Deliberation Without Democracy in Multi-stakeholder Initiatives: A Pragmatic Way Forward

Political CSR scholars argue that multi-stakeholder initiatives (MSIs) should be designed to facilitate deliberation among corporations, civil society groups, and others affected by corporate conduct for their decisions to be considered democratically legitimate. However, critics argue that decisions reached within deliberative MSIs will lack democratic legitimacy so long as corporations are granted a role in helping to make them. If the critics are correct, it leads to a paradox. Corporations must be excluded from holding decision-making authority within MSIs if they are to function as democratically legitimate regulatory institutions at a global level. However, this risks severely diminishing the incentive of corporations to support and participate within MSIs, which often depend heavily for their success on the visibility provided them by corporate participants. In this paper, I argue that this apparent paradox should be considered irrelevant to the future study of MSIs since it is both unnecessary and impractical for researchers to focus on establishing democratically legitimate systems of governance within them. Instead, I recommend an approach informed by three touchstones of pragmatic philosophy to guide their future study—a criterion of usefulness, wariness of category disputes and commitment to experimentalism. I conclude by drawing on research in political science and social psychology that demonstrates an important practical role for deliberation within such organizations, arguing that researchers must zero in on the role that inclusive deliberation can play in bolstering their effectiveness as regulatory instruments.

Corporations today have joined national governments and international governmental organizations to play an increasingly politicized role as global governance actors involved in the process of establishing international rules and standards for the conduct of business (Matten & Crane, 2005;Scherer & Palazzo, 2007;Scherer et al., 2006). In light of these changes and with the declining ability of the nationstate to police corporate action in a globalized world, "political CSR" scholars have argued that corporations must seek to understand their increasingly state-like role in terms of democratic political legitimacy typically reserved for formally representative institutions (Matten & Crane, 2005;Scherer & Palazzo, 2007;Scherer et al., 2006). Drawing in particular on Jurgen Habermas's deliberative democratic theory to structure their ideas about the legitimation of these activities, many of these scholars argue that a deliberative democratic "institutionalization of accountability" can "provide a substitute, or at least a complement, for problems that have a non-territorial nature and cannot be solved by national governments" (Martens et al., 2017;Scherer & Palazzo, 2007;Scherer et al., 2006).
Proponents of these views claim that inclusive and authentic deliberation among corporations, civil society actors, and other stakeholders regarding the regulations that should bind corporations has two effects. First, it ensures that these decisions should be considered democratically legitimate despite the absence of authorization through the formal channels of representative government. Second, it will consolidate corporate moral legitimacy in the face of public resentment and questions about their contributions to the common good (Hahn & Weidtmann, 2016;Mena & Palazzo, 2012;Scherer & Palazzo, 2007).
Several recent articles have focused on how multi-stakeholder initiatives (MSIs) might be designed according to principles of deliberative democracy and thereby ensure that 1 3 both standards of legitimacy may be realized (Hahn & Weidtmann, 2016;Martens et al., 2017;Mena & Palazzo, 2012). This work has generally not addressed important questions about whether the participation of corporations in collective decision-making within MSIs might impact the democratic legitimacy of their decisions. However, deliberative democrats have typically held that, like other associations in a democracy, the interests of corporations only deserve consideration insofar as they will impact those of rightsbearing individuals. Hussain and Moriarty (2018) take up this strain of criticism in an incisive recent essay, in which they note that corporations are neither sentient, rights-bearing individuals entitled to direct decision-making authority nor politically representative organizations authorized to act on behalf of citizens. As a result, they claim, MSIs can only issue regulations with legitimate democratic authority if corporations are excluded from a say in making final decisions. At best, corporations might serve as technical experts who may be consulted by deliberators but cannot play a role in issuing final decisions if those decisions are to be considered democratically legitimate.
In light of these arguments, it appears that political CSR scholars are faced with a paradox: we can only establish democratically legitimate regulatory institutions at a global level by severely diminishing the incentive of corporations to participate within them by eliminating their role in decision-making. However, this in turn risks undermining their authority as regulatory instruments, which often depends heavily on the visibility provided them by corporate participants. Democratic legitimacy, to which political CSR scholars have attached tremendous importance and in the name of which they have defended the significance of internally deliberative procedures, can only come at the risk of diminishing the effectiveness of these institutions.
In this paper, I present two main arguments against continuing to focus on the role of democratic legitimacy within the administration of MSIs and other such organizations. First, I demonstrate that the apparent paradox need not concern us: it is unnecessary for MSIs to fully satisfy deliberative democratic legitimacy requirements given that, unlike representative governments, their methods of securing regulatory compliance do not involve the use of coercion backed by the threat of force. Furthermore, their effectiveness as regulatory instruments, and therefore their ability to solve problems of collective concern and secure corporate moral legitimacy, depends on corporate participation. This leads to the second main argument of the paper: that the future of MSIs is very much in doubt, and focusing on such category disputes detracts from the role that scholars might play in bolstering their power to improve human lives. I argue that, rather than continuing to debate about such matters, we must adopt a more practical approach, applying core insights from pragmatist philosophy to the practical study of such organizations. By adhering to a criterion of usefulness in our approach, we can avoid such category disputes by retaining a principled focus where it belongs: on improving human lives through concrete action that resolves immediate injustices and leads to comparative advances in organizational performance (Farjoun et al., 2015;Sen, 2018;Wicks & Freeman, 1998). To this end, we need not abandon a potential role for deliberation, but must shift our focus away from questions about the strict democratic legitimacy of MSIs and zero in on better understanding the role that inclusive deliberation can play in bolstering their effectiveness as regulatory instruments and securing corporate moral legitimacy.
Empirical research supports the idea that inclusive deliberation within MSIs can improve the decision-making capacities of individual participants and increase the quality of their collective output. At the same time, it helps to cultivate prosocial attitudes that not only improve their functioning over time but foster broader awareness of and engagement with business challenges of collective concern. In the process, deliberative MSIs can provide the added benefit of contributing to the broader promotion of democratic attitudes and competences in society.
Focusing exclusively on the pragmatic effects of inclusive deliberation in this way means that its use within MSIs should be considered open to criticism when other methods would lead to superior outcomes. As we shall see, however, research indicates that deliberation has a potentially important practical role to play within such institutions. I begin the paper by introducing the distinction between moral and democratic legitimacy and exploring how these ideas have been deployed as normative standards in the literature on political CSR. Next, I consider arguments for the idea that organizing MSIs according to principles of deliberative democracy will allow us to realize these standards, concluding that they cannot be simultaneously realized given the dependence of MSIs on corporate participation for their effectiveness. I then use the prisoner's dilemma model to show why the regulatory systems deployed by MSIs do not require full democratic legitimation. This leads to a broader discussion of the need for a more practical approach to the study of MSIs informed by pragmatist philosophy. Taking such an approach will help re-invigorate interest in the practical possibilities opened by the proliferation of MSIs and ensure that the attention of business scholars remains squarely focused on improving their performance as effective regulatory organizations that can help promote corporate moral legitimacy. To this end, I argue that we should redirect our attention away from questions of democratic legitimacy and towards how deliberation affects the way MSIs function, drawing on work from political science and social psychology to demonstrate deliberation's positive practical effects.

Two Models of Legitimacy
Political CSR scholars begin from the premise that the circumstances arising from globalization have led to a "governance gap" (Edward & Willmott, 2008;Matten & Crane, 2005;Scherer & Palazzo, 2007). The causes of this gap are well-known. In the international sphere, there is no central government with the authority to develop and enforce legal restrictions on the behavior of global corporations, leading to widespread exploitation. Among other things, in the absence of enforceable global regulations, corporations are often free to engage in corruption, particularly in developing contexts, and face incentives to drive down human rights (including labor rights) and environmental standards by inciting competition between nations for the jobs, tax revenue, and other benefits they can provide (Bakan, 2012;Klein, 2009;Stiglitz, 2002). These downsides of globalization have been explored in many places and there is no pressing need to dwell on them here.
Where some legitimate debate persists is over who is best suited to tackle these challenges and how best to do so. Political CSR scholars take a strong stand on this front, arguing that, in many cases, these problems can be solved by corporations themselves, who have the power and reach to regulate themselves in partnership with governments, international organizations and civil society groups. Some go even further than this, arguing that corporations must take such steps if they are to secure their global standing into the future Schneider & Scherer, 2013).

