The Theoretical Framework

Joseph Nye assessed the common themes connecting neoliberal thinkers from the pioneering works of Ernst Haas, Karl Deutsch, and Nye to the present. All share

a focus on the ways in which increased transactions and contacts changed attitudes and transnational coalition opportunities, and the ways in which institutions help foster such interaction. In short, they emphasized the political process of learning and of redefining national interests, as encouraged by institutional frameworks and regimes.^[6]

Nye's statement of the common thread in neoliberal analyses embraces the present study. The key analytical theme in examining European telematics collaboration is precisely the "political process of learning and of redefining national interests," with a focus on how international institutions can stimulate and lead in that process.

[6] Joseph S. Nye, Jr., "Neorealism and Neoliberalism," 239.

Nation-states must rely principally on their own means to ensure their security, survival, and prosperity. Leaders of nations seek to maximize their autonomy from other states. Assuming such a preference for autonomy does not require that one build on it the same theoretical superstructure that realists do. Rather, it seems more fruitful to think of the preference for autonomy as leading to a problem of collective action at the international level. Cooperation necessarily imposes constraints on national leaders, foreclosing some options, requiring others, or both. Cooperation implies some degree of policy adjustment. By this definition, when states pursue policies that are mutually beneficial and do not require any policy adjustments, that is harmony of interests, not cooperation.^[7] State leaders in general, therefore, prefer autonomy to cooperation. Students of international politics face two major tasks. First, we must show why state leaders choose courses that limit their international autonomy; we must account for the demand for cooperation. Second, we must address the collective-action problem at the international level.

To explain the paths from a preference for autonomy to international cooperation, I propose a three-part theoretical framework. The first two parts deal with demand and supply: the demand for cooperation and the supply of political leadership. Robert Keohane employed the same terminology in explaining the demand for regimes.^[8] Like Keohane, I use the notions of demand and supply heuristically, not to imply a market for international cooperation. My approach differs from his in two respects, though. First, I analyze the demand for cooperation not by how regimes meet functional needs of states but rather by how the preferences and beliefs of state leaders shift so as to make cooperation possible. Second, unlike Keohane, I devote equal attention to the supply of leadership.

The third stage of analysis is to examine the arrangements for distributing the costs and benefits of cooperation. The theoretical proposition involved is that state leaders must perceive an adequate balance between contributions and rewards, or they will not participate. With technological collaboration this question takes on the specific meaning of finding solutions to the problem of juste retour . I will propose two potential mechanisms for resolving that problem.

[7] Keohane makes a similar argument in After Hegemony .

[8] Robert O. Keohane, "The Demand for International Regimes."

What follows is not a general theory of international cooperation. I doubt that a single theory could adequately explain the various species of the phenomenon. The analytical framework developed below builds on existing theories and binds them into a whole that is, I suggest, larger than the sum of its parts. It is a strategy for analysis, a guide for structuring explanations.

Cognitive Change

The demand for cooperation involves cognitive change. For cooperation to emerge state leaders must come to believe that unilateral means will not achieve valued ends. This explanation implies rationality. I assume that policy-makers and organizations can learn. All the players in the drama of international cooperation (politicians, bureaucrats, officials of international organizations, industry executives) are rational actors. But their rationality is the bounded rationality described by Herbert Simon. They are rational in that they choose among options so as to maximize the attainment of values on the basis of available information and resources. They are not capable of ranking all preferences, perceiving every alternative, foreseeing all consequences, or making all relevant value tradeoffs.^[9] Actor preferences are not deducible solely from bureaucratic position, electoral considerations, personal economic interests, or any other single, reductionist source. The sources of preferences are complex and include, in addition to the factors just mentioned, ideologies and beliefs about causality (the relation between ends and means).

Two different kinds of cognitive change can lead policy-makers to conclude that unilateral strategies are insufficient. Each cognitive process corresponds to a different class of international problem. For some kinds of problems it is clear that other states are inextricably involved in any solution—that is, solutions require policy adjustments in more than one state; unilateral approaches possess little or no value. In some instances, this need for coordination may be obvious, as in the handling of international mail. Sometimes it is not at all obvious that a problem is multilateral in nature. Indeed, it may not even be clear that a problem exists. For instance, before there was a technical understanding of the degradation of the ozone

[9] Herbert A. Simon, Administrative Behavior .

layer, it was not the clear candidate for multilateral concern that it became with the 1987 Montreal Protocols to limit emissions of chlorofluorocarbons. I label this class of issues problems of scope because their resolution inherently requires the participation of more than one state.

