Multinationals as Global Institution: Power, Authority and Relative Autonomy

This article aims to inform the long-standing and unresolved debate between voluntary corporate social responsibility and initiatives to impose binding legal obligations on multinational enterprises. The two approaches share a common feature: neither can fully specify its own scope conditions, that is, how much of the people and planet agenda either can expect to deliver. The reason they share this feature is also the same: neither is based on a foundational political analysis of the multinational enterprise in the context of global governance. Such an analysis is essential for providing background to and perspective on what either approach can hope to achieve, and how. This article begins to bridge the gap by illustrating aspects of the political power, authority, and relative autonomy of the contemporary multinational enterprise. The conclusion spells out some implications for the debate itself, and for further research.


Introduction
According to a leading handbook, the "phenomenal rise" of corporate social responsibility (CSR) reflects a journey "that is almost unique in the pantheon of ideas in the management literature" (Crane et al. 2008, p. 3). Precise definitional issues are still debated if not contested. But in a lengthy etymological and philosophical discussion Sheehey (2015, p. 625) concludes that CSR is best understood as "international private business self-regulation," where private may include a role for non-governmental organizations (NGOs). In short, Sheehy affirms, CSR falls onto the voluntary side of the regulatory scale.
There is also broad consensus that "global CSR" differs from its counterpart at national levels. As Scherer and Palazzo note, national-level CSR is "based on the assumption that responsible firms operate within a more or less properly working political framework of rules and regulations which are defined by governmental authorities" (2008, p. 414). This condition, however, does not hold globally: "The global framework of rules is fragile and incomplete" (Scherer & Palazzo 2008, p. 414). Globally, there is no central regulator and national laws where multinationals operate may be weak, poorly enforced, or simply do not exist. Therefore, they propose a new paradigm, one in which the firm is drawn into greater political roles, performing tasks that we traditionally associate with the state.
At the official level, attempts to devise a comprehensive and legally binding international instrument to govern multinational enterprises go back to the UN Code of Conduct negotiations, which began in the 1970s and continued for more than a decade before they were abandoned (Sagafi-nejad 2008). They also include the Norms on the Responsibilities of Transnational Corporations and other Business Enterprises with Regard to Human Rights (United Nations [UN] 2003), produced by an expert subsidiary body of the (then) UN Commission on Human Rights, which the Commission itself declined to adopt (UN 2004); and most recently, the proposal by Ecuador and South Africa to establish an intergovernmental Working Group within the UN Human Rights Council "to elaborate an international legally binding instrument to regulate, in international human rights law, the activities of transnational corporations and other business enterprises" (UN 2014). The proposal to establish such a group was endorsed by a small plurality of states and is supported by an NGO "treaty alliance." 1 Thus far, the Working Group has held two annual informational sessions.
In short, the voluntary/mandatory debate has been a constant feature of corporate responsibility discussions for decades. However, what is missing from that discussion is a current and systematic political analysis of the multinational enterprise in the context of global governance. Such a portrayal seems essential for providing perspective and guidance to what both voluntarism and binding treaty advocates can hope to achieve, and how. This article helps to bridge that gap by illustrating aspects of the political power, authority, and relative autonomy of today's multinational enterprise. The implication, which I elaborate briefly in the conclusion, is that there is no silver bullet that would fully achieve the objectives of either the voluntary and mandatory side in the global CSR debate as it is currently conceived and practiced. A paradigmatic shift is required.
The discussion proceeds as follows. The first section describes core features of today's multinational enterprise in terms relevant to the corporate responsibility debate. The sections that follow address the multinational's power, authority, and relative autonomy in that order.

The multinational enterprise
Surprisingly, there is no legally precise and universally accepted definition of the multinational enterprise. The Organisation for Economic Co-operation (OECD) Guidelines for Multinational Enterprises, the oldest such international instrument dating back to 1976, even now employ only a minimalist conception: They [MEs] usually comprise companies or other entities established in more than one country and so linked that they may coordinate their operations in various ways. While one or more of these entities may be able to exercise a significant influence over the activities of others, their degree of autonomy within the enterprise may vary widely from one multinational enterprise to another." (Organisation for Economic Co-operation [OECD] 2011, p. 17) As a first step before delving more deeply into the status of the multinational enterprise as a global institution, I briefly describe it first as an economic and then as a legal entity.
The multinational as economic entity broadly speaking takes one of two forms: one is actor-based, the other network-based. To illustrate the actor-based view, consider Total, the French oil and gas company whose global headquarters is just a short taxi ride from the OECD in Paris, and whose organization illustrates the conventional integrated multinational. By market capitalization, Total ranks as the fourth largest publicly traded integrated oil and gas company in the world (Total 2016). It operates in 130 countries and has 100,000 direct employees. Its business segments cover every aspect of the oil and gas industry, from exploration, development, production, refining, and petrochemicals to marketing, trading, and shipping. It is also active in specialty chemicals and aims to become a global leader in new energies.
The Total "group" comprises nearly 900 subsidiaries and equity affiliates. 2 But the reach of the enterprise doesn't stop there. As part of its marketing business, Total's service station network includes more than 16,000 outlets in 110 countries, all carrying the Total brand. Reflecting local circumstances and sometimes requirements, they fall into three categories: owned and operated by an in-country subsidiary, owned by a subsidiary but operated by independent dealers, and dealer owned and operated. Beyond that set of relationships are countless suppliers of products and services, as well as distributors other than service stations that are contractually connected to individual corporate entities within the Total group.
Things become more complex when we look at multinationals through the lens of the contractual ecosystems they continue to generate. These are variously called transnational production networks, supply chains, or global value chains. 3 Take Starbucks as an example of a buyer-led supply chain. 4 It lies at the simple end of that spectrum.
