Governing spaces: a multi-sited ethnography of governing welfare reform at close range and at a distance

ABSTRACT Governing at a distance has been much discussed in relation to welfare reforms but debates tend toward abstract theorization, neglecting Foucault’s instruction to study mundane practices in specific sites. Although sociological concepts of place, positioning and boundaries carry particular resonance for public policy, ethnographic studies are scarce. A multi-sited ethnography of welfare reform reveals how seemingly discrete governance sites turned out to be linked in a complex policy assemblage. Findings suggest that governing at a distance may be ineffective and may necessitate governing at close range, although scales may fold over. Local spaces of network governance may not be autonomous but imbricated with national and local government and broader scales of governance. Apparently inclusive spaces exhibited exclusionary features, with spaces doing representational work that was simultaneously political, material and symbolic. Complex, shifting socio-spatial relationships thus influence the uneven development of welfare reform.


Introduction
This article derives from an ethnographic study of welfare reform and presents an analysis of governing both at a distance and at close range. The ethnographic study goes beyond Rose and Miller's (2010) much cited work on governing at a distance and draws on Allen's work on socio-spatial relationships (Allen 2011;Allen and Cochrane 2010;Allen 2003). The aim is to addresses gaps in the critical policy studies literature in understanding how governing at close range persists alongside governing at a distance and how, while certain technologies of rule may achieve powers of reach, stability and permanence, others are relatively ephemeral. I argue that not enough is understood about how relationships of reach and proximity are negotiated in the context of neoliberal welfare reform.
It is a truism to state that boundaries demarcate and denote inclusion and exclusion (Bowker and Leigh Star 1999) but I follow Allen (2003) in arguing that insufficient attention has been paid to forms of power inherent within socio-spatial relationships and the ways in which borders and exclusion zones are established, enforced or transgressed. This empirical study analyzes physical and symbolic aspects of architectural and administrative boundaries and policy implementation practices that took place within and across governmental and non-governmental sites. Three microcase studies in the form of ethnographic encounters are selected from an existing dataset and re-analyzed using primarily Allen's (2003Allen's ( , 2011) work on geography and power but also other critical writing on space (Massey 2005) as an analytical lens. This spatial lens enables an exploration of how governing at a distance is entwined or imbricated with governing at close range. The ethnography illustrates how actors manoeuvre within three very different governance spaces. The central argument is that governing at a distance may be subverted at a local level and is likely to occur alongside more proximal spatial relationships, i.e. governing at close range. The analysis of welfare reform policy implementation thereby adds to an accumulating evidence base of ethnographies of neoliberal governmentalities (Brady 2014). A broader objective is to present an empirically based argument for further research to take socio-spatial relationships more seriously.

Socio-spatial relationships of government and governance
Foucault's concept of governmentality has been used to argue that within neoliberal societies, governing increasingly takes place at a distance (Rose and Miller 2010). Associated with this premise is the suggestion that traditional forms of hierarchical state power, termed 'sovereign power' by Foucault, is supplemented (if not supplanted) by practices and governmental technologies that enable the 'conduct of conduct' (Rose and Miller 2010;Allen and Cochrane 2010), 'governance of the soul' (Rose 1990) and the 'responsibilization' of citizens (Clarke et al. 2007), resulting from a form of power that produces subjects even as they are constrained (Burchell, Gordon, and Miller 1991).
Foucault had little to say about town halls or politically devolved systems of governmental power but he did emphasize the local and contingent nature of power relationships. In systems of representative democracy, the electorate is differentiated and 'emplaced' (Dale and Burrell 2008) through designated boundaries, constituted from population statistics. Representative democracy is dependent on these relatively stable geographic and political boundaries, together with the political division between citizens and their elected representatives. States, of course, are not monolithic and socio-spatial relationships of the electorate and their representatives are complicated by neoliberal shifts from government to governance, with new scales and new forms of non-representative or participatory democracy on offer (Rose and Miller 2010;Durose, Justice, and Skelcher 2015). Fraser notes that in an era of globalization 'social ordering now occurs simultaneously at several levels' (Fraser 2003, 165). This echoes Allen's (2011) emphasis on 'topological twists' whereby spaces and scales of governance are not hierarchically ordered or laterally distinct but overlap and fold over, and this also fits conceptually with Newman and Clarke's insistence on imbrication and open systems of complex policy assemblages (Newman and Clarke 2009). As well as central and local government, a series of networks, partnerships and 'quangos' 1 can be found at a variety of imbricated spaces, from local communities or neighborhoods to supra-national bodies (Clarke and Cochrane 2013).
The turn to governance entails a move from analyses of traditional forms of power assumed to operate in hierarchical, sovereign, zero-sum fashion, to study networks where power is distributed with associational forms of horizontal accountability (Hirst 2013).Theorists of governance have tended to de-center the state, moving away from state-focused explanations of social and political reform to focus on the activities of these wider networks of actors (Bevir 2013;Rose and Miller 2010;Newman and Clarke 2009;Rhodes 2007). While some claim that government has given way to governance (Rhodes 2007;Kooiman 2003), others argue that government (including local government) remains a source of power, authority and legitimacy (Hill and Lynn 2005;Davies 2014). Analytically, I adopt Newman's premise that the role of government persists and relates to neoliberalism in ways that are not straightforwardly predictable, linear or rational (Newman 2014).
