© A. Hasenclever, P. Mayer, V. Rittberger, Tübingen 1996
ISBN 3-927604-18-6
Redaktion: Thomas Nielebock, Uta Reichardt, Frank Schimmelfennig
References
Appendix I: Cases Selected for In-Depth Study and Comparison
Appendix II: A Note on the Projected Evaluation Procedure
States create international regimes by agreeing upon a set of principles, norms, and rules which they undertake to respect when deciding on their policies in a particular issue area. In addition, they usually lay down procedures for the further development and adaptation to changed circumstances of the substantive norms and rules of the regime (Krasner 1983b: 2). When states form (or join) a regime they accept certain constraints on the exercise of their sovereign rights (e.g. their right to raise or lower tariffs at will) in exchange for others doing the same. Typically, these exchanges, in which international regimes are rooted, generate both benefits and burdens for the states that enter into them (and often for outsiders as well).
Adherents to the dominant school of thought in regime analysis, which we will here refer to as "rationalistic institutionalism", argue that states create, join, and maintain regimes in order to further their self-interests. This proposition suggests that at least in effective regimes (i.e. institutional arrangements which enjoy their members' support in word and deed) benefits will in general outweigh burdens both collectively and individually.[Note 2] Even assuming net benefits for all actors, however, leaves open the question of how the aggregated benefits accruing under a regime are allocated among its members . To denote this particular aspect of international regimes we will henceforth use the term "distributional consequences (of a regime)".
At the most general level, distributional consequences are a function of the specific content of the principles, norms, and rules of the regime, on the one hand, and of the distribution of resources among the member states, on the other.[Note 3] (So, e.g., under a non-discriminatory free trade regime states with highly competitive economies will generally do better than economically less developed ones.) Distributional consequences of regimes can be described and classified under various aspects. For example, a given regime may be said to allot gains and costs (roughly) symmetrically (or asymmetrically) among its members.[Note 4] Often, however, distributional consequences of institutions are not just described, but also evaluated according to some normative criterion such as efficiency or equity (Young 1982: ch. 5, 1992: 164).
The distributional consequences of international regimes (as well as the factors that determine their variation) are a topic which regime analysis has not yet systematically explored, although this lacuna has not gone unnoticed among students of regimes (see, e.g., Snidal 1985b: 612-3, 1986: 46). In this paper we report on an ongoing research project which tries to shed some light on this understudied aspect of international regimes.[Note 5] Our concern in this project is not with the distributional consequences of regimes for their own sake, however. Rather, we wish to study them within a broader theoretical and empirical context, asking whether, and if so how, the way a regime allocates benefits and burdens to its members affects its robustness , i.e. its ability to withstand exogenous challenges without its effectiveness being diminished.[Note 6] (A prime example for an exogenous challenge in this sense is a sharp deterioration of the political relationship between some of the regime's members for reasons which are unrelated to the regime.)
Earlier studies have found evidence that even regimes whose membership is composed of states (or alliances) which are involved in intense political and ideological competition (such as the former Eastern and Western blocs) can prove surprisingly robust (Rittberger 1990, Rittberger/Zürn 1991, see also Nye 1987: 400-1). This finding suggests that international regimes may serve as building blocs of a "stable peace" (Boulding) among nations. Much as this result is welcome, from a broader normative point of view, it immediately provokes the question of whether this peace comes at a price and, if so, how high this price is. Do international stability and order - grounded in and fostered by robust international regimes - have to be bought at the expense of justice? Many authors in the "classical (realist) tradition" (Holsti) of International Relations seem to have thought so.[Note 7] On the other hand, there is also the ancient idea (to be found as early as in the Prophets of the Old Testament) that no peace is stable unless it is founded on just social relations. By treating regime robustness as a variable rather than a constant and by asking, inter alia , what difference (if any) a just (or nearly just) distribution of benefits and burdens makes for a regime's robustness, we seek to make an empirically-grounded contribution to this long-standing controversy.
First and foremost, however, we believe that our research may contribute to a more recent and more focused debate within the study of international institutions, which involves rationalistic institutionalists such as Robert Keohane and Joseph Grieco, on the one side,[Note 8] and the proponents of a reflective approach to understanding international regimes, on the other. Distinct hypotheses regarding the relationship of the distributional consequences and the robustness of regimes can be derived from, or at least plausibly attributed to, these two schools of thought in institutional theory. A confrontation of these hypotheses with a set of carefully chosen cases can therefore be expected to shed some light on the relative merits of these rivalling perspectives on international politics.
Our purpose in this paper is not to report however preliminary results of our project. Empirical research is under way, but, at this stage, no findings could be presented with a reasonable degree of confidence. What we do wish to do, however, is to report the measures we have taken in order to maximize the likelihood of arriving at such findings - hence the subtitle of this paper: "A Research Design". We proceed in the following steps: In the next section (2.) we explore the concept of regime robustness, i.e. our dependent variable. Subsequently (sec. 3), we introduce the hypotheses which will guide our research and place each of them into a broader theoretical context. This is followed by a section (4.) in which we address questions of method. Against the backdrop of these considerations, the subsequent section (5.) presents our control variables. Finally (sec. 6), we discuss some problems arising in connection with the selection of cases for in-depth study and the strategies we have fallen back on to solve them.
Our project is intended as an inquiry into the robustness of international regimes. Our focus is not on regime robustness as a generic phenomenon, i.e. we do not ask what general features of regimes (such as sunk costs or bureaucratic inertia) account for the fact that regimes often turn out surprisingly robust (Krasner 1983a: 360- 1).[Note 9] Rather we consciously treat regime robustness as a variable, i.e. as a property which different regimes may possess in different degrees, and ask which factors might explain these differences. In doing so, we concentrate on distributional consequences as a potential source of variation.
By robustness we understand the capacity of a regime to withstand exogenous challenges (or "shocks") without its effectiveness decreasing.[Note 10] A notable implication of this definition is that regime robustness is a dispositional concept, i.e. it refers to a property that shows - or fails to show - only under specific conditions. The one condition that must be met in order for the question to be answerable if and to what extent a given regime is robust is the occurrence of an exogenous challenge to the regime. As a result, in our research, we have to ignore regimes which have not yet confronted such a shock. Moreover, challenges will vary in strength, which suggests that we compare regimes which have undergone similar (or even identical) challenges. Given two such regimes, the one whose effectiveness is least affected by the shock will have proven the more robust.[Note 11]
Two terms in the above definition call for explanation: "effectiveness" and "exogenous challenge". By the effectiveness of a regime we understand the measure of verbal and, most of all, practical support the regime receives from its members. In general, it is assumed, this support will be adequately expressed by the extent to which the members abide by the (comparatively precise) rules of the regime.[Note 12] The concept of regime robustness does not imply constancy of rules. Thus, a change of rules in the wake of some exogenous shock as such does not indicate low robustness. Any conclusion to this effect would be premature as long as we have not inspected the level of compliance with the modified set of rules.
