental purposes … by a procuring entity and that is not otherwise excluded from the coverage schedules of each Party's annexes to Appendix I’.⁹⁸

Although the previous GPA negotiations and revisions do not literally provide sufficient criteria for an entity to be subject to the GPA, those documents do establish the underlying doctrines to define ‘covered entity’.

5.1 The Implied Consensus on the Definition: Findings from the Annexes of the GPA Parties

5.1.1 Control Doctrine

In the GATT Tokyo Round Code on Government Procurement and its revised amendments, the definition of a government entity was based on the ‘control doctrine’. The footnote to Article 1 of the Tokyo Round Code states that ‘entities include agencies, no matter whether the agencies had independent legal personality or not’.⁹⁹

Article I:1(c) of the Revised Code states that government procurement refers to procurement by ‘the entities under the direct or substantial control of Parties and other designated entities, with respect to their procurement procedures and practice … the entity coverage listed subject to further negotiation’.¹⁰⁰

The ruling of the dispute US – Sonar Mapping concluded that the wording of Article I: 1(a) suggests that the entity shall have ‘some form of controlling influence over the obtaining of the product such as payment by the government, government use of or benefit from the product, government possession, etc.’¹⁰¹ Accordingly, in fact, the ‘control doctrine’ has been established as a substantial constituent of the definition of any covered entity, and the formal identity (such as a legal personality under public law) of the purchaser has no relevance.

Although a list approach has completely replaced the general ‘control doctrine’ since the GPA 1994, each Party’ entity coverage has de facto applied the ‘control doctrine’.¹⁰² For example, Armenia's Annex 3 included three groups of legal persons governed by public law: State or community non-commercial (non-profit) organizations; commercial organizations with over 50% of government or community shareholding; and legal persons providing public service.¹⁰³ Ukraine employs similar descriptions in the footnote to its Annex 3. It states that a covered entity must meet one of the following criteria: ‘the public authorities, authorities of Crimea, local self-governance bodies or other administrators of public funds have more than 50% in its authorized capital, or a majority of the votes in its supreme governing body, or have the right to appoint more than half of its executive or supervisory body, or in possession of special or exclusive rights granted by the authorities’.¹⁰⁴

The criteria of ‘control’ from the aspects of ownership, financial source, and managerial and personnel influence reflect the common denominator in the understanding of ‘government control’ and should be included in the definition of a covered entity.

5.5.2 Competition Doctrine

‘Controlled by government’ is the positive factor of ‘covered entities’. As indicated by the Parties, procurement by entities characterized by ‘commerciality/competition’ should not be subject to the GPA. For example, Armenia's Annex 3 clearly states that a covered entity shall be a non-commercial/non-profit legal person.¹⁰⁵ Moldova, with the same intention, states in its Annex 3 that the entities should be legal entities governed by public law and ‘not having an industrial or commercial character’.¹⁰⁶ In addition, there are also Parties that make similar exceptions in the notes after the covered entities. The EU, Liechtenstein, and Montenegro state in the notes that even though an entity is generally a public body, if its procurement in pursuit of listed activity was exposed to competitive forces in the market concerned, the public body shall not be covered by the GPA.¹⁰⁷ Similarly, Norway states that when directly exposed to competition in the market to which access is not restricted, the procurement by the entity shall not be subject to the GPA.¹⁰⁸

The competition doctrine is inherently consistent with the general aims and purpose of the GPA. If any evidence indicates that the ‘public controlled’ entity is committed to a policy of open, fair, and non-discriminatory procurement and, in practice, had procured goods or services in a market where competition is indiscriminately open to any suppliers, irrespective of nationality, the entity should not be subject to the GPA because that entity does not restrict competition opportunity.

Although the ‘control doctrine’ and ‘competition doctrine’ provide us with two guiding principles to recognize a ‘covered entity’, the article will not stop and start again to define a ‘covered entity’.

5.2 A General Premise for the Definition of ‘Covered Entity’: The Line between the ‘Public’ and the ‘Private’

As previously mentioned, the EU proposed a general definition for covered entities during the negotiations for the OECD code on Government procurement.¹⁰⁹ The proposal rests its suggestions on a clear public‒private dualism.

