Show Summary Details
Page of

(p. 23) 2. Theories of EU Governance 

(p. 23) 2. Theories of EU Governance
(p. 23) 2. Theories of EU Governance

Simon Bulmer

, Owen Parker

, Ian Bache

, Stephen George

, and Charlotte Burns

Page of

PRINTED FROM OXFORD POLITICS TROVE ( © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Politics Trove for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: null; date: 19 January 2021

Chapter Overview

As European integration progressed, the academic focus began to shift from explaining the integration process to understanding the European Union (EU) as a political system (see Table I.1). EU scholars increasingly drew on approaches from the study of domestic and comparative politics, in part to escape the supranational–intergovernmental dichotomy. These contributions broadened the study of the EU beyond the traditional international relations (IR) debate, with many scholars referring to a ‘governance turn’ in EU studies in the 1990s. This chapter surveys a number of approaches that focus on the EU as a political system. These approaches are quite varied and include new institutionalism, governance, and policy network approaches. While they diverge in some important ways, what brings them together is their focus on the EU as a system of governance. At the end of the chapter we focus on some of the overall characterizations of EU governance that also offer valuable insights: supranational governance; new intergovernmentalism; and differentiated integration.

A Shift of Focus

Approaches from the study of domestic and comparative politics turn away from the focus of IR theories on the process of European integration, and instead treat the EU as a political system, and try to explain ‘the nature of the beast’. This shift of attention to concerns with institutional and policy analysis first appeared with Lindberg and Scheingold’s book Europe’s Would-Be Polity in 1970 (Lindberg and Scheingold 1970). It gathered pace with the acceleration of integration in the 1980s through the single market programme (see Chapter 19), and was strengthened by Simon Hix’s (1994) call to scholars within the discipline of comparative politics to wake up to the existence of the EU as a suitable subject for study using their established concepts and toolkits. Instead of asking questions such as how far the EU was dominated by the member states and how far it operated as an autonomous entity, Hix (1999: 1) asked questions that derived from the study of comparative politics:

How is governmental power exercised? Under what conditions can the Parliament influence legislation? Is the Court of Justice beyond political control? Why do some citizens support the central institutions while others oppose them? How important are political parties and elections in shaping political choices? Why are some social groups able to influence the political agenda more than others?

(p. 24) Comparative politics approaches succeeded in shifting the focus away from the study of European integration as a process to ‘how the EU works today’ (Hix 1999: 1; emphasis in original).

An increasing number of concepts are applied to the EU that originated from the study of comparative and domestic politics. Here, we look at three influential sets of analytical approaches that come under the headings of new institutionalism, governance, and policy networks. While we identify these as separate approaches, and there are variants within each approach, there is also overlap between them. We pull these approaches together at the end of the chapter as a way of highlighting broad characteristics of governance in the EU today. In this final section, we look at supranational governance, new intergovernmentalism, and differentiated integration. This final section is therefore concerned with characterizations of EU governance rather than the more widely applicable analytical approaches earlier in the chapter.

New Institutionalism

One approach that was originally applied to the study of domestic and comparative politics that found favour among EU scholars was new institutionalism (NI) (March and Olsen 1984, 1989, 1996). It developed as a reaction to the behavioural approaches that had come to dominate political science in the 1960s and 1970s. It was a return to older concerns with institutions but it departed from the formal, constitutional, and rather descriptive character of earlier institutionalist analysis by widening the scope of analysis. It was one way in which the governance turn in EU studies was operationalized.

Bulmer (1993) set down a marker in arguing that institutions matter in the EU context, identifying the different institutions and instruments of EU governance and how they shape political action. The article further argued that the profile of instruments and institutions differed within discrete ‘governance regimes’ situated within an overarching EU structure that had pronounced regulatory characteristics. In work with a lawyer, Kenneth Armstrong, he applied NI to the governance of the single market, including a set of six case studies that featured discrete governance regimes (Armstrong and Bulmer 1998).

As NI became more prevalent in political analysis generally and in the literature on EU studies, its ‘value added’ became clear. NI not only argued that institutions matter in understanding EU politics and policy making, but also sought to explain how they matter. What was ‘new’ about NI was that institutions were not only defined as formal organizations as they were traditionally understood—such as parliaments, executives, and courts—but also extended to categorize informal institutional features: rules, norms, and even institutional culture. Understood in this way, institutions constrained or shaped political action: whether at a macro (EU) level or in particular governance regimes concerned with specific policy issues. New institutionalists argued that institutions thus defined were not neutral arenas, since they biased access to the political process in favour of some actors and societal groups over others. Second, they argued that institutions could be autonomous political actors in their own right.

While, collectively, this reassertion of institutionalist perspectives is known as NI, it became common to distinguish between three varieties: rational choice (p. 25) institutionalism; historical institutionalism; and sociological institutionalism (Hall and Taylor 1996). These approaches emerged at around the same time, but in relative isolation from each other. In particular, the three variants display different approaches that derive from divergent understandings of institutions. Rational choice institutionalism has the ‘thinnest’ understanding of institutions: typically, as formal rules. At the other end of the spectrum is sociological institutionalism, which assumes the widest understanding of institutions, including culture, norms, and values. Moreover, sociological institutionalism assumed that political action followed a logic of appropriateness: following conventions, for instance, rather than simply trying to maximize self-interest, as with rational choice institutionalism. Historical institutionalism occupied a hybrid position between the two poles but was often especially concerned with developments across time: the causes and effects of policy change, for instance, or the ‘stickiness’ of policy; that is, its resistance to reform. NI therefore opened up new research agendas on the EU (see Schneider and Aspinwall 2001: Chapter 1).

Rational Choice Institutionalism

Rational choice institutionalism focuses on the constraints that formal institutional structures impose on actors. It suggests that, in trying to understand the behaviour of political actors, it is important to identify the parameters that are set by the specific framework of rules within which politics takes place. So, for example, the activities of interest groups reflect the procedures that prevail for the passage of the legislation that affects them, and the access points that are available to them in that process. Thus, in the case of the EU, whether interest groups choose to try to influence legislation through national governments or through the Commission and the European Parliament (EP) would reflect:

  • the extent to which the process is supranational or intergovernmental (e.g. what decision rules apply within the Council of Ministers);

  • in the case of a supranational process, what role the EP has in the final decision; and

  • in the case of a more intergovernmental process, access to national governments.

For policy areas that are more supranational in character, such as the EU’s trade policy (see Chapter 24), interest groups are more likely to turn to the Commission because it has the key powers in negotiating with other states; for instance if such states are using trade practices that conflict with international rules. Similarly, if qualified majority voting (QMV) in the Council of Ministers applies to the particular policy issue, the potential influence of the Commission is greater, making it a more attractive target for lobbying. Moreover, as the powers of the EP have increased over successive reforms (see Chapter 15, ‘The Struggle for Power’), interest groups began to lobby the EP more extensively than they had previously.

By contrast, in a policy such as taxation policy, where unanimity applies in the Council of Ministers, it is important for interest groups to lobby at the national level because one member state’s vote against an EU proposal could block it. There are consequently strong incentives for lobbying at the national level. Interest groups therefore are ‘shooting where the ducks are’; that is, they lobby where they are most likely to get the desired results—a classic assumption of rational action.

