The Old Governance: Informal Institutions in the EU Andrew Moravcsik
Princeton University1 In international politics, informal governance is the norm, not the exception. Even at their most centralized and autonomous, as in the European Union (EU), international institutions are generally more flexible, consensual, and open-ended than their domestic counterparts. Over the past 25 years, the EU has evolved toward even greater reliance on informal institutions. A strengthened European Council, norms of consensus voting instead of qualified majority voting, multi-speed arrangements in money and free movement, expanded intergovernmentalism in foreign policy, and the Open Method of Coordination (OMC) are only a few examples. In many ways, this is not the new governance; it is a return to the old. This invites us to ask: What motivates governments and their constituents to select informal institutions? Under what conditions might we expect this to lead to voluntary informal policy diffusion, policy convergence, and international cooperation? Who wins or loses from such processes, and is thus likely to support or oppose them?
To help answer these questions we can draw on a large empirical literature on the OMC and other informal mechanisms. Often this literature presents these mechanisms as exemplars of one or another ideal-typical theory—most often of some form of learning. Yet a number of alternative theories might plausibly explain the nature of informal international governance. This memo considers four: regime theory, two-level games, simple learning, and democratic experimentalism. In each case, I review the theory, suggest hypotheses that might apply to the EU, and briefly sketch what the empirical literature on OMC and other informal processes implies about its validity. Since I expect other contributors to this workshop to focus primarily on democratic experimentalism, I focus primarily on the other three.
My central argument is that old governance is explained well by old theories. The extensive empirical literature on the OMC strongly confirms rationalist, agent-based theories—here represented by regime theory and, to a lesser extent, two-level games analysis. In both models, rational state leaders and officials choose institutions and policies to maximize the national or governmental interest. They seek to reduce cross-border negative externalities that stem from sub-optimal uncoordinated policies in a way that suits particular coalitions of domestic actors. The empirical evidence for the basic predictions of both models is strong, even if we draw on studies by those sympathetic to other approaches. Explanations focusing on learning—whether simple learning or democratic experimentalism—seem to explain modest deviations from the predictions of this core theory of integration. It would thus be instructive for future studies of OMC and informal governance not to be exploited as exemplars of single theories, but to involve more explicit tests among alternative theories and to catalyze efforts to synthesize them creatively.
Theory 1:Regime Theory – “The European Constitutional Settlement” From a regime theoretical perspective, international cooperation aims to reduce cross-border negative policy externalities stemming from uncoordinated national policies. Governments choose between formal and informal commitments based on costs and benefits. Specifically, they calculate the potential benefits of substantive policy coordination andthe possible costs that might accrue if they and their constituents are subjected to distributive shocks. Informal institutions, as opposed to uniform, binding formal commitments, permit governments to retain more flexibility and autonomy to deal with distributional shocks, but at the cost of less effective coordination to eliminate sub-optimal policies. Informal institutions like the OMC emerge where governments prefer to engage in sporadic and ad hoc policy convergence because the incentives are insufficient to overcome concern about future distributive shocks. The critical concern driving informality is not just the magnitude of distributive shocks (which could be offset by linkages or side-payments) but uncertaintyabout their future magnitude and direction.2 Informal institutions govern areas where states expect cooperation to be uneven or sporadic. Throughout, the causal chain runs from government interests to the OMC policies, not (as learning models have it) in the reverse direction. This regime theoretical approach generates the following hypotheses.
Hypothesis 1A: The smaller the substantive benefits of policy coordination and the greater the potential for asymmetrical, politically salient, and distributional shocks, the more governments will favor informal institutions.
Hypothesis 1B: All other things equal, governments are more likely to accept policy transfers consistent with their current or desired policies, while seeking to avoid policy “misfits” that impose adjustment costs.
Hypothesis 1C: When distributional effects of bargaining and enforcement are uncertain, governments prefer informal cooperation, voluntary enforcement, and flexible “coalitions of the willing.” When future distributional consequences of policy convergence are uneven but predictable ex ante, governments favor multi-speed arrangements, vanguard groups, and opt-out clauses.
Hypothesis 1D: Informal governance will expand as the EU takes on new issues and “drains the swamp” of the most promising areas for cooperation (i.e. those with the greatest functional benefits and lowest distributive risk). Hypothesis 1E: All other things equal, informal policy transfer is more likely to occur in “the shadow of hierarchy,” i.e. when financial incentives, side-payments, quid pro quos, formal institutional mandates and oversight, and linkages to formal EU membership encourage adoption. This is most likely in transitional, newer or poorer member-states, and in issue areas linked to enlargement, structural funding, or participation in formal arrangements like the Euro.
Hypothesis 1F: Because governments select informal institutions where they expect to act only sporadically or unevenly, the OMC is far more likely to trigger change in domestic discourses and agendas than in policies.
