What is the decentered state?

Group theory

Group theory was in part an offshoot of American pluralism (Bartelson, 2001; Bevir, 2012; Eisenberg, 1995). British pluralists, such as G. D. H. Cole, generally sought to promote the role of guilds and trade unions within a future socialist society (Laborde, 2000; Nicholls, 1994; Stears, 2006). American pluralists placed much more emphasis on the inherently pluralistic nature of the state (Gunnell, 2004). Arthur Bentley (1908) famously argued, for example, that “all phenomena of government are phenomena of groups pressing one another” (269). This argument provided the basis for the study of pluralism and policy networks in the U.S. It shifted attention to the ways in which groups make decisions and the ways in which the relationships between groups impact policy.

The concept of a subgovernment is just one example of the bridges that led from an interest in pluralism to the study of policy networks. Subgovernments are generally informal and they consist of the actors that dominate an area of policymaking. Douglass Cater (1964) developed the concept in his analysis of power in Washington (17). Cater argued that Bentley’s group theory should be modified to reflect the fact that pressure groups affect all aspects of policymaking, not just legislation. In his view, Washington was increasingly fragmented with actors constantly trying to retain and extend their influence. More generally, subgovernment theory rewrote pluralism; a democratic optimism based on localism, diversity, and participation gave way to a pessimism based on the persistent role of hidden elites in policymaking. The state was dominated by iron triangles—coalitions of Congress, special interests, and bureaucrats—that took little note of public opinion (Overman and Simanton, 1986: 584).

In the late 20th century, then, group theory oscillated back and forth between optimism and pessimism (Jordan, 1981; Jordan and Schubert, 1992). Hugh Heclo challenged the pessimism of the concept of iron triangles. He argued it ignored the vast majority of policymaking process and “the fairly open networks of people that increasingly impinge upon government” (Heclo, 1978: 263). Heclo highlighted the role of issue networks composed of a large number of actors with varying degrees of dependence on one another and so with different abilities to move independently. Issue networks operate across many different venues, from the White House to city planning hearings. They are more fluid and permeable than are iron triangles and subgovernments with their stable relations among fixed members. They can be informal and temporary. Heclo did not just expand the analysis of networks to include informal actors; he also placed the network—that is the relations among the actors—at the center of analysis, not the actors and their agendas. On the one hand, Edward Laumann and David Knoke (1987) followed his lead, highlighting “resource exchange networks” (5) in which all the actors depend on one another. On the other, Randal Ripley and Grace Franklin (1987: 9) argued that the federal government generally remains the dominant actor within any network, drawing other actors together and defining their role.

By this time, group theory had been extended to the study of British politics. Grant Jordan and Jeremy Richardson explicitly grounded their influential study, Governing Under Pressure, in Bentley’s group theory. They claimed that the “interplay of group pressures is the dominating feature of the policy process in Western democracies” (Richardson and Jordan, 1979: 3). In their view, policy is made within issue communities, and it is almost always segmented due to the presence of so many groups within these communities. British government has been eroded; there is no longer a clear distinction between the central administration and “bodies outside of that perimeter” (Richardson and Jordan, 1979: 57).

Interorganizational theory

While Jordan and Richardson drew on group theory, other social scientists were using interorganizational theory to reach similar conclusions about the importance of networks in policymaking and so the fragmented nature of British government (Jordan, 1990; Thatcher, 1998). William Evans, writing in the mid-1960s, had shifted the focus of some organizational theorists away from intra-organizational issues toward interorganizational ones, prefiguring the way Heclo would shift the attention of group theorists toward the relations among groups. Evans argued that interorganizational analysis was crucial to understanding boundary relation problems and the relation of an organization to its environment. He thus introduced the idea of an “organization set” as an analytical tool (Evan, 1965: 220). The social scientist adopts a “focal organization” and examines the relations among the organizations that compose its set. Evans suggested that the organization set might be analyzed in terms of variables such as input versus output, comparative versus normative, size, concentration of resources, overlap in membership, overlap in goals and values, and boundary personnel. He proposed several hypotheses about the relations between these variables. For example, the greater the size of the organization set, the lower the autonomy of the focal organization. Other organizational theorists focused specifically on the implications of the dependence of any one organization upon other organizations. James Thompson (1967) in particular drew on exchange theory to posit that any organization is dependent upon other organizations.