(p. 26) Rational choice institutionalists have also made a significant contribution to understanding the ways in which supranational actors might obtain a degree of autonomy from national governments, allowing them to make their own input to the policy process (Pollack 1997). Applying what is known as principal–agent theory, rational choice institutionalists have highlighted the difficulties of principals (the national governments) in keeping control over the activities of their agents (the supranational institutions).

  • As the range of delegated tasks has increased, so the difficulties of monitoring what the agents were doing has increased.

  • As the number of principals increases with successive enlargements of membership, so the agents may play off the preferences of different coalitions of principals against the attempts of other principals to restrain them.

  • As QMV has expanded, so the constraints on the Commission in constructing a winning coalition in support of its proposals have been reduced.

Pollack (2003) applied rational choice theory to test hypotheses about the delegation of power by the member states to the EU’s supranational organizations (mainly the Commission, the Court of Justice, and the Parliament) and the efforts of these organizations to shape the process of European integration. Pollack’s first concern was with the types of function that governments delegate to supranational organizations and the conditions under which greater or less discretion is allocated to these agents. Here, principal–agent theory suggested that governments would delegate to supranational agents to reduce the transaction costs of EU policy making. The evidence suggested that this was the case with the Commission and the Court of Justice, but delegation to the EP was motivated more by governments’ attempts to reduce the EU’s democratic deficit (see Chapter 3, ‘Democracy and Legitimacy’) and was thus more effectively explained in terms of the logic of appropriate behaviour identified by sociological institutionalists (see ‘Sociological Institutionalism’ below), rather than by rational choice explanations.

The principal–agent debate extends to a concern with how control is maintained over regulatory agencies that are a key component of modern governance in the EU and beyond (see ‘Regulatory Governance’ below). The proliferation of agencies in the EU context has been important to the development of governance approaches to understanding the EU (see ‘Regulatory Governance’ below) and has also raised important issues relating to the accountability of EU decision making (see Chapter 3, ‘Democracy and Legitimacy’).

Historical Institutionalism

Historical institutionalists place emphasis on the argument that political relationships have to be viewed over time. Their approach argues that decisions are not made according to an abstract rationality, but according to perceptions and within constraints that are structured by pre-existing institutional relationships. Historical institutionalists also take a broader definition of institutions to incorporate informal constraints on behaviour such as values and behavioural norms. As Hall and Taylor (1996: 938) note, ‘they define them as the formal or informal procedures, routines, (p. 27) norms and conventions embedded in the organizational structure of the polity or political economy’.

Central to historical institutionalism is the concept of ‘path dependence’. This is the argument that once one decision is made, it tends to make it more likely that policy continues to develop in a particular direction. Or, as Hall and Taylor (1996: 941) put it, ‘forces will be mediated by the contextual features of a given situation often inherited from the past’. In extreme cases, path dependence can turn into ‘lock-in’ (Pierson 1996), whereby other avenues of policy are entirely blocked off by the bias towards the existing route that is built into the system. This could be one explanation of why policies such as the Common Agricultural Policy (CAP) proved so resistant to reform even after their negative effects (such as the policy’s cost and distortion of trade with non-EU states) had become obvious in the 1980s (see Chapter 21).

Applying historical institutionalism to the EU led to a critique of liberal intergovernmentalism (LI), suggesting that national governments might not be entirely in control of the process of integration (Pierson 1998: 34–50). Intergovernmental analyses tended to focus on the major order decisions, represented mainly by revisions to the Treaties (see Chapter 1, ‘Theories of European Integration’). Intergovernmentalists treated what happened between these historic decisions as simply the working through of the decisions. Historical institutionalists argued that, after the decision had been taken, it would be likely to produce unanticipated and unintended outcomes. This might be because of a simple failure to think through the implications—but it might be for one or both of two other reasons (Pierson 1998: 41).

First, the preferences of governments might change over time. For example, the preference for a CAP based on price support might have been a rational response to conditions in the 1960s in Europe, when security of food supplies was a paramount concern, but no longer appropriate in the changed circumstances of the 1990s, when technological advances had removed this concern. Also, national governments might change. For example, the EC directives on social policy to which a British Labour government had agreed in the 1970s were not to the liking of the Conservative governments between 1979 and 1997. However, even where preferences change, governments would find it extremely difficult to change decisions: policy sectors became institutionalized and various incentives emerged for maintaining the status quo.

A second reason why governments find it difficult to change decisions is because the voting rules in the Council of Ministers might make it difficult to reform the character of a policy once it has been agreed. Where the rule is subject to unanimity, it is impossible to make major reform so long as one member state benefits from the status quo and refuses to move from it. For example, France was reluctant to reform the CAP in the 1980s; yet its agreement was necessary due to the need for unanimity in the Council of Ministers. This situation is sometimes called ‘the ratchet effect’. Even where QMV applies, in order to effect change it is necessary to construct a coalition representing more than a simple majority of states (the exact number depending on the weighting of the votes of the members of the coalition, and therefore on the identity of the states involved: see Chapter 12, ‘The Council of the EU (the Council’). Not only do the necessary votes for reform need to be gathered but there also has to be agreement on the preferred policy revision.

In a similar vein, and working from the understanding of the EU’s evolution as a timescape, Bulmer (2009: 313–14) also argued that intergovernmental bargains (p. 28) represented only one dynamic of integration; what happened between the historic decisions represented another, while the impact on politics and policy of the rulings of the Court of Justice of the EU introduced a third that was neglected by LI. Bulmer (2009: 314–17) also took issue with the critique often directed at historical institutionalism, namely that it could only explain continuity but less so policy change. He noted that some characteristics of EU politics were cyclical, such as the five-year elections to the EP now aligned with changes to the Commission composition and other posts. This cycle refreshed the internal politics of the EU, with the potential for policy change. Moreover, some policies are cyclical, such as the EU budget and the structural and investment funds. Both are subject to seven-yearly reviews. The policies therefore have inbuilt opportunities for reform, and contrast with others that incorporated no such opportunity for change, notably the CAP.

Sociological Institutionalism

The emergence of sociological institutionalism was closely linked with the ‘constructivist turn’ in the study of the EU and international politics (see Chapter 4, ‘Social Constructivism’). Like constructivism, sociological institutionalism takes as its starting point a rejection of the rationalist approach to the study of politics that characterizes rational choice institutionalism and some contributions to historical institutionalism, and places more emphasis on broadly ‘cultural’ practices (see Chapter 4 for more on this debate).

Hall and Taylor (1996: 947) identified three features of sociological institutionalism that distinguished it from the other NIs. First, the definition of what constitutes ‘institutions’ is considerably broader than in the other approaches, so that it includes not only formal rules, but also ‘symbol systems, cognitive scripts, and moral templates that provide the “frames of meaning” guiding human action’. This definition blurred the lines traditionally separating the notions of ‘institutions’ and ‘culture’.

Second, sociological institutionalism takes a distinct position on the relationship between institutions and individual action that flows from the ‘cultural’ approach. In particular, it suggests that institutions do not simply influence the ‘strategic calculation’ of actors, but have a deeper effect on their preferences and very identity. As Hall and Taylor (1996: 948) put it:

The self images and identities of social actors are said to be constituted from the institutional forms, images and signs provided by social life.