Moral Legitimacy
According to many of these authors, corporations have an urgent need to develop these regulatory regimes due to the declining corporate legitimacy conditions in our globalizing world. Palazzo and Scherer (2006) argue that corporations must initiate a public practice of deliberative engagement with government, civil society groups, and affected populations in order to secure public acceptance of their moral legitimacy or face threats to their continued existence. They build their claims around Mark Suchman's influential tripartite typology of organizational legitimacy, which includes pragmatic, cognitive, and moral forms. Each of these is a form of empirical legitimacy, which Suchman describes as "a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions" (Suchman, 1995, p. 574).
Pragmatic legitimacy occurs when the actions of an organization benefit its audience. For example, if the presence of a corporation within a community is clearly beneficial for the members of that community, it will achieve a degree of pragmatic legitimacy among them.
Cognitive legitimacy is the product of background conditions that hold a "taken-for-granted" character, against which the organization's role is deemed proper or desirable. For example, the system of banking and investment enjoys a high degree of cognitive legitimacy against the backdrop of capitalist society, where its basic principles go unquestioned.
Finally, moral legitimacy "reflects a positive normative evaluation of the organization and its activities." However, moral legitimacy is "sociotropic", meaning "it rests not on judgments about whether a given activity benefits the evaluator, but rather on judgments about whether the activity is 'the right thing to do'" (Aldrich & Fiol, 1994;Parsons, 1960;Suchman, 1995, p. 579). It describes a condition under which authority is seen as being morally justified by those over whom it is exercised, thereby preserving a stable social order (Beetham, 1991;Lipset, 1959;Suchman, 1995;Weber, 1958).
Drawing on Habermas's social theory, Palazzo and Scherer argue that the complexity of modern pluralist societies coupled with pressures wrought by globalization combine to undermine the background conditions from which the corporation once took cognitive legitimacy. "Globalization not only macerates the cultural background of the nation-state, it furthermore leads to a vivid debate on the interplay of state, economy, and civil society." As a result, they claim, "the once more or less homogenous cultural lifeworld background becomes fragmented… Values, interests, goals, and lifestyles are pluralizing and societies struggle with growing complexity and heterogeneity" (Palazzo & Scherer, 2006, p. 77).
A similar erosion of public confidence in the pragmatic benefits of corporate activity has arisen as the power of the state to promote social welfare continues to recede and people find themselves receiving less and less of the benefit from housing successful multinationals within their own countries (Piketty, 2015;Piketty & Zucman, 2014).
In this fragmented environment, they argue, moral legitimacy is essential to secure the corporation's global standing. Yet, corporate strategy is generally detached from these realities, as most corporations take an instrumental approach to CSR that does not adequately reflect these changing social conditions, focusing instead on the connection between good corporate conduct and profitability (Freeman et al., 2007;Porter & Kramer, 2011).

Democratic Legitimacy
Political CSR scholars often argue that the demand for a second, normative form of legitimacy is created by the role corporations play in establishing standards of conduct in 1 3 the international sphere. Hahn and Weidtmann distinguish empirical forms of legitimacy like those described by Suchman from normative legitimacy, noting that while empirical forms focus "on the actual acceptance of social rules, organizations, or structures," normative forms involve establishing "under which conditions such rules, organizations, or structures can be perceived as legitimate" (Hahn & Weidtmann, 2016, p. 96). Scherer et al. claim that, by developing and enforcing regulatory requirements, corporations are engaging in a form of political governance. Therefore, the normative legitimacy of such activities should be judged against the same kinds of democratic standards as the ordinary business of formally representative governments . In other words, the procedures through which these regulations are developed need to satisfy the normative requirements of democratic legitimacy.
As a normative concept, democratic legitimacy describes the terms under which the exercise of political authority is justified according to the moral standards of democracy. These standards are typically connected with a particular understanding of the individual rights, such as freedom and equality that democracy is intended to protect. While order and stability are important social goods for any society as are a range of other virtues associated with democracy, the reasons theorists offer for thinking it is justified generally go beyond its capacity to bring such benefits. In addition, they refer to the ways it preserves the freedom and equality of rights-bearing individuals who face conditions of complex interdependence requiring the use of authority to preserve order and stability (Cohen, 1997;Gutmann & Thompson, 1996;Rawls, 2005). As Alex Levitov notes, "it is important that the two notions be kept conceptually distinct, and that legitimacy in one sense of the term is not taken automatically to imply legitimacy in the other sense" (Levitov, 2016, pp. 1-2). 1 The risk in doing so in this case is to conflate the appearance of normative justification or the conviction among members of the population that a governance arrangement is justified with its actual justification.
Some political CSR scholars have claimed that, like a formally representative government, any institution designed for the purposes of establishing a self-regulatory regime in the international sphere must satisfy both kinds of legitimation requirements (Martens et al., 2017;Mena & Palazzo, 2012;Scherer, 2018). Furthermore, this work internalizes an assumption shared by most democratic theorists, that in a society where the exercise of authority is democratically legitimate, citizens will see that exercise of authority as morally legitimate, meaning the former will bring about the latter (Hahn & Weidtmann, 2016;Schrempf-stirling et al., 2016). Therefore, these authors have generally agreed that deliberative democratic theory offers an attractive combination of normative features and practical applicability to the international sphere, particularly given the impossibility of elected representation and the supraterritorial nature of these exercises in global governance. As a result, they propose it should be adopted as the standard of democratic legitimacy against which efforts at international corporate self-regulation are institutionalized and measured.