[Starbucks] directly employs 150,000 people; sources coffee from thousands of traders, agents and contract farmers across the developing world; manufactures coffee in over 30 countries, mostly in alliance with partner firms, usually close to final market; distributes coffee to retail outlets through over 50 major central and regional warehouses and distribution centres; and operates some 17,000 retail stores in over 50 countries across the globe. (United Nations Conference on Trade and Development [UNCTAD] 2013, p. 142) The Apple iPhone 6 illustrates a producer-led production network. As of 2014, its components were produced by 785 suppliers in 31 countries (comparecamp.com 2014). The product is designed in the United States (US) and assembled in China, which also had the largest number of suppliers in 2014 at 349, nearly half the total. Some 60 suppliers were US-based, several themselves multinationals, some headquartered in other countries. Many US suppliers also outsourced fabrication of components to companies in Japan, South Korea, and Taiwan, which in turn are sourced from yet other (and lower cost) locations in South East Asia.
Other forms exist, but these two examples suffice for the present purposes. They illustrate a fundamental transformation that has taken place in the production process of many if not most sectors. Multinational enterprises "are increasingly able to fine-slice activities and operations in their value chains and place them in the most cost-effective location, domestically and globally" (UNCTAD, 2013, p. 135). The electronics and automobile industries led the way, largely because components can be broken down into so many discrete parts and are easy to transport. But garments and footwear were also early movers, and the list today includes food and beverages, chemicals, mining, furniture, and a host of others. In view of the complexity of managing these transactions, the role of services looms large (logistics, telecommunications, legal services, data processing, accounting, and human resources management, among others). Although a single product emerges at the end, production networks are inherently multi-sectoral, drawing upon inputs from several sectors simultaneously. In a word, these are miniand in many cases not-so-minitransnational economic systems. Product and process standards are set by the lead firm and then cascade across its network and down the supply chains. The links among the various parties can take a variety of legal forms: they can be among subsidiaries or affiliates of the same corporate group, joint ventures, different types of non-equity relationships (contract manufacturing, licensing, franchising), and arms-length purchasesor any combination thereof.
One striking consequence is that "about 80% of global trade (in terms of gross exports) has become linked to international production networks of TNCs [transnational corporations]" (UNCTAD, 2013, p. 135). This is because of the multiple times intermediate goods are imported, a step or two is taken in their processing, and then they are exported again as an intermediate product to the next stop, where the cycle is repeated until the final product is assembled. World trade in intermediate goods is now greater than all other non-oil traded goods combined (International Labour Organization [ILO] 2016, p, 18).
Another striking consequence reported in an International Labour Organization study is that perhaps as many as one out of seven jobs in the world is supply-chain related (ILO 2015). The report goes on to stress that this estimate does not capture informal or "non-standard" forms of work. Informal work may include specific tasks outsourced by a subcontractor that end up being performed in the home, typically by women and often children; stitching embroidery onto garments is an oft-cited example. Non-standard work also includes the entire range from casual and temporary employment to bonded and forced labor. If the ILO figures are correct and we make a rough guess about non-standard work, then global supply chain-related jobs may involve more than half a billion people. The world's total labor force in 2013 was estimated at 3.3 billion (indexmundi.com). And if we assume that many of those workers have families, the number of directly affected people easily exceeds 1 billion. Intense cost pressures and tightly synchronized delivery schedules throughout these networks and various layers of subcontractors can lead to violations of workplace standards and pose environmental hazards.
These, then, are two snapshots of the 21st century multinational seen in terms of the organization of their transnational economic activity. But importantly, this is not how the law conceives of themfar from it. When we think of a multinational we think of the likes of Exxon, GE, Unilever, Sinopec, Google, Coca-Cola, Toyota, Novartis, and so on down the list of well-known names as one company, with unity of command, operating under a single global vision and strategy, optimizing worldwide operations for efficiencies, market share, and profits. We are not wrong to think this. Those organizations do exist in the everyday world of economic activity. But they don't exist as entities in the law.
Where no equity nexus exists, as in the Starbucks and iPhone examples, relations between producers/buyers and their suppliers are governed by the private law of contracts. In the case of an integrated firm like Total, each of its 900 subsidiaries and affiliates is a separate legal entity, subject to the laws of the particular jurisdiction in which it is incorporated. Through equity relations they are ultimately linked to a "corporate parent," which is itself a separate legal entity. Hence, Robé (2016, p. 14) insists that the first step in fully understanding the multinational in relation to governance issues is to grasp this fundamental disjuncture between economic reality and legal form. The multinational as economic organization orchestrates and controls the entirety of its global operations. The group of firms or enterprises that make up the multinational as an economic organization is structured using the corporate form; but legally the group itself is not a corporation. Why is this distinction so crucial? Because national law for the most part governs the separate legal entities, not the single economic enterprise.
The "parent company" enjoys limited liability even if it wholly owns all of its subsidiaries. This means that the corporate parent is generally not liable for risks incurred by a subsidiary, or monetary damages imposed on a subsidiary, beyond the extent of its investment in it. Moreover, a subsidiary or affiliate may have subsidiaries and affiliates of its own, based on the same principle of limited liability. Some subsidiaries may be listed on stock exchanges in their own right, with the corporate parent remaining the majority or controlling shareholder. In all such cases, the parent company is not liable for harm caused by subsidiaries, other than in exceptional situations such as demonstrable negligence, fraud, or other illicit conduct that the corporate parent directed or of which it had knowledge and did nothing to stop. That would provide grounds for what is known as "piercing the corporate veil." But it remains the exception domestically and even more so across national borders. As one widely quoted criticismand witticismhas it: "Piercing seems to happen freakishly. Like lightning it is rare, severe, and unprincipled" (Easterbrook & Fischel 1985, p.89).
The main body of national law governing corporations is domestic corporate law and securities regulation, plus whatever civil and criminal provisions in other areas of substantive law and regulations may be applicable to corporations. But domestic law is only able to reach beyond its national borders in limited circumstances (Zerk 2006(Zerk , 2010. One such exception is the US Foreign Corrupt Practices Act, the scope of which includes the overseas conduct of US firms, as well as foreign firms, if their furtherance of a corrupt act takes place in or through the US. Anti-trust law is another exception in the US and the European Union (EU).
