Policy to reality: evaluating the evidence trajectory for English eco-towns

Can evidence reconcile idealized policy formation processes with the messy reality of large-scale sustainable projects? Was the main driver of the UK eco-towns project the national housing need or a local version of sustainability? To answer, this paper traces the progress of the eco-town policy: from a government-funded initiative to start a cohort of new towns, via the first pilot sites, to the delivery of a single eco-town. It maps the national policy intent, which incorporated industry and academic expertise into a strategic policy vision. Then the local interpretation of this generic eco-town definition is reviewed, along with how the authoritative knowledge on innovative eco-planning is distilled into guidance. By revisiting the original eco-town objectives, comparing them with the recently published garden cities prospectus, the limitations for intra-national implementation of national policy frameworks are considered, asking whether any lessons have been incorporated in the current round of policy formation. The success or failure of such policy implementation is found to be more of a socio-political exercise than an empirical or rationalist process. The eco-towns were less an attempt to establish new forms of sustainable habitations than a political attempt to use eco-planning to justify new large-scale housing settlements.


Introduction
The nature of urban planning policy, particularly sustainability policy, is an iterative, discursive process full of conflict, contradiction and ambiguity (Freeman, 2012;Healey, 2010). The experience of the UK eco-towns programme between 2007 and 2010 fits this messy, conversational, combative yet collaborative depiction. The programme required a consensual definition of what an eco-town might be, not least to provide assessment criteria to evaluate and select pilot schemes for funding. Nevertheless, even this seemingly simple aim required months of argument and consultation to become codified into planning policy (DCLG, 2009). Considering just one criteria discussed in this paper -should eco-towns be stand-alone or extensions of existing settlements?demonstrates such varied flexible interpretation by each of the four selected eco-town sites that any attempt at policy fixity seems spurious. Yet Baker, Kousis, Richardson, andStephen (1997, cited in Rydin, 1999, p. 468) argue that the lack of clarity defining sustainable development has unexpected benefits precisely because it requires those with different, even conflicting, interests to 'reach some common ground upon which concrete policies can be developed' (Baker et al., 1997, p. 5). The search for consensually agreed criteria starts with what can be jointly considered as robust evidence. This paper traces how evidence was used in generating the policy framework for the eco-towns, but was excluded from any of the outcomes of the policy itself, or subsequent policy. Since the millennium there have been four major policy attempts to scale up housing growth: Growth Areas (2003), Growth Points (2006), the eco-towns initiative , and a fourth round of policy activity unfolding in the autumn of 2014 to stimulate private sector-led construction of large 'sustainable' housing settlements. The first policy attempt was an extension of the regional strategy process, the second an invitation for local authorities to bid for funding towards the infrastructure needed for new housing (Cleary, 2014). This paper concentrates on the third initiative -eco-towns -but also asks which lessons are reappearing in the most recent re-envisioning of new towns as garden cities.
The discussion focuses on the interwoven networks that gathered information, and knowledge on sustainable lifestyles and construction used to shape Baker et al.'s (1997) common ground, providing en route: a critique of the national policy intent and framework; a description of the policy evolution from a high-level strategic vision into deliverable targets through the drafting of guidance documents; and how delivery of these plans has been affected by the local particularity of each eco-town demonstrating the horizontal (centralist) and vertical (national to local) processes of policy coordination (Tomozeiu & Joss, 2014). Drawing on Rydin and Moore's (2008) research on the creation and use of sustainable construction information within planning departments, and on Cooper's (2006) reflection on the extent that this expert evidence gathering is a socially constructed interaction of insiders (local government, project teams) and outside 'experts', this paper looks for the conditions required for that knowledge to be recognized as (1) legitimate and useable and (2) how the political context affects knowledge capture and deployment from these multiple sources (Cooper, 2006). The eco-towns episode can thus be considered as a tentative example of evidence-based policy-making with government engaging with experts to identify 'what works' in both policy and practice (Blunkett, 2000;Solesbury, 2001), and as an example of an emerging sustainable evidence base promoted via the different policy and practice knowledge networks.