The concept of regime effectiveness poses various problems of measurement and, more specifically, commensurability.[Note 13] For example, some rules, by their very nature, give actors more opportunities for violating them than others. Thus, a rule that requires its addressees to take a specific measure within a certain time-frame (e.g. scrap 50% of your tanks of type X by year Y) can be broken only once by any one actor. In contrast, other rules, because they regulate a more or less continuous activity (e.g. announce in advance military manoeuvers which involve more than Y thousand troops), can be violated much more frequently. Problems of this sort can be mitigated or circumvented by restricting direct comparison to regimes in which the same type of rule prevails. Furthermore, in assessing the level of rule compliance it seems necessary somehow to take into account the unequal importance of rules for the cooperative venture (Young 1989a: 71, n. 52). In practice, however, distinguishing between central and peripheral rules as well as weighing these differences in a reasonable way may often require difficult decisions.[Note 14]
The other concept in need of clarifying is that of an exogenous challenge (or shock).[Note 15] It is this concept that gives this project a profile of its own, as far as the dependent variable is concerned.[Note 16] An exogenous challenge is an event in the environment of the regime which tends to reduce the willingness (or ability) of states to cooperate in the issue-area in question. The paradigm case for East-West regimes before 1989 was the Second Cold War at the beginning of the 1980s (Rittberger 1990). But how do we know in general whether a given event is to be regarded as an exogenous shock to the regime under study? (After all, not all shocks cause a breakdown of the regime. If they did our project would be pointless.) A conclusive answer to this question could only be provided by a well corroborated theory of the dynamics (change and stability) of international regimes. No such theory exists to date. What is more, our project is an attempt to contribute to the development of such a theory, which gives it a tinge of circularity. The circle is not vicious, though: there is no reason to assume that the desired theory cannot be built up in a piecemeal fashion ("theory" meaning no more in this context than a catalogue of determinants).[Note 17]
At this stage we could have resorted to "general experience" which suggests various rules of thumb linking certain events to pressures on, or tensions within, regimes.[Note 18] We could have screened the empirical literature on regimes as well as our minds in order to compile a list of such rules which, in turn, could have helped us to identify exogenous challenges in (other) concrete cases. The trouble with such an inductive approach, however, is that it is unsystematic and (most likely) could not have provided a clue as to how the various rules of thumb (and, correspondingly, the shocks these rules identify) are related to one another. It would have been difficult to tell how many (essentially) different kinds of shocks there are, what their relative weights are or how they may derange or disrupt an ongoing cooperation. For this reason, we have opted for a more theoretically-informed approach to concretizing the idea of an exogenous challenge. In the following, we sketch the simple "theory of exogenous shocks" we are going to base our empirical analyses on. This theory goes at least some way to ameliorate the problems of a purely inductive or common-sense approach to identifying such shocks. As far as this theory is concerned, we do not claim a great deal of originality. In fact, it is firmly rooted in the interest-based theory of international cooperation that has emerged from the work of scholars such as Robert Jervis (1978), Arthur Stein (1983, 1990), Charles Lipson (1984), Duncan Snidal (1985a), Kenneth Oye (1986), and Michael Zürn (1992).[Note 19]
Central to this literature which draws heavily on game theory is the distinction between different types of payoff structures (or constellations of interests). In particular, coordination and Prisoners' Dilemma (PD) situations have often been contrasted. A crucial difference between these two kinds of situation is that only in coordination situations are the Pareto-efficient cooperative outcomes Nash equilibria. As a result, coordination regimes are inherently stable, while collaboration regimes (i.e. regimes which are responses to PD situations)[Note 20] are not. Only in collaboration regimes do actors have an incentive unilaterally to depart from cooperation (or to reduce their level of cooperation).[Note 21] Another fact about these two games is of similar importance to us. As Snidal (1985a: 939-40) has observed, whether or not a given event has a potentially destabilizing effect on cooperation is not independent of the type of situation at hand (coordination or PD). Consequently, a comprehensive theory of exogenous shocks to regimes would have to be a conditional one, including special propositions for coordination, PD, and other situations, respectively. In this project we have decided to focus on collaboration regimes and, consequently, to develop only that part of the theory which is relevant to this type of regime. Two reasons account for this: First, collaboration regimes are widespread in international relations and are comparatively well understood.[Note 22] Second and even more important, stability being not, as it were, an in-built feature of such regimes, collaboration regimes appear to be the more promising object of study for a project that looks into the sources of variation in regime robustness.[Note 23]
The (partial) theory of exogenous shocks to regimes which we intend to premise our empirical research upon rests on the simple principle that a given event counts as a challenge if it tends to undermine the willingness (or ability) of actors to cooperate in PD situations. The challenge is an exogenous one if it emanates from the environment of the regime, i.e. if it is neither logically nor causally related to the regime.[Note 24] Finally, the decision as to what types of events do tend to have this destabilizing effect is made in light of (interest- based) cooperation theory.
(1) The existence of at least some incentive for actors to cheat their partners is characteristic of PD situations. Consequently, there is always some concern about being double-crossed. To the extent that this concern becomes acute, even those actors may no longer be willing to cooperate (at the previous level of intensity) who, so far, have been prepared to forego (short-term) maximization of gains for the sake of successful cooperation - whether it is because they are practicing a strategy of reciprocity or simply because they want to avoid the sucker's payoff. For this reason the presence or absence of mutual trust among the partners can be an important determinant for the success and the stability of cooperation (Jönsson 1993: 205-6). Consequently, any event in the environment of the regime that can be expected to damage or destroy this important common resource must be regarded as an exogenous challenge. Thus, it is plausible to interpret the Second Cold War or rather the events that have triggered it (such as the Soviet Union's intervention in Afghanistan) as exogenous shocks to the then-existing East-West regimes.