Problems of scope like the handling of international mail are obvious and hence only marginally interesting. Other problems of scope require cognitive evolution on the part of national leaders. The type of cognitive change involved is learning as defined by Ernst Haas: the use of consensual knowledge to redefine causal relations so as to affect public policy. Learning entails the "sharing of larger meanings among those who learn." Consensual knowledge refers to beliefs about cause/effect linkages that are derived from "information, scientific and nonscientific, available about a given subject and considered authoritative by the interested parties." It is social knowledge, subject to evolution and revision.^[10]

Scientists and experts generate new information and theories; sometimes these acquire broad agreement in the relevant scientific community. Of course, different political actors can use different bodies of scientific knowledge (or different interpretations of a single body of knowledge) to push for contrary policy options. So scientific facts and theories become consensual knowledge for public decision-making only when they are accepted as valid for policy-making purposes by the relevant politicians, interest groups, and bureaucrats. International cooperation can occur when national leaders reach a consensus on the relevant technical information. One way to apply the notion of learning is by focusing on epistemic communities, the networks of actors who develop and diffuse consensual ideas on policy problems.^[11]

Problems of scale constitute a separate category. They do not inherently require multilateral solutions. The crucial concern with problems of scale is not how the problem transcends national boundaries but rather how to muster the level of resources needed to address the problem. Some states are large enough or wealthy enough to go it alone on many issues. Other states cannot marshal the resources needed at acceptable political cost. The United States has so far been able to sustain two companies producing civilian

[10] Ernst B. Haas, When Knowledge Is Power , 21–24.

[11] An interesting example of this approach is Peter M. Haas, "Do Regimes Matter? Epistemic Communities and Mediterranean Pollution Control."

airliners, Boeing and McDonnell Douglas. But four European countries found that they could not individually sustain aircraft industries and they banded together in Airbus.

A crucial difficulty is that there is no objective level of resources needed to address any given problem of scale. Ambiguity clouds the issue because so much depends on the political component, on how many resources political authorities are able to devote to the problem. In problems of scale, the viability of unilateral solutions depends both on the scale of resources potentially available and on the capacity of political actors to mobilize and focus those resources. For example, a small or poor country can autonomously pursue large-scale projects if the political system is capable of focusing resources. India's space program may fit in this category. Much depends on the value governments place on autonomy in a given area and on the opportunity costs they are politically able to accept in order to pursue it.

In other words, with problems of scale the resource threshold is in part a matter of perception and political capacity. Furthermore, national leaders may not know the cost of unilateral policies in advance, much less whether they can afford them. The outside observer cannot know either; countries with apparently limited resources can be quite dogged in pursuing autonomy. As a consequence, I propose, state leaders frequently learn through experience the price of autonomy and the political limits of their resources.^[12] The Mitterrand government in France learned from experience in 1981–1983 that it could not afford a unilateral policy of reflation. Because states prefer autonomy, I hypothesize that they will generally attempt unilateral strategies first and surrender the goal of autonomy only when unilateral means have proven to be impossible or too costly.

The process by which states discover the practical limits to autonomy is a species of cognitive change that I call adaptation. Adaptation is a less far-reaching cognitive shift than learning, as defined above. It entails a reconsideration of the means chosen to pursue policy objectives. In adaptation decision-makers do not revalue their ultimate ends in light of consensual knowledge; they search for new

[12] Technically countries or states do nothing; people acting in or for them do. When I write that states "choose" or "learn" something, I use the word state as shorthand for the leaders of the state.

means to achieve those ends.^[13] Of course, state leaders link objectives in hierarchies or chains, with lower-level goals serving as means to higher-level ones.^[14] At the most abstract level (for example, "economic growth" as a national goal) these policy ends probably never change. A lower-level goal might be "national capacity in high-technology sectors," seen as necessary to achieve economic growth. Adaptation at this link in the ends/means chain is probably extremely gradual. Subordinate to high-tech capacity might be the goal of sustaining on national soil an enterprise capable of competing in global markets, a national champion. Adaptation is probably more common at this lower level than at higher levels. The adaptations examined in this study take place among subordinate means. Not even the mid-level goal of high-tech capacities comes into question.

