|
|
|
Email this citation
The Path to European Integration:
A Historical Institutionalist Analysis
Center for German and European Studies, University of California at Berkeley
November 1996
Abstract
Many European and American observers of the EC have criticized "intergovernmentalist" accounts for exaggerating the extent of member state control over the process of European integration. This essay seeks to ground these criticisms in a historical institutionalist" account that stresses the need to study European integration as a political process which unfolds over time. Such a perspective highlights the limits of member state control over long-term institutional development. Losses of control result from member state preoccupation with short-term concerns, the ubiquity of unintended consequences, and processes that "lock in" past decisions and make reassertions of member state authority difficult. Brief examination of the evolution of EC social policy suggests the limitations of treating the EC as an institutional "instrument" facilitating collective action among sovereign states. It is more useful to view integration as a path-dependent process that has produced a fragmented but still discernible multi-tiered European polity.
The evolution of the European Community has long fascinated political scientists. For four decades, some of the world's most enduring nation states have conducted an extraordinary political experiment. Progressing sporadically but in a consistent direction, the member states of the European Community have "pooled" increasing areas of policy authority, introducing prominent collective institutions. The creation of these institutions initiated a process which has transformed the nature of European politics.
How the evolution of these arrangements of collective governance can be explained and the nature of the current system understood remain matters of considerable controversy. Within American political science, it has been students of international relations who have maintained the most theoretically-driven discussions of the EC. Despite significant internal disputes, the dominant paradigm in international relations scholarship regards European integration as the practice of ordinary diplomacy under conditions creating unusual opportunities for the providing collective goods through highly institutionalized exchange (Garrett 1992; Moravcsik 1993). From this "intergovernmentalist" perspective the EC is essentially a forum for interstate bargaining. Member states remain the only important actors at the European level. Societal actors exert influence only through the domestic political structures of member states. Policy-making is made through negotiation among member states or through carefully circumscribed delegations of authority. Whether relying on negotiation or delegation, Chiefs of Government ("COGs") are at the heart of the EC, and each member state seeks to maximize its own advantage. Debate within this perspective has concerned such questions as why member states desired certain observed outcomes, which member states have the most influence on collective decision-making, and which alignment of member state interests can best explain policy or institutional development in the EC (Moravcsik 1991; Lange 1993; Martin 1993).
This perspective has not been without its challengers. European scholars have generally depicted the EC as a more complex and pluralistic political structure, less firmly under member state control. Much of this scholarship is not particularly concerned with advancing broad theories of integration, concentrating instead on the detailed investigation of day-to-day policy development in areas where the EC's role is prominent. From this perspective, the Community looks more like a single (if highly fragmented) polity than the site of diplomatic maneuvering among autonomous member states. Within Europe, analyses that treat the European Community as a quasi-federal system -- "an obvious reference point for the European Community" in the words of one prominent analyst (Dehousse 1994, p. 103) -- are now quite common (Scharpf 1988; Majone 1992).
This is equally true within the ranks of comparativists who have turned their attention recently to the European Community (Sbragia 1992; Marks 1993; Anderson 1995; Pierson and Leibfried 1995b). The principal reason for this new interest is revealing: students of a wide range of government activities, including industrial, regional, social, and environmental policies, have found that they can no longer understand the domestic processes and outcomes that interest them without addressing the role of the EC. These investigations also portray a complex and pluralistic political process, not firmly under member state control and not explicable in terms of simple diplomatic bargaining. Coming from the detailed investigation of particular domestic policy arenas to address a strikingly new phenomenon, however, comparativists possessed few theoretical tools that appeared directly applicable. Like European analysts, they have tended to depict the Community as a quasi-federal, "multi-level" or "multi-tiered" political system. Yet these terms are used more to describe the current state of affairs than to explain it. Thus, if a growing body of detailed research reveals considerable unease about the dominant IR models of EC politics, critics have so far had little to offer as an alternative to intergovernmentalist accounts.
In practice, the critics of intergovernmentalism have tried to move forward in two ways. Some have continued to investigate particular policy areas, content to reveal the density and pluralism of actual policymaking while simply observing that the focus of international relations theory on grand diplomacy among sovereign member states does not square with what is actually occurring "on the ground." However, it is almost always possible, ex post, to posit some set of member state preferences that reconciles observed outcomes with the image of near total member state control. Where policy outcomes do not conform to the expected preferences of member states, they may be explained as part of a "nested game" or as an instance of side payments. Drawing on rational choice theory, intergovernmentalism possesses flexible conceptual tools that can "explain" why member states would favor the observed outcomes (Green and Shapiro 1984). 1 Thus, absent a theoretically-based explanation for the constraints on member states, these detailed investigations will not persuade proponents of intergovernmentalism.
More theoretically-oriented critics have drawn on aspects of the "neo-functionalist" tradition in international relations, showing how "spillover" processes and the autonomous actions of supranational actors (including the Commission and European Court of Justice) contribute to European policymaking. Recent efforts to update neo-functionalism have successfully highlighted important limitations in intergovernmentalist accounts, and I will rely in part on these arguments in developing my own analysis. Yet neo-functionalism has serious problems of its own. Given the strong institutional position of member states in the EC, neo-functionalists seem to attribute greater autonomy to supranational actors than can plausibly be sustained. Although neo-tunctionalist arguments about the independent action of the Commission and Court of Justice have some merit, there is little doubt that the member states, acting together in the Council, remain the most powerful decision-makers. In most cases, it seems equally probable that these decision-makers act to secure their own interests, whatever those are deemed to be. Crucially, these "principals" retain the legal authority to rein in their "agents" if they find it in their interests to do so. Thus at any given point in time the key propositions of intergovernmentalist theory are likely to hold.
This essay seeks to lay the foundation for a more persuasive account of member state constraint. My focus is on the reasons why gaps emerge in member state control over the evolution of European institutions and public policies, on why these gaps are difficult to close, and on how these openings both create room for actors other than member states to influence the process of European integration while simultaneously constraining the room for maneuver of all political actors. The basis for this challenge to intergovernmentalism lies in insights from what I will term "historical institutionalism" (Steinmo and Thelen 1992; Ikenberry 1994). The label covers a diverse range of scholarship, much of it with little theoretical focus. Indeed, a principal goal of this essay is to strengthen the theoretical foundations of historical institutionalism. There are, however, two unifying themes within this broad research orientation. This scholarship is historical because it recognizes that political development must be understood as a process that unfolds over time. It is institutionalist because it stresses that many of the contemporary implications of these temporal processes are embedded in institutions -- whether these be formal rules, policy structures or norms. 2
The crucial claim I derive from historical institutionalism is that actors may be in a strong initial position, seek to maximize their interests, and nevertheless carry out institutional and policy reforms that fundamentally transform their own positions (or those of their successors) in ways that are unanticipated and/or undesired. Attempts to cut into on-going social processes at a single point in time produce a "snapshot" view that is distorted in crucial respects. Central parts of my analysis emphasize temporal aspects of politics: the lags between decisions and long-term consequences, and the constraints that emerge from societal adaptations and shifts in policy preferences that occur during the interim. When European integration is examined over time, the gaps in member state control appear far more prominent than they do in intergovernmentalist accounts.
In contrast to the functional account of institutions which underpins intergovernmentalism, historical institutionalism stresses the difficulties of subjecting institutional evolution to tight control. Two brief historical examples can illustrate the broad point explored in this essay. The first concerns the changing institutional position of state governments in the United States (Riker 1955). Because approval of the American constitution required state ratification, the interests of states received considerable attention in the process of institutional design. The framers intended the Senate to serve as a strong support of state interests. In an arrangement that partly echoes the EC's emphasis on member state participation in collective deliberations, state legislatures were to appoint senators, who were expected to serve as delegates representing states in the formation of policy. Over time, however, Senators seeking greater autonomy were able to gradually free themselves from state oversight. By the early 1900s, the enactment of the 17th Amendment requiring popular election of Senators only ratified the result of a lengthy erosion of state legislative control.
The development of Canadian federalism provides another example (Watts 1987). The designers of the Canadian federation sought a highly centralized form of federalism -- in part as a reaction to the ways in which decentralization contributed to the horrors of the American Civil War. Yet the Canadian federation is now far less centralized than the American one. Among the reasons: the Canadian federation left to the provinces sole responsibility for many activities that were then considered trivial. With the growing role of government in social policy and economic management, however, these responsibilities turned out to be of tremendous importance.
