The legitimacy of global governance institutions.

Buchanan, Allen ; Keohane, Robert O.

"Legitimacy" has both a normative and a sociological meaning. To say that an institution is legitimate in the normative sense is to assert that it has the right to rule--where ruling includes promulgating rules and attempting to secure compliance with them by attaching costs to noncompliance and/or benefits to compliance. An institution is legitimate in the sociological sense when it is widely believed to have the right to rule. (1) When people disagree over whether the WTO is legitimate, their disagreements are typically normative. They are not disagreeing about whether they or others believe that this institution has the right to rule; they are disagreeing about whether it has the right to rule. (2) This essay addresses the normative dimension of recent legitimacy discussions.

We articulate a global public standard for the normative legitimacy of global governance institutions. This standard can provide the basis for principled criticism of global governance institutions and guide reform efforts in circumstances in which people disagree deeply about the demands of global justice and the role that global governance institutions should play in meeting them. We stake out a middle ground Between an increasingly discredited conception of legitimacy that conflates legitimacy with international legality understood as state consent, on the one hand, and the unrealistic view that legitimacy for these institutions requires the same democratic standards that are now applied to states, on the other.

Our approach to the problem of legitimacy integrates conceptual analysis and moral reasoning with an appreciation of the fact that global governance institutions are novel, still evolving, and characterized by reasonable disagreement about what their proper goals are and what standards of justice they should meet. Because both standards and institutions are subject to change as a result of further reflection and action, we do not claim to discover timeless necessary and sufficient conditions for legitimacy. Instead, we offer a principled proposal for how the legitimacy of these institutions ought to be assessed--for the time being. Essential to our account is the idea that to be legitimate a global governance institution must possess certain epistemic virtues that facilitate the ongoing critical revision of its goals, through interaction with agents and organizations outside the institution. A principled global public standard of legitimacy can help citizens committed to democratic principles to distinguish legitimate institutions from illegitimate ones and to achieve a reasonable congruence in their legitimacy assessments. Were such a standard widely accepted, it could bolster public support for valuable global governance institutions that either satisfy the standard or at least make credible efforts to do so.

"Global governance institutions" covers a diversity of multilateral entities, including the World Trade Organization (WTO), the International Monetary Fund (IMF), various environmental institutions, such as the climate change regime built around the Kyoto Protocol, judges' and regulators' networks, the UN Security Council, and the new International Criminal Court (ICC). (3) These institutions are like governments in that they issue rules and publicly attach significant consequences to compliance or failure to comply with them--and claim the authority to do so. Nonetheless, they do not attempt to perform anything approaching a full range of governmental functions. These institutions do not seek, as governments do, to monopolize the legitimate use of violence within a permanently specified territory, and their design and major actions require the consent of states.

Determining whether global governance institutions are legitimate--and whether they are widely perceived to be so--is an urgent matter. Global governance institutions can promote international cooperation and also help to construct regulatory frameworks that limit abuses by nonstate actors (from corporations to narcotraffickers and terrorists) who exploit transnational mobility. At the same time, however, they constrain the choices facing societies, sometimes limit the exercise of sovereignty by democratic states, and impose burdens as well as confer benefits. For example, states must belong to the WTO in order to participate effectively in the world economy, yet WTO membership requires accepting a large number of quite intrusive rules, authoritatively applied by its dispute settlement system. Furthermore, individuals can be adversely affected by global rules--for example, by the blacklists maintained by the Security Council's Sanctions Committee (4) or the WTO's policies on intellectual property in "essential medicines." If these institutions lack legitimacy, then their claims to authority are unfounded and they are not entitled to our support.

Judgments about institutional legitimacy have distinctive practical implications. Generally speaking, if an institution is legitimate, then this legitimacy should shape the character of both our responses to the claims it makes on us and the form that our criticisms of it take. We should support or at least refrain from interfering with legitimate institutions. Further, agents of legitimate institutions deserve a kind of impersonal respect, even when we voice serious criticisms of them. Judging an institution to be legitimate, if flawed, focuses critical discourse by signaling that the appropriate objective is to reform it, rather than to reject it outright.

It is important not only that global governance institutions be legitimate, but that they are perceived to be legitimate. The perception of legitimacy matters, because, in a democratic era, multilateral institutions will only thrive if they are viewed as legitimate by democratic publics. If one is unclear about the appropriate standards of legitimacy or if unrealistically demanding standards are assumed, then public support for global governance institutions may be undermined and their effectiveness in providing valuable goods may be impaired.

ASSESSING LEGITIMACY

The Social Function of Legitimacy Assessments

Global governance institutions are valuable because they create norms and information that enable member states and other actors to coordinate their behavior in mutually beneficial ways. (5) They can reduce transaction costs, create opportunities for states and other actors to demonstrate credibility, thereby overcoming commitment problems, and provide public goods, including rule-based, peaceful resolutions of conflicts. (6) An institution's ability to perform these valuable functions, however, may depend on whether those to whom it addresses its rules regard them as binding and whether others within the institution's domain of operation support or at least do not interfere with its functioning. It is not enough that the relevant actors agree that some institution is needed; they must agree that this institution is worthy of support. So, for institutions to perform their valuable coordinating functions, a higher-order coordination problem must be solved. (7)

Once an institution is in place, ongoing support for it and compliance with its rules are sometimes simply a matter of self-interest from the perspective of states, assuming that the institution actually achieves coordination or other benefits that all or at least the more powerful actors regard as valuable. (8) Similarly, once the rule of the road has been established and penalties for violating it are in place, most people will find compliance with it to be rational from a purely selfinterested point of view. In the latter case, no question of legitimacy arises, because the sole function of the institution is coordination and the choice of the particular coordination point raises no issues on which people are likely to disagree. Global governance institutions are not pure coordination devices in the way in which the rule of the road is, however. Even though all may agree that some institution or other is needed in a specific domain (the regulation of global trade, for example), and all may agree that any of several particular institutions is better than the noninstitutional alternative, different parties, depending upon their differing interests and moral perspectives, will find some feasible institutions more attractive than others. The fact that all acknowledge that it is in their interest to achieve coordinated support for some institution or other may not be sufficient to assure adequate support for any particular institution.

The concept of legitimacy allows various actors to coordinate their support for particular institutions by appealing to their common capacity to be moved by moral reasons, as distinct from purely strategic or exclusively self-interested reasons. If legitimacy judgments are to perform this coordinating function, however, actors must not insist that only institutions that are optimal from the standpoint of their own moral views are acceptable, since this would preclude coordinated support in the face of diverging normative views. More specifically, actors must not assume that an institution is worthy of support only if it is fully just. We thus need a standard of legitimacy that is both accessible from a diversity of moral standpoints and less demanding than a standard of justice. Such a standard must appeal to various actors' capacities to be moved by moral reasons, but without presupposing more moral agreement than exists.

Legitimacy and Self-Interest

It is one thing to say that an institution promotes one's interests and another to say that it is legitimate. As Andrew Hurrell points out, the rule-following that results from a sense of legitimacy is "distinguishable from purely self-interested or instrumental behaviour on the one hand, and from straightforward imposed or coercive rule on the other." (9) Sometimes self-interest may speak in favor of treating an institution's rules as binding; that is, it can be in one's interest to take the fact that an institution issues a rule as a weighty reason for complying with it, independently of a positive assessment of the content of particular rules. This would be the case if one is likely to do better, from the standpoint of one's own interest, by taking the rules as binding than one would by evaluating each particular rule as to how complying with it would affect one's interests. Yet clearly it makes sense to ask whether an institution that promotes one's interests is legitimate. So legitimacy, understood as the right to rule, is a moral notion that cannot be reduced to rational self-interest. To say that an institution is legitimate implies that it has the right to rule even if it does not act in accordance with the rational self-interest of everyone who is subject to its rule.