In the late 20th century, these developments in interorganizational theory influenced the study of urban politics, policy networks, and governance. Students of urban politics, such as Stephen Elkin (1975), combined the insights of Evans and Thompson, highlighting the ways focal organizations manage their dependencies within their environment thereby enhancing their domains. Elkin outlined several strategies that organizations can adopt, including coalition, co-optation, exchange, socialization, and supra-organization. He concludes that the adoption of a strategy and its success depend primarily on an organization’s resources.

Unsurprisingly, the same themes—organizations, dependencies, resources, and exchange—are equally prominent in studies of policy networks (Börzel, 1998; Thatcher, 1998). Dave Marsh and R. A. W. Rhodes (1992) argue that policy networks consist of governmental and societal actors whose interactions with one another give rise to policies. Typically the actors are linked through informal practices as well as formal institutions, or even instead of such institutions. Typically they are interdependent; they can secure the outcomes for which they hope only by collaborating with one another.

Marsh and Rhodes (1992) developed an influential typology of policy networks. At one extreme, there are “policy communities.” These have a limited number of participant groups, with some others being deliberately excluded. The participants share broad values, beliefs, and preferences. They often meet frequently, with all of them interacting closely on any topic related to the policy area. All of them have significant resources or power, so their interactions consist of institutionalized forms of negotiation and bargaining. They are usually organized hierarchically so the leaders can secure the acquiescence of the members in whatever policies are agreed upon. At the other extreme, there are “issue networks.” These typically have far more participants. The participants disagree with one another so conflict, not consensus, is the norm. They also have unequal levels of power, and widely varying degrees of access, so their interactions are often primarily consultative.

So, by the late 20th century, numerous social scientists conceived of the state as composed of groups or networks. Some of these social scientists, most notably Rhodes, brought this concept of the state to bear on the changes associated with Thatcherism or, if we prefer, neoliberalism. Here Rhodes argued that there had arisen a new governance in which networks had proliferated. The state, he claimed, has been hollowed out (Rhodes, 1997; also Patterson and Pinch, 1995; Weller et al., 1997). The state has shifted from hierarchy—the bureaucracies of the traditional welfare state—by way of the market reforms of the New Right to a contemporary era of networks. The role of the state has changed from making policy decisions to coordinating the delivery of services. The state has become increasingly dependent on other actors with which it forms networks. Governance through networks constitutes an alternative to hierarchies and markets as a way of allocating resources and securing co-ordination. Networks rely on trust and co-operation, whereas hierarchies rely on administrative orders and markets rely on price competition. Networks are characterized by diplomacy, reciprocity, and interdependence.

Postfoundationalism

Group theory and interorganizational theory were part of the modernist revolt against the idealism and developmental historicisms of the 19th century (Adcock et al., 2007; Bevir, 2017). Certainly, some early advocates of pluralism and group theory drew on idealism. Nonetheless, by the middle of the 20th century, pluralism had become strongly associated within the behavioral revolution and its focus on informal topics like public opinion and interest groups as distinct from the formal institutions depicted in constitutions (Adcock, 2007). Critics even complained that group theory and interorganizational theory exemplified the behavioralists’ quest for a general theory that neglected agency, history, and institutions.

Modernist tropes have now become so dominant that many social scientists take them for granted as a kind of default common sense. Modernist social scientists typically seek mid-level, or general, theories by which to explain particulars such as the adoption, operation, and effects of a policy. They prefer explanations that are formal, as opposed to historical, precisely because they conceive of explanation as involving the subsumption of particulars under either a mid-level or general theory (Brady and Collier, 2004; King et al., 1994). Many social scientists would not even accept that an account of a particular case could count as an explanation, no matter how broad and abstractly that case is defined. In their view, explanations must be synchronic accounts of patterns that persist across multiple cases. They have used much ink discussing how to select appropriate cases in order to arrive at a valid mid-level or general explanation (George and Bennett, 2005).