The empirical literature on informal EU institutions such as OMC offers strong empirical confirmation for these hypotheses. Governments tend to have rational reasons to guard their sovereignty and strike a calculated balance between cooperation and control (1A).3 The issues where member states select OMC—social and employment policies, for example—have two distinctive characteristics. First is a relatively low level of the cross-border policy externalities (1A). Citi and Rhodes note: “In the absence of a much higher levels of cross-border labour flows, or trends for countries to diminish welfare spending in the search for competitive advantage, or greater evidence than presently exists for widespread tax competition, there are equally few incentives to compel serious national actor engagement in even the softer forms of social and employment policy coordination advanced under the EES/OMC.”4 Exceptions, such as efforts to impose budgetary control as a flanking policy to the Euro, are rare, and generally induce a higher degree of formal institutional backing.
The second distinctive characteristic of “new” areas of informal and OMC governance (national social and labor policies, in particular) is that they tend to be politically salient, deeply embedded in specific national political compromises and subject to distributional shocks imposed unpredictable economic and social circumstances (1C). Following Fritz Scharp, Citi and Rhodes observe: “Regulations would be opposed by both national governments and actors, for at least three reasons: uniform policy convergence would undermine the 'social contracts' enshrined in their different welfare regimes; it would have deleterious effects on economic growth…; and the inherent differences and political salience of national health and pension systems render EU harmonization unrealistic. Thus, [the classic] method has been abandoned…”5 As regime theory predicts, there is scant enthusiasm for Europe-wide legislation in the areas to which OMC has been applied, particularly those involving redistribution (1A).6 For example, there are surprisingly few viable proposals (let alone widely supported ones) for a European social policy, and little significant pressure for any policy outside of current mandates (1D).7 The limitations on cooperation in areas of informal governance are not contingent on institutional design, lack of information, insufficient passage of time, or domestic blockages. They are inherent. This has two further implications for the OMC. The first is that the OMC is likely to be relatively ineffective at transferring policy. A consistent finding across the entire literature, even among the most positive analysts, is that OMC often changes elite bureaucratic discourse, but has little impact on major public policy (1F).8 This is not surprising because, as micro-studies reveal, the binding constraint on policy transfer is not lack of information about superior policies, but adjustment costs (“misfit”), compatibility with broader domestic social and political compromises, and distributional costs, based on pre-existing interests (1B).9 In any case, governments established informal institutions in places where they intended only to transfer policy only sporadically.
The second implication for OMC is that exceptional cases of successful policy transfer within informal institutions often rest ultimately on traditional EU policy-making instruments and incentives (1E).10 A disproportionate number successful policy transfer cases (as opposed to cases of change in discourse or agendas) involve EU enlargement, climate change or monetary union—areas where informal mechanisms of policy learning are used as flanking mechanisms for explicit conditionality, concrete financial incentives, and formal legislative mandates.11 The sad fate of the “stability pact” and macroeconomic discipline in the Eurozone after the removal of those incentives, as exemplified by the current Greek crisis, is the exception that proves the rule.
A final implication of this analysis is OMC, rather than marking—as some have argued—a dynamic new constitutional paradigm that will progressively expand to take over the EU or global governance, marks a return to traditional international politics. From the perspective of European integration, it signals arrival at an institutional equilibrium: the “European Constitutional Settlement.”12 This settlement broadly defines some issues (trade, business regulation, some environmental policies) as EU competences, some issues (social policy, pensions, taxation, third-country immigration, education, cultural policy, taxation) as national and local, and other marked for traditional intergovernmental treatment. Though the lines sometimes blur in specific cases, this division has remained stable over the past 20 years, resulting in a stable situation in which about 90% of European policy decisions are handled at the national level and 10% in Brussels. Institutional flexibility has been required to achieve even this level of EU policy-making. No major EU initiative of recent decades—not just the nascent social policy, but the Euro, Schengen, Justice and Home Affairs, immigration policy, foreign policy-making—applies to all member states uniformly. All involve de jure or de facto “coalitions of the willing” (1C). EU institutions have evolved since the 1970s away from the classic uniform, centralized “community method” (if it ever existed) toward a more intergovernmental system reliant on informal intergovernmental mechanisms like the European Council (1D). The steeper trade-off between scope and scale of policy-making, visible in every aspect of the EU, suggests that the OMC, rather than a new and innovative form of policy-making, is another step along the EU’s well-trodden path toward constitutional equilibrium.
Theory 2: Two-Level Games – “Membership has its Privileges” The theory of “two-level games,” like regime analysis, views the primary causal chain as running from government interests to the OMC policies, not (as learning models have it) in the reverse direction. The critical characteristic of the OMC and other informal arrangements, in this view, lies in their ability to set domestic informational, institutional, and ideational agendas in ways that governments prefer. Informal arrangements tend to bolster the domestic power of those with privileged access to information, institutional agendas, and ideas at the international level, which they can manipulate domestically to achieve policy outcomes. In this view, the leading actors are national executives—officials backed by the national political leadership—who can best manipulate and control access to the process, and it tends to benefit them at the expense of other groups in domestic society, notably national parliaments and some broader civil society groups.13 Governments favor informal arrangements in circumstances where they are unsure, as we saw above, that they want to act, and we should not forget that they play an important role in defining the OMC agenda and the nature of the performance criteria.14 This general perspective suggests the following hypotheses.
Hypothesis 2A: The OMC and other informal procedures, and specific policy transfers, are likely to emerge where governments seek to implement unpopular or controversial distributive policies, either to “shift blame” to the EU or to cover for domestic policy with contrary symbolic rhetoric.