In both these cases, the current functioning of institutions cannot be derived from the aspirations of the original designers. Processes evolving over time led to quite unexpected outcomes. Similarly, I will argue that what one makes of the EC depends on whether one examines a photograph or a moving picture. Just as a film often reveals meanings that cannot be discerned from a single photograph, a view of Europe's development over time gives us a richer sense of the nature of the emerging European polity. At any given time, the diplomatic maneuvering among member states looms large, and an intergovernmentalist perspective makes considerable sense. Seen as a historical process, however, the scope of member state authority appears far more circumscribed, and both the interventions of other actors and the cumulative constraints of rule-based governance more considerable.
My argument is developed in three stages. In the first, I review the main features of intergovernmentalist analyses of the EC. In the second, I develop a historical institutionalist critique. In section three, I briefly apply these historical institutionalist arguments to one aspect of European integration, the development of social policy. This application is hardly intended as a full test of my approach. Nonetheless, historical institutionalism's applicability in an area where intergovernmentalist analysis ought to be on strong ground provides further evidence of its theoretical promise.
The accelerated activity of the EC in the past decade coincided with a growing focus among international relations scholars on international regimes, which were conceptualized as institutionalized forms of collective action among nation-states (Krasner 1983; Keohane 1984; Haggard and Simmons 1987). While some analysts of European integration have continued to echo the earlier international relations literature on neo-functionalism, the dominant intergovernmentalist perspective has treated the EC as a standard (albeit unusually well-developed and multi-faceted) international regime. It would be unrealistic to attempt a thorough review of this diverse and sophisticated literature here. Instead, I focus on three core features of intergovernmentalism: the emphasis on member state preoccupation with sovereignty; the depiction of institutions as instruments; and the focus on "grand bargains" among member states.
The Centrality of Sovereignty.
Intergovernmentalism itself generally takes member state preferences as given, focusing instead on how member states seek to pursue those preferences.
3
Yet despite this apparent openness, intergovernmentalist accounts tend to stress member state preoccupation with preserving sovereignty. As Keohane maintains, "...governments put a high value on the maintenance of their own autonomy, [so] it is usually impossible to establish international institutions that exercise authority over states" (Keohane 1984, p.88).
Of course, much of the writing on international regimes, of course, arose as a reaction against realist perspectives that were seen as putting too much weight on sovereignty concerns --suggesting that collective action among states should almost never be possible. Regime theorists have argued that in contexts where security concerns have diminished, nation states may care about absolute gains as well as relative ones. Nonetheless, the realist focus on sovereignty carries over into intergovernmentalist treatments of the EC. Most intergovernmentalist analyses suggest that member state preferences are heavily weighted towards preserving sovereignty, leading Chiefs of Government to be vigilant guardians of national autonomy in evaluating proposals for international cooperation. The issue is often posed in principal-agent terms (Garrett 1992). The principals (member states) may delegate certain responsibilities to agents (international institutions), but only with the strictest oversight. The core calculation for member states is whether the benefits of collective action outweigh any possible risk to autonomy. According to Moravcsik, "[i]n the intergovernmentalist view, the unique institutional structure of the EC is acceptable to national governments only insofar as it strengthens, rather than weakens, their control over domestic affairs..." (Moravcsik 1993, p. 507, emphasis added).
The Instrumentality of institutions.
Work on international regimes has drawn heavily on the insights of Transaction Cost Economics (TCE), which analyzes institutions in functional terms (Williamson 1975; Keohane 1984; North 1990). As Moravcsik summarizes, "modern regime theory views international institutions as deliberate instruments to improve the efficiency of bargaining between states" (Moravcsik 1993, p.507). Drawing on sophisticated work in game theory and economic theories of organizations; intergovernmentalists note that collective action among autonomous nation states is often desired yet enormously difficult. A critical issue concerns problems of information. Uncertainty about the preferences, intentions, and reliability of other actors makes agreements difficult to execute and enforce. Institutions can help surmount these problems, reducing information asymmetries, monitoring compliance, and creating linkages across issues that diminish the likelihood of defection. According to Keohane:
"Far from being threats to governments (in which case it would be hard to understand why they exist at all), they permit governments to attain objectives that would otherwise be unattainable. They do so in part by facilitating intergovernmental agreements. Regimes facilitate agreements by raising the anticipated costs of violating others' property rights, by altering transaction costs through the clustering of issues, and by providing reliable information to members. Regimes are relatively efficient institutions, compared with the alternative of having a myriad of unrelated agreements, since their principles, rules, and institutions create linkages among issues that give actors incentives to reach mutually beneficial agreements" (Keohane 1984, p.97).In intergovernmentalist accounts, self-conscious, maximizing actors (member states) create institutions because these institutions help them surmount collective action problems and achieve gains from exchange. The best way to understand the development of international institutions is to identify the functions that they fulfill, especially the lowering of bargaining costs and the reduction of uncertainty through the provision of "a forum and vocabulary for the signalling of preferences and intentions" (Stone 1994, p.456).
The Centrality of Intergovernmental Bargains.
Students of the EC frequently distinguish between the intermittent "grand bargains" (e.g. the Treaty of Rome, the Single European Act, Maastricht) that establish basic features of institutional design and the "day-to-day" policy-making in the Community that occurs between these agreements. For intergovernmentalists, the grand bargains are where the action is. Since, as Moravcsik puts it, "functional regime theory view[s]... international institutions as passive, transaction-cost reducing sets of rules", it is the design of those rules that is central. The EC, he adds, "has developed through a series of celebrated intergovernmental bargains, each of which sets the agenda for an intervening period of consolidation. The most fundamental task facing a theoretical account of European integration is to explain these bargains" (Moravcsik 1993, p.508, 473, emphasis added). The intergovernmentalist research agenda clearly reflects this line of thinking, focusing overwhelmingly on explaining aspects of two grand bargains: the Single European Act (Moravcsik 1991; Garrett 1992) and the Maastricht Treaty (Garrett 1993; Lange 1993; Martin 1993). Political developments during the periods between these bargains, or that concern matters that are not hotly-contested during those bargains, have received almost no attention.
These three aspects of intergovernmentalist accounts are closely connected. The depiction of member states as profoundly concerned about sovereignty contributes to a functional view of regimes. Given the preoccupation with sovereignty, the institutional underpinnings for cooperation will only be created or extended after a careful weighing of long-term costs and benefits. The "benefits" are the transaction-cost reducing functions that regimes perform, while the costs often relate to any risk of lost autonomy. Similarly, the emphasis on member state bargains follows logically from the functional analysis of institutions. If the EC is an international regime in which member states have carefully designed passive instruments to allow them to carry out collective goals, periods of "consolidation" are of little interest. It is the bargains themselves that create or change the rules of the game, and that therefore demand attention. The "post-bargain" period simply plays out the implications intended in the grand bargains. Together, these three positions have contributed to a powerful argument about the process of European integration. As I suggest in the next section of this essay, however, all three are open to serious challenge.
"Historical institutionalism" is a loose term covering a range of scholarship that has tried to combine social science concerns and methods with a recognition that social processes must be understood as historical phenomena (Steinmo and Thelen 1992; Ikenberry 1994). In my own usage, historical institutionalism cuts across the usual sharp dichotomy between rational choice and non-rational choice work, drawing instead on research within both traditions that emphasizes the significance of historical processes. Thus it includes rational choice analyses that consider issues of institutional evolution and path dependence crucial (North 1990; Knight 1992). It excludes much "historical" research in political science that uses history only as a technique for widening the universe of available cases.
The core arguments of historical institutionalism contrast with a more common view in the social sciences, which, as March and Olson observe, assumes (often implicitly) that "institutions and behavior... evolve through some form of efficient historical process. An efficient historical process ... is one that moves rapidly to a unique solution, conditional on current environmental conditions, and is independent of the historical path" (March and Olson 1989, pp.5-6). Given this orientation, Theda Skocpol notes, "[a]nalysts typically look only for synchronic determinants of policies -- for example, in current social interests or in existing political alliances. In addition, however, we must examine patterns unfolding over time..." (Skocpol 1992, p.58).
Recent research focusing on institutional evolution and path dependence has challenged the expectation that institutions embody the long-term interests of those responsible for original institutional design (Krasner 1989; North 1990). Where the legal authority of the institutional designers is as unquestionable as that of the member states in the EC, I will argue that such a challenge must be based on two sets of claims. First, there must be an account of why "gaps" -- by which I mean significant divergences between the institutional and policy preferences of member states and the actual functioning of institutions and policies -- would emerge. Second, critics must explain why, once such gaps emerge, they cannot reliably be closed. One can find scattered elements of such accounts in recent theoretical treatments of institutional change. When brought together, they provide a compelling response to the claim that institutional development in the European Union can be understood in functional terms.
I focus first on the factors which are likely to create considerable gaps in member state control. Four are of fundamental importance: the autonomous actions of European institutional actors, the restricted time horizons of decision-makers, the large potential for unintended consequences, and the likelihood of changes in COG preferences over time. Each of these factors requires more detailed discussion.