The paper explores the initial phases of the eco-town story, the national policy formulation, the selection of the pilots and the beginning of the local-level discussions. It also concentrates on the master plan and locational planning issues rather than the detailed building construction, CO 2 emissions or technical sustainability standards. In doing so it risks falling into the trap identified by Rydin and Moore (2008) that thinking on sustainable development is crudely bifurcated into spheres of technical construction or sustainable planning. Similarly, there is not space to follow Healey's (2010) advice to scrutinize the micro-practices of policy and planning processes at their most mundane daily level of implementation. Nonetheless, Healey's recommendation is a useful reminder that bottom-up analysis is essential to counter the traditional top-down evaluation of policy transfer. 1 The literature on policy mobilities has to date concentrated on this top-down international transfer of ideas as inter-state exchanges, rather than the more complex intra-state flows between national-level policy, regional-level delivery agencies, and local planning authorities working with local delivery partners and local communities (McCann, 2011;Peck & Theodore, 2010;Stone, 2004). The fruitful learning derived from international examples of sustainable development, and the process of 'learning from Europe', demonstrably shaped the eco-towns initiative (the repeated citing of favoured best-practice examples such as Amersfoort, Freiburg or Hammarby Sjö stad is a familiar policy transfer trope) (PRP, URBED, & Design for Homes, 2008;HCA, 2009). 2 However, the preliminary query here, whether the main driver of the eco-towns project was driven by national UK housing need or a local version of sustainable development, requires a 'middle-out' consideration of how a national-level policy vision impacts on local-level decision-making, alongside consideration of the practical mechanisms that have been used to drive the vision forward in a variety of local contexts. Freeman (2012) notes that an essential, if routine and mundane, component of the complex process of policy formulation is meeting to talk through ideas. These transient exchanges are codified in various forms of writing; notes of meetings, papers or reports which enable the imperfect contingent process of translating and articulating nascent policy ideas into actionable guidance that can be shared:

Methods
The meeting [. . .] is where policy concepts are built, tested and rebuilt in communicative interaction. Meetings invariably generate documents; in being written down, committed to paper, a policy concept is both fixed and made mobile. (Freeman, 2012, p. 14) This 'fixing' of the matter under discussion might only be partially successful, making it available to others for use (or to study) while retaining a degree of ambiguity and fluidity of interpretation. This research similarly scrutinized documents, traced ideas, reconstituted meetings and rehearsed explanations through conversation. Methodologically, this paper applied aspects of discourse analysis (Fairclough, 1992) to frame the context and significance of statements found in selected government reports, the grey literature and press articles covering the past decade, combined with a dozen semi-structured interviews with a diverse range of participants in the eco-town process: civil servants, policy-makers, architects, urban designers, staff of participating organizations, authors of best practice, and delivery partners for two of the eco-town projects. 3 The written accounts of historic meetings and encounters varied (from recollections, to meeting minutes, to verbatim Hansard reports 4 ), and the research process was discursive and narrative driven, looking for repeated citations of reports, documents and other circulated evidence. Yet one of the most informative interview questions was 'Who was at the meeting?' exploring the networks of individuals and how their relationships shifted and rearranged. 5 A timeline of key events, activities and publications was drawn up for the initial policy and selection . This identified five spheres of activity: firstly, key government announcements acting as political catalysts; secondly, identifying the actors/groups involved; thirdly, the discussion and debates that led to policy statements; fourthly, guidance publications; and finally, the bidding process, assessment and selection of potential sites and early development activity. There was an (unrealized) expectation that each historic phase would inform and feed into subsequent ones, with the distillation of the research, discussion, expert scrutiny and other knowledge-generation activities summarized into a logical suite of publications, which would in turn be used and recognizable in the built schemes. The timeline showed a more sporadic drawn-out and disconnected narrative of evidence selectively emerging and evaporating. It also reinforced the initial policy formation as the earliest stage of an extremely long process: the rapid timetable for selecting the pilots and a similar short master-planning period providing the foundation for a 20 -30-year delivery phase. The current study is clearly unable to assess the success or failure of the policy implementation played out during and after construction of the new towns, merely speculating on the early translation and enactment of the policy intent. Yet this early eco-towns experience can be seen as a microcosm of the 'policy to reality' trajectory: a relatively 'fast track' version of the learning and thinking that are needed for remaking new and existing cities.
Initial national policy vision for eco-towns derived from existing evidence The third of these objectives was the most contentious, even at a point where the planning and development community had become suspicious of the Planning Policy Guidance 3: Housing (PPG3) (DETR, 2000) 'sequential test' 6 for prioritizing housing development on brownfield sites, arguing for the use of greenfield sites (Adams, 2011). Thus, the initial definition of an eco-town was distanced from existing unpopular plans for town extensions.
To be considered for eco-town status, proposals needed to be new settlements with a separate and distinct identity but with good links to surrounding towns and cities in terms of jobs, transport and services. A minimum size was 5000 -10 000 homes arranged in mixed communities, with affordable housing comprising between 30% and 50% of the total. The target was zero carbon for homes and reliance on low carbon renewable energy and exemplary innovation in at least one area of environment technology. A wide range of facilities was required, including a secondary school, shops, business and leisure areas, with a suitable delivery/governance organization to manage the town, its residents, businesses and community services. The four sites selected for eco-town status would benefit from additional DCLG Growth Point funding for infrastructure and community projects, while the housing development would be largely funded by the private sector, supported by the newly formed Homes and Communities Agency (HCA).