For nearly two decades, the US Alien Tort Statute (ATS) was also an exception, providing a means for foreign plaintiffs to bring suit in federal courts for egregious human rights abuses committed abroad. 5 A California district court first agreed to extend it to corporations in 1997 (Joseph 2004). More than 150 such cases were subsequently brought. The net result? The only case to go to a jury trial was won by the corporation. Two were settled for modest sums (the ATS is a civil statute, resulting in payment for damages if successful). The rest were dismissed on various procedural grounds, primarily the forum non convenience doctrine (that US courts are not the appropriate jurisdiction) or with the judiciary deferring to executive branch claims that foreign policy interests would be adversely affected. One case, Kiobel v. Royal Dutch Petroleum, made it all the way to the Supreme Court. Nigerian plaintiffs charged that the Anglo-Dutch parent companies and the Nigerian subsidiary aided and abetted the Nigerian military in its suppression of widespread protests against Shell, which led to egregious human rights violations, including extrajudicial killings of protest leaders. In a divided 2013 ruling, the Court held that a presumption against the extraterritorial application of US laws applies unless Congress has specified otherwise, a test the Court said the ATS did not meet (Kiobel v. Royal Dutch Petroleum Co 2013). The Court left open the possibility that some claims might "touch upon and concern" US territory, but they would have to "do so with sufficient force to displace the presumption against extraterritorial application." Uncertainty currently exists about precisely where the "sufficient force" threshold might lie, and also because the Supreme Court did not address the Second Circuit Court Kiobel ruling that had triggered the appeal, which held that international law does not recognize corporate liability in the first place (Kiobel v. Royal Dutch Petroleum Co 2010). However this plays out, as a general rule, unilateral extraterritorial jurisdiction is typically also contested and resented by the other states into whose jurisdiction it reaches.
Finally, the multinational enterprise barely exists under international law; indeed, some scholars have gone so far as to describe it as "invisible" (Johns 1994;May 2006). For its part, international human rights and humanitarian law generally applies to states, and the latter has been expanded to include natural persons accused of war crimes and certain crimes against humanity (Ruggie 2007). International human rights law may "contemplate" multinational corporations, as Knox (2008) has put it, and in some instances even "specify" appropriate conduct, as ILO labor conventions, for example, clearly do. But it imposes the correlative duties on states, not on companies directly. States then have to ratify the relevant instrument, enact it within their individual jurisdictions, and enforce it as a matter of domestic law. States face considerable collective action problems, however, being economic actors themselves, whether in promoting or attracting foreign investment, which goes a long way toward explaining the relative weakness of public international law in this domain.
Thus, Larry Catá Backer concludes, in a masterful understatement: "from a public law perspective, the framework for the regulation of multinational enterprises can be viewed most charitably as in flux" (2007, p. 507). The fact that public law (national and international) does not generally encompass the economic unity of the multinational firm is the single most important contextual factor shaping its power, authority, and relative autonomy. Twenty-first century corporate globalization is built on foundational principles of corporate law that date back to the 19th century when they were intended to facilitate capital formation among natural persons: attributing legal personhood to corporations, investors' limited liability, and permitting one corporation to own another while still construing them to be separate legal entities.
I turn next to the matter of power.

Power
In political science, power is not treated as a "thing:" it is inherently relational, typically defined as the ability of A to get B to do something that B otherwise would not do (Dahl 1956). This can occur even when A does not undertake specific actions to achieve that outcome in any particular circumstance. To illustrate the political power of multinational enterprises, I draw on a typology developed by Doris Fuchs (2007). 6 The first is "instrumental power," the most traditional form of which is business lobbying. The second is referred to as "structural power," which may include companies' locational choice sets, the ability to transfer risks to suppliers, and generally the ways in which business gets things onto or keeps them off the policy agenda. The third has come to be called "discursive power," which refers to the ability by business and business associations to frame and define public interest issues in their favorthat is, to shape ideas that then come to be taken for granted as the way things should be done, even for non-business entities like governments. I briefly illustrate each in turn.

Instrumental power
Among the most commonly discussed forms of businesses deploying instrumental power are political campaign contributions; the asymmetry in knowledge and privileged access to decisionmaking that business may have, particularly where standards are set by private bodies or "clubs" of state representatives (Büthe & Mattli 2011); and lobbying. In short, instrumental power involves the employment of specific resources to achieve one's aims. I focus here on lobbying because its evolution closely tracks corporate globalization itself.
Lobbying is undoubtedly as old as politics itself. Its particular forms have varied over time, for example, as pure patronage politics by political bosses was reformed out of existence, while today large-scale fundraising needs of politicians coupled with the revolving door from government regulator to corporate lobbyist occupy center stage. Certain key industries, led by financial institutions, were at the vanguard in advocating the deconstruction and reconfiguration of the international economic architecture that enabled the current wave of corporate globalization to take root and expand since the 1980s, the phenomenon widely depicted as neoliberalism (Crouch 2011). Here I want to draw attention to three current features of business lobbying that relate to global governance issues. The first two may be peculiarly prevalent in the US, but they may also involve foreign corporations listed in the US as well as having other international spillover effects. The third is a direct result of corporate globalization everywhere.
The first feature is the sheer magnitude of corporate lobbying in the US, both direct and indirect (through business associations, which may have foreign corporations as members, and through political action committees). The Washington Post cites data from the Center for Responsive Politics indicating that publicly reported corporate lobbying in the US was $3.24 billion in 2013, more than double the $1.57 billion in 2000 (Becker 2014). Yet the Post also reports that this undoubtedly understates the total because "a good chunk" of lobbying activity "has gone dark, which leaves expenditures invisible to the public or to regulators" (Becker 2014). According to Drutman (2015), business lobbying expenditures are more than 30 times the total lobbying spending for all labor unions and groups representing public interests combined.