Although the proposal was not accepted, it is worthwhile to stress that the line between the public and the private is blurry and that a definition of ‘covered entity’ should avoid relying on the ‘public law’ status. In the late 1930s, Carl Schmitt observed that there are two great dualisms in modern legal systems: the dualism of international law and domestic law and the dualism of public law and private law.¹¹⁰ The two dualisms have been criticized for inconsistency with the context of globalization and its inability to provide a comprehensive explanation for why public bodies, such as governments, have become increasingly involved in economic activities, especially in the public service market through cooperation with the private sector.¹¹¹

The more interconnection and interdependencies there are between the public and the private, the greater the possibility of breaking the wall between the public and the private. Categorizing matters as either ‘public’ or ‘private’ in nature may be partly attributable to the rise of the theory of ‘State’ in the nineteenth century.¹¹² However, this legal approach became contestable with the rise of State capitalism after the financial crisis in 2008. That year witnessed the prominent resurgence of State investment in enterprises, especially in developing countries. As a result of the expansion of the State's role in world trade and investment, a number of developing countries have increasingly invested in and controlled SOEs rather than privatizing them. Therefore, SOEs operate in a commercial way and are transformed into subjects of private law, despite their legal status in public law. While traditional SOEs with prominent levels (over 50%) of government ownership and correspondingly high levels of control still exist, public and private mixed ownership/control prevail as the dominant ownership/governance structure of SOEs.¹¹³ Despite the differences among SOEs, SOEs are generally regulated under public law, regardless of their features of transcendence.

In this sense, a clear division between the public and the private is obsolete. Consequently, the regulation of public‒private mixed entities (especially some commercial SOEs) without a black-and-white line between the public and the private is truly relevant in GPA accession negotiations.

5.2.1 The Regulation of Public‒Private Mixed Entities in the WTO Context

As a part of public law (international economic law), WTO Agreements have specific provisions dealing with the trade practices of public bodies or their equivalent bodies, such as State-owned trading enterprises, State monopolies, and public bodies in GATT Article XVII, GATS Article XVIII: (h) or SCM Article 1.

Owing to the generality of those articles as well as the complexity and novelty of current world economic activity, the WTO adjudicatory body has also tried to explore and explain the meaning of those notions by distinguishing the relationship between public and private. What can be clearly seen in those efforts is that attempts have been made to attribute the responsibility to a non-State actor as an instrumentality of the State by referring to the principle of State responsibility in Article 5 of the International Law Commission, although it is also recognized that Article 5 does not attempt to identify the scope of ‘government authority’.¹¹⁴

More interestingly, it is recognized that ‘substantial impact on competition’¹¹⁵ is the ultimate concerning factor in considering whether a private body should be subject to WTO discipline. One of the main difficulties with the line of reasoning in the above articles is that the WTO agreements do not clearly define public‒private mixed entities that carry features of both sides (the public and the private), such as SOEs, nor does the adjudicatory body have the competence to provide a reference for identifying public‒private mixed entities. Nevertheless, those efforts demonstrate the intriguing interconnection and interdependencies between the public and private sectors in international trade. In view of this, the discipline of public‒private mixed entities is better for not adopting a rigid formalism approach based on its legal status.

5.2.2 A Functionalist Understanding of SOEs

A definition of ‘covered entities’ should be based on the common denominator of the current coverage schedules to win political acceptance by the GPA Parties.¹¹⁶ The common denominator reflects the consensus of the GPA Parties, namely, the ‘control doctrine’ and the ‘competition doctrine’, as concluded in the previous section.

In addition, the definition should be inherently consistent with the aims and purposes of the GPA/WTO. An equivalent of Article I and Article III of the GATT can be found in GPA Article III, entitled ‘National treatment and non-discrimination’.¹¹⁷ Besides, the recital of the agreement is crucial, as it usually states the aims and purpose of the agreement. The recital of the GPA clearly states that the agreement is ‘with a view to achieving greater liberalization and expansion of, and improving, the framework for, the conduct of international trade’. These words expressly indicate the goal of gaining market access for further trade liberalization.

It is under this general premise that legal interpreters of Korea–Procurement have inferred that an entity should not be subject to the GPA rules if it has been privatized and is without government control or has effectively eliminated government influence.¹¹⁸ An inclusive definition should try to capture all the entities that could jeopardize the liberalization of government procurement, while at the same time exempting those entities that have no adverse effect on free competition. Therefore, the definition is supposed to potentially deepen the integration of the international government procurement market.