In one sense lower-level adaptation occurs constantly, in the process Charles Lindblom calls "muddling through," by which existing policies are altered in a piecemeal way.^[15] Keohane and Nye distinguish between "incremental" and "crisis-induced" learning, though the phenomenon they refer to is not learning as defined by Haas but adaptation in the sense I use here.^[16] Muddling through equates with incremental adaptation. Crisis-induced adaptation involves more than just muddling through. Failures, breakdowns, and crises provoke leaders to search for new policies.^[17]

In the search for new policy instruments decision-makers abandon the means but not the ultimate ends behind a policy that has been invalidated by experience. In this sense, as Martin Landau has shown, policies are like hypotheses and can be falsified.^[18] In the case

[13] This definition corresponds to that employed by Ernst Haas, who draws a similar distinction between learning and adaptation. The principal difference is that Haas applies the notions to international organizations and I apply them to national decision-makers. See Haas, When Knowledge Is Power , 33–34.

[14] On hierarchies of ends and means in decision-making, see Simon, Administrative Behavior , chaps. 3–4.

[15] Charles E. Lindblom, "The Science of 'Muddling Through.'"

[16] Robert O. Keohane and Joseph S. Nye, Jr., "Power and Interdependence Revisited," 751.

[17] By crisis I do not mean confrontations involving the threat of military force, as in the "crisis literature." Rather, I refer to a condition described by Ernst Haas as "the compounding of uncertainty in the minds of actors engaged in collective decision-making—uncertainty about the adequacy of cause-and-effect links carried over from past experience, about the proper ranking of values in competition, about the future toward which one should be working," though I emphasize the first element. See Ernst B. Haas, The Obsolescence of Regional Integration Theory , 25.

[18] Martin Landau, "On the Concept of a Self-Correcting Organization."

of telematics collaboration in Europe, tinkering with the national-champion strategies involved incremental adaptation. Adding a collaborative component to policy displayed crisis-induced adaptation. Crisis-induced adaptation leads to purposive searching. I shall use adaptation as shorthand for crisis-induced adaptation from now on and shall call the searching it provokes the adaptive mode .

The difference between adaptation and learning consists of this: Learning has to do with the nature of the problem; adaptation has to do with the limits of the material and political resources of the state. The notion of adaptation implies that the failure of unilateral strategies provokes a rethinking of policy and a search for new approaches. When several states simultaneously face a problem of scale, potential demand for cooperation exists.

Leadership and International Organizations

The second part of this analytical schema confronts the question, Who organizes cooperation? Policy adaptation in a set of countries may create an opportunity for cooperation, as state leaders react to failed unilateral strategies by searching for new approaches. But cooperation does not necessarily emerge self-created out of the soup of failed unilateral strategies. Some political actor (or actors) must propose cooperation and sell the idea to potential collaborators. Such a political leader or entrepreneur must mobilize a coalition in favor of cooperation. Without a leader, the demand for cooperation is likely to remain latent.

Some theorists make the case that, in certain circumstances, cooperation can emerge spontaneously from the unorganized, self-interested behavior of rational actors. Axelrod demonstrates that cooperation can evolve, provided a sufficiently large core of actors chooses the right kind of strategy (tit-for-tat) and their interactions extend over an indefinite period of time.^[19] This line of analysis has been elaborated and applied to a variety of international interactions. The contributors to the volume Cooperation under Anarchy , for example, emphasize three aspects of the structure of the strategic game as "favoring" cooperation: mutuality of interest (the structure of payoffs), iteration of the game, and the number of actors.^[20]

[19] Axelrod, Evolution of Cooperation .

[20] Kenneth A. Oye, ed., Cooperation under Anarchy .