A second feature involves asymmetries in whose views get heard in the policy process, and who wins in court when regulators are challenged. Two political scientists have addressed the first question in the US context, using a data set of 1,779 policy issues between 1981 and 2002. They differentiate among four types of political constituencies, one being business groups that typically include foreign firms listed or operating in the US. Their conclusion is sobering: "Multivariate analysis indicates that economic elites and organized groups representing business interests have substantial independent impacts on U.S. government policy, while average citizens and mass-based interest groups have little or no independent influence" (Gillins & Page 2014, p. 564). That long-standing staple of American politics, the "median voter," today appears to have "only a miniscule, near-zero, statistically non-significant impact upon public policy." By all accounts, the center is not holding in many European countries either. 7 Moreover, it seems that litigation may have become simply an extension of lobbyingand of course law firms are involved in both. Lose in Congress when legislation is written, then pressure the agency drafting the regulations, and finally, sue the regulator for issuing regulations that are depicted as being too "intrusive," too "costly," or violate some constitutional right that the same Supreme Court has attributed to corporations as legal persons. This, too, can have significant international consequences. For example, the Dodd-Frank Wall Street Reform and Consumer Protection Act was aimed at repairing the US financial sector. But it included other provisions as well. Its Section 1504 requires US corporations in the oil, gas, and minerals industry to report their payments to foreign governments at the level of the individual project, in an effort to stem official corruption and the so-called resource curse. The American Petroleum Institute (API), the US Chamber of Commerce, and others brought suit against the Securities and Exchange Commission on some of the grounds enumerated above. The only API member to publicly dissociate itself from the lawsuit was Norway's Statoil, still partially state-owned (Geman 2013).
The most direct change in corporate lobbying that globalization has brought about is, of course, to expand its locus to the international level. Brussels has become a major magnet for corporate lobbying, given the regulatory scope of the EU. Fuchs references a German study that found: "75% of all associations represented in Brussels are business associations (to which one has to add all of the representatives of individual corporations…), while unions make up less than 5%" (2013, p. 83). Microsoft reportedly spends more money on lobbying in Brussels, the EU's capital, than any European companyalthough apparently primarily to promote grievances against its competitor, Google, on anti-trust grounds rather than to undermine specific public interests (Hakim 2015). Geneva has also become a magnet thanks to the presence of the World Trade Organization (WTO) and the World Health Organization (WHO). For example, pharmaceutical companies lobby to get their products onto the WHO Model list of Essential Medicines because they are then adopted as "must have" by more than 130 countries.
Among the businesses at the egregious end of the lobbying spectrum in Geneva in the past few years have been tobacco companies. With sales in the industrialized world plummeting, the major tobacco companies turned their attention to developing countries. WHO led a successful effort to negotiate a Framework Convention on Tobacco Control (FCTC), its first treaty ever, in order to create a level playing field in combatting what the FCTC describes as "the globalization of the tobacco epidemic" (World Health Organization [WHO] 2003). Opened for signature in 2003, the FCTC has now been ratified by some 170 countries. In essence, it asks signatories to develop adequate domestic regulatory frameworks. It lays out two types of recommended measures to reduce demand: price and tax measures, and such non-price measures as regulating the content of tobacco products, packaging, and labeling; tobacco advertising and sponsorship of events; and awareness raising and education. No specific content is provided for any of these recommendations so as to provide states with flexibility as how to best implement these measures given their particular circumstances. Furthermore, issues of liability are also left to individual signatory states to decide.
The major tobacco companies (British American, Philip Morris, RJ Reynolds, and Brown and Williamson) responded aggressively to this effort, and its very comprehensiveness attracted attention as a possible sign of things to come in other controversial global policy areas. Thanks to exhaustive research by a group of academic public health specialists who reviewed thousands of pages of documents and conducted numerous interviews, a public record exists of the lobbying strategies the companies used (Weisshaar et al. 2008). Only a brief summary of findings is possible here. WHO's right to convene a treaty negotiation was challenged. Claims were made that the FCTC would conflict with other international treaties. Developing countries were warned that the treaty would undermine national sovereignty, good governance, and prospects for economic development. Even some of the scientific evidence of the adverse effects of tobacco was questionedand some scientists willingly enlisted for this purpose. Specific national delegations were targeted for pressure in Geneva and national capitals. Alliances were sought with unrelated business associations that might have an interest in resisting the potential contagion effects of this instance of international regulation. Joint ventures with governments were proposed, weak national policies were encouraged, CSR initiatives were promoted as an alternative, and freedom of expression in labeling invoked. Needless to say, such an effort required sophisticated professional coordination. Multinational firms specializing in such services were enlisted. Essentially, those same services were provided by public relations and crisis management firms advocating climate change denial on behalf of their clients.
The tobacco multinationals at first had sought to prevent the treaty negotiations from taking place at all, then to undermine them, and failing that, to weaken the final product. They did not succeed in their first aim, and more research is needed to determine how "well" they did on the other two dimensions. At the end of the day, there now is a widely ratified treaty that leaves implementation measures to national discretion. But the story doesn't end with the FCTC. As individual countries began to act on its provisions they faced lawsuits by those same companies under bilateral investment treaties (BITs): for trademark infringements and regulatory takings if the required packaging turned out to be particularly gruesome, as indeed some are, or if plain packaging was required. Several suits are still ongoing but Philip Morris lost cases against Australia (Hurst 2015) and Uruguay (Mander 2016).
In sum, lobbying is a standard form by means of which business exercises instrumental power. There is more of it in major centers of governance with international influence, be it Washington or Brussels; it has gone global, as in the case of the FCTC; and if business loses at the legislative level, legal actions against regulators (national, regional, and global), has become a routine extension of lobbying.

Structural power
Structural power exists but it is easy to reify. James Carville, a long-time advisor to President Clinton, is reported to have said that if there is reincarnation "I would like to come back as the bond market. You can intimidate everybody" (The Economist 1992). However, except in special cases, bond markets haven't done much intimidating of late. The Institute for Policy Studies in Washington, a left-of-center think tank, appears to have come up with the idea of measuring the relative structural power of multinationals and states by comparing their annual turnover with states' gross domestic product, concluding that, in 2000, 51 of the world's 100 biggest economies were enterprises (Anderson & Cavanagh 2000). Economists and others have pointed out ever since that this is an apples and oranges comparison that overstates the economic "size" of multinationals vis-à-vis states by perhaps as much as four-fold. The reason is simply that gross domestic product measures value added, whereas annual turnover measures final sales without subtracting the intermediate costs that went into producing the ultimate good or service. Measuring the latter is not easy, but a United Nations Conference on Trade and Development (UNCTAD) study estimated that conducting roughly comparable calculations left only two multinationals on the list of the top 50 "economies" (UNCTAD, 2002)and questioned how much substantive sense such a comparison makes in the first place. However, this has not deterred the continued use of this comparison to "demonstrate" how much more powerful multinational firms are than many of the word's states.