Postfoundational philosophies have inspired some social scientists to reject modernism (Barry et al., 1996; Burchell et al., 1991; Finlayson, 2007; Gibbons, 2006; Glynos and Howarth, 2007; Howarth, 2013). Postfoundationalism is often overtly historicist or genealogical, emphasizing contingency and context (Bevir, 2010b; Biebricher, 2008; Foucault, 1984; Walters, 2012). It rejects the hubris of mid-level or comprehensive explanations that claim to unpack the essential properties and necessary logics of social and political life. It suggests that neither the intrinsic rationality of markets nor the path dependency of institutions determines whether policies are adopted, how they coalesce into patterns of governance, or what effects they have. Postfoundationalists conceive of public policies, instead, as contingent constructions of actors inspired by competing beliefs or discourses themselves rooted in different traditions or epistemes. It replaces aggregate concepts that refer to objectified social laws with historical narratives that explain actions by relating them to the patterns of meaning that produce them. Finally, this postfoundationalism has led some scholars to revise earlier concepts of the plural or hollow state. They have argued that postfoundationalism implies the state is stateless (Bevir and Rhodes, 2010).

The decentered state

So, to sum up, we can distinguish three different versions of the decentered state in the historical order in which they emerged:

The pluralist state: Pluralism suggests that the state was made up of different groups and actors. The state does not have the formal unity that idealists ascribed to it and that many earlier political scientists had treated it as having at least for heuristic purposes. As time went on, pluralism became associated with the empiricism and even positivism of the behavioral revolution. Yet, its claim is actually generally universal: the state is always and inherently plural.

Governance and the hollow state

As we have seen, governance scholars often argue, first, that the existence of policy networks means the state is fragmented or differentiated, and second, that the proliferation of policy networks since the 1970s has led to a hollowing out of the state. Clearly these governance scholars insist on the importance of networks. Other social scientists are, however, skeptical of the claims that governance scholars make about network governance. They argue that the concept of a network is unhelpfully vague and that a focus on fragmentation can obscure the continuing power of the state (Capano, 2011; Colebach, 2009; Marsh, 2011; Offe, 2008). Both governance scholars and their critics generally rely on formal explanations. So, for example, governance scholars explain the rise and content of network governance by appealing to a functional logic of differentiation (Rhodes, 1997). Decentered theory replaces these formal explanations with historicist genealogies and narratives. It thereby enables governance scholars to respond to the main challenges they face but in doing so it transforms their account of governance.

One challenge facing governance scholars dates back to the criticisms made by rational choice theorists of the literature on policy networks. Rational choice theorists asked: What does a policy network refer to other than the actions of individuals? How do policy networks explain anything? (Dowding, 1995). These questions are about social ontology and social explanation. Network theorists and governance scholars typically respond to these questions by appealing to institutionalism and mid-level theory as alternatives to rational choice and micro-level theory (Marsh and Smith, 2000; Rhodes, 1997). Unfortunately they do not spell out the philosophical content of their mid-level commitments. Sometimes they just wave the flag of “critical realism”, as if that phrase itself could magically answer the awkward philosophical questions. At other times they just evoke institutionalism as a longstanding and common approach, as if longevity and popularity could substitute for philosophical argument. Generally they appear to want to wish away the philosophical questions posed by rational choice theory in order to return to familiar empirical topics. Nonetheless, the implicit commitments of their mid-level theories are fairly clear. Mid-level theories involve a commitment to institutions or structures as existing apart from actors and their activity and as exercising a causal influence on actors and their activity. Mid-level ontologies typically reify norms, conventions, ideal types, and structures. Mid-level explanations typically appeal to formal systems, formal functions, ahistorical logics, and ahistorical mechanisms.

Too many governance scholars try to ignore the awkward questions about micro-theory that rational choice poses. They cling forlornly to mid-level theories that drift inexorably toward reification and formalism. In contrast, decentered theory unpacks governance as meaningful activity. To discuss and explain this meaningful activity is to ascribe desires and beliefs to the relevant actors. Actions can be understood only in terms of the conscious, unconscious, and subconscious intentionality of the actors. Unlike rational choice theory, however, decentered theory emphasizes the holistic and contingent nature of intentionality. Social scientists have to do the empirical work of finding out what beliefs and desires people actually hold in any given case. They have to rely less on formal models than on contextual and historical explanations. Thus, decentered theory concentrates not only on the construction of practices as people act on beliefs but also on the narratives and traditions that provide the context and historical background to people’s beliefs and actions. Decentered theory provides governance scholars with an alternative micro-theory to that associated with rational choice theory.