The Partial Autonomy of EC Institutions.
The main contribution of recent neo-functionalist analysis has been to emphasize the autonomous role of supranational actors, especially the Commission and the Court (Sandholtz 1992; Burley and Mattli 1993; Ross 1994). I begin by summarizing these arguments and suggesting why, by themselves, they constitute an inadequate response to intergovernmentalism.
The central objections raised by neo-functionalists can be cast in terms of the same principal-agent framework used in many intergovernmentalist accounts. Member states created the European Community, and they did so to serve their own purposes. In order to carry out collective tasks, however, the member states felt compelled to create new institutions. As Terry Moe has argued, the results are predictable:
A new public agency is literally a new actor on the political scene. It has its own interests, which may diverge from those of its creators, and it typically has resources -- expertise, delegated authority -- to strike out on its own should the opportunities arise. The political game is different now: there are more players and more interests to be accommodated (Moe 1990, p.121).In the European context, the member states' problem has been especially difficult. They have needed to create arrangements that would allow reasonably efficient decision-making and effective enforcement despite the involvement of a large number of governments with differing interests, and despite the need for decision-making, implementation, and oversight on a wide range of complex and tightly-coupled policy arenas. These considerations generated pressure to grant those who run these institutions considerable authority. Thus, the political organs of the EC are not without resources; as a result, they are not simply passive tools of the member states.
Over time, EC organizations will seek to use grants of authority for their own purposes, and especially to increase their autonomy. They will try to expand the gaps in member state control, and they will use any accumulated political resources to resist efforts to curtail their authority. The result is an intricate, on-going struggle that is well-known to students of the European Union but would also be familiar to American observers of; say, relations between congressional committees and administrative agencies (McCubbins and Schwartz 1984; Moe 1987; Kiewiet and McCubbins 1991). Member states generally (but not always) seek to rein in EC institutions. They recognize, however, that these crucial collective organizations cannot function without significant power, and that the authority required must grow as the tasks addressed at the European level expand and become more complex.
For their part, European institutions such as the Commission, the European Court of Justice, and the European Parliament are always looking for opportunities to enhance their powers. Neo functionalist analyses have emphasized the significant successes of these supranational actors. The Council, to be sure, continues to stand watch over proposed legislation, and actively protects member state interests. Yet the Commission, Parliament and Court possess considerable ability to advance their own interests. For the Commission, two assets are particularly important (Ross 1994). The first concerns the setting of agendas, a source of influence which it frequently shares with the European Parliament (Tsebelis 1994; Garrett and Tsebelis 1995). Choosing which proposals to consider is a tremendously important (if frequently unappreciated) aspect of politics, and here European institutional actors often have primacy. Obviously, this power is far from unlimited; the Commission cannot expect to pass proposals that ignore the preferences of member states. Usually, however, it will have some room for maneuver (Pollack 1994; Pollack 1995). Entrepreneurial European actors, such as the Delors Commission, may be able to frame issues, design packages, and structure the sequence of proposals in ways that maximize their room for independent initiative (Riker 1986). The expansion of qualified majority voting has widened the range of possible "winning coalitions," further increasing the agenda-setting powers of the Commission and Parliament. Neo-functionalists have argued persuasively that the Commission's effective use of agenda-setting powers has advanced European integration and increased its own role in policy reform (Ross 1994).
The Commission's second major asset is its role as what Volker Eichener calls "process manager" (Eichener 1993; Peters 1992). Policy-making at the EC level, as many have noted, is heavily tilted toward regulation -- a type of policy-making with its own distinctive qualities (Lowi 1964; Majone 1989). The development of complex social regulations requires the assembly and coordination of dense networks of experts. This task falls to the Commission, and with it comes additional room for influence. Especially in the labyrinths of regulatory policy-making, this role may give the Commission significant power.
The political resources of the European Court are at least as significant. If the United States in the nineteenth century had a "state of courts and parties" (Skowronek 1982), the EC looks at times like a "state of courts and technocrats" (Leibfried 1992, p.249). In the process of European integration, the European Court has taken an active, even forcing stance, gradually building a remarkable base of authority and effectively "constitutionalizing" the emerging European polity (Weiler 1991; Burley and Mattli 1993; Alter 1996). The Court has more extensive powers of judicial review than most of its national counterparts, and fewer impediments to action than other EC decision-making bodies. If the Council is prone to gridlock, the necessity of deciding cases inclines the ECJ to action. This inclination is strengthened by rules allowing simple majority decisions, and by a secrecy (neither actual votes nor dissenting views are made public) that shelters judges from member state and popular pressures. ECJ judges also share a common professional background, legal culture (at least on the continent), and sense of mission that seems to effectively limit the influence of the member states in judicial decision-making.
Neo-functionalist accounts of these supranational institutions have certainly demonstrated their prominent role in the EC, as even some intergovernmentalists have acknowledged. 4 Yet the true influence of the Court, Commission, and Parliament on policymaking and future institutional development remains uncertain. Do these organizations create genuine gaps in member state control, or do they simply act as agents, fulfilling monitoring, information-gathering, and implementation roles under tight member state scrutiny? As Lisa Martin, among others, has suggested, autonomy may be more apparent than real:
Politicians and academic observers often infer from such a pattern [of activity] autonomy of the Commission and/or of government leaders. However, consideration of institutional constraints leads us to examine delegation of authority because of the costs of exercising tight control over agents, an optimal structure of delegation may be one with little active oversight or overt interference in the negotiating process from principals. Agents rationally anticipate the responses of those they represent. The law of anticipated reactions suggests that we cannot infer a lack of political influence from a lack of observed oversight activity. (Martin 1993, p.135)Thus, what appears to be autonomy may simply reflect the principals' deft use of oversight. Relying on the disciplining power of anticipated reactions and the use of "fire alarms" -- signals derived from reporting requirements or interest group montoring activity -- to identify significant problems, member states can stay in the background while remaining firmly in charge (McCubbins and Schwartz 1984).
Again, given the ease of assembling plausible ex post accounts of why given outcomes served member state interests, these arguments about delegation are difficult to refute, although they are equally difficult to demonstrate (Garrett 1995; Burley and Mattli 1995). To foreshadow a point pursued at length below, the intergovernmentalist claim that supranational actors are agents rather than autonomous actors is strengthened if we believe that member states can react powerfully to observed losses of control. If the Commission, Court and Parliament anticipate that their efforts to produce or exploit gaps will be detected, punished, and reversed, they are indeed unlikely to strike out on their own. Thus a crucial problem with neo-functionalism is that it lacks a coherent account of why the threat of such a member state reaction is not always credible. I address this problem in a moment.
Before proceeding to that issue, however, the case for constraints on member state control can be greatly strengthened if other sources of gaps can be identified. Here, the historical institutionalist focus on the temporal dimension of politics is invaluable. It highlights three additional sources of gaps: the short time-horizons of decision-makers, the prevalence of unanticipated consequences, and the prospect of shifting member state policy preferences.
The Restricted Time Horizons of Political Decision-makers.
A statement attributed to David Stockman, Reagan's budget director, is unusual among political decision-makers only for its candor. Asked by an advisor to consider pension reforms to combat social security's severe long-term financing problems, he dismissed the idea out of hand, exclaiming that he had no interest in wasting "a lot of political capital on some other guy's problem in [the year] 2010..." (quoted in Greider 1982, p.43).
Many of the implications of political decisions -- especially complex policy interventions or major institutional reforms -- only play out in the long-run (Garrett and Lange 1994). Yet political decision-makers are frequently most interested in the short-term consequences of their actions; long-term effects are often heavily discounted. The principal reason is the logic of electoral politics. Keynes once noted that in the long-run we are all dead; for politicians in democratic polities, electoral death can come much faster. Since the decisions of voters that determine political success are taken in the short-run, politicians are likely to employ a high discount rate. They have a strong incentive to pay attention to long-term consequences only if these become politically salient, or when they have little reason to fear short-term electoral retribution.
The gap between short-term interests and long-term consequences is often ignored in arguments about institutional design and reform. As a number of critics have noted, choice-theoretic treatments of institutions often make an intentionalist or functionalist fallacy, arguing that the long-term effects of institutions explain why decision-makers introduce them (Bates 1987; Knight 1992; Hall and Taylor 1994). Instead, long-term institutional consequences are often the by-products of actions taken for short-term political reasons. The evolution of the Congressional committee system in the United States -- a central institutional feature of contemporary American governance -- is a good example. As Kenneth Shepsle notes, Henry Clay and his supporters introduced the system to further their immediate political goals without regard to long-term consequences: "[T]he lasting effects of this institutional innovation could hardly have been anticipated, much less desired, by Clay. They were by-products (and proved to be the more enduring and important products) of self-interested leadership behavior" (Shepsle 1989, p. 141). In this case, the system's long-term functioning was not the goal of the actors who created it. By the same token, the reasons for the institution's invention cannot be derived from an analysis of its long-term effects.