In general terms, structural power refers to the ability of "business" or a specific business entity to affect outcomes in its favor without having to exercise instrumental power to achieve it (Fuchs 2007(Fuchs , 2013. Multinationals enjoy a number of intrinsic sources of structural power. Above all, states are territorially fixed entities competing for investments while multinationals typically have locational options, in some measure even in the extractive sectors. As an UNCTAD report puts it: "For many GVC [global value chain] segments, tasks and activities, there are relatively few "make or break" locational determinants that act as preconditions for countries' access to GVCs" (2013, p. 144). In addition, the juridification of private international commercial relations has expanded immensely through the so-called new lex mercatoria or merchant law, which has the effect of "delocalizing" not only commercial transactions but also related lawmaking and enforceable dispute settlement (Cutler 2003;Sassen 2006;Callies & Zumbansen 2010;Ruggie & Sherman 2015;Robé 2016). Particular outcomes, of course, are determined by a host of additional factors. In the following, I provide several illustrative forms of multinationals' structural power in relation to state-based governance.
The first concerns the right of foreign investors to sue governments under binding international arbitration, included in BITs and investment chapters of free trade agreements. This is a right not granted to states. Arbitrators are required to apply "the law applicable," which is the investment agreement, and not concern themselves with broader public interest concerns. Unlike in the WTO, no appellate procedure exists to challenge rulings. The awards for violating an agreement and losing the case can be steep, and they are legally enforceable. 8 The largest award in the history of the International Center for the Settlement of Investment Disputes, the most widely used venue, is $2.3 billion handed down in 2012 in a case brought by Occidental Petroleum Corp. against Ecuador (Chen & Bento 2012). A standard view is that states win at least as many cases as multinationals, which could suggest that no structural power is in play (Franck 2009). But as Mann (2015) notes, investors win over 70 percent of jurisdictional determinationswhether the treaty preconditions for arbitration have been met, which could be interpreted as a low bar to entry for investors. Moreover, once jurisdictional decisions against them are subtracted, investors go on to win some 60 percent of cases on the merits: The combination of low bars to entry and high win rates…is perfect for motivating investors, counsel and thirdparty funders to initiate cases. Importantly, it also increases the effectiveness of threats by investors against states in response to potential measures states may be considering adopting-the regulatory chill impact of the investor-state arbitration system.
In any event, Mann concludes, "states never win; they only do not lose," in which case at best they may receive an award of some of their legal costs. It is worth noting that China, whose outward foreign investment is on the verge of exceeding inward flows, has broadened the scope of issues it has agreed to subject to international arbitration in its most recent BITs (Zhang 2016).
A second illustration concerns transfer pricing through related-party trade (Rugman & Eden 1985; Grubert & Mutti 1991). This may enable firms to administer or negotiate prices that differ from market prices. Fine-slicing production by multinationals and the resulting increase in intermediate goods trade increases the opportunities for cross-border transfer pricing and also makes it more difficult to combat. The result can be "leaking" of fiscal revenues from some of the countries involved (Christian Aid 2009; UNCTAD 2013, p. 156). Official statistics are of only limited use; most countries collect few, some none. But here are two things we do know. First, how closely transfer pricing tracks arms-length market prices varies depending on whether comparable market prices exist. Where they don'tfor example, where some form of intellectual property is involved, or where the service component in transactions is highmultinationals have considerable pricing discretion (UNCTAD 2013). Second, studies across different industries using US government data show that intra-firm prices deviate significantly from arms-length prices if major tax and tariff considerations are in play. Transfer prices tend to be higher for goods sent to countries with low corporate tax rates and high tariff rates (Lanz & Miroudot 2011). Unless and until tax and tariff rates are harmonized, this remains a source of structural power for multinationals.
Third, multinationals' structural power is greatly augmented by the existence of tax havens. These have increased rapidly in number and magnitude since the 1970s. Their number now may exceed 50 (Palan et al. 2010). According to President Obama's Chairman of the Council of Economic Advisers, American-controlled corporate profits retained in the Cayman Islands alone amount to 1,430 percent of the island's gross domestic product (Cohen 2014). Tax havens are not simply a place to park profits, however. They are also a vehicle for companies to book investments to third countries, making the tax haven the investors' "home country," and the profits from the investments can then be returned to the low/no tax haven, continuing the cycle. That is why, for example, much of the money that flows into the British Virgin Islands as foreign investments flows out again to third countries (UNCTAD 2015, p. 188). A nominal presence in tax havens for the purposes of tax avoidance is not limited to any particular sector. For example, in the extractive and infrastructure industries a shell company registered in a tax haven may sell or lease capital equipment to an affiliate belonging to the same corporate group and operating in a third country, offering transfer pricing opportunities. Or instead of the parent company repatriating its overseas profits, its shell company affiliate in the tax haven can issue loans to the parent company, which then permits the parent company to deduct the interest payments from its home country tax obligations. The OECD and G20 refer to such practices as "base erosion and profit shifting," and have been developing guidelines to reduce it (OECD 2015), while the European Commission has found that Ireland's tax breaks for Apple constitute illegal state aid and has demanded that Ireland claw back 13 billion euros in taxes plus interest from Apple (Kanter & Scott 2016). That case is under appeal. But the ability of multinationals to augment their structural power through tax havens is unlikely to disappear anytime soon.
