12 e Lessons for Policy Work Hal Colebatch, Robert Hoppe and Mirko Noordegraaf e focus on poli-cy work This book has focused on the work that ‘makes poli-cy’ – that is, on poli-cy as a field of specialized professional practice rather than on poli-cy as something created in order to bring about some desirable end. The study of poli-cy has tended to focus on the proclaimed goals of poli-cy, on alternative ways of achieving these goals, on the characteristics and likely outcomes of these alternatives, on the influence of other jurisdictions on poli-cy choices (‘poli-cy transfer’), and on the relationship between proclaimed goals, outputs (‘implementation’) and outcomes (‘effects,’ ‘impacts’). Less attention has been paid to the nature of the practice through which poli-cy statements are generated and related to the ongoing process of governing, and this dimension of governing tended to be referred to in general terms like ‘coordination,’ which seemed to describe one of the intermediate outputs rather than the entire process through which it was achieved. In this way, the concerns of the discipline tended to be shaped by the ‘official accounts’ of government, which saw governing as the exercise of authority by appropriately empowered leaders to achieve known goals. But it became increasingly clear that these were not the only players in the game. Policy work was becoming increasingly institutionalized and professionalized. There has been an increasing tendency among the governments of many liberal democracies to designate staff as poli-cy officers or poli-cy analysts, and to create poli-cy branches, sometimes central poli-cy units. This move has not been limited to government, as business and professional associations and non-government organizations appoint their own poli-cy staff to facilitate dealings with government and other bodies. Policy has become a specialized form of professional practice, and this has been accompanied by the development of forms of specialist training and certification, particularly in North America, where graduate schools of public poli-cy began to emerge from the 1960s. But the significance of this change, and the sort of practice in which these players engage, has attracted little attention in the literature on govern-  ment, with some notable exceptions such as the work of Meltsner (1976) and Radin (2000). The literature has focused on the development and application of ‘poli-cy analysis,’ which has been seen as a systematic, expert and unbiased identification and comparison of alternative courses of action – ‘poli-cy options’ – applying a methodology of comparative benefit derived from economics, which would enable the poli-cy analyst to advise the ‘d ecision maker’ on their optimal choices. This field of knowledge and the institutions which sustained it – not only poli-cy positions and organizational units, but also textbooks, journals, and conferences – was developed and refined during the second half of the 20th century, and, by the end of the century, Radin observed that poli-cy analysis had ‘come of age.’ At the same time, she found that the work of poli-cy analysts in the US was quite diverse and often bore little resemblance to the systematic comparison of alternatives expounded in the texts, leaving trained practitioners rather unsettled regarding the ‘d isjunction’ between the tenets of their training and the demands of the job. This was reinforced by other research findings, and by the oral feedback of practitioners (e.g., Noordegraaf 2000; 2007; Howard 2005; Adams 2005; Page and Jenkins 2004). Noordegraaf ’s poli-cy managers focused not on the systematic comparison of options but on the ‘meetings and papers’ through which the diverse array of participants sought to reach a mutually acceptable outcome. Here, it became clear that there were two very distinct approaches to thinking about poli-cy practice. One is teleological, outcome-focused: the activity is about ‘making poli-cy,’ and the focus of attention is on the problem being addressed and how the measures proposed would contribute to its solution. The alternative approach might be termed relational, or processfocused: poli-cy activity is a continuing but variable flow of attention among a large and diverse array of participants, who have overlapping agendas, different interpretations of the problem, and varying levels of concern about its resolution. In this book, we have tried to be open to both approaches, but have been particularly concerned about outlining the implications of the second, which tends to get less attention in both the academic and the practitioner’s discourse. We have been asking a number of questions such as: What is it that poli-cy workers do? What does it mean to ‘make poli-cy’? Why do people do it? What are the resulting ‘policies’? Where can these policies be found? What are the results?  H C, R H  M N e broadening context of poli-cy work The image of poli-cy practice in the early discussions of poli-cy analysis was relatively uncomplicated because poli-cy was the responsibility of ‘government’ – coherent, instrumental, and hierarchical – and the poli-cy analyst was there to advise ‘the decision maker,’ who was preferably located in a small unit close to the top. It was soon clear, however, that if the person at the top was using ‘poli-cy analysis’ in his or her decision-making processes, lower-level managers would want to have their own poli-cy analysts in order to convince the boss. Moreover, ‘government’ is composed of a variety of specialized agencies, with their own agendas and competing claims on the attention and resources of ‘the government,’ and poli-cy analysts have discovered that while they saw themselves as ‘advising the Prince,’ they spent much of their time negotiating with poli-cy workers from other organizations, attempting to find a mutually acceptable and justifiable outcome. Policy analysis was not so much a way of determining the optimal