This article addresses the theoretical issues raised by Rose and Miller's (2010, 281) problematics of how contemporary governing beyond the state occurs and addresses several of the phenomena on their list of: . . . humble and mundane mechanisms by which authorities seek to instantiate government: techniques of notation, computation and calculation; . . . the invention of devices such as surveys and presentational forms such as tables; . . . building designs and architectural forms.
As Foucaultian scholars are aware, building designs and techniques of calculation and computation may instantiate rule and affect the 'conduct of conduct'. Scholars such as Allen (2003), Newman (2014), Williams, Goodwin, and Cloke (2014), Brady (2014) and Teghtsoonian (2015), however, warn that the stability and the effects of mundane tools, techniques and devices should not be presumed a priori, but rather should form the topic of empirical inquiry. Remaining 'open to surprise' (Brady 2014) encourages researchers to venture beyond their academic environs and beyond the analysis of texts to study at 'close range', i.e. ethnographically, how governmental techniques and discursive technologies are enacted in practice in particular sites.
Governing at a distance enacts particular forms of power relationships but Allen's work challenges the notion that power relationships take only one form. He shows how manipulation, seduction, authority and domination have different effects and the implications of this are that power may have variable transmission with discourses, buildings and controlling devices able to exert more or less extensive 'reach' to affect practice. Techniques of rule thus manifest differently across the boundaries of different sites (Allen 2003), and although certain devices or technologies may achieve relative permanence and reach across long distances to become 'immutably mobile' (Latour 1987), less well understood is where disciplinary devices might have more ephemeral effects and encounter stoppages, rather than constituting a ceaseless flow of capillary power. Allen (2003, 116) refers to this as the 'fixture and fluidity' of resources.
While this study was completed in 2009, this analysis has contemporary relevance because current policy developments concerning welfare reform, governance and devolution (Clarke and Cochrane 2013) mean that issues of how spaces enact governance and how spaces are governed are highly pertinent today. Geddes and Sullivan (2011, 393) note that: while neoliberal tendencies are observable everywhere, actual practices are uneven and contingently produced in place-specific ways . . . as they interact with 'inherited landscapes' of political, economic and social conditions. I will show how 'inherited landscapes', building designs and architectural forms are influential with respect to governance, how authorities used computation and calculation with unpredictable consequences and how mundane mechanisms operated with exclusionary consequences. I will also show how less mundane coercive practices operate alongside governmentality but before doing so, the following section provides the policy context for the three micro cases.

Policy context
The original study was initially designed to focus on the changing governance of welfare in the United Kingdom in the late 2000s with an emphasis on child poverty and early year's childcare. Long before current 'austerity' regimes, a trajectory of neoliberal policies has promoted 'self-reliance' and 'encouraged' welfare benefits recipients to move from 'dependence' to paid employment (Newman and Clarke 2009;Dean 2012). It is important to acknowledge that the neoliberal emphasis on economic productivity means that welfare reform is likely to have differential effects across the life course (Moffatt et al. 2012;Fraser 2003). Socio-spatial relationships are also relevant as the UK Government administers the 'cash' and 'care' elements of welfare at different spatial scales. Tax credits and welfare benefits are administered on a national basis according to a complex set of rules set by central government, while provision of child welfare, social services (including the protection of vulnerable adults) and education have traditionally been the responsibility of local authorities, that exercise relative autonomy over the provision of family support services (Glendinning and Kemp 2006;Carter 2011). Policy tools and levers (alongside discursive rhetoric) targeted at unemployed parents with pre-school aged children in the United Kingdom have included provision of subsidized childcare, creation of multi-agency Sure Start Children's Centres (modeled on aspects of the US Head Start program) and the use of tax credits to incentivize unemployed parents to move from welfare to work. The UK Treasury has been at the heart of welfare reform, governing from a distance by fiscalizing social policy, linking policy investment to hypothecated funding, strictly defined performance targets and a 'social investment' approach to policy (Lister 2003;Cantillon and Van Lancker 2013). This approach focuses on enabling productive citizens to undertake paid work in the labor market rather than protecting vulnerable people, ideologically positioned as a 'drain' on society.
Alongside the neoliberal welfare-to-work agenda, however, there have been progressive strategies to reduce the numbers of children living in poverty and enhance their well-being, with rates of child poverty monitored globally by international bodies such as the OECD (CPAG 2012) producing statistics that ought to be embarrassing for highincome countries. This policy agenda focuses on the needs of children qua children (Lister 2006) rather than as future worker citizens. A further discourse that may or may not support neoliberal workfare tendencies is the localism agenda that situates collective responsibilities but also collective action at the neighborhood or community level. Rather than locate agency with the 'soul' (Rose 1990) or the subjectivity of the 'responsibilised' individualized worker citizen (Clarke et al. 2007), 'community' is produced as a locale for governance and for effecting social change (Griggs and Sullivan 2014).