There are advantages in achieving coordinated support for institutions on the basis of moral reasons, rather than exclusively on the basis of purely self-interested ones. First, the appeal to moral reasons is instrumentally valuable in securing the benefits that only institutions can provide because, as a matter of psychological fact, moral reasons matter when we try to determine what practical attitudes should be taken toward particular institutional arrangements. For example, we care not only about whether an environmental regulation regime reduces air pollutants and thereby produces benefits for all, but also whether it fairly distributes the costs of the benefits it provides. Given that there is widespread disagreement as to which institutional arrangement would be optimal, we need to find a shared evaluative perspective that makes it possible for us to achieve the coordinated support required for effective institutions without requiring us to disregard our most basic moral commitments. Second, and perhaps most important, if our support for an institution is based on reasons other than self-interest or the fear of coercion, it may be more stable. What is in our self-interest may change as circumstances change and the threat of coercion may not always be credible, and moral commitments can preserve support for valuable institutions in such circumstances.

For questions of legitimacy to arise there must be considerable moral disagreement about how institutions should be designed. Yet for agreement about legitimacy to be reached, there must be sufficient agreement on the sorts of moral considerations that are relevant for evaluating alternative institutional designs. The practice of making legitimacy judgments is grounded in a complex belief--namely, that while it is true that institutions ought to meet standards more demanding than mere mutual benefit (relative to some relevant noninstitutional alternative), they can be worthy of our support even if they do not maximally serve our interests and even if they do not measure up to our highest moral standards. (10)

Legitimacy requires not only that institutional agents are justified in carrying out their roles, but also that those to whom institutional rules are addressed have content-independent reasons to comply with them, and that those within the domain of the institution's operations have content-independent reasons to support the institution or at least to not interfere with its functioning. (11) One has a content-independent reason to comply with a rule if and only if one has a reason to comply regardless of any positive assessment of the content of that rule. For example, I have a content-independent reason to comply with the rules of a club to which I belong if I have agreed to follow them and this reason is independent of whether I judge any particular rule to be a good or useful one. If I acknowledge an institution as having authority, I thereby acknowledge that there are content-independent reasons to comply with its rules or at least to not interfere with their operation. Legitimacy disputes concern not merely what institutional agents are morally permitted to do but also whether those to whom the institution addresses its rules should regard it as having authority.

The debate about the legitimacy of global governance institutions engages both the perspective of states and that of individuals. Indeed, as recent mass protests against the WTO suggest, politically mobilized individuals can adversely affect the functioning of global governance institutions, both directly, by disrupting key meetings, and indirectly, by imposing political costs on their governments for their support of institutional policies. Legitimacy in the case of global governance institutions, then, is the right to rule, understood to mean both that institutional agents are morally justified in making rules and attempting to secure compliance with them and that people subject to those rules have moral, content-independent reasons to follow them and/or to not interfere with others' compliance with them.

If it becomes widely believed that an institution does not measure up to standards of legitimacy, then the result may be a lack of coordination, at least until the institution changes to conform to the standards or a new institution that better conforms to them replaces it. Thus, it would be misleading to say simply that the function of legitimacy judgments is to achieve coordinated support for institutions; rather, their function is to make possible coordinated support based on moral reasons, while at the same time supplying a critical but realistic minimal moral standard by which to determine whether institutions are worthy of support.

Justice and Legitimacy

The foregoing account of the social function of legitimacy assessments helps clarify the relationship between justice and legitimacy. Collapsing legitimacy into justice undermines the valuable social function of legitimacy assessments. There are two reasons not to insist that only just institutions have the right to rule. First, there is sufficient disagreement on what justice requires that such a standard for legitimacy would thwart the eminently reasonable goal of securing coordinated support for valuable institutions on the basis of moral reasons. Second, even if we all agreed on what justice requires, withholding support from institutions because they fail to meet the demands of justice would be self-defeating from the standpoint of justice itself, because progress toward justice requires effective institutions. To mistake legitimacy for justice is to make the best the enemy of the good.

COMPETING STANDARDS OF LEGITIMACY

Having explicated our conception of legitimacy, we now explore standards of legitimacy: the conditions an institution must satisfy in order to have the right to rule. In this section we articulate three candidates for the appropriate standard of legitimacy--state consent, consent by democratic states, and global democracy--and argue that each is inadequate.

State Consent

The first view is relatively simple. Global governance institutions are legitimate if (and only if) they are created through state consent. In this conception, legitimacy is simply a matter of legality. Legally constituted institutions, created by states according to the recognized procedures of public international law and consistent with it, are ipso facto legitimate or at the very least enjoy a strong presumption of legitimacy. (12) Call this the International Legal Pedigree View (the Pedigree View, for short). A more sophisticated version of the Pedigree View would require the periodic reaffirmation of state consent, on the grounds that states have a legitimate interest in determining whether these institutions are performing as they are supposed to. (13)

The Pedigree View fails because it is hard to see how state consent could render global governance institutions legitimate, given that many states are non-democratic and systematically violate the human rights of their citizens and are for that reason themselves illegitimate. State consent in these cases cannot transfer legitimacy for the simple reason that there is no legitimacy to transfer. To assert that state consent, regardless of the character of the state, is sufficient for the legitimacy of global governance institutions is to regress to a conception of international order that fails to impose even the most minimal normative requirements on states. Indeed, once we abandon that deeply defective conception of international order, it is hard to see why state consent is even a necessary condition for legitimacy.

It might be argued, however, that even though the consent of illegitimate states cannot itself make global governance institutions legitimate, there is an important instrumental justification for treating state consent as a necessary condition for their legitimacy: doing so provides a check on the tendency of stronger states to exploit weak ones. In other words, persisting in the fiction that all states--irrespective of whether they respect the basic rights of their own citizens--are moral agents worthy of respect serves an important value. This conception of the state, however, is not a fiction that those who take human rights seriously can consistently accept.

The proponent of state consent might reply as follows: "My proposal is not that we should return to the pernicious fiction of the Morality of States. Instead, it is that we should agree, for good cosmopolitan reasons, to regard a global governance institution as legitimate only if it enjoys the consent of all states." Withholding legitimacy from global governance institutions, no matter how valuable they are, simply because not all states consent to them, however, would purport to protect weaker states at the expense of giving a legitimacy veto to tyrannies. The price is too high. Weak states are in a numerical majority in multilateral institutions. Generally speaking, they are less threatened by the dominance of powerful states within the institutions than they are by the actions of such powerful states acting outside of institutional constraints.

The Consent of Democratic States

The idea that state consent confers legitimacy is much more plausible when restricted to democratic states. On reflection, however, the mere fact of state consent, even when the state in question is democratic and satisfies whatever other conditions are appropriate for state legitimacy, is not sufficient for the legitimacy of global governance institutions.

From the standpoint of a particular weak democratic state, participation in global governance institutions such as the WTO is hardly voluntary, since the state would suffer serious costs by not participating. Yet "substantial" voluntariness is generally thought to be a necessary condition for consent to play a legitimating role. (14) Of course, there may be reasonable disagreements over what counts as substantial voluntariness, but the vulnerability of individual weak states is serious enough to undercut the view that the consent of democratic states is by itself sufficient for legitimacy.

There is another reason why the consent of democratic states is not sufficient for the legitimacy of global governance institutions: the problem of reconciling democratic values with unavoidable "bureaucratic discretion" that plagues democratic theory at the domestic level looms even larger in the global case. The problem is that for a modern state to function, much of what state agents do will not be subject to democratic decisions, and saying that the public has consented in some highly general way to whatever it is that state agents do is clearly inadequate. The difficulty is not in identifying chains of delegation stretching from the individual citizen to state agents, but rather that at some point the impact of the popular will on how political power is used becomes so attenuated as to be merely nominal. Given how problematic democratic authorization is in the modern state and given that global governance institutions require lengthening the chain of delegation, democratic state consent is not sufficient for legitimacy.

Still, the consent of democratic states may appear to be necessary, if not sufficient, for the legitimacy of global governance institutions. Indeed, it seems obvious that for such an institution to attempt to impose its rules on democratic states without their consent would violate the right of self-determination of the people of those states. Matters are not so simple, however. A democratic people's right of self-determination is not absolute. If the majority persecutes a minority, the fact that it does so through democratic processes does not render the state in question immune to sanctions or even to intervention. One might accommodate this fact by stipulating that a necessary condition for the legitimacy of global governance institutions is that they enjoy the consent of states that are democratic and that do a credible job of respecting the rights of all their citizens.