(2) According to a central theorem of cooperation theory, cooperation among non-altruistic actors, in a large measure, depends on the existence of a sufficiently long "shadow of the future" (Axelrod 1984). This metaphor stands for what economists call discount rate, i.e. the rate by which actors devalue future gains relative to present ones. In a regime context this means that strategies that are designed to induce and maintain cooperation in a decentralized setting such as the international system may fail to take effect, because some actors are preoccupied with short-term advantages (i.e. they are acting under a short shadow of the future). The expectations held by actors concerning the length of the (super-)game have an analoguous significance. Their willingness to cooperate in any given round of play will be the greater, the more they are convinced that the (sequence of) game(s) will continue beyond all foreseeable future. As a result, events outside the regime that can be expected to drive up the discount rates of players or to nourish doubts as to the (virtual) infinity of the game must be regarded as exogenous shocks to that regime. Thus, a sharp economic recession can often be assumed to pose a challenge to international regimes in economic and environmental issue-areas, as in such situations pressures on decisionmakers mount to come up with fast solutions (Lipson 1984: 21-2). Whereas here the challenge is one of a shrunk shadow of the future, shocks of the second kind (i.e. events that lead actors to believe that the game they are involved in is close to an end) now and again occur to regimes which regulate the use of international common property resources (e.g., when as a result of environmental change or previous overfishing the stocks in a particular fishery decline dramatically).[Note 25]
(3) Once more following Axelrod's analysis of the "evolution of cooperation", many students of regimes argue that the application of a strategy of reciprocity is central to the initiation and maintenance of regime-based cooperation (Keohane 1984, Axelrod/Keohane 1986, Zürn 1992: 204-5). Keohane (1989e) has distinguished two variants of reciprocity - specific and diffuse reciprocity - and has shown that neither of the variants can induce cooperation reliably, but a combination of both is often most successful in practice. This means that cooperation on the basis of reciprocity is far from easy (in particular where the number of actors is quite large) and will often have succeeded only after actors have undergone a lengthy and perhaps painful learning process. As a result, the emergence of new powerful actors which could not yet have been socialized into the complicated rules of the cooperation game will often disturb or even threaten an ongoing cooperative practice. Examples can be found in oligopolistic markets, when new suppliers are unwilling to play by the established rules of the game. This type of exogenous shock is also relevant to international politics, however. Thus, when highly efficient Japanese suppliers appeared at the international market for automobiles, the international regime for the regulation of trade in cars which was run by a couple of Western industrialized countries faced a serious challenge (Dunn 1987). A case in point in the field of international security is the disintegration of the Soviet Union at the beginning of this decade which left four new nuclear powers (including Russia) to the nonproliferation regime (Schimmelfennig 1994).
(4) Finally, we should not fail to mention a background condition of cooperation in PD situations which has played a very prominent role in the development of regime theory: the distribution of power . Drawing on the theory of collective goods (Olson 1965) the so-called theory of hegemonic stability states that the effectiveness of regimes vanishes if the power gap between the hegemon (under whose leadership the regime has been created) and the other members of the regime closes (Kindleberger 1976, Keohane 1980). On this account, hegemonic decline, particularly when it occurs rapidly and is accompanied by visible signs of weakness, poses a prima facie exogenous challenge to those regimes which have largely arisen as a political investment of the (previously) dominant state.[Note 26]
Table 1: A Taxonomy of Exogenous Challenges to Regimes
| Category of Shock | Example |
| Breakdown of Mutual Trust | Events that led to the Second World War |
|
Shrinking
Shadow of the Future a. Rapidly increasing discount rate b. Growing doubts as to the continuation of the game |
a. Sharp
recession b. Rapidly declining fishery |
| Emergence of New Powerful Actors not yet Socialized | Disintegration of the Soviet Union |
| Hegemonic Decline | U.S. in the 1960s and 1970s |
In our project we examine two hypotheses regarding the causal relationship between the distibutional consequences and the robustness of international regimes. One, which we refer to as the symmetry hypothesis , is derived from rationalistic institutionalism and states that ceteris paribus regimes which are characterized by a roughly equal (even, balanced) distribution of benefits and burdens among their members are more robust. The other, which we have labeled the equity hypothesis , is compatible with reflective institutionalism and states that regimes whose distributional consequences meet the normative standard of equity (fairness, justice)[Note 27] tend to exhibit a higher degree of robustness.
Foundations of Rationalistic Institutionalism
When, in the early 80s, the study of international regimes emerged as a new research program within the discipline of International Relations (IR), much of the most innovative theoretical work was characterized by a mixed attitude towards (neo-)realism, which had dominated the field since the end of the Second World War. This is particularly evident in Robert Keohane's (1983, 1984, 1989b) contribution. Far from arguing for a wholesale rejection of realism, Keohane deliberately adopted realist assumptions about the nature of international actors and their environment. At the same time, however, he tried to show that the traditional realist skepticism about the ability of international institutions to further cooperation among states is not warranted. Thus, Keohane acknowledged that states are still the most important actors in world politics, and that anarchy as well as the distribution of power among them are highly significant features of their (external) social environment. Following neorealist Kenneth Waltz's (1979) lead, he opted for a systemic approach to theory-building, and, last but not least, he shared realism's commitment to rationalism as guiding meta-theory.
Rationalism in international relations theory attempts to explain international politics on the basis of the assumption that states behave as utility-maximizers. Foreign policy decisions and international institutions are to be reconstructed as the results of individualistic calculations of advantage made by governments. Rationalists usually assume that the utility functions that states maximize do not attach any (positive) value to the welfare of others: in other words, altruism is never a motivating force for states. Rationalists postulate that the behavior of states reflects strategic rationality (taking into account the rationality of their opponents), but not moral considerations. As Keohane and others (Stein 1983, Axelrod and Keohane 1986) have shown, this postulate does not imply that international rules can have no bearing on foreign policy decisions. It is assumed, however, that states do not respect rules because they feel morally obliged to do so. Rather, they comply with international norms and rules when, e.g., the short-term gains of failing to do so would be outweighed by losses in the long run.
Two Strands of Rationalistic Institutionalism: Neoliberal Institutionalism and Modern Realism
In recent years, it has been a matter of intense dispute whether Keohane's agenda has been successful (Baldwin 1993). Realist critics, most notably Joseph Grieco (1988, 1990), have argued that Keohane and the other "neoliberals" have not fully appreciated the implications of international anarchy for states and that, consequently, their attempt to challenge realism as the dominant perspective in IR theory has failed. Neoliberals, in turn, have sought to refute Grieco's central argument regarding the significance of relative-gains concerns for international cooperation (Snidal 1991a, 1991b). The ensuing debate has been highly stimulating and has deservedly attracted a great deal of scholarly interest. At the same time, the very vigor with which it has been conducted may have veiled notable points of convergence.