Hypothesis 2B: To strengthen domestic political support for and the credibility of commitment to policies, governments may approve the “uploading” of current or favored future policies to the EU level and then “download” the same policies.
Hypothesis 2C: Officials, politicians, and other “gatekeepers” to the EU-level policy process will be selective in policy transfer, with their actions depending on whether they favor the goals of the policy, consider it relevant to local circumstances, and anticipate it politically viable.
Hypothesis 2D: Domestic actors may empower officials to alter domestic elite discourses and agendas, but governments have no rational incentive to delegate authority to change policy in politically salient and controversial areas unless governments and powerful domestic political forces approve.
Hypothesis 2E: Informal arrangements will work best in countries where “Europe” is popular, creating support for specific policy changes by linking them to an ideologically salient ideal. In such countries, governments that favor specific policy goals have greater incentive to promote informal arrangements and policy transfers. Elsewhere, actors may quietly offer diplomatic support for cynical reasons (e.g. to undermine formal cooperation).
Hypothesis 2F: Informal processes are supported by countries that seek to implement reforms favored by a consensus of European governments, or have already implemented similar reforms. Member states with other, outlier preferences may view them with some suspicion. As the consensus changes, the policies will change.
The empirical literature on the OMC offers moderately strong empirical support for the “two-level games” theory. European social policy emerged in the late 1990s—a period of social democratic dominance of the EU—in the absence of any consensus about viable policies. Social democratic parties, then as now, faced two problems. First, the EU was unpopular among their electorate. With the EU having recently moved forward on issues like monetary union, it still lacked a “social dimension.” This placed pro-European parties social democratic parties in a bind Second, many of these parties had recently committed to unpopular “third way” domestic social and labor reforms, such as increased labor market flexibility, linking welfare to work, and other such “market-oriented” mechanisms. These triggered strong domestic opposition in their base. The OMC process in social and labor areas “emerged…parallel” to the prior European of “third way” reform programs, which policy documents suggest it was designed to support (2A).15 Such policies are candidates for informal institutionalization because the benefits from cooperation are relatively low, compared to the uncertainty and risk of potential domestic and transnational distributional implications; governments preferred institutions that would permit sporadic, ad hoc policy convergence, over which they were careful to maintain sovereign control (2D). Zeitlin has argued that “governments often use references to OMC processes as a source of legitimation and blame-sharing in order to advance their own domestic agenda, sometimes irrespective of their real influence on policy decisions.”16 An example is the German use of the EES as rhetorical backing for unpopular changes (2A, 2E). 17 Micro-studies of policy diffusion within the OMC suggest that, far from acting in an uncontrolled way, officials carefully calculate and manipulate such mechanisms (2B, 2C). The empirical literature suggests that national officials enjoy autonomy, but it is limited—as the “two-level” theory predicts. Studies of decisions whether or not to engage in “peer review,” for example, suggest that officials feel constrained, acting to transfer policies and models in anticipation of domestic political imperatives and the broader domestic social model.18 In other, often low-salience cases, they appear to act freely to promote national discourses and set elite agendas—yet in few, if any, cases are they able to exploit this access to the EU level in order to obtain substantive policy goals outside of scattered low-salience issues (2D). The OMC literature consistently notes that national officials are quite selective about the policies they “upload” and “download”, or consider seriously for peer review or adoption. Governments are often frankly manipulative in how they doing this, using redefinition of goals to avoid potentially costly “misfits.”19 Governments hedge by “uploading” flatly contradictory elements into EU documents, so they can choose those that suit their domestic system (2B).20 Much of the micro-evidence for the positive effect of OMC proposals on domestic discourse does not suggest any sort of technocratic learning is occurring, but instead points to linkage to the European ideal in pro-European polities (such as Germany) (2E). In the “two-level” approach, “shaming,” “peer pressure,” and “mimicking” occur not because of technocratic consensus, as learning theorists would have it, but because of linkage to shared European ideals, or because of manipulation of domestic information by committed elites. These mechanisms have been shown not to work correspondingly well in less pro-European settings (such as France or Britain).21 Similarly, the observation that the OMC seemed to decline in influence on discourse and policy when Social Democratic parties lost their hold on EU policy seems consistent with the two-level games interpretation (2F).
Theory 3: Simple Learning – “Leaders and Laggards” OMC and other informal arrangements offer a conduit by which governments can learn what their counterparts in other EU governments are doing and selectively transfer policies they deem useful. Simple or one-shot learning of this type requires not just dissatisfaction on the part of the target state about current policy; it also requires that a successful and easily appropriated alternative policy be readily available.22 It also assumes that governments do not have the resources, analytical capacity or inclination to engage in long-term, controlled, repeated play “democratic experimentalist” observation of policies in multiple settings—which might be more reliable. Instead, they must decide on a more limited empirical basis what foreign models are successful and similar enough to emulate. Such a process can be at most only boundedly rational, and sometimes resembles socialization or “mimicking” more than rational learning.23 It implies that there are already knowledgeable states in the EU, some of which have already adopted successful policies (“leaders”), and who transfer their knowhow to backward states (“laggards”). Since a successful policy innovation might not be relevant in a very different setting, we might expect that successful policy innovations will tend to be copied and shared among societies and polities that are broadly similar in important ways. In this context, governments may also “shame,” apply “peer pressure,” and encourage “mimicking” and “discursive diffusion.”24 Simple learning of this kind within transgovernmental networks in the EU is an alternative to learning from domestic experience, academic studies, the media, consultants, professional networks, other international organizations, countries outside the EU, or hearsay.