Recognizing the importance of policy-makers' high discount rates raises a challenge for intergovernmentalist theories of the European Community. As noted above, most international relations approaches to European integration stress the tenacity with which nation states cling to all aspects of national sovereignty. The design of collective institutions is assumed to reflect this preoccupation. Yet in democratic polities, sustained power requires electoral vindication. Under many circumstances, the first concern of national governments is not with sovereignty per se, but with creating the conditions for continued domestic political success. By extension, where the time-horizons of decision-makers are restricted, functional arguments that are central to transaction-cost views of international regimes also come into question. Rather than being treated as the goals of policymakers under such circumstances, long-term institutional effects should often be seen as the by-products of their purposive behavior.
Unanticipated Consequences.
Gaps in member state control occur not only because long-term consequences tend to be heavily discounted. Even if policymakers do focus on long-term effects, unintended consequences are likely to be widespread. Complex social processes involving large number of actors always generate elaborate feedback loops and significant interaction effects which decision-makers cannot hope to fully comprehend (Hirsch 1977; Schelling 1978; Van Parijs 1982; Perrow 1984; Jervis 1993). While social scientists possess limited tools for dealing with such outcomes, many models -- such as core neo-classical arguments about the dynamics of market systems -- are based on them.
Unanticipated consequences are likely to be of particular significance in the European Union because of the presence of high issue density (Pierson and Leibfried 1995a). In sharp contrast to any existing international organization, the range of decisions made at the European level runs almost the full gamut of traditionally "domestic" issues, from the setting of agricultural prices to the regulation of auto emissions and fuel content to the negotiation of international trade agreements. In the past decade, there has been a massive expansion of EC decision-making, primarily because of the single market project. The sheer scope of this decision-making limits the ability of member states to firmly control the development of policy.
The profound implications of expanded policy activity need to be underlined. As the number of decisions made and the number of actors involved grows, interactions -- among actors and among policies -- increase geometrically (Beer 1978). Figure 1 offers a simple illustration. The connections between the points could represent either relationships among actors or among governmental activities. With 2 points (Figure la), there is only one connection; with an expansion to four points (Figure lb), there are six connections; with an expansion to eight points (1c), there are 28. This is the kind of process that has been underway in the EC.
As has already been discussed, problems of overload are especially consequential when member states must contend with supranational organizations which are eager to extend their authority. In the development of complex regulatory judgments and the legal determination of what previous decisions actually require, essential policy-making authority is often in the hands of bodies of experts, where the Commission plays a crucial role, or in the hands of the Court. This is, of course, one of the central insights of principal-agent theory. Agents can use their greater information about their own activities and the requirements connected to their work to achieve autonomy from principals. Asymmetrical access to information, which is ubiquitous in complex decision-making processes, provides a foundation for influence (Moe 1984).
The second consequence of issue density is the oft-cited process of spillover: the tendency of tasks adopted to have important consequences for realms outside those originally intended, or to empower actors who generate new demands for extended intervention (Haas 1958). One of the key arguments in much writing on contemporary political economies stresses precisely the embeddedness of economic action within networks of tightly-coupled social and political institutions (Hall 1986; North 1990; Garrett and Lange 1994). Efforts to integrate some aspects of complex modern societies without changing other components may prove problematic because the sectors to be integrated cannot be effectively isolated. The more "tightly coupled" government policies are, the more likely it is that actions in one realm will have unanticipated effects in others (Perrow 1984). McNamara, for example, has demonstrated the significance of such interaction effects in the cases of monetary and agricultural policies (McNamara 1993). Similar connections between the single market initiative and social policy development have also been documented (Leibfried and Pierson 1995). As the density of EC policymaking increases, such interaction effects become more prevalent, unintended consequences multiply, and the prospect of gaps in member state control will grow.
Shifts in COG Policy Preferences.
Intergovernmentalist theories tend to treat the institutional and policy preferences of the member states as essentially fixed. This is one of a number of crucial respects in which intergovernmentalism involves a too-easy translation from the world of economic organizations to the world of politics. It may make some sense to assume stable preferences when studying firms, or even when one discusses the enduring issues of grand diplomacy. However, as one moves from traditional foreign policy issues such as national security toward the traditionally "domestic" concerns where the EC has become quite significant, this becomes a more dubious premise.
The policy preferences of member states may shift for a number of reasons. Altered circumstances or new information may lead governments to question previous arrangements. Equally important, changes in government occur frequently, and governments of different partisan complexions often have quite distinct views on policy matters dealt with at the EC level. Governments come and go. Each inherits a set of arrangements from the past; each tries to place its own imprint on this heritage. The result, over time, is that evolving arrangements will diverge from the intentions of original designers, while any newly-arriving COG is likely to find institutional and policy arrangements considerably out of synch with its own preferences.
Thus, there are a number of reasons why gaps in member state control are likely to emerge. Two general points about these sources of gaps deserve emphasis. First, most of these processes have a temporal quality that makes them invisible to a synchronic analysis of institutional and policy choice. The role of restricted time horizons, unintended consequences, and shifting member state preferences will only be evident if we examine political processes over time. Second, most of the processes highlighted are much more likely to be prevalent in the European Community than in the more purely "international" settings which were the subject of original efforts to develop and refine regime theory. Because many of the more "domestic" issues which the EC considers have significant electoral implications, the time-horizons of decision-makers are likely to be shorter. Unanticipated consequences are also more prevalent, because unlike a typical international regime, the EC deals with many tightly-coupled issues. Electoral turnover is more likely to cause shifts in COG preferences on the more domestic issues which the EC considers than on the traditional diplomatic agenda of most international regimes. In short, the EC's focus on core concerns of traditional domestic politics makes it more prone to all the sources of gaps in member state control which historical institutionalism identifies.
At this point, however, the claim of member state constraint is incomplete. Transaction cost approaches are compatible with the possibility of at least some sorts of gaps, although these are rarely addressed in practice. After all, while it has not emphasized unanticipated consequences (Williamson 1993, p. 116), TCE is based in large part on how uncertainty about future events provokes particular organizational responses. It is not enough to demonstrate that gaps emerge; one must also show that once such losses of control take place they often cannot be corrected.
For intergovernmentalists, even where the possibility of gaps is acknowledged, these losses of control are considered theoretically unproblematic. Should outcomes occur which principals do not desire, TCE describes two routes to restored efficiency: competition and learning. 5 Competitive pressures in a market society mean that new organizations with more efficient structures will develop, eventually replacing suboptimal organizations. Learning processes among principals can also lead to correction. According to Williamson, one can rely on
"...the 'far-sighted propensity' or 'rational spirit' that economics ascribes to economic actors Once the unanticipated consequences are understood, these effects will thereafter be anticipated and the ramifications can be folded back into the organizational design. Unwanted costs will then be mitigated and unanticipated benefits will be enhanced. Better economic performance will ordinarily result" (Williamson 1993, pp.116-17).Both these corrective mechanisms, however, are of limited applicability when one shifts from Williamson's focus on firms in private markets to the world of political institutions (Moe 1984; Moe 1990). This is clearest for the mechanism of competition. Political institutions rarely confront a dense environment of competing institutions which will instantly capitalize on inefficient performance, swooping in to carry off an institution's "customers" and drive it into "bankruptcy." Political environments are typically more "permissive" (Krasner 1989; Powell and DiMaggio 1991). Within Europe, there is nothing like a marketplace for competition among international regimes, in which new market entrants can demonstrate that their efficiency (however that might be defined and measured) is greater than the EC's.
While arguments based on competition are weak, learning arguments would appear to be more applicable to political environments. Indeed, Gary Marks, who has pointed to the significance of unanticipated consequences in limiting member state control, concedes that the use of such arguments "is tricky in the context of ongoing political relationships where learning takes place" (Marks 1993, p.403). The process through which actors "learn" about gaps in control and how to address them has received little attention (McCubbins and Schwartz 1984). However, at least on the biggest issues, intergovernmentalists can reasonably assert that member states will gradually become aware of undesired or unanticipated outcomes, and will become more adept at developing effective responses over time. Learning thus seems to offer an effective mechanism for closing gaps and returning institutional and policy designs to an "efficient" (from the point of view of the member states) path.