This does not mean that actors are not ‘rational’ in the pursuit of the goals and objectives; rather, it means that these goals and objectives are constituted differently (that is, socially) and are more broadly defined than rationalist theory would suggest.

Finally, sociological institutionalism contrasts with rationalist explanations on how institutions are formed and developed. For rational choice institutionalists, institutions are developed by rational actors to meet particular ends efficiently, such as reducing transaction costs. For sociological institutionalists, institutions are often created and developed because they contribute to social legitimacy rather than efficiency. In some cases, this may mean that the formal goals of an organization are overridden by these broader social goals. Here, a distinction is made between the rationalist ‘logic of instrumentality’ with the sociological ‘logic of appropriateness’. Thus, in the case of (p. 29) delegation to supranational institutions discussed above, it may not be in the instrumental interests of national governments to enhance the EP’s powers, but the need to enhance the democratic legitimacy of the EU provides a powerful logic of appropriateness.

Rosamond (2003) suggested that the application of sociological institutionalism may be particularly useful in understanding why the Commission’s Directorates General operate in very distinct ways. It may also provide insights into whether ‘formally intergovernmental processes … conform to established patterns of interstate interaction, or whether they bring about new norms of exchange between the envoys of member states’ (Rosamond 2003: 117). Another application relates to the persistence of different national varieties of capitalism despite the centripetal forces or convergent effects of the EU single market and monetary union. Nationally rooted cultural characteristics help explain the persistence of different patterns of capitalism across the eurozone, and these in turn explain the difficulties of following a single set of rules in the eurozone: one of the underlying problems that led to the eurozone crisis (see Chapter 11, ‘The Unfolding Eurozone Crisis’; Chapter 20, ‘The Eurozone Crisis’).

Where change takes place, sociological institutionalists place emphasis on the role of ‘norm entrepreneurs’: key politicians or member states that may be able to change the existing policy norms. For example, attention might be focused on the role of Germany in reinforcing ‘sound money’ principles into eurozone governance during the crisis in the early 2010s (Matthijs and McNamara 2016). In other words, the German government acted as a norm entrepreneur.

New Institutionalism Assessed

Put concisely, NI in its three variants reveals that, respectively, formal rules matter to rational action (rational choice institutionalism), time matters to policy and institutional evolution (historical institutionalism), and social context matters to political behaviour within institutions (sociological institutionalism).

For Hall and Taylor (1996: 95), none of the NIs is particularly ‘wrong-headed’ or ‘substantially untrue’; rather, ‘each seems to be providing a partial account of the forces at work in a given situation or capturing different dimensions of the human action and institutional impact present there’. The three approaches rest on different assumptions about the world that led them to focus on different questions. This is illustrated in relation to EU enlargement (see Chapter 26, ‘Explaining Enlargement’), where sociological institutionalists have placed more emphasis on why the EU decided to enlarge and how we can account for the subsequent negotiations, while rational choice institutionalists have devoted greater attention to the impact of enlargement on the EU’s institutional arrangements and historical institutionalists on the process of reform in central and eastern Europe (Pollack 2004: 151).

The various strands of NI have been refined over time, through both empirical research and through engagement with the other strands and with other conceptual approaches to the EU. Other variants of NI have emerged that space prevents us from outlining, such as discursive institutionalism (e.g. Schmidt 2014). New institutionalism has made a significant contribution to the literature on Europeanization (see Chapter 3, ‘Europeanization and the New Institutionalisms’) by providing sensitivity to how domestic institutions (formal and informal) mediate EU pressures. The NIs also link up with research on governance and network approaches, to which we now turn.

(p. 30) Governance

The practice and study of governance developed in the 1990s and the similar timing to that of the development of NI was not just chance. What was distinctive about governance was a move away from government: a development that reflected new ‘softer’ forms of rule in domestic politics and the engagement of a wider range of actors in these processes, as well as the growth of the EU itself and other global organizations internationally. New institutionalism’s analytical shift away from formal institutions thus reflected trends in the practice of politics and policy making.

Definitions of governance abound (for different perspectives, see Pierre 2000). A particularly useful definition was offered by Kohler-Koch and Rittberger (2006: 28), identifying governance as:

a process and a state whereby public and private actors engage in the intentional regulation of societal relationships and conflicts. Governance is thus different from government, the latter stressing hierarchical decision making structures and the centrality of public actors, while the former denotes the participation of public and private actors, as well as non-hierarchical forms of decision making.

The term ‘governance’ therefore recognized that public policy making was increasingly being characterized by wide participation of public, private, and voluntary-sector actors. In the context of the EU, the multi-level governance framework (see ‘Multi-Level Governance’ below) brought together this increased ‘horizontal’ mix of societal actors from different sectors with increased ‘vertical’ interactions between actors organized at different territorial levels (supranational, national, and sub-national).

Some of the literature on governance entailed a normative dimension, termed ‘good governance’. The EU itself promoted the notion of good governance and published a White Paper (European Commission 2001) entitled ‘European Governance’. The Commission’s work entailed exchanges with academics working on governance and the European Commission seeking to improve its own processes.

Arguably the main ‘takeaway’ from the governance literature is the recognition that governing is not just achieved by the traditional hierarchical means of ‘command and control’, backed up by legal means (see Börzel 2019). Governance and policy making can still be undertaken by this traditional route—governance by hierarchy—but also through two other methods: governance by markets or by networks (see Table 2.1).

Table 2.1 Forms of Governance in the EU

Form of governance




Method of governance

Command and control: authoritative decision backed by law

Competition amongst rules

Negotiation, learning

Policy examples

Competition policy, monetary policy in the eurozone

Corporate taxation

Open method of co-ordination (OMC), Common Foreign and Security Policy (CFSP)

Single market

Governance by hierarchy typically takes place in the most centralized policy areas, where the European Commission and Court of Justice have strongest powers and can rely on European law as a means of enforcing policy. The European Central Bank (ECB) exercises equivalent power in the eurozone. Governance by markets applies where member states have competing rules within a loose EU framework, such as with corporate tax policy. Ireland, for instance, seeks to attract technology firms like Google and Apple to invest there, helped by low corporate taxation rates, while the EU’s control over such policy detail is constrained. The cornerstone of the EU—the single market (see Chapter 19)—is a hybrid policy. Barriers to trade have been removed and policy discriminating against EU-wide trade are prohibited (hierarchy). In general, however, it is down to companies to compete following market dynamics. Economic behaviour is driven by competition: a familiar situation in national politics as well, for (p. 31) instance with the opening up of sectors such as mobile and landline telephony to competing providers. Governance by networks, finally, entails negotiation. This typically takes place in policy areas where the supranational institutions’ power is weak. Bargaining or the exchange of ideas between governments and other actors is what drives governance in this third case.

Three particular applications of the governance literature are worth exploring because of their significance to how EU policy making takes place: multi-level governance, regulatory governance, and new methods of governance.