A fourth source of structural power stems from how little is known about trade flows at the firm level, and what impact this may have on official trade policy. Since the very first General Agreement on Tariffs and Trade round a half century ago, the prevailing trade discourse has been about "free trade." Of course trade is far freer today in the sense that tariff barriers are lower, markets are more open, and more goods and services are exchanged than ever before. But at the same time, a rapidly shrinking portion of trade takes place internationally as traditionally understood: among nations, at arms-length, and at prices determined by demand and supply in the marketplace. What we call "international" trade today, as we saw earlier, is increasingly "internal" trade within multinationals and with parties related to them contractually. But no public institution anywhere has the mandate or capacity to collect systematic and universal data on such trade. That this raises potentially serious questions about the efficacy of trade policy did not escape then WTO Director General Pascal Lamy in a plea to corporate leaders: It no longer suffices that you trade while relying on governments to craft the regulatory framework for you in the WTO through which your trade relations take place. You must provide the "evidence," through your trade experience, of what is actually happening on the ground, and must guide us in how to make things better. (Lamy 2011;also see Maurer & Degain 2010) A strong case can be made that the WTO itself, not knowing "what is actually happening on the ground" when it comes to firm-level trade activities, constitutes a significant source of multinationals' structural power.

Discursive power
Discursive power is the ability to influence outcomes through promoting ideas, setting social norms and expectations, and even shaping identities. Its exercise involves persuasion and emulation, not coercion. "Empirical analyses of the discursive power of global companies are difficult, due to the subtle nature of this power…" writes Fuchs, a leading exponent of the subject (2013, p. 86). Her references range from companies conducting focus groups and sponsoring TV advertisements to promoting the idea of free markets and limited government. For the purposes of the present discussion, I offer three broader observations.
The first is that corporate globalization has benefited from a massive shift in discursive power that favored business even if it was not always directly driven by business itself. This is the story the neoliberalism narrative tells (Crouch 2011). It involved displacing prevailing ideas, norms, and identities. Ideas, for example, in the intellectual assault on Keynesianism; norms in promoting the shareholder primacy norm; and identities in Thatcher's dismissal of the "nanny state" together with Reagan's powerful reframing in his first inaugural address: "Government is not the solution to our problem; government is the problem (Reagan 1981)." This shift had the result of normatively privileging the market and market actors, accompanied by deregulation, privatization, and outsourcing. The subsequent "third way" efforts by Clinton, the "New Democrat," and Blair, leader of "New Labour," struggled mightily to reposition their political parties and policies in alignment with this shift. By then, globalization had attained a near transcendent status. "Change is upon us. We can do nothing about that," President Clinton conceded in a speech. "[A] new global economy of constant innovation and instant communication is cutting through our world like a new river, providing both power and disruption to the people and nations who live along its course" (Clinton 2016, as quoted in Frank 2016. Both Blair and Clinton became ardent promoters of CSR initiatives so as to "put a human face on the global economy," in President Clinton's words (Clinton 1999). Corporate globalization benefited from this shift, and in turn reinforced it.
My second observation is that certain ideational elements of this broader shift can be closely related to the power of business. For instance, Teles' book, The Rise of the Conservative Legal Movement, takes us some way toward understanding the origins of key ideas concerning regulation in the US. It is a detailed scholarly analysis of deliberate intellectual institution building, beginning with the new field of Law and Economics. This sought nothing less, according to Teles, than "to undermine the intellectual foundations on which [legal liberal] arguments, and its claim to represent the public interest, were based" (2008, p, 90). Over time, the effort fundamentally altered dominant conceptions about when regulation was called for, how regulations were to be designed, and how regulatory agencies were expected to calculate the costs and benefits of proposed actions. The creation of that movement was followed by the founding of the Federalist Society, now the most important organization of conservative legal professionals in the US, of which a significant number have found their way into the upper reaches of US courts. Moreover, conservative think tanks were established, including the American Enterprise Institute, the Heritage Foundation, and the libertarian Cato Institute and Competitive Enterprise Institute. Their active research agenda promotes limited government, lower taxes, fewer regulations, and it seeks to cast doubt on global challenges, such as climate change, which might require policies at variance with their preferences. Beyond the think tanks lie such entities as the American Legislative Exchange Council, which among other things drafts model legislation for individual state legislatures to undermine and reverse progressive laws on the books. Initially, this investment in generating new content for corporate discursive power and driving it into the policy and judicial process was the brainchild of conservative and libertarian owners of large and typically familyowned companies, but it soon began to enjoy broader corporate financial support (Mayer 2016). Some of this thinking inevitably spilled over into the international realm.
My third observation is of a "dog-not-barking" kind. Raymond Vernon, a pioneer in the study of multinational enterprises going back to the 1970s, published a book in 1998 entitled In the Hurricane's Eye: The Troubled Prospects of Multinational Enterprises. His decision to write the book, he states in the preface: …grew out of a sense that the world was slipping into a period in which the inescapable clashes between multinational enterprises and nation-states might be growing in frequency and intensity, evoking responses from both the public and the private sectors that would substantial[ly] impair their performance. (Vernon 1998, pp. vii-viii) Yet today, the multinational enterprise is the standard mode of organizing economic activities across countries. Of course there exist different national variants of multinational firms, as well as different types of ownership and governance structures. But the convergence around the multinational corporate form is universal. Perhaps nowhere is it more remarkable than in the case of China. As Wilks puts it: Multinational corporations are now welcome in China and Chinese corporations have become the dominant domestic economic organization and have established power in global markets…Indeed a revolutionary change, but not quite the one that Marxist-Leninist-Maoist theory anticipated. (2013, p. 58) Nor have the "inescapable clashes" between multinationals and nation states that Vernon foresaw occurred. Among policymakers, the multinational has been normalized. This may simply reflect their greater efficiency in providing access to investments and markets, thereby creating economic opportunities and stimulating growth. The economic transformation in emerging market countries and many developing countries lends strong support to that view. But that the convergence has occurred so rapidly and so thoroughly suggests that mimetic and normative factors have also been in play. "Mimetic pressures take the form of copying or contagion, of borrowing models that appear effective and offer convenient solutions" (Wilks 2013, p. 58). The normative dimension comes into play when such consequentialist considerations are supplemented or even yield to the logic of appropriatenessthat this, not that, is the appropriate and expected form of conduct. Whatever combination of factors best explains the outcome, it endows the multinational with a reservoir of discursive power that it can draw upon in pursuing its interests.