Yet the efficacy of learning argument depends crucially on the capacity of member states to fold new understandings "back into the organizational design." Put differently, once gaps appear and are identified, how easy is it for principals to regain control? Here the distinction between economic and political institutions becomes crucial. In economic organizations, owners (or principals) may face few barriers to such efforts. In the political world, however (and the case of the EC in particular), incorporating new understandings into institutions and policies is no simple task The next stage of the argument, then, is to consider why gaps, even when identified, might be hard to close. There are three broad reasons: the resistance of EC institutional actors; the institutional obstacles to reform within the EC, and the sunk costs associated with previous actions. If these barriers are sufficiently high, learning will not provide a sufficient basis for correction, and member state control will be constrained.
The Resistance of Supranational Actors.
To the extent that neo-functionalism has had an implicit argument about the difficulty of closing gaps, it has centered on supranational actors. The Court, Commission and Parliament have accumulated significant political resources. They can be expected to use these resources to resist member state efforts to exercise greater control over their activities. Yet neo-functionalism has failed to address the question of why, in an open confrontation between member states and supranational actors, the latter could ever be expected to prevail. Member states, after all, have substantial oversight powers, along with control over budgets and appointments. More fundamentally, they possess the legal authority to determine (and alter) the basic rules of the game, including those affecting the very existence of the EC's supranational organizations. The resources of the Court, Commission and Parliament, such as the capacity to play off one member state against another in the agenda-setting process and perhaps exploit information asymmetries, are not trivial, but they are clearly modest by comparison. A persuasive account of member state constraint must draw on more than the political resources of supranational actors.
Institutional Barriers to Reform.
The efforts of principals to reassert control will be facilitated if they can easily redesign policies and institutions. In the economic realm, principals are generally in a strong position to remake their organizations as they choose. Lines of authority are clear, and the relevant decision-makers are likely to share the same broad goal of maximizing profits. In politics, however, the temporal dimension raises distinct problems. Political decision-makers know that continuous institutional control is unlikely. This lack of continuous control has implications both for how institutions are designed and for the prospects of changing institutions once they are created. In particular, those designing institutions must consider the likelihood that future governments will be eager to overturn their designs, or to turn the institutions they create to other purposes. As Moe notes, the designers of institutions "... do not want 'their' agencies to fall under the control of opponents. And given the way public authority is allocated and exercised in a democracy, they often can only shut out their opponents by shutting themselves out too. In many cases, then, they purposely create structures that even they cannot control" (Moe 1990, p.125).
Thus, political institutions are often "sticky" -- specifically designed to hinder the process of institutional and policy reform. This is, of course, far more true of some national polities than others (Weaver and Rockman 1993). Yet the barriers in most national political systems pale in comparison to the obstacles present in the EC. In principle, the member states decide: they have the authority, if they so choose, to reform or even abolish the Court, Commission, or Parliament. But in fact, the rules of the game within the Community were designed to inhibit even modest changes of course. The same requirements that makes initial decision-making difficult also makes previously-enacted reforms hard to undo, even if those reforms turn out to be unexpectedly costly or to infringe on member state sovereignty.
Efforts to employ the most radical vehicle of institutional redesign, a Treaty revision, face extremely high barriers: unanimous member state agreement, plus ratification by national parliaments and (in some cases) electorates. Given the chances for disagreements among COGs, let alone the problems connected to ratification, the chances of achieving such an extraordinary degree of consensus are generally quite low. Use of this process is now widely recognized to be extraordinarily difficult and unpredictable. As Mark Pollack notes, "the threat of Treaty revision is essentially the 'nuclear option' -- exceedingly effective, but difficult to use -- and is therefore a relatively ineffective and non-credible means of member state control" (Pollack 1995, p.30).
Efforts to produce more modest changes in course confront more modest hurdles, but these remain far tougher than the obstacles facing, for example, a Congressional committee trying to rein in a rogue federal agency. Member states will often be divided on significant issues, but in many policy areas change require a unanimous vote of the member states. In other cases, Qualified Majority Voting is the rule. This makes reform easier, but the standard -- roughly 5/7ths of the weighted votes of member states -- still presents a threshold that is considerably tougher to cross than that required in most democratic institutions (Pollack 1994; Pollack 1995).
The extent to which these barriers constrain member states has recently been questioned. Where it was once understood that participation in the EC was an all-or-nothing proposition, Maastricht has enhanced the prospects for a Europe "a la carte," or a Europe of "variable geometries." Britain and Denmark received opt-outs on monetary union; the eleven other member states circumvented the British veto by opting "up and out" with the Social Protocol. As Jeff Anderson summarizes the new situation, "[Maastricht] and attached protocols established an important precedent, opening the door to a multitrack Europe in which the treaties and resulting secondary legislation do not apply uniformly to each member" (Anderson 1995, p.449). This new flexibility, however, refers only to additional treaty obligations. Member state governments may be able to obtain opt-outs from future treaty provisions. Unless they succeed in navigating the difficult EC decision rules for reversing course, however, they are not free to review and discard the commitments of previous governments, even if those earlier governments were pre-occupied by short-term goals, had quite different policy preferences, or acted in ways that produced many unanticipated consequences. And as new policies are enacted, the scope of this restrictive acquis communautaire continues to grow.
The rules governing institutional and policy reform in the European Community create what Fritz Scharpf calls a "joint-decision trap", making member state efforts to close gaps in control highly problematic (Scharpf 1988). The extent of the institutional obstacles will vary from issue to issue. Obviously, if the benefits of acting are high enough, member states will be able to act. But often the benefits must be quite high. In shutting out their potential successors, COGS have indeed shut themselves out as well.
Sunk Costs and the Rising Price of Exit.
The evolution of EC policy over time may constrain member states not only because institutional arrangements make a reversal of course difficult when member states discover unanticipated consequences or their policy preferences change. Individual and organizational adaptations to previous decisions may also generate massive sunk costs that make policy reversal unattractive. When actors adapt to the new rules of the game by making extensive commitments based on the expectation that these rules will continue, previous decisions may "lock-"in member states to policy options that they would not now choose to initiate. Put another way, social adaptation to EC institutions and policies drastically increases the cost of exit from existing arrangements for member states. Rather than reflecting the benefits of institutionalized exchange, continuing integration could easily reflect the rising costs of "non-Europe."
Recent work on path dependence has emphasized the ways in which initial institutional or policy decisions -- even suboptimal ones -- can become self-reinforcing over time (Krasner 1989; North 1990). These initial choices encourage the emergence of elaborate social and economic networks, greatly increasing the cost of adopting once-possible alternatives and therefore inhibiting exit from a current policy path. Major initiatives have major social consequences. Individuals make important commitments in response to government actions. These commitments, in turn, may vastly increase the disruption caused by policy shifts or institutional reforms, effectively "locking in" previous decisions (Pierson 1992; Pierson 1993).
Work on technological change has revealed some of the circumstances conducive to path dependence (David 1985; Arthur 1988; Arthur 1989). The crucial idea is the prevalence of increasing returns, which encourage a focus on a single alternative and continued movement down a specific path once initial steps are taken. Large set-up or fixed costs are likely to create increasing returns to further investment in a given technology, providing individuals with a strong incentive to identify and stick with a single option. Substantial learning effects connected to the operation of complex systems provide an additional source of increasing returns. Co-ordination effects occur when the individual receives increased benefits from a particular activity if others also adopt the same option. Finally, adaptive expectations occur when individuals feel a need to "pick the right horse" because options that fail to win broad acceptance will have drawbacks later on. Under these conditions, individual expectations about usage patterns may become self-fulfilling.
As North has argued, all of these arguments can be extended from studies of technological change to other social processes, making path-dependence a common feature of institutional evolution (North 1990, pp.93-5). Path dependence may occur in policy development as well, since policies can also constitute crucial systems of rules, incentives, and constraints (Pierson 1993, pp. 607-8). In contexts of complex social interdependence new institutions and policies will often generate high fixed costs, learning effects, coordination effects, and adaptive expectations. For example, housing and transportation policies in the United States after World War II encouraged massive investments in particular spatial patterns of work, consumption, and residence. Once in place, these patterns sharply constrained the alternatives available to policymakers on issues ranging from energy policy to school desegregation (Danielson 1976; Jackson 1985). Many of the commitments that locked in suburbanization were literally cast in concrete, but this need not be the case. Social Security in the United States became gradually locked-in through its financing system, which created a kind of rolling intergenerational contract (Pierson 1992). Institutions and policies may encourage individuals and organizations to develop particular skills, make certain investments, purchase particular goods, or devote time and money to certain organizations. All these decisions generate sunk costs. That is to say, they create commitments. In many cases, initial actions push individual behavior onto paths that are hard to reverse.