Hypothesis 3A: Simple learning is likely to be asymmetrical, with “leaders” teaching “laggards.” Those who adopt policies are most likely to occur in transitional governments (e.g. enlargement countries) forced to adopt new policies, after catastrophic policy failure, or in new policy areas where only a few countries have yet developed competence. In such cases, policy change is inevitable, prior national learning is unlikely to have taken place, the political equilibrium may be more malleable, and established models exist elsewhere.
Hypothesis 3B: Learning is more likely to influence domestic discourses and agendas than policies, particularly when backed by material incentives. Discourses and agendas are translated into policy, however, only where the issue does not have major distributive implications or is in the interest of dominant domestic political actors.
Hypothesis 3C: Rational EU learning can influence domestic discourse and agendas only where EU bureaucratic networks enjoy a distinct comparative advantage or monopoly in information transmission, analysis and dissemination over other networks and analytical methods, such as policy think-tanks, academia, media, classical international organizations, bilateral contacts, professional associations, multinational business, consultants, non-EU countries, and national systems of evaluation and experimentation. Learning in any other context is likely to be less efficient, successful, and sustainable.
Hypothesis 3D: Rational officials will learn best about small, discrete policy innovations not dependent on broader national political, social, cultural or economic structures. In cases where micro- and macro-structures are connected, rational officials will learn best from governments with similar macro-structures but different policies. Learning in any other context is likely to be less efficient, successful, and sustainable.
Hypothesis 3E: Learning in any given area is cyclical, rising as lessons are learned, but falling thereafter until new policies are available.
The empirical literature on learning in informal EU mechanisms suggests that it is a mechanism of rather limited scope and power. In almost all cases that have been observed, only a modest effort is made to draw broad policy conclusions different than those that motivated creation of the OMC in the first place (3B). Officials only consider a very small proportion of possible policies.25 “Peer review” and other mechanisms function on a very selective basis, with officials selecting for visits and adoption narrow policies perceived as successful from specific countries perceived as having domestic systems similar enough to their own to be appropriate models (3D). The reason for this is fundamental. National systems vary widely, and empirical studies of the motivations of officials engaged in peer review suggest that they appreciate that the up-front political, social, and economic costs of adopting new policies are often high. Learning appropriate lessons from foreign examples requires that observers already have a mental model of the relationship between a policy instrument and its political, social, economic and cultural context—issue-specific knowledge that is difficult to generate on a one-shot basis. In simple learning, diversity is a mixed blessing. Hence learning tends to take place in cases where there are clear “leaders” and “laggards” (3A).
Most of the examples of possible transnational learning cited in the literature on OMC appear to be examples of simple learning: one-way, one-shot, and hierarchical. As we have seen, even sympathetic analysts have found very few prima facie cases of potential OMC influence on policy outputs (let alone social outcomes) within social policy or employment areas, where it has been most widely employed. Nearly all the exceptional examples where policy innovation or compliance may have taken place are found in new member states and the EU periphery, or where backed by clear inducements.26 The most striking cases are convergence around the single currency norms in the 1990s and 2000s, and European enlargement norms in the 2000s. In social policy, lessons were asymmetrical, disproportionately being taught by countries like Sweden and Denmark to countries like Greece.27 The simple learning model is consistent with the observation that OMC seems to have influenced, at most, a series of small, discrete policy innovations, while having little impact on broad policies, and with the observation that it seems to have expanded until about 3-5 years ago and currently seems to be declining in influence (3E). It is unclear whether OMC enjoys comparative advantages over other forms of transnational learning: As far as I know, this issue has never been systematically studied empirically (3F).
Simple learning is a risky model of policy transfer if extended beyond small discrete policies (3D). Alberto Alesina and Roberto Perotti have viciously criticized OMC learning models for implicitly assuming that the EU knows the “correct” policy model. They have argued that “this Kafkian exercise has contributed to set back the level of the debate and of understanding by the public by giving the impression that some EU institution actually knows how to solve the problem of European unemployment, if only national governments cooperated….”28 Whereas this may be exaggerated, there is some truth to it, as we learn from the best example of this sort of policy transfer at work: the adoption of Economic and Monetary Union. The convergence of countries toward the model of low inflation, central bank independence and, ultimately, a single currency, was not simply an inference drawn from academic writings or domestic experience; it was also an effort to extend the “German” model to other countries. By transferring monetary policy discretion to an independent European Central Bank—a non-majoritarian institution—the expectation was that price stability, economic competitiveness, and macroeconomic discipline would follow. This seemed plausible because it drew on a successful model, namely the German Bundesbank.29 In retrospect, however, this seems to have been an overambitious exercise, insufficiently attentive to the basic monetary economics of optimal currency areas and the political science of the “varieties of capitalism”—as European governments are currently discovering in seeking to salvage the Greek economy. Typical of the dynamics of simple learning, this was in most respects a top-down decision, taken by national officials and politicians, and supranational technocrats, over the objections of those officials closest to the informal functional networks.30 In the end, they are the decision-makers who matter when the issues are salient.
