1. Introduction

State-owned enterprises (SOEs) in China are among the largest firms in the world, and yet we do not know much about them. The top three SOEs in China are ranked second, third, and fourth on the Fortune 500 list of the world's largest companies.Footnote ¹ The share of SOEs among the top ten Chinese firms at 96% exceeds that of any other country (OECD, 2015a: Figure 8.1). The tens of thousands of SOEs in China (144,700 at the end of 2011 – see OECD, 2014: 147) are a major force in the Chinese economy and a growing presence in international trade and investment. The assumption of this special issue is that these SOEs pose a challenge for the legal regime of the World Trade Organization (WTO). The challenge is commercial, given concerns that state ownership, combined with their sheer size and their share of Chinese output, may affect the terms of competition. The challenge is political, given concerns that trade and investment by SOEs may be driven by public policy goals rather than, or in addition to, commercial considerations. Both challenges may be exacerbated by the murky world of Chinese SOEs, in which even the number of such firms is uncertain, in part because of ambiguity about the degree of state ownership or control needed to be called an SOE.

In this article, I first review whether Chinese SOEs are a problem for the WTO, and whether more sunshine on their operations might be a useful corrective. I then ask what we know about SOEs inside the WTO, including in the Trade Policy Review Mechanism. Since the answer is, not much, I consider whether mega-regional trade negotiations offer a better approach. My answer being negative, I finally consider whether an attempt to negotiate a WTO Reference Paper on SOEs might help. I conclude that transparency is likely to be a better discipline on the spillovers associated with SOEs than a search for binding rules, while also helping everyone better understand the efficiency effects.

2. What is the problem with Chinese SOEs?

Are SOEs a problem? And are they a trade policy problem? If one is asking about national welfare – that is, the economic efficiency of SOEs as a vehicle for achieving policy objectives in the many countries where they exist – then one might easily conclude that SOEs are a problem to be addressed by the International Monetary Fund (IMF), which can obtain data from finance ministries, or by the Organisation for Economic Co-operation and Development (OECD), which has an excellent set of guidelines. The Chinese government itself sees an efficiency problem, and has been reforming the sector for years – the objectives and process if not the results are well described in successive WTO Trade Policy Review (TPR) reports on China. On the other hand, if one were asking about the trading system – that is, the potential for negative spillovers affecting other firms – then that would be a trade policy problem in the domain of the WTO. And either way, are Chinese SOEs a special case? These questions can appear to be ideological rather than empirical in the absence of data, which leads to the first reason to seek more transparency – we do not necessarily know enough about SOEs, especially in China, to be sure whether we have a problem, and hence what ought to be done.

Chinese SOEs are known to be highly indebted. The IMF is attentive to them in the Article IV consultations process, but mostly because of increasing worries about the overall impact of high levels of corporate debt – the SOE share of corporate debt in China is much higher than their share of economic output (IMF, 2016). If an SOE has preferential access to credit, on the one hand, the credit not the recipient is the issue. On the other hand, the Secretariat does not know if the SOEs repay their loans, which could disadvantage foreign competitors.

The OECD is also attentive to Chinese SOEs. A recent report provides an excellent summary of the history and evolution of Chinese SOEs, including their role in the economy, governance structures, and the policy objectives assigned to them (OECD, 2015a: Chapter 5). The OECD has an active program on SOEs, in which China often participates (e.g. OECD, 2016b). The orientation is towards good corporate governance – how SOEs should be managed, and how the state can exercise oversight. The guidelines start with an expansive set of criteria that governments should consider when deciding whether an enterprise is ‘state-owned’ (OECD, 2015b: 14–16). Transparency enters the guidelines only in the context of a particular company providing the same high quality reporting on its financial results as required of other companies – something that can be complicated when a firm is assigned both commercial and policy objectives (for the detailed suggestions on reporting, see OECD, 2015b: 24). From a trade policy standpoint, however, we mean something else by transparency. We want to know the extent to which state action, the only thing subject to WTO rules, alters the terms of competition.