Multi-Level Governance

Multi-level governance (MLG) has assumed considerable value in exploring policy areas that bring together actors from different levels of political authority. In fact, MLG was first developed from a study of EU structural policy and was later developed and applied more widely. An early definition by Gary Marks (1993: 392) referred to:

the emergence of multi-level governance, a system of continuous negotiation among nested governments at several territorial tiers—supranational, national, regional and local.

While accepting that integration involved intergovernmental bargains, MLG theorists reasserted the neofunctionalist critique of realism that individual governments were not as firmly in control as suggested.

Marks et al. (1996) were writing at a time when LI had just become dominant in theorizing integration (see Chapter 1, ‘Liberal Intergovernmentalism’). Accordingly, they chose to contrast the claims of MLG against those of LI. They therefore made three key points against the intergovernmental view, as follows (Marks et al. 1996):

  1. 1. Collective decision making involves loss of control for the governments of individual states.

  2. 2. Decision making competencies in the EU are shared by actors at different levels, not monopolized by the governments of states.

  3. (p. 32) 3. The political systems of member states are not separate from each other, as Moravcsik assumed (see Chapter 1, ‘Theories of European Integration’), but are connected in various ways.

While Marks et al. accepted the central role of the Council of Ministers in EU decision making, they pointed to a number of constraints on the ability of individual governments to control the outcomes of such collective decision making. The use of QMV in the Council was an obvious constraint: any individual government might be outvoted. So, while it was true that governments might be able to attain desired objectives by pooling their sovereignty, this was not the same as arguing that their control of decision making remained intact.

Supranational institutions might be created by member governments to assist them, as Moravcsik argued, but these did not remain under close national government control. For intergovernmentalists, national governments could ultimately choose to rein in the power of these institutions. For MLG theorists, this was difficult in practice because changes to the role of supranational institutions required unanimous agreement, which was difficult to secure with so many member states.

Marks et al. also criticized Moravcsik’s argument that national interests were determined purely through the interplay of domestic actors (see Chapter 1, ‘Theories of European Integration’). Instead, they argued that component parts of national governments, and also non-state actors, could form alliances with their counterparts in other member states, which influenced national governments’ negotiating positions on EU matters. These alliances would not be under the control of the core institutions of the central government, such as the Foreign Office or the Prime Minister’s Office, and the Commission in particular would be able to exploit the existence of these transgovernmental and transnational networks of actors to promote its policy preferences within the ‘domestic’ politics of member states.

What this attempt to achieve MLG’s ‘brand differentiation’ from LI rather skated over in relation to the specific case of cohesion policy was the fact that they might both be correct. The seven-yearly revision of cohesion policy tends to be intensely intergovernmental in character and to be focused around negotiation in the Council of Ministers. LI could be helpful in explaining these negotiations. Thereafter, the policy had to be implemented. At this stage it was possible for the European Commission and subnational authorities to work together—perhaps bypassing national governments—in the manner identified by the MLG literature.

Rather than a coherent theory, MLG was initially an amalgamation of perspectives that were primarily directed at what its advocates saw as the misrepresentation of the nature of the EU by the intergovernmental theorists. It did contain some elements of an explanation for the development of the EU, but it was primarily concerned with the analysis of the nature of the EU. In fact, it was to be later that Hooghe and Marks (2009) incorporated the insights of MLG as part of the foundations of postfunctionalism (see Chapter 1, ‘Postfunctionalism’).

In an elaboration of their approach, Hooghe and Marks (2003, 2004) developed a twofold typology of MLG to capture the EU (see Table 2.2). Type I MLG effectively describes the formal institutions of government at various territorial levels (supranational, national, sub-national), which have multiple tasks and responsibilities and have jurisdictions that are clearly distinct from each other. In other words, Type I MLG (p. 33) describes the system-wide governing architecture. Co-existing with Type I MLG, Type II describes the many smaller bodies of governance that are generally set up with a specific purpose, sometimes for a limited time period, and which are deliberately flexible in membership and organization to deal with specific public policy challenges.

Table 2.2 Types of Multi-Level Governance

Type I

Type II

General-purpose jurisdictions

Task-specific jurisdictions

Non-intersecting memberships

Intersecting memberships

Jurisdictions at a limited number of levels

No limit to the number of jurisdictional levels

System-wide architecture

Flexible design

Source: Hooghe and Marks (2004: 17). Reprinted with the permission of Oxford University Press.

Type II MLG was arguably the variant with greatest value for understanding EU policy areas such as the European Structural and Investment Funds that require task-specific cooperation across levels of government. Type I MLG was not too distant from the established literature on federalism, which is itself interested in tiers of government and the distribution of power between them. Federalism, it will be recalled. had acted as a motivating force behind integration itself (see Chapter 1, ‘Spinelli and Federalism’). In those early years, advocates of a new form of European international relations to transcend the nation state advocated a federal Europe, and sought a political strategy to achieve it. Federalism is also an analytical literature, and we take the opportunity to point it out in passing. It is a body of work that some analysts use to highlight weaknesses in the EU as a political system or to compare it to the federal system in the United States and others (e.g. McKay 2001; Nicolaïdis and Howse 2001; Fabbrini 2005, 2010; Menon and Schain 2006). Comparison of the EU with American federalism is a popular way of teaching the former in the United States.

The MLG literature has sparked different theoretical debates. For example, Peters and Pierre (2004) suggested that MLG’s flexible and informal modes of co-ordination might bring dangers. In particular, the purported advantages of MLG in terms of functional efficiency might be traded for core democratic values as authority seeps away from the formal institutions in which democratic accountability is exercised, or as political actors use these informal and opaque processes to escape accountability for their decisions (see also Chapter 3, ‘Democracy and Legitimacy’). To meet this challenge, Jan Olsson (2003) suggested both abolishing the institutions of MLG (multi-level and cross-sectoral partnerships) and allocating their functions to elected institutions, or, more realistically, allowing elected institutions to play a greater role in regulating partnerships.

Building on Olsson’s work, Bache and Chapman (2008) put forward three models of democracy—electoral, pluralist, and elite-democratic—through which to evaluate the democratic credentials of MLG. Their case study of the structural funds in South (p. 34) Yorkshire (UK) illustrated that, among the expected complexity and technocracy at this stage of policy making, there were also experiments in local democracy that had not previously been identified in the academic literature. As such, in the context of deep MLG, there is evidence that, while traditional mechanisms of accountability may be undermined, other mechanisms may provide a valuable alternative.

MLG sparked a lot of debate (see Jordan 2001; George 2004). Critical questions were posed. Was it descriptive of the EU rather than a theory? Did it only help explain governance in the cohesion policy area? Did it overestimate the role of subnational authorities?

In summarizing the state of the debate, Bache and Flinders (2004: 197) identified four common strands in the literature on MLG that raised hypotheses for future research:

  1. 1. that decision making at various territorial levels is characterized by the increased participation of non-state actors;

  2. 2. that the identification of discrete or nested territorial levels of decision making is becoming more difficult in the context of complex overlapping networks;

  3. 3. that, in this changing context, the role of the state is being transformed as state actors develop new strategies of co-ordination, steering, and networking to protect, and in some cases to enhance state autonomy;

  4. 4. that, in this changing context, the nature of democratic accountability has been challenged.