Theory Four: Democratic Experimentalism – “Unity in Diversity” In this view, OMC and other informal arrangements are a method to overcome the limitations of local and single-shot learning by encouraging policy learning that is high-n, iterative, and transnational. If we have enough cases and if learning is recursive, we can overcome the impediments to learning what might work in Italy by observing Sweden. More countries could learn from each other. As the OMC developer Mario Joao Rodrigues argues, “diversity in Europe should be treated as an asset (a natural laboratory for policy experimentation) rather than as an obstacle to integration.”31 The distinctive advantage of OMC vis-à-vis classical policy-making lies not in its ability to transfer learned policies, which could, after all, be achieved by classical means. It is to encourage deliberation about correct policy and policy goals through decentralized epistemic communities of officials and experts.
Hypothesis 4A: Same as 3B. (Informal arrangements change discourse and agendas, not policy, except in non-distributive areas.)
Hypothesis 4B: Same as 3C. (Informal EU networks must enjoy a distinct informational comparative advantage.)
Hypothesis 4C: Rational policy transfer will be proportional to the range of available cross-national data and analysis. Over time, it is likely to develop slowly within an issue area, but to grow more intensive over time.
Hypothesis 4D: The high data requirements and constraint imposed by distributional consequences mean that democratic experimentalist processes should work best in small, bounded, iterated and technical processes in which there are very high numbers of iterations of similar cases. (e.g. legal reasoning, compliance in established sectors)
Hypothesis 4E: Rational officialsmay be able to consider broader reforms if they can elicit information from a broader range of social actors.
Hypothesis 4F: The impact of the OMC should co-vary with the availability of peer review, benchmarking, explicit standards, public shaming, and other instruments.
Here I will be brief, as other discussants in this workshop will surely treat democratic experimentalist theory in more detail. We have seen that there is a consensus in the literature, including even the authors most sympathetic to democratic experimentalism, that OMC processes have had some impact on elite discourse and agendas, but little on policy.32 In some ways, this seems consistent with the predictions of the theory. This may be so because only a small proportion that goes on within OMC satisfies the core assumptions of democratic experimentalism. Sabel and Zeitlin posit that “The possibility conditions for experimentalist governance are arguably quite minimal: strategic uncertainty, meaning that policy makers recognize that they cannot rely on their strategic dispositions (e,g. more market vs more plan) to guide action in a particular domain (or equivalently that they do not know how to achieve their declared goals); and a multipolar or polyarchic distribution of power, in which no single actor has the capacity to impose her own preferred solution without taking into account the views of the others….Together, these conditions open up the possibility for transforming distributive bargaining into deliberative problem solving through the institutional mechanisms of experimentalist governance.”33 Yet the requirements for successful democratic experimentalism are in fact more onerous than the tone of this quotation suggests, and the empirical literature suggests they are not often met.34 Other necessary conditions include that EU networks be a relatively efficient source of information, information not be a decisive strategic resource even if their ability to manipulate it is imperfect, adjustment costs to new policies be low, superior common solutions exist and be knowable, etc.
As we have seen, all but the most detailed OMC issues are perceived by participants as involving domestic distributive concerns that generally outweigh the impact of international policy externalities (4A). In many cases, moreover, there is no obvious common policy solution, even after discussion. Nor is it obvious in some cases, to the extent that new solutions may exist, that an efficient way to find them is for countries to converse with one another via informal institutions. Policy-making concerning new rules involves a number of iterations that usually seems to approximate the assumptions of simple learning more closely than the high data requirements of democratic experimentalism. This mechanism has now been in place for over a decade and it is difficult to find little evidence of rational policy transfer being directly related to the range and quality of available data and analysis (4B, E).35 Few if any studies compare the learning opportunities in EU networks to other opportunities, however, so it is difficult to assess hypotheses about the relative efficiency of transgovernmental means of complex learning (4B). The impact of OMC has not increased consistently over time, as predicted; most analysts view its influence as having been cyclical (4C). As we have seen, studies tend to disconfirm the view that changes in discourse, agendas and policy correlate with the availability of OMC instruments. Dozens of studies have analyzed this problem—cited above—almost all of them attributing observed variation to functional imperatives and domestic politics (4F).