Lock-in arguments have received relatively little attention within political science, in part because these processes have a tendency to depoliticize issues. By accelerating the momentum behind one path, they render previously viable alternatives implausible. The result is often not the kind of conflict over the foregone alternative that political scientists would quickly identify, but the absence of conflict. Lock-in leads to what Bachrach and Baratz called "non-decisions" (Bachrach and Baratz 1962). This aspect of politics can probably be identified only through careful, theoretically-grounded historical investigation of how social adaptations to institutional and policy constraints alter the context for future decision-making.
Over time, as social actors make commitments based on existing institutions and policies, the cost of "exit" from existing arrangements rises. Within the European Community, dense networks of social, political, and economic activity have grown up around past institutional and policy decisions. In speculating about a hypothetical effort to stem the power of Court and Commission, member states must ask themselves if this can be done without, for instance, jeopardizing the single market project. Thus, sunk costs may dramatically reduce a member state government's room for maneuver. In the EC, one can see this development in the growing implausibility of member state "exit threats." While "sovereign" member states remain free to tear up treaties and walk away at any time, the constantly increasing costs of exit in a densely integrated polity have rendered this option virtually unthinkable for EC member states.
Williamson's confident assertion that learning allows firms to adjust to unanticipated consequences applies far less well to an analysis of politics. Member state learning from past events may lead, as it did at Maastricht, to greater restrictions on supranational actors in new initiatives (Dehousse 1993). Recapturing ground in previously institutionalized fields of activity, however, will often be quite difficult. Member states do not inherit a blank slate that they can remake at will when their policy preferences shift or unintended consequences become visible. Decision rules hamper reform, while extensive adaptations to existing arrangements increase the associated costs. Thus a central fact of life for member states is the acquis communautaire, the corpus of existing legislation and practice. As Michael Shackleton notes, "[h]owever much Member States might deplore certain aspects of Community policy, there is no question that all find themselves locked into a system which narrows down the areas for possible change and obliges them to think of incremental revision of existing arrangements" (Shackleton 1993, p.20). Just as has always been true in domestic politics, new governments in member states now find that the dead weight of previous institutional and policy decisions at the European level seriously limits their room for maneuver.
The need to examine political processes over time is the crucial feature linking all the arguments presented in this section. None of these processes are likely to be captured by a "snapshot" view. Historical institutionalism provides a clear account of why gaps emerge in member state authority. Member states are often preoccupied with short-term outcomes. Their decisions are certain to produce all sorts of unanticipated consequences. The preferences of member states may also shift, leaving them with formal institutions and highly-developed policies that do not fit their current needs. At least as important, historical institutionalism provides a coherent account of why learning processes and "fire alarms" may not be sufficient to prompt the reassertion of member state control. If member states decide that their agents have captured too much authority, they may well seek to rein them in. Gaps, however, open possibilities for autonomous action by supranational actors, which may in turn produce political resources that make them more significant players in the next round of decision-making. Decision rules and the proliferation of sunk costs may make the price of reasserting control too high.
In short, historical institutionalist analysis can incorporate key aspects of neo-functionalism while offering a stronger and expanded analytical foundation for an account of member state constraint. There are important points of compatibility between the two approaches. Both suggest that unintended consequences, including spillover, are likely to be significant for institutional development. Both point to the significance of supranational actors. A crucial difference is that neo-functionalism sees political control as a zero-sum phenomenon, with authority gradually transferred from member states to supranational actors, while historical institutionalism emphasizes how the evolution of rules and policies along with social adaptations create an increasingly structured polity which restricts the options available to all political actors. What has been missing from neo functionalism-- and what historical institutionalist arguments can supply -- is a more convincing analysis of member state constraint. Intergovernmentalists challenge neo-functionalism with two key questions: why would member states lose control, and even if they did why would they not subsequently reassert it? Historical institutionalism gives clear and plausible answers to both.
The crucial contrasts between an intergovernmentalist and a historical institutionalist account can be seen in Figure II. While intergovernmentalists focus on the initial bargain at time T0, historical institutionalism emphasizes the need to analyse the consequences of that bargain over time. Doing so reveals the potential for considerable gaps in member state control to emerge (T1). When the time of the next "grand bargain arrives (T2), member states will again be central actors, but in a considerably altered context. Member states may dominate decision-making in these intergovernmental bargains, and actively pursue their interests, but they do so within constraints (frequently unplanned and often hardly visible) created by their predecessors and the micro-level reactions to those preceding decisions. Studying processes of policy and institutional change over time reveals that gaps may well be extensive and the prospects for recapturing lost control are often quite limited. As the next section demonstrates, this has been true in at least one significant (and unexpected) area of European policy development.
The obstacles to an activist role for Brussels in social policy development have always been formidable. As noted, EC institutions make it much easier to block reforms than to enact them. The social forces most sympathetic to European-level activity -- labor unions and social democratic parties -- have had relatively little influence in the past fifteen years. The member states themselves, which serve as gatekeepers for initiatives that require Council approval, jealously protect social policy prerogatives. Economic and geopolitical changes since World War II have gradually diminished the scope of national sovereignty in a variety of domains. The welfare state remains one of the few key realms of policy competence where national governments still appear to reign supreme. Given the popularity of most social programs, national executives will usually resist losses of social policy authority. The agendas of member states generally created only narrow, market-related openings for social legislation, and then only by super-majority vote.
Yet even in this area -- where an intergovernmentalist account seems highly plausible -- a historical institutionalist perspective casts the development of European policy in quite a different light. Those seeking a more thorough analysis should look elsewhere (Leibfried and Pierson 1995). Here I discuss three aspects of policy development which point to significant initiatives that have extended beyond the firm control of member states: (1) interventions on issues of gender equality; (2) the expansion of health and safety regulations; and (3) the recent enactment of the "Social Protocol." Each of these developments illustrates important aspects of an historical institutionalist account.
The EC and Gender Equality.
The European Community has assumed a central role in the development of policies to promote gender equality. It is clear that member states did not seek this outcome. Rather, the EC's extensive role must be considered an unintended by-product of the Community's original institutional design. The key development was the inclusion of Article 119 of the Treaty of Rome, requiring member states to "... ensure and maintain the application of the principle that men and women should receive equal pay for equal work." This provision grew out of a lengthy fight between Germany and France over the more general harmonization of social policy. The Germans, who rejected calls for harmonization, eventually won. Article 119 was considered "merely hortatory" -- a face-saving concession to France rather than a basis for policy (Hoskyns 1986, p.305; Milward 1992, pp.209-16; Moravcsik 1994, p.27). Indeed, the mandate to address pay inequities lay dormant for almost two decades.
Article 119's broad wording, however, offered untapped potential. An opening occurred when the policy preferences of member states shifted in the early 1 970s -- the high tide of social democratic sentiment in the EC and a time when women's movements were gathering strength in many countries. Politicians eagerly sought a symbolic response to these new demands. In this context, the Council agreed to several directives which gave the "equal treatment" provision some content. Catherine Hoskyns summarizes the atmosphere at the time:
.... directives were passed without much awareness of their consequences. Time and again interviews with national officials have shown that those who negotiated the original provisions had no idea what force they would prove to have or the legislative upheaval they would provoke. This is undoubtedly one of the reasons why governments have been so reluctant since 1978 to adopt new directives in this field (Hoskyns 1986, p.306).This growing reluctance reflected both growing member state awareness of unintended consequences and yet another shift in member state policy preferences, this time accompanying the rightward drift in the ideological complexion of member state governments after 1979.
If the member states soon became hesitant, however, their own initiatives had pushed the EC far down a path where member states could no longer fully control the evolution of policy. The passage of the directives, backed by the now far from symbolic Article 119, transferred considerable influence over gender policy to the ECJ. Over the past fifteen years, the European Court of Justice (ECJ) has played a crucial activist role, spurred on by women 5 groups which saw a significant opportunity to advance their agendas. The Court has turned Article 119 and the directives into a broad set of requirements and prohibitions related to the treatment of women workers (Ostner and Lewis 1995).