This does not mean that all such states must consent. A few such states may willfully seek to isolate themselves from global governance (Switzerland only joined the UN in 2002). Furthermore, democratic states may engage in wars that are unnecessary and unjust, and resist pressures from international institutions to desist. It would hardly delegitimize a global governance institution established to constrain unjust warfare that it was opposed by a democratic state that was waging an unjust war. A more reasonable position would be that there is a strong presumption that global governance institutions are illegitimate unless they enjoy the ongoing consent of democratic states. Let us say, then, that ongoing consent by rights-respecting democratic states constitutes the democratic channel of accountability. (15)

However valuable the democratic channel of accountability is, it is not sufficient. First, as already noted, the problem of bureaucratic discretion that attenuates the power of majoritarian processes at the domestic level seems even more serious in the case of global bureaucracies. Second, not all the people who are affected by global governance institutions are citizens of democratic states, so even if the ongoing consent of democratic states fosters accountability, it may not foster accountability to them. If--as is the case at present--democratic states tend to be richer and hence more powerful than nondemocratic ones, then the requirement of ongoing consent by democratic states may actually foster a type of accountability that is detrimental to the interests of the world's worst-off people. From the standpoint of any broadly cosmopolitan moral theory, this is a deep flaw of domestic democracies as ordinarily conceived: government is supposed to be responsive to the interests and preferences of the "sovereign people"--the people whose government it is--not all people or even all people whose legitimate interests will be seriously affected by the government's actions. (16) For these reasons, the consent of democratic states seems insufficient. The idea that the legitimacy of global governance institutions requires democracy on a grander scale may seem plausible.

Global Democracy

Because democracy is now widely thought to be the gold standard for legitimacy in the case of the state, it may seem obvious that global governance institutions are legitimate if and only if they are democratic. And since these institutions increasingly affect the welfare of people everywhere, surely this must mean that they ought to be democratic in the sense of giving everyone an equal say in how they operate. Call this the Global Democracy View.

The most obvious difficulty with this view is that the social and political conditions for democracy are not met at the global level and there is no reason to think that they will be in the foreseeable future. At present there is no global political structure that could provide the basis for democratic control over global governance institutions, even if one assumes that democracy requires little direct participation by individuals. Any attempt to create such a structure in the form of a global democratic federation that relies on existing states as federal units would lack legitimacy, and hence could not confer legitimacy on global governance institutions, because, as has already been noted, many states are themselves undemocratic or lack other qualities necessary for state legitimacy. Furthermore, there is at present no global public--no worldwide political community constituted by a broad consensus recognizing a common domain as the proper subject of global collective decision-making and habitually communicating with one another about public issues. Nor is there consensus on a normative framework within which to deliberate together about a global common interest. Indeed, there is not even a global consensus that some form of global government, much less a global democracy, is needed or appropriate. Finally, once it is understood that it is liberal democracy, democracy that protects individual and minority rights, that is desirable, the Global Democracy View seems even more unfeasible. Democracy worth aspiring to is more than elections; it includes a complex web of institutions, including a free press and media, an active civil society, and institutions to check abuses of power by administrative agencies and elected officials.

Global governance institutions provide benefits that cannot be provided by states and, as we have argued, securing those benefits may depend upon these institutions being regarded as legitimate. The value of global governance institutions, therefore, warrants being more critical about the assumption that they must be democratic on the domestic model and more willing to explore an alternative conception of their legitimacy. In the next section we take up this task.

A COMPLEX STANDARD OF LEGITIMACY

Desiderata for a Standard of Legitimacy

Our discussion of the social function of legitimacy assessments and our critique of the three dominant views on the standard of legitimacy for global governance institutions (state consent, democratic state consent, and global democracy) suggest that a standard of legitimacy for such institutions should have the following characteristics:

1. It must provide a reasonable public basis for coordinated support for the institutions in question, on the basis of moral reasons that are widely accessible in spite of the persistence of significant moral disagreement--in particular, about the requirements of justice.

2. It must not confuse legitimacy with justice but nonetheless must not allow that extremely unjust institutions are legitimate.

3. It must take the ongoing consent of democratic states as a presumptive necessary condition, though not a sufficient condition, for legitimacy.

4. Although the standard should not make authorization by a global democracy a necessary condition of legitimacy, it should nonetheless promote the key values that underlie demands for democracy.

5. It must properly reflect the dynamic character of global governance institutions: the fact that not only the means they employ, but even their goals, may and ought to change over time.

6. It must address the two problems we encountered earlier: the problem of bureaucratic discretion and the tendency of democratic states to disregard the legitimate interests of foreigners.

The standard of legitimacy must therefore incorporate mechanisms for accountability that are both more robust and more inclusive than that provided by the consent of democratic states.

Moral Disagreement and Uncertainty

The first desideratum of a standard of legitimacy is complex and warrants further explication and emphasis. We have noted that a central feature of the circumstances of legitimacy is the persistence of disagreement about, first, what the proper goals of the institution are (given the limitations imposed by state sovereignty properly conceived), second, what global justice requires, and third, what role if any the institution should play in the pursuit of global justice. Moral disagreement is not unique to global governance institutions, but extends also to the appropriate role of the state.

There are two circumstances in the case of global governance institutions, however, that exacerbate the problem of moral disagreement. First, in the case of the state, democratic processes, at least ideally, provide a way of accommodating these disagreements, by providing a public process that assures every citizen that she is being treated as an equal, through the electoral process, while, as we have seen, democracy is unavailable at the global level. Second, although there is a widespread perception, at least among cosmopolitans broadly speaking, that there is serious global injustice and that the effective pursuit of global justice requires a significant role for global institutions, it is not possible at present to provide a principled specification of the division of institutional labor for pursuing global justice. In part the problem is that there is no unified system of global institutions within which a fair and effective allocation of institutional responsibilities for justice can be devised. How responsibilities for justice ought to be allocated among global institutions and between states and global institutions depends chiefly on the answers to two questions: What are the proper responsibilities of states in the pursuit of global justice, taking into account the proper scope of state sovereignty (because this will determine how extensive the role of global institutions should be), and what are the capabilities of various global institutions for contributing to the pursuit of global justice? But neither of these questions can be answered satisfactorily at present, in part because global governance institutions are so new and in part because people have only recently begun to think seriously about achieving justice on a global scale. So the difficulty is not just that there is considerable moral disagreement about the proper goals of global governance institutions and about the role these institutions should play in the pursuit of global justice; there is also moral uncertainty. (17) A plausible standard of legitimacy for global governance institutions must somehow accommodate the facts of moral disagreement and uncertainty.

Three Substantive Criteria

We begin with a set of institutional attributes that have considerable intuitive appeal: minimal moral acceptability, comparative benefit, and institutional integrity.

Minimal Moral Acceptability. Global governance institutions, like institutions generally, must not persist in committing serious injustices. If they do so, they are not entitled to our support. On our view, the primary instance of a serious injustice is the violation of human rights. We also believe that the most plausible conception of human rights is what might be called the basic human interest conception. This conception, which we can only sketch in broad outlines here, builds on Joseph Raz's insight that rights generally are normative relations (in particular, duties and entitlements), which, if realized, provide important protections for interests. (18) On this view, to justify the claim that R is a right, one must identify an interest, support the claim that the interest is of sufficient moral importance to ground duties, explain why the duties are owed to the right holders, and make the case that if the normative relations in question are satisfied, significant protection for the interest will be achieved. Certain rights are properly called human rights because the duties they entail provide especially important protections for basic human interests, given the standard threats to those interests in our world.

What the standard threats are can change over time. For example, when human societies create legal systems and police and courts to enforce laws, they also create new opportunities for damaging basic human interests. For this reason, the content of particular human rights, and even which rights are included among the human rights, may also change, even though the basic interests that ground them do not. For example, all human beings, regardless of where or when they exist, have a basic interest in physical security, but in a society with a legal system backed by the coercive power of the state, adequate protection of this interest requires rights of due process and equal protection under the law.