In sum, the institution of the multinational has considerable transnational instrumental, structural, and discursive powerand of course the three are dynamically related. Contrary to early theorizations of these developments, however, this does not necessarily come at the expense of the territorial state as a political institution (Strange 1996); the two are too closely interwoven to support that argument. Examining multinationals' source and exercise of authority sheds further light on this relationship.

Authority
The boundary between power and authority is blurry. The key difference between them lies in the voluntary suspension of individual judgment based on a widely accepted and institutionalized belief that the authoritative entity is entitled or has the right to prescribe. Weber's (1978) classic categorization of the sources of authority differentiated between charismatic, traditional, and legal sources. He had the long-term evolution of domestic western society in mind. Leaving charisma aside, it seems clear that multinational enterprises today draw upon and embody a combination of traditional and legal authority. The "traditional" sources of multinationals' authority are the principles of private property rights and freedom of contractsacrosanct and codified in the liberal societies in which they first emerged as constitutional prerogatives of autonomous natural persons. Today they form the foundational authoritative basis for the agglomeration of worldwide multinational enterprises (Robé 2009). Crucially, even states that lack political liberal institutions domestically, such as China, adhere to this transnational authority structure so as to be in a position to participate in and benefit from the global economic system. Core elements of this traditional source of authority are enshrined in, elaborated by, and enforced through public and private law, including obligations under the WTO and international investment agreements.
Let me not be misunderstood. The host state is the public authority in any particular country. It has the right to determine certain parameters of the operations of a multinational's local subsidiaries, affiliates, and contractual parties. Compliance with applicable laws is a formal requirement, although it is not always rigorously enforced or elements may be waived altogether, as in export processing zones. The state can require the multinational to take on local joint venture partners. It can require permits the company needs in order to operate. It negotiates the taxes and royalties the company must pay. It can refuse access to particular sites. Competition authorities can regulate mergers and acquisitions. A number of states have gone so far as to demand that multinationals share encryption keys of their communication systems with local authorities, or use Internet servers located in the host country. But this isn't the whole story of the relationship between the public authority of the state and the private authority of the multinational.
To take a specific example, Disney had to share the keys to the Magic Kingdom to seal its deal with the Chinese government for the new $5.5 billion Shanghai Disney Resort. According to a New York Times report (Qin 2016), Disney had to dial back on several demands, including a TV channel it was eager to have for cross-branding purposes, and it had to give the government a majority stake in the resort, as well as a minority share in the Disney management company that runs the property. Disney's CEO Robert A. Iger notably describes the arrangement as "authentically Disney and distinctly Chinese." For good measure, when Disney's blockbuster musical "The Lion King" opened at the resort, it included a new character, the Monkey King. This is a figure of Chinese legend recently popularized in a Hong Kong fantasy film, but based on a classic 16th century novel by Wu Cheng'en entitled Journey to the West. The Chinese government, in turn, committed to providing Disney special protection from intellectual property piracy, which remains common in China.
These examples illustrate the obvious point that states possess authority in their dealings with multinationals. Specific outcomes will vary depending on the balance of interests and power; not every country is China, and not every company Disney. But they also illustrate a fundamental institutional fact: a dynamic interplay between two different centers of power, each with its own basis of authority. One is transnational and rests on private property rights, and the other is territorial and rests on sovereignty. The power of each in some measure is constrained by the authority of the other. Neither supplants the other. They co-exist: "authentically Disney and distinctly Chinese." To elaborate on the scope and scale of multinationals' authority, I address two further questions: over what or whom do they have authority? And in the next section I ask: in whose name or on whose behalf do they exercise that authority?
Multinationals have authority over themselves. This is not as trite as it may sound when we consider the number of multinationals in the world today; the number of countries in which many operate; the range of activities they encompass; the already vast and still expanding private transnational legal order they have generated; and their capacity to affect workplace conditions, the welfare of communities, and even national economic prospects around the world. The difference between the multinational as a single economic organization and the group of separate entities recognized under the law is critical to understanding the scope of multinationals' authority over "themselves." The economic organization, acting through one legal "self" (often called the corporate "parent"), has the authority to create and structure the other legal "selves" (which of course are integral parts of the single economic organization) in such a way as to optimize the entire group's interests throughout its transnational sphere of operation, as well as to limit its liabilities. The economic organization sets the strategic objectives for the entire enterprise. It decides where and how to allocate its assets, which is based not only on such factors as market size, labor costs, or promising natural resource deposits, but also on selecting or constructing favorable regulatory environments through the global market for legal norms. For some purposes, this now includes the possibility of combining the most favorable regulations of different countries within in a single contract (Eidenműller 2011). The economic organization sets terms of employment. It sets product and performance standards. It institutes codes of conduct. In doing so, multinationals have authority that states lack. For example, through their codes of conduct they can require suppliers in host states to adhere to social and environmental standards that, if imposed by the country importing those products, the host state could challenge as a non-tariff barrier under the WTO. The ability of multinationals to enforce such "internal" decisions across countries is the envy of states and intergovernmental organizations.
The authority of multinationals is not limited to their subsidiaries and affiliates. It also extends to contractors, franchisees, and other types of non-equity counterparties. In some industries, such as consumer electronics, toys, and ready-made garments, suppliers (and their subcontractors) may serve multiple multinationals simultaneously, and thus are subject to multiple and sometimes conflicting corporate authorities. Thus, workers in the same factory can be subject to different labor standards, for example, while each may differ from poorly enforced local labor laws, if they exist at all (Locke 2013).
The vast web of non-equity relationships in global production networks are structured through the private law of contracts and are subject to private international commercial arbitration. Through its International Court of Arbitration, the Paris-based International Chamber of Commerce hosts the world's leading facility. This has transformed business lawyers and arbitration centers into institutionalized global "governors." In sum, saying that multinationals have authority over themselves is actually saying a lot. Their authority is clearest and most direct when it is administered internally within the corporate group. From there it radiates outward across networks and down supply chains via the private law of contracts. The ultimate boundaries of multinationals' authority, therefore, are difficult to discern, not unlike the gradual frontier zones in early modern Europe before clear national boundaries were fixed (Ruggie 1993).