Still, there may be some reason to believe that there are distinctive pockets of EU activity in which the conditions of low distributional consequences, technical autonomy, and high iteration demanded by democratic experimentalism are approximated. Two promising areas would be legal enforcement and compliance oversight committees such as comitologie.36 While we do not want to draw too complete a distinction between legislation and enforcement, in the latter case the scope of immediate policy impact is often narrower, a number of similar cases arise over time, and a community or network of experts with semi-autonomy can function over periods of time. To be sure, the claims for the independence of the epistemic community of lawyers and comitologie offiicials can be, and have been, exaggerated—sometimes quite strikingly. The dominant discourse in comitologie remains that of national interest, and radical arguments of “social constructivist” socialization have been empirically disconfirmed. Yet there is clear evidence that officials in this context enjoy a certain autonomy. That autonomy is larger in the legal system. The striking evolution of the European Court of Justice that, over time, such systems can deepen in the exponential way predicted by democratic experimentalist theory (4C), as the enormous literature on the European Court of Justice suggests—the one area where a historical institutionalist account of the EU has real empirical purchase. There is thus a case to be made in these areas, and the literature might benefit from focusing on democratic experimentalism in areas where its theoretical preconditions are most obviously fulfilled.37
Conclusion A decade of research has generated a large empirical literature on OMC and other informal processes in the EU. Three results consistently emerge: (1) OMC is having a modest aggregate impact on European public policy and macro-political indicators. (2) The causal mechanisms conjectured to lead to effective policy learning within OMC can be reasonably well-explained by traditional factors that influence classical interstate cooperation in the EU: convergent national policy preferences, permissive domestic institutions, positive-sum interstate bargaining, and formal compliance mechanisms. (3) Where simple and complex (democratic experimentalist) learning takes place, it appears to do so within tight constraints set by rationalist theory. This is not to say that decentralized learning is unimportant in explaining some of what we see in European integration, or learning theories are flatly disconfirmed. It is to say that these processes can only function within relatively narrow exogenous constraints set by interests, power and institutions. This is what a decade of extensive empirical research shows.
This is a heartening result for social scientists and policy analysts—as well, I think, for democratic theorists!—because it is consistent with most of what we already know about how the European Union actually works. EU policy today reflects a carefully crafted compromise among national interests carefully designed to balance a commitment to common policy-making with the maintenance of national flexibility. For five decades, alternative theoretical views—neo-functionalism, historical institutionalism, constructivism, federalism, post-modernism, Habermasian discourse theory, supranational agenda-setting theory, theories of supranational-subnational alliances—have sought to show that unintended consequences and indirect processes dominate this underlying rationalist logic of European integration. None of these has been empirically successful as a general or foundational theory of European integration, though each explains some subordinate part of the puzzle within specific scope conditions. The way forward, then, is for studies of informal governance in the EU to be explicitly multi-causal, both in the sense of involving tests among alternative theories, studies more carefully designed to ascertain the scope conditions of claims, and efforts to develop creative new syntheses among theories. In sum, the research program should go the way of EU studies and the study of international organization in general, namely considering how alternative mechanisms combine to explain a complex reality.38
email@example.com. I am grateful for the research assistance of Justin Simeone in preparing this memo.
2 Mareike Kleine, All Roads Lead Away From Rome: A Theory of Informal Norms (Dissertation, Freie Universität Berlin, 2009).
3 Member states jealously guard their sovereignty and may have rational reasons for resisting further Europeanization—even in its 'soft' OMC guise. Armin Schafer, “Beyond the Community Method: Why the Open Method of Coordination was Introduced to EU Policy-Making”, European Integration Online Papers 8:13 (2004), 1-19.
4 Martin Rhodes, “Employment policy: between efficacy and experimentation', in H. Wallace, W. Wallace and M. Pollack, eds., Policy-making in the European Union (Oxford, 2005), 279-304; Manuele Citi and Martin Rhodes, “New Modes of Governance in the European Union,” in Knud Erik Jørgensen, Mark A. Pollack and Ben Rosamond, eds. Handbook of European Union Politics (Sage, 2007), 469; Andrew Moravcsik, "The European Constitutional Compromise and the Neofunctionalist Legacy," Journal of European Public Policy 12:2 (2005), 366, passim.; Fritz Scharpf, “The European Social Model: Coping with the Challenges of Diversity”, Journal of Common Market Studies 40:4 (2002), 645-670.
5 Top down regulation and uniform application of EU law in areas such as welfare and fiscal policy would not only be illegitimate but “unthinkable” (Scharpf 2002). Elsewhere, in innovation policy, for example, the complexity and uniqueness of national systems render centrally-determined policies dysfunctional. R. Kaiser and H. Prange, “Managing diversity in a system of multi-level governance: The open method of co-ordination in innovation policy”, Journal of European Public Policy 11:2 (2004), 249-266.
6 Even in the EU Constitutional convention, where hundreds of proposals were considered, there were almost no concrete proposals for a European social policy. On the lack of consensus, see Citi and Rhodes: “Neither a compulsory range of tax rates nor a single European tax system has been adopted by EU member states; nor has there been much progress in moving beyond an EU social or employment policy based on intergovernmental bargaining and the standard Community Method…. By contrast with internal market competition policy and the EMU/SGP deficit criterion, negative externalities and the potential for downward policy spirals in other policy areas is actually quite limited. A rule of thumb for assessing the success of novel mechanisms of coordination in the EU is the presence of such externalities. In the absence of a much higher levels of cross-border labour flows, or trends for countries to diminish welfare spending in the search for competitive advantage, or greater evidence than presently exists for widespread tax competition, there are equally few incentives to compel serious national actor engagement in even the softer forms of social and employment policy coordination advanced under the EES/OMC.” Citi and Rhodes 2007, 475. See also Scharpf 2002; Rhodes 2005; Moravcsik 2005.