There is disagreement among basic interest theorists of human rights as to exactly what the list of human rights includes and how the content of particular rights is to be filled out. There is agreement, however, that the list includes the rights to physical security, to liberty (understood as at least encompassing freedom from slavery, servitude, and forced occupations), and the right to subsistence. Assuming that this is so, we can at least say this much: global governance institutions (like institutions generally) are legitimate only if they do not persist in violations of the least controversial human rights. This is a rather minimal moral requirement for legitimacy. Yet in view of the normative disagreement and uncertainty that characterize our attitudes toward these institutions, it might be hard at present to justify a more extensive set of rights that all such institutions are bound to respect. It would certainly be desirable to develop a more meaningful consensus on stronger human rights standards. What this suggests is that we should require global governance institutions to respect minimal human rights, but also expect them to meet higher standards as we gain greater clarity about the scope of human rights.

For many global governance institutions, it is proper to expect that they should respect human rights, but not that they should play a major role in promoting human rights. Nonetheless, a theory of legitimacy cannot ignore the fact that in some cases the dispute over whether a global governance institution is legitimate is in large part a disagreement over whether it is worthy of support if it does not actively promote human rights. A proposal for a standard of legitimacy for global governance institutions must take into account the fact that some of these institutions play a more direct and substantial role in securing human rights than others.

When we see the injustices of our world and appreciate that ameliorating them requires institutional actions, we are quick to attribute obligations to institutions and then criticize them for failing to fulfill those obligations. It is one thing to say that it would be a good thing if a particular global governance institution took on certain functions that would promote human rights, however, and quite another to say that it has a duty to do so and that this duty is of such importance that failure to discharge it makes the institution illegitimate. There are two mistakes to be avoided here. The first is "duty dumping," that is, arbitrarily assuming that some particular institution has a duty simply because it has the resources to fulfill it and no other actor is doing so. (19) Duty dumping not only makes unsupported attributions of institutional responsibility; it also distracts attention from the difficult task of determining what a fair distribution of the burdens--among individuals and institutions--for protecting the human rights in question would be. The second error derives from the first: if one uncritically assumes that the institution has a duty to provide X and also assumes that X is a central matter of justice (as is the case with human rights), then one may conclude that the institution's failure to provide X is such a serious injustice as to rob the institution of legitimacy. But the fact that an institution could provide X and the fact that X is a human right does not imply that in refraining from providing X the institution commits a serious injustice. That conclusion would only follow if it were established that the institution has a duty of justice to provide X. Merely pointing out that the institution could provide X--or even showing that it is the only existing institution that can do so--is not sufficient to show that it has a duty of justice or any duty at all to provide X.

We seem to be in a quandary. Contemporary institutions have to operate in an environment of moral disagreement and uncertainty, which limits the demands we can reasonably place on them to respect or protect particular human rights. Furthermore, to be sufficiently general, an account of legitimacy must avoid moral requirements that only apply to some global governance institutions. These considerations suggest the appropriateness of something like the minimal moral acceptability requirement, understood as refraining from violations of the least controversial human rights. On the other hand, the standard of legitimacy should somehow reflect the fact that part of what is at issue in disputes over the legitimacy of some of these institutions is whether they should satisfy more robust demands of justice. In other words, the standard should acknowledge the fact that where the issue of legitimacy is most urgent, there is likely to be deep moral disagreement and uncertainty.

In our view, the way out of this impasse is to build the conditions needed for principled, informed deliberation about moral issues into the standard of legitimacy itself. The standard of legitimacy should require minimal moral acceptability, but should also accommodate and even encourage the possibility of developing more determinate and demanding requirements of justice for at least some of these institutions, as a principled basis for an institutional division of labor regarding justice emerges.

Comparative Benefit. This second substantive condition for legitimacy is relatively straightforward. The justification for having global governance institutions is primarily if not exclusively instrumental. The basic reason for states or other addressees of institutional rules to take them as binding and for individuals generally to support or at least to not interfere with the operation of these institutions is that they provide benefits that cannot otherwise be obtained. If an institution cannot effectively perform the functions invoked to justify its existence, then this insufficiency undermines its claim to the right to rule.

"Benefit" here is comparative. The legitimacy of an institution is called into question if there is an institutional alternative, providing significantly greater benefits, that is feasible, accessible without excessive transition costs, and meets the minimal moral acceptability criterion. The most difficult issues, as discussed below, concern trade-offs between comparative benefit and our other criteria. Legitimacy is not to be confused with optimal efficacy and efficiency. The other values that we discuss are also important in their own right; and in any case, institutional stability is a virtue. Nevertheless, if an institution steadfastly remains instrumentally suboptimal when it could take steps to become significantly more efficient or effective, this could impugn its legitimacy in an indirect way: it would indicate that those in charge of the institution were either grossly incompetent or not seriously committed to providing the benefits that were invoked to justify the creation of the institution in the first place. For instance, as of the beginning of 2006 the United Nations faced the issue of reconstituting a Human Rights Commission that had been discredited by the membership of states that notoriously abuse human rights, with Libya serving as chair in 2003. (20)

Institutional Integrity. If an institution exhibits a pattern of egregious disparity between its actual performance, on the one hand, and its self-proclaimed procedures or major goals, on the other, its legitimacy is seriously called into question. The United Nations Oil-for-Food scandal is a case in point. The Oil-for-Food Program was devised to enable Iraqi oil to be sold, under strict controls, to pay for food imports under the UN-mandated sanctions of the 1990s. The purpose was both to prevent malnutrition in Iraq and to counter Iraqi propaganda holding the United Nations responsible for the deaths of hundreds of thousands of Iraqi children, without relieving the pressure on Saddam Hussein's regime to get rid of its supposed weapons of mass destruction. Yet it led to a great deal of corruption. Oil-for-Food became a huge program, permitting the government of Iraq to sell $64.2 billion of oil to 248 companies, and enabling 3,614 companies to sell $34.5 billion of humanitarian goods to Iraq. Yet more than half of the companies involved paid illegal surcharges or kickbacks to Saddam and his cronies, resulting in large profits for corporations and pecuniary benefits for some program administrators, including at least one high-level UN official. (21) The most damning charge is that neither the Security Council oversight bodies nor the Office of the Secretary-General followed the UN's prescribed procedures for accountability. At least when viewed in the light of the historical record of other, perhaps less egregious failures of accountability in the use of resources on the part of the UN, these findings have raised questions about the legitimacy of the Security Council and the secretariat.

It also appears that an institution should be presumed to be illegitimate .if its practices or procedures predictably undermine the pursuit of the very goals in terms of which it justifies its existence. Thus, for example, if the fundamental character of the Security Council's decision-making process renders that institution incapable of successfully pursuing what it now acknowledges as one of its chief goals--stopping large-scale violations of basic human rights--this impugns its legitimacy. To take another example, Randall Stone has shown that the IMF during the 1990s inconsistently applied its own standards with respect to its lending, systematically relaxing enforcement on countries that had rich and powerful patrons. (22) Similarly, if the WTO claims to provide the benefits of trade liberalization to all of its members, but consistently develops policies that exclude its weaker members from the benefits of liberalization, this undermines its claim to legitimacy. If an institution fails to satisfy the integrity criterion, we have reason to believe that key institutional agents are either untrustworthy or grossly incompetent, that the institution lacks correctives for these deficiencies, and that the institution is therefore unlikely to be effective in providing the goods that would give it a claim to our support.

Integrity and comparative benefit are related but distinct. If there are major discrepancies between an institution's behavior and its prescribed procedures and professed goals, then we can have little confidence that it will succeed in delivering the benefits it is supposed to provide. Integrity, however, is a more forward-looking, dynamic virtue than comparative benefit, which measures benefit solely in terms of the current situation. If an institution satisfies the criterion of integrity, there is reason to be confident that institutional actors will not only deliver the benefits that are now taken to constitute the proper goals of institutional activity, but also that they will be able to maintain the institution's effectiveness if its goals change.