But none of the analysts developing or applying these notions claims that these game-theoretic dimensions are either necessary or sufficient for international cooperation to arise. And though it is certainly worthwhile specifying some conditions that make cooperation likely, their analyses leave open an immense array of other variables that are crucial in determining the effects of the gametheoretic factors. In fact, Axelrod and Keohane, in concluding Cooperation under Anarchy , point out that the context of an issue frequently has "a decisive impact on its politics and outcomes." They argue that the perceptions of actors and international institutions can be of particular import.^[21]

The cooperation-under-anarchy approach needs shoring up in another way as well. In many (if not most) cases costs are associated with organizing cooperation. States must frequently be persuaded to appear at the discussion table. Someone must propose mechanisms for implementing cooperation and must broker compromises. Political leaders are actors who are willing to assume the costs of organizing. This insight derives from another (and earlier) branch of rational-choice theorizing, that which pioneered the analysis of public goods. As Norman Frohlich, Joe Oppenheimer, and Oran Young note:

Except in the unusual case of the single individual who supplies a collective good, it is generally agreed that some sort of organization is required to collect resources and to supply the good in question. Yet discussions of collective goods seldom pay much attention to the process through which such an organization can or will come into existence in a social structure.^[22]

Political leadership is the key to organizing cooperation.

Frohlich and his collaborators argue that organizing to provide collective goods entails certain costs. In other words, there are costs beyond those of providing the collective good itself, and those costs have to do with mobilizing participants. The cost of providing a collection organization can be assumed by a political leader or political entrepreneur. The political leader does not pay the costs of organizing out of altruism; she expects to derive some sort of surplus or net gain. The surplus could be psychic (like enhanced prestige)

[21] Robert Axelrod and Robert O. Keohane, "Achieving Cooperation under Anarchy," 227.

[22] Norman Frohlich, Joe A. Oppenheimer, and Oran R. Young, Political Leadership and Collective Goods , 6.

or pecuniary (material profits). The notions of political leadership and the costs of organizing have been developed with reference to domestic politics and policy processes but have not been exploited in the study of international politics.

To be sure, students of international politics have constantly analyzed international political leadership. But we have not always been cognizant of the costs of organizing cooperation and by whom those costs can be assumed. The predominant mode of thinking has been that single powerful states play the crucial role of organizing cooperation. In fact, this argument goes back as far as Mancur Olson, one of the pioneers in the analysis of collective action. Olson argued that public goods could be supplied when one powerful actor either valued the good more than the cost of providing it (and so supplied it for all) or could make private side payments in order to enlist contributions.^[23] Kindleberger made the case that a liberal international economic order depended on the existence of a hegemon to lead it, such as Britain in the nineteenth century and the United States in the mid-twentieth.^[24] Indeed, the notion of single-power leadership became the heart of the hegemonic-stability theory of international regimes.^[25] Thus the economic preeminence of the United States was seen as explaining the creation and stability (into the 1970s) of the Bretton Woods system. More recently, scholars have shown that, in principle, a small number of states can jointly provide international leadership.^[26]

The study of international cooperation requires a more generally useful notion of political leadership than that provided by hegemonic-stability theory. I propose revising the concept of international political leadership in two ways: first, by decoupling it from public goods so as to apply it to private goods as well; and, second, by including international organizations among the potential sources of political leadership. Before proceeding with those arguments, however, I will specify the functions of international political leadership

[23] Mancur Olson, Jr., The Logic of Collective Action .

[24] Kindleberger, World in Depression .

[25] Key early statements were Robert Gilpin, U.S. Power and the Multinational Corporation; and Stephen D. Krasner, "State Power and the Structure of International Trade."

[26] See, for example, Keohane, After Hegemony; Snidal, "Limits of Hegemonic Stability Theory"; David A. Lake, "Beneath the Commerce of Nations: A Theory of International Economic Structures."

in the initiation of cooperation. International political leaders perform some or all of the following functions:

All these functions require investments of time, personnel, energy, and financial resources. All relate to the earliest stage of initiating collaboration.^[27]

One would derive a different notion of leadership from the traditional, realist-inspired literature on international politics. To be sure, scholarship on international politics does not generally employ explicit concepts of leadership. Rather, initiative in world politics is presumed to be the prerogative of the powerful. In effect, leadership means the ability to coerce or compel other actors to go along with what the "leader" desires. Leadership is implicitly linked to power, whether power means control of resources that others need or the ability to obtain desired outcomes in the face of opposition.^[28] Keohane and Nye presumably had in mind this notion of leadership when they wrote that "leadership will not come from international organizations, nor will effective power."^[29]