Democratic Legitimacy Within MSIS
Deliberative democrats offer a particular normative interpretation or refinement of the idea of democracy itself, which is founded on the basic idea that a society should be governed collectively by all of its members. They believe that the normative force of the laws and policies enacted by the state comes, not from the counting of votes (alone), but from the fact that they are the product of reasoned deliberation among free and equal citizens or their representatives who aim to establish shared terms of association that all similarly motivated citizens will find mutually acceptable (Benhabib, 1996;Cohen, 1997;Gutmann & Thompson, 1996;Habermas, 1996;Rawls, 2005).
Political CSR scholars have primarily referred to the pioneering work of Habermas in thinking about how the concepts underlying the theory of deliberative democracy might be applied in the international sphere to practices of corporate public engagement and standard-setting (Gilbert & Rasche, 2007;Hahn & Weidtmann, 2016;Hussain & Moriarty, 2018;Martens et al., 2017;Mena & Palazzo, 2012;Scherer & Palazzo, 2007;Scherer et al., 2006). In an early piece that did much to shape the subsequent agenda of political CSR research, Scherer and Palazzo "challenge the liberal conception of CSR," in which corporations are treated purely as private actors that "do not need to expose their decisions to public scrutiny, as long as they comply with the law and moral customs" (Scherer & Palazzo, 2007, p. 1106. By contrast, they argue, the principles of deliberative justification underlying Habermas's theory of democratic legitimacy should be applied to corporations in much the same ways as they are the state, since corporations are likewise involved in defining the permissible limits of their activities in ways that will broadly impact the public. They claim that "the deliberative model assumes that corporate as well as governmental actors depend on processes of civic self-determination (Habermas, 1996) and that there is no reason to exclude some spheres of society from democratic scrutiny" (Scherer & Palazzo, 2007, p. 1106).
Earlier work often described this idea in relation to more abstract conceptions of "civil society engagement" (e.g., Scherer & Palazzo, 2007, p. 1109). However, more recent work on the application of deliberative democratic principles of corporate self-regulation has focused almost exclusively on the practices undertaken within MSIs (e.g., Hahn & Weidtmann, 2016;Hussain & Moriarty, 2018;Martens et al., 2017;Mena & Palazzo, 2012). Often, these institutions are well-suited as vehicles for the kind of deliberative engagement recommended in the political CSR literature.
There is no universally agreed-upon definition of an MSI, which come in many different forms. As a first cut, MSI Integrity and the Duke Human Rights Center at the Kenan Institute for Ethics, authors of the largest database that currently exists on MSI structures and activities, define them as "voluntary initiatives in which more than one stakeholder group (i.e., industry, civil society, government, or communities affected by business operations) collaborates in the primary decision-making processes of the initiative" (MSI Integrity, 2017a). Beyond these initial criteria, MSIs must possess a pair of additional features of particular interest to political CSR scholars in order to be included in the database: (1) they will engage in standard-setting behavior, meaning they will not only "set standards for company or government conduct [but] require members to work towards compliance with the standard," and (2) the initiative will be "transnational, defined as implementing its standard in more than one country" (MSI Integrity, 2017a). 2 From the standpoint of political CSR, this definition is appropriately restrictive, since only MSIs meeting these additional criteria will be involved in the kind of regulatory activities believed to necessitate democratic legitimation.
Corporations take up the standards and regulations issued by MSIs on a voluntary basis. They are not legally binding nor are they enforceable in the traditional sense. Instead, they are instances of "soft law" regulation, meaning that their force comes primarily from the reputational benefits they afford to compliant participants or reputational costs to which they can lead for those who fail to comply or choose not to participate. In this instance, reputational benefits and costs refers to how firms are benefited or harmed by abiding (or failing to abide) by voluntary corporate responsibility standards about which there is broad public knowledge. When consumers who are aware of these standards can and do base their consumption decisions on whether these standards are met, the impact on corporate behavior will often mimic that of laws that impose formal punishment for noncompliance, but without need for the threat of force. I will return to this issue in a moment, but for now I will not question the assumption, held throughout much of the political CSR literature, that "soft" and "hard" types of regulatory power need equivalent forms of democratic legitimation in order to secure their normative validity and practical value.
Work that has taken on the task of applying deliberative democratic theory to the institutionalization of MSIs has generally assumed that, apart from the absence of formal elections, the actual structure of deliberative decisionmaking within them should mirror those in idealized representative governance arrangements. In this respect, they follow Habermas, who envisions the practice of establishing legitimate law in terms of a "two-track" model. According to that model, the legitimacy of modern law originates in the requirements of a discourse principle, D, which states that: "Just those action norms are valid to which all possibly affected persons could agree as participants in rational discourses" (Habermas, 1996, p. 107). However, as Scherer and Palazzo point out, the development of political institutions through which to secure conditions of modern democratic legitimacy does not merely involve a "(utopian) attempt at a large-scale application of the criteria of the ideal speech situation" described in Habermas's moral theory (Scherer & Palazzo, 2007, p. 1107. Instead, he calls for the institution of formal processes of rational deliberation among the members of representative bodies responsible for the development of modern law, combined with a relatively "wild" public sphere largely unbounded by the constraints of formal deliberative requirements. Habermas might have been led to these conclusions on grounds of practicality: modern societies are constrained from engaging in mass deliberation by problems of scale alone. However, his reasons are ultimately more principled than this. A public sphere featuring open-ended public discussion is unbounded by the demand to find immediate solutions to collective problems and is therefore a better vehicle for the discovery of social issues and concerns that are "problematized" and fed into the formal system of deliberative governance (on this point, see Habermas, 1996, p. 356). In Joshua Cohen's phrase, "This interplay between discovery and justification supports the presumption that the results will conform to idealized, discursive problem-solving. Because the two phases of reasoning in the actual process conform to idealized reasoning, the actual process will generate results like those that idealized discourse would generate" (Cohen, 1999, p. 401).
Habermas's overall approach to understanding the democratic legitimacy of modern law is therefore systemic in nature. He assumes that the "output" of formally deliberative legislative bodies is only democratically legitimate by virtue of the collection of social knowledge that takes place within the broader, relatively unstructured public sphere. This basic model of normative justification, in which legitimate decision-making processes involve a "filtering" of public discourses into formally deliberative bodies that are in some sense "representative" is repeated in several recent articles on the presence-or absence-of democratic legitimacy within deliberative MSIs. Mena and Palazzo (2012) offer one early example of this model in action. They compare the "input-legitimacy" of decision-making procedures within MSIs with those of the formal governing institutions in democratic nation-states, starting from the somewhat controversial premise that the primary way that elections and democratic representation contribute to the input legitimacy of democratic governance within the nation-state is in how they guarantee inclusion. In the absence of elections within MSIs, they argue that close attention must therefore be paid to establishing highly inclusive collective decision-making procedures. To compensate, they recommend that deliberative decisions should involve "all stakeholders affected by the issue, where affected means 'that decisions and policies significantly condition a person's options for action'" (Mena & Palazzo, 2012, p. 538). Most importantly, they argue, it is not the sheer number of stakeholder representatives involved in such decision-making processes that confers democratic legitimacy on their activities, but that it involves the "relevant ones" (Mena & Palazzo, 2012, p. 539).

Representation in MSIs
These actors lack the formal authorization and accountability established through elections. Nonetheless, Mena and Palazzo claim that their standing as legitimate deciders is rooted in the fact that, taken together, they represent an appropriately descriptive cross section of those who will be regulated by the procedures (corporations) and society (civil society actors and "all affected"). This serves as their defense of the idea that there is a domain of formal decision within MSIs paralleling the formal decision-making bodies of the nation-state, through which relevant issues holding "action potential" might be filtered from the public sphere into formalized regulation.
In a later article, Hahn and Weidtmann (2016) pursue a similar strategy, for similar reasons. They note that "a radical democratic representation (in terms of full participation of every single actor) is impossible to achieve." Therefore, "a representation of some form appears necessary in order to ensure the effectiveness of deliberative processes through information transfer as well as opinion and preference formation" (Hahn & Weidtmann, 2016, p. 104). The criteria for inclusion they endorse involves expert selection of qualified participants from different stakeholder categories, controlled to ensure balanced regional representation (Hahn & Weidtmann, 2016, p. 117). They too envision this formal decisionmaking apparatus as a venue for deliberation about ideas, complaints and problems that will be transmitted from the less structured discourses of civil society.
Hussain and Moriarty (2018) pose a strong critique of these ideas. Their primary concern is with whether businesses can really serve as "politically representative organizations" (PROs) who are legitimately authorized to engage in deliberation and regulatory decision-making on behalf of citizens. As an initial premise they establish that, unlike human beings, corporations lack sentience, meaning they cannot have interests entitling them to participate in the formal deliberative process of public decision-making. This is in keeping with how deliberative democrats have typically envisioned corporate participation rights. Such theorists generally deny that corporations hold "selfauthenticating claims" to see their own interests directly addressed in the process of such deliberations (see, for example, Mansbridge et al., 2010). Instead, they hold that, like other associations in a democracy, corporate interests only deserve consideration to the extent they will impact the rights and welfare of those human individuals who do hold such claims. Therefore, decision-making procedures that treat corporate interests as having fundamental importance equivalent to those human individuals are not consistent with deliberative democracy's moral foundation, which treats the rights-bearing individual as its exclusive locus of concern.
This distinction marks a recognition among deliberative democratic theorists that, while public organizations and associations are not only valuable but necessary for organizing our social and political lives, deliberative public decision-making practices must ultimately place primacy on the interests of rights-bearing individuals (Dryzek & Niemeyer, 2006;Habermas, 1996;Mansbridge et al., 2010). Provided their influence within such deliberations is suitably constrained to reason-giving, it is potentially valuable for corporations to participate in public deliberation in complex modern democratic contexts as information-providers and advocates on behalf of those whose interests are potentially at stake. But to grant them any more fundamental standing corrupts the ideal of democratic self-governance at the core of deliberative democratic theories.
Hussain and Moriarty provide a simplified version of this argument in their own account. In their words, …individual human beings have interests, and the individual human beings who are stakeholders in a corporation have interests. These individual human beings -e.g., shareholders, workers, and managersmay have a claim to participate in social deliberation, but any talk of the interests of corporations only makes sense as a proxy for the interests of individual human beings who have stakes in a corporation (Hussain & Moriarty, 2018, p. 527). 3 As a result of these factors, they argue that their best case for assuming an active role in organized deliberation regarding their own regulation is as PROs.
This leads to the more important claim for the present discussion, that unlike some other public associations that have a better claim to play such a function, corporations do not qualify as PROs, meaning they have no claim to speak on behalf of their "members" as representatives in public deliberation. At best, they argue, they might serve as "technical experts" within multi-stakeholder governance processes without participation rights in any decision (Hussain & Moriarty, 2018, p. 530). This is because, unlike political parties and activist NGOs like the NRA or Amnesty International, citizens do not "join" corporations as employees or shareholders with the primary reason in mind that the organizations match their social and political commitments. While this may enter their calculations on occasion, it is not the primary reason for the association's existence nor for most people's participation within it. 4 Taken together, these arguments rule out treating businesses as "supervising authorities" with a right to participate in formal deliberative decision-making processes within MSIs alongside legitimate representatives. 5