Epistemic Aspects of Legitimacy

Minimal moral acceptability, comparative benefit, and institutional integrity are plausible presumptive substantive requirements for the legitimacy of global governance institutions. It would be excessive to claim that they are necessary conditions simpliciter, because there might be extraordinary circumstances in which an institution would fail to satisfy one or two of them, yet still reasonably be regarded as legitimate. This might be the case if there were no feasible and accessible alternative institutional arrangement, if the noninstitutional alternative were sufficiently grim, and if there was reason to believe that the institution had the resources and the political will to correct the deficiency. How much we expect of an institution should depend, inter alia, upon how valuable the benefits it provides are and whether there are acceptable, feasible alternatives to it. For example, we might be warranted in regarding an institution as legitimate even though it lacked integrity, if it were nonetheless providing important protections for basic human rights and the alternatives to relying on it were even less acceptable. In contrast, the fact that an institution is effective in incrementally liberalizing trade would not be sufficient to rebut the presumption that it is illegitimate because it abuses human rights. (23)

Our three substantive conditions are best thought of as what Rawls calls "counting principles": the more of them an institution satisfies, and the higher the degree to which it satisfies them, the stronger its claim to legitimacy. (24)

There are two limitations on the applicability of these three criteria, however. The first is the problem of factual knowledge: being able to make reasonable judgments about whether an institution satisfies any of the three substantive conditions requires considerable information about the workings of the institution and their effects in a number of domains, as well as about the likely effects of feasible alternatives. Some institutions may not only fail to supply the needed information, however; they may, whether deliberately or otherwise, make such information either impossible for outsiders to obtain or make obtaining it prohibitively costly. Even if the institution does not try to limit access to the relevant information, it may not be accessible, in suitably integrated, understandable form.

The second difficulty with taking the three substantive conditions as jointly sufficient for legitimacy is the problem of moral disagreement and uncertainty noted earlier. Even if there is sufficient agreement on what counts as the violation of basic human rights, there are ongoing disputes about whether some global governance institutions should meet higher moral standards. As emphasized above, there is not only disagreement but also uncertainty as to the role that some of these institutions should play in the pursuit of global justice, chiefly because we do not have a coherent idea of what the institutional division of labor for achieving global justice would look like.

Furthermore, merely requiring that global governance institutions not violate basic human rights is unresponsive to the familiar complaint that rich countries unfairly dominate them, and that even if they provide benefits to all, the richer members receive unjustifiably greater benefits. Although all parties may agree that fairness matters, however, there are likely to be disagreements about what fairness would consist of, disputes about whether fairness would suffice or whether equality is required, and about how equality is to be understood and even over what is to be made equal (welfare, opportunities, resources, and so on). So, quite apart from the issue of what positive role, if any, these institutions should play in the pursuit of global justice, there is disagreement about what standards of fairness they should meet internally. There is also likely to be disagreement about how unfair an institution must be to lack legitimacy. A proposal for a public global standard of legitimacy must not gloss over these disagreements.

In the following sections we argue that the proper response to both the problem of factual knowledge and the problem of moral disagreement and uncertainty is to focus on what might be called the epistemic-deliberative quality of the institution, the extent to which the institution provides reliable information needed for grappling with normative disagreement and uncertainty concerning its proper functions. To lay the groundwork for that argument we begin by considering two items that are often assumed to be obvious requirements for the legitimacy of global governance institutions: accountability and transparency.

Accountability. Critics of global governance institutions often complain that they lack accountability. To understand the strengths and limitations of accountability as a gauge of legitimacy, we start with a skeletal but serviceable analysis of accountability. Accountability includes three elements: first, standards that those who are held accountable are expected to meet; second, information available to accountability holders, who can then apply the standards in question to the performance of those who are held to account; and third, the ability of these accountability holders to impose sanctions--to attach costs to the failure to meet the standards. The need for information about whether the institution is meeting the standards accountability holders apply means that a degree of transparency regarding the institution's operations is essential to any form of accountability.

It is misleading to say that global governance institutions are illegitimate because they lack accountability and to suggest that the key to making them legitimate is to make them accountable. Most global governance institutions, including those whose legitimacy is most strenuously denied, include mechanisms for accountability. (25) The problem is that existing patterns of accountability are morally inadequate. For example, the World Bank has traditionally exhibited a high degree of accountability, but it has been accountability to the biggest donor countries, and the Bank therefore has to act in conformity with their interests, at least insofar as they agree. This kind of accountability does not ensure meaningful participation by those affected by rules or due consideration of their legitimate interests. (26) A high degree of accountability in this case may serve to perpetuate the defects of the institution.

So accountability per se is not sufficient; it must be the right sort of accountability. At the very least, this means that there must be effective provisions in the structure of the institution to hold institutional agents accountable for acting in ways that ensure satisfaction of the minimal moral acceptability and comparative benefit conditions. But accountability understood in this narrow way is not sufficiently dynamic to serve as an assurance of the legitimacy of global governance institutions, given that in some cases there is serious disagreement about what the goals of the institution should be and, more specifically, about what role if any the institution should play in the pursuit of global justice. The point is that what the terms of accountability ought to be--what standards of accountability ought to be employed, who the accountability holders should be, and whose interests the accountability holders should represent--cannot be definitively ascertained without knowing what role, if any, the institution should play in the pursuit of global justice.

Therefore, what might be called narrow accountability--accountability without provision for contestation of the terms of accountability--is insufficient for legitimacy, given the facts of moral disagreement and uncertainty. Because what constitutes appropriate accountability is itself subject to reasonable dispute, the legitimacy of global governance institutions depends in part upon whether they operate in such a way as to facilitate principled, factually informed deliberation about the terms of accountability. There must be provisions for revising existing standards of accountability and current conceptions of who the proper accountability holders are and whose interests they should represent.

Transparency. Achieving transparency is often touted as the proper response to worries about the legitimacy of global governance institutions. (27) But transparency by itself is inadequate. First, if transparency means merely the availability of accurate information about how the institution works, it is insufficient even for narrow accountability--that is, for ensuring that the institution is accurately evaluated in accordance with the current terms of accountability. If information about how the institution operates is to serve the end of narrow accountability, it must be (a) accessible at reasonable cost, (b) properly integrated and interpreted, and (c) directed to the accountability holders. Furthermore, (d) the accountability holders must be adequately motivated to use it properly in evaluating the performance of the relevant institutional agents. Second, if, as we have suggested, the capacity for critically revising the terms of accountability is necessary for legitimacy, information about how the institution works must be available not only to those who are presently designated as accountability holders, but also to those who may contest the terms of accountability.

Broad transparency is needed for critical revisability of the terms of accountability. Both institutional practices and the moral principles that shape the terms of accountability must be revisable in the light of critical reflection and discussion. (28) Under conditions of broad transparency, information produced initially to enable institutionally designated accountability holders to assess officials' performance may be appropriated by agents external to the institution, such as nongovernmental organizations (NGOs) and other actors in transnational civil society, and used to support more fundamental criticisms, not only of the institution's processes and structures, but even of its most fundamental goals and its role in the pursuit of global justice.

One especially important dimension of broad transparency is responsibility for public justification. (29) Institutional actors must offer public justifications of at least the more controversial and consequential institutional policies and must facilitate timely critical responses to them. Potential critics must be in a position to determine whether the public justifications are cogent, whether they are consistent with the current terms of accountability, and whether, if taken seriously, these justifications call for revision of the current terms of responsibility. To help ensure this dimension of broad transparency, it may be worthwhile to draw on, while adapting, the notice and comment procedures of administrative law at the domestic level. (30)

Earlier we noted that although comparative benefit, minimal moral acceptability, and integrity are reasonable presumptive necessary conditions for legitimacy, it may be difficult for those outside the institution to determine whether these conditions are satisfied. We suggest that broad transparency can serve as a proxy for satisfaction of the minimal moral acceptability, comparative benefit, and integrity criteria. For example, it may be easier for outsiders to discover that an institution is not responding to demands for information relevant to determining whether it is violating its own prescribed procedures, than to determine whether in fact it is violating them. Similarly, it may be very difficult to determine whether an institution is comparatively effective in solving certain global problems, but much easier to tell whether it generates--or systematically restricts access to--the information outsiders would need to evaluate its effectiveness. If an institution persistently fails to cooperate in making available to outsiders the information that would be needed to determine whether the three presumptive necessary conditions are satisfied, that by itself creates a presumption that it is illegitimate.