The other articles in this issue help us understand the challenges posed by Chinese SOEs for the WTO legal regime. We know that state-trading enterprises (STEs), which may or may not be SOEs, are subject to WTO rules, and why (Mastromatteo), though we know little about how well China complies with those rules. We know that SOEs are complex entities (Lin), that are deeply entwined in China, Inc., a complex web of overlapping networks and relationships, which is hard for outsiders to understand (Wu), making even the definition of SOE contestable. The problem was foreseen but not well handled in the 2001 Protocol of Accession (Levy), leading to conflict under the WTO subsidies and antidumping agreements (Qin, Prusa, Wu). That many provisions of the WTO are applicable to SOEs is not in doubt (Biau et al., Reference Biau, Capobianco and Chammas2016: 11–12), but these rules do not apply specifically to SOEs (Kawase, Reference Kawase2014). Mavroidis (Reference Mavroidis2016b) barely even mentions SOEs because, until recently they have not been seen as a WTO problem. The actions of SOEs have occasioned some disputes, but the evolving WTO jurisprudence has simply confirmed the complexity of what is at stake. Here I am thinking of whether SOEs are part of the state, and hence subject to the same disciplines as the rest of the state apparatus, or whether an SOE is not necessarily a ‘public body’. The Appellate Body (US–Anti-Dumping and Countervailing Duties (China) Footnote ² and US–Countervailing Measures Footnote ³ ) seems to think that, in the terms of the Agreement on Subsides and Countervailing Measures (ASCM), we know a body exercising ‘governmental authority’ by what it does, not by who owns it (Prusa and Vermulst, Reference Prusa and Vermulst2013: 199). The distinction may matter in the case of an SOE providing a subsidy, but does not help in assessing other ways in which SOEs might affect the commercial interests of foreign firms. Moreover, without more information on the universe of SOEs, and what they do, or even just their objectives, the distinction has little operational utility.

Nevertheless, we know that the United States perceives a trade policy problem. Discipline of SOEs is an explicit objective articulated by the trade promotion authority granted the President in 2015, where the aim is to eliminate both trade distortions and unfair competition ‘through disciplines that eliminate or prevent discrimination and market-distorting subsidies and that promote transparency’ (United States, 2015). Many analysts also believe that Chinese SOEs distort the terms of competition. We know that SOEs receive large state transfers, but we know less than we should about the rationale, or purposes to which the money is put (that is, the extent to which losses are being covered, on the one hand, or excess investment is encouraged for some public purpose, on the other). The concern with spillovers goes beyond the fact that SOEs do not confront a hard budget constraint, or perhaps have a monopoly position, because governments can attain similar outcomes through a variety of policy instruments, such as discriminatory entry regulation or implicit guarantees, that limit the ability of private firms to contest the market. Subsidies can take many forms, from targeted taxation and access to credit through calibrated regulation, guided sourcing, golden shares, and policies that promote conglomerates.Footnote ⁴ Without more information, we cannot understand the extent to which competition is affected by the intervention of the Chinese state; hence, concerns about spillovers and efficiency persist. Such concerns are widespread about SOEs in general (Sultan Balbuena, Reference Sultan Balbuena2016), and not just Chinese ones.

Increasingly, we also care about the spillovers from government influence when an SOE invests abroad. We want to know two things: whether that SOE brings unfair advantages to competition in foreign markets (for example, if it has access to capital at home at below market rates that it uses to invest in the host market), and whether its decisions in host markets will be taken on non-discriminatory commercial terms. The worry is that the policy objectives of a host government can be undermined by a foreign SOE acting in support of the policy objectives of its state owner. The latter is often phrased as a worry about whether SOE investment compromises national security in some way. Should Canadians worry, for example, if a Chinese SOE became the owner of a major Canadian oil company? Would that company operate at the behest of the Chinese government in ways that might be inimical to Canadian security (Canada, 2012: 19)? Since 2012, Canadian policy under the Investment Canada Act requires ‘that the investor satisfies the Minister of the investment's commercial orientation; freedom from political influence; adherence to Canadian laws, standards and practices that promote sound corporate governance and transparency; and positive contributions to the productivity and industrial efficiency of the Canadian business' (Sauvant and Nolan, Reference Sauvant and Nolan2015).

In trying to understand how SOEs are a trade policy problem, I wondered if Canadian firms mention unfair competition from Chinese SOEs in the complaints they file under Section 20 of Canada's Special Import Measures Act (SIMA). Such complaints can be a source of information in themselves as well as a means of revealing that firms see a problem, and they can indicate, if only by inference, when lack of transparency is part of the problem. Section 20 of SIMA is the legal basis for the Canada Border Services Agency (CBSA) to treat an industry or sector in a designated country as operating on a non-market economy basis with respect to antidumping investigations if ‘domestic prices are substantially determined by the government of that country and there is sufficient reason to believe that they are not substantially the same as they would be if they were determined in a competitive market’. An SOE may be considered to constitute ‘government’ for the purpose of a subsidy determination under SIMA ‘if it possesses, exercises, or is vested with, governmental authority’.Footnote ⁵