Another challenge facing governance scholars is the analysis of power. Mid-level theorists often want to ignore the micro-level and to focus on institutions and structures precisely because they believe that power is an important structuring force within social relations. Some mid-level theorists argue that concepts such as “differentiated polity” and “network governance” do not allow for the way in which power structures governance (Marsh, 2008, 2011). Decentered theory offers a response to this challenge in so far as it encourages social scientists to rethink power as a force lacking any center (Foucault, 1982; Torfing, 2009). If power refers to the ways in which the actions of others define what any individual can and cannot do, then power appears throughout state action. Power appears wherever people interpret and respond to one another. Every actor is both enabled and constrained by the actions of others. Prime ministers, senior civil servants, doctors, police officers, and everyday citizens all find their possibilities for action restricted by what others do. So, concepts like “differentiated polity” and “hollow state” rightly emphasize the diverse ways in which other actors thwart the intentions of the center. Governance scholars show how local actors—ministerial barons, Whitehall bureaucrats, doctors, and police officers—intentionally and unintentionally resist the core executive.

As decentered theory responds to criticisms of accounts of network governance, so it also transforms them. First, the decentered narrative of governance is not based primarily on policy networks. It is based on the claims that modernist social science inspired reforms promoting markets and networks, and that these reforms produced complex patterns of public action and organization (Bevir, 2010a). Because decentered theory presents the new politics as a product of modernist social science, it allows for varied consequences of public sector reform irrespective of whether, in any given case, they do or do not include the fragmentation of the state and the proliferation of networks.

Second, when decentered theory invokes a fragmented state or differentiated polity, therefore, it is not appealing to a functional logic of increasing specialization. Decentered theory points instead to a postfoundational critique of reified concepts of the state for their neglect of the varied contingent meanings and activities that make up the state. As we have seen, it implies less that bureaucracy has declined and networks grown than that the state is and always has been stateless. States have no essence, structural quality, or power to determine the actions of which they consist. The state is just an aggregate description for a vast array of meaningful actions that coalesce into contingent, shifting, and contested practices.

Third, when decentered theory addresses changes in the state, it does not presuppose a decline in hierarchies and a rise in networks (Capano, 2011; Capano et al., 2015; Davies, 2000; Goetz, 2008; Holliday, 2000; Richardson, 2017; Taylor, 2000). Indeed, it prompts suspicion of the implicit attempt to reify these concepts, individuate them, and so count their number. Decentered theory is interested, instead, in how the spread of new ideas about markets and networks led to changes in the state. On one level, decentered theory here engages issues of governmentality, notably the discourses and policies of political elites. On another level, however, decentered theory encourages studies of the myriad ways in which local actors have interpreted these discourses and policies, responded to them, resisted the intentions of the elites, and forged their own practices of governance.

Finally, decentered theory transforms the literature on network governance in a way that blurs the distinction between that literature and other leading accounts of the state, including, for example, asymmetric power, metagovernance, straightjacketed, and congested (Clifton, 2015; Jessop, 2013; Marsh et al., 2003; Skelcher, 2000). Decentered theory insists the state is stateless in that it should not be reified, but it does not imply that the state always has a particular level of power that it uses to act in a particular way in a particular relationship to other organizations. Many social scientists treat different accounts of the state as if these cut nature at the joints and captured its essence. In contrast, decentered theory rejects the idea that the state has an essence. Decentered theory implies that state activity is disparate, probably containing fluid examples of network governance, asymmetric power, metagovernance, and congestion.