Legitimate global governance institutions should possess three epistemic virtues. First, because their chief function is to achieve coordination, they must generate and properly direct reliable information about coordination points; otherwise they will not satisfy the condition of comparative benefit. Second, because accountability is required to determine whether they are in fact performing their current coordinating functions efficiently and effectively requires narrow transparency, they must at least be transparent in the narrow sense. They must also have effective provisions for integrating and interpreting the information current accountability holders need and for directing it to them. Third, and most demanding, they must have the capacity for revising the terms of accountability, and this requires broad transparency: institutions must facilitate positive information externalities to permit inclusive, informed contestation of their current terms of accountability. There must be provision for ongoing deliberation about what global justice requires and how the institution in question fits into a division of institutional responsibilities for achieving it.

Overcoming Informational Asymmetries

A fundamental problem of institutional accountability is that insiders generally have better information about the institution than outsiders. Outsiders can determine whether institutions enjoy the consent of states, and whether states are democratic; but it may be very difficult for them to reach well-informed conclusions about the minimal moral acceptability, comparative benefit, and integrity conditions. Our emphasis on epistemic institutional virtues is well suited to illuminate these problems of asymmetrical information.

First, if institutional agents persist in failing to provide public justifications for their policies and withhold other information critical to the evaluation of institutional performance, we have good reason to believe the institution is not satisfying the substantive criteria for legitimacy. (31) Second, there may be an asymmetry of knowledge in the other direction as well, and this can have beneficial consequences for institutional accountability. Consider issue areas such as human rights and the environment, which are richly populated with independent NGOs that seek to monitor and criticize national governments and global governance institutions and to suggest policy alternatives. Suppose that in these domains there is a division of labor among external epistemic actors. Some individuals and groups seek information about certain types of issues, while others focus on other aspects, each drawing on distinct but in some cases overlapping groups of experts. Still others specialize in integrating and interpreting information gathered by other external epistemic actors.

The fact that the information held by external epistemic actors is dispersed will make it difficult for institutional agents to know what is known about their behavior or to predict when potentially damaging information may be integrated and interpreted in ways that make it politically potent. The institutional agents' awareness of this asymmetry will provide incentives for avoiding behavior for which they may be criticized. A condition of productive uncertainty will exist: although institutional agents will know that external epistemic actors do not possess the full range of knowledge that they do, they will know that there are many individuals and organizations gathering information about the institution. Further, they will know that some of the information that external epistemic actors have access to can serve as a reliable proxy for information they cannot access. Finally, they will also know that potentially damaging information that is currently harmless because it is dispersed among many external epistemic agents may at any time be integrated and interpreted in such a way as to make it politically effective, but they will not be able to predict when this will occur. Under these conditions, institutional agents will have significant incentives to refrain from behavior that will attract damning criticism, despite the fundamental asymmetry of knowledge between insiders and outsiders.

This is not to say that the effects of transparency will always be benign. Indeed, under some circumstances transparency can have malign effects. As David Stasavage points out, "open-door bargaining ... encourages representatives to posture by adopting overly aggressive bargaining positions that increase the risks of breakdown in negotiations." (32) When issues combine highly charged symbolic elements with the need for incentives, conflicts between transparency and efficiency may be severe. Our claim is not that outcomes are necessarily better the more transparent institutions are. Rather, it is that the dispersal of information among a plurality of external epistemic actors provides some counterbalance to informational asymmetries favoring insiders. There should be a very strong but rebuttable presumption of transparency, because the ills of too much transparency can be corrected by deeper, more sophisticated public discussion, whereas there can be no democratic response to secret action by bureaucracies not accountable to the public.

Furthermore, if national legislatures are to retain their relevance--if what we have called the democratic accountability channel is to be effective--they must be able to review the policies of global governance institutions. (33) For legislatures to have information essential to performing these functions, they need a flow of information from transnational civil society. Monitoring is best done pluralistically by transnational civil society, whereas the sanctions aspects of accountability are more effectively carried out by legislatures. With respect both to the monitoring and sanctioning functions, broad transparency is conducive to the principled revisability of institutions and to their improvement through increasingly inclusive criticism and more deeply probing discussion over time.

Institutional agents generally have incentives to prevent outsiders from getting information that may eventually be interpreted and integrated in damaging ways and to deprive outsiders of information that can serve as a reliable proxy to assess institutional legitimacy. The very reasons that make the epistemic virtues valuable from the standpoint of assessing institutional legitimacy may therefore tempt institutional agents to ensure that their institutions do not exemplify these virtues. But institutional agents are also aware that it is important for their institutions to be widely regarded as legitimate. Outsiders deprived of access to information are likely to react as does the prospective buyer of a used car who is prevented from taking it to an independent mechanic. They will discount the claims of the insiders and may conclude that the institution is illegitimate. So if there is a broad consensus among outsiders that institutions are not legitimate unless they exemplify the epistemic virtues, institutional agents will have a weighty reason to ensure that their institutions do so.

Contestation and Revisability: Links to External Actors and Institutions

We have argued that the legitimacy of global governance institutions depends upon whether there is ongoing, informed, principled contestation of their goals and terms of accountability. This process of contestation and revision depends upon activities of actors outside the institution. It is not enough for the institutions to make information available. Other agents, whose interests and commitments do not coincide too closely with those of the institution, must provide a check on the reliability of the information, integrate it, and make it available in understandable, usable form, to all who have a legitimate interest in the operations of the institution. Such activities can produce positive feedback, in which appeal to standards of legitimacy by the external epistemic actors not only increases compliance with existing standards but also leads to improvements in the quality of these standards themselves. For these reasons, in the absence of global democracy, and given the limitations of the democratic channel described earlier, legitimacy depends crucially upon not only the epistemic virtues of the institution itself but also on the activities of external episternic actors. Effective linkage between the institution and external epistemic actors constitutes what might be called the transnational civil society channel of accountability.

The needed external epistemic actors, if they are effective, will themselves be institutionally organized. (34) Institutional legitimacy, then, is not simply a function of the institution's characteristics; it also depends upon the broader institutional environment in which the particular institution exists. To borrow a biological metaphor, ours is an ecological conception of legitimacy.

All three elements of our complex standard of legitimacy are now in place. First, global governance institutions should enjoy the ongoing consent of democratic states. That is, the democratic accountability channel must function reasonably well. Second, these institutions should satisfy the substantive criteria of minimal moral acceptability, comparative benefit, and institutional integrity. Third, they should possess the epistemic virtues needed to make credible judgments about whether the three substantive criteria are satisfied and to achieve the ongoing contestation and critical revision of their goals, their terms of accountability, and ultimately their role in a division of labor for the pursuit of global justice, through their interaction with effective external epistemic agents.

The Complex Standard frames the legitimacy of global governance institutions as both dynamic and relational. Its emphasis on the conditions for ongoing contestation and critical revision of the most basic features of the institutions captures the exceptional moral disagreement and uncertainty that characterize the circumstances of legitimacy for this type of institution. While acknowledging the facts of moral disagreement and uncertainty, the Complex Standard includes provisions for developing more robust moral requirements for institutions over time. The Complex Standard also makes it clear that whether the institution is legitimate does not depend solely upon its own characteristics, but also upon the epistemic-deliberative relationships between the institution and epistemic actors outside it.

A Place for Democratic Values in the Absence of Global Democracy

Earlier we argued that it is a mistake to hold global governance institutions to the standard of democratic legitimacy that is now widely applied to states. We now want to suggest that when the Complex Standard of legitimacy we propose is satisfied, important democratic values will be served. For purposes of the present discussion we will assume, rather than argue, that among the most important democratic values are the following: first, equal regard for the fundamental interests of all persons; second, decision-making about the public order through principled, collective deliberation; and third, mutual respect for persons as beings who are guided by reasons.