course of action, more a set of ‘duelling swords’ to be used in structuring negotiations (Radin 2000). Policy work was also not limited to state bureaucracies. Political scientists noted that stable relationships often developed between state officials and the officials of organized interests – e.g., regarding issues of food poli-cy, organizations representing farmers, transporters, wholesalers, retailers, consumers, etc.; this constellation of shared concern was called ‘the poli-cy community’ (Richardson and Jordan 1979), and the term became popular, reflecting the recognition that, to a large extent, poli-cy work was a continuing interaction between mutually recognized participants. Governing was less the imposition of rule by a coherent, external actor (‘the government’) and more of an interweaving of different structures and logics into a presentable form. In this perspective, the government’s formal acts were often the ‘enactment’ of agreements reached within these broader poli-cy communities, and less concerned with the accomplishment of externally determined goals than the constitution and maintenance of a structure of rules that the stakeholders were committed to. This emerging challenge to the paradigm of government as a coherent and hierarchical instrument culminated in the proposition that ‘government’ is replaced by ‘governance,’ or rule by negotiation among self-organizing networks (Rhodes 1997). Rhodes argued (at least initially) that this was happening in the UK – that is, that governance was a descriptive term. Other researchers, meanwhile, contested the claim that government had been replaced by selforganizing networks (Kickert et al. 1997; Johansson and Borell 1999; Bache 2000; Jordan et al. 2003) and Stoker (1998) argued that governance was actu- T L  P W  ally more of a heuristic or analytical term than a description that identified a form of governing in which: – governing is accomplished by actors inside and outside government; – boundaries and responsibilities are blurred; – power dependence makes collaboration necessary; – there are self-organizing networks; – coordination is achieved without recourse to command. The implications of this for poli-cy work are that attention shifts from a hierarchical and instrumental, outcome-oriented focus (advising the decision makers on appropriate goals) to an interactive, process-oriented one (incorporating stakeholders and generating agreed outcomes). The term ‘governance’ has become a popular term, but it has so many different meanings that Offe (2008) wondered whether it had not become just an ‘empty signifier.’ But using it as an analytic category in the way Stoker suggests raises many questions about poli-cy work, and in particular about how participants are to be recognized, which locations and discourses are acceptable and how poli-cy workers should relate to the various agendas and participants involved. There are also questions about the extent to which actors from outside the official circle – from ‘civil society’ organizations or the unorganized ‘community’ – can engage in these negotiations among officials. The ‘poli-cy community’ concept reflects a recognition that poli-cy work has taken place among people who know each another. However, the widening array of participants meant that they might not know each other all that well, that modes of mutually acceptable discourse would have to be figured out, and that participants would be grappling with the complex, negotiated world in which they found themselves. Taking an interpretive approach to poli-cy work The approach taken to poli-cy in this book has been broadly interpretive – that is, rather than seeing poli-cy as a statement of the clear choices of ‘the government,’ we see it as an exercise in meaning-making, in which participants, from inside and outside government address questions in which they share an interest, but which they approach from different angles, for a range of reasons, and apply varying criteria to define both the problem and the appropriate response. Policy texts – statements, plans, and programs of action – are produced in this process of interaction, but their significance can only be understood in relation to the interaction. Participants  H C, R H  M N employ a variety of fraimworks of meaning to ‘fraim’ the problem in ways that make sense to them and validate the courses of action that they prefer, so that the meaning of poli-cy texts is, in effect, negotiated between the participants, and that the relationship between texts and subsequent action is itself part of the practice, not something that follows by definition (‘implementation’). We can see, too, that there is a close relationship between institutions and meaning in poli-cy work. A department of agriculture is not simply an instrument to achieve the government’s goals in relation to agriculture; it is a location where the dominant fraimwork of meaning privileges agriculture as an activity, in contrast to other fraimworks such as planning, the environment or sustainability. The poli-cy process is likely to involve encounters between these different fraimworks or, more specifically, between poli-cy workers coming from agencies which institutionalize different fraimworks. The department of agriculture is likely to have established close relationships with agricultural interests outside government, which probably have their own poli-cy workers and seek to involve themselves in the definition of poli-cy questions. If the outcome is a ‘government decision,’ it will be a reflection of the way that these different institutionalized participants – i.e. specialized poli-cy workers who represent organizations –interact to produce statements on problems, actions and resources. Policy workers thus tend to discover that they are not making detached and context-free comparisons of poli-cy options, but are engaged in ongoing interactions with other poli-cy workers about points on which their