In a review of a selected number of recent cases, we found some where a firm alleged a subsidy connected with state ownership.Footnote ⁶ In Certain Fabricated Industrial Steel Components,Footnote ⁷ the complainants allege that purchases by SOEs affect the domestic price in China, which is in effect a subsidy. In Certain Welded Large Diameter Line Pipe,Footnote ⁸ the complainant provides evidence that eight of the top ten steel companies in China are state-owned and that the government ‘has continuously pressured state-owned steel mills to avoid cutbacks in bids to maintain economic growth and employment, which in turn has been a contributing factor behind plummeting Chinese steel prices'. In Certain Copper Tube Originating in or Exported from the People's Republic of China,Footnote ⁹ the complainant ‘provided evidence to demonstrate that the nonferrous industry in general and the copper sector in China are largely state-owned and/or controlled by the central government or by regional and local authorities', hence the government ‘through extensive use of industrial policies and measures, strongly influences the supply and price of refined copper’. In addition, ‘The CBSA compared the interest rate on the loans received with the benchmark interest rates and determined that ZJ Hailiang had received preferential interest rates from the state-owned banks.’ Finally, in Galvanized Steel Wire,Footnote ¹⁰ the CBSA observed that the Chinese government ‘has continuously pressured state-owned steel mills to avoid cutbacks in bids to maintain economic growth and employment, which in turn has been a contributing factor behind plummeting Chinese steel prices'.

Canadian firms evidently see a problem of anti-competitive conditions in the Chinese market exacerbated by state ownership. In one case, an SOE may have itself provided something that looks like a subsidy to another SOE. Officials agree on the existence of a problem, although the picture of that problem in the cases mentioned above is murky.Footnote ¹¹ And the solution is second best – unilateral trade remedy measures imposed by the Canadian authorities only offer relief against Chinese imports in Canada; such measures can do nothing about competition in China itself, which is also a concern expressed by Canadian firms, including Canadian investors in China. More transparency about SOEs would make it easier to understand their impact on the relevant markets, hence the extent to which SOEs are a trade policy problem. Would more transparency also enhance the disciplines on SOEs?

3. Sunshine as a principle and a tool

The core worry about SOEs is that they are subsidized by their governments. Formally binding discipline on subsidies is notoriously difficult, but informal law can help (Shaffer et al., Reference Shaffer, Wolfe and Le2015). Can more sunshine help address the problems associated with SOEs? The first use of sunshine as a metaphor for transparency as a policy tool is attributed to the US jurist, Louis Brandeis. In writing about efforts to regulate finance, Brandeis (Reference Brandeis1914) stated, ‘Publicity is justly commended as a remedy for social and industrial diseases. Sunlight is said to be the best of disinfectants; electric light the most efficient policeman.’ The Brandeis view assumes that agents whose actions are subject to public scrutiny will hew more closely to shared understandings of the common good. If not, other agents provided with information can exercise appropriate discipline. In this view of agency, sunlight contributes more to social order than coercion.

The purpose of transparency, an essential element of WTO institutional design, is to illuminate trade policy practices to the benefit of both governments and traders. Such transparency reduces the inherent information asymmetry when a government knows more about its domestic policies than do its trading partners. Here too the WTO role is analogous to that of any regulator, who needs information from the regulatee without which the regulator cannot know enough about a problem to be sure how to proceed. It begins with the procedures Members ought to follow at home, such as publication of all measures affecting trade and impartial administration of trade-related rules; if they do not follow such procedures, efforts to mitigate domestic obscurity by actions in Geneva will be more difficult. Transparency in Geneva takes three forms. The one that attracts the most notice of scholars is largely outside the scope of this paper – a complaint in the dispute settlement system draws attention to the trade policy practices of another Member. The other two forms are notification and surveillance through the committee structure, where consistency with obligations is the main consideration, and the Trade Policy Review Mechanism (TPRM), where the objective is simply better understanding of the trade policies of Members.

I have claimed elsewhere that transparency, in the form of good data and a forum for surveillance, can reduce the propensity to resort to dispute settlement (Wolfe, Reference Wolfe2013), but the institutional design of transparency matters for its success as a policy tool. Two dimensions are especially important in the WTO – how the other Members are notified of a new policy action, the subject of dozens of notification requirements in the agreements; and how a notification is discussed in Geneva, which happens in the many sectoral committees. Such meetings are opportunities for Members to learn more about the incidence of a particular policy, and to understand the rationale for its use. Transparency practices in the WTO are based on the principled belief that democratic governance and efficient markets are both enhanced when participants know what is going on, and when administrative agencies have a degree of autonomy, or independence from political interference, yet can be asked to give reasons for their actions. The one is effectively a constraint on the other: administrators must be free to get on with the job, but openness is a constraint on abuse of discretion.