Surveying Options
This article began by outlining two forms of legitimation that political CSR scholars claim to value: the democratic legitimation of organizations devoted to developing regulations and the moral legitimation of corporations who are subject to their requirements. Based on this recent work, it appears we are left with two main options for institutionalizing deliberative MSIs, neither of which can appropriately satisfy these dual functions.
Option 1: MSI designers might develop regulatory regimes within which corporations participate in deliberative decision-making processes alongside other civil society groups and representatives from affected populations, in which case they should be considered democratically illegitimate organizations. Assuming they are well-institutionalized, however, participation within such organizations may convey moral legitimacy to corporations by virtue of their level of inclusivity, the quality of deliberation involved in developing their standards, and their effectiveness in achieving publicly beneficial outcomes (Hahn & Weidtmann, 2016;Mena & Palazzo, 2012).
Option 2: MSI designers might develop regulatory regimes that exclude corporations from formalized processes of deliberative decision-making and retain some hope of satisfying democratic legitimacy criteria. In this case, however, it becomes far less likely that such MSIs will function in an effective way, meaning they are far less likely to have a positive impact on the corporation's moral legitimacy. The most important reason for their probable failure is that, so long as corporations remain capable of joining or initiating competing processes of establishing regulatory standards, there is little reason to think they will choose to participate in those within which they lack any definitive say. The generation of "downward pressure" on legitimacy criteria is already a problem for established MSIs that do include corporations within their decision-making framework.
For example, Moog et al. describe the impact of competing standards enforcing less restrictive criteria on the success of the Forest Stewardship Council (FSC), noting that industry groups "perceived…the new global benchmark for sustainable forestry as a potential threat to their independence and standard operating procedures" (Moog et al., 2015). Within a decade of its development, over 50 different alternative forestry certification schemes had been developed, serving to "confuse the landscape for consumers wanting to make environmentally responsible purchases" and creating an incentive for the FSC to drive down its own standards (Gullison, 2003;Moog et al., 2015).
Given their own standards, it appears that political CSR scholars should find neither of these outcomes particularly desirable. However, the apparent undesirability of Option 1 is primarily a function of the fact that these scholars assume a false moral equivalence between the use of coercive and non-coercive modes of generating regulatory compliance. In the next section I argue that, since MSIs and similar global institutions cannot exert their authority through coercive force in the manner of nation-state governments, it is not strictly necessary to secure democratically legitimate decision-making processes within them.

Democratic Legitimation and Coercion
Earlier, I noted the potential significance of the distinction between hard and soft law, intimating that "soft" and "hard" types of regulatory power may not require equivalent forms of democratic legitimation in order to secure their normative validity and practical value. Given the unsatisfactory options left available in the preceding analysis, it is worth investigating this idea further. If soft law regulations are less in need of democratic legitimation than political CSR scholars have previously thought, it will render Option 1 more palatable. The success or failure of soft law regulations depends heavily on public awareness of their existence and consumer demand for businesses to comply with their requirements. In this respect, their effectiveness as governance instruments depends on a very different set of conditions than those in the nation-state, which possesses the capacity to issue compulsory "hard law". The differences originate in how these mechanisms are used to manipulate corporate incentives.