7 Empirical research shows this is not a function of enlargement. Voting patterns do not track old and new state splits. It is a function of a lack of consensus in new, redistributive issue areas. There remains, of course, incremental movement within existing constitutional mandates, which blurs the lines. See, for example, progress in gender equality, asylum, services deregulation, and foreign policy. This does not undermine the point.
8 The best single study is Jonathan Zeitlin, Philippe Pochet, eds., with Lars Magnusson, The Open Method of Co-ordination in Action (P.I.E.-Peter Lang, 2005), written by three authors quite sympathetic to OMC. Yet, even though the studies are not controlled, do not test theories systematically, and do not use quantitative methods, the authors still conclude with admirable honesty that little evidence of policy change exists. Héritier points to a consistent trade-off between the level of member state consensus and the formulation of precise targets. A. Heritier, “New Modes of Governance in Europe: Increasing Political Capacity and Policy Effectiveness?” in T. Börzel and R. Cichowski, eds., State of the European Union VI: Law, Politics, and Society (Oxford, 2003), 119. In the area of pension policy, member states have achieved little standard-setting, information-gathering, and behavior modification. Martin Lodge, “Comparing Non-Hierarchical Governance in Action: The Open Method of Co-ordination in Pensions and Information Society”, Journal of Common Market Studies 45 (2007), 343-365.
9 In the peer review mechanisms of the most-developed OMC, the European Employment Strategy, heavy political and social costs are “routinely referred to” in surveys to explain why 'best practices' are so hard to identify and transfer. B.H. Casey and M. Gold, “Peer review of labour market programmes in the European Union: what can countries really learn from one another?”, Journal of European Public Policy 12:1 (2005), 23-43. Other scholars have emphasized the importance of policy “misfit.” See, for example, Mikkel Mailand, “North, South, East, West: the implementation of the European Employment Strategy in Denmark, the UK, Spain, and Poland”, in Martin Heidenriech and Jonathan Zeitlin, eds., Changing European Employment and Welfare Regimes, (Routledge, 2009). From this perspective, state compliance is only a result of pre-existing interests; consequently, member state policy convergence is unattainable. Claire Kilpatrick, “New EU Employment Governance and Constitutionalism”, in Grainne de Burca and Joanne Scott, eds., Law and New Governance in the EU and the US (Hart, 2006), 134.
10 Citi and Rhodes 2007, 469. See also: C.M. Radaelli, “The Europeanization of Public Policy”, in K. Featherstone and C.M. Raedelli, eds., The Politics of Europeanization (Oxford, 2003); Susana Borras and Bent Greve, “Concluding remarks: New method or just cheap talk?” Journal of European Public Policy 11:2 (2004), 329-336; Susana Borras and Kerstin Jacobsson, “The open method of coordination and new governance patterns in the EU”, Journal of European Public Policy 11:2 (2004), 185-208. Scharpf (2002) claims effectiveness in employment or social policy is largely due to policy linkages with Treaty-based legal standards and framework directives. Citi and Rhodes summarize the literature as follows: “There have been few attempts to subject the OMC and other new modes of governance to systematic, variable-based analysis. Nonetheless, existing work does alert us to two critical factors for understanding potential impact: the presence of sanctions and the extent of institutionalization” (2007, 469).
11 See, for example, Radaelli 2003; Borras and Greve 2004; Borras and Jacobsson 2004. Citi and Rhodes observe, “Benchmarks are at their strongest when used to monitor and publicize national indicators and when linked to the potential for 'harder' policy interventions (as with deficit levels in the case of the Stability and Growth Pact.... They have 'moderate strength' when they take the form of non-binding guidelines and indicative targets and are also accompanied by monitoring and reporting requirements and the threat of legislation, as in the European Climate Change Programme. They are at their weakest, as in education policy, where there are no concrete targets for individual countries, but merely 'reference levels' of average European performance. The latter have little real meaning…” (2007, 470).
12 Kathleen McNamara and Sophie Meunier, "The European Constitutional Settlement," in Kathleen McNamara and Sophie Meunier, eds. Making History: European Integration and Institutional Change at 50, State of the European Union, Vol. 8 (Oxford, 2007).
13 Andrew Moravcsik, "Why the European Community Strengthens the State: Domestic Politics and International Institutions," Center for European Studies Working Paper Series 52 (Center for European Studies, 1994).
14 Büchs asserts, “Member state governments know that the OMC objectives are not legally binding, that their national parliaments do not have to agree to them or transpose them into national law and that the EU cannot impose any formal sanctions if the objectives are not achieved. It is therefore no political risk to governments to agree to ambitious objectives…they may even hope to “use” the OMC objective as an additional reform lever.” Milena Büchs, New Governance in European Social Policy: The Open Method of Coordination (Palgrave Macmillan, 2007), 31.
15 Büchs, p. 44 and Chapter 3 passim.
16 Zeitlin and Pochet, with Magnusson 2005, 455.
17 Büchs 2007, 82.
18 When compliance is evident, this often results from the fact that member states may strategically “up-load” and selectively “down-load” preferred policy goals during the benchmark setting phase (Zeitlin 2005: 451-452). Jonathan Zeitlin, “The Open Method of Coordination in Action”, in Zeitlin and Pochet, with Magnusson 2005.