These western ideals of democratic governance are not shared by China. The government attempts to comply in a mechanical fashion with WTO publication requirements, including the translation into English of trade-related laws and regulations (WTO, 2016b: 69), and it is doing reasonably well in meeting its notification obligations. But transparency is far from being an embedded principle of governance (Potter, Reference Potter2014: 24). Whether or not what Coglianese (Reference Coglianese2009) calls ‘fishbowl transparency’ (where the public can be a fly on the decision maker's wall) is a good idea, and he is dubious, such extreme transparency is not something the Chinese state will provide for in the foreseeable future. What Coglianese calls ‘reasoned transparency’ (asking officials to provide reasons for their decisions) is a worthy objective, though that too will not come quickly. And the problem goes far beyond SOEs. The US Food and Drug Administration faces the same information problem with respect to China, and it matters more: US regulators need to know about the safety of imports, including drugs and medical devices, but confidence in Chinese regulatory transparency and effectiveness was shaken by scandals involving unsafe food, unsafe toys, and much else in the middle of the last decade (Delisle, Reference Delisle, Coglianese, Finkel and Zaring2009). The tension between western notions of the ‘responsible agency’ of officials with traditional ‘patrimonial sovereignty’ where officials are accountable to their superiors will not be resolved soon (Potter, Reference Potter2014: 37). When governance is relational, western process-oriented ideas have little purchase, which may explain why WTO transparency provides less discipline on Chinese SOEs than one might hope.

4. How does WTO transparency actually work in this area?

It follows from the discussion so far that discipline on SOEs ought to begin with notification of any subsidies that they receive or offer. Collins-Williams and Wolfe (Reference Collins-Williams and Wolfe2010) showed how the record of industrial subsidies notification under the ASCM was poor. It still is, in general, and especially with respect to SOEs. Why do Members not notify subsidies, in general? Four reasons can be advanced (Wolfe, Reference Wolfe2013). The first is bureaucratic incapacity. Second, Members might worry about providing adverse information for a potential legal dispute, perhaps about a measure they suspect might be illegal. By notifying, they provide information that a trading partner might not have and they admit that the measures might be actionable. Third, Members' trade authorities find it easier to notify actions taken by themselves than data on subsidies offered by other ministries, or other levels of government, or by SOEs. The fourth reason, and perhaps most important, is ambiguity about what requires notification.

This ambiguity is evident in China's notifications. On 29 July 2016, China tabled its latest ‘new and full notification’ of subsidies under the ASCM (G/SCM/N/95/CHN/Suppl.), covering the period to the end of 2014. Given the distinction made in recent Appellate Body decisions, this notification is unlikely to contain any subsidies provided by SOEs, since China does not see them as ‘public bodies' whose measures would be subject to notification. Whether any SOE would be on the list as a recipient of subsidies is also unlikely, since the discipline applies to the source of funds, not the recipient. A reverse notification would be possible if a trading partner wished to argue that a Chinese subsidies program discriminated in favor of SOEs.

As a result of the absence of any specific obligations applying to SOEs, little is known about them from regular notifications, or discussion in WTO committees. Given the missing incentives to notify, Mavroidis and Wolfe (Reference Mavroidis and Wolfe2015) argue, one solution might be to enable a disinterested party to acquire and disseminate information as the common agent of all participants in the trading system. Some Members may not need the help with respect to Chinese SOEs – the Americans obtain a lot of information from their offices in China, and many of their people speak Chinese, as does the staff of the American Chamber of Commerce in China (AmCham). The result is their periodic report on China's WTO compliance (United States, 2014), but this report serves American interests. The WTO has multiple principals – 164 member governments, but also citizens, and hundreds of millions of traders. Sunshine as now practiced does not properly enable these agents to discipline other agents who may not be playing by the rules. In order to ensure that such discipline is possible, two things are required, both of which can only come from a common agent: active efforts to (1) increase the amount of information available; and (2) make that information widely available in a useful way. A move toward increasing the role of the WTO Secretariat could be a useful response to the failure of Members to submit notifications, which implies a greater role for the TPRM. Is it up to the task?

5. What do we know from TPR reports?

The central objective of the TPRM is ‘to contribute to … the smoother functioning of the multilateral trading system, by achieving greater transparency in, and understanding of, the trade policies and practices of Members'. In the periodic Trade Policy Review (TPR) of each Member, which in the case of China is completed every two years, the Secretariat sometimes warns or expresses concerns on the basis of its analysis, but never criticizes Members explicitly, and never comments on their rights and obligations under the WTO agreements. It follows from the TPR mandate that Members need to know enough about the overall picture of SOEs in a given country to understand their effect on the trading system. What we learn there is less than we might hope.

The core of each TPR report is based on notifications from Members, but each report builds from a far wider range of information. If a country has been less than fully transparent in regular work, for whatever reason, it may not provide more information in the TPR. The Secretariat collects data from official sources (in addition to questionnaires to the Member under review) and non-official sources, including from other international organizations, media reports, academic analysis, and NGOs. To ensure accuracy, the Secretariat seeks verification of the data from non-official sources when discussing the draft of its report with the Member. The nature of the TPR process creates an ongoing debate between the Chinese authorities, the WTO Secretariat, and other Members (who pose detailed questions when the report is discussed) about what counts as an SOE, and what information about them ought to be provided. The authorities learn from the debate about how to find and aggregate information in WTO terms, often because they want to dispute the information the TPR contains from other sources.