Pluralism and democratic theory

Decentered theory transforms not only governance scholars’ discussions of fragmentation and hollowing, but also pluralists’ discussions of democracy. As we have seen, it replaces formal modernist accounts of the state with a historical narrative suggesting that the currently dominant approaches to social organization embody a contingent modernist form of expertise, and that this modernist expertise is flawed since it does not properly allow for its own historicity. Modernist social science inspires formal defenses of state planning, markets, free markets, and networks. It suggests that one or other of these organizational types is, at least under specified circumstances, ideally rational. In sharp contrast, decentered theory foregrounds the inherent contingency and contestability of human activity and so the variety and unpredictability of organizations.

Crucially, as decentered theory is an alternative to modernist social science, so it supports participation and dialogue as alternatives to hierarchies, markets, and networks (Bevir, 2010a; Griggs et al., 2014). Yet decentered theory, like postfoundationalism more generally, provides no great optimism about the prospects for this democratic alternative. On the contrary, it can inspire a bleak vision of a misguided modernism, biopower, or neoliberalism colonizing more and more of life (Bevir, 2010a; Foucault, 2008; Rose, 1989). The developmental historicists of the 19th century could appeal to teleological principles that they believed were guiding history to a benevolent end. In contrast, postfoundationalists have no philosophical grounds on which to postulate agents or processes of change that will end modernism. They may hope that the constant failures of modernist expertise eventually will lead policymakers to try more democratic alternatives, but that hope resides mainly in its performance as an argument.

Nonetheless, there is a bit more to say about what postfoundationalists might hope for. Because decentered theory rejects the mantle of modernist expertise, it cannot inspire a utopian blueprint in which a particular type of organization or action provides a cure-all. If social scientists want individuals to make choices for themselves, social scientists should typically leave it to the relevant actors to decide how best to promote participation and resolve policy issues. The decentered vision of a democratic future is thus a largely unspecified one. It has specific content mainly as a result of its involving a break with modernist expertise. A world after modernism requires a new type of knowledge. Social scientists should adopt a noticeably more interpretive approach in which practices appear as patterns of contingent activity explained by reference to the meanings within them and the historical contexts of these meanings. They should champion participation and dialogue as interpretive approaches to decision making more than as particular practices and institutions. This democratic future would not necessarily involve an end to bureaucracy. It would just require bureaucracies, or whatever replaced them, to rely on historicist and humanist ways of knowing rather than modernist ones. Policymakers should treat people as agents who can act for reasons of their own, rather than as dupes acting in accord with a fixed economic or sociological rationality. Policymakers should recognize the contingency of the stories they tell, and they should engage the targets of their policies in dialogue.

A more detailed democratic theory might do as much harm as good. Much of the existing literature prescribes detailed institutional arrangements and concrete practices. There is, for example, a growing literature that attempts empirically to identify causal factors that allegedly determine if and when deliberative democracy and collaborative governance are effective (Bächtiger and Hangartner, 2010; Jackman and Sniderman, 2006; Landa and Meirowitz, 2009). This literature has an ambiguous relationship to decentered theory and its democratic ideals. Almost all of this literature is sympathetic to these democratic ideals. Some of it may be compatible with them. Nonetheless, parts of the literature on deliberative democracy and collaborative governance ape the modernist expertise of which decentered theory would rid us. Some democratic theorists seem to aspire to formal classifications and correlations between deliberation, self-governing institutions, and specific outcomes. They claim that deliberation and self-governance have such and such effects at least under such and such conditions. They cloak themselves in the mantle of modernist expertise.

A menu of democratic innovations instead leaves it to democratic actors to decide which innovations to adopt in which contexts. Social scientists can just describe the innovations without purporting to have identified formal correlations and underlying mechanisms that explain the success and outcomes of these innovations. Instead of offering policymakers laws and models that seem to prescribe what practices or policies they should adopt to get certain outcomes, social scientists can encourage policymakers to learn by analogy from particular cases and stories (Turnbull, 2018b). Although the stories might involve generalizations about practices, the generalizations can be descriptive and historical rather than attempts at a formal and comprehensive theory.

Of course proponents of democratic innovations may need to show relevant constituencies that these innovations work. However, decentered theory leads to a different view of how to show that an innovation works. The case for an innovation can be made by telling stories about cases and learning analogically from those stories (Turnbull, 2018b). Even correlations and models do not offer secure predictions, but rather are themselves best thought of as stories. Whatever limits social scientists build into their predictions, people could arrive at new beliefs and actions outside those limits. So, social scientists cannot make predictions. All they can offer are informed conjectures that seek to explain practices and actions by pointing to the conditional connections between actions, beliefs, traditions, and dilemmas. Their conjectures are stories, understood as provisional narratives about possible futures.