Incentive Transformation in Hard vs. Soft Law Regulation
MSIs regulate harmful corporate conduct by issuing soft law regulations that make everyone better off by transforming corporate incentives. It is important to note that we need not assume it is bad intentions on the part of corporate owners or managers that leads to the need for such regulations. In fact, competitive pressures often ensure that, whatever values they may hold, these actors face powerful incentives to pursue harmful strategies that make everyone worse off. These incentives make collective action difficult or impossible and, absent a mechanism to incentivize compliance, the participants will often find themselves unable to self-regulate for mutual benefit. Viewed as a game theory problem, the scenario is a classic prisoner's dilemma. (see Appendix 1 for a detailed description of the prisoner's dilemma scenario, which is represented on the left-hand side of Fig. 1). Though softand hard law approaches both aim to transform the incentives leading to such problems, they differ in how they do so. The differences illustrate why the administration of hard law by states demands strict normative standards of democratic authorization while the soft law approaches of MSIs generally do not.
The most important feature of a prisoner's dilemma is that both players have a "strictly dominant" strategy: no matter what one player does, it is always in the other player's interest to defect. As a result, the outcome is always the same: both players will fail to cooperate or "defect", leading to a worse situation for both than if they had each cooperated. Given the assumptions of the game, trust is not enough to solve the problem. The more convinced one player is that she has secured the other's trust, the greater her incentive to defect. A strategic problem in forest management (drawing on examples from the FSC) helps to illustrate the different ways such problems are typically solved through the use of hard and soft law regulation. The FSC Principles and Criteria for Forest Stewardship is comprised of ten principles with which all certified members must abide. Principle 10.1 states the following: After harvest or in accordance with the management plan, the Organization shall, by natural or artificial regeneration methods, regenerate vegetation cover in a timely fashion to pre-harvesting or more natural conditions (FSC, 2015).
It is in the public interest for logging companies to follow this standard. Forests are an important public asset and their regrowth following harvesting is an important part of forest management. In fact, over the long-term, it is in the logging industry's interest to ensure this practice is sustained in order to preserve the availability of wood as a resource for future generations. 6 Now imagine the scenario in the absence of any regulation. In the modern day, logging companies will be aware that it is in everyone's long-term interest (including their own) for them to engage in a practice of regenerating vegetation cover after harvesting trees in order to sustain forest growth. However, doing so increases costs, and those who choose not to do so can reap pricing advantages or increased per unit profitability relative to those who do. In the absence of enforcement, this means that forest regeneration is unlikely to be adopted as a widespread practice. This is because every individual company has an incentive to "free-ride" on the participation of others in the practice, despite the fact that all lumber companies would be better off over the long-run if everyone was in compliance. Moreover, when considering whether to replant, firms will be aware that their competitors are incentivized to defect, generating fear of becoming non-competitive that may motivate defensive non-compliance. It is a classic prisoner's dilemma scenario.
For the sake of illustration, it is helpful to describe the stakes in more formal terms. Assume there are two forestry companies, labeled Player 1 and Player 2. If Player 2 chooses to "cooperate" by regenerating forest lands, then Player 1's best strategy is to defect. By doing so, the company gains a competitive advantage in the market, either via pricing or profitability advantages. Likewise, if Player 2 chooses to defect, then Player 1's best strategy is also to defect in order to remain competitive from a price and profitability standpoint. The same is true for all lumber companies participating in the market. As a result, the model predicts that nobody will cooperate and everybody will be left worse off than if they had all done so. Even if some companies defy these perverse incentives or find ways to overcome them, the parameters of the situation ensure widespread non-compliance in the absence of regulation.
There are two possible approaches to regulating the situation that will encourage widespread compliance. The first is illustrated on the right-hand side of Fig. 1 and involves using state coercive power ("hard law") to increase the costs of failing to comply. In our hypothetical example, imagine that the state will impose large fines, backed by the threat of more serious criminal sanction, on any company found not to participate in forest regeneration practices. Assume that the state has complete ability to police and enforce these penalties. This change will affect the parameters of the game by increasing the costs of failing to cooperate. As a result, if Player 2 chooses to cooperate by regenerating forest lands, Player 1's best strategy is now to cooperate as well. If the company fails to do so, the large penalties it receives from the state will make it impossible to price their goods competitively and maintain profits. The same is true if Player 2 chooses to defect, in which case Player 1 can seize the advantage by maintaining compliance. By introducing a penalty regime, the state has ensured that all parties have an incentive to comply.
The alternative is to focus on increasing the benefits of cooperation rather than increasing costs of defection. The impacts on the game's parameters will be the same (also illustrated on the right-hand side of Fig. 1), but the mechanism will not involve the threat of physical force. This is how "soft law" approaches such as the certification scheme deployed by the FSC (and other standard-setting MSIs) generally work to generate behavioral change. The aim of such approaches is to publicize good practice, thereby capitalizing on consumer preferences for ethical behavior and hopefully disincentivizing bad forest management practices. If a standard is sufficiently well-publicized and the consumer demand for ethical behavior is sufficiently high, then firms will pay a market penalty for failing to cooperate-they will lose market share or be forced to cut prices to remain competitive, leading to diminished profits. 7 Strategically, the situation plays out in the following way. By publicizing standards and generating a consumer preference for ethically sourced lumber, regulation increases the payoff for receiving certification, resulting in increased sales 6 Some readers may wonder why the logging companies in this scenario cannot be expected to coordinate in order to realize their shared long-term interests over time. There are two main impediments to this occurring. First, given contemporary responses to global warming, we have seen that firms tend to severely discount the future. Second, whereas firms might be expected to coordinate around a cooperative solution due to "repeated play", the concept does not really apply in this case, where they are not pitted in direct encounters. 7 Consider the approach taken by Home Depot, which has devoted itself to purchasing sustainably harvested wood and claims to "sell more Forest Stewardship Council (FSC)-certified wood than any other North American retailer" (Home Depot Eco-Options: Sustainable Forestry, 2020). Presumably, this position is partly grounded on principle. However, the widespread promotion of this fact in Home Depot's public documentation is a clear signal to consumers who relative to uncertified competitors and (potentially, if public awareness is sufficiently high) the ability to charge a price premium for otherwise equivalent goods. This is the strategy deployed by the most successful certification schemes (e.g., Fair Trade International in the coffee industry), and the FSC seeks to generate similar effects on consumer demand in the forestry sector. If consumers respond and begin using such signals to direct their consumption decisions, it places uncertified competitors at a competitive disadvantage. Under such conditions, if Player 2 chooses to cooperate and abide by the standards, then Player 1's best strategy is to cooperate and do the same, since the company will be at a competitive disadvantage if it does not. If Player 2 chooses to defect and forgo certification, then Player 1's best strategy is also to cooperate, since the company will gain a competitive advantage by doing so. Ideally, the end result is that certification schemes like that deployed by the FSC will generate identical results to those produced by the state through enforcement of law, but without the use of coercion backed by the threat of physical force.
While nation-states have a choice of tools by which to generate compliance-they might use positive incentives such as public certification or use the threat of force-MSIs have only the former at their disposal. One of the core arguments offered by political CSR scholars to defend the importance of establishing democratically legitimate MSIs has been that they are engaging in practices of governance analogous to those undertaken by the modern state. But given that MSIs lack the right or capacity to use physical force as a means of generating compliance-a fundamentally important attribute of all legitimate states-the comparison suffers from a clear deficiency, calling into question whether the democratic legitimation requirements these scholars recommend are truly necessary.
It is undoubtedly a good thing for MSIs to adopt inclusive procedures that empower a diverse range of stakeholders to have a say in the final setting of appropriate regulations. Unlike democratic states, however, the purpose of MSIs is not to establish and enforce legitimate law applying to all citizens backed by the threat of force. Rather, their primary useful function as organizations in civil society is to construct standards and generate opportunities for companies to reap the reputational benefits of ethical compliance, thereby generating a positive incentive for good behavior. While these are worthy aims, they do not give rise to the same democratic legitimation requirements as the functions of the modern democratic state. In fact, among democratic theorists in general and deliberative democrats in particular, the fundamental justification for democracy has typically involved explaining how the coercive application of law can be rendered compatible with the liberty of individual citizens. 8 In an early statement of this view that inspired much subsequent work in deliberative democratic theory, Rousseau describes the challenge in terms of pursuing "a form of organization that will defend and protect the person and goods of each associate with the full common force, and by means of which, each uniting with all, nevertheless obey only himself and remain as free as before" (Rousseau, 1997, p. 1.6.4). In a similar vein, Joshua Cohen argues that "The fundamental idea of democratic legitimacy is that the collective authorization to exercise state power must arise from the collective decisions of the members of a society who are governed by that power" (Cohen, 2009, p. 154). And in Between Facts and Norms, Habermas himself makes the point that "the positivity of law," that is, its actionability through coercive force, is bound up with the promise that democratic processes of lawmaking justify the presumption that enacted norms are rationally acceptable. Rather than displaying the facticity of an arbitrary, absolutely contingent choice, the positivity of law expresses the legitimate will that stems from a presumptively rational self-legislation of politically autonomous citizens (Habermas, 1996, p. 33).
This backdrop of ideas further reinforces the intuition that there are clear and fundamental differences between the legitimation requirements applying to hard and soft law regulation.
In tandem with the conclusions drawn earlier, these findings should lead us to prefer forgoing the pursuit of democratic legitimacy and continuing to allow corporate participation in MSI decision-making processes (i.e., pursuing Option 1). We have reason to doubt that soft law regulation requires full democratic legitimation in the manner that many political CSR scholars have proposed. Furthermore, excluding corporations from procedures of MSI governance in order to pursue democratic legitimacy as Hussain and Moriarty recommend (i.e., pursuing Option 2) is not only unnecessary, but likely to come at the cost of their practical effectiveness. Given the probability that diminished effectiveness will undermine the potential contribution MSIs might make to bolstering corporate moral legitimacy by Footnote 7 (continued) would prefer to purchase ethically sourced wood, and their own purchasing decisions are partly shaped by a desire to cater to this audience. This, in turn, generates the kinds of pressures on lumber providers that I have discussed in this section. coordinating the provision of socially beneficial regulations and goods, there is little practical reason to take this route.