The major problem for the TPR of China is how much the Secretariat does not know. Unlike the reports on other large emerging economies, such as India, a highly transparent country, the reports on China may have significant gaps. Of course, comparability of data on SOEs is a general problem given the different forms they take, and the sometimes divergent national classifications (Kowalski et al., Reference Kowalski, Büge, Sztajerowska and Egeland2013: 64). The Secretariat tries to estimate transfers to SOEs but the Chinese government disputes the accuracy of the data. Even efforts to get a simple count of the major SOEs is hard because China distinguishes between state-owned, state-controlled, and state-invested enterprises (WTO, 2016a: Table 3.19), hence the TPR uses OECD data. The Chinese authorities are most comfortable providing data about SOEs that are under the State-owned Assets Supervision and Administration Commission of the State Council (SASAC), but in the Reference He2016 TPR report, only 106 SOEs are said to be under SASAC, and yet tens of thousands of SOEs are known to exist. The successive TPR reports on China present a shifting and confusing picture of the role of SASAC and its influence, due, in fairness, to the active reform program underway.

What then do we learn about such spillovers from the Reference He2016 TPR of China (WTO, 2016a)? Not much. This year's TPR seems more descriptive than in the past, perhaps in response to pressures from the Chinese government. The section on STEs (3.172) notes that a new notification was submitted in 2015, but provides no detail, simply repeating the policy objectives as stated by the authorities. The report describes the responsibility of China's official export credit agency (3.115) without assessing whether its support is provided at market rates, or rates under the OECD Export Credit Arrangement. One paragraph stands out. By digging through corporate filings the Secretariat is able to say (3.179) that:

The previous TPR (WTO, 2014) had only been able to say

When the report was discussed in the Trade Policy Review Body (TPRB), this section caught the eye of the delegates of Chinese Taipei, the US, Canada, and Australia, who asked questions of China designed to elicit more information. In its replies (WT/TPR/M/300), the government denied that banks received any such direction from the government, or that any such support is provided to SOEs. The official Discussant of the report, Ambassador Joakim Reiter of Sweden, said that

Two years later, a new discussant reiterated the same concerns (WTO, 2016b: 3.19). While many questions were asked orally and in writing about SOEs, the answers again do not add much detail (WTO, 2016c).

In sum, we learn from the TPR process that trading partners worry about the spillovers associated with Chinese SOEs, but the picture remains murky. I then wondered if this obscurity in the TPR is unique to China. For comparison, I looked at the section on SOEs in the TPR report on the European Union, home to a great many such firms. The last report (WTO, 2015b) contains two pages on state enterprises. The report describes the EU rules on ‘public undertakings', but all of its quantitative information comes from the OECD. That data are reasonably complete, because the EU member states are also members of the OECD, and participate in its work, including its database of SOEs (OECD, 2014). I then looked at the annual monitoring reports by the Director-General. That report mentions SOEs only when they come up in TPR reports on individual countries (WTO, 2015a). The report does have horizontal sections on various areas of WTO work, but always on the basis of activity in the relevant WTO body. Absent explicit WTO obligations in an area, the TPR reports are mute.

In a sense that is as it should be. Members can only be accountable in Geneva for promises that they have made there (Wolfe, Reference Wolfe2015). If Members of the WTO have not agreed on a definition, in the sense of which SOEs ought to be subject to discipline, and have not developed a shared understanding of how SOEs do or do not offend the non-discrimination norms of the trading system, then they and the Secretariat will flounder in knowing what to be transparent about.

6. Could mega-regional negotiations improve SOE transparency?

With WTO negotiations stalled, regional negotiations became a forum for trying to advance disciplines on new issues, often explicitly designed to develop new rules without China (Wolfe, Reference Wolfe, Van Assche, Tapp and Wolfe2016). While the Trans-Pacific Partnership (TPP) agreement of early 2016 is now dead, or at least in limbo, it does contain the most extensive attempt of any preferential agreement to draft rules for SOEs, based in part on the provisions of the Singapore–US FTA (Sylvestre and Marcoux, Reference Sylvestre and Marcoux2016: 17ff). Chapter 15 of NAFTA is much more limited, while the focus of Chapter 18 of CETA is on non-discrimination and commercial considerations in the operation of state enterprises and monopolies, not on subsidies, and it does not mention transparency. The Transatlantic Trade and Investment Partnership (TTIP) SOEs chapter would probably be similar to the TPP, with a more sophisticated answer to the question of how much and what kind of state involvement would make an enterprise subject to the disciplines. The Trade in Services Agreement (TiSA) chapter is similarly modeled on TPP. Indeed, US officials explicitly saw TPP texts on new issues as being a basis for further multilateral or plurilateral negotiations (Inside US Trade, 2016). Australia and Canada also supported the TPP chapter; Canada is likely to use it during the exploratory talks now underway with China on a possible FTA. In this section, I consider what we can learn from the TPP about possible rules that could be adapted in some other agreement involving China.Footnote ¹²