There are, however, other notions of leadership. Leadership need not imply the use of power to coerce or compel. At one extreme, in fact, leadership can rest wholly on the charisma and personal appeal of an individual—entirely without coercion or side payments.^[30]

[27] The initiation of collaboration is the focus of this study. Other important functions and variables may be involved in maintaining or managing cooperation once it is underway, but they require separate development.

[28] See David Baldwin, "Interdependence and Power: A Conceptual Analysis."

[29] Robert O. Keohane and Joseph S. Nye, Jr., Power and Interdependence , 240.

Students of organizations have long recognized different kinds of authority and leadership. For instance, Amitai Etzioni proposed that within organizations there could be different "means of control": coercive, utilitarian, and social.^[31] Etzioni's definitions may not be directly applicable to relations among states, but they certainly suggest that organizational interactions, even at the international level, might involve forms of leadership other than those based on power and coercion.

Keohane and Nye, though they reject international organizations as sources of power or leadership, suggest that they can exercise initiative. They argue that "international organizations may therefore help to activate 'potential coalitions' in world politics, by facilitating communication between certain elites; secretariats of organizations may speed up this process through their own coalition-building activities."^[32] I suggest that this kind of activity on the part of international organizations constitutes international leadership. It is leadership defined by the functions outlined above: persuading, setting agendas, mobilizing coalitions, promoting consensus, and pushing compromises. Such noncoercive leadership can, I will argue, be effective in initiating international cooperation.

As we have seen, the theoretical literature dealing with leadership and collective action grew around the problem of public goods. Pure public goods satisfy two criteria: jointness of consumption, meaning that consumption of a unit of the good by one actor does not reduce its availability to others; and impossibility of exclusion, meaning that once the good is provided, consumption cannot be limited to those who contributed to its provision. The public-goods theorists showed that these characteristics mean that public goods will always be underprovided, if they are provided at all. In this theory, political leadership is the means by which public goods can be supplied and the costs can be shared.

The analytical shortcoming with this approach is that pure public goods are rare in the real world. Indeed, many instances of international cooperation probably involve the provision of goods that cannot be jointly consumed. Furthermore, most international collective action deals with goods for which exclusion is feasible and

[30] See Max Weber, From Max Weber: Essays in Sociology .

[31] Amitai Etzioni, Modern Organizations , chap. 6.

[32] Keohane and Nye, Power and Interdependence , 240.

routinely practiced. The free-trade regime centered on the General Agreement on Tariffs and Trade (GATT)—presumably one of the most nearly "public" of international goods—excludes scores of nations.

I propose that costs of organizing exist even for private goods that require cooperation for their provision. In other words, even for goods that are not joint and for which exclusion is feasible, there is still a problem of collective action whenever a single actor cannot provide the good for herself. This point has also been made by Russell Hardin, but it is frequently overlooked.^[33] The upshot is that the analytical framework I propose in this study, centered on cognitive adaptation and political leadership, should be useful for any kind of international collective action, for public or private goods. A latent demand for cooperation will not become collective action until a political leader assumes the cost of organizing cooperation. And the costs of organizing exist for private as well as public goods: Someone must propose, persuade, mobilize, prepare an agenda, and broker compromises.

For the cases of European technological collaboration, this point becomes important. In principle, technological knowledge is joint—my possession of it does not diminish the amount of it available for you. However, in practice, the more possessors there are of technological knowledge, the more competitors there are likely to be in the relevant commercial market. Market shares are not joint or are at the least extremely prone to crowding. Thus technological knowledge is one short step away from being nonjoint. Companies tend to think of technical information as nonjoint and hence seek to protect it. Patents and copyrights exist for precisely that purpose. Thus the broad notion of political leadership I propose is crucial for the empirical cases analyzed here and for all other instances of collective action for nonpublic goods.