A Pragmatic Way Forward
This leaves us to focus on what might have been our starting point: given that MSIs are a potentially powerful medium for organizing public certifications that incentivize good corporate conduct, what can we do to make them work better? This may serve as a point of transition, not only within the paper but within the larger debate around MSIs and political CSR. The argument until now has been offered as a contribution to that debate, but its result invites us to step past it. The need for a renewed and intense focus on MSIs and their functioning is increasing: faith in their potential as a means of combating global corporate misconduct is waning, and it is doing so despite a relative dearth of comparative, theorydriven research into how the design of MSIs and the systems of which they are a part may affect their functioning (on this point see De Bakker et al., 2019).
For example, MSI Integrity recently released a more than 200-page long report entitled "Not Fit for Purpose", excoriating MSIs for their ineffectiveness in promoting human rights and for their inability "to hold corporations accountable for abuse, or to provide rights holders with access to remedy for abuses" (MSI Integrity, 2020, p. 4). They pledge to abandon MSIs as a primary focus and turn towards the goals of advocating for worker and community ownership of the modern corporation. The report itself involves in-depth analysis of how MSIs currently tend to function and of how corporations have frequently used them to gain a measure of popular credibility on ethical issues without driving significant changes in corporate behavior. However, theory-driven research into the conditions under which such institutions tend to be more or less successful in promoting pro-social outcomes has received scant scholarly attention (De Bakker et al., 2019). Empirical work on the outcome of MSI interventions has often instead focused on the general success or failure of MSIs to produce meaningful changes in corporate conduct within different areas of industry (Gulbrandsen, 2009;McCarthy, 2012;Pattberg, 2006;Ponte, 2012), with particular focus on the prevalence of "policy-practice decoupling", in which companies adopt formal restrictions on their own conduct but fail to implement appropriate practices to see them realized (Barrientos & Smith, 2007;Hatanaka, 2010;Schwartz & Tilling, 2009;Selfa et al., 2014). Far less work focuses on what separates comparatively more or less successful interventions within and across industries, and this must change if such research is to generate useful policy recommendations.
The way forward must have a more practical emphasis that will allow us to isolate conditions separating success and failure, and elements of pragmatist philosophy provide us a useful theoretical frame to guide us in doing so. More than twenty years ago, Wicks and Freeman (1998) proposed a turn to pragmatism as a means of finding a natural approach to integrating ethics into the study of business and of resolving disputes among positivists and anti-positivists of the day. More recent work in organization science has drawn similar attention to pragmatism's appealing aspects (Farjoun et al., 2015). Three features of such work might serve a particularly valuable purpose as guiding lights in the next phases of MSI research.

Criterion of Usefulness
Pragmatists suppose that the separation of epistemological (what we know about how things are) and normative (what we believe about how things should be) dimensions within positivist social science is counterproductive for purposes of solving specific social problems and promoting practical interventions that will improve human lives. They propose that this distinction can be overcome with a more direct focus on evaluating the pursuit of research knowledge about organizations in terms of its usefulness (Wicks & Freeman, 1998). Research satisfying this standard will be framed and developed from the ground up in terms of the specific value its findings can provide to practitioners dedicated to making MSIs work better as tools of corporate regulation. Furthermore, it will often be comparative in emphasis, focusing on whether interventions lead to improvements in resolving specific challenges, improving the quality of outcomes relative to alternative arrangements, and establishing theoretical reasons to explain such practical differences.

Wariness of Category Disputes
Pragmatists are wary of reified concepts and categories, viewing them as tools that emerge out of experience to solve specific problems but become sedimented and then shape and define our interpretation of the world, drawing attention from more practically relevant concerns in the present. They view "category disputes"-such as those over the categories "legitimate" and "illegitimate"-as posing the risk of distraction from a more appropriate focus on the processes and problems that make up the substance of our experience (Dewey, 1930;Farjoun et al., 2015;Wicks & Freeman, 1998). They advocate continual, recursive attention to the context within which concepts are used and focus on the value of their usage. Setting aside disputes over terminology and subtle theoretical distinctions in favor of a renewed focus on designing research into the tools that may promote superior MSI functioning and useful results is essential.

Commitment to Experimentalism
A spirit of practical experimentation is at the core of pragmatist philosophy. Wicks and Freeman argue that if we are to "find meaningful ways of incorporating ethics…into mainstream management research, then something like pragmatic experimentation is required" (Wicks & Freeman, 1998, p. 138). To this end, researchers must view the current generation of MSIs as practical experiments in a new form of regulatory governance and avoid spending too much time trying to understand how they fit into existing categories. Isolating the factors that differentiate such efforts from one another and identifying the contributions those differences make to success or failure is a crucial part of this effort. Successful experimentation will involve promoting the proliferation of new, innovative, and theoretically well-informed efforts to "get the formula right" while treating the introduction of new tools and organizational variables as potential "levers" in the toolkits of MSI designers. Finally, an increased focus on more local-level systems of corporate self-regulation and community engagement can provide potentially valuable lessons for larger scale interventions at a global level while multiplying the opportunities available to researchers to observe and help design systems and processes for developing policy and standards.

The Benefits of Deliberation Within MSIs
Taking a more practical approach informed by these ideas will allow business scholars to focus with greater intensity on questions about how the design of MSIs can and does influence their practical effectiveness as regulatory instruments. In the closing parts of the paper, I will shift away from a focus on normative theories of deliberative democracy and towards empirical and policy literatures in political science and social psychology on deliberation and its effects with the aim of contributing to this effort. Relatively recent work in these traditions suggests a range of ways that inclusive deliberation can both improve the quality of decisions made within MSIs and help to generate broader changes in the social beliefs, attitudes, and competences of participants. In my discussion I will focus on the specific use to which such research might be put, blending work from different theoretical traditions and drawing on experimental findings as a resource. Taken together, these benefits can facilitate better, more representative practical outcomes and help bolster the moral legitimacy of MSIs and their corporate participants.
These and similar findings are important resources for political CSR scholars to exploit as they seek to better understand the practical value of deliberation within MSIs and, in what follows, I will suggest some ways they might be used to do so through a targeted analysis of several key findings. While far from an exhaustive catalogue of relevant work, the discussion provides some indication of the potential contribution deliberation within MSIs can make to the quality of the decisions made within them, the moral legitimacy ascribed to corporate participants, and the broader contribution such exercises may make to cultivating pro-social beliefs, attitudes, and competences among those involved. This turn away from focusing on the democratic legitimacy of such institutions and towards the role that deliberation can play in enhancing their usefulness takes the debate about the practical importance of deliberative MSIs in a more pragmatic direction without abandoning the overall commitment to deliberative public engagement.

Deliberation and the Quality of Group Decision-Making
A growing body of research focuses on how deliberation affects the quality of collective decision-making. Two types of research are particularly relevant to our concerns here. One focuses on how small-scale exercises in deliberation affect the decision-making capacities of individual participants while a second concerns the benefits of deliberation for the problem-solving and predictive capacity of groups making collective decisions.