Regulatory Governance

The literature on regulatory governance in the EU arose from two developments (Insight 2.1). First, changing patterns of state involvement in economic policy resulted in the state playing a more arm’s-length role as a regulator. This trend was global in nature, influenced by the policy agendas of the Reagan presidency in the USA (1981–89) or of Mrs Thatcher’s privatization and deregulation policies in the UK (1979–90). The second origin came from creating the single European market in the late 1980s (see Chapter 19) and its consequences for EU governance (Majone 1991, 1996). The original laborious efforts to harmonize thousands of European regulations regarding multiple products was set aside in favour of a pattern of mutual recognition of standards at member state level, provided they met broad safety standards. In addition, setting product standards in some new technologies was outsourced to private-sector bodies or EU regulatory agencies. The number of these agencies proliferated with the single market (see Chapter 13, Box 13.1, ‘Independent Agencies’). The European Medicines Agency, for instance, monitors the safety of drugs for sale in the EU ( The emphasis in this proliferation of agencies was upon policy expertise and policy efficiency. In most cases, committees of national regulators oversaw policy implementation. Indeed, EU and member-state regulators began to build issue-specific networks.

There is, it should be pointed out, a downside to delegating power to expert bodies, and that is that it all adds up to reducing political choice. This situation creates specific problems (see the example of the ECB’s accountability in Insight 2.1) but cumulatively it is a contribution to the EU’s democratic deficit (see Chapter 3, ‘The Democratic Deficit’). (p. 35)

Insight 2.1 The EU as a Regulatory System

The idea of the EU as a regulatory system is most closely associated with the work of Giandomenico Majone (1996). Rather than a single analytical approach, it offers a terrain of study focusing on the EU’s increased use of regulatory instruments and agencies to secure policy objectives. This focus accompanied the changing character of the EU from the late 1980s, with the extensive regulatory activity aimed at completing the single market, and it continues to be important for a polity with limited budgetary policy instruments (see also Chapter 19, ‘Evaluating the Single Market’).

Like MLG, regulatory agencies threw up some governance problems; specifically concerning democratic accountability. The public might not be so interested in the technicalities of medicine regulation but other agencies had much greater political salience. Arguably, the most significant body of this kind is the ECB, which sets interest rates for the eurozone (see Chapter 20). At the time of the Maastricht Treaty (1992), when the provisions for the ECB were agreed, keeping interest rate and money supply decisions away from politicians was deemed a good thing. It was seen as a way of keeping short-term political thinking out of these decisions. It was an approach to policy that was strongly influenced by German practice. In the 2010s, by contrast, with the ECB setting negative interest rates and selling government bonds to boost the eurozone economy, its policies have faced significant criticism. Paradoxically, the greatest criticism has come from inside Germany because of the unorthodox nature of these policy measures by German standards. Specifically, they are eroding the value of savings and consequently providing fuel for populist criticism of the EU in Germany. However, by design the process for holding the ECB to account is very indirect.

New Methods of Governance

From the 2000s, there was political momentum behind coordinating policy in several policy fields, such as employment policy, or in seeking to enhance the global competitiveness of the EU. However, because the EU did not have clear authority over such matters it could not proceed by passing legislation. Instead, the EU set guidelines, encouraging member governments and societal actors to develop their own action plans. These plans were to involve other stakeholders from the member states. The policy dynamic was to be one of policy learning through exchange of good practice between member states and other ‘softer’ methods (governance by network) by contrast with the ‘command-and-control’ route of legislation (governance by hierarchy). This approach became termed the open method of co-ordination (OMC). It was most closely associated with the Lisbon Strategy (Chapter 10, ‘The Lisbon Strategy’) and the successor programme, Europe 2020 (Chapter 11, ‘Europe 2020’).

For some analysts these new methods of governance, with their emphasis on benchmarking, peer review, and league tables as ways of evaluating policy success, seemed to offer new opportunities for the EU to help member states enhance their domestic policies through cooperation at EU level. Sabel and Zeitlin (2010; Zeitlin 2015) advanced the OMC as part of a new form of ‘experimentalist governance’ that was (p. 36) characterized by less interventionist forms of governance. Yet, at the same time the process of learning between member states offered the prospect of benefits.

This form of governance has also been susceptible to criticism, however. Policy performance in relation to the Lisbon Strategy, for instance, was very mixed (Copeland 2012). Furthermore, the European Parliament was distant from the processes and it was not clear that national parliaments were closely involved either, thus raising questions of democratic accountability (Parker 2018).

Governance Assessed

The governance turn was important in revealing how patterns of rule were changing in the EU, just as they were in national politics. The governance literature therefore introduced new ways of exploring the EU as a political system. Multi-level governance, regulatory governance, and new methods of (or experimentalist) governance are by no means confined to the EU alone. However, they offer important insights into the EU’s evolution as a political system, supplementing the original hierarchical approach of policy making through the institutions and law (see also Chapter 18).

Policy Networks

In the same way that some analysts advocated treating the EU from a comparative politics perspective, such as Simon Hix (see ‘A Shift of Focus’, above), those working in public policy analysis also sought to shine a light on EU governance. Richardson (1996) advocated organizing analysis of EU governance around the concepts that he and others had been using for some time to study the policy-making process within member states. He argued in particular for the application of two concepts: policy networks and epistemic communities. The policy network concept was originally developed in studies of public policy making in the United States and later became prominent in Britain, particularly through the work of Rhodes (1981, 1988). It is a mid-range or ‘meso-level’ concept, aimed at explanation of particular policy sectors or issues, rather than the characteristics of the political system as a whole (the ‘macro level’).

According to the ‘Rhodes model’, a policy network is a set of resource-dependent organizations, meaning that each of the groups that makes up the policy network needs something that the others have in order to fulfil its own objectives (Rhodes 1988). The types of resources that organizations bring to a policy network to exchange in the process of bargaining include constitutional-legal, organizational, financial, political, and informational resources. These ‘resource dependencies’—the extent to which organizations depend on each other for resources—are the key variable in shaping policy outcomes. As Peterson and Bomberg (1993: 28) put it: ‘They set the “chessboard” where private and public interests manoeuvre for advantage.’ However, interdependence between network participants is ‘almost always asymmetrical’ and in some cases it is possible to talk of ‘unilateral leadership’ within networks (Rhodes 1986: 5).

The policy-networks approach does not constitute a predictive theory of policy making, but contributes to explaining policy outputs. It is actor-centred and therefore (p. 37) contrasts with the institution-focused approaches considered in new institutionalism above. For Peterson and Bomberg (1993: 31):

Policy networks are essentially descriptive theoretical tools which simply help order facts and evidence in novel ways. However, policy networks can be used to anticipate and explain policy outputs by providing insights into how and why decisions were taken which produced them.

Policy Networks and the Study of the EU

There are two very useful contributions made by the policy network literature to the study of the EU. The first is to propose that the structure of the policy network may impact on the resultant policy. The second is to identify the different scales of decision making in the EU: from the grand bargains reached by member governments on Treaty reform to the much more specific agreements reached in a group of policy specialists on an EU issue such as the total allowable catch of fish in the Baltic Sea. We take each of these insights in turn.

