parliamentary and presidential federations, and found a consistent trend, albeit to different degrees, toward the evolution and use of MLG governance patterns and networks in each. But we were unable to identify one type of federation (e.g. a parliamentary federation) as being more likely to generate these MLG patterns than another (e.g. a presidential federation). And we could not yet indicate with any precision the specific institutional, cultural and socio-economic conditions that produce and shape these MLG patterns.
MLG versus multinational federalism
In recent years, this debate has shifted somewhat to a comparison between multilevel governance and multinational federalism. The exponents of the latter (multinational federalism) view MLG as a concept which gives priority to the underlying values of efficiency and political stability rather than community or justice for its minority national groups and citizens. Therefore, as an analytical tool, MLG is viewed as biased in favour of the dominant elites and the status quo (Gagnon 2011). This critique of MLG is one which is focused on normative rather than analytical and empirical differences between the two concepts. Although important, due to lack of space, we will be unable to address these normative concerns in this paper.
In our view, “federalism” and “multilevel governance” share a number of common characteristics. They both can be understood in at least three distinct ways: 1) as descriptive terms applied to concrete political systems, 2) as analytical/empirical constructs, and 3) as normatively positive or desirable forms. With respect to the first (concrete systems) they each perform similar political functions, including that of dividing power in order to combat authoritarianism, managing conflict and promoting cooperation between internal groups, and protecting minority rights. They have also both increasingly abandoned an exclusive emphasis on territoriality in devising methods of political representation, and now opt for a combination of territorial and functional bases of representation in the performance of political decision-making tasks. There has also been a notable trend toward increased emphasis on ethical and normative [36] concerns (or both) in decision-making, increased stress on the whole person rather than on the atomistic individual in societal action, and less attention to statist and hierarchical models of administration. However, as analytical and empirical constructs, the two concepts still have some important differences, such as: 1) federalism has a narrower and more restrictive reach than MLG, 2) federalism encourages greater formalisation of policy decisions and processes, 3) federalism fosters a broader climate of competitiveness, and 4) federalism tends to produce a more efficient, less costly and less time-consuming problem-solving governance process than MLG (Stein and Turkewitsch 2008).
MLG versus centralised unitarism (the Westminster Model)
The centralised unitary system model (or Westminster Model) refers to parliamentary unitary systems that are centred in political power terms on the national political executive or cabinet. This is the focus of a study by Bache and Flinders on Multilevel Governance and the British State (2004) with which we largely concur. They accept the point of view of most political analysts of the UK that that political system prior to 1998 can be described as a prototype of the centralised unitary system in the pattern of intergovernmental relations that operated between the national government, based in the Westminster Parliament in London, and the country’s various political and bureaucratic decisionmakers representing its territorial regions in England, Scotland, Wales and Northern Ireland during most of the nineteenth and twentieth centuries But their rather different argument, is that MLG works better as an analytical construct when applied to post-devolution UK intergovernmental relations since 1998. They maintain that the UK since that time has been transformed from a centralised unitary state of the “Westminster model” prototype into a decentralised unitary state in its intergovernmental relations involving Scotland, Wales and Northern Ireland (but not England) (Bache and Flinders 2004). Thus, Charlie Jeffery and Daniel Wincott refer to it after 1998 as a “lopsided state,” and Hogwood et al. (2000) describe it as “asymmetrical devolution.” Others described it as a “quasi-federal” system. There will be more discussion of this issue in later sections of the paper.
[37] Other Recent Theoretical Contributions to MLG
The concept of the “joint-decision trap” revisited
In an early contribution to MLG theory in 1988 that was subsequently widely cited, Fritz Scharpf had identified stalemating commonalities between the decision-making structures of the German federation and those of the multilevel EC/EU system arising from their mutual dependence on unanimous or nearunanimous voting procedures, which he as a “joint-decision trap”. In a 2006 paper (republished as a chapter in his edited 2009 volume), he continues to claim, with justification, that his original seminal analysis is “still basically valid”, even though the European Union had diluted its unanimity voting requirements to that of near-unanimity in 1992. But he concedes that this analysis needs to be complemented by a similar account of non-governmental policy-making processes in the “supranational-hierarchical” modes of governance by the European Community Bank (ECB) or European Court of Justice (ECJ) in the financial and judicial policy sectors.
Scharpf (2009) also reviews some theoretical extensions and modifications that he made to his joint-decision trap thesis in recent years. He notes in particular his focus on and use of a hybrid model during this phase, containing three different modes of EU-intergovernmental relations: 1) the “intergovernmental mode” (i.e. an “applied negotiation mode”) in which the “joint-decision trap” thesis readily applies and in which institutionalising national governments remain in full control at the lowest level of policy-making, 2) a combined or “mixed mode” of joint-decision-making that includes aspects of both intergovernmental negotiations and supranational centralisation) in which the “joint-decision trap” may or may not apply, and 3) a “supranational hierarchical mode” (exemplified by the European Court of Justice and the European Bank), in which the “joint-decision trap” does not apply. He also suggests ways that these impediments to efficient EU policy-making may be mitigated or overcome. He acknowledges that “the effectiveness of problem-solving in policy-making at the national as well as the European and international levels varies considerably from one policy field to another” (Scharpf 1997). And he agrees with other advocates of MLG that “the complexity of the multilevel European polity is not adequately represented by the single-level theoretical concepts of competing ‘intergovernmentalist’ and ‘supranationalist’ approaches.” But he also warns that “empirical research that focuses on multilevel interaction overemphasises the uniqueness of its objects of study or attempts to create novel concepts which are likely to remain contested by Europeanists and over-isolates this area from general theory and the political science mainstream” (Scharpf 2009). These arguments are, in our opinion, well [38] founded, and should be incorporated into an updated and current assessment of the relative strengths and weaknesses of MLG as an analytical concept.
The strengths and weaknesses of MLG
Simona Piattoni, in both her initial brief (2009) historical and conceptual overview of multilevel governance, and in her subsequent (2010) full-length volume on the conceptual, empirical and normative challenges posed by the theory of multilevel governance, defines MLG in broad and abstract terms as a dynamic three-dimensional concept involving 1) the relationship between the centre and periphery, 2) the relationship between state and society, and 3) the relationship between the domestic and the international. Each of these dimensions involves changes that occur at three analytical levels: 1) political mobilisation 2) policy-making and 3) policy restructuring. In this way, she claims to be able to generate a three-dimensional space within which to gauge MLG’s empirical scope and desirability in normative terms (Piattoni 2009:1).
But there are several major criticisms that one can level at the MLG concept, as Piattoni, among others, also notes. First, empirically it is unclear which phenomena MLG encompasses, and whether it can be significantly distinguished from “governance” in general or other similarly broad concepts. Secondly, its epistemological and ontological meaning is unclear, and it is not evident where it stands on the ladder of abstraction (Piattoni 2009: 1, citing Sartori 1984). Thirdly, normatively, it is uncertain whether political