The eco-town criteria and intended policy outcomes in the Prospectus (low carbon living, low impact transport use, community participation and empowerment, high-quality design following active design principles to encourage health and well-being, and land use to mitigate the impact on biodiversity) were unexceptional, conventional sustainable development responses. The location/land-use objectives were constrained by the double challenge of locating sizable sites connecting to the existing infrastructure network of towns, whilst making use of brownfield sites where there were 'good opportunities to do so' (DCLG, 2007a, p. 18 ) a year later. Exemplary urban developments in Sweden, Germany and the Netherlands were also key influences and were used as best-practice case studies in all these publications. So in urban design terms, eco-towns were clearly part of an ongoing, established large-scale planning tradition despite being far less ambitious in scale or number than the post-war New Towns. The establishment of legitimate sustainability criteria to justify additional housing construction was salient, not just with regard to planning authorities, but also to prevailing political aspirations. Figure 1 shows the initial analysis tracing the nationallevel political statements and discussions, leading to the eco-towns definitions/criteria which, following a formal consultation process, were codified in the Planning Policy Statement: Eco-towns (PPS:ET) (DCLG, 2009). The definitional task was drawn out as the initial criteria were challenged by schemes proposed by bidders. The range of guidance documents (produced for planners and developers/eco-town promoters) implied separate and distinct criteria, but one must question the extent that the policy was evidence based. As 'socially constructed' knowledge, the guidance conflated exploration and research onto 'derived best practice' (Moore, 2005). This knowledge was interpreted in different ways at the local level by the tier of players delivering the plans. It is at this vertical transfer that the greatest disconnect between the original policy intent and its interpretation was apparent. For example, the TCPA best-practice guidance emphasized the importance of decisions for new towns or urban extension being made at a regional or sub-regional level rather than fixed specification at a national level (TCPA, 2007) without acknowledging the difficulties of setting context-relevant local standards or how the policy unravelled in practice under the dynamics and pressures of local politics.

Shifting targets, diluted visions
Politically, eco-towns were a top-down initiative, with the 2007 announcement to expand the funding to 10 locations made by the new Labour Prime Minister Gordon Brown rather than the DCLG. Brown's background as Chancellor of the Exchequer ensured rapid buy-in at the Treasury, but unusually it was a cross-Whitehall project, involving not only DCLG, Department of Energy and Climate Change (DECC) and Department for Environment, Food and Rural Affairs (DEFRA), but also the Environment Agency, the Department for Transport (DfT), the Highways Agency and English Heritage in an inter-departmental strategic review of the proposals (DCLG, 2007a). Tomozeiu and Joss (2014) identify the bias towards spatial planning and housing issues over environmental aspects arising from the subsidiary consultee roles of DECC, DEFRA and the Environment Agency to Figure 1 Long-term timeframe and analysis frame Policy to reality for English eco-towns DCLG. Whilst the cross-department interest implied united governmental support, in practice it initiated a slow bureaucratic process of ministerial consultation, sign off and clearance, with a secretariat of civil servants within a further department, Office of Government Commerce (OGC), to manage the Gateway project management process. The doubling in size of the initiative between March and September 2007 was due in part to concern over the diminishing number of housing completions (Adams, 2011;Goodchild & Walshaw, 2011). In November 2007, DCLG received 57 bids, which were winnowed down by the Challenge Panel over the following 18 months, with the four selected pilot sites (St Austell, Cornwall; Rackheath, Norfolk; North West Bicester, Oxfordshire; and Whitehill Bordon, Hampshire) announced in July 2009.

Transfer routes for eco-town knowledge
The timeline highlighted the large number of diverse industry and professional groups involved, emphasizing the extent that DCLG was reliant on external agencies (see Appendix A in the Supplemental Data online for a partial list). These groups formed the major mechanisms for policy transfer, with the documents forming summaries, or occasionally provocative responses, to the discussions, workshops and meetings that took place. Expert individuals played key roles, for example a highly visible TCPA official named in the DCLG prospectus (DCLG, 2007a) providing status and credibility to the material. Circulation of knowledge in and between these organizations was fluid and informal depending on the policy tasks they were addressing. Expert networks of professionals, academics, researchers and policy actors collaborated on the explicit task (stated in the Prospectus) of establishing robust criteria for sustainable development, devising an accessible, comprehensible and collectively justified description of what constituted an eco-town. These criteria included the appropriate location, scale, urban form and provision of amenities/technologies to promote sustainability. Two generations of expert groups contributed: existing groups such as the TCPA (who located eco-towns within their historical continuum promoting sustainable settlements) and new specifically formed groups such as the Eco-towns Challenge Panel. Other networks were directly engaged in the more technical details of a proposed eco-town (for example, local urban designers, developers and partners undertaking the feasibility studies and proposals, or the Commission for Architecture and the Built Environment (CABE) design review panellists). Primarily collaborating on the practicalities of delivering a complex large-scale mixed-use development, they also gathered localized knowledge about sustainable development. The Eco-Development Group was an example of a more bottom-up local-level community of practice and an effective mechanism for sharing learning and testing the proposed policy approaches. Formed in 2009 as a learning and support network for the pilot schemes, it is now funded by participating local authorities and arranges visits, seminars and workshops to facilitate the exchange of European best practice. Rebranded in 2013 as the New Communities Group, it expanded to support non eco-town local authorities, delivery bodies and other groups working on large-scale sustainable developments of different models: from garden city-inspired new towns and villages to urban regeneration and extension schemes.