In outlining the importance of different structural characteristics of different types of network, Rhodes (1988) distinguished between different types of network, ranging from highly integrated policy communities to loosely integrated issue networks (see Insight 2.2). These different ‘structural characteristics’ of networks have different effects on both the internal dynamics of the networks and on the ability of networks to resist external pressures for change. However, policy outputs are generally not just a function of internal network characteristics, but are shaped also by changes in the broader political and economic environment (Rhodes et al. 1996). As such, the approach is often at its strongest when used in conjunction with a macro-level theory of politics or policy making that seeks to explain the broader political context within which the network is situated.

One basic contention evident in the policy network literature is that the more integrated a policy network—i.e. a policy community (Insight 2.2)—the more likely it is that major policy change may be obstructed. A close policy community might also mean that interest groups have quite a degree of influence over the policy decision. The automobile and pharmaceutical sectors were identified as having policy communities (Peterson and Bomberg 2003). By contrast, issue networks are looser and sectoral interests more difficult to enforce. Consequently, the EU institutions have more ability to develop policy solutions without obstruction by interest groups. Hence issue networks are more likely in social and consumer policy (Peterson 1995b).

Insight 2.2 Policy Communities and Issue Networks

A policy community is marked by:

  • limited membership;

  • stable membership over long periods of time;

  • a high level of interaction between the members;

  • shared values between members;

  • some degree of equality in the distribution of resources;

  • a relative balance of power and influence between members.

An issue network, in contrast, is marked by:

  • large and diffuse membership;

  • frequent shifts in the membership;

  • fluctuating frequency of contact between members;

  • lack of shared values;

  • marked inequality in the distribution of resources;

  • marked inequality in power and influence within the network.

The applicability of the policy-networks framework to the analysis of the EU was questioned by Kassim (1994). He argued that the multinational character and institutional complexity of the EU made it difficult to delimit policy networks, and particularly to identify the relevant public sector actor. Sometimes national agencies would be key; sometimes the Commission; sometimes other EU-level institutions. The institutions themselves often acted as lobbyists in the EU in pursuit of their own objectives. Peterson (1995a) responded that while some sectors remained fluid, others had settled into more stable patterns. Indeed, he argued that the Commission was so under-resourced that it had to try to enter into stable relationships with partners that it could trust, which had information that it could use. While he accepted that the delineation (p. 38) of policy networks at the EU level was a difficult task, Peterson insisted that this did not make it a less important one.

The second major contribution of the policy-networks literature related to the need to be clear about the level (or scale) of the policy-making process that was being analysed. Peterson argued that the policy-networks model was best able to explain what he called ‘the policy-shaping decisions’, when proposals were being formulated and before a political decision was taken that ‘set’ the policy (Peterson 1995a: 400). In a subsequent article, he expanded on this argument (Peterson 1995b). He identified three ‘levels’ of analysis in EU decision making. The highest level he termed the ‘super-systemic’ or ‘history-making’ decisions. These were mainly decisions taken by national governments in the European Council or at intergovernmental conferences (IGCs), and were most fruitfully analysed using intergovernmental ideas. The second level he termed the ‘systemic’ or ‘policy-setting’ stage. At this level, a combination of intergovernmental and inter-institutional analysis was needed to understand outcomes. The third level he called the ‘sub-systemic’ or ‘policy-shaping’ stage. At this level, policy networks were a useful concept for understanding how policy options were formulated in bargaining between the Commission Directorates General, national civil servants, and private actors (see Table 2.3). Levels here are not meant in the (territorial) sense used in MLG (see ‘Multi-Level Governance’ above). Rather, they are about the scope of the decision and political salience.

Table 2.3 Levels of Analysis in EU Decision Making


Type of decision

Dominant actors



History making

European Council, national governments in IGCs, Court of Justice

Political, legalistic


Policy setting

Council of Ministers, Committee of Permanent Representatives (COREPER)

Political, technocratic, administrative


Policy shaping

Commission, committees, Council groups

Technocratic, consensual, administrative

Source: Reproduced from Journal of European Public Policy, 2 (1995): 71. Reprinted with the permission of the Taylor & Francis Group.

Subsequently, Peterson (2009: 109) put his case more strongly, suggesting that ‘policy network analysis is never more powerful an analytical tool than when it is deployed at the EU level’. He set out a three-pronged argument for the applicability of the policy-networks approach to the study of the EU: first, that there is considerable variation in how different EU policy sectors operate; second, that much of the EU’s policy (p. 39) making is highly technical; and, third, that EU policy making is ‘underpinned by an extraordinarily complex labyrinth of committees that shape policy options before policies are “set” by overtly political decision makers such as the college of Commissioners, Council of Ministers, or European Parliament’ (Peterson 2009: 118).

Jachtenfuchs (2001: 254) made a complementary case for the advantages of network analysis:

It appears that, on the whole, the fragmented and fluid institutional structure of the EU and the lack of a strong power centre lead to an increase of channels of access and a larger variation of participants in the policy-making process as compared to governance systems in territorial states.

Related Literatures

Parallel to the Anglo-American policy-networks literature exemplified by the work of Rhodes is a related strand of conceptual and empirical literature on network governance that is most strongly associated with German scholars—especially Beate Kohler-Koch and her associates (see Kohler-Koch 1996; Kohler-Koch and Eising 1999). The key difference is that, while the British approach is primarily an analytical model seeking to understand state–society relations in a given policy area, the German approach treats networks as an alternative form of governance to states and hierarchies (Börzel 1998). It is therefore directly linked to the governance literature (see ‘Governance’ above), and specifically to governance by networks (Table 2.1).

Another literature used in EU policy analysis—this time drawn from IR—is that of epistemic communities. These are epistemic communities, defined by Peter Haas (1992: 3) as:

a network of professionals with recognised expertise and competence in a particular domain and an authoritative claim to policy-relevant knowledge within that domain or issue-area.

The members of an epistemic community share expertise and knowledge: that is, they hold a common set of understandings about what is right and desirable on that basis, (p. 40) and a common set of assumptions about how to achieve those goals. Epistemic communities are likely to exercise particular influence over policy when policy makers face conditions of uncertainty about the likely consequences of policy choice. Epistemic communities might therefore be used by supranational actors such as the European Commission as a means of furthering the integration of policy. The expert analysis that they provide, and the policy prescriptions that they advocate, if they point in the desired direction, could form a powerful lever for supranational actors to move states in the direction of common European solutions to problems that confronted them.

While developed separately from the Rhodes model of policy networks, the notion of epistemic communities is a compatible approach that provides a way of understanding how professionals can come to dominate policy making. Sabatier’s advocacy coalition framework, which offers an explanation for how policy change is brought about by coalitions within networks bound together by a shared belief system, has similar potential (Sabatier 1988, 1998). Moreover, Peterson (2004: 121) suggested that alliances of epistemic communities and advocacy coalitions may form to influence policy making and provides the example of the EU’s ‘quite radical liberalization of its agricultural sector during the Uruguay Round which gave birth to the World Trade Organization in the early 1990s’.