Two such points of convergence are especially important to us. Together they may, to some extent, justify our perhaps somewhat premature use of the term "rationalistic institutionalism" to refer to both the realist and the neoliberal ideas about international cooperation, downplaying the differences that continue to exist between them. (1) Even pronounced realists do no longer dispute the possibility, apart from military alliances, of sustained rule-based cooperation among states (i.e. the possibility of effective international regimes in our parlance). What is more, they concede that international institutions can have a significant part in the process of launching and maintaining such cooperation, and they now even call for a realist theory of international institutions (Grieco 1990: 22-3, 233-4, 1993: 335). (2) The propositions that realists and neoliberals put forward with respect to the relationship between the distributional consequences and the robustness of international regimes are largely congruent, even though the reasons they give are quite different. More specifically, they both subscribe to what we have called the symmetry hypothesis.
The Realist Case for the Symmetry Hypothesis
The realist argument runs as follows: The international behavior of states is crucially shaped by the anarchical structure of the international system. For realists anarchy first and foremost means that no central institution guarantees the survival and/or independence of the units. This is why the international system is one of self-help (Waltz 1979). Ultimately, each state depends for its safety and its survival on its own (especially military) capabilities. These capabilities, however, cannot serve their purpose independently of the capabilities that are controlled by other states: it is relative , not absolute power that counts in the international system. As a result, each state must take an interest not only in the dynamics of its own capabilities, but in those of its neighbors as well (where the spatial dimensions of international neighborhood, to a large extent, are capability-dependent themselves).
These basic facts of international life have important consequences for states' ability to cooperate. States engage in regime-based cooperation in order to realize mutual gains. Such gains, however, have repercussions on the (relatively valued) capabilities of the cooperating parties (as well as on those of outsiders). As a result, states which have to decide whether or not to (continue to) cooperate with others on some issue[Note 28] are concerned not only with absolute gains, but with relative ones as well. More specifically, according to Grieco, they are sensitive to relative losses, i.e. gaps in gains in favor of their partners. Although this sensitivity is not a constant, but is contingent on certain situational attributes,[Note 29] concerns about relative losses never completely disappear and may prevent states from embarking on, or pursuing, cooperation with others even where this means forgoing substantial (absolute) gains. Consequently, the symmetry hypothesis applies: Since no relative losses accrue in regimes which allocate benefits and costs evenly among its members, such regimes are ceteris paribus more robust.[Note 30]
The Neoliberal Case for the Symmetry Hypothesis
Neoliberals[Note 31] have not arrived at similarly clear propositions regarding the distributional aspects of regimes and their connection with regime robustness. Nor have they developed a comparable deductive argument. Nevertheless, their (especially Keohane's) pertinent considerations and suggestions add up to a very similar picture (see also Milner 1992: 470-2). Neoliberals argue that cooperation among self-seeking and morally indifferent states crucially depends on the application of reciprocal strategies such as tit-for-tat (Axelrod 1984, Axelrod and Keohane 1986). According to Keohane (1989e: 135), reciprocity has two essential attributes: contingency and equivalence of the (not necessarily material) "goods" exchanged. As we have seen, regimes are (based upon) a certain kind of exchange as well: at the most abstract level, the goods that are exchanged in a regime are the (net) benefits that states make available to one another by means of their rule-governed cooperation. Regimes are not supposed to supplant the reciprocal strategies necessary for maintaining cooperation. Rather, one of their central functions is to help remove obstacles to the successful operation of such strategies at the international level (Oye 1986: 16-7, 20). We see now why the symmetry hypothesis follows: in view of the functional logic that informs neoliberal thinking on international regimes (Keohane 1984: 80-1, 1989c: 167) it must be concluded that regimes which do no longer allow for reciprocity (which implies, as we have seen, roughly equivalent gains for the participants) should be valued less by the states that have maintained them so far and, consequently, should be less robust.[Note 32]
A Rival Hypothesis: Power
It must be added that the rationalistic view on robustness and distribution in regimes is less homogenous than our discussion suggests. Whereas Grieco tends to treat power (and its conservation) exclusively as a goal - states' actions resulting from their basic motivation of "defensive positionalism" -, other realists have emphasized power as a means . For example, Stephen Krasner (1991) has stressed the role of power differences in the resolution of the distributional conflicts that regularly emerge in connection with international cooperation. Hence, we would have to expect the distribution of benefits from any given regime to reflect the distribution of (presumably, issue-specific) power among its members. Moreover, we might predict that regimes which, in their distributional consequences, do no longer accord with these external realities must either change or, if change cannot be organized, collapse. We might, in other words, suggest an alternative rationalistic hypothesis regarding the relationship of interest to us: Regimes tend to be more robust if external power asymmetries are "respected" in the allocation of benefits to its members.[Note 33]
Two considerations have led us to focus on the symmetry hypothesis (regarding it as representative of rationalistic institutionalism) rather than on this power hypothesis (as it may be called). (1) Due to the controversy between realists and neoliberals the symmetrical (or "balanced") gains hypothesis seems to enjoy special prominence in the current scholarly debate (even though our above analysis has suggested that, ironically, this is actually a point of convergence rather than divergence between these two strands of rationalistic institutionalism). (2) Part of Krasner's (1991: 360-6) argument for his "power-oriented research program" is the expectation that Battle of the Sexes (a coordination game) is more relevant for international politics than Prisoner's Dilemma, which has informed neoliberal thinking about international regimes. For reasons which have been explained in the preceding section (2.), however, in our empirical research, we are going to focus not on coordination, but on collaboration regimes.[Note 34]
The Symmetry and Equity Hypotheses as Competing Hypotheses
Inquiring into the connection (or non-connection) of regime robustness on the one hand and distributional consequences of regimes on the other, we do not focus on just one hypothesis. Rather, we intend to pursue a competitive approach. More specifically, we are contrasting the symmetry hypothesis with an alternative conjecture according to which it is not equality but equity in the distribution of benefits and burdens that enhances the robustness of regimes. Despite their seeming similarity, the symmetry and the equity hypotheses are in direct competition with one another. Since they rely on the same independent variable (distributional consequences), which they categorize differently, it is logically impossible that, with respect to the same universe of cases, both of the specified factors ('equality of the distribution' and 'justice of the distribution') have an impact. (To put it differently, at least one of the following statements must be false: (1) Regimes are ceteris paribus the more robust, the more symmetrical their distributional consequences. (2) Regimes are ceteris paribus the more robust, the more just their distributional consequences.)[Note 35]
Characteristics of Reflective Institutionalism
Unlike the symmetry hypothesis the equity hypothesis cannot be located in the prevailing rationalistic strand of regime analysis. It does, however, display at least a strong affinity to a set of alternative approaches to the study of international institutions which, for several years, have been discussed under the summary name of reflective approaches (Keohane 1989c: 170- 4). The least common denominator of these approaches is their critical posture vis-à-vis both the ontological commitments and the epistemological stance of rationalistic institutionalism. So reflective institutionalists reject both the (ontological) atomism and the (epistemological) positivism which they perceive as characteristic of mainstream regime analysis.[Note 36]
In reflective institutionalism the focus of research shifts from the analysis of rational choice behavior in view of given actor preferences to an analysis of the cognitive and social processes in which such preferences form and change. Definitions of self-interest and perceptions of behavioral options are no longer treated as stable and theoretically unproblematic "givens" of research and theory-building. Rather, they have to be accounted for as both product and subject (matter) of inter- and transnational interaction and communication. Reflective institutionalism, in other words, calls for the endogenization of the fundamental interests and identities of states and other actors on the international scene. Furthermore, reflectives point out that, by necessity, actors must interpret reality in order to act upon it and that, consequently, their actions are as much determined by their interpretations of the situation at hand as by the objective properties of that situation (if there is such a thing). For reflectives, this simple insight has important epistemological consequences: The meaning- contingency of action (as it might be called) suggests that the distanced outside perspective as is characteristic of positivism (or its ideal) is misguided. Whatever states do with respect to international institutions (create regimes, break rules, justify deviations, etc.), their actions are not mechanical responses to external stimuli, but deeply infused with 'sense'. Detecting the - not strictly private, but intersubjective - meanings that structure the choices and interactions of states no less than those of other social actors, however, requires a recourse to hermeneutic techniques of research (Kratochwil/Ruggie 1986).