The analytical reach of the notion of international political leadership can be expanded in another way. The potential sources of leadership in the international arena are broader than is usually recognized. International leadership can come from single powerful states or small groups of states. But international organizations (IOs) can, under certain circumstances, also exercise the leadership functions

[33] Russell Hardin, Collective Action .

outlined above. This flatly contradicts the realist assertion that "international institutions are unable to mitigate anarchy's constraining effects on inter-state cooperation."^[34]

My argument also contradicts the neorealist hypothesis advanced by Andrew Moravscik, who argues that EC institutions had no impact on the definition of interests or the political bargaining that produced the Single European Act (SEA). Moravscik contends that the SEA was the product solely of bargaining among France, Germany, and the United Kingdom. Unfortunately, this hypothesis is at once obvious and misleading. Ultimately, all important decisions in the Community involve bargains among EC states, and the large states have a greater say in those bargains than do the small states. But Community institutions can shape the agenda, promote the redefinition of interests, and take the lead in proposing solutions to common problems. In the case of the SEA, the European Parliament started the process. The Parliament's Draft Treaty placed European union on the agenda for national governments; national parliaments were examining and voting on it. Thus Mitterrand's campaign for renewal of the Community did not emerge out of a void; the French president attached his prestige and his ambitions to the Parliament's proposals. Moravscik states that the Commission's White Paper "was a response to a mandate from the member states"; but, in fact, the member states were only responding to the initiative of the European Parliament.^[35]

One important line of theorizing is that international institutions in general (including regimes as well as specific IOs) can sustain cooperation by reducing the costs of transactions and information. Reduction of transaction costs lays a basis for reciprocity and stable expectations about behaviors.^[36] Oran Young argues that IOs can be a source of innovative policy ideas and can act at times almost like pressure groups.^[37] My proposition is similar but goes still farther. Under some conditions IOs can exercise political leadership and influence the adaptive behaviors of states. Reducing transaction and information costs are relatively passive functions; IOs can also actively lead.

[34] Grieco, "Anarchy and the Limits of Cooperation," 485.

[35] Andrew Moravscik, "Negotiating the Single Act: National Interests and Conventional Statecraft in the European Community."

[36] A clear statement of this argument is in Keohane, After Hegemony .

[37] Oran Young, International Cooperation: Building Regimes for Natural Resources and the Environment , 54.

What are the conditions under which international leadership can come from IOs? First, the greater the initial grant of decision-making authority to the IO, the greater its ability to lead in new areas. IOs with limited mandates and specific issues will not have much latitude for leadership; the International Civil Aviation Organization is an example of this type. The Commission of the European Communities is one of the IOs best endowed with some autonomous capacities (for example, competition policy) and broad powers to propose.

Second, when representatives of an IO have substantive knowledge and relevant information, they can help shape the technical discussions and agreements. One of the most dramatic examples of this effect is the economic analysis generated by Raul Prebisch and his fellow economists at the United Nations Economic Commission for Latin America (ECLA). ECLA thinking was incorporated into the philosophical foundations of Latin American regional organizations, as well as the United Nations Conference on Trade and Development. In the case of telematics in Europe, the Commission was armed with technical studies and data that allowed Commission representatives to structure and contribute to the technical discussions leading to ESPRIT and RACE. The Commission acted as an informed participant in the discussions, influencing the direction of discussions and the consensus that emerged.

Third, IOs can exercise initiative, proposing solutions to common problems and rallying support among member governments. Their ability to perform these functions depends in part on the personal characteristics of IO leaders. As Robert Cox and Harold Jacobson suggest, IO leaders possess different sources of influence, some inhering in their position (legitimacy, authority) and some deriving from personal characteristics. Personal attributes that enhance the influence of an IO official include charisma, expert knowledge, negotiating ability, personal achievement outside the IO, and administrative competence. Thus, "high international officials may command information and recognition, which allows them the initiative in proposing action or resolving conflict."^[38]

Fourth, and most important, I submit that international organizations will register the greatest impact on interstate cooperation during periods of crisis (as defined previously in this chapter). When

[38] Robert W. Cox and Harold K. Jacobson, The Anatomy of Influence , 20.

national leaders confront policy failures that compel them to rethink their objectives or the means chosen to pursue them (or both), they are in the adaptive mode—searching for alternative approaches. IO actors can seize the initiative under such circumstances to supply new models and strategies. In addition, crises provide opportunities for activist, entrepreneurial IO leaders to marshal states behind a cooperative solution. The activist IO leader not only can sponsor transnational contacts and consensus building but can sometimes appeal directly to heads of state. In other words, the efforts of IO officials will count for more when member countries are simultaneously in an adaptive mode. Ernst Haas provides examples of this effect in the United Nations.^[39] In this study Commissioner Davignon plays the role of the entrepreneurial IO leader.