Policy ethnography methods and methodology
The state is not monolithic and governance occurs within and across many different spaces so multi-sited ethnography is appropriate for studying how such spaces form part of a complex policy assemblage (Newman and Clarke 2009). This approach also lends itself to studying how certain techniques of rule achieve governing at a distance and how power is also exercised within more proximal spaces. The original research that this article derives from explored how the government's multi-agency strategy to address the 'wicked issue' of child poverty was being addressed through childcare and family support services in local neighborhoods. The study design was exploratory and so rather than reifying policy, an inductive approach led to a focus on how policy was interpreted, represented and re-enacted (Peck and Theodore 2012).
Ethnographic-and anthropologically oriented methods (Rhodes 2013;Shore and Wright 1997) are recognized as particularly valuable for the study of process and policy implementation (Exworthy et al. 2010). Multi-sited ethnography (Nadai and Maeder 2005) offers the opportunity to trace theoretical concernsin this instance the sociospatial relationships of governancewithin and across the boundaries of sites. Foucaultian inspired analyzes have been critiqued by social policy scholars with a methodological commitment to ethnography (McKee 2009, Brady 2014) and a normative orientation to understanding oppression (Smith 2005;Massey 2005;Fraser 2014). Such critiques point out that, especially where studies begin and end with a discursive analysis of policy texts, the experience of the effects of governmentality and the perspective of local actors are neglected, despite Foucault's own insistence on the relevance of understanding mundane practices in actual sites (Foucault 1982). While discourse, in the form of policy texts at national and local level, was a key focus of this study, equally important areas of interest were the practices of policy implementers as they translated, interpreted and in some cases re-inscribed policy. Peck and Theodore (2012, 24) point out that this approach: is a far cry, however, from taking policy designs at face value and then analyzing their diffusion and emulation within a rational-actor framework. . . . [Rather,] it necessitates critical investigations of those multi sited social processes through which policy rationales, rationalities, and routines are constructed and reconstructed, made and unmade. Rather than a machinic process of replication, policy mobility is inescapably associated with policy mutation: if the form and effects of policies vary with context and shift while in transit, becoming embedded in both networks and within multiple 'local' milieux, the 'thing' that is being followed is evidently not itself an immutable object.
The connections and differences between a governmentality approach and Dorothy Smith's institutional ethnography have been usefully articulated by Teghtsoonian (2015) but rather than slavishly adhere to one or other of these approaches, this article attempts to work productively with aspects of both. Studying governmentality presumes an analysis of the exercise of rule, irrespective of whether this is analyzed as just or legitimate rule, or indeed coercion (Fraser 1981). In contrast, institutional ethnography (Smith 2005) drives the analytical and normative interest in exclusionary practices in situ and in how power relationships are reproduced or contested. The analysis for this study went beyond an empirical description of spaces to focus in particular on what are theorized as processes of exclusion and oppression taking place within and across interconnected spaces.
Data collection for the original study took place over the course of 9 months and meetings relating to the focus of inquiry at the timeearly years childcare and child povertywere accessed both purposively and opportunistically. For the original study, interviews were carried out with a range of actors including teachers, voluntary sector representatives, family support workers and childcare managers. Ethical approval was obtained and anonymity formed part of the agreement with the council (pseudonymized here as Whereham) that kindly permitted access. Findings reported here in the form of three micro case studies are drawn from previously unpublished data gathered at events that took place at three sites of governance. A limitation of this article is that these micro-ethnographic encounters may appear to be extremely fleeting. However, the case studies have been selected from the broader dataset using the criteria of salience (Buetow 2010)i.e. they are pertinent to the particular research questions that are addressed here. The cases are each instances where socio-spatial relationships of governance and government and associated processes of inclusion and exclusion were particularly apparent. While the three contrasting cases may appear to be disparate, analysis reveals relationships. The table below sets out key features of each case.

Architectural space Symbolic use of space in practices
Socio-spatial relationships Case 1 = Treasury Historic building with armed guards.
Informal meeting but a podium and dais separate the interlocutors and instantiate hierarchical relationships. GIS subverted by local knowledge, adhocracy, informal rules-in-use. Self-representation of CLP as allinclusive but exclusionary practices occur.
Relative autonomy of local CLP, constrained by national policy and local government strategy. Local knowledge in implementation key to understanding unintended consequences of policy reform.

Governing spaces
Three separate but inter-linked ethnographic encounters reveal how techniques of governing at a distance were imbricated with governing at close range. I witnessed a variety of 'humble and mundane technologies' (Rose and Miller 2010, 281) and other power modalities operating within these spaces and I observed spatial features of particular environments and the rules-in-use (Ostrom 1999) and rituals. Some boundaries proved to be permeable and malleable, others less so.