In such a sense, the EU Public Procurement Directives are a good case at the point of positive integration of government procurement regulations and therefore can provide experience to borrow. In the EU Public Procurement Directive, the ‘contracting bodies’ consist of three categories: ‘government/public authorities, public undertakings, and other undertakings granted special or exclusive rights operating in the utilities sectors’. In addition to clear definitions in the Directive, the Court of Justice of the European Union (CJEU) has vigorously interpreted those definitions with a functional approach on a case-by-case basis under the general aim and purpose of eliminating market barriers between Member States and promoting competition in the EU public market. Clear definition together with the functional interpretation of the definition of purchasing bodies under EU public procurement law enhances the integration process of the EU public procurement market.

Unlike the EU, which has developed integration among Member States, the WTO itself has no authority to take initiatives in trade policy-making or trade agreement negotiation. The Dispute Settlement Body does not have the same competence in law-making as the EU legislators do. It can only rely on treaty texts, customary international law, and the general principles of international law for literal interpretation but is not allowed to actively harmonize its members’ trade policies or establish the meaning of the rules by purposive interpretation.

Therefore, on the one hand, it is imperative that the GPA Parties can introduce a definition of the ‘covered entity’ reflecting their political will and, on the other hand, take a functional approach to incorporate the ‘control doctrine’ and the ‘competition doctrine’ into the structure of the legal definition. Otherwise, with the current list approach, entity coverage is under-included because some public SOEs are not covered, while with a functionalist approach, the ‘competition’ assessment can exclude commercial SOEs from GPA obligations, avoiding the risk of over-inclusiveness.

In the following section, the author will explain how to incorporate the two doctrines into a rule for establishing a definition.

5.3 The Legal Structure of the Definition: General Scope and Escape Clause

In the above sections, the author explained the premise that should be kept in mind when making a definition as well as the two guiding principles governing the definition process. Then, we move to the following problem: what should the definition look like? Namely, as a rule to be suggested for the GPA, what is the legal structure of the suggested definition?

5.3.1 Harmonizing Approach

Generally, there are three WTO rule-making approaches to bridge divergences among the members.

If the divergence is over technical standards or a procedural issue, WTO law will take a harmonizing approach,¹¹⁹ since these problems need minimum standards for the application of an integral rule. This creates a prominent level of efficiency and guarantees legal security in international trade. For example, the TBT (Agreement on Technical Barriers to Trade) Committee has developed a set of detailed recommendations and decisions regarding the notification of regulations, procedures for assessing conformity, and mechanisms for responding to information provided and requests regarding domestic regulatory programs.¹²⁰ Similarly, the SPS (Sanitary and Phytosanitary Measures) Committee has been developing a set of procedures to enhance the transparency of special and differential treatment in favour of developing countries.¹²¹

The harmonization approach aims to produce an integrated rule that applies with fewer exceptions. Despite the advantage of thoroughness by means of harmonization, international rule-making, of which the WTO law is part, has to respect sovereignty, and thus, it is not always easy to obtain across-the-board agreement on legal harmonization. If the divergence is immediately related to a specific trade commitment (for example, tariff reduction), negotiators implicitly or explicitly rely on an assessment of national market forces to ensure that markets are nationally beneficial. In this situation, a second approach is preferable to moderate divergences between countries, such as bilateral trade agreements or regional trade agreements. Continuous bilateral ‘trades’ or ‘swaps’ generate trade creation in more extensive trade areas (where there is a lack of universal consensus in the overall WTO community), as a result of which it is possible for the divergences between signatories to be significantly narrowed.

Statistics prove that FTAs successfully lower tariff rates and reduce by 30–60%non-tariff barriers in manufacturing industries through the employment of bilateral tariffs and non-tariff barriers.¹²² Moreover, FTAs are often signed by a smaller number of countries that share more extensive common interests. The extensive common concerns or interests enable a stronger sense of ‘community’, which lays the foundation for a broader consensus on tariffs, non-tariff barriers (including discriminatory government procurement policies), intellectual property, labour standards, e-commerce, and so forth.

Particularly in cases of divergences in the GPA's expansion of its coverage, interestingly, trade commitments made by GPA Parties under bilateral trade agreements are usually more generous than those made under the GPA. The history of GPA negotiation indicates a lack of consensus, and a lack of interest in even a general harmonization in the field of government procurement, and its plurilateral nature will not change in the near future.¹²³ However, a number of FTAs have included provisions for more flexible or advanced government procurement procedures¹²⁴ or even further trade commitments. For example, the EU has also signed several bilateral agreements providing more extensive coverage on government procurement than it dictated under the GPA.