Decentered theory encourages democratic innovations partly because it draws on an open historical ontology rather than a formal structural one. Yet, while decentered theory involves historical genealogies of contingent activity, some other postfoundationalists analyze democracy in quasi-structural terms. Jacques Derrida (1986, 1990) in particular has inspired some postfoundationalists to identify democracy mainly with extraordinary moments and eruptions that allegedly reveal aporias inherent in political life, such as the way law and authority inevitably carry traces of their absence (Honig, 2009; Mouffe, 2000). These postfoundationalists define democracy in terms of antagonistic struggles that forge and institutionalize stable regimes or that challenge seemingly stable regimes. They neglect democracy as an everyday activity of collective decision making. Derrida (1990) viewed collective decisions as leaps of madness lacking any justification, and his followers often treat collective decisions as products of agonistic struggles that cannot be adjudicated by fair procedures and shared moral values (Mouffe, 2000). In contrast, decentered theory focuses on the contingent beliefs on which people have acted to make and remake the organizations and practices through which they reach collective decisions. Democracy thus appears as a series of changing everyday practices (lacking any quasi-structural essence) by which people make collective decisions about how to govern themselves.

The stateless state

There seems little point in returning again to the arguments for a pluralist, hollow, or stateless view of the state. These arguments and the empirical evidence supporting them would surely have triumphed already were it not for the modernist desire to impose scholarly order on a messy world. Besides, Paul Cairney reminds us in his article of the classic arguments for viewing the state as plural and policymaking as decentralized. Policymakers have limited knowledge of and control over their environment. Of course some policymakers are typically more powerful than others, but that does not mean they are able to exercise control and realize their intentions. In addition, Matt Beech reminds us in his article that the era since the 1970s seems to have been characterized, at least relatively speaking and in some areas of policymaking, by a hollowing out of the British state. Various shocks, schisms, and ruptures associated with the global economy, neoliberalism, social norms, and global institutions have fragmented the state. So, we have both theoretical and historical reasons to believe that the British state is characterized by the kind of informal governance and collaborative networks that take center stage in the articles by Sarah Ayres and by Justin Waring and his coauthors.

Narratives

Decentered theory suggests that social scientists should pay more attention to the narratives or discourses by which elites make sense of the world and their interests in it. Moreover, the central elite need not be a uniform group, all the members of which see their interests in the same way, share a common culture, or speak a shared discourse.

The British state can seem to be dominated by adversarial political parties and their competing ideologies. Here the dominant narratives have been the Tory, Whig, liberal, and socialist ones. Beech shows in his article how Brexit, irrespective of its policy effects, has fed on and also transformed these narratives. The narratives of ardent remainers, pragmatic remainers, and leavers cross party lines and could remake British politics.

Yet, arguably, the narratives of British parties are not that important for policymaking. Historically, the most important narrative of the British state has surely been the Westminster Model. Cairney stresses the continuing importance of the Westminster Model in shaping the way policymakers think about evidence-based policymaking. Evidence-based approaches appeal to policymakers because they fit with the Westminster Model’s story of central government control. This story creates dilemmas for policymakers as the practice of policymaking constantly belies such control. Consequently, policymakers constantly try to project control in its absence in part by presenting choices as technical matters. Ayres too stresses the role of the Westminster Model, showing how it both framed and created problems for English devolution. Finally, Waring and his coauthors argue that the master frame for collaboration within health research networks relies on the key Westminster Model’s stress on local actors enacting national policy.

If the Westminster Model remains a crucial narrative in the British state, it has been challenged, or at least modified, by managerialism. The rise of a managerial narrative informs the dominant rationalities of policymaking I discuss in the next sub-section. The managerialist narrative was initially imposed by neoliberal politicians on civil servants who remained attached to the Westminster Model and the generalist tradition. Today, however, many civil servants and local managers adhere to managerialism. Waring and his coauthors write of tensions between, on the one hand, older professional narratives of clinical and academic autonomy, and on the other, managerial narratives of efficiency and productivity.