Relative autonomy
My second question is in whose name or on whose behalf multinationals exercise the authority they have. Common sense suggests that it must be their owners. Some multinationals are family owned or privately held. Others are state owned. Publicly traded companies, it has traditionally been said, are owned by their shareholders, thus presumably exercising power and authority on their behalf. But there are two problems with this premise in today's world.
First, many investors move into and out of individual stocks of large corporations several times a day or hold them very short periods of timesometimes mere secondsusing a variety of trading algorithms and automated means to make a quick return. Hedge funds are the single largest trader in US equity markets (Sethi 2015). Five of the world's most heavily traded equity securities are exchange-traded funds (Authers & Newlands 2016). These and other such investors are not "owners" in any traditional sense of the term; they have no per se interest in the corporations whose shares they are trading. They are speculators and arbitrageurs, day traders, flash traders, or indexers who buy the whole market or some slice of it without "valuing" the price of any one stock as a normal owner would want to do.
The second problem runs deeper. Legal scholars in particular have long struggled with the question of shareholder ownership of the public corporation and what exactly it means (Pickering 1968). From a very different analytical point of view, economics Nobel laureate Eugene Farma stated long ago: "ownership of capital should not be confused with ownership of the firm" (1980 p. 290). However, Friedman's (1970) property conception of the firm, coupled with principal-agent theory (Jensen & Mecklin 1976), and the shareholder primacy norm (Hausman & Kraakman 2001) dominated theory, and to some extent, US practice, for the past generation. It now seems to be on the wane (Smith & Rnnegard 2016). Robé, for one, rejects the idea that anyone owns the large public corporation: After the process of incorporation, shareholders have no right of access to the assets of the corporation; they do not enter into any contract in its name. No liability can arise for them from the corporate activity. They do not run the corporation and do not own it." (Robé 2012, p. 6, emphasis in original) And unless a shareholder has a controlling interest, shareholders' rights are limited: "a right to dividends; and the right to participate in shareholder assemblies and vote on rare decisions…But owning shares does not give title to the corporation or to a portion of the corporation" (Robé 2009, p. 868). Even the right to dividends, should the enterprise generate a profit, is subject to the discretion of directors, acting in the corporation's interest. All of this is true of national firms, not only multinationals. The difference, of course, is that multinationals exercise the power and authority they have transnationally, in a global system that lacks a central regulator, and in some cases across more than 200 national jurisdictions.
Clearly, directors and management of a publicly traded firm need to keep investors sufficiently happy to avoid having them sell off their shares, causing a drop in the company's market value (and in executive compensation). The owners of shares thereby constitute a market force that constrains directors' and management decisions. Moreover, holders of large blocks of shares, such as institutional investors, can exercise influence through board elections and more informal means. But neither of these situations makes them owners of the firm.
Thus, there appear to be two possible answers to the question of who owns publicly traded firms: they own themselves, or no one doesand in effect they amount to the same thing. There appears to be only one answer to the question on whose behalf multinationals exercise their authority: their own. In sum, the institution of the multinational has come to constitute not only a significant center of global power but with the exception of state-owned enterprises, also a relatively autonomous transnational authority structure.

Conclusion
This article set out to elucidate the multinational enterprise as a global institution in terms of its power, authority, and relative autonomy. To better understand the status of the multinational in these terms, it is critical to differentiate between the economic organization that is able to act under unity of command across its entire spheres of operations, and the separate legal entities within the multinational corporate group. With only limited exceptions, law governs the separate entities, not the enterprise as a whole. This disjuncture between economic reality and legal convention is the single most important contextual factor shaping the global institutional status of multinationals. And it creates a global governance gap that defies standard and simple solutions. This is true both of CSR as a voluntary form of self-regulation or co-regulation, and attempts to devise a mandatory treaty-based regime to govern the conduct of multinationals.
The scope conditions of CSR have never been adequately defined: just how far can we expect it to reach, even under ideal circumstances? How much can it help to constrain the excesses of lobbying, and litigation as its extension? Is tax evasion and transfer pricing within its scope? What about the asymmetrical investor/state dispute settlement system, which may have a chilling effect on progressive legislation? Or equating financial clout with which to influence elections with constitutionally protected free speech? And so on. All of these have significant consequences for the people and planet agenda that CSR seeks to address.
At the same time, creating an overarching legal regime, whether within human rights law as the current Ecuador and South Africa led initiative has framed it, or some other framing, seems highly implausible, not only on political but also on sheer logical grounds. It would involve harmonizing aspects of often vastly different bodies of national, sub-national and international lawfor starters, investment law, trade law, corporate law and securities regulation, tax laws, consumer protection law, labor law, anti-discrimination law, other areas of human rights law, and criminal law, and impinge on underlying conceptions of property rights and private contracts. The point is not that these are unrelated, but that they embody such extensive problem diversity, institutional variations, and conflicting interests, not only across states but even within them, that any attempt to aggregate them into a general treaty, a global constitutional order of sorts, would have to be pitched at such a high level of abstraction that it would be without practical meaning.
This leads to two implications for future work in this area. The first is that scholarship and practice should strive to better understand the limits of both CSR and the pursuit of international treaties governing multinational enterprises. As measured against multinationals' power, authority, and relative autonomy, the former currently underreaches while the latter overreaches. CSR by itself is highly unlikely to take us far enough, and the repeated pursuit of an overarching constitutionalizing treaty is doomed to repeated failure. The second implication is that much greater attention should be paid to the dynamic interplay between the two spheres, and its potential cumulative effects. Under what conditions can CSR satisfactorily substitute for mandatory measures? What public regulatory measures can generate cascading effects in the CSR sphere? When can soft-law standards lead to legal hardening through issue-specific focused juridification at national and international levels? Scholars are only beginning to address these questions. 9 One conclusion seems inescapable: in light of the multinational's power, authority, and relative autonomy, the time-worn mandatory/voluntary dichotomy inhibits rather than advances our coming to grips with the challenges posed by corporate globalization.