Governing at close range: the treasury invites partners in for a workshop A free event was arranged by the Treasury to discuss progress on ameliorating child poverty, open to delegates from across the country but hosted in the Treasury building itself located in the capital. The delegate list included local and national-level 'stakeholders', including delegates from children's charities and local government officers, thus bringing together in one space officials who were responsible for the 'cash' elements of tax credits with those working at a local level, responsible for more of the 'care' elements of childcare and family support. Inside the Treasury, approximately 80 delegates, including myself were offered name badges and invited to sit around tables, with around ten people per table. A podium was positioned on a dais running along one side of the modern, somewhat sterile meeting room. The opening address was a speech by the Chief Secretary to the Treasury in which he attempted to explain what was needed to tackle the 'wicked issue' of child poverty. At the lunch break, I attempted to go outside into a courtyard with my coffee but was prevented by an armed security guard and so, obediently, I milled around inside with the other delegates. I shall return to this somewhat awkward moment as an instance of governmental control in the discussion section. Lunch was followed by a workshop session in which we broke into smaller groups, going off into side rooms to discuss issues surrounding child poverty. I chose to go into the group discussing 'work'. The session was facilitated by an eager young civil servant, keen to share ideas using the familiar scribbling-on-the-flip chart technique (Carter 2012) so that we could all discover 'what was working' in relation to tackling child poverty. Local government workers and children's charity representatives were in attendance. Issues raised by participants included problems with the inefficient tax credit system, difficulties with sustaining children's centers because of uncertainty around revenue beyond the short term, and challenges faced by local authorities attempting to work with Job Centre Plus. 2 There appeared to be no lack of enthusiasm for the strategy to reduce child poverty, with an array of projects and initiatives described by local authority delegates. There was talk about developing unemployed parents' 'soft skills' such as time management and 'readiness for work', and some acknowledgment of the paradox of local authorities' seeking to secure 'value for money' in their outsourcing procurement processes, recognizing that this might in itself depress local wages.
When groups reconvened in the larger room for the plenary session one group's rapporteur appeared to raise a challenge to the neoliberal 'work first' discourse, pointing out that 'work doesn't always pay', that there could be disincentives for people moving from benefits into low paid work. He criticized the welfare to work strategy on the basis that 'some government targets seemed to be a higher priority than reducing child poverty'. This dissent was glossed over as another rapporteur asserted that he had anticipated that 'ideological barriers' might have prevented local authorities from fully engaging with 'the agenda'. He said he had found it 'very refreshing' that 'this was not the case'. The speaker's de-politicization denied dissent and reinforced a narrative of consensus. My analysis suggests that this discursive enrolment and manipulation was aimed at ensuring that local authorities were compliant with the neoliberal welfare to work agenda and would be making every effort to 'conduct the conduct' or affect the behavior of their local populace. The day was 'summed up' by a government minister. My field notes record that it was said that 'at the Treasury we sweat our assets'. This colloquial, managerialist phrase did not make it into the public report of his speech, but the public record did acknowledge problems associated with the tax credit system, and the written text exhorted local authorities to offer support: it is crucial that authorities continue to help out where possible, targeting families who they know are eligible, providing advice where possiblesometimes through other means, such as [Sure Start] Children's Centres. (Treasury 2006) In this way, provision of childcare via Sure Start Children's Centres was discursively incorporated into the central government policy frame of tackling child poverty and intertextually linked with the hegemonic discourse of 'work first' (Fairclough 2000;Rose and Miller 2010;Carter 2011;Cantillon and Van Lancker 2013). Local and regionally dispersed policy implementers were invited in and this 'relationship of proximity' (Allen 2003) was used to 'check out' (in an apparent attempt to root out) ideological opposition to the welfare reform agenda. The child poverty meeting did not make decisions and so did not implement policy in the narrow sense. Rather, the meeting provided an opportunity for Ministers and senior civil servants to govern discursively, to reach and meet face to face with people responsible for implementationto 'see the whites of their eyes'. This allowed them to work out how policy actors at the local level were making sense of and interpreting the child poverty strategy and in particular, whether they were able to make discursive and practical links between the child poverty strategy and the welfare to work agenda thus attempting to suture the neoliberal agenda. The meeting itself had few formal rules; free speech apparently took place in the 'workshops'. The oral discourse of the workshops was, however, edited by rapporteurs in the plenary, by speakers with a privileged place on the dais, then reedited for the formal written report. The armed guards remained in place at the courtyard doors; the neoliberal welfare-to-work agenda did not shift as a result of the discussion.
I turn now to the second case. At the time of the study (and increasingly since then), local authorities with social services responsibilities have found themselves under severe financial pressure. Whereham council's social services budget was in deficit, putting pressure on services, including traditional child protection services. Paradoxically, at the same time new hypothecated finance was available to Whereham for child care, family learning and family support; i.e. for the welfare to work agenda. Consequently, I found myself in a local government meeting where my principal empirical interest in child poverty was displaced by wider issuesbut in a way that proved analytically interesting as I explored the dynamics of multi-level governance.
Local government: inside and outside a local authority council chamber Whereham council had taken a decision to respond to the 'new' hypothecated funding and weave together separate central government strategies of school based initiatives, family learning and pre-school child care with its own espoused ambition to devolve power and respond to community needs. While there was no overt reference to welfareto-work or child poverty in the resultant strategy, there was an acknowledgment of the need to enhance social mobility and invest in early years support as well as family learning within local communities. An agenda for a council meeting was published in advance and the meeting was open to the public. Included on the agenda was the Community Learning Partnership (CLP) strategy containing a list of over 30 CLPs designed to enable Whereham to achieve central government targets through implementing local action plans strategically linked to a discourse of community empowerment.