The Equity Hypothesis as a Reflective Hypothesis
The equity hypothesis cannot just be logically derived from reflective institutionalism. (There is no such thing as the reflective theory of international institutions, after all.) Still, there are at least two strands in reflective theorizing that are open to this hypothesis.[Note 37]
(1) Reflective institutionalism acknowledges the significance of ideas and beliefs for the way that actors assess cause-effect relationships. Such ideas have been termed "causal beliefs". Causal beliefs suggest opportunities for action to states, but do not determine them. Part of the reason for this is the presence of yet another type of ideas that informs actors' perceptions of their options: beliefs which are not causal, but normative in nature (Goldstein/Keohane 1993). If those normative ideas (shared among states at a given time and place) were to include concerns about fairness we could expect governments to be hesitant to destroy or threaten regimes that embody the value of justice in their principles, norms, rules, and procedures, i.e. the equity hypothesis would apply.
(2) Having emerged well before the "rationalistic turn" in the study of international institutions in the early 1980s, the so-called English School certainly cannot be interpreted as a response to mainstream regime analysis.[Note 38] Nevertheless, it shares many concerns with the current wave of reflective scholarship in international relations theory (Wendt/Duvall 1989, Hurrell 1993, Buzan 1993). This is most obvious in the very notion that gives coherence and identity to this literature: the idea that states form an international society whose members deliberately uphold certain institutions and also share a (however limited) sense of community. This communitarian interpretation of the state system is in direct opposition to the individualistic, atomistic ontology that underlies rationalistic institutionalism. The international society school directs our attention to the fundamental, deeply embedded international norms and institutions. As constitutive rules, they serve as "transcendental" conditions for meaningful, purposive action by, and interaction between, states, i.e. they make such (inter)action possible in the first place. As regulative rules, they operate as meta-norms (or higher order norms) rendering international behavior orderly and consistent over the whole range of international issues. This suggests that the robustness of specific institutions such as international regimes to a large extent depends on their compatibility with these higher order norms.[Note 39] The institution of sovereignty (Krasner 1988) or the pacta sunt servanda principle (Chayes/Chayes 1993: 185) are among the most prominent examples for such fundamental "conventions" (Keohane) of international society. But regulative principles with a somewhat stronger moral content such as the norm of reciprocity have been pointed to in this connection as well (Keohane 1989d: 4).[Note 40] The informal character of many of these deeply engrained conventions and practices makes it difficult to prove their existence in special cases or to determine their precise content.[Note 41] We would argue, however, that an empirical corroboration of the equity hypothesis would suggest that certain norms of justice have to be counted among these basic norms of international society.[Note 42]
The Meaning of "Equity" Within a Reflective Framework
The preceding discussion, of course, raises the question of what we understand by equity in this connection. Both students of regimes and political theorists have failed so far to develop a normative theory of international institutions.[Note 43] There is, however, an extensive and sophisticated literature on the ethics of domestic political and social institutions which may be assumed to offer important clues for the development of international theory as well.[Note 44] But, as the complex discussion that John Rawls' modern classic "A Theory of Justice" has given rise to has once more demonstrated, consensus in this field seems even more difficult to achieve than in other disciplines dealing with human affairs. Moreover, international ethics (just as is positive international theory) is plagued with a level- of-analysis problem that general ethics can widely ignore. This problem centers upon the question of whether international justice can (or should) be conceived of in inter-state terms, or whether the individual is the only legitimate claimant to transboundary justice as well (Hoffman 1994).
Daunting as these problems may be, it may not be necessary to solve them in order to be able to address the questions that guide our project. Thus, we could have decided the question of what standard is appropriate to judge the normative standing of a given regime simply by fiat . For example, we could have settled upon an international version of Rawls's difference principle[Note 45] using it as criterion to range international regimes along the dimension of fairness. In normative theory such an approach would amount to begging the question and would thus be surely illegitimate. Yet, our recurrent reference to terms such as justice and equity notwithstanding, the sets of issues we are interested in are not normative, but empirical. So this objection would not apply. We could have operationalized the concept of an equitable distribution of benefits and burdens from regime-based cooperation in terms of (the distance from the distribution required by) the difference principle and then tested whether the equity hypothesis, thus understood holds up empirically. Whatever the outcome of this test, it would not be affected by the twists and turns of ethical theorizing or even by the emergence of a final consensus on the nature of international justice. (In fact, the worst thing to happen is that we would have to add quotations marks to the word "equitable" in the hypothesis.)