The TPP chapter on State-Owned Enterprises and Designated Monopolies has important provisions on non-discriminatory treatment in commercial transactions of SOEs in their own or other markets. But it is above all a subsidies chapter, although the word is not used there or anywhere else in the TPP. The chapter attempts to address the problems that have arisen in the WTO by clarifying what an SOE is, based on the amount of state ownership or control, and what sort of measures ought to be constrained. It has one conventional notification obligation (Article 17.10.1), requiring publication of a list of a Party's state-owned enterprises. Other Parties can challenge the absence of an entity from such a list, notably by raising the matter in the Committee, but it may be hard for anybody except the putative owner to know if an entity meets the definition.

An ongoing description of the universe of SOEs could be useful, but the SOE chapter does not contain any further notification requirement, for example on non-commercial assistance offered, perhaps reflecting US frustration with the inadequacies of WTO notifications, as discussed above. What the US pushed for instead was a provision that allows one Party to make a request for more information on another Party, in writing, with an explanation of how the activities of the entity may be affecting trade or investment between the Parties (Article 17.10.3 and 17.10.4). The type of information required in response to a written request (Article 17.10.5) tracks the requirements in the ASCM notification template (WTO, 2003), and is similar to that specified in the OECD Guidelines discussed above, although making it public could be difficult for some countries. The requirement for another Party to explain the possible effects on its interests eliminates the need for guesswork inherent in the part of the ASCM notification template that asks the drafter of a notification to provide ‘Statistical data permitting an assessment of the trade effects of the subsidy’. Notifications rarely contain such information, perhaps, Collins-Williams and Wolfe (Reference Collins-Williams and Wolfe2010: 564) argued, because trade effects are best assessed by the affected actors.

The provision of information in response to a question does not prejudge the legal status of any assistance (Article 17.10.8), but all the usual worries about notifications discussed above may arise. Moreover, the mandate of the Committee on State-Owned Enterprises and Designated Monopolies (Article 17.12) does not provide that it should receive or discuss notifications – which would be hard to do in any event since the TPP would have no Secretariat. While the Committee could discuss entities that ought to have been mentioned on the list, such reverse notification will be ad hoc, and difficult for all but the best-informed governments. This weak institutional design of the TPP SOE transparency provisions (which is typical of the institutional provisions of the TPP) means that observers who think it can ‘name and shame’ the providers of subsidies (Miner, Reference Miner, Schott and Cimino-Isaacs2016: 341) are misinformed, because the supplementary information in the response to a question is not actually a notification; indeed, the Party supplying the information can request that it be kept confidential (Article 17.10.9). It is not clear, therefore, that any of the TPP provisions except the list of SOEs (Article 17.10.1) would actually increase transparency. The Americans would be glad of any information received in response to a request, but transparency best provides discipline when sunshine illuminates the practices in question for all affected actors.

Would China accept TPP provisions on SOEs? Some scholars think these provisions might help push the reform process along (Yun, Reference Yun2016: 32), which probably gets it backwards: when the reform process is sufficiently advanced, the authorities will happily accept new disciplines. Other scholars think that the reform process now underway is too limited to allow Chinese leaders to agree to adopt the TPP's high standards for SOEs (He, Reference He2016: 16): although it is said that they have sought technical briefings to ensure that they understand the SOE provisions. What about the transparency provisions? It may be unrealistic to expect the Chinese to create an online list of every SOE, and even more unrealistic for any of their partners to try to provide written evidence of the trade effects of most SOEs on any eventual list in order to justify a request for more information – would even the Americans know enough about the tens of thousands of SOEs assumed to operate in China to pose such questions?

Even if TPP comes into effect, some day, the transparency it could provide will be imperfect. Moreover, I think Wu (Reference Wu2016: 60) takes the wrong tack when he argues that even if China will never join, mega-regionals ‘offer a chance to establish norms, linked with additional preferential trade access for those that choose to accept them’. Norms and new rules ultimately arise in social interaction – if China is not part of the process, new texts will be a dead end as a means for dealing with China Inc.