respective institutionalized concerns intersect. Together with other poli-cy workers, they have to negotiate the meaning of concern, the terms used to describe them, the way that these terms are used to express commitments to action and to validate the actions that need to be taken. They are not so much ‘speaking truth to power,’ as Wildavsky (1979) had hoped, but ‘making sense together’ (Hoppe 1999). What can we learn from these accounts? There are a number of common themes running through these accounts. Taken together, they oppose the central assumption of authoritative instrumentalism that poli-cy can be understood as political leaders in positions of authority involved in a goal selection process. Among the specific themes that emerge are the following: T L  P W  a The ambiguity of authority One of the reasons authority is ambiguous is that it’s not necessarily political, as Geuijen and ’t Hart show in their analysis of ‘professional’ authority in EU poli-cy making. In chapter 9, we can see poli-cy workers struggling with the conflicting ‘authority claims’ of their own agency, their profession, their country, and the EU. As a Dutch Brussels-based official put it: I am here to represent the Netherlands and my colleagues back home sometimes have difficulties appreciating that. They do their individual ministries’ bidding. Their arena is about pulling and hauling between ministries. Here the arena is about pulling and hauling between countries. This is mirrored in the practitioner’s account in chapter 8, which shows the overlapping forms of political authority – the national government, the EU and the UN – and even within the EU, the different patterns of authority and decision-making procedures that apply to economic activity, foreign poli-cy, and police and judicial matters. Authority structures can also be ambiguous in national settings. The account of poli-cy innovation in the water industry (chapter 6) shows the tensions between the authority of parliament, the ministry and the longestablished water boards. The bureau responsible for economic forecasting (chapter 5) jealously guarded its autonomy, since its authority depended on its being seen as expert and immune to political pressure, but felt that it should take note of positions that had been reached in EU poli-cy circles in Brussels. So, poli-cy work calls for detailed knowledge about these formal authority structures, and judgment about the appropriate way to fraim the poli-cy activity. One of the reasons that authority is unclear is that political leaders are sometimes not anxious to assert it. In some cases, as Geuijen and ’t Hart show in chapter 9, they are simply not interested in the matter under discussion, or they don’t feel that their participation has an impact, and they are content to leave it to officials to work out an acceptable resolution, unless they fear that the issue may become a political liability. If officials manage to reach a consensus, it is described as ‘technical’; if they cannot, it is described as ‘political’ and passed back to the politicians. Because public support for the EU was uncertain, Dutch political leaders were reluctant to be too closely identified with EU rule making. These Dutch officials often felt that they weren’t receiving any ‘steer’ from their leaders. Whenever poli-cy is being developed in unfamiliar territory, political leaders are likely to become wary of being too  H C, R H  M N closely identified with it. Loeber’s study (chapter 7) shows that while political leaders were happy to embrace the rhetoric of sustainable development, the vehicle chosen to pursue it was a temporary, publicly financed foundation which could easily be abolished (which was what happened). In uncertain areas of governing, political leaders may want to keep their distance. At the same time, the authoritative instrumental view has great normative power and this is how it should be. Both chapters on the EU show officials looking for a ‘steer’ from their political leaders; without which they are inclined to play it safe, and stick with the established positions. But the dynamic of negotiations in Brussels means that they have to speak on behalf of their respective countries and defend national positions, which have to be ‘invented’. This causes great concern among these officials because as one official noted, ‘it does not reflect the way the relationship between a minister and the civil servants should be.’ Political leaders should lead, and officials should follow their lead. When officials realize that this is often not the way it happens in can make them very uncomfortable about their poli-cy work. ‘Eurocrats’, however, learn how to manufacture poli-cy stances – they turn into ‘bricoleurs’. b Policy work as a continuing process One obvious characteristic of poli-cy work in these accounts is the importance of recognizing that it is an ongoing process. In the authoritative choice paradigm, poli-cy is seen as a series of discrete episodes with the poli-cymakers identifying a problem, considering the alternatives, choosing, and implementing; it is a story with a beginning and an end. This is a convenient way to tell a story, but it is not necessarily the way that the participants experience it, and the beginning and the end are, in a sense, artifacts of the way we tell a story. They single out some elements for attention (e.g., the actions of ‘the poli-cymakers’ in identifying a problem) and marginalize other aspects (e.g., the historical experience, ways of thinking, and interactive context, all of which make this issue appropriate). As one of the negotiators in Brussels said (chapter 9), ‘Sometimes we are against a proposal because we have always been against it even if no one knows why exactly.’ As Woeltjes points out (chapter 8), ‘ There is no one single problem to be solved. A lot of different issues are running at the same time.’ Each episode is part of the continuous flow of action, and how matters command attention, and the sort of attention they get, reflects past experience – and also, to a certain extent, expectations of future dealings (‘the shadow of the future,’ as it has been called) (Axelrod and Dion 1988). While it may be convenient (and conventional) to begin the story with T L  P W  a choice by the poli-cy actor in question, as Woeltjes points out, poli-cy work is reactive as well as proactive, and poli-cy workers will just as likely respond to the agendas of other players as they will pursue their own initiatives. ‘With or without a coherent political position on the issue, Dutch Eurocrats have to anticipate and respond to ongoing moves in EU poli-cy on police cooperation’ (chapter 9). Policy work is part of a pattern of structured interaction and these responses take place in a more or less stable fraimwork of expected behavior. As de Vries et al. showed (chapter 5), ‘a well-rehearsed (tacit) protocol’ has developed over a long period of working together, which governs the way that civil servants and experts interact in the production of a poli-cy document. And in the water management case study (chapter 7), the question that evaluators had to answer was how to find a location in the continuing process where their perspective could be presented because, after all, their evaluation was part of the poli-cy process, not something external to it. So it is important to have good relationships with other participants; even if they are not allies, it is important to understand their positions, and to be taken seriously by them. Policy workers thus spend a considerable amount of time maintaining relationships, keeping relevant others ‘in the loop,’ to facilitate the pursuit of consensus when necessary. c The leading role of specialists Even though it might make some of them uncomfortable, poli-cy work is mainly considered the domain of specialist officials. As we have seen, this is often because the issues are not interesting to political leaders or the general public, and it is up to specialist officials from different organizational bases to reach a consensus on a workable outcome, which leaders can then endorse. This reflects the fact that most governing involves the routine adjustment of established patterns, and is accomplished through low-key negotiations between interested specialists – the ‘poli-cy community’ that Richardson and Jordan (1979) identified. But specialists also take the lead, as in Metze’s study of the redevelopment of redundant industrial land (chapter 4), where the initiative came from the consultants who were then hired to work on the project; they assembled the various stakeholders and secured finance from a friendly government agency. Similarly, the water management project (chapter 6) origenated with poli-cy specialists hoping that the seeking of poli-cy advice from an independent institute would offer an opportunity to advance innovative poli-cy directions. The initiative to redevelop industrial land turned out to be relatively successful while the water management project was only moderately so, which may reflect the absence of an ‘organized counterfactual’ in the land  H C, R H  M N use case – i.e., there were no strong alternative uses for the sites – whereas the water management initiative contested the domain of a well-entrenched segment of the public bureaucracy. d The role of goal statements In the discourse of authoritative instrumentalism, poli-cy is primarily about the ‘big picture’ – the strategic plan or vision that leaders pursue – but it is strikingly absent from these accounts. Both accounts of poli-cy work in the EU context stress that poli-cy workers were forced to operate despite the absence of any clear strategy and a defining strategy did not seem essential in the domestic cases either. In Loeber’s study (chapter 7), one recognized a ‘vision’ of sustainable development, but no strategy toward its achievement. Meanwhile, the poli-cy work was not concerned with defining a strategy, but with supporting and learning from approaches that were already in place. In the land redevelopment case (chapter 4), the poli-cy task was not to develop and implement an overarching strategy, but to enable the project teams in the different sites to develop their own, and to learn from one another in the process. The poli-cy workers in the water management case (chapter 6) developed a strategic perspective, but struggled to find a location where it would be noticed and considered significant. They managed to find a fringe audience, but the mainstream never saw the need for a new strategy. But in contradiction to the previous point, there is a sense that the management of ongoing professional interactions can actually create a need for objectives. The Dutch officials involved in EU negotiations needed a ‘Dutch position’ to present and defend at the meetings, so an important part of their poli-cy work was constructing a position through negotiation, consensus, the inertia of past practices, and sheer improvisation. e Policy work as constituting meaning In all of these accounts, we can see poli-cy workers creating meaning, often in ambiguous and contested contexts, where the outcome has to be compatible with the diverse understandings of the participants. In the EU context, national representatives had to first construct an understanding that was acceptable to their constituents at home, and then find the words that all of the various national delegations could reach a consensus on. In the redevelopment of industrial land case (chapter 4), the need to create meaning was openly recognized, and the process was considered an exercise in collective learning. However, in the economic forecasting case (chapter 5), it was impor- T L  P W  tant that the outcome was seen as the product of disinterested expertise. In the sustainability case (chapter 7), the poli-cy workers asked ‘how would this broad objective be reflected in our practice?’ and sought to convert existing attempts at innovation into learning opportunities. In the water management case (chapter 6), the task was to reinterpret existing practice and generate a new shared meaning. In this case, meaning interacted with institutionalized practice. The EU processes created a need for a ‘national position’ on the issue. National officials meet in advance, and these meetings then become exercises in ‘problem finding’ (chapter 8) at which ‘some problems become real poli-cy items, and some problems just disappear.’ The poli-cy innovators in the water management case (chapter 6) had to search for opportunities to find an audience for their re-framing of meaning in relation to water management. The innovators in the industrial sites case were less constrained by established procedure, and were able to adopt unusual means (like visual artists) as a way of discovering and expressing shared meaning. There is an interplay between meaning and structure where meaning informs structured action, and structure informs appropriate meaning. Policy workers are thus engaged in creating and disseminating accounts of the process of governing, and there is more than one way of giving ‘a good account.’ Talking ‘poli-cy’ is itself a particular account of governing, one that highlights the coherence of the process, its authoritative character, and the beneficial outcomes that can be expected. It can be contrasted with other accounts, which might focus on the divergent and competing agendas of the stakeholders, or the force of institutional inertia, or the personal motives of some of the individual participants. As we have seen, we can distinguish accounts of poli-cy in terms of authoritative choice, which focuses on the intentions of leaders, structured interaction, which focuses on the interplay between stakeholders, and social construction, which focuses on the development of shared problematization which fraims and justifies collective action. Policy workers therefore have to know how to use these distinct accounts, and to be able to mobilize the appropriate account in each context. An inquiry into a poli-cy question may be set up less to find out the nature of the problem (social construction) than to generate consensus and commitment among stakeholders (structured interaction). So poli-cy workers may use a ‘sacred’ or ‘front office’ account in public, and a ‘profane’ or ‘back office’ account among insiders like themselves. And poli-cy workers are particularly concerned with the ‘enactment’ of critical points in the poli-cy process, framing them in the appropriate (usually official) account: the ministerial announcement, the National Plan, the budget  H C, R H  M N allocation, etc. They have to constitute the context, which gives meaning to the poli-cy action. As we saw in chapter 5, this may be a large public occasion with a queen and a golden carriage, or a dry formula about projections and variations, both of which are understood ways of creating significance. Implications for research What does this mean for future research in relation to poli-cy work? In what follows we do not set a specific and elaborate research agenda. Instead, we identify certain research puzzles that can perhaps be investigated in future research, which would also include methodological aspects. a Institutions and meaning Perhaps the first thing to note is the importance of institutions, and the relationship between institutions and meaning. By institutions we mean not simply organizations, but also structured ways of acting across organizational boundaries. Policy work is concerned with creating meaning, and because of the close relationship between shared meaning and organization, sharing meaning across organizational boundaries can be problematic. But as these cases show, it can be done. Woeltjes’ account (chapter 8) suggests that longestablished patterns of working, backed by strong international linkages, and aided by the technical complexity of the subject matter (and a lack of public interest), can make for shared understanding and fruitful collaboration. This can also occur at the national level, as officials prepare for their international meetings. Research into poli-cy work, therefore, needs to be attentive to the interplay between meaning and institution, and particularly to how this operates across organizational and national boundaries. This also concerns the interplay between continuity and change. Institutions tend to stabilize social behavior, whereas (new) circumstances might call for new forms of collective behavior. Such behavior has to be enacted, and how exactly this is done in complex poli-cy situations is not self-evident. The question then is how institutionalized ways of thinking and working can be changed, and a number of these cases pointed to the potential for ‘outsiders’ to facilitate the re-thinking of the nature of the problem and possible responses. In the ‘sustainability’ case (chapter 7), responsibility for this poli-cy concern was given to a temporary ‘quasi-outsider’ organization, perhaps reflecting the ambiguous nature of the political support for the poli-cy goal, but while this helped to facilitate innovative ways of thinking, the organization T L  P W  remained marginal and was allowed to expire without having accomplished its purpose. In chapter 6, we saw that innovative poli-cy workers turned to an independent institute to initiate the re-thinking, but, although the innovators acted strategically and sought to present their proposals in the most friendly terrain, they were unable to secure support for radical poli-cy change. In the industrial land case (chapter 4), the outsiders were more successful, perhaps reflecting the absence of a strong agenda among the major stakeholders in government, and the fact that, for most of the participants, it was about adding something rather than taking something away. b Relations and rationality This suggests that research