Several national and regional level advisory panels were formed to assess and improve proposed schemes. The Challenge Panel commented on the bids across all issues: transport, people, health, jobs, economics and selecting suitable sites. The CABE panel reviewed the master-planning, sustainability and delivery aspirations of all pilot schemes and continued to comment as subsequent phases were submitted for planning. The policy expected that ecotowns would explore novel ways of living, but the interviewees who reviewed the schemes felt that few proposals promoted transformational lifestyles. The initial master plans struggled to unite visions for place-making with suitably challenging sustainability targets. There was a pressure to bring designs to review at an early, possibly too unformed, stage, and as a result many recycled volume house builders' standard solutions, reviving familiar planning schemes. Panel participants were also drawn from a small pool of experts. Key individuals playing multiple roles re-emerged as advisors, or decision-makers for selected schemes, moving between organizations and groups. So as well as the explicit process of drawing up guidance, translating knowledge across the panels was highly reliant on the transfer role of these 'knowledge brokers' (Cooper, 2006). The interconnected nature of the expert and practice communities that policy and practice actors were part of led to the construction and promotion of 'favoured models'. This inherently political aspect of transferring concepts was reinforced by interaction between particular policy-makers and particular sites such as the study trips to sustainable settlements in the Netherlands and Germany or visits by DCLG officials to Sweden. 7 The gathered evidence and tacit knowledge from all these groups and individuals was codified into a hierarchy of written guidance. The Eco-towns Prospectus signposts existing planning policies with which the eco-towns were expected to comply (DCLG, 2007a). It listed the documents required to describe the potential eco-town bids. Applications were to be accompanied by detailed master plans, design and access statements, sustainability action plans, longterm governance proposals, calculations of carbon emissions, financial assessments, as well as transport, flood, and other written assessments and construction specifications (and later subject to legal planning Section 106 agreements 8 ). These drawings and written assessments were used as 'evidence' in at least two ways: the DCLG evaluation/selection process, and then for the Challenge Panel/CABE eco-towns design review panel, both meetings where the drawings and implied visions were subject to a process of interpretation and scrutiny. Some of these policy documents can be considered as boundary objects (like 'knowledge brokers'), moving critical concepts across networks. One example was the report What Makes an Eco-Town? (CABE & BioRegional, 2008) that summarized the metrics for sustainability discussed in the Challenge Panel, design reviews and was used as a foundation for the PPS:ET. Checklists and toolkits such as these are ways of creating hybridized knowledge (sitting somewhere between technical depth and bureaucratic organizational knowledge) to be contextualized to a particular locality and scheme.

Criticisms of the eco-towns concept, and evolving criteria
Submission and assessment of proposals were played out under a public spotlight, and met differing forms of opposition. There were several high-profile NIMBY 9 campaigns against the Oxfordshire and Stratford proposals, and lack of local authority support for proposals resulted in planning rejections for the government-shortlisted Middle Quinton and Ford proposals in 2008. The architect Richard Rogers argued that the eco-towns were anti-urban (Olcayto, 2008). This open criticism led to many of the eco-town criteria being questioned, at both a theoretical generic level (by the TCPA, CABE, BioRegional, etc.) and then more pragmatically around specific bids. Any generic support for the strategic vision of more sustainable housing quickly became an argument over detailed numbers of housing, distance from existing settlements and the unambitious sustainability targets.
The ideal size of a new eco-town was disputed. There was consensus around the ambition to construct complete new settlements, but the proposed 5000 homes (sufficient for supporting a secondary school) was felt to be substantially too small-scale to provide a critical occupational density for the place to be self-supporting. A new stand-alone settlement would require 15 000-30 000 homes to be viable. Terry Farrell (Farrells & Hyder, 2010) also related the size of the towns to their location within the existing urban/ rural network, recommending that they be attached to an existing town, not just to attract sufficient residents but to provide the critical mass of population necessary to kick start the new community. The idea of separate distinct 'eco enclaves' was derided: A PPS which directs separation of eco-towns into isolated eco-enclaves is folly and will grind to a standstill in the planning system. Eco-towns need to be a more integrated initiative and they need to be seen to give something back to the existing community. (Twinn, 2008) So despite the evidence behind the criteria, several of the proposals (particularly in the second wave of bids) were considered even if they were below the size thresholds or attached to existing settlements, showing how the criteria thresholds mutated and evolved as the theoretical national criteria were tested against local circumstances; for example, both Whitehill Borden and North West Bicester brown field regeneration schemes were 'within existing municipal boundaries' (Tomozeiu & Joss, 2014).