Whereham planned to further devolve the hypothecated resources for childcare and family learning that flowed to it from central government, not to traditional borough or town councils but to specially constructed locality networks, designated by Whereham as CLPs. Using a geographic information system (GIS), Whereham council carved up its administrative boundaries into novel CLPs. The strategy set out how these CLPs would be expected to allocate the funds in response to local community need, as long as the objectives of childcare and family support were met, with Whereham remaining the legal accountable body. At this time, 'community' featured heavily in discourses of empowerment (Schofield 2002). Subsequently, localism has re-emerged in the context of debates about devolution (Clarke and Cochrane 2013).
At the council meeting I attended, the CLP strategy featured on the agenda just after another, contentious, item: the proposed closure of elderly people's council-run homes. There had been a vigorous debate over this closure during recent months in the local media. As I entered County Hall, I noted oil paintings of past civic dignitaries hung on the wall. It was unclear whether these were past council leaders or former Lord Mayors, and how many might still have been alive, but all were white men. Tiered, polished and upholstered wooden seating fanned outwards and upwards around three sides of the council chamber. At the back of the room was a gallery intended for press and the public. At the front, there was a platform with high backed, carved chairs placed for cabinet members to face the chamber. Microphones were permanently positioned by the seats. Refreshments were not permitted inside the chamber and this mundane ritual associated with formal proceedings created distance and a tone of formality.
At the meeting council, officials shepherded me, together with approximately 50 or 60 agitators who were protesting against the elderly home closures, into the upstairs public gallery. There was not enough room in the gallery, so some people were seated in the main council chamber directly behind councillors and their advisers. The council leader addressed the gathering in an authoritarian manner: 'be assured that I do not suffer fools gladly', pointing out that this was a Cabinet meeting, not an opportunity for members of the public to have their say. This 'rule' or directive was disobeyed, and the leader's attention was distracted by unauthorized interventions including derisory snorts, requests for clearer speaking and angry shouts. The lead officer responsible for management and administration of the proposed change for older people's services presented her case, framing the need for change as being at least in part due to 'demographic pressures'. The decision was taken by the Cabinet to implement the closures. In interview, a local authority officer explained: 'I mean it's like anything else, if there's an issue you'll get the turnout' (LA 05). Contradicting the notion of citizen apathy, people turned out in force to protest against this element of welfare reform. They were treated by the council leader not as equal power sharing 'participants' or as responsible, engaged citizens with rights, but as troublemakers.
When the agenda item on older people had been dealt with, the Cabinet took a 10minute break and reconvened in a separate, smaller room. The agitators left as their issue had been formally addressed but they re-grouped with their banners in the street outside so that the protest could still be heard and attentions were distracted. Inside the building, I sat alongside a local press reporter taking notes. Child poverty was not referred to, but the report recommending the establishment of 32 CLPs was tabled and 'nodded through' by council members with approving comments. There were no dissensions and no further debatethe agenda item took up about 3 minutes in total of council time before moving on to the next agenda item, at which point I left.
Governing a local community network CLP implementation got underway and I was granted permission to attend Goldborough CLP meetings as a non-participant observer. Meetings took place initially in a community center and later in a meeting room at a local high school. I had found a Whereham council officer who had responsibility for GIS and interviewed her, asking questions about her CLP mapping work. She explained how she utilized the mapping software and gave me copies of the maps pertaining to CLPs. This drew on combined administrative data derived from school locations, statistical indices of multiple deprivation and on whether there were existing childcare facilities in the area thus demonstrating many features of a governmental technology designed for governing at a distance. The maps were notated with dots, triangles, stars, population figures, etc. (see Figure 1). As the legend shows, a bureaucratic, mainly educational system overlays this representation of a 'community'; which is configured using the 'joined up government' technology of the Index of Multiple Deprivation ranking the super output areas (SOAs). 3 Some detail has been removed to protect anonymity.
A different representation of the CLP manifested in the form of a slide presentation. The first slide used in CLP consultation events was headed 'What is a Community Learning Partnership?' An anonymized version is reproduced below: Figure 2 slide from Community Learning Partnerships District Consultation Event (anonymized).
The slide frames and represents CLPs as fully inclusive communities. As long as 'representatives' are interested in 'ensuring positive outcomes for children, families and the wider community', then they are welcomed into the CLP. Unlike the diagrammatic GIS representation (yet like the life-sized map of the world in the story by Borges) (Borges 1999), there seems to be no distinction between representation and the represented, between who is a member and who is excluded. Whereham council's strap-line appears on the Goldborough CLP slide but during fieldwork, I became aware that there was considerable latitude within the CLP strategy for local implementers to exercise their discretion and that there was some misunderstanding about the form and function of the CLP. When I interviewed a manager working in the voluntary sector, I asked: 'So in terms of these Community Learning Partnerships do you feel sceptical about what they've got to offer?' She answered: 'Yeah, I think it's a load of old quangos to be honest' (LA5). I interviewed a primary school head teacher who attended CLP meetings and asked:  I: So if I can move on to the Community and Learning Partnershipwho do you think it is accountable to? Headteacher: (10 sec pause) I really do not know. I think I would say as a guess I would think it was accountable in financial terms to the financial regulations of the local authority who then would be accountable to -I presume the [Department for Education] because that's where the funding comes from. I think it should be accountable to the people in the local area. . . .. I: Sodo you think it works democratically? Headteacher: I don't know -I mean I don't know what democratic means in this situation. I don't know who the people if you like are entitled to vote. I don't know how many people who, because they haven't had opportunity to go to meetings -I don't know how many people are actually out there who should be in the meeting or involved and so in a sense they're outside the democracy of it because they've not been able to come for whatever reason or they don't realize the value of it. So I think that's a very difficult one to say. (LA 53) It was difficult for this CLP network 'member' to articulate the boundaries that might differentiate the inside from the outside of the network group and to understand control mechanismsnote her 10 second pause before she was able to reply to the question. What her discourse did reveal was an awareness of funding emanating from central Government via Whereham Council but rules and accountability mechanisms were not made clear and transparent, resulting in a sense of frustration. Rather than a regime of governmentality, governance mechanisms appeared to take the form of an 'adhocracy' (Courpasson 2000).