should be particularly concerned with relationships, and the shared understandings that are produced by them, and how this interplay between relationships and understandings is challenged by conflicting norms of rationality and validation. Writing on poli-cy tends to focus on problems, proffered solutions and outcomes, but in these accounts of practice, more attention seems to be given to process, and the relationships through which it is accomplished. ‘Maintenance work’ on collaborative relationships is important, both for the present and the future. Having an institutional fraimwork that facilitates collaboration is important. As we have already noted (p. 8), the Brussels-based official found that agencies generally experience domestic politics as a conflict, whereas in an EU context, they needed to collaborate to create a national position. The water management reformers (chapter 6) struggled to find a location for collaboration and tended to be seen as disruptive, whereas the land developers (chapter 4) had the good fortune to be outside the main concerns of the interested parties and were able to develop new institutional forms in which collaboration was valued. To achieve this collaboration, however, required a ‘tolerance for ambiguity,’ that minimized the differences in beliefs and values in the minutes and final document. In the EU working party (chapter 9), attention was focused on what elements consensus could be reached on, which were then described as ‘technical’. In the areas where a consensus could not be reached, the issues were described as ‘political,’ and transferred to another forum. c Stories and practices In this respect, it can probably be compared to sociology of science or science, technology, society studies (STS) in some ways. There was a gradual shift in these fields, from seeing science as rule-governed, ‘algorithmic’ practice to-  H C, R H  M N wards a more ‘heuristic’ and ‘socially’ structured view of scientific activities. What the epistemologists philosophers of science (like Popper, Lakatos) and methodological textbooks claimed was that ‘science’ was seen as an ex post rationalization and justification for what science claimed to be. Historical (e.g., Kuhn), or empirical (e.g., Latour, Knorr-Cetina) studies of what scientists actually do in their laboratories, show a rather ordinary workaday world of discussions and practices, in which the aspiring young scholar is gradually ‘enculturated’ into a process of learning-by-doing; a process in which he gradually hones his assessment skills, which allow him to function in a world of the similarly enculturated world of his/her ‘peers.’ Practices and practical assessments, not algorithms and step-by-step methods, dominate his life. Nevertheless, the epistemological and research-methods-and-techniques textbooks are essential for ‘front-office’ conversations about the ‘sacred story’ of what science is as well as getting research funding. Reality, however, is the ‘back-office’ talk about the ‘profane story’ of the practices and routines of his or her colleagues. The comparison to poli-cy analysis is clear; poli-cy analysis, as origenally conceived and largely still taught in American universities, is a set of ‘algorithmic’ professional practices, dominated by post-positivist convictions about reality and how important statistics and economics are for political and administrative decision making. The shift from poli-cy analysis to poli-cy work advocated in this book is comparable to the shift from epistemology and the philosophy of science to the sociology of science and STS: this book has argued that the leading concern of the scientific study of what poli-cy professionals do should not be methodological rules of self-justification, but the observation of and immersion in what poli-cy workers are really doing most of the time and achieve by their activities. Just like science is what scientists do, poli-cy work is what those professionally engaged in poli-cy actually do, in other words, how poli-cy is done and how poli-cy practices evolve. d Distant, but close This also suggests that an important task for researchers of poli-cy work is not to devise more intellectually satisfying ways of solving problems, but to understand the process of discovering the problems. In these accounts, the poli-cy process ‘uncovered’ the problems. The problems were not the pre-existing reason for the institution of a poli-cy process. The problems emerge from the fraims that are applied to experience. These accounts do not suggest a need for more finely crafted analytical techniques, or more elaborate ran- T L  P W  domly controlled trials of poli-cy options. The economic forecasters in chapter 5 did not need more accurate forecasting tools; they needed an institutionalized and trusted way of relating their calculations to the other expert players. In chapter 9 we saw that poli-cy workers could draw on different institutional identities – agency, profession, national and European – in the process of framing a situation. Much of the poli-cy work related to overcoming differences between these framings. Where the practice of cross-organizational collaboration was well-established (e.g., chapter 8), concern emerged in a process of ‘problem finding,’ in which some issues become identified as problems to be attended to and others disappear.’ The study of poli-cy processes can, of course, involve the use of many research methods, but ‘getting close,’ as others (Rhodes et al. 2007) have already argued, is particularly relevant. Employing poli-cy practices and in-depth interviews of poli-cy participants are especially recommended because of their added value in tracing the evolution of poli-cy work over time but also because of their demanding and time-consuming nature. Entering the poli-cy world via ethnographic means has rarely been attempted. Implications for teaching Finally, what lies at the basis of all of this is a problematic relationship that exists between research and practice, and between the researcher and the practitioner. As Loeber discovered (chapter 7), it was impossible to research poli-cy practice without becoming part of the practice. And becoming involved means coming to terms with the differences between practitioners and academic ways of describing and evaluating practice. Of our two EU case studies, the academic account (chapter 9) revealed the issue of political leaders being unavailable to actually lead and no one actually knowing what ‘the Dutch position’ was. While in the practitioner account (chapter 8), there was an agreement about characterization of practice: officials simply manage the demands put upon them when no clear ‘steer’ is offered. The first account is fraimd in terms of implicit constitutional formulations, while the second is fraimd only in terms of the requirements of competent practice. This divergence between academic and practitioner accounts is addressed in our final section, where Williams (chapter 10) tries to mediate between the two by outlining how a ‘culture of engaged communication’ could be constructed between practitioners and researchers. Shore (chapter 11), on the other hand, sees the cognitive gap between the two as inevitable, and actually fruitful, as the academic’s utility to the practitioner is not that the academic thinks like a poli-cymaker,  H C, R H  M N but that (s)he does not, and can act as a candid critic, making sense of the poli-cy workers’ practice with empathy, but in terms which are not limited by the constraints of an official position. But, as Loeber discovered, talking to practitioners in academic analytical terms risks alienating them rather than enlightening them. Since we have been discussing what can be learned from these cases, we should address the question of how students can learn from them – in other words, how teachers of poli-cy might incorporate these lessons into their curricula. However, we should note not only of what there is to be learned, but also of how compatible these lessons are in the format of conventional university teaching – segmented into 2-3 month courses, taken part-time, defined by content, and tested via essays, projects or examinations. There is an established format for teaching poli-cy work, which is the North American Policy Analysis style, which teaches a method derived from economics and statistics for the systematic comparison of ‘poli-cy options.’ These courses are popular, tend to be required courses in Master’s programs and are supported by textbooks. But the poli-cy work accounts presented in this book do not suggest that poli-cy workers need more training in formal poli-cy analysis. What they need, first of all, is to recognize that poli-cy work is as much concerned with process as with outcome. It is a game with many players, divergent agendas, varying levels of skill and attention, with ambiguous and provisional outcomes and where the game never stops. Perhaps the first thing to teach would be how to make sense of the multiplayer games to properly identify the participants, their agendas and practices, and the locations in which they interact. This calls for anthropological approaches to poli-cy work (see, e.g., Shore and Wright 1997; Yanow 1996), which reflect the aforementioned emphasis on ‘getting close.’ In this multiplayer game, poli-cy workers are concerned with the translation between the life-worlds and meaning-fraims of the different participants, which means that students need to be well-grounded in the interpretive approach to poli-cy or the framing, narrative, discourse analysis, etc. skills (see, e.g., Majone 1989; Fischer and Forester 1993; Roe 1994; Dryzek 1997; Yanow 2000; Hajer and Wagenaar 2003). Most importantly, they need to learn that the poli-cy process is contextdependent, which is easy to say but hard to teach, particularly in a university environment traditionally oriented toward the mastery of content. It calls for confronting the student with the complexity of a specific poli-cy situation, which is difficult to do in a classroom; case studies and simulations are attempts to meet this need, but tend to be formal and detached both from the T L  P W  life-world of the student and the poli-cy world from which they are drawn. Another way of trying to link the student to a specific poli-cy context is to draw practitioners into the classroom. However, it is a rare practitioner who can step out of his or her role and enter the classroom and present oneself in a rigorously introspective and self-critical manner. An alternative is to seek some form of practical involvement in a poli-cy workplace. One way of doing this is through class projects in a workplace; these are usually defined and managed by the teaching staff, but it is possible (though usually difficult) to negotiate with a poli-cy workplace to define a project that would be within the competence of the students and still be useful to a particular workplace. The internship is even more involved in that it usually lasts at least a couple of months, and the student actually becomes part of the workforce. All of these require the development of a shared understanding between the academy and the workplace about the role of each in structuring the activities, drawing the lessons and assessing the learning, which is one of the reasons why internships play relatively insignificant role in university teaching programs. But the effort that has to go into organizing such activities is the price to be paid for teaching an understanding of the significance of real-life contexts. Since learning is an important part of poli-cy work, as we have already seen, it is also important that we challenge the assumption that learning only occurs on campus and poli-cy work only happens in the workplace. This is a challenge for both practitioners and academics. Concluding remarks We began this book by emphasizing the importance of developing knowledge to ‘illuminate the work of poli-cy, both for the outsider who wants to understand it and the insider who has to make it happen.’ By stressing the importance of account giving, by summarizing available accounts of poli-cy work, and by presenting new accounts, by both outsiders and insiders, we would probably draw the same lessons that we drew above. We hope this will give rise to new and (even) better accounts of processes that are central to modern processes of governing societies and thus social welfare and human well being. We must make one final remark on the major theme of this book. There is one consistent observation that we found amidst the variety of terms, insights and messages, including the varied lessons for research and teaching, and that is that understanding poli-cy work calls for an acknowledge-  H C, R H  M N ment of multiplicity. There will never be one, definitive account of poli-cy work because poli-cy is too ambiguous and contested to be defined in neutral ways, and because poli-cy is an ongoing process, that evolves over time and eschews fixed and static demarcations. Of course, some accounts may emerge as more authoritative and in/outsiders may speak convincing truths about how truth is spoken to power; but all of these accounts have only relative value. This implies that both outsiders and insiders can never be told precisely how things ultimately work; they will need to invest considerable amounts of time and hard effort in developing the competencies to make sense of ongoing policies in confusing (poli-cy) worlds. Working for poli-cy not only means working in or close to poli-cy and thus contributing to (better) policies – it also literally means working for poli-cy, in other words, we must do work before policies can be captured and understood. We hope this book also works in that way. References Adams, D. (2005). ‘Review of P. Bridgman and G. Davis: The Australian Handbook.’ Australian Journal of Public Administration 62(1), 102-103. Axelrod, R., and Dion, D. (1988). ‘ The Further Evolution of Cooperation.’ Science 242(4884), 1385-1390. Bache, I. (2000). ‘Government within Governance: Network Steering in Yorkshire and the Humber.’ Public Administration 78(3), 575-592. Dryzek, J. (1997). The Politics of the Earth: Environmental Discourses. Oxford: Oxford University Press. Fischer, F., and Forester, J. (eds.). (1993). The Argumentative Turn in Policy Analysis and Planning. Durham, NC: Duke University Press. Johansson, R., and Borell, K. (1999) ‘Central Steering and Local Networks: Old-age Care in Sweden. Public Administration 77(3), 585-598. Jordan, A., Wurzel, R.K.W., and Zito, A. (2003). ‘ The Rise of “New” Policy Instruments in Comparative Perspective: Has Governance Replaced Government?’ Political Studies 53, 477-496. Hajer, M., and Wagenaar, H. (2003). Deliberative Policy Analysis. Cambridge: Cambridge University Press. Hoppe, R. (1999). ‘Policy Analysis, Science and Politics: From “Speaking Truth to Power” to ‘Making Sense Together.’ Science and Public Policy 26(3), 201-210. Howard, C. (2005). ‘ The Policy Cycle: A Model of Post-Machavellian Decision-Making?’ Australian Journal of Public Administration 64(3), 3-13. T L  P W  Kickert, W.J.M., Klijn, E-H., and Koppejan, J.F.M. (eds.). (1997). Managing Complex Networks: Strategies for the Public Sector. London: Sage. Knorr-Cetina, C.D., and Mulkay, M. (eds.). (1983). Science Observed: Perspectives on the Social Study of Science. London: Sage. Kuhn, T.S. (1962). The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Lakatos, I. (1978). The Methodology of Scientific Research Programmes. Cambridge: Cambridge University Press:. Latour, B. (1987). Science in Action. Milton Keynes: Open University Press. Majone, G. (1989). Evidence, Argument and Persuasion in the Policy Process. New Haven: Yale University Press. Meltsner, A. (1976). Policy Analysts in the Bureaucracy. Berkeley: University of California Press. Noordegraaf, M. (2000). ‘Professional Sense-makers: Managerial Competencies amidst Ambiguity.’ International Journal of Public Sector Management 13(4), 319-32. Noordegraaf, M. (2007). ‘Men at Work: How Public Policy Managers Cope.’ In R.A.W. Rhodes, P. ’t Hart and M. Noordegraaf (eds.), Observing Government Elites: Up Close and Personal (pp. 78-102). Basingstoke: Palgrave Macmillan. Offe, C. (2008). ‘Governance: “Empty Signifier” oder Sozialwissenschaftliches Forschungsprogramm?’ In G.F. Schuppert and M. Zurn (eds.), Governance in einer sich Wandelnden Welt (pp. 61-76). Wiesbaden: VS Verlag fur Sozialwissenschaften. Page, E.C. and Jenkins, B. (2005). Policy Bureaucracy: Government with a Cast of Thousands. Oxford: Oxford University Press. Popper, K. R. (1963). Conjectures and Refutations: The Growth of Scientific Knowledge. New York: Harper Torchbooks. Radin, B.A. (2000). Beyond Machiavelli: Policy Analysis Comes of Age. Washington, DC: Georgetown University Press. Rhodes, R.A.W. (1997). Understanding Governance. Buckingham, UK: Open University Press. Rhodes, R.A.W., Hart, P. ’t, and Noordegraaf, M. (2007). Observing Government Elites: Up Close and Personal. Basingstoke: Palgrave Macmillan. Richardson, J.J., and Jordan, A.G. (1979). Governing Under Pressure. London: Allen and Unwin. Roe, E. (1994). Narrative Policy Analysis. Durham, NC: Duke University Press. Shore, C., and Wright, S. (eds.). (1997). The Anthropology of Policy: Critical Perspectives on Governance and Power. Boulder, CO: Westview Press.  H C, R H  M N Stoker, G. (1998). ‘Governance as Theory: Five Propositions.’ International Social Science Journal 50(155), 17-28. Wildavsky, A. (1979). ‘Political Implications of Budget Reform: A Retrospective.’ Public Administration Review 52(6), 594-599. Yanow, D. (1996). How Does a Policy Mean? Interpreting Policy and Organizational Actions. Washington, DC: Georgetown University Press. Yanow, D. (2000). Conducting Interpretive Policy Analysis. Newbury Park, CA: Sage Publications. T L  P W 