5.3.2 Buffering/Escape-Clause Mechanism

In contrast to the general harmonization approach, the second approach tries to obtain any possible consensus in a much smaller group, which is less ambitious but very practical. If general harmonization is the ultimate goal, the second approach is preferable as a starting point for the achievement of further extended consensus in the future. Those consensuses in FTAs indicate that there could be improved rules or criteria that help to broaden the degree of harmonization in the international government procurement market; at least practically, the trade commitment made in bilateral agreements proves that they are willing and able to offer broader market access.

If consensus is rare and the divergence is difficult to reconcile (for example, in intractable trade topics, such as e-commerce and agriculture), buffering/escape-clause mechanisms are useful for rule-making. This approach is based on the recognition that the world economy always consists of different economic systems. In order to push nations towards greater cooperation, the interface between different economies could be realized by institutional or legal means, seeking common ground while reserving differences.¹²⁵ Where there is full consensus, a general applicability of rules would be envisaged, while where there is substantial divergence, there would be an escape clause.

The buffering or escape-clause mechanisms allow ‘derogations’ from the general WTO obligation of non-discrimination on the basis of a legitimate domestic policy.¹²⁶ Apart from those justifiable exemptions, the WTO obligation should generally apply. Those exceptions could be included in a general rule as a justification (not falling within the scope of application of a general obligation). The buffering mechanism accommodates the diversities of national economies and thereby is more acceptable by WTO members than the other two approaches.

5.3.3 The Mixed Approach for GPA Entity Coverage: General Scope and Escape Clauses

The above three approaches are, of course, not mutually exclusive. In the case of the GPA, a general harmonization has been realized in some aspects of the procedural rules, such as tendering procedures and other transparency notification obligations (since it is required that every acceding Party ensures that its domestic government procurement law is consistent with the GPA rules).¹²⁷

With regard to the coverage and scope of the GPA, there are always divergences on whether harmonized criteria on the sub-government entities and ‘other entities’ should be adopted.¹²⁸ As previously mentioned, in the bilateral agreements between WTO members – for example, the bilateral agreements between the EU and Singapore, Canada, the US, etc. – the coverage of sub-government entities and other entities has been further expanded.¹²⁹ It is argued that the expanded entity coverage in bilateral agreements concerning the government procurement market represents the possibility of further harmonization on the issue of entity coverage among the GPA Parties.

Although it would be unlikely in the immediate future to achieve a multilateral consensus on the liberalization of the international government procurement market,¹³⁰ the GPA entity coverage could achieve further expansion by adopting a definition with a general scope and an escape clause. For the reason that there are divergences in the coverage of ‘other entities’ among the GPA Parties and disparities typically in the regulations of SOEs and their equivalents between the WTO members, a general rule can assume that all publicly controlled or influenced entities are presumably covered by the GPA, with an escape clause that the entity can rebut this presumption if it can prove that, despite being under public control or public influence, its procurement activities are exposed to effective market competition. This escape clause or justification could accommodate new members with large state sectors, such as China, Poland, Russia, and Vietnam, or large sovereign wealth funds, such as the UAE or other similar kinds of large public‒private mixed entities. Therefore, a primary conclusion is that the legal structure of the definition of a ‘covered entity’ consists of two parts: a general scope clause and an escape clause.

5.4 The Primary Function of the Definition: Trade Rules or Competition Rules?

In addition to the legal structure, the next fundamental question worth asking is whether the primary function of the definition of the ‘covered entity’ serves the purpose of eliminating trade barriers, promoting competition in the field of the government procurement market, or both. The question is fundamental because there should always be a legislative purpose to guide rule drafting.

Whether government procurement rules are ‘trade rules’ or ‘competition rules’ is conceptually and practically relevant. First, trade-related rules seek to remove barriers to trade and safeguard trade opportunities, while competition-related rules focus on avoiding competition distortion and protecting the competitive process in the market.¹³¹ From a practical perspective, a definition with the effect of promoting competition inherently requires an assessment of the impact of the competition particular markets, whether potential or actual, and therefore, it demands a more refined and sophisticated set of criteria. These differences have implications for the means and mechanisms employed by the two kinds of rules.

The EU public procurement rules are a good example to demonstrate the difference between trade rules and competition rules. As an integrated part of the Internal Market rules, the EU public procurement regulations serve to establish a ‘public market’ within the EU; accordingly, the EU public procurement rules primarily safeguard market access.¹³² The definition of ‘contracting entities’ (including ‘contracting authorities’, ‘the bodies governed by public law’, ‘public undertakings’, and ‘undertakings with special or exclusive rights’) has properly encompassed entities that could set trade barriers. A large amount of CJEU jurisprudence has established a broader understanding of this concept to encompass all entities that could hinder the integration of the EU Internal Market.¹³³ Therefore, overall, the EU public procurement rules are trade rules.