These tensions are prominent in local government where managerialist executives confront professionals. From one side, chief executives can obviously use the managerialist narrative to interpret “crisis” in ways that give them opportunities. In this narrative, crises call for creative disruptions that legitimize chief executives introducing programs of modernized management that they might otherwise not be able to. From the other side, Ben Clifford shows how local authority planners continue to draw on professional narratives and identities in order to interpret and adapt policies associated with managerial attempts to impose austerity.

Rationalities

Modernist social science inspired the public sector reform agendas that promoted markets and networks. However, this general pattern hides wide variation across cases. Other traditions of social science have influenced public policy in ways that vary across time, space, and sector. The state encompasses clashing rationalities, that is technologies that derive from social science and seek to govern conduct. These rationalities include nudge, resilience, collaboration, localism, and evidence-based policymaking.

I suspect that a focus on an allegedly monolithic neoliberalism obscures the diverse rationalities that now inform policymaking. The most important task for decentered theory today is arguably, therefore, to identify these rationalities and trace their different genealogies. Although it is impossible to undertake this task here, I can highlight examples from the articles in this special issue.

For a start, Waring and his coauthors highlight rationalities of collaboration in health research. These rationalities differ significantly from neoliberal ones that promote markets and competition. They derive from sociological and institutional traditions of social science that oppose the neoliberal faith in neoclassical economics and rational choice theory. Institutionalists argue not only that networks can be efficient in reducing transaction costs, but also that networks and partnerships foster collaboration and innovation and so promote economic growth especially in hi-tech environments.

In addition, Ayres shows English devolution drew on rationalities according to which decentralization and localism boost economic productivity and growth. These claims about localism are by no means intrinsic to the neoclassical and monetarist theories that gave rise to neoliberalism. Clearly, therefore, we need to look for the more specific social science theories that justified and promoted localism and trace their contingent histories. We might trace localism back through late 20th century communitarianism to sociologists such as Ferdinand Tönnies. Or we might trace it through the environmental movement to economists such as E. F. Schumacher. Perhaps we might conclude that the devolution agenda was a muddle containing disparate strands including both of those just mentioned.

Finally, allow me to sketch a brief genealogy of the evidence-based rationality discussed in Cairney’s article. Epidemiological studies of the effectiveness of drugs often rely on induction divorced from causation. Although there is little reason to assume that public policies resemble drugs or that social problems resemble diseases in any of the relevant respects, an evidence-based rationality has been extended to public policy. Randomized control trials identify a new policy intervention, determine the anticipated outcomes, and specify ways of measuring them. The investigator chooses control groups, whether comprised of individuals or institutions. The policy intervention is randomly assigned to the target groups with a designated control group. Using a randomly assigned control group enables the investigator to compare the impact of an intervention with a group where nothing has changed.

Resistance

As it is important to explore rationalities that have little to do with neoliberalism, so it is important to question the idea that official narratives constitute or correspond to reality. Other actors can resist, transform, and thwart elite agendas. State activity is a site of struggles not just between strategic elites, but between elites and other actors. Subordinate actors can resist the intentions and policies of elites by consuming them in ways that draw on their local traditions and local reasoning.

Resistance is, of course, far more variable than are official policy narratives and social science rationalities. Central governments often have relatively consistent stories to tell about the policies they favor. Further, they often draw on distinct trends in the social sciences to defend their policies. In contrast, local actors often draw on traditions that are distinctive to a particular profession, policy area, or territory. It is, therefore, hard to provide a schematic outline of key traditions and narratives that constitute resistance today. Nonetheless, resistance is always present between and within policy actors, and if we fail to allow for it, we will fail to grasp the way a policy works and also the ubiquity of policy failure. Resistance is present in Ayres’ account of the big delivery departments—health, education, and welfare—being reluctant to decentralize their budgets and activities. It is present in Waring’s account of regional actors’ differing interpretations of the policy of collaboration within health research networks. And it is present in Clifford’s account of planners defending and protecting their ideal of professional autonomy and their norms of good practice.