19 Büchs 2007, 31. As Büchs recounts, “Jobelius (2003) analysed how the employment guidelines in 2003 were formulated and concluded that the Commission was not successful in introducing more rigorous quantitative indicators due to opposition by member state governments. This can also be illustrated by the fact that the EU Commission included job counseling into its definition of measures of active labour market policy. This was demanded by the UK government, who strongly opposed the Commission’s view that the UK did not comply with the objective of providing a measure of active labour market policy for young people after six months of unemployment at the latest.... The UK government argued that it actually fulfilled this objective if the EU Commission accepted job counseling as a measure of active labour market policy. Due to this intervention, the UK is now the only EU country reported to be fully complying with these activation objectives” (2007, 30).
20 For example, a 2003 Commission report on The Future ofthe European Employment Strategy recommends both greater flexibility in contractual arrangements and restrictions on the segmentation of the labour market into different categories of workers—inconsistencies that may account for “asymmetric” policy learning across different member states. Milena Buchs, “Asymmetries of policy learning? The European employment strategy and its role in labour market policy reform in Germany and the UK”, ESPAnet Conference, University of Oxford (9-11 September 2004). See also: G. Raveaud, “The European employment strategy: from ends to means”, in R. Salais and R. Villeneuve, eds., Europe and the Politics of Capabilities (Cambridge, 2005); Rhodes 2005; Citi and Rhodes 2007, 476.
21 Büchs 2007, 92.
22 Citi and Rhodes 2007, 474-475.
23 P. DiMaggio and W. Powell, “The iron cage revisited: institutional isomorphism and collective rationality in organizational fields”, in W. Powell and P. DiMaggio, eds., The New Institutionalism in Organizational Analysis (Chicago, 1991).
24 Kerstin Jacobsson, “Soft regulation and the subtle transformation of states: the case of EU employment policy”, Journal of European Social Policy 14:4 (2004), 359-365; D. Trubek, P. Cottrell, and M. Nance, “Hard and Soft Law in European Integration”, in Joanne Scott and Grainne de Burca, eds., New Governance and Constitutionalism (Hart, 2005), 78-80.
25 Casey and Gold 2005.
26 Jacobsson and West contend that the EES promoted compliance via advice, monitoring, and peer review in the Baltic countries. Kerstin Jacobsson and Charlotte West, “Joining the European Employment Strategy: Europeanization of labour market policy making in the Baltic states”, in Martin Heidenreich and Jonathan Zeitlin, eds., Changing European Employment and Welfare Regimes (Routledge, 2009).
27 In regard to the EES, several authors support this claim. See, for example: Zeitlin, in Zeitlin and Pochet, with Magnusson 2005, 454; M. Heidenreich and G. Bischoff, “The Open Method of Coordination: A Way to the Europeanization of Social and Employment Policies?” Journal of Common Market Studies 46 (2008), 497-532. In regard to social policy, some authors observe a “North-South learning direction.” See, for example, Kristin Edquist, “EU social-policy governance: advocating activism or servicing states?” Journal of European Public Policy 13:4 (2006), 500-518.
28 Alberto Alesina and Roberto Perotti, “The European Union: a politically incorrect view”, National Bureau of Economic Research, Working Paper 10342 (2004), http://papers.nber.org/papers/w10342.pdf, pp. 6-7.
29 C.M. Raedelli, “Policy transfer in the European Union: institutional isomorphism as a source of legitimacy”, Governance 13:1 (2000), 25-43; K. McNamara, “Rational fictions: central bank independence and social logic of delegation”, West European Politics 25:1 (2002), 47-76.
30 McNamara 2002; Moravcsik 2005.
31 Jonathan Zeitlin and David Trubek, Governing Work and Welfare in a New Economy (Oxford, 2003), 17. Cohen and Sabel contend, the OMC allows actors to “transform diversity and difference from an obstacle to cooperative investigation of possibilities into a means for accelerating and widening such enquiry.” J. Cohen and C. Sabel, “Sovereignty and solidarity: the EU and the US”, in Jonathan Zeitlin and David Trubek, Governing Work and Welfare in a New Economy (Oxford, 2003), 362-373.
32 Zeitlin and Pochet, with Magnusson 2005.
33 Charles Sabel and Jonathan Zeitlin, “Learning from Difference: The New Architecture of Experimentalist Governance in the EU”, European Law Journal 14:3 (2008), 280.
34 The two cited by Sabel and Zeitlin seem neither necessary nor sufficient. There are cases of OMC learning in which actors are not strategically interdependent.
35 Büchs 2007, 92.
36 Gráinne de Búrca and Joanne Scott, eds. Law and new governance in the EU and the US (Oxford ; Portland, Oregon : Hart Publishing, 2006); Christian Joerges and Ellen Vos, eds. EU committees : social regulation, law and politics Oxford ; Portland, Ore. : Hart, 1999.
37Laffan and Shaw (2005).
38 E.g. Citi and Rhodes 2007, 475; Linsemann, Meyer, and Wessels 2007).