More obviously, the WTO rules are trade rules to ensure equal competition opportunities (in a liberalized trade market). Article III GATT represents the most important legal reference for securing this ‘equal competition’.¹³⁴ The GPA, as a set of regulations of the non-tariff barriers, is intended to achieve the general aim of ‘equal competition conditions’. Therefore, the definition of the ‘covered entity’ under the GPA should certainly remain compatible with the aim of general trade liberalization, as was expressed in a mandate to expand the entity coverage of the GPA.¹³⁵

On the basis of the above discussion on trade rules, theoretically, the GPA could expand its entity coverage to all entities that are able to set trade barriers, as the EU public procurement regulations do. In the context of this research, the ‘theoretically expansive’ entity coverage is exactly defined as the ‘general scope’ of the proposed definition of the ‘covered entity’, which presumably includes all entities that set trade barriers in the international government procurement market.

5.4.1 The Interplay between Trade Rules and Competition Rules

Furthermore, the GPA/WTO and the EU public procurement regulations are at the interface of trade rules and competition rules. Trade policies and competition policies should be inextricably linked as required to promote the equality of competitive opportunities for WTO members, and the boundaries between the two have to some extent merged.¹³⁶ As a result, trade liberalization policy and general competition policy usually coexist, and the convergence of trade rules and competition rules in the two legal systems is increasing.¹³⁷ The consensus on the interaction of trade policy and competition policy within the multilateral trading system has been expressed in many contexts¹³⁸ and provides a basis for the GPA to incorporate a competition legal technique in the definition of the ‘covered entity’.

With the aim of maintaining an integrated ‘public market’ within the EU, EU Public Procurement Directives borrow from EU competition law in defining a ‘contracting entity’, particularly the explanation of ‘bodies governed by public law’, ‘public undertakings’, and ‘undertakings granted with special or exclusive rights’, operating in the utilities sectors. These definitions elaborate detailed criteria on whether the procuring entity is under government control and detailed tests to discern whether the control could lead to a loss of commercial freedom to the extent that the entity may not make a procurement decision on the basis of considerations of competition. These criteria of public control are specifically derived from managerial, financial, or personnel aspects, and the distortion of competition refers to the EU competition rules.¹³⁹

Both the WTO and the EU legal systems maintain an aim to promote trade liberalization among their members; moreover, there are many signs of the convergence of trade rules and competition rules in both the WTO and the EU legal systems.

Therefore, the GPA can borrow the experience from the EU by making a reference to competition rules. For example, in the EU context, both Internal Market rules and competition rules are convergent under the goal of liberating the economic forces in the Internal Market. Moreover, the aim of creating an Internal Market constitutes a unifying thread or, at least, an interface between EU internal trade law and EU competition law.¹⁴⁰ This conclusion also applies to the GPA. In the neoliberal context, trade liberalization and the principle of non-discrimination are compatible with the promotion of competition. Trade rules and competition rules could be applied consistently for the same purpose without conflicting with each other. Thus, the GPA/WTO should take the EU public procurement rules as a model to incorporate competition rule techniques where appropriate.

In the next section, the elements of competition law will be introduced into the GPA definition of ‘covered entity’ to avoid the over-inclusion of entities that compete under normal market conditions, thereby avoiding the imposition of trade barriers in the international government procurement market.

5.5 The Conceptual Framework of the Definition

A conceptual framework for the identification of features and characteristics of ‘covered entities’ in procurement activities in the WTO legal system offers a benchmark for outlining the features of a government-controlled, non-competing entity that can serve as a reference for accession negotiations.

5.5.1 The General Scope: Public Control

To date, there is only one WTO appellate body report on trade measures affecting ‘government procurement’. This adjudicatory report raised questions about the coverage of an entity under governmental control and provided a reference for clarifying the notion of a ‘government entity’ in the GPA.