At a CLP strategy meeting, a manager (pseudonym Louise) urged the gathered network of people from multiple agencies to share in some responsibilities for developing the CLP. There was no representative from Job Centre Plus in attendance to put forward the welfare to work discourse as they were too busy. Louise began a slide presentation that explained the Community Learning Partnership concept. The slides mirrored the strategy document that was adopted at the Whereham Cabinet meeting, except that Louise added her own slant with graphic images. One of these was a picture of interlinked metal chains. Louise explained that she had selected this image to represent interlinking bonds of strong partnership. She pointed out that we were not to see these as 'tortuous chains that bind us'. She was directing the audience to see the image of the CLP as positive and enabling, not constraining. She explained in her presentation that the government aims to 'knock child poverty on the head' by aiming to get 'one worker in every household'. People in the room seemed attentive as they sat around rectangular tables that had been placed as near as possible in a circle shape. Unlike at the Treasury plenary session (or the council meeting), there was no platform, dais or separately delineated space for the chair. Members of the CLP were identified, selected and appointed by one another on the basis that they had time to attend meetings, a stake in the CLP by virtue of their locally based organization, and a local reputation and history of working in partnership (consonant with the depiction of a strong interlinked chain) from previous multi-agency initiatives.
According to the CLP strategy document (version 12) one of the head teachers, Mary, was not in the proper place. As represented on the CLP map and the list produced by Whereham, her school should have belonged in a different CLP. Mary made a case that, although her school was 'over the hill', she felt she should be at this meeting and part of Goldborough CLP, explaining that most children moved on from her primary school to travel to the high school 'down the bank' so she felt her school more naturally related to this local schools cluster. It became apparent during the course of the meeting that schools had working relationships, families had local habits and traditions and teachers had friendships and community ties that cut across the administrative boundaries defined by Whereham council; as Scott (1998) put it, 'the map is not the terrain'. For the most part of its business, Goldborough CLP ignored the GIS data, making decisions by apparent consensus, guided by Louise on how to allocate its resources and enhance a sense of community. Rather than the technical tool of the GIS map designed to specify and produce a bureaucratic form of community, established working relationships and knowledge of the local area, including local history, influenced decisions so that resources were allocated to a range of community and voluntary organizations, including out-of-school clubs. By the time of the final published version of the CLP strategy (version number 18), a high school had metaphorically shifted location and become included in Goldborough CLP. The technological GIS designed to govern at a distance and to definitively place communities according to a specification by Whereham Council was rarely referred to in meetings, but at a meeting held 6 months after the CLP commenced, Louise addressed 'those of you who know this area'. Despite the ambiguity about the school over the hill, Louise claimed to speak for all: 'we know which area we're talking about'. This discursive enrolment enacted a form of spatial closure that diverged from the GIS technology.
This was the last meeting I attended as part of formal fieldwork but I remained on the e-mail distribution list and added this 'postscript' message to my dataset. It was sent from a Whereham councillor who represented Goldborough Borough (who had not been to any of the meetings) in response to CLP minutes that had been circulated to the list: Despite a directive from Whereham council mandating CLPs to include local councillors, Goldborough CLP had gone about its network business without this particular elected member. Unlike the council's Cabinet meetings, the CLP met in borrowed community spaces, did not publicize its agendas and meetings were not open to the public but remained exclusive to the self-appointed members. Power appeared to be diffuse and accountability seemed on the surface to be club like or associational (Hirst 2013) and ad hoc. However, Whereham retained a role as 'lead accountable body' and reported to central government how the hypothecated finance had been allocated. Autopoeisis is a term used by Kooiman (2003) to categorize autonomous, 'self-steering' networks, and during fieldwork often it appeared as though the CLP partnership was only accountable horizontally to itselfas the teacher put it: 'a load of old quangos'. However, the multi-sited ethnography allowed insight into the socio-spatial relationships that linked the CLP to Whereham, to the Treasury and broader scale processes of neoliberalism as well as to global concern about child poverty and highly localized geographical knowledge.