Impact of the National Planning Policy Framework (NPPF) and growth of localism
After the selection of four pilots, a second wave of bids was announced in December 2009, but they were rapidly cancelled immediately following the election of the new Conservative/Liberal Democrat Coalition Government in June 2010 (Hardman, 2010). The major policy shift was the increased support for localized visions through the growing localism agenda and more specifically the NPPF's relaxation of national planning policy. A consequence of the NPPF (DCLG, 2012) enactment in April 2012 was that the PPS:ET was no longer mandatory. Strategically, the eco-town pilots should have had little difficulty conforming to the five extremely high-level and generic principles quoted within the NPPF (living within environmental limits; ensuring a strong, healthy and just society; achieving a sustainable economy; promoting good governance; and using sound science responsibly) (DEFRA, 2005, cited in DCLG, 2012) to assess whether a scheme was acceptable under the presumption in favour of sustainable development. Yet the eco-town pilots demonstrated very differing responses to the removal of this national-level guidance and the growth in devolved local responsibility. At one of the pilots the easing of centralized direction was met with relief and enthusiasm. 10 A alternative response was seen at North West Bicester where the council and team remained committed to implementing the PPS:ET guidance. Developers involved responded positively to the detail in the PPS:ET as it provided them with a clear set of aims and clarity for expected standards. At the point the NPPF was enacted, the local authorities involved were struggling to get their core strategies adopted. As a way of retaining its content, Cherwell District Council bravely translated the PPS:ET (almost word for word) into a single shared vision for North West Bicester to apply across all scales of local decision-making by the county, district and town councils (Eco Bicester Strategic Delivery Board, 2010). There were risks relying on this local-level adoption process, but it was strongly supported by the council. This confidence in the achievability of the PPS:ET standards, embedded at a local level, suggested that the sustainability standards it embodied were deliverable and not overly prescriptive. 11 Limitations to the eco-town process The early history of the eco-town pilots has been subject to criticism (Cleary, 2014;Hardman, 2010;Manns, 2008;Tomozeiu & Joss, 2014), yet positive practical and theoretical lessons have emerged. As well as suggesting criteria for monitoring and disseminating findings, the What is an Eco-Town? report (CABE & BioRegional, 2008) acknowledged that while the initiative was an unparalleled learning process for all participants, lessons could be identified from both its successes and failures. It is too soon to assess the success of the pilot schemes being built out, although North West Bicester is making good progress and positively delivering 'early-win' projects for existing residents (McGregor, 2014). Nonetheless, the following constraints on the initial policy formation phase are observable, which future policy should strive to avoid: . The initial eco-towns policy was highly interventionist, dictated by an intense and high-profile process that altered and re-altered (Cleary, 2014).
. The concept had insufficient resilience to changes in economic and political context. Despite recent improvement in the economy, the financial viability of sustainable development continues to be questioned, particularly the financial uncertainties of innovation (Morton, 2011).
. Too rapid policy development. Despite the expert advice and structured challenge process, there was a sense of disjuncture between the policy vision and the hastily prepared central government policy guidance emerging from DCLG, particularly on issues such as size thresholds and how these might adjust to suit local situations. Insufficient time was allowed for consideration of the complex issues around sustainable development, especially for large-scale new settlements.
. Too many agencies were involved. There was additional blurring of roles through allocation of specific tasks: DCLG's drive as a coordinating delivery hub was reliant on TCPA's pivotal intellectual contribution.
. Practical local advice for schemes occurred too late in the bidding process. So CABE's technical advice happened in parallel rather than in preparation for the review panels. There was no long-term coordinated provision of technical support for the selected eco-town sites.
. Failure to set exemplar sustainability targets. Conflict arose from pitching a distinct eco-towns policy with one (unambitious) definition of sustainable development against the mainstream incremental stepping-up of housing sustainability standards via the Building Regulations or CSH. The initial eco-towns vision lacked sufficiently demanding sustainability standards with the PPS:ET only requiring housing at CSH Level 4 (although the pilots individually had more aspirational targets with an aim for CSH level 5+ set at North West Bicester) (Eco Bicester Strategic Delivery Board, 2010).
In April 2011, the Coalition Government (elected in 2010 and led by Prime Minister David Cameron) announced that only one of the proposed eco-towns, North West Bicester, would now be built to the originally proposed sustainability standards. The other proposed eco-towns will only need to meet current building regulation requirements as applied to any new build dwelling. The sustainable construction targets were further diluted by the reductions of zero carbon homes targets by excluding unregulated carbon emissions.

Garden cities revive a vision for green living
The concept of constructing new communities has recently been revived by DCLG's garden cities initiative, again with the TCPA as a lead proponent of the concept. At first glance the TCPA definition of a garden city as: a holistically planned new settlement which enhances the natural environment and offers high-quality affordable housing and locally accessible work in beautiful, healthy and sociable communities (TCPA, 2014b, p. 3) is little different from the aspirations for the eco-towns.