On 17 March 1999, the United States requested that the Dispute Settlement Body establish a panel to examine certain procurement practices of entities concerning the procurement of airport construction for Inchon International Airport (IIA) in Korea, pursuant to Article XXII of the GPA. Three entities were responsible for the IIA procurement: KAA, KOAC, and IIAC.¹⁴¹ One of the debated issues was whether the entities that had borne procurement responsibility for the project since its inception were ‘covered entities’ under the GPA.¹⁴²

The US argued that a ‘control’ test must be applied to determine whether the unlisted entity should be subject to the Agreement. If the unlisted entity was controlled by a listed entity, the GPA rules must apply. ‘Direct or substantial control’ is intended to encompass not only governmental entities but also quasi-governmental purchasing agents.¹⁴³

Korea contested the proposed test, arguing that there was no normative rule relating to direct control, either in the Korean Schedule or in GPA provisions. The question was further narrowed to whether there were some criteria exogenous to the Schedules that could serve as a normative rule for direct guidance on what is covered by a GPA signatory's commitments and for the scope of legitimate expectation.¹⁴⁴

The Panel did not adopt the US position that a ‘control test’ should be read into the GPA. However, the Panel accepted that ‘control’ of one entity over another could be a relevant criterion for determining the coverage of the GPA and could also be relevant to the determination of which is a ‘governmental entity’.¹⁴⁵

It was argued that applying a broadly defined ‘control’ test to expose the control structure of the entity could result in the Parties being unable to access sufficient evidence to substantiate a claim of violation under the GPA obligations. Transparency would be a further problem. In addition, if the judgement of ‘control’ was made by referring to national laws, Parties could be burdened with disputes over the understanding or interpretation of rules. Thus, if ‘control’ criteria were included in the GPA as one of the determinants but not determinative of GPA coverage, the criteria must be clear and avoid normative notions that are too abstract.

A government has a variety of ways to influence or control an entity, and no two governments exercise that influence or control in the same way. Thus, a multi-dimensional framework is relevant in demonstrating the existence of effective government control over an entity while avoiding a focus on one single core characteristic at the expense of neglecting other relevant characteristics. In contrast, a formal approach that regards one specific dimension, such as formal legal status or ownership, as essential to identifying public control is not reliable, as the absence of an express legal status or statutory delegation of government authority does not necessarily preclude a finding that an entity is a public body or government body.¹⁴⁶

There is a minor difference in the criteria employed to identify the existence of public control in the approaches of the US, the EU, UNCITRAL, and the WTO Committee on Government Procurement. The government procurement regulations in the four above-mentioned approaches (namely, those of the US, the EU, the GPA, and the UNCITRAL) generally identify the existence of government control in the form of two tiers:

1. (1) The external tier. This includes legal and formal constraints over the entity, namely, legal statutes, hierarchical or bureaucratical supervision by governments, public financing, and provision of public goods.¹⁴⁷

2. (2) The internal tier. This includes the freedom of management and appointment of personnel. Although all four approaches cite these two factors, only the EU has elaborated further details. For example, in both the definition and in the relevant CJEU jurisprudence, the EU refers to a ‘body governed by public law’ and a ‘public undertaking’ operating in the utilities sector and stipulates that half of the members of the administrative, managerial, or supervisory board are appointed by a public authority. ‘Public control’ is easily recognizable among the above factors, but under the GPA obligation, the simple existence of such control does not suffice; it must also be demonstrated that the causal link between the control and the outcome of loss of commerciality prevents competition under normal market conditions.

The author wishes to emphasize the significance of this point, especially in cases where recognition of the coverage of public SOEs and commercial SOEs under the GPA obligation is in question. Both kinds of SOE match the characteristics of the external tier and the internal tier. However, it is not enough to distinguish their obligation under the GPA based only on the ‘control test’. With regard to entity coverage, it is necessary to conduct a further test to see whether the control is ‘effective’ to the extent that the entity cannot compete under normal market conditions (presumably not setting trade barriers); therefore, it must be excluded from the GPA application.

5.5.2 The Escape Conditions: Assessment of Competition

Government control/invention often has a positive effect only when the control/invention promotes competition instead of obstructing it. Therefore, the competition paradigm, as an accredited proxy, can be used to examine the economic effects of public or political influence over an entity. It is especially useful as a central criterion in assessing whether public control or influence is so extensive or effective that it directs the entity to apply discriminatory or preferential practices in government procurement activities as governments do.

Taking ‘competition’ as a benchmark, if an enterprise competes in a contestable market, it is unlikely to afford the luxury of procuring intermediate service inputs from any but the most competitive seller. Thus, anon-competitive market would seem to be a necessary condition for a distorted procurement decision. The distortion could have two sources: de jure distortion, namely, through government regulatory policies, such as granting special or exclusive rights in the relevant market, or de facto distortion, namely, anti-competitive practices on the part of enterprises enjoying significant market power.