Discussion
This article has examined a variety of socio-spatial relationships that influence welfare reform. Findings support Newman's (2014) and Rose and Miller (2010) argument that government and governance are not distinct, that (as Foucault acknowledged) sovereign power has not been completely replaced by governmentality. The theoretically informed ethnography revealed that governing at a distance may not achieve intended effects, so that 'conduct of conduct' is contingent and variable rather than inevitable. Allen's work is a key to understanding the variety of power modalities in play within socio-spatial relationships that affect processes of neoliberal welfare reform (Allen 2003). The distinction, as well as connections, between 'cash and care' elements of welfare provision are recognized within the traditional academic discipline of social policy (Glendinning and Kemp 2006) so that the distant reach of welfare benefits set at national level in the United Kingdom, contrasts with the local delivery of childcare, family support and other welfare services via local authorities and partner organizations. In what follows, I discuss governing at a distance, but as this has been well theorized (including by Rose and Miller 2010), I pay greater attention to the process of governing at close range, before concluding that socio-spatial relationships of governance and governance are inextricably entwined in topological twists and that further research is warranted to understand the effects of persistent sovereign power alongside governmentality.

Governing at a distance
The Foucauldian notion of governmentality is drawn on by Sterling (2005, 145) to show how '. . . territorial identity is created deliberately through governance . . . "territory is rendered visible" so as to be acted upon. ' Rose and Miller write (1992, 12) Programmes presuppose that the real is programmable, that it is a domain subject to certain determinants, rules, norms and processes that can be acted upon and improved by authorities. They make the objects of government thinkable in such a way that their ills appear susceptible to diagnosis, prescription and cure by calculating and normalizing intervention. . . . Government inspires and depends upon a huge labour of inscription which renders reality into a calculable form. Written reports, drawings, pictures, numbers, charts, graphs and statistics are some of the ways in which this is achieved. . . . By means of inscription, reality is made stable, mobile, comparable, combinable. The Treasury's fiscal program attempted to govern local authorities and citizens from a distance and the 'worker in every household' mantra was reproduced in discourse and semiotics circulating alongside a community empowerment and strong partnership discourse within the local space of the CLP. GIS was deployed to govern at a distance but the flat cartography was contradicted at the Goldborough CLP meeting by historical and topographical understandings of children's downhill walk to school so that one high school changed its administrative location while remaining physically in the same place. Thus, the technocratic GIS system belied the human subjective work of data interpretation (Bowker and Leigh Star 1999). Small area statistics delineated the catchment area but could not pinpoint or map individual benefit claimants. Despite the 'spurious precision' (McLaughlin 2006) implied by the software, the statistical data reproduced in the form of blobs and triangles did not translate simply and directly into decision-making. This particular GIS tool designed for governing at a distance was not stable nor an example of what Latour describes as an 'immutable device' (Latour, cited in Allen 2011); local knowledge of the terrain and the adhocracy of rules-in-use (Carter 2012, Ostrom 1999) proved more powerful. Goldborough CLP had no dedicated space of its own but used meeting rooms that were proffered as part of the networked partnership. This sharing demonstrates the resource dependency (Malatesta and Smith 2014) and degree of collaboration amongst partners as represented in the image of strong chains, but the borrowed spaces also portrayed the relative informality and temporary nature of the network. The CLP was informally constituted via the routines that became established rather than having any more formal governing instruments. Welfare reform, child poverty, local history and community empowerment discourses competed and interacted in a complex and contingent set of sociospatial relationships.

Governing at close range
Whereham's council chamber was designed for the express purpose of local democracy with a clearly demarcated space for those with the elected mandate to speak on behalf of others. In this instance, the authoritarian cabinet leader attempted to define who belonged in which zones of the space. Despite the close proximity of the electors and elected in the overcrowded chamber, there was a gulf in understanding, what Scott (1998) terms a 'dialogue of the deaf', with little prospect of reconciliation through deliberation. There were no attempts at 'culture governance' (Dean 2003), dialogic deliberation or governmentality in this formally democratic but symbolically authoritarian, patriarchal space. Citizens' legal entitlement to occupy the space did not guarantee them a welcome or a hearing.
In the guarded historic space of the Treasury, the child poverty meeting was also an instance of governing at close range. Allen (2003, 135) analyzes how the 'far off' may be 'dissolved into the close by'. Governing welfare reform from a distance using fiscal measures of tax credits and subsidized childcare faced inevitable limitations, with the possibility (acknowledged by senior Ministers) that local authorities may not have been as enthusiastic as others in Westminster and Whitehall regarding the welfare reform agenda. My analysis indicates that the meeting was an opportunity for Whitehall and Westminster policy makers to close the 'implementation gap' that is inevitable in techniques of governing at a distance (Rose and Miller 2010;Barrett 2004). Allen (2003, 137) notes the importance of emotion and the significance of facial expressions and vocal intonations for communication. The Treasury meeting provided an opportunity for face-to-face communication so that policy makers could understand localized policy effects (such as the problems experienced with the tax credit system) and assess 'ideological barriers', as the civil servant put it, within local authorities that might have militated against the Government's welfare to work agenda. The neoliberal complex assemblage of child care initiatives and tax credits did not, for many of those in the workshops, readily translate into solving the problem of child poverty at local level. 'Transcripts from below' (Scott 1998) were surfaced by the ethnography, revealing that some participants in the workshop spaces challenged the hegemonic neoliberal 'work first' discourse by pointing out that work did not always pay enough for parents to forego their benefit entitlements (Dean 2012). The meeting might have seduced some local actors by the apparent openness of accessing the inside of the historic, prestigious building and being physically close to MPs and Central Government officials, but the agenda was set and the record was produced backstage. The close encounter and the attempt at seduction (Allen 2003) was not reflected in the authoritative report on the event that left out the reference to overcoming 'ideological barriers' made by the civil servant and thus, can be seen as a manipulation (Allen 2003), or at least a selective representation, of the proceedings.