7. Would negotiations on an SOE Reference Paper be timely?

The Americans and others evidently believe that SOEs are a trade policy problem, and that more transparency would help. If TPP is a dead end, how can we improve the WTO? The solution will not come through further dispute settlement cases on the meaning of ‘public body’, nor should Members wait to address the problem until it forms part of the single undertaking for the eventual successor to the Doha Round. In this section, I justify plurilateral negotiations as a means to make explicit what is already implicit: since WTO disciplines do indeed cover SOEs, more transparency is possible.

The WTO is a government-to-government contract, but some private behavior is regulated in the agreements – the antidumping agreement, for example, applies to the measures governments apply to foreign firms. The most salient example in the SOE context is anti-competitive pricing of essential telecommunications facilities in telecoms. The relevant legal obligations are assumed by WTO Members, who must enforce them on private entities through the so-called ‘Reference Paper’ that is incorporated into Members' Schedules to the General Agreement on Trade in Services (GATS) in accordance with the 1997 Agreement on Trade in Basic Telecommunications Services. In Japan–Semiconductors,Footnote ¹³ a GATT panel held that activities not performed by a state but which can be attributed to it should come under the ambit of the GATT. The key to attribution is the response to the question ‘would the challenged behavior have occurred absent incentives by the government to behave in this way’? Governments thus do not need to compel behavior. It suffices that they ‘incentivize’ private agents to behave in a particular way in order to be themselves accountable before the WTO (Mavroidis and Wolfe, Reference Mavroidis and Wolfe2017).

If we want more transparency through the WTO, we need an agreed understanding of the problem, hence agreement on what to be transparent about. In the WTO, Members only notify what it is agreed they should notify, and not always even that. The absence of specific WTO disciplines on SOEs, including the absence of a definition, leads to the absence of transparency requirements about who they are and what they do. Multilateral consensus on new rules might prove elusive, but an SOEs Reference Paper on the model of the Telecoms Reference Paper, or the one Mavroidis and Wolfe (Reference Mavroidis and Wolfe2017) propose for private standards, might be a way to make progress. A plurilateral negotiation under the umbrella of the WTO as opposed to a PTA would be non-discriminatory, allowing all interested countries to participate, which would ensure that the focus is on issues that all agree are a problem.

The virtue of a Reference Paper as a negotiating modality is that the results can be inscribed in the Schedules of participating Members. It enters into force when a pre-defined critical mass is reached, with no need for consensus among all Members, or formal ratification at home.Footnote ¹⁴ Critical mass would require the participation of key OECD countries, plus the major emerging market homes of SOEs. The negotiations would need to agree on a definition of the domain of the Reference Paper, decide on transparency procedures, and create a committee. The committee would be crucial – it could exert pressure on participants to comply with their notification requirements; it would also provide occasions to ask about errors and omissions in notifications due to ambiguous definitions, which will be the case with SOEs.

A Reference Paper might also offer a simpler approach to creating disciplines for what after all are the anti-competitive actions of firms, which goes beyond the possibility of subsidization. If using national competition policy to ensure ‘competitive neutrality’ is the most important objective (Pérez Motta, Reference Pérez Motta2016), then transparency about SOEs would be part of any WTO negotiation on competition policy, whenever it might start.

Negotiations on SOEs would also need to consider linkages to procurement disciplines. Levy in his article in this issue draws attention to concerns expressed during China's WTO accession negotiations regarding SOEs as purchasers – do their government mandates include a requirement to purchase locally in ways that discriminate against foreign suppliers? Are Chinese purchasers making decisions solely on standard commercial considerations, such as quality and price, or does the government provide direction? Other WTO Members would be more inclined to believe the assurances China made on this point when it joined the WTO (WTO, 2001: para 46) if we had more transparency about the operations of SOEs. TPP excludes government procurement from the scope of the SOEs chapter (Article 17.2.7), but has many provisions requiring that SOEs use commercial considerations in making their purchases.

In the WTO, we might achieve such transparency if China accedes to the Government Procurement Agreement (GPA), and if SOEs would then be listed in Appendix 1 of the GPA. If the major SOEs were subject to the GPA, under its transparency procedures we would know a great deal about their purchasing decisions, both in general and with respect to specific procurements (Mavroidis, Reference Mavroidis2016a: section 10.7). During the negotiations on accession to the GPA, which have yet to conclude, including the SOEs in China's revised offers was a key ask among the more engaged members, according to one negotiator. Perhaps the reason China was reticent to include substantive obligations with regard to regard to SOEs was precisely to shelter them from opening up large procurement contracts to foreign bidders – major infrastructure projects, for example are procured not by ministries but by SOEs they control (Wang, Reference Wang2009: 678). Including them in the GPA might also constitute an admission that they are ‘public bodies', which could affect how they are treated in trade remedy investigations. In any event, in 2015 China said that it would not be able to make a new GPA offer until ongoing reforms of SOEs were complete (Inside US Trade, 2015).