Under these four conditions—broad initial grant of authority, technical preparedness, activist IO officials, and policy-making crisis in member countries—IOs enjoy the broadest possible opportunity to exercise international political leadership. Political leadership, as I have argued, depends on the leader's receiving some form of net benefit. For IOs the gains take the form of enlarged mandates, increased prestige, perhaps even expanded budgets and staffing. With regard to paying the costs of organizing new cooperations, IOs possess certain advantages over states. What may be a cost to states trying to organize cooperation—in time, resources, energy—is a matter of course to IOs. IOs exist to foster cooperation. In other words, what may be a cost to a national government is a part of the mission of IOs.

The Problem of Juste Retour

The third element in this theoretical framework bears particularly on instances of collective action for private goods. I take it as given that states will seek assurances of a fair return on their investment in the cooperation. The general theoretical point, enunciated by James March and Herbert Simon with respect to all organizations, is that each member expects a balance between contributions and rewards, between what it puts into and what it receives from the organization.^[40] I propose that governments in organized cooperation expect

[39] Haas, When Knowledge Is Power .

[40] James G. March and Herbert Simon, Organizations . I am grateful to Judith Haymore Sandholtz for bringing this line of theorizing to my attention.

the same. Governments are unlikely to agree to collaborate unless they have some confidence in obtaining a satisfactory ratio of contributions to benefits. What constitutes a satisfactory ratio is entirely subjective. Furthermore, the juste retour problem is not one with a unique solution.

To begin abstractly, two extremes bound a continuum of solutions. Lindblom suggests that problems of resource allocation can have market and political solutions.^[41] Though Lindblom writes of national economies, a similar logic is relevant here. At the market extreme, individual choices interact to produce a distribution of resources; at the opposite extreme, a central authority decrees the distribution.^[42] In technological collaboration, the market approach would be the pure `a la carte system, in which countries bid to participate only in those parts of a program in which they are interested. At the political end of the continuum, a central authority would assign agreed-upon shares.

In practice, both forms of allocation exist. Pure à la carte participation is the basis of the EC's Coopération Scientifique et Technologique (COST) program and of the EUREKA program: States participate only in projects of their choosing. Indeed, the programs are really umbrellas for intergovernmental R&D agreements. In contrast, shares of work in the Airbus consortium are set by contract. Similarly, the members of the European Space Agency (ESA) agreed in January 1985 to a hybrid approach: Each country should receive industrial contracts worth 95 percent of its financial contribution, yet countries participate in only those major programs that they choose to.^[43]

I hypothesize a straightforward connection between the method of handling juste retour and the mission of the cooperative effort. Collaboration for a single, discrete aim will allocate shares authoritatively because later adjustments are likely to be difficult. Collaboration for a large and differentiated set of objectives will be likely to distribute shares by self-selection. If a collaborative effort is ongoing (iterated) and composed of myriad bits and pieces, states can pick and choose which ones they want to participate in and still be confident, over the long run, of achieving a fair share. A one-shot project cannot offer that assurance; the initial shares will also be

[41] See Lindblom, Politics and Markets .

[42] See ibid.

[43] I examine ESA at length in Chapter 5.

the final shares, barring a painful renegotiation. Put differently, the more pies, the easier it is to see that everyone gets enough slices of the right sizes.

In technological collaboration single-purpose projects will be likely to display authoritative allocation of shares. This is certainly the case with Airbus and with cross-national jet-engine collaboration. Each participant receives a set share, negotiated at the outset. In contrast, programs that aim to advance technological capacities across a broad front (rather than to develop a specific item) will display `a la carte participation. This is the case in programs like ESPRIT and EUREKA.