Policy Networks Assessed

Policy network analysis provides a valuable tool in understanding decision making at the policy-shaping level. The policy community–issue network distinction provides a good way to understand the different character of relations between the EU institutions and lobbyists and other actors across different policy areas. Like the epistemic community literature, that on policy networks is best in explaining technocratic policy making. However, the bigger-scale issues such as how the EU was given authority over the policy in the first place, or how policy goals were set, requires a different set of analytical tools.

EU Governance Approaches

When Hix made the plea to turn to comparative politics approaches, he was concerned to treat the EU like other political systems. There are, however, three approaches to understanding EU governance that do not comply. They have been developed to understand the specificities of EU governance. They are: supranational governance, which was developed in the 1990s; new intergovernmentalism, developed in the 2010s; and differentiated integration, a further analytical approach of the 2010s.

Supranational Governance

Starting from the intergovernmentalism versus supranationalism debate (Chapter 1), a team of scholars claimed to offer an alternative that cut through the dichotomy (Stone Sweet and Sandholtz 1997; Sandholtz and Stone Sweet 1998; Stone Sweet et al. 2001). Supranational governance was an approach that drew on the transactionalism of Karl Deutsch (Deutsch 1953; Deutsch et al. 1957) and on NI as applied to the EU (see (p. 41) ‘New Institutionalism’ above), although the authors themselves located the origins of their approach in neofunctionalism. Following Deutsch and working with the trend of globalization, Stone Sweet and Sandholtz (1997) argued that transactions across national boundaries were increasing. As they increased, so a supranational society of relevant actors would emerge. These actors would favour the construction of rules to govern their interactions at the supranational level because nationally based rules would be a hindrance to them.

Although with aspirations to be an integration theory, supranational governance’s strength was in understanding the EU as a series of regimes for different policy sectors. The authors therefore sought to explain the different extents of supranationalism that existed in different policy sectors. In two edited volumes (Sandholtz and Stone Sweet 1998; Stone Sweet et al. 2001) contributors explored the development of supranational governance in sectors such as air transport, telecommunications, the environment, gender equality, and monetary policy at a time when European integration was developing strongly, with the single market and moves to monetary union.

Supranational governance offered valuable insights into the specific dynamics of supranational rule making in individual policy areas. However, it was criticized for giving a privileged role to transnational business actors, to supranational actors, and to the operation of supranational rules of governance (Branch and Øhrgaard 1999). Indeed, supranational governance is susceptible to the same kind of critique that postfunctionalism has made of neofunctionalism and LI; particularly with respect to its excessive emphasis on the interests and influence of elite actors to the detriment of broader processes of politicization (see Chapter 1, ‘Postfunctionalism’).

New Intergovernmentalism

New intergovernmentalism, by contrast, has developed explicitly to capture the character of the EU in the period since the Maastricht Treaty of 1992. As its title indicates, the emphasis is on the enduring importance of national governments. New intergovernmentalism is based on six assumptions (Bickerton et al. 2015b, 2015c: 29–39):

  • that deliberation and consensus are the guiding norms of EU decision making;

  • that supranational institutions are not ‘hard-wired’ to seek ever closer union;

  • that delegation of new powers is now tending towards ‘de novo’ institutions (i.e. agencies) rather than traditional supranational ones, i.e. the European Commission;

  • that domestic preference formation has taken on a wider significance to the European integration process, extending out to include populism (and Euroscepticism);

  • that the difference between high politics and low politics has become blurred (contrasting with Hoffmann’s formulation of intergovernmentalism: see Chapter 1, ‘Intergovernmentalism’); and

  • that the EU is in a state of disequilibrium, reflecting both contemporary socio-economic problems and heightened political contestation around the EU itself.

New intergovernmentalism argues that co-operation has occurred in certain policy areas but without a clear reduction in national authority or a clear increase in supranational authority. The result has been that governance has been based on the institutionalization (p. 42) of intergovernmental exchanges and mutual oversight at the EU level, such as through new methods of government (see ‘New Methods of Governance’ above).

New intergovernmentalism has enabled insightful analysis of governance (institutional and policy aspects) in the more contested political environment of the 2010s (Bickerton et al. 2015b). Its assumptions contrast with those of supranational governance and highlight the way in which the changing political environment within which the EU is operating can impact on its governance.

Differentiated Integration

Differentiated integration has developed in European integration since the UK’s non-participation in the exchange rate mechanism of the European monetary system in March 1979. It set a trend that became more explicit with the UK’s opt-outs from monetary union and the social chapter of the Maastricht Treaty (Chapter 9, ‘Towards Maastricht’). It has now become an established pattern of integration where a member state cannot sign up to a Treaty reform or does not meet the conditions for joining a policy: a situation applying to various member states absent from the eurozone or the Schengen passport-free zone.

Leuffen et al. (2013) have taken differentiated integration one step further and used it as the basis to analyse governance in the EU. Instead of using the term just to highlight the way that in some policy areas not all member states are full participants—termed by them ‘horizontal differentiation’—they also look at the different extent of integration across policy areas (termed by them ‘vertical differentiation’). In this way, they develop differentiated integration as an analytical tool for comparing governance across different policies of the EU. Of course, they are not the first to conduct a comparison of policy making across sectors—see Wallace et al. (2020), the latest edition of a book first published in 1977. However, in today’s EU with its diverse membership, differentiation has become an important way of understanding varying patterns of integration and how they impact on EU governance and policy making (see Chapter 18, ‘The European Union’s Policy Agenda’).

EU Governance Approaches Assessed

The three literatures considered in this section of the chapter depart from the principle of utilizing theories that have applicability in other political systems. Instead, they provide frameworks based around the EU political system itself. They flirt with the risks that arise from being based around one set of circumstances, the EU: for instance, that they may lose relevance if circumstances in the EU change. Nevertheless, they constitute an important part of the theoretical literature on EU governance.


If theories of EU governance were relative latecomers to the study of the EU, their proliferation has more than made up for lost time. There are a number of overlapping and related approaches competing for the same space in analysing the operation of the EU system. We have distinguished (p. 43) between NI, governance, policy networks, and EU governance approaches. However, the reality is less clear-cut, not only between these categories, but also within them.

What brings together the approaches and theories discussed in this chapter is that they are primarily mid-range: they seek to explain developments at a sub-system or sectoral level and are not attempts at theorizing the EU system more broadly. As the EU has become more complex, there has been greater application of mid-range theories. Governance theories broadly defined have become more prominent with the increased practice of regulatory and new methods of governance.

After seven decades of governance in the EU and the European Communities, it is scarcely surprising that this chapter contains a diversity of approaches. EU governance has changed over that time frame, as have comparative and EU-specific theories. These theories certainly help the understanding of how EU governance works (see Table I.1)

However, this body of literature is subject to a broad critique; they have failed to address the questions addressed in Chapters 3 and 4 (see Table I.1). Thus, in joining the mainstream of EU studies—like integration theory before it—governance approaches have failed sufficiently to explore the political contestation within the EU. They have failed to understand some of the ‘givens’ in EU studies, such as underlying questions about power that have become more evident during the crises in the EU during the 2010s. They can be considered to be too descriptive, not asking enough normative/critical questions. In that sense it is also important to consider the consequences of the EU for member states and democracy: covered in Chapter 3. Whitman and Manners (2016: 14) go further and advocate alternative ‘dissident’ voices that:

do more than simply critique the existing discourses and practices of EU studies; they raise the possibility of speaking a different language of Europe, one that is critically aware that socio-economic power structures, systems of difference and narratives of exclusion are potentially embodied in all politics.