If the Complex Standard of legitimacy we propose is satisfied, all three of these values will be served. To the extent that connections between the institutions and external epistemic actors provide access to information that is not restricted to certain groups but available globally, it becomes harder for institutions to continue to exclude consideration of the interests of certain groups, and we move closer toward the ideal of equal regard for the fundamental interests of all. Furthermore, by making information available globally, networks of external epistemic actors are in effect addressing all people as individuals for whom moral reasons, not just the threat of coercion, determine whether they regard an institution's rules as authoritative. Finally, if the Complex Standard of legitimacy is satisfied, every feature of the institution becomes a potential object of principled, informed, collective deliberation, and eligibility for participation in deliberation will not be restricted by institutional interests. (35)

Consistency with Democratic Sovereignty

One source of doubts about the legitimacy of global governance institutions is the worry that they are incompatible with democratic sovereignty. Our analysis shows why and how global governance should constrain democratic sovereignty. The standard of legitimacy we propose is designed inter alia to help global governance institutions correct for the tendency of democratic governments to disregard the interests and preferences of those outside their own publics. It does this chiefly in two ways. First, the emphasis on the role of external institutional epistemic actors in achieving broad accountability helps to ensure more inclusive representation of interests and preferences over time. Second, the requirement of minimal moral acceptability, understood as nonviolation of basic human rights, provides an important protection for the most vulnerable: if this condition is met, democratic publics cannot ignore the most serious "negative externalities" of their policy choices. Global governance institutions that satisfy our standard of legitimacy should not be viewed as undermining democratic sovereignty, but rather as enabling democracies to function justly.

A legitimate global order will include human rights institutions that promote the conditions for the proper functioning of democracy (the right to basic education, the right to freedom of expression and association, and so on) in countries that are democratizing and help sustain these conditions in countries that already have democratic institutions. Critics of global governance institutions that claim they are illegitimate because they constrain democratic sovereignty either beg the question by assuming that the "will of the people" should not be constrained so as to take into account the interests of those outside their polity or they underestimate the extent to which democracy depends upon global governance institutions.

Having articulated the Complex Standard, and indicated how it reflects several key democratic values, we can now show, briefly, how it satisfies the desiderata for a standard of legitimacy we set out earlier.

1. The Complex Standard provides a reasonable basis for coordinated support of institutions that meet the standard, support based on moral reasons that are widely accessible in the circumstances under which legitimacy is an issue. To serve the social function of legitimacy assessments, the Complex Standard only requires a consensus on the importance of not violating the most widely recognized human rights, broad agreement that comparative benefit and integrity are also presumptive necessary conditions of legitimacy, and a commitment to inclusive, informed deliberation directed toward resolving or at least reducing the moral disagreement and uncertainty that characterize our practical attitudes toward these institutions. In other words, the Complex Standard steers a middle course between requiring more moral agreement than is available in the circumstances of legitimacy and abandoning the attempt to construct a more robust, shared moral perspective from which to evaluate global governance institutions. In particular, the Complex Standard acknowledges that the role that these institutions ought to play in a more just world order is both deeply contested and probably not knowable at present.

2. In requiring only minimal moral acceptability at present, the Complex Standard acknowledges that legitimacy does not require justice, but at the same time affirms the intuition that extreme injustice, understood as violation of the most widely recognized human rights, robs an institution of legitimacy.

3. The Complex Standard takes the ongoing consent of democratic states to be a presumptive necessity, though not a sufficient condition for legitimacy.

4. The Complex Standard rejects the assumption that global governance institutions cannot be legitimate unless there is global democracy, but at the same time promotes some of the key democratic values, including informed, public deliberation conducted on the assumption that every individual has standing to participate and the requirement that key institutional policies must be publicly justified.

5. The Complex Standard reflects a proper appreciation of the dynamic, experimental character of global governance institutions and of the fact that not only the means they employ but even the goals they pursue may and probably should change over time.

6. The Complex Standard's requirement of a functioning transnational civil society channel of accountability--an array of overlapping networks of external epistemic actors--helps to compensate for the limitations of accountability through democratic state consent.

The central argument of this essay can now be summarized. The Complex Standard provides a reasonable basis for agreement in legitimacy assessments of global governance institutions. When the comparative benefit condition is satisfied, the institution provides goods that are not readily obtainable without it. These goods, however, can be reliably provided only if coordination is achieved, and achieving coordination without excessive costs requires that the relevant agents regard the institution's rules as presumptively binding--that is, that they take the fact that the rule is issued by the institution as a content-independent reason for compliance. The instrumental value of institutions that satisfy the comparative benefit condition also gives individuals generally a content-independent reason not to interfere with the functioning of the institutions. Satisfaction of the minimal moral acceptability condition rules out the more serious moral objections that might otherwise undercut the instrumental reasons for supporting the institution. Satisfaction of the other conditions of the Complex Standard, taken together, provides moral reasons to support or at least not interfere with the institution. Among the most important of these reasons is that the institution has epistemic virtues that facilitate the development of more demanding standards and the progressive improvement of the institution itself. Thus, when a global governance institution meets the demands of the Complex Standard, there is justification for saying that it has the right to rule, not merely that it is beneficial.

CONCLUSION

In this essay we have offered a proposal for a public standard of legitimacy for global governance institutions. These institutions supply important benefits that neither states nor traditional treaty-based relationships among states can provide, but they are quite new, often fragile, and still evolving. Politically mobilized challenges to the legitimacy of these institutions jeopardize the support they need to function effectively, in spite of the fact that these challenges are typically unprincipled and possibly grounded in unrealistic demands that confuse justice with legitimacy. A principled global public standard of legitimacy could facilitate more responsible criticism while at the same time providing guidance for improvement, through a process of institutionalized, collective learning, both about what it is reasonable to expect from global governance institutions and about how to achieve it. Our hope is that the proposal offered in this paper serves these purposes.

Allen Buchanan and Robert O. Keohane *

* The authors are grateful to Sahar Akhtar, Christian Barry, Thomas Christiano, Michael Doyle, Nicole Hassoun, Andrew Hurrell, Nan Keohane, Avery Kolers, Joseph S. Nye, John Tasioulas, and two anonymous referees for their helpful comments on earlier versions of this paper, and to William Alford, Ryan Goodman, and Gerald L. Neuman for valuable criticisms and suggestions when such a version was presented at Harvard Law School, November 3, 2005. We are particularly grateful to comments by Charles Beitz and a number of other colleagues made at a workshop on the normative and empirical evaluation of global governance, Princeton University, February 17-18, 2006. Further useful comments were made at the conference on "Legitimacy and International Law" at the Max Planck Institute for Comparative Public Law and International Law, Heidelberg, Germany, June 13-14, 2006.

(1) A thorough review of the sociological literature on organizational legitimacy can be found in Mark C. Suchman, "Managing Legitimacy: Strategic and Institutional Approaches," Academy of Management Review 20, no. 3 (1995), pp. 571-610.

(2) For an excellent discussion of the inadequacy of existing standards of legitimacy for global governance institutions, see Daniel Bodansky, "The Legitimacy of International Governance: A Coming Challenge for International Environmental Law?" American Journal of International Law 93, no. 3 (1999), pp. 596-624. For an impressive earlier book on the subject, see Thomas Franck, The Power of Legitimacy Among Nations (New York: Oxford University Press, 1990). Franck's account focuses on the legitimacy of rules more than institutions and in our judgment does not distinguish clearly enough between the normative and sociological senses of legitimacy.

(3) A large and growing literature exists on global governance. See, for example, Aseem Prakash and Jeffrey A. Hart, eds., Globalization and Governance (London: Routledge, 1999); Joseph S. Nye and John D. Donahue, eds., Governance in a Globalizing World (Washington: Brookings Institution Press, 200o); and David Held and Anthony McGrew, eds., Governing Globalization (London: Polity Press, 2002).

(4) Erika de Wet, "The Security Council as Legislator/Executive in Its Fight against Terrorism and against Proliferation of Weapons of Mass Destruction: The Question of Legitimacy" (presentation at the conference "Legitimacy and International Law," Max Planck Institute for Comparative Public Law and International Law, Heidelberg, Germany, June 14, 2006).

(5) The emphasis here on the coordinating function should not be misunderstood: global governance institutions do not merely coordinate state actions in order to satisfy preexisting state preferences. As our analysis will make clear, they can also help shape state preferences and lead to the development of new norms and institutional goals.

(6) Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton: Princeton University Press, 1984 [20th anniversary edition, 2005]).