In regard to the source of de jure distortion, the exclusivity arrangement has the most obvious impact on the contestability of the market. Therefore, an entity vested with special or exclusive rights by the State should generally be subject to the GPA regulations. Similarly, due to its distorting effect on trade, the granting of special or exclusive rights is also subject to regulation under Article XVII: 1(a) GATT (State trading enterprise) and Article VIII:1 GATS (monopoly supplier). However, if the entity with special or exclusive rights could otherwise prove that despite its vested rights, it competes in the market without any difference from other commercial entities, the entity could be exempt from the GPA obligations with a justification.

In addition to the exclusivity arrangement, if public monopolies or public-controlled or public-influenced entities enjoy a dominant market position in a highly concentrated market, it would preferably subject these kinds of entities to regulation under the GPA as they supposedly face little competitive pressure and would have no motivation to compete in the market.¹⁴⁸ Furthermore, the higher the concentration is, the lower the possibility that these entities compete in the market and the higher the possibility that public control or influence can effectively lead these entities to apply a discriminatory procurement policy. For this concern, the structure of the market in which the entity participates is a core indicator of its contestability.

A crucial prerequisite of market analysis is defining the relevant product and the geographical market. Plainly, the economic definition of a ‘market’ is ‘the area of economic activity in which buyers and sellers come together, and [in which] the force of supply and demand affects prices’.¹⁴⁹ However, as the definition of a ‘market’ serves to support a specific legal analysis, that legal analysis of the market must take precedence over an economic analysis.¹⁵⁰ Specifically, for the purpose of rule-making, the relevant market should generally be delineated in legal terms rather than economic terms. For that reason, the Appellate Body indicated under the SCM agreement that the term ‘market’ is used to refer to a particular Party, even if it is clear that there is a global market for the product at issue.¹⁵¹ Likewise, the relevant market here generally refers to the market of the relevant GPA Parties. In regard to a specific context, a specific refinement should be left to the Parties of a specific dispute.

In addition to the scope of the ‘market’, the market shares of an entity and the overall level of concentration in that market must be investigated to support the assessment of the entity's competitive situation in that market. If a publicly controlled entity has dominant market power in a concentrated market, it is highly likely that the entity could apply discriminatory or preferential measures in making government procurement decisions rather than making the most cost-effective procurement decisions based on commercial considerations.

The utilities market is a typical case in point. Utilities sectors are characteristically network industries, in which the market is usually concentrated for the purpose of having a large-scale economic effect. As a concentrated market, the competition in the utilities market is unlike that of other private markets. Moreover, as utilities are essential services that play a vital role in economic and social development, they are often regulated by government control or influenced by government policies through public ownership, price control, licensing, and so forth. As a result, those legal or policy barriers limit the number of competitors that can serve the market; moreover, technical barriers to entry, including sunk costs, could further limit the number of new competitors.

Accordingly, utilities operators should presumably be covered under the GPA due to its non-competitive nature. These arguments can also find supporting evidence in the entity coverage schedules of GPA Parties, such as the EU, Liechtenstein, and Montenegro. Moreover, the EU Directive 2014/25/EU specifically subjects the procurement in utilities sectors by public undertakings and undertakings with special or exclusive rights to a non-discrimination obligation.

Another condition for the invocation of the GPA obligation is that the procurement of the non-competing entity must be larger than the domestic supply under non-discriminatory equilibrium. Under this condition, shifting the procurement of this non-competing entity towards domestic suppliers (by exclusionary or preferential government procurement measures) may have a trade effect, namely, by increasing domestic supply and reducing imports. The economic analysis of Robert and Baldwin demonstrated that when government demand does not exceed domestic supply and when domestic products and foreign supplies are perfect substitutes for one another in a perfectly competitive market, the government needs to import from a foreign supplier. In this situation, excluding imports for domestic supply is merely a process of replacement (i.e., government imports are replaced by private imports).

Therefore, if the procurement demands are smaller than the domestic supply, discriminatory government procurement does not have a distorting effect. The conclusion also applies to the situation when the procurement can also be supplied from a foreign oligopoly and when the domestic supply and foreign supply are heterogeneous.¹⁵² Thus, discriminatory government procurement would have a trade distortion effect, which is of concern to the GPA. Governments have traditionally espoused the belief that reserving procurement contracts for domestic suppliers could serve as a catalyst for stimulating domestic supply. This, in turn, was expected to bolster national output and concurrently curtail imports, especially within the sphere of government demand. However, navigating the intricate relationship between government demand, domestic supply, and imports constitutes a nuanced and delicate balancing act. Under conditions of perfect competition, wherein government demand aligns seamlessly with domestic supply for homogeneous goods, the impact of discriminatory procurement policies on both domestic supply and imports is marginal. Although in an oligopolistic setting results are ambiguous in some cases, a prevailing proposition suggests that discriminatory procurement policies tend to be generally ineffective in fostering domestic supply or reducing imports. Remarkably, such policies might even contribute to an increase in imports.¹⁵³ This is particularly evident when favoritism discrimination is applied, entailing premium ad valorem (proportional to the price) benefits for domestic suppliers compared to import prices. The efficacy of these policies is further compromised in the presence of barriers to competition in domestic markets that hinder the entry of foreign competitors.