The ECJ's expansive interpretations of Article 119 and the various Directives have required extensive national reforms of social security law and corporate employment practices. The impact on one member state was described by Ireland's Joint Committee on Secondary Legislation of the EC:
The Community has brought about changes in employment practices which might otherwise have taken decades to achieve. Irish women have the Community to thank for the removal of the marriage bar in employment, the introduction of maternity leave, greater opportunities to train at a skilled trade, protection against dismissal on pregnancy, the disappearance of advertisements specifying the sex of an applicant for a job and greater equality in the social welfare code. After farmers, Irish women in employment have probably benefited most from entry to the EEC (quoted in Mangan 1993, p.72).This conclusion may be generous, but few doubt the broad impact of EC interventions. To take just one important recent example, ECJ decisions have had a dramatic impact on public and private pension schemes. The Court's insistence on equal retirement ages in public pension schemes forced reform in a number of countries. When in Barber the ECJ made a similar ruling for occupational pensions, fear that the ruling might be applied retroactively (at a cost to private insurers estimated at up to &sterl;40 billion in Britain and 35 billion DM in Germany) fueled "what is probably the most intense lobbying campaign yet seen in Brussels" (Mazey and Richardson 1993, p. 15). In this instance, member states were in fact able to limit the damage, but it took a lot to do so: a unanimous agreement to add a "Barber protocol" to the Maastricht Treaty. The protocol states that Barber will not be applied retroactively, thus preempting an interpretation that many agreed would be a dubious application of European law. The prospective impact of the Court's ruling, however, remains profound. 6
There continue to be considerable constraints on EC gender policy (Ostner and Lewis 1995). As with much European regulation, a close connection to the "market-building" project limits the range of possible interventions. Issues of gender equity which cannot be linked directly to the workplace remain out of bounds. Member state implementation has been uneven. The point, however, is not to praise or criticize EC gender policies. Rather, it is to take note of their considerable impact and to demonstrate the historical processes through which this extension of supranational authority took place.
It is also true that member states retain the capacity, as in the case of Barber, to modify outcomes when these are so unacceptable that they mobilize unanimous member state opinion. Gaps, in other words, invariably have some limit (Pollack 1994). Such compensatory steps, however, are likely to be rare. Member states may well wish that the Community had never become active in pursuing issues of gender equality. It is quite another thing, however, to publicly stop or reverse such efforts once they have been enacted and incorporated in national laws, have motivated thousands of firms to adjust their labor market practices, and have enhanced the monitoring and mobilizing capacities of national and transnational interest groups.
The EC has thus come to play a major role in the development of gender policies, and it is the ECJ that now determines what the often vague EC rules require. This outcome cannot have been intended or desired by either the makers of the Treaty or the current COGs of member states. Institutional designers, both in the 1950s and 1970s, were often preoccupied with the short-term and symbolic consequences of their actions; many long-term effects were either ignored or unanticipated. Changes in member state preferences at later dates led to unexpected shifts in course that have proven hard to reverse. Other actors (notably the ECJ but also the Commission and European women's groups) were quick to seize these opportunities, and member states have found it difficult to close the resulting gaps. Indeed, the case of gender equality reveals all of the features of institutional evolution stressed in Part II of this essay.
Workplace Health and Safety.
Another instance of gaps in member state control can be seen in the development of health and safety regulations.
7
Openings for health and safety regulation came with the Single European Act, which allowed qualified majority voting on these issues. Policymakers were concerned that national restrictions could be trade barriers in disguise. The expansion of EC activity in this domain has been remarkable (Eichener 1993). By late 1994, 29 new directives had been passed under the new procedures introduced with the SEA. Many of these were broad "framework" directives covering a range of more specific regulatory activity (Martin and Ross 1994).
Even more surprisingly, a very high level of standards has generally been achieved -- often higher than that of any member state. To be sure, the use of qualified majority voting has been crucial. Yet the outcome of very high harmonization seems difficult to explain in terms of simple intergovernmental bargaining. Constructing the single market might require harmonization of health and safety standards for products. There is, however, no clear need for harmonized standards for production processes. Here as well, however, the European Union has been highly interventionist. Nor is it clear why member states with low standards should accede to significantly higher ones.
As Eichener's detailed investigation documents, the Commission's "process manager" role --a delegation of authority required to pursue complex regulatory policies -- appears to have been critical in this low-profile environment (Eichener 1993). Much of the crucial decision-making took place in committees composed of policy experts. Representatives within these committees were often interested in innovation, having gravitated towards Brussels because it seemed to be "where the action is" on regulatory issues. In this technocratic context, "best practices" from many member states (and from countries then outside the EC, such as Sweden) were pieced together to form a quite interventionist structure of social regulation. At the same time, the Commission played a central part in joining together the work of different committees and incorporating concerns of other actors such as the European Parliament -- all the while actively promoting particularly innovative proposals.
Throughout, the member states appear to have played only a loose supervisory role. This was especially true for the low-standard states of the EC's southern rim. These states had the most to lose from the enactment of high standards, since their adjustment costs would be highest (Lange 1993). Yet these member states found their limited supplies of specialists either co-opted in the "consensual" committee process or overwhelmed by the enormity of the regulatory task.
Health and safety policy reveals how the complexity of regulatory policy-making in a setting of high issue density may generate considerable gaps in control. Thus while the Commission, like other actors in the EC, operates under considerable constraints, it will often be able to advance its own agenda. As Eichener concludes, "[t]he complex, opaque and Commission-dominated decision-making process leads to results which would never be expected from simple inter-governmental bargaining within the Council" (Eichener 1993, p.64).
The Maastricht Social Protocol.
A final illustration of the historical dynamics of institutionalization can be seen in the "Social Protocol" enacted as part of the Maastricht Treaty negotiations. The Social Protocol grew out of continuing efforts to modestly enhance the capacity for activist social policy at the EC level. The Protocol itself allows qualified majority voting on a range of important issues, including working conditions, gender equality with regard to labor market opportunities and treatment at work, and the integration of persons excluded from the labor market. Already, the Commission has used the Social Protocol track to push through the long-stalled European Works Council Directive and is pursuing other initiatives (Falkner 1995).
While it is impossible to know at this stage how the Social Protocol will play out, this exercise in institutional reform again supports key parts of the historical institutionalist argument. The enactment of the Social Protocol is very difficult to reconcile with a simple model of intergovernmentalist bargaining among sovereignty-focused member states (Ross 1994, p.191). The eleven member states acceding to the Protocol were not introducing a carefully-designed "instrument." The member states had in fact expected Britain to sign a much-watered-down clause on social policy, but the Major government rejected all proposed versions. Faced with the prospect that British intransigence would prompt a disastrous breakdown of the Maastricht conference, the member states rushed to adopt (at Delors' suggestion) a last-minute solution. The hastily-cobbled together agreement, which excludes Britain, committed the other eleven member states to a much more ambitious earlier draft on social policy. This version had been designed as a bargaining chip in the expectation that Britain would eventually accede to a contentless compromise. Member states, which had exploited Britain's expected position to engage in cheap talk, suddenly found themselves exposed.
That Britain preferred to block any agreement rather than accept the largely symbolic alternative waiting in the wings is equally instructive. It illustrates how the long-term institutional consequences of the Protocol should be seen as the by-products of a decision made to meet various short-term domestic objectives. If Major had truly wished to preserve social policy autonomy, a solution was readily available. Britain's refusal to agree appears to have had less to do with some long-sighted views of sovereignty than with Major's need to placate right-wing Tories by taking a tough public stance.
Indeed, the Major government's strategy clearly created a greater long-term threat to British autonomy. Choosing to opt Out means that Britain will not participate in decisions that it will have to abide by if a future government joins the Protocol. 8 The Conservatives will likely resist pressures to sign the Protocol, but a Labour government would probably reverse that choice. Thus the status quo can be maintained only if it is ratified at every British election. A single Labour victory would produce an institutional change that could not subsequently be reversed without provoking a constitutional crisis in Europe. In short, the Major government accepted a considerable long-term threat to British sovereignty in the sphere of social policy in return for an important short-run symbolic victory, which it needed for domestic political reasons.
Finally, one should note that the Social Protocol leaves tremendous room for unanticipated consequences (Rhodes 1995). Rather than being an example of Williamson's "far-sighted propensity" in institutional design, the arrangement clearly reflects a harried and desperate effort to keep the Maastricht negotiations from coming unravelled altogether. Legal ambiguities abound. Not only is the whole legal basis of the Protocol open to challenge, but, as Martin Rhodes notes, "the boundaries are blurred between areas subject to QMV, those subject to unanimity and those where the agreement eleven have no competence at all" (Rhodes 1995, p.114). It is, of course, the ECJ that will determine how these ambiguities are resolved. Further uncertainties include whether and when Britain will "opt-in" to the agreement, and what the consequences will be if it remains on the outside. Even with several years hindsight, these uncertainties remain. They were clearly very much a part of the atmosphere in the short period during which the Protocol agreement was reached.
It is only now becoming possible to study the stream of consequences flowing from the Protocol's enactment -- an important aspect of an historical institutionalist investigation (Falkner 1995). Yet the process of institutional design itself appears to be quite in line with the general framework advanced in this essay. Indeed, along with the earlier discussion of Article 119, the case of the Social Protocol reveals that historical institutionalist arguments are relevant not only during the day-to-day activities between the "grand bargains", but for understanding the grand bargains themselves.