Individual Capacities
Researchers have found that small-scale exercises in public deliberation can help participants to gain broader knowledge about politics, society, and their fellow deliberators, and continue to explore the impact of that knowledge on the rationality of their preferences. In a particularly notable surveybased study of participants in the British Columbia Citizens' Assembly on Electoral Reform-a practical experiment in collective self-governance in its own right- Blais et al. (2008) demonstrate the capacity of well-designed deliberative "mini-publics" to help fulfill these aims. In 2004, representatives were chosen at random from each electoral riding in the Canadian province of British Columbia to assemble and decide on whether to recommend a new electoral system to the people in a provincial referendum. The 160 representatives were ordinary citizens who traveled to Vancouver for a series of weekend meetings over the course of several months, first to engage in a series of educational sessions concerning the effects of electoral systems on politics, then to deliberate in small and large-group sessions about their concerns and preferences. Blais et al. used survey data issued over the course of the months-long process to track the shifting priorities and preferences of participants. They found that ordinary citizens were very capable of learning about the complexities of the different systems. Furthermore, as they were exposed to the concerns of fellow members about local representation and other issues during the deliberative phase, participants' priorities shifted, and so too did their electoral system preferences (Blais et al., 2008). These results are consistent with findings from studies of experimental Deliberative Polls and subsequent citizens' assemblies in Ontario and the Netherlands suggesting that participants are capable of developing thoughtful and well-founded judgments based on information gained through deliberation with others (cf. Crosby, 1995Crosby, , 2005Dienel & Renn, 1995;Fishkin, 1997Fishkin, , 2009Fung & Wright, 2001). They indicate significant learning and preference transformation among participants due specifically to the process of public deliberation in which they participated with others from around the province.
These findings demonstrate that, under the right conditions, MSI participants might genuinely learn about one another and the issues at stake. This accumulation of knowledge will usefully improve the quality of their collective decisions and improve the ability of participants to function as critics of and evangelists for voluntary standards in society at large.
A more technical body of research considers how deliberation affects the coherence of participant preferences and how this, in turn, affects the rationality of decision-making processes among deliberators. For example, List et al. demonstrate that deliberation can rationalize the ordering of cyclical preferences, suggesting these problems can often be overcome through deliberation prior to voting. 9 Using data from Deliberative Polls, they find that structured deliberation tends to induce "single-peakedness" in preferences across the population of participants. Single-peakedness is a condition of "meta-agreement" along the underlying issue dimension about which there is dispute. It requires "the existence of a left-right ordering of the alternatives such that each individual has (1) a most preferred alternative and (2) a decreasing preference for other alternatives as they get more distant in either direction from it" (List et al., 2013). The result is a more rational ordering of collective preferences, an avoidance of majority cycling, and restoration of conditions under which further deliberation may be conducted in rational terms and majoritarian decision should be considered meaningful. These benefits will not only improve the quality of rational decision-making within MSIs but help to cultivate better decision-making practices among participants across the rest of their political engagements.
In a related vein, Dryzek and Niemeyer (2006) argue for the importance of achieving "meta-consensus" on different dimensions of debate, making the case that establishing meta-consensus is an important step along the way to potentially achieving consensus through deliberation. While their study is not empirical in nature, it provides a more general account of the potentially significant effects of deliberation on the rational structuration of preferences. Whereas consensus produced through deliberation involves agreement on the "right answer", meta-consensus involves agreement on the shared space over which debate about that answer might occur.
Dryzek identifies three types of consensus and corresponding forms of metaconsensus: Normative consensus involves agreement on the values that should predominate while its meta-counterpart involves recognition of the legitimacy of disputed values.
Epistemic consensus involves agreement about a particular belief concerning the impact of a policy while its metacounterpart involves agreement "on the credibility of disputed beliefs, and on their relevance to the norms that define the issue at hand" (Dryzek & Niemeyer, 2006, p. 640).
Finally, preference consensus involves agreement on an expressed preference for a policy, while its meta-counterpart involves agreement on both the range of alternative preferences that are considered acceptable (if not ideal) and "the validity of different ways that choices across alternatives can be structured" (Dryzek & Niemeyer, 2006, p. 641).
Achieving meta-consensus within MSIs is a crucial first step towards hopefully achieving consensus on mutually satisfactory outcomes. When consensus is impossible, however, it will serve an important use by helping facilitate better resolution of continuing conflicts by ensuring that participants hold a shared understanding of the terms of discourse. Improving the quality of decisions implemented by MSIs by improving individual capacities and establishing common ground for mutual understanding among stakeholders stands to have a positive impact on the moral legitimacy of MSIs and their corporate participants.

Group Capacities
An equally relevant body of work concerns the useful benefits of deliberation for the problem-solving and predictive 9 List et al (2013) focus more narrowly on how participation in deliberation impacts the rationality of decision-making processes within deliberative bodies by helping to prevent "majority-cycling". Majority-cycling is a problem resulting from differences in the way that individuals order their preferences. In their words: "If a third of a (electorate, committee, etc.) prefers x to y, another third prefers y to z, and the remaining third prefers z to x, then majorities prefer x to y, y to z, and yet z to x" (List et al., 2013, p. 80). Beyond the fact that cyclical majority preferences exhibit collectively irrational properties, they can also lead to situations in which pairwise majority voting fails to produce a stable winning outcome. Perhaps most importantly, when preferences are ordered in this way and pairwise votes are taken sequentially, the outcome will depend on the order in which they are taken, meaning they are subject to agenda manipulation. As a result, some have argued that preferences ordered in this way can "undermine the meaningfulness of democracy" (List et al., 2013;Riker, 1982). 1 3 capacity of groups making collective decisions. Early studies on this issue were primarily theoretical and premised on the Condorcet jury theorem. According to the jury theorem, provided the average competence of participants in a joint decision-making process is sufficiently high (and certain other conditions are met), a voting majority is more likely to be right than the average individual. As the number of individuals involved in voting for or against a proposal increases, the theoretical expectation is that moderately competent groups of sufficient size should do better than experts. 10 The origins of this idea as a political principle go back to Rousseau, but have been formalized in a range of additional places, most notably in the work of Cohen (1986), Coleman and Ferejohn (1986), Grofman and Feld (1988), and Goodin (2005; see also Goodin and Spiekermann, 2018).
A similar body of work builds upon these insights while adding a focus on the benefits of cognitive diversity as an epistemic asset within deliberative decision-making. Drawing on the work of Page (2008) and Hong and Page (2001), Landemore (2012) identifies the cognitive diversity of inclusive deliberative bodies as a virtue through which their decision-making competence is often considerably enhanced. Specifically, she shows that, in a problem-solving context, "cognitive diversity actually matters more to the production of smart collective solutions than individual ability does" (Landemore, 2012). Using the tools of formal theory, she argues that deliberation oriented towards consensus within cognitively diverse groups can be particularly well-suited to solving specific problems, where problem-solving "involves identifying or formulating the best solution from a set of possibilities" (Landemore, 2012). She then goes on to demonstrate that deliberation followed by majority rule serves better as a prediction device, oriented towards trying to derive "some estimate…or belief about a future state of the world" (Landemore, 2012;Landemore & Page, 2015).
Taken together, there is strong evidence that designing MSIs around an ideal of inclusive deliberation empowers the decision-making capacities of participants both individually, and collectively, when taken as a whole. Further experimentation in the institutional design of such organizations will help researchers to better understand these dynamics and their contribution to long-term changes in corporate behavior and social outcomes. As the effects of these improved capacities are experienced by participants and the broader community of affected individuals and groups, we should expect a parallel increase in the perceived moral legitimacy of MSIs and their corporate members. This perception will only be enhanced to the extent that their experiences within such institutions empower individuals to reason and deliberate with one another as democratic citizens in the broader sphere of civil society.

Effects on Social Beliefs and Attitudes
The idea that civil society organizations can and should function as "schools of democracy" within which citizens acquire the beliefs and attitudes required to sustain the integrity of democratic political life has a rich scholarly history. Frequently, this work includes reference to the importance of deliberation within the framework of associational decisionmaking (Cohen & Rogers, 1993;Fung & Wright, 2001). Empirical work in this area suggests that deliberative MSIs are ideal sites for generating the same cultural effects on participants. When properly designed, participants tend to attach increasing value to the pursuit of collective aims through such institutions and develop the tolerant beliefs and attitudes necessary to sustain their smooth operation.

Perception of Value
One important mechanism by which this may occur is via the "self-reinforcing properties" of deliberation. Individuals who participate in successful deliberative enterprises report experiencing fulfillment from having participated and show increased support for deliberation as a method for solving collective problems (Burkhalter et al., 2002;Levine et al., 2005). Moreover, practice at deliberating is crucial to maintaining the capacities and skills required to do so. Fung and Wright report that "Individuals' capacities to deliberate and make public decisions atrophy when left unused," meaning that practice in diverse contexts is essential. Furthermore, they note, participants "have incentives to develop the capacities and master the information necessary to make good decisions because they must live with the consequences of poor ones" when they participate in collective decisionmaking processes with binding consequences like those promoted within deliberative MSIs (Fung & Wright, 2001, p. 28). This suggests that giving people the responsibility for directly making such binding decisions will tend to motivate greater dedication to doing so and foster increasing levels of commitment among participants over time.
Furthermore, a crucial "background enabling condition" through which to spur active participation is secured by ensuring a "rough equality of power, for the purposes of deliberative decision, [among] participants" (Fung & Wright, 2001, p. 25). If the goal of generating widespread participation is to be realized within MSIs, it is important to ensure that the various types of contributors possess a sense of ownership over the final result. This can only be achieved if they are treated as equally valued participants within the deliberative decision-making process. If the participation of some groups is treated as merely advisory or as having significantly less intrinsic weight than that of others, the MSI will lack the buy-in from less valued participants that is necessary to bolster broader social enthusiasm for such participation (Fung & Wright, 2001, p. 28).
More generally, it is not hard to imagine how the proliferation of MSIs may promote broader democratic participation in society at large by actively drawing in new and previously unheard from voices to discuss the social consequences of corporate behavior and work on finding solutions. At the same time, provided they are heard and treated with respect, the participation of these groups in the collective decisionmaking process will pay dividends in terms of the public moral legitimacy attributed to both the MSIs themselves and to participating corporations. In particular, Fung and Wright argue that the devolution of authority structures to "local action units" (like MSIs) is a natural incentive to broader participation (Fung & Wright, 2001, p. 21). They find that groups included within the regulatory process are more likely to see political participation as being worthwhile, leading to broader citizen engagement within civil society.