Still, we have chosen not to resort to this nominalistic expedient, as to do so would have been incompatible with the (meta-)theoretical context that we have placed the equity hypothesis in. Interpreting this hypothesis from a reflective point of view suggests that the relevant "data" are subjective and that, consequently, any simple application of some external standard of justice - whether grounded in a sophisticated ethical argument or arbitrarily chosen - is misguided. What is required in such an inquiry is not the solution of a philosophical dispute, but a careful reconstruction of the subjective conceptions of justice that are shared by the actors (see also Hurrell 1993: 69). Nevertheless, moral philosophy is not irrelevant to our undertaking, but our concern shifts from its normative to its meta-ethical branch, as it were. The reason is that we need to know about the formal characteristics of the concept of justice to be able to recognize a given conception of justice as such.[Note 46]
Two "signposts" are supposed to help us identify those statements and arguments with regard to distributional aspects of regimes which are moral in content and, more specifically, relate to issues of justice.[Note 47] (1) Distributive justice is a social norm whose function is to settle, and mediate between, competing claims to scarce social goods. Consequently, the practical discourse on justice is not detached from individual interests, it depends on their existence. In the absence of conflicts of interest there would be neither justice nor injustice. (2) The concept of justice is not identical with the concept of equality, but it implies equality in the sense that like cases are to be treated alike. To put it differently, in a cooperative venture the allocation of gains and costs must not be arbitrary. It must accord with a rule of distribution (or a consistent system of such rules).[Note 48] These considerations allow for the following specification of the equity hypothesis: A regime is ceteris paribus more robust if its distributional consequences satisfy a rule of distribution which is acknowledged as valid by the participating states and regarded as pertinent to the type of situation that is exemplified by the situation at hand.[Note 49]
In principle, this proposition should suffice to guide an empirical analysis of cases. On the other hand, it is still somewhat formal. We therefore decided to resort to a third "signpost", which we could also refer to as a piece of foreknowledge ( Vorverständnis ) in the sense of philosophical hermeneutics. We have just observed that in a sense justice implies equality. The prescription that like cases are to be treated alike is not of much help, though, as long as we are not told the respects in which persons (or groups of persons) have to be alike to deserve equal treatment. For, obviously, there are countless respects in which any given set of persons are "equal" or similar to one another (and just as many in which they are not). Hence, what we need to identify are those properties of individuals (or collectivities) which are morally relevant (Hart 1961: 153-63). In common morality the idea of justice is closely linked to certain precepts. Different (and in part even contradictory) in content these precepts are very similar in form: Each requires institutions (or decisionmaking units) to distribute benefits (or, in other cases, burdens) in proportion to some variable such as merit or needs in the subjects of justice (Perelman 1990: 31-46). The reflective (or subjectivist) approach that we intend to take to studying the impact of the normative quality of distributional consequences on the robustness of regimes suggests that we use these common morality notions as points of departure in our attempt to decipher the conceptions of justice that are shared by a group of states in a particular issue area.[Note 50]
Nothing in the literature on regime stability and effectiveness suggests that all variation in the robustness of regimes may be put down to differences in their distributional consequences. Hence, both the symmetry hypothesis and the equity hypothesis include the ceteris paribus clause. Attempting to test these hypotheses empirically we must take this qualification into account and seek to control (i.e. keep constant) as much as possible other explanatory variables, i.e. factors apart from distributional consequences that may influence the robustness of regimes.
Basic scientific methods differ in the way they handle the problem of control (Lijphart 1971, 1975). Among the methods that are applicable to the research problems of IR the statistical method, which achieves control by means of partial correlations,[Note 51] is most effective in this regard. Statistical procedures, however, cannot be applied unless a sufficiently large number of cases are available. In principle, this prerequisite does not rule out studying regimes with the statistical method. The universe of cases seems to have grown enormously within the last decades.[Note 52] Moreover, the number of cases can often be further enlarged by adopting a longitudinal perspective, splitting up, as it were, one "case" into a plurality of cases. Measuring the kind of variables that are discussed in regime theory (Levy/Young/Zürn 1995) is rarely a straightforward matter, however. In fact, a considerable familiarity with the case at hand is usually indispensible, driving up the expenses (and imponderabilities) of standardized "coding". At any rate, this applies to the variables of immediate interest in this project (distributional consequences, robustness) and extends to many of our prospective control variables as well (see sec. 5 below). As a result, aiming at a statistical test of our hypotheses would have been no viable strategy in the absence of extensive resources.[Note 53]
In order to secure, under these circumstances, a maximum of reliability for our results, we have embarked on a method that Lijphart and others (e.g. George 1979) have recommended as a substitute for the statistical approach. This method, which Lijphart (1975: 163-5) calls the "comparable-cases strategy", tries to solve (or mitigate) the problem of control by means of a careful selection of cases.[Note 54] Accordingly, the cases which are chosen for comparison should meet the following conditions: 1) They should display as much variation as possible with respect to the independent variable of interest (here: distributional consequences). 2) They should be as similar as possible with respect to other explanatory variables (here: other variables influencing regime robustness).[Note 55] Moreover, in order not to decide the empirical test beforehand, the respective values on the part of the dependent variable should be completely ignored in the selection process (which is most easily achieved if they are unknown to the researcher when she makes her choice).
One of the disadvantages the comparable-cases strategy has relative to the statistical method is that it permits only "conditional" results: the hypothesis is tested against a special constellation of background conditions (defined by the values of the controlled variables) which cannot be varied at will through partial correlating. As a result, interaction effects might go unnoticed.[Note 56] As Lijphart (1975: 165-172) points out, though, the comparable-cases strategy is not merely a "second-best", but enjoys special advantages of its own. These advantages derive from the fact that, within limits, it offers the opportunity to combine the strengths of the statistical method (achieving a considerable measure of control) with those of the case study method (intensity, accuracy, descriptive richness) (Eckstein 1975). Case studies allow for the application of process tracing techniques, which, in turn, can help to assess the plausiblity of specific causal claims or detect causal pathways (George 1979: 46, George/McKeown 1985: 34-41). In other words, case studies may suggest answers to question of the type: How is it possible for X to occur? as opposed to questions of the type: Under what conditions can we expect X to occur? (Dessler 1991) Thus, even if we succeeded in establishing (with statistical means) that just regimes tend to be comparatively robust, we would still be ignorant of why this is so. Only case studies promise non-speculative answers to this question.