We explore some of these ‘dissident’ approaches in Chapter 4.

Key Points

A Shift of Focus

  • Approaches from the study of domestic and comparative politics turn away from the focus of IR theories on the process of European integration, and instead treat the EU as a political system, and try to explore ‘the nature of the beast’.

New Institutionalism

  • New institutionalism argues that analysts have lost sight of the importance of institutions in structuring political action.

  • Three varieties of NI can be distinguished: rational choice institutionalism, historical institutionalism, and sociological institutionalism.

  • Rational choice institutionalism emphasizes the argument that the behaviour of political actors is shaped by the specific framework of formal rules within which they operate.

  • Historical institutionalism emphasizes the argument that political relationships have to be viewed over time and that decisions are shaped by the nature of pre-existing institutional relationships.

  • Sociological institutionalism emphasizes the argument that the behaviour of political actors is shaped by informal norms and values.

(p. 44) Governance

  • Definitions of governance commonly emphasize the proliferation of non-state actors in the policy process.

  • Analytically it is possible to distinguish between governance by hierarchy, markets, and networks.

  • Multi-level governance, regulatory governance, and new methods of governance offer important concepts for understanding EU practice.

  • Multi-level governance argues that national governments have lost some control over policy to the supranational level, and is particularly useful for understanding terrritorially organized policies such as cohesion policy.

  • Regulatory governance captures the arm’s-length pattern of rule in the single market and related policy areas.

  • New methods of governance have been utilized in policy areas such as the Lisbon Strategy and Europe 2020, where the EU lacks clear Treaty powers.

Policy Networks

  • Policy networks offer a ‘meso-level’ approach that can offer insights into different policy outcomes, contrasting the pattern of relations between actors and institutions in policy communities with that of issue networks.

  • Policy networks are concerned with more specific policy issues. The origins of a policy and its main objectives require the use of other theories reflecting the level or scale of what is at stake.

  • The literatures on network governance and epistemic communities are complementary to the policy-networks approach.

EU Governance Approaches

  • Supranational governance helps understand the accumulation of governance responsibilities at EU level in the late 1980s and 1990s, but seemed less apposite by the 2010s.

  • New intergovernmentalism has been developed around the more contested character of EU governance in the 2010s, where national governments and domestic politics have re-asserted themselves.

  • Differentiated integration helps understand the complexity of EU governance: the different balance of powers across policy areas; and the pattern whereby not all states are full participants in all EU policies such as the eurozone.


  1. 1. Why did analysts turn to comparative politics approaches to explain developments in the EU?

  2. 2. How, and in what ways, do institutions matter in the EU?

  3. 3. Explain the ‘governance turn’ in the analysis of the EU.

  4. 4. How useful are policy networks in explaining EU politics?

  5. 5. Are ‘EU governance approaches’ helped or hindered by being based solely on the experience of the EU? (p. 45)

Go to the online resources to test your understanding with multiple-choice questions.

Further Reading

The call for a shift of focus to comparative politics approaches was set out in S. Hix, ‘The Study of the European Community: The Challenge to Comparative Politics’, West European Politics, 17 (1994): 1–30. His call has been implemented in book form as S. Hix and B. Høyland, The Political System of the European Union, 3rd edn (4th in preparation) (Basingstoke: Palgrave Macmillan, 2011).

Useful reviews of the new institutionalist research agenda in EU studies are offered by M. Aspinwall and G. Schneider, ‘Same Menu, Separate Tables: The Institutionalist Turn in Political Science and the Study of European Integration’, European Journal of Political Research, 38 (2000): 1–36 and M. Pollack, ‘Rational Choice and Historical Institutionalism’, in A. Wiener, T. Börzel, and T. Risse (eds), European Integration Theory, 3rd edn (Oxford: Oxford University Press, 2019), 108–27. The three institutionalisms are applied to cases in G. Schneider and M. Aspinwall (eds), The Rules of Integration (Manchester: Manchester University Press, 2001), 73–96. Useful for an overview of institutionalist and governance approaches is S. Saurugger, Theoretical Approaches to European Integration (Basingstoke: Palgrave Macmillan, 2013), Chapters 4 and 5.

In relation to the adoption of governance approaches in EU studies, see B. Kohler-Koch and B. Rittberger, ‘The “Governance Turn” in EU Studies’, Journal of Common Market Studies, Annual Review, 44 (2006): 27–49, and T. Börzel, ‘Governance Approaches to European Integration’, in A. Wiener, T. Börzel, and T. Risse (eds), European Integration Theory, 3rd edn (Oxford: Oxford University Press, 2019), 87–107. Also recommended is F. Scharpf, Governing in Europe: Effective and Democratic? (Oxford: Oxford University Press, 1998).

On MLG, the most complete statement of the approach is contained in L. Hooghe and G. Marks, Multi-Level Governance and European Integration (London: Rowman and Littlefield, 2004). The collection by I. Bache and M. Flinders (eds), Multi-Level Governance (Oxford: Oxford University Press, 2004) undertakes a critical assessment of both the potentialities and limitations of MLG, drawing both on theoretical contributions by scholars from different academic traditions and fields, and on different policy studies. On regulatory governance, see G. Majone (ed.), Regulating Europe (London; Routledge, 1996). On new methods of governance, see C. Sabel and J. Zeitlin (eds), Experimentalist Governance in the European Union: Towards a New Architecture (Oxford: Oxford University Press, 2010).

For an overview of the policy networks approach and its application to the EU, see J. Peterson, ‘Policy Networks’, in A. Wiener and T. Diez (eds), European Integration Theory, 2nd edn (Oxford: Oxford University Press, 2009), 105–24. The approach is applied at greater length in J. Peterson and E. Bomberg, Decision Making in the European Union (Basingstoke: Palgrave Macmillan, 1999).

Supranational governance is outlined in A. Stone Sweet and W. Sandholtz, ‘European Integration and Supranational Governance’, Journal of European Public Policy, 4 (1997): 297–317. It is developed in W. Sandholtz and A. Stone Sweet (eds), European Integration and Supranational Governance (Oxford: Oxford University Press, 1998). New intergovernmentalism is outlined in C. Bickerton, D. Hodson, and U. Puetter, ‘The New Intergovernmentalism: European Integration in the Post-Maastricht Era’, Journal of Common Market Studies, 53 (2015): 703–22, and developed at greater length in C. Bickerton, D. Hodson, and U. Puetter (eds), The New Intergovernmentalism: States and Supranational Actors in the Post Maastricht Period (Oxford: Oxford University Press, 2015). On the theory and analysis of differentiated integration, see D. Leuffen, B. Rittberger, and F. Schimmelfennig (eds), Differentiated Integration: Explaining Variation in the European Union (Basingstoke: Palgrave Macmillan, 2013).


Visit the online resources for a selection of web links to further sources of information on EU politics.