Cultivation of Tolerance
Finally, there is strong evidence that deliberation and broader "cross-cutting engagement" like that promoted within deliberative MSIs leads to increases in tolerance for social outgroups. The cultivation of such beliefs and attitudes is clearly not only important for the ongoing functioning of MSIs themselves, but will contribute to the broader promotion of democratic values of tolerance and mutual respect in society. MSIs are particularly well-suited kinds of institutions for promoting such effects due to the "weak tie" relationships that bind participants within them together. This finding draws on the resources of experimental work in social psychology on intergroup contact theory. The latter finds that the forging of weak tie relationships among individuals from different social groups who engage in pursuit of a common goal can lead to increased levels of tolerance, not only among the specific individuals involved but towards the social or demographic groups to which they belong (Allport, 1979;Pettigrew, 1998).
Mutz builds on this insight in a wide-ranging study of political tolerance, in which she demonstrates that "Hearing the other side" leads participants in weak tie relationships to.
…see that there is more than one side to an issue, that a political conflict is, in fact, a legitimate controversy with rationales on both sides [which] translates to greater willingness to extend civil liberties to even those groups whose political views one dislikes a great deal (Mutz, 2006). This implies that deliberation and exposure to rational argument from "the other side" facilitates meta-consensus of the kind Dryzek and Niemeyer describe (Dryzek & Niemeyer, 2006). However, an even more dramatic effect has been found to result from the development of affective ties between participants in group tasks. As affective attachment to those with whom one disagrees become stronger, so too does the extent of tolerance one feels for the groups to which they belong (Mutz, 2006).
The impact of this effect on the functionality of MSIs over time and the role such associations might play in bolstering the communicative resources of individuals in civil society are not to be underestimated. Participation in such enterprises holds the potential to ease the internal processes of MSIs, enhance the broad ability of MSI participants to process and solve problems through deliberative means in society, and to promote mutual respect among MSI participants. This will usefully contribute to enhancing corporate moral legitimacy over time to the extent that the collective decisions arising out of deliberations represent genuine accommodations of stakeholder needs and preferences, an outcome that well-designed deliberative engagements are likely to promote.

Conclusion
A body of recent political CSR scholarship has focused on the potential of MSIs to realize the dual goals of satisfying democratic legitimacy requirements while securing the moral legitimacy of corporate participants. However, if these are their goals, then MSI designers appear to face a paradox. It is essential that corporations be invited to participate in the process of making decisions within MSIs if they are to serve as useful regulatory instruments and therefore to facilitate corporate moral legitimacy. Yet, any institution within which corporations play such a role will lack democratic legitimacy. Therefore, if they are to be effective, MSIs cannot achieve the democratic legitimacy that is possible within formally representative institutions of governance. I have argued that this is not a serious problem since the mechanisms through which MSIs regulate corporate behavior do not involve coercion backed by the threat of force, meaning it is not essential that they should meet this standard. Though they will not satisfy democratic legitimacy standards on their own, MSIs hold great potential to fill unmet needs for regulation. Therefore, we should redirect our energies away from improving the democratic legitimacy of their implementation and towards understanding how we can improve their effectiveness as regulatory institutions.
To this end, I argue that a broader shift towards a more pragmatic orientation is called for in the academic discussion about MSIs and their functioning. In the absence of focused research into the different factors and conditions that promote or hinder their likelihood of producing socially beneficial outcomes, we risk seeing MSIs dismissed as a passing trend or fad and missing out on opportunities to learn from them and take steps to use them as a means of improving human lives. To avoid that eventuality, I have proposed adopting three elements of pragmatist philosophy-a criterion of usefulness, wariness of category disputes, and an embrace of experimentalism-as touchpoints to guide future MSI research.
With these ideas as backdrop, I have aimed to demonstrate the practical ways that strengthening the role of deliberation within MSIs can bolster their effectiveness as institutions and reinforce corporate moral legitimacy while contributing to the broader adoption of deliberative attitudes among participants. At present, however, these are only potential effects. Few MSIs are currently institutionalized in ways that will lead to these effects, meaning that most will require significant reforms to realize their potential. The MSI Database reports that "the overwhelming majority of MSIs are failing to meaningfully engage communities affected by the operations of participating companies" (MSI Integrity, 2017b, p. 3). In fact, as of 2016, only 14 percent of the MSIs surveyed allowed representatives from affected communities to participate in decision-making and only 49 percent involved them in any way at all (MSI Integrity, 2017b). In the absence of such input, the standards on which these certifications are based are less likely to promote socially beneficial outcomes due to the low information and biased input on which their development has been based. In turn, this lack of inclusion diminishes the moral legitimacy of these institutions and their corporate members.
On the plus side, the 45 MSIs surveyed in the dataset operate in over 170 countries on six continents, engage with over 50 national governments, and regulate over 9,000 companies with combined annual revenues of more than 5.4 trillion dollars (MSI Integrity, 2017b). However, fully 20 percent of these lack power to sanction members who violate standards by suspending or expelling them from the initiative and only 40 percent have a formal complaint process that would allow members of affected communities to raise grievances when standards are violated (MSI Integrity, 2017b). A major cause of these deficits is that corporations and civil society organizations currently lack incentives to ensure that MSIs include a diverse array of stakeholders. This is in large part because consumers lack the knowledge necessary to differentiate effective and well-designed regulatory approaches and certification schemes from those that are lacking. This is less a product of blameworthy ignorance on their part than of the difficulty involved in verifying the institutional quality and practical effectiveness of MSIs. Nonetheless, MSIs will not be driven to improve their internal procedures until consumers demand that they do so. There are a range of practical ways to approach solving this problem and further experimentation in this domain is sorely needed.
One possible route deserving of further experimentation is the use of government packaging or labeling requirements, which would require MSI certifications to receive a stamp of approval from government regulators before it could be used in advertising or on product packaging. Though there will always be a risk of corruption involved in conscripting governments to such purposes, they have often proven better than the private sector at regulating such standards in ways that promote the public good. Government might also encourage development of certification schemes in new areas of industry where their implementation currently lags: at present, 90 percent of industry-specific MSIs are clustered in the agriculture, forestry and fishing, mining and energy and consumer goods sectors.
Alternatively, private organizations devoted to "certifying the certifiers" might take up the slack. These organizations might follow the lead of givewell.org in the philanthropic sector, which tracks the efficiency and spending habits of philanthropic organizations and reports on them to the public (Givewell.org). Similar organizations might do the same for MSIs, issuing seals of approval verifying the quality of their certification schemes. In fact, one organization has already taken steps to meet this challenge: ISEAL Alliance is a new venture that has grown out of the Embedding Project, a system of resources originally designed to help companies better embed sustainability within their operations. ISEAL is a membership organization that, like givewell and similar groups, investigates and validates the credibility and performance of sustainability standards and systems and issues recommendations about the value of different certifications based on rigorous social science research.
Whatever form they may take, dramatic design improvements are necessary before MSIs will fully achieve their potential to close unmet regulatory needs and facilitate corporate moral legitimacy, and experimentation with different approaches will be necessary to make these changes happen. In general, future research in the area must be guided by a more pragmatic outlook if it is to generate genuine social benefits. To this end, researchers must redouble efforts to identify useful avenues of research and experimentation focused on understanding the reasons for existing failures and how they might be corrected.