The case for in- depth studies is further strengthened by the (meta)theoretical context of the equity hypothesis. Reflective scholars have criticized rationalistic institutionalism for sticking to a (positivist) epistemology that is out of tune with the intersubjective nature (or conceptualization) of their subject matter. Consequently, they have called for the introduction of hermeneutic research techniques into the study of international institutions and regimes (Kratochwil/Ruggie 1986: 764-6). As noted above, we intend to heed this criticism, even though the general outline of our research design is surely still overly "positivistic" (and self-consciously so).[Note 57] Bringing to bear interpretative elements in empirical research, however, requires engaging in some sort of case study method.
Needless to say, the comparable-cases strategy is not a panacea Simple as the rules of selection may be, the search for appropriate cases can be a tedious and frustrating process, in particular if the similarity of cases is not just inferred from some broad indicator (such as geographical region), but actually to be checked case by case. In this way, a considerable part of the time and labor that is saved due to the small-n dimension of the study, may get lost again in the course of case selection. Furthermore, the closer one looks at cases before they are eventually selected for test purposes or left aside, the more difficult it is to remain in ignorance about the values of the dependent variable.[Note 58] Not unexpectedly, therefore, our project has been riddled with these problems as well.
The method we have selected to guide the empirical evaluation of our hypotheses seeks to secure a maximum of reliability of results by requiring a focus on comparable cases. The process of case selection has to heed this requirement.
This process is further complicated by the nature of our dependent variable. As we have observed above, regime robustness is a dispositional concept such that this property only shows under certain conditions, which we call exogenous challenges. Furthermore, attempts to measure regime robustness should take into account that challenges to regimes can vary in strength. Hence a first set of conditions of comparability asserts itself: The regimes that are to be examined in detail and compared with others must have confronted at least one exogenous shock each, and those shocks the responses to which are compared should not differ much in strength (which is presumably best achieved if the shock is the same in all cases).
As a frame of reference for assessing the comparability of challenges we have developed a theory of exogenous shocks to regimes. In this connection, we have noted that the type of strategic situation (PD, coordination, etc.) that underlies a given regime has important repercussions on its overall stability (or robustness, for that matter). It is therefore appropriate to control for this variable when trying to assess the robustness of different regimes. In our project we have done so by restricting our universe of cases to collaboration regimes.
Against the backdrop of this decision the following rationale can be used to detect further control variables: Since it is difficult (if not impossible) to draw a sharp dividing line between variables that influence the robustness of regimes, on the one hand, and the events that put this robustness to the test (i.e. what we call exogenous challenges), on the other,[Note 59] we can resume the considerations we have developed in section 2 in connection with our theory of exogenous challenges. To be sure, exogenous shocks as such are not variables, but events. However, we have "deduced" them as (causes of) changes in certain situative variables (such as the distribution of power), and, as we wish to argue, any such variable is a plausible control variable for our study .[Note 60]
Table 2: Control Variables for Collaboration Regimes
| External Variables | Institutional Variables |
|
|
One important addendum is necessary, though. Control variables are not restricted to aspects of the environment of the regime. On the contrary, institutional attributes can be expected to be important determinants of regime robustness (possibly) in addition to distributional consequences. The literature that touches upon issues of effectiveness and stability stresses two such variables inter alia : (1) The existence and effectiveness of monitoring arrangements or other compliance mechanisms. It has to be kept in mind, though, that the crucial (intervening) variable is transparency of (others') behavior , which is not a function of certain institutional devices alone, but depends, in a large measure, on the nature of the regulated activity (Young 1992: 176). (2) Finally, students of regimes, rationalists as well as reflectivists, have emphasized the significance of the nesting patterns of regimes (Keohane 1984, Aggarval 1985, Müller 1993).
One of the criteria for the selection of cases as specified by the comparable-cases strategy is that they vary with respect to the independent variable under study. Variation in a given sample, in turn, depends partially on how the variable is categorized. Each of the two hypotheses under study suggests a minimum dichotomous categorization: {symmetrical, asymmetrical} and {just, unjust}, respectively. As we have pointed out above, these two categorizations are incommensurable: knowledge of the value of one variable does not permit inferring the value of the other. This means that, ideally, we should make sure that the set of cases chosen for closer investigation displays variation on both dimensions of the independent variable (symmetry and equity). Doing so in this case, however, would have meant clashing with an important procedural implication of the comparable-cases strategy: the sequence of first selecting cases according to certain criteria and then testing the hypothesis against the cases selected. (The sequence is important, because it contains the solution to the problem of control.) Yet, the data requirements of the equity hypothesis (in its reflective reading) appeared to be simply too demanding to allow for such a sequence to be carried out. Consequently, we confined ourselves to trying to secure sufficient variation with respect to (the categories implied by) the symmetry hypothesis, simply hoping that our expectation to find both (relatively) just and (relatively) unjust regimes included in our sample will not be disappointed.[Note 61]
This decision had consequences for another issue raised by the comparable-cases strategy's first criterion of selection. Obviously, we cannot know whether a given set of cases vary with respect to the independent variable unless we can measure it. How then can distributional consequences of regimes be measured or, at least, assessed in a plausible way? A first problem arising in this connection is conceptual: Should we try to establish the actual distribution of gains (i.e. adopt an "objectivist" perspective), or should we attempt to find out how the actors themselves perceive it (i.e. adopt a "subjectivist" perspective)? Once more, this question had to be answered in light of the theoretical context of the hypothesis to be tested. Having already decided to disregard the equity hypothesis in the case selection stage, only the broader theoretical context of the symmetry hypothesis was relevant. As opposed to reflective institutionalism, rationalistic approaches to the study of regimes do not problematize actors' perceptions (see sec. 3.2. above).[Note 62] Consequently, the objectivist perspective seemed to be appropriate. On the other hand, the theoretical justification of the symmetry hypothesis makes reference to certain motives on the part of the actors (in particular, the endeavor to avoid relative losses) which clearly operate on the basis of some perception of the distribution of benefits and burdens. Consequently, there is a certain danger of oversophistication in designing measurement instruments for distributional outcomes: we should not try to be more precise than the actors themselves can be plausibly expected to be.
The latter consideration casts doubt on the usefulness (as far as our project is concerned) of some of the techniques that are discussed in the literature dealing with the problem of regime effectiveness. Thus, Thomas Biersteker (1993) has made the case for the application of the strategy of counterfactual argument (Fearon 1991) in the empirical study of regime consequences. This method requires the researcher to perform a disciplined thought experiment. We are supposed to imagine what would have happened with regard to some aspect of the world (or variable) if the event whose causal role we are interested in (e.g. the establishement of a particular regime) had not occurred. The difference between the actual and the counterfactual course of affairs (with respect to that variable), then, can be regarded as (part of) the effect of that event. The thought experiment is a disciplined one, because it does not legitimize free-hand speculations, but requires the analyst to base the evaluation of the counterfactual scenario on empirically corroborated theories (laws, regularities, principles of behavior, etc.). But not only is it often very difficult to meet this requirement in practice, given the present state of IR theory. As far as the special needs of this project are concerned, it is doubtful to assume that decisionmakers really go about things this way when evaluating advantages and disadvantages of a given institution for them and others.