Information is not free, and disclosure can be awkward. Are the incentives of the major players aligned? Put differently, why should China participate in a Reference Paper on SOEs? At the broadest level, restrictions imposed because trading partners fear the motives of Chinese SOEs disrupt the investment strategies of those firms, and hence their participation in sophisticated global value chains. If everybody involved knew more, and had a space to discuss what they know, they might be more relaxed. (One can see the potential for such balancing in OECD recommendations that call for ‘high standards of governance, disclosure, accountability and transparency for SOEs' while also seeing the need for host policies to be ‘non-discriminatory and transparent towards SOEs (OECD, 2016a: 15)’. More narrowly, as we saw above in the case of Canada, and as we have seen in decisions taken by the United States, the fact of SOEs is part of the story about how China does not operate as a market economy, which becomes a justification for continued application of the provisions in Article 15 of China's accession protocol (WT/L/432) permitting an alternative methodology in trade remedy cases. Here too more information might help to resolve uncertainty. Everyone needs to know more about SOEs, and the WTO provides neutral ground.

8. Why do we need more sunshine over Shanghai?

Every tool of governance has an (implicit) conception of the definition of the problem, the objective, the nature of society, and legal authority. Instruments can be embodied in public agencies, SOEs, or private firms. Sometimes economic actors will seem like agents of government in implementing a particular policy, for example when commercial banks implement a government student loan program or a construction company builds and runs a new toll highway. It is a commonplace to define as ‘public’ anything that is not ‘private’, but as we see in the TPP, TTIP, and TiSA, debates about what makes an entity sufficiently state-owned to be subject to discipline, the distinction breaks down in practice. The TPP chapter has exceptions for non-conforming measures, which countries could list in an Annex. Canada exempted the public broadcaster (CBC), the mortgage insurance agency (CMHC), and certain bridge authorities. The US exempted federal funding for housing and infrastructure. That is, countries have views on the legitimate role of the state, and views on which of those purposes should be subject to international discipline. As Canada's Minister of International Trade said in praise of improvements to the investment provisions of CETA (Canada, 2016b: 12), ‘We clarified that it was the job of democratically elected states, and not of a trade agreement, to choose which parts of the economy should be in the private sector and which should be in the public sector.’ Acceptance of such difference is embedded in the norms and practices of the trade regime.

The US tends to see the SOE challenge in regulatory terms. As Canada's Chief Negotiator for the TPP told a House of Commons committee (Canada, 2016a), ‘All of these rules that we've put in the TPP are rules that our state-owned enterprises have to play by. All of our crown corporations have to play by these rules already. We have put them in the TPP in an effort to have other countries also have a similar standard.’ But does that objective concern the end state for the role of SOEs in the economy, or the means to get there? Trade people prefer performance over design standards, but neither approach may be appropriate given the diversity of SOEs, the objectives assigned to them by states, and the ways in which states support them. Maybe just being better able to observe their actions would be enough. We might like alignment between OECD countries and China on how SOEs are regulated, but shared guidelines would be an improvement. But mere publication of a list of SOEs may not help us understand the process through which SOEs exercise their functions. We need something like a committee, where OECD SOEs are also on the table, so that consensual understanding about best practices, and relations of trust, can emerge.

I agree with Mavroidis that the GATT/WTO system is a set of solutions to problems in commercial relations between states, which fits well with the familiar notion that the system is a contract (Mavroidis, Reference Mavroidis2016b). But what kind of contract? The framers knew that they could not specify the thousands of domestic policies to which the rules might apply, nor could they imagine the evolution of such policies, hence the WTO is necessarily an incomplete contract. For many economists, the idea is that because we are repeat players, we do not have to be specific and we can also tolerate some anti-contractual behavior either because we believe in future reciprocity, or because occasional deviations matter less than keeping the contract alive. The trading system is also, Mavroidis therefore says, a relational contract. The Members' relationship is ongoing and hence is dependent on building and maintaining relations of trust. The explicit terms of the contract are just an outline as implicit terms, and understandings determine the behavior of the parties. Conversely, in areas where the parties lack consensual understanding of the issues, and relations of trust are yet to emerge, premature efforts to complete the contract, to create binding rules through negotiations or formal disputes, are unlikely to succeed (Wolfe, Reference Wolfe2015). Regulation of SOEs is just such an issue.

Transparency is likely to be a better discipline on the spillovers associated with SOEs than a search for binding rules (Shaffer et al., Reference Shaffer, Wolfe and Le2015), while also helping everyone better understand the effects on economic efficiency. And it might help with generating the information needed to assess what may be the biggest challenge with Chinese SOEs – the possibility that a sudden inability to pay their possibly massive debts causes a global financial crisis, which is why we need to illuminate the murky world of Chinese SOEs.