The new Coalition Government first mooted a new policy for large-scale extensions in early 2011 as part of the Housing Strategy (DCLG, 2011b). The NPPF referred directly to these settlements as following the principles of garden cities (DCLG, 2012, para. 52). But progress was extremely slow, and 18 months later, in November 2012, the unresolved debate between the coalition partners continued on the nature of these larger scale developments. In November 2013 Kriss Hopkins, then Housing Minister, dismissed the 'last Government's failed so-called ecotowns programme which built nothing but resentment' (Dunton, 2014, p. 2;Hopkins, 2013, col. 85 W) claiming that the coalition's planned new towns would lead to the creation of 69 000 new homes. Stripping back the confrontational parliamentary rhetoric and distancing it from previous government policy, three ecotown elements remained in the rebranded garden cities version: working with local communities, building more homes and promoting sustainable development. Similarly unsurprising following the NPPF and Localism Act was the reframing of the policy as locally led: However brave the launch, and however noble the concept, a Garden City or new settlement scheme is unlikely to happen without a delivery route that has local legitimacy. (Cleary, 2014, p. 3) This 'lively debate' continued until January 2014 when the Daily Telegraph newspaper reported the suppression of the 'long-awaited' prospectus alongside leaked locations of potential sites in Oxfordshire, Buckinghamshire or Warwickshire (Daily Telegraph, 11 January 2014, reported in Dunton, 2014). When the prospectus was published in April 2014, it endorsed the TCPA's nine garden city principles (TCPA, 2012b, p. 3). 12 However the TCPA emphasized how the policy-making agenda had retreated: It is important to note that the prospectus does not commit to these principles as Government policy. Instead it leaves the detail for the process open to local determination. Nonetheless the prospectus makes clear that the principles are the foundation for such local decision-making. (TCPA, 2014a, p. 5) Local authorities were invited to submit their garden cities proposals in August 2014, yet the initial two announced by the government are not new proposals but a rebranding of two schemes already underway: Ebbsfleet in Kent (despite failing to include any of the TCPA's garden city principles within the consultation document for the scheme; Robertson, 2014); and the North West Bicester eco-town (Spurr, 2014). In parallel, the TCPA has provided greater detail on the practical implementation of their nine principles (TCPA, 2014a, 2014b) including a 'Garden Cities Myth-buster' jointly published with developers Crest Nicholson (TCPA & Crest Nicholson, 2014) in response to misconceptions about garden cities. A detailed comparison of the criteria and principles for both Prospectuses is contained in Appendix C in the Supplemental Data online, and while the inherited similarities are clear, the garden city criteria are even less exacting or prescriptive. One can question whether a similar disjuncture between the generic principles and the local versions will emerge as with eco-towns. There has been a plethora of reports and conferences discussing and proposing alternative approaches to constructing new settlements (Adonis, Rogers, & Sims, 2014;McFarlane, Pearce, Ciaglia, Czischke, & Pattison, 2013;Morton, 2011) implying that the discussion on what constitutes a garden city may be broader and more inclusive than the ecotowns process. However, the types of networks, the individuals involved and the favoured European examples remain the same.
The TCPA guidance now describes the 'art' of building garden cities, a subtle decline in credibility from the earlier technical instructions. The current Garden Cities Prospectus (2014) is a very different document to the eco-towns one. It is plain text unadorned by colour or photographs and contains far fewer references to guidance -with none earlier than TCPA's (2008) eco-town worksheets except for the nostalgic centenary reprint of Raymond Unwin's How the Garden City Type of Devolvement may Benefit both Owner and Occupier (TCPA, 2012b). There are a handful of links to external agencies, and other than a mention of Ebbsfleet (the proposed site for a garden city) it promotes no exemplar projects or case studies. It opens, however, with a statement reinforcing that the focus of this initiative is housing supply and the need to build more homes after the economic crash.

Conclusions: learning lessons for future policy
The former DCLG Head of Growth Areas Division leading on the eco-towns argued that a critical lesson of the inappropriateness of top-down processes has been heeded. His recommendations for subsequent policy are: greater consistency given the extended timeframes; precise criteria, such as selecting areas of highest housing need, or most pressing affordability; informed local debate about local housing choice; serious consideration given to alternative options despite NIMBYism; and government infrastructure brokerage (including an early 'show-stopper review' to let promoters veto scheme if say transport is overly expensive), and for bids to relate to realistic local economic priorities. (Cleary, 2014, p. 2) The Highbury Group of housing academics also identified several practical preconditions required for any new garden cities based on their critique of the TCPA report Creating Garden Cities and Garden Suburbs Today (TCPA, 2012a). Conditions required include: Policy to reality for English eco-towns local employment potential -most likely to arise from extensions to existing towns; a mix of housing types and tenure balance with sufficient social rented properties, and avoiding too many 'upmarket' homes; a higher density of 50 -120 dwellings per hectare, greater than the typical garden suburb density of 12 homes per acre (30 per hectare); and land acquisition by a public sector body or trust (Highbury Group, 2013). However, their most telling recommendation is to refute the idea that sustainable development can be a discrete activity from house building in general: There is no reason for having a specific set of standards applicable to 'eco-towns' or 'garden cities' or 'garden suburbs' which do not apply to other forms of significant new residential-led development.