Alongside this form of discursive governance, the presence of the armed guards was at once symbolic but also a stark reminder of the State's capacity to exert violence (Clausewitz cited in Frazer and Hutchings 2011), indicating that governance has not entirely displaced government and that subtle techniques of governmentality that rely on freedom and consent (Rose and Miller 2010) may be supplemented by more direct domination (Davies 2014). This fleeting encounter clearly demonstrated that while power may have reach and flow across the socio-spatial relationships of networks (Allen and Cochrane 2010), it may also encounter key stopping points so that some boundaries, even if they are not physically fortified, persist in protecting highly exclusive and exclusionary zones.

Topological twists
The case of the CLP showed that any assumption that the delineated spaces and governance systems of representative and communitarian or associational democracy (Hirst 2013;Griggs and Sullivan 2014) are binary and separate is problematic. In their analysis of the UK Coalition government localism agenda, Clarke and Cochrane (2013, 13) write: Localities are produced through distanciated relations. They are nodes in networks, meeting places for mobilities, assemblages of parts from elsewhere. They are dynamic, becoming, and contingent. They are plural, heterogeneous, contested.
Academic, policy and practice debates have all focused on how to engage communities in 'civic renewal', in politics and in public services and participatory democracy has been discussed in relation to network governance (Fraser 2003;Clarke et al. 2007;Martin 2011;Bovaird et al. 2015). The case of the CLP operated as, in Clarke and Cochrane's (2013) terms, a nodal space, where the 'parts from elsewhere' comprised a potentially progressive global child poverty agenda and local community partnership governance arrangements. Use of the terms 'community' and 'partnership' and the use of interlinked chains were semiotic and discursive devices that portrayed a strong, cohesive network and attempted to enroll members in a sense of belonging but the ethnography revealed exclusionary practices and doubts about legitimacy were expressed in the accounts of some members. Boundaries of the techno-terrain were shifted by a sense of place derived from local knowledge with admittance to the CLP determined, not by formal citizenship rights, as happened at Whereham, but via the tightly bounded network of people who had established working relationships. Kooiman (2003) notes that governance involves a process of 'image formation' and we saw how the image presented on the CLP slide presentation was of an all-inclusive community with no acknowledgment of difference, distinction or of any conflict of interest between, for example, the 'private sector services' and other partners in the CLP. Although the CLP attempted to represent itself in its own local image, the inclusion of the Whereham logo in the slide presentation indicated that, although CLP boundaries were bent out of shape and a Councillor was excluded through the adhocracy of the rules-in-use, Whereham council retained a legal status as the lead accountable body thus formally linking the CLP to local and national government.
The spaces I accessed at close range did representational work that had real and symbolic effects (Massey 2005;Allen 2003). Their design and use presumed relationships between governors and they governed through designated zones of membership, inclusion and exclusion. The spatial codes conveyed by spaces suggested '. . .who should be seen and heard and who should not' (Allen 2003, 162) with the portraits of patriarchs suggesting that it was white men's voices that had historic influence in the particular environment of Whereham and beyond. As Smith argues, As an institutional ethnography works from the specific local settings from which it takes its direction, it discloses beyond them not just specific institutional forms but also the ruling relations in which institutions are embedded. (Smith 2005, 200) Multi-sited ethnography revealed attempts at discursive closure and exclusionary boundaries operating in all three cases. Although interstitial spaces of subversion were found in each site, some boundaries remained impermeable.

Conclusion
In the introduction, I set out the intention of this article to study governing at a distance and to respond to Rose and Miller's problematics of governing. What the analysis reveals is that governing at a distance and governing at close range may occur simultaneously. Besides governing at a distance, governing at close range also took place using techniques of reconfiguring territorial boundaries, discursive manipulation, enrolment and exclusion, as well as domination with the threat of force. While mundane practices of governing at a distance such as GIS and tax credits do link different institutional sites, disciplinary effects are less certain and more leaky than panoptic surveillance (Foucault 1975). Close range proximal relationships should not be overlooked as these may be vital both to interpersonal and inter-organizational discursive communication and to more or less mundane threats of violence. As Allen's work on topological twists predicts (Allen 2011), proximal and distant spaces overlap in complex and contingent assemblages.
For those engaged with current localism agendas where novel forms of governance are developing, the power of invitation processes, the symbolic meaning conveyed by the rituals, routines and architecture of spaces, what occurs inside and outside the boundaries of traditional government spaces and less formal network sites could all be matters of concern. Perhaps especially in apparently open arenas, who is invited to become a member and is thereby entitled to occupy the space, who feels welcome and who is excluded, as well as whether invitations (or rights) to participate in governance or government result in genuine dialogue, all deserve greater attention. Notes 1. Quango quasi-autonomous non-governmental organization. 2. The UK arms-length agency outside local authority jurisdiction, created from what were separate state institutions of the Benefits Agency and the Employment Service. 3. Super output areas are statistical boundaries defined at a lower level than an electoral ward.