The Evolution of European Social Policy.
A more complete review of social policy issues would reveal further constraints on member state autonomy. Among the most important have been spillovers from the single market project. Encouraging the mobility of labor has not been a high-profile issue in the EC, but it has gradually prompted an incremental expansion of Community regulations and, especially, court decisions that have seriously eroded national welfare state sovereignties (Leibfried and Pierson 1995). Social policy cases account for a growing share of the rapidly rising ECJ caseload, increasing from 3.3% of total cases in 1968 to 8.1% in 1992 (Caporaso and Keeler 1993, Table 1). Member states are now prohibited from pursuing a range of social policy options because their actions would be incompatible with the single market project. Furthermore, as individuals, firms, and non-profit organizations adapt to new opportunities made available by the single market (e.g., In private pensions and health services), these micro-level commitments will further restrict the policy options available to national welfare states. Member state social policies are increasingly firmly embedded in a dense set of "hard" Community requirements and prohibitions, as well as "soft" incentives and disincentives.
Thus European integration has generated a partial but nonetheless significant development of European-level social policies. The processes that produced this outcome provide powerful illustrations of the institutional dynamics discussed in Section II of this essay. In a number of instances, the short-term preoccupations of institutional designers have led them to make decisions that undermined long-term member state control. Unanticipated consequences have been widespread, especially as the density of EC activity has grown. Shifts in member state preferences led to unexpected exploitation of opportunities created earlier (e.g., Article 119 in the 1970s) as well as growing frustration with previous commitments (the Gender Equality Directives in the 1980s). In short, even though social policy is widely seen as an area of firm member state control with a minimal EC role, a historical institutionalist perspective highlights the growing significance of European policy, the influence of actors other than member states, and the mounting constraints on member state initiative.
The arguments advanced in this essay present major challenges for an intergovernmentalist account of European integration. By providing explicit micro-foundations for an analysis that places much more emphasis on member state constraint, historical institutionalism increases the pressure on intergovernmentalists to offer convincing evidence that the causal processes they posit are actually at work. Rather than simply inferring preferences post hoc from an examination of outcomes, intergovernmentalists will need to show that the desire to achieve these functional outcomes actually motivated key decision-makers. 9
In principle, important aspects of an historical institutionalist analysis could be integrated with intergovernmentalism. Indeed, this essay accepts the starting point of intergovernmentalism: that member states are the central institution-builders of the European Community, and that they do so to serve their own purposes. Although it has rarely been done in practice, many intergovernmentalist arguments could incorporate a temporal dimension. Keohane, for instance, has recognized the possibility that COGs might anticipate the potential for preference shifts in successor governments (Keohane 1984, p.1 17) Other challenges, however, will not be so easy to reconcile, such as the possibility that COGs employ a high discount rate in making decisions about institutional design, that unintended consequences are ubiquitous, and that gaps that emerge are difficult to close. It is hard to see how these factors could be systematically incorporated into intergovernmentalism without undermining the three pillars of that approach: the emphasis on member state sovereignty concerns, the treatment of institutions as instruments, and the nearly exclusive focus on grand bargains.
The challenge for those wishing to advance a historical institutionalist account is also daunting. The temporal processes outlined here would have to be carefully specified to generate clear hypotheses concerning such matters as when we should expect policy-makers to employ short time-horizons, when to expect that unintended consequences will be widespread, or how particular decision rules influence the prospects for closing gaps in control. As Mark Pollack has persuasively argued, such analyses should focus on the factors that can explain variation in outcomes across issues and among institutional arenas, as well as over time (Pollack 1994; Pollack 1995). To develop the historical institutionalist line of argument will require difficult efforts to trace the motivations of political actors in order to separate the intended from the unintended. Determining the impact of sunk costs on current decision-making also represents a considerable challenge. Studying political arenas in detail over long periods of time is arduous. The evidentiary requirements encourage a focus on detailed analyses of particular cases, rendering investigations vulnerable to the critique that the cases examined are unrepresentative. However, if one accepts the conclusion that intergovernmentalists must now show that the processes they hypothesize are actually at work, rather than simply inferring those processes from observed outcomes, it is not clear that their research tasks are any less formidable.
The purpose of the current investigation is not to pursue these difficult questions but to set an agenda by identifying plausible causal processes that can lead to growing constraints on COGs over time. While only the first step, such an effort can be a prelude to empirically-grounded research, as demonstrated by the brief discussion of EC social policy. Indeed, this first step is a significant one. Historical institutionalist arguments can provide a compelling account for a remarkable development that is widely accepted by European scholars and most Americans working in the field of comparative politics: the European Community is no longer simply a multilateral in strument, limited in scope and firmly under the control of individual member states. Instead, the EC possesses characteristics of a supranational entity, including extensive bureaucratic competencies, unified judicial control and significant capacities to develop or modify policies. Within Europe, a wide range of policies classically seen as "domestic" can no longer be understood without acknowledging the European Community's role within a highly fragmented but increasingly integrated polity. Historical institutionalism provides the analytical tools for thinking of the EC not as an international organization, but as the central level -- albeit still a weak one -- of an emergent multi-tiered system of governance. The power of the member states in this polity is not merely "pooled" but increasingly constrained.
It would be folly to suggest that the member states do not play a central part in policy development within the European Union. Rather, my point is that they do so in a context that they do not (even collectively) fully control. Arguments about intergovernmental bargaining exaggerate the extent of member state power. In their focus on grand intergovernmental bargains, they fail to capture the gradually unfolding implications of a very complex and ambitious agenda of shared decision-making. While the member states remain extremely powerful, tracing the process of integration over time suggests that their influence is increasingly circumscribed. The path to European integration has embedded member states in a dense institutional environment that cannot be understood in the language of inter-state bargaining.
Note *: Paul Pierson, Center for European Studies and Dept. of Government, Harvard University Back.
Note 1: A good example (among many possible ones) is Peter Lange's analysis of the Maastricht Social Protocol (Lange 1993). Lange may well be correct in arguing that poor member states signed the Protocol because of side payments, but he provides no actual evidence that this was the case. Back.
Note 2: Throughout, I rely on North's definition of institutions: "...the rules of the game in a society or, more formally,... the humanly devised constraints that shape human interaction" (North 1990, p.3). Back.
Note 3: Moravcsik has outlined a "liberal intergovernmentalist" view in which "liberal" theories of member state preference formation are used to supplement intergovernmentalist theories of member state bargaining (Moravcsik 1993). Back.
Note 4: Consider for instance Moravcsik's striking acknowledgment of the growing power of the ECJ: "...the decisions of the Court clearly transcend what was initially foreseen and desired by most national governments. The 'constitutionalization' of the Treaty of Rome was unexpected. It is impossible, moreover, to argue that the current system is the one to which all national governments would currently consent, as recent explicit limitations on the Court in the Maastricht Treaty demonstrate." Moravcsik, "Preferences and Power," p.513. Back.
Note 5: An alternative way to discount the significance of unintended effects would be to treat them as random "noise." Yet while this may be appropriate in studying mass populations (e.g., the dynamics of public opinion), it seems inappropriate when single unintended effects may be quite large and processes may be path dependent. There is little reason to think that such effects will somehow "balance out", leaving an analyst free to study the "systematic" elements. To take an example discussed later in this essay, it would be difficult to examine the dynamics of gender issues in Europe by treating the role of Article 119 as "noise." Back.
Note 6: The ECJ upheld the solution found in Maastricht in Ten Oever (Case 109/91 of October 6, 1993). Estimated costs of full retroactivity for Germany are from Claus Berenz, "Hat die betriebliche Altersversorgung zukunftig noch eine Chance?" Neue Zeitschrift fur Arbeitsrecht, 11 (no. 9/1 0, 1994), pp.385-390 (part 1), 433-438 (part 2), pp.437; for Britain, from Mazey and Richardson, "Transference of Power," p. 15. Back.
Note 7: The following account draws heavily on Eichener 1993. See also Martin and Ross 1994. Back.
Note 8: ln this sense, the current situation may parallel the development of the Common Agricultural Policy, where Britain's long absence from the development of policy left it in a weak position to pursue its goal of policy liberalization within the EC (Keeler 1995). Back.
Note 9: Moravcsik 1991 provides a good example of such an effort. Historical institutionalist arguments, however, suggest the need to go beyond even Moravcsik's ambitious attempt to supplement intergovernmentalism with a "liberal" theory of COG preference formation. Moravcsik's account considers only the synchronic domestic sources of COG preferences, ignoring the possibility of significant feedback effects from previous rounds of institutionalization. For a critique of his interpretation along these lines see Cameron 1992. Back.