(Highbury Group, 2013, p. 2) Henry Cleary called for a 'robust discussion' on ways to 'address the housing supply shortage including learning from the recent past' (Cleary, 2014, p. 2). Yet there is little evidence that the current government reviewed or reflected on the eco-towns experience. By relying on the TCPA's historical ideas as precedent (which even recommended a settlement size based on Ebenezer Howard's idealized 1898 populations; TCPA, 2012b), the garden city programme downplays the importance of devising innovative approaches. Disappointingly, the Lyons Review (2014) 13 and other recent trawls for housing evidence leave an impression that there are few novel solutions to the ongoing housing crisis that remain unconsidered without an explicit programme of speculative and evaluative research. From the first, the eco-town programme had the potential to be a unique showcase for pioneering radical sustainable development in the UK. However, the most influential lessons may not occur from the extreme innovations but from addressing the drivers and challenges of an established market for housing, and exploring a sustainable alternative to dense urban development as is happening at North West Bicester. Here housing is being built with medium levels of parking and a conventional midscale density, but by constructing the Middle England version of sustainable suburbs, with space for food, integrated energy generation and a sustainable urban drainage system (SUDs) or rainwater collection as standard, a more mainstream, achievable vision is being promoted (Farrells & Hyder, 2010). This seems closer to the garden city/garden suburb concept currently being promoted.
Tracing the transfer of policy concepts from eco-towns to the garden cities initiative showed how the technical content was introduced and applied differently across vertical and horizontal, national and local networks. Tracking influential 'knowledge brokers', those actors who reappeared promoting favoured concepts in new local contexts, was a similar process to the discourse analysis undertaken with the careful reading of guidance, political speeches and the descriptions of the eco-towns. By noting phrases that reappeared as a short hand for sustainable settlements it was possible to follow the more resilient concepts (prioritizing climate proofing, balanced communities, collaboration between local authorities, or local leadership) during the evolution from 'eco-communities', 'green-cities', 'eco-cities', 'eco-towns', 'garden cities' and 'garden suburbs' to 'sustainable town extensions'. This evolution of definition indicated the subtle shifts in intent as the concept diluted in scale and ambition, and as other more negative phrases ('building resentment', 'imposing edge-of-town developments') reappeared more frequently.
The broad thinking and initial findings from eco-towns has informed other projects. Yet ultimately one can question whether the evidence that supported the policies was presented in a format that could actively influence and implement change in practice, or whether it remained at a level of generic solutions. Throughout the initiative the aspiration for innovation was ambiguous: were the eco-towns test-bed demonstrators for one-off novel ideas or pilots to trial the roll out of replicable approaches? How realistic was the hope that learning from newly constructed urban extensions could be repeated in existing settlements? Little in the initial eco-towns guidance discussed transferring learning to improve the sustainability of existing places, and there has been scant assessment of the overall carbon reduction impact of a series of newbuild eco-towns against a programme of improved existing towns (even to a less ambitious sustainability standard).
More broadly this narrative has uncovered an underscrutinized area of policy formulation and transfer. The repetition of familiar examples/technical standards as subsequent initiatives that emerged suggests that the barriers to delivery are at a political/ideological level more than at a technical one. Greater theoretical consideration is needed of how some ideas on sustainable development persist despite premeditated or unintentional policy derailment. Stone portrays a 'benign view' of 'policy transfer as a voluntary process undertaken by civil servants and politicians seeking to emulate "best practice"' (Stone, 2004, p. 485). The 'evidence'-gathering supporting the ecotowns or TCPA's garden city principles aligns to this positive version of best practice. The eco-towns initiative functioned under the belief that if sufficiently persuasive evidence was presented, it would be taken up and used. King and Crewe (2013) take a more realistic view of the fluctuating power relationships that shape the shifting ideological/theoretical landscape that policy-makers work within (see Simmons, 2015, for an elaboration of these barriers). Political expediency, ministers pursuing personal agendas, the increasingly powerful role of special advisors and informational asymmetries are all visible in the unsuccessful promotion of new large-scale sustainable settlements in the UK. However much governments call for policyrelevant research, and policy-relevant knowledge to support 'evidence-based policy', the gap between policy-making and practice remains, highlighting the disconnect between policy authors and those who implement it.