(7) James D. Fearon, "Bargaining, Enforcement, and International Cooperation," International Organization 52, no. 2 (Spring 1998), pp. 269-306.

(8) This is a major theme of Russell Hardin, Liberalism, Constitutionalism, and Democracy (Oxford: Oxford University Press, 1999).

(9) Andrew Hurrell, "Legitimacy and the Use of Force: Can the Circle Be Squared?" Review of International Studies 31, supp. S1 (2005), p. 16.

(10) Legitimacy can also be seen as providing a "focal point" that helps strategic actors select one equilibrium solution among others. For the classic discussion of focal points, see Thomas C. Schelling, The Strategy of Conflict (Cambridge: Harvard University Press, 1960), ch. 3. For a critique of theories of cooperation on the basis of focal point theory, and an application to the European Union, see Geoffrey Garrett and Barry Weingast, "Ideas, Interests, and Institutions: Constructing the European Community's Internal Market," in Judith Goldstein and Robert O. Keohane, eds., Ideas and Foreign Policy: Beliefs, Institutions and Political Change (Ithaca: Cornell University Press, 1993), esp. pp. 178-85.

(11) Most contemporary analytic philosophical literature on legitimacy tends to focus exclusively on the legitimacy of the state and typically assumes a very strong understanding of legitimacy. In particular, it is assumed that legitimacy entails (1) a content-independent moral obligation to comply with all institutional rules (not just content-independent moral reasons to comply and/or a content-independent moral obligation to not interfere with others' compliance), (2) being justified in using coercion to secure compliance with rules, and (3) being justified in using coercion to exclude other actors from operating in the institution's domain. (See, for example, Christopher Heath Wellman and A. John Simmons, Is There a Duty to Obey the Law? For and Against [Cambridge: Cambridge University Press, 2005]). It is far from obvious, however, that this very strong conception is even the only conception of legitimacy appropriate for the state, given what is sometimes referred to as the "unbundling" of sovereignty into various types of decentralized states and the existence of the European Union. Be that as it may, this state-centered conception is too strong for global governance institutions, which generally do not wield coercive power or claim such strong authority. For a more detailed development of this point, see Allen Buchanan, "The Legitimacy of International Law," in Samantha Besson and John Tasioulas, eds., The Philosophy of International Lave (Oxford: Oxford University Press, forthcoming).

(12) This view was forcefully expressed by Professor Yoram Dinstein of Tel Aviv University, in comments on a draft of this essay.

(13) For a more detailed discussion, see Allen Buchanan, Justice, Legitimacy and Self-Determination: Moral Foundations for International Law (Oxford: Oxford University Press, 2003), esp. ch. 5.

(14) For a perceptive discussion of how consent to new international trade rules in the Uruguay Round (1986-94) was merely nominal, since the alternatives for poor countries were so unattractive, see Richard H. Steinberg, "In the Shadow of Law or Power? Consensus-based Bargaining and Outcomes in the GATT/WTO," International Organization 56, no. 2 (2002), pp. 339-74.

(15) How the requirement of ongoing consent should be operationalized is a complex question we need not try to answer here; one possibility would be that the treaties creating the institution would have to be periodically reaffirmed.

(16) Buchanan, "The Legitimacy of International Law."

(17) For a valuable discussion that employs a different conception of normative uncertainty, see Monica Hlavac, "A Developmental Approach to the Legitimacy of Global Governance Institutions" (unpublished paper).

(18) See Joseph Raz, The Morality of Freedom (New York: Oxford University Press, 1986), n. 17.

(19) Allen Buchanan and Matthew DeCamp, "Responsibility for Global Health," Transnational Medicine (forthcoming).

(20) In March 2005, Secretary-General Kofi Annan called for the replacement of the Commission on Human Rights (fifty-three members elected from slates put forward by regional groups) with a smaller Human Rights Council elected by a two-thirds vote of members of the General Assembly (see his report 'tin Larger Freedom," A/59/2005, para. 183).

(21) For the report of the Independent Inquiry Committee into the United Nations Oil-for-Food Program (the Volcker Committee), dated October 27, 2005, see www.iic-offp.org/story27oct05.htm.

(22) Randall W. Stone, "The Political Economy of IMF Lending in Africa," American Political Science Review 98, no. 4 (2004), pp. 577-91. See also Randall W. Stone, Lending Credibility: The International Monetary Fund and the Post-Communist Transition (Princeton: Princeton University Press, 2002).

(23) We are indebted to Andrew Hurrell for this example.

(24) John Rawls, A Theory of Justice (Cambridge: Harvard University Press, 1971).

(25) Ruth W. Grant and Robert O. Keohane, "Accountability and Abuses of Power in World Politics," American Political Science Review 99, no. 1 (2005), pp. 29-44. See also Robert O. Keohane and Joseph S. Nye, "Redefining Accountability for Global Governance," in Miles Kahler and David A. Lake, eds., Governance in a Global Economy: Political Authority in Transition (Princeton: Princeton University Press, 2003), pp. 386-411.

(26) For a discussion, see Ngaire Woods, "Holding Intergovernmental Institutions to Account," Ethics & International Affairs 17, no. 1 (2003), pp. 69-80.

(27) Ann Florini, The Coming Democracy (Washington, D.C.: Island Press, 2003).

(28) For a discussion of the role of critical revisability in practical reasoning, with parallels to theoretical reasoning, see Mien Buchanan, "Revisability and Rational Choice," Canadian Journal of Philosophy 5, no. 3 (1975), pp. 395-408.

(29) For an illuminating account of the legitimacy of health care institutions that emphasizes responsibility for justifications, see Norman Daniels and James Sabin, "Limits to Health Care: Fair Procedures, Democratic Deliberation, and the Legitimacy Problem for Insurers," Philosophy & Public Affairs 26, no. 4 (1997), pp. 303-50.

(30) See Richard B. Stewart, "Administrative Law in the Twenty-First Century," New York University Law Review 78, no. 2 (2003), pp. 437-60; and Benedict Kingsbury, Nikon Kirsch, and Richard B. Stewart, "The Emergence of Global Administrative Law," Law and Contemporary Problems 68, nos. 3 and 4 (2005). See also Daniel Esty, "Toward Good Global Governance: The Role of Administrative Law" (paper presented at a conference on global administrative law, New York University, April 21-23, 2005). See also John Wickham, "Toward a Green Multilateral Investment Framework: NAFTA and the Search for Models," Georgetown International Environmental Law Review 12, no. 3 (2000), pp. 617-46; James Salzman," Labor Rights, Globalization, and Institutions: The Role and Influence of the Organization for Economic Cooperation and Development," Michigan Journal of International Law 21, no. 4 (2000), pp. 769-848; and OECD, Getting to Grips with Globalization: The OECD in a Changing World (Paris: OECD Publications, 2004).

(31) The analogy in the economics of information is to the market for used cars. A potential buyer of a used car would be justified in inferring poor quality if the seller were unwilling to let him have the car thoroughly examined by a competent mechanic. See George A. Akerlof, "The Market for Lemons: Quality Uncertainty and the Market Mechanism," Quarterly Journal of Economics 84, no. 3 (1970), pp. 488-500.

(32) David Stasavage, "Open-Door or Closed-Door? Transparency in Domestic and International Bargaining," International Organization 58, no. 4 (2004), pp. 667-704.

(33) On the role of legislatures with respect to the legitimacy of an international legal order, see Rudiger Wolfrum, "Legitimacy in International Law: Some Introductory Considerations" (paper prepared for the conference "Legitimacy in International Law" at the Max Planck Institute for Comparative Public Law and International Law, Heidelberg, Germany, June 13-14, 2006).

(34) We use the term "external epistemic actor" here broadly, to include individuals and groups outside the institution in question who gain knowledge about the institution, interpret and integrate such knowledge, and exchange it with others, in ways that are intended to influence institutional behavior, whether directly or indirectly (through the mediation of the activities of other individuals and groups).

(35) On our view, the legitimacy of global governance institutions, at present at least, does not require participation in the critical evaluation of institutional goals and policies by all who are affected by them; but if the standard of legitimacy we recommend were accepted, opportunities for participation would expand.