In the face of such a dilemma pragmatic considerations must gain the upper hand. Thus, in the case selection stage we drew from various sources when trying to estimate the symmetry or asymmetry of distributional consequences of particular regimes. (1) We relied on the assessments of issue-area experts. (2) We analyzed the principles, norms, rules, and procedures of the regime with regard to the opportunities and restrictions they involve for actors, whose (initial) situations differ. (3) We contrasted pre-regime formation and post-regime formation performance of actors with respect to the relevant values (e.g. welfare in economic and environmental regimes), attempting to rectify distortions by cautiously introducing (informal) counterfactual considerations.
Not surprisingly, measurement problems arose on the part of the control variables as well, which often made difficult (and perhaps even, at least from an ideal point of view, questionable) decisions necessary. Some forty international regimes were included in a time-consuming and yet often too superficial screening procedure. This undertaking was guided and structured by a set of questions covering most of the theoretically relevant variables we have discussed in this paper. Through various "qualifying rounds" five groups of cases emerged as objects for in-depth study and as empirical base for the competititve hypothesis test. Each of these groups consists of cases that belong to the same broad issue-area. Direct comparisons will be made within groups and not across groups, in order to secure a maximum of comparability. The range of (broad) issue-areas that are covered by these groups is wide and spans humanitarian issues, economic affairs, environmental politics, and international security.[Note 63] Since we do not intend to do comparisons across groups, this diversity should not affect the reliability of our results. At the same time, it opens up the opportunity for distinguishing between (sets of) easy and (sets of) hard cases relative to a particular hypothesis, which allows for a more differentiated evaluation of the test results: failure to account for a set of easy cases weighs more than making erronous predictions with respect to hard ones. For example, it may be argued that for the equity hypothesis humanitarian affairs are a comparatively easy case, while security cases are hard ones (and that just the opposite is true for the symmetry hypothesis).[Note 64]
A. Humanitarian Affairs:
The Food Aid Regime
B. Protection of the Environment:
European Regimes for the Protection of Regional Seas including the North Sea, Baltic Sea, and Mediterranean Sea Regimes
C. Economic Issues I:
Fisheries Regimes (possibly including regimes for the protection of marine mammals)
D. Economic Issues II:
International Commodity Trade Regimes
E. International Security:
The Nuclear Non- Proliferation Regime
The Regime for the Limitation of Conventional Forces in Europe
Note that, contrary to appearance, the food aid regime is a group of cases. During its life-cycle the regime has undergone important changes which have significantly affected its distributional consequences . This fact gives us the opportunity to carry out a longitudinal study.
As we have seen, equality and justice are not identical concepts. Yet, although the intensions (meanings) of the two concepts are different, their extensions (i.e. the classes of distributions which come under the two concepts) overlap: a regime that generates an equal distribution of benefits and burdens is not necessarily unjust. Whether or not a symmetrical allocation of gains also satisfies the criterion of justice, depends on the precise circumstances of the case, in particular on the distribution of those properties which are specified as morally relevant by the appropriate precept of distributional justice (e.g. merit or needs).
This logical relationship between justice and equality suggests an evaluation procedure to be applied to our case studies which at first focuses on the symmetry hypothesis and only afterwards brings in the equity hypothesis. Evaluating the record of the symmetry hypothesis will yield either of following two results: 1) The hypothesis proves incapable of explaining the variation on the dependent variable, regime robustness. 2) It proves capable.
In the first case we conclude that the rationalistic hypothesis has been disconfirmed or perhaps even falsified. We then check the performance of the equity hypothesis. Again, only two possibilities exist:
a) The second hypothesis also fails to account for the phenomenon under study. In this case we conclude that, at least for the time being, there is no evidence that our fundamental intuition that robustness is systematically related to distributional properties of regimes is substantiated. (Since other hypotheses associating distributional patterns and degrees of robustness do exist, the conclusion has to be this cautious. In fact, should both justice and equality prove to be incapable of accounting for robustness, the logical next step would be to probe the empirical standing of the alternative rationalistic hypothesis which we have referred to as the "power hypothesis".)
b) The second hypothesis (equity) does account for the variation observed. In this case we conclude that the equity hypothesis is superior to the symmetry hypothesis as an explanation for regime robustness. In addition, we take this result as an important piece of evidence indicating that reflective institutionalism holds more promise than the now conventional rationalistic approaches to international regimes. Finally, we can explain why the rationalistic hypothesis sometimes seems to be supported by the data: as we have noted above, there are likely to be cases in which the criteria of equality and justice require the same distribution.
In the second case - i.e. robustness turns out a function of symmetry - the rationalistic perspective can be said to have been corroborated. Due to the overlap of equality and justice, though, this result is inconclusive with respect to the competition of our two hypotheses. Consequently, the performance of the equity hypothesis has to be checked as well. Once again, two outcomes are possible:
a) The equity hypothesis fails to account for the variation on our dependent variable. In this case we conclude that the symmetry hypothesis offers a better basis for an explanation of the robustness properties of regimes than does its reflective rival and that, consequently, the results of our project strengthen rather than weaken the mainstream perspective on international institutions.
b) The equity hypothesis correctly predicts the differences in the robustness of our test cases. The possibility of this somewhat odd result - both hypotheses, though mutually inconsistent, seem to have been confirmed - once more reflects the special logical relationship between justice and equality we have pointed to. The cases selected do not discriminate between the two hypotheses, because, in these cases, justice requires a symmetrical distribution of gains and burdens. If this were to happen (unlikely though it is), we would need to look for, and to analyze, an additional, "crucial" case, i.e. a case in which (1) fairness commands a distributional pattern that clearly deviates from symmetry and (2) the actual distribution of gains is either symmetrical or just, such that the case-specific predictions that can be inferred from the two hypotheses will differ accordingly. Only the outcome of this study would determine which of the two hypotheses is truely corroborated by our project.[Note 65]
Figure: Research Design and Evaluation Scheme
Address of the Authors:
Abteilung Internationale Beziehungen/Friedens- und Konfliktforschung