Given the intricate nature of determining optimal discriminatory policies, a pragmatic rule of thumb surfaces – avoidance of discrimination. The rationale behind this approach stems from the anticipation that favoritism is prone to being more costly than the alternative of non-discrimination. ¹⁵⁴ Rational decision-making, therefore, inclines towards a non-discriminatory stance against international competitors. This inclination gains further weight when considering the complex dynamics of import reductions and welfare increases, intricately influenced by market structures, national competition policies, and comparative advantages. The lack of transparency in information pertaining to these factors adds an additional layer of complexity, potentially rendering government attempts to formulate optimal discriminatory procurement policies.

The immediate negative domestic political impacts linked with opening market access contribute significantly to governments’ reservations about adopting long-term, beneficial non-discriminatory procurement policies. This hesitancy is particularly pronounced in countries with limited trading power in international trade, where the prospect of short-term compensation is perceived as unlikely.¹⁵⁵ This decision-making inertia is further exacerbated in developing countries that are often disinterested in initiating accession negotiations. Moreover, certain existing Parties, such as the US, exhibit a proclivity towards protectionist policies within the framework of the GPA.

In addition to economic considerations, governments are compelled to weigh non-economic objectives and exercise discretion in their interventions. This involves a nuanced evaluation of the political dimension, entailing considerations of social values and welfare. This implies that any definition and rules within the GPA must harmoniously blend political, economic, and jurisprudential concerns. Claims of derogation, exclusion, or justification from the non-discrimination principle necessitate a balanced approach of definition rooted in both political considerations and economic expertise.

6. Conclusion

The evolution of the GPA has produced a list approach, whereby Parties list their ‘covered entities’ in a series of schedules. Unfortunately, the list approach has complicated accession negotiations and discouraged Parties from providing a ‘wider’ range of entity coverage. This turn of events has limited the effectiveness of the MFN obligation and thereby handicapped the GPA expansion. Moreover, this approach raises some tensions and a lack of legal certainty over which entities are covered by the GPA, especially in connection with entities that are not strictly ‘government entities’, such as SOEs. This problem is exacerbated in the case of modern SOEs in some developing countries, many of which can bear both public and private features.

To replace the current list approach, the author proposes a definition for the ‘covered entities’.

First, a review of the history of the GPA and the study of the coverage schedules of the GPA Parties reveals that the factors of ‘public control’ and ‘competition’ are politically accepted as two attributes of ‘covered entities’.

Second, it is observed that a ‘black and white’ division between the public and the private is obsolete and not problem-solving for defining entities bearing both features. Therefore, a functionalist approach that examines the nature of the entities’ economic activities in the market on case-by-case analysis is appropriate to identify whether a ‘public–private mixed entity’ should be covered under the GPA.

Third, it is desirable to take a general harmonizing approach together with escape clauses for the definition of ‘covered entities’. That legal structure takes into account the integration of the government procurement market among the GPA Parties and encompasses a generally broad entity coverage as well as offering flexibility for the GPA Parties and prospective Parties.

Fourth, it is important to know that the GPA is at the interface of trade rules and competition rules. Therefore, the definition of the ‘covered entities’ can incorporate some competition law elements, just as EU public procurement regulations already do.

In the end, a conceptual framework is developed to provide further reference to the GPA Parties, by which the existence of government control is identified from both the internal and external tiers of the relationship between the entity and the government. Whether the entity competes with commercial freedom in the market where it carries out activities can be judged by examination of the contestability of the market where the entity carries out its activities, the market power of the entity, and whether governments restrict the entrance to this market of new market competitors.

Last but not least, generally, where the procurement of the non-competing entities is larger than the domestic supply under non-discriminatory equilibrium, shifting the procurement responsibilities of this non-competing entity to domestic suppliers (by exclusionary or preferential government procurement measures) would produce a trade distorting effect, which is increasing domestic supply and reducing imports.