The eco-town initiative did successfully stimulate a wide-ranging discourse about the criteria for a sustainable development, for example the sustainability of stand-alone settlements against urban extensions taking advantage of existing infrastructure, regenerating town centres, accommodating growth ensuring employment and other sustainable aims. The TCPA policy officer (who became the DCLG official responsible for the selection process) portrayed the initiative as a way of broaching the difficult topic of town extensions, using sustainability as the guise to 'hammer out' more sensitive concerns. While this was less successful at the micro-level it did allow strategic conversions to begin: 'the big bold visions enabled the issues to be aired and dialogue started'. 14 However, if the initial political vision had taken a more explicit stance on the role of evidence, experimentation and learning, then alternative policy approaches and delivery models might have been considered, perhaps focusing efforts into a single national pilot similar to Hammarby Sjö stad in Sweden, or the alternative distributed local approach employed in the Vorarlberg region of Austria. UK versions of these alternative top-down/ bottom-up approaches can be imagined. Even proceeding with four extremely diverse and varied pilots as planned would have benefited from additional (nationally) coordinated expertise and support (for example, if the Eco Development Group had supported a genuine action learning programme), with committed plans for future roll-out out across a range of locations and circumstances. Within the current garden cities programme the government is again relying on the expertise of the TCPA to gather and disseminate existing and future learning.
Nonetheless, substantial novel thinking and protopolicy are now being generated by industry, from communities of practice formed around specific projects or via speculative idea generation such as the Wolfson Prize 2014 which asked entrants to speculate on the challenge of creating a new visionary, economically viable and popular garden city (Anon., 2014). The Wolfson competition re-stimulated industry debate towards positive solutions to overcome local 'NIMBYism'. The positive media response was a reminder of the quiet popularity and public support for the concept of garden cities. The short-listed schemes presented varied alternative models, providing a range of inspirational approaches not only supported by financially rigorous development models (as expected for an economics prize) but also carefully considering the impact on existing legislation and governance powers. By suggesting plausible adjustments to The New Towns Act, the winning scheme demonstrated that the legal and policy mechanisms to construct new sustainable settlements are in place, but the long-term governmental commitment to delivery is lacking (URBED, 2014).
In addition, the outsourcing of policy development enabled successive governments to distance themselves from unpopular or contentious policy where politically desirable. A day after the announcement of URBED as winners of the Wolfson Prize, Housing Minister Brandon Lewis was able to declare 'Yesterday's proposal from Lord Wolfson's competition is not Government Policy and will not be taken up' (Clarke, 2014, para. 4). Characterizing URBED's proposal, which doubled the residential population of Oxford in a sustainable manner, as 'urban sprawl', the Minister deflected the discussion towards the contentious issue of building within the Green Belt rather than on ways to encourage local support for sorely needed housing. Repeating his predecessor's sound bite: 'Picking housing numbers out of thin air and imposing them on local communities builds nothing but resentment' (para. 9), Lewis overlooked the extent that URBED's analysis of existing social and physical infrastructure dictated an appropriate housing mix and density and that a locally led proposition was a key component of all of the prize-winning entries.
This local participation is essential, and this paper shows how the UK eco-town policy evolved and mutated as the national political and economic context altered, and local players implemented their own versions of the national policy vision. Yet despite the disappointing disintegration of the initiative, there are several policy lessons from it that could helpfully inform any future attempts at constructing large-scale sustainable settlements. Three key lessons echo the triple-bottom-line balance of social, environmental and economic sustainability. The failure of eco-towns as a top-down national policy coincided with the birth of localism, reinforcing the principle that any new settlement must be locally led; the insufficiently innovative sustainability approaches in the bids were a response to the undemanding targets set in the policy vision; and the hard financial realities after the Global Financial Crisis of 2008 resulted in economic reassessment again with risk limiting funding innovation. Tomozeiu and Joss (2014) argue that the eco-town initiative unevenly concentrated on the socioeconomic aspects of sustainability. They evidence this via the initiative's emergence from the Sustainable Communities Plan (ODPM, 2003) (itself concentrating more Policy to reality for English eco-towns on delivery of greater numbers of homes than addressing climate change) as well as the dominance of the DCLG over other departments. They conclude that this overreliance on one agenda contributed to the downfall of the initiative, as DCLG was heavily affected by the decentralizing shift of the new Coalition Government. Yet there is a greater risk than an inability to weather inevitable financial and political shifts. Referring to the three-legged stool of sustainability in The Eco Towns: Scoping Report (2007), TCPA Chairman David Lock noted: 'miss out any one of these three aspects, and the project will very likely fall over' (TCPA & Lock, 2007a, p. 1). With the increasing reliance on private funding, the pendulum of the garden cities programme has swung far towards economic dominance. As a result, the three-legged approach to sustainable settlement continues to wobble. The hardest lesson from the eco-towns initiative remains unanswered: that sustainable development (however well evidenced, beautifully designed or inclusively delivered) is unlikely to provide sufficient incentive to win the approval of local communities. Nor can it survive as the driver for national housing policy. To reconcile the national and local policies recognition is needed that, unlike a more traditional linear view of policy transfer, the process of formulating and implementing sustainability policy is less one of direct imitation and replication, but more complicated and piecemeal. The eco-towns story highlights that a critical step in the policy cycle is understanding what is relevant 'here' in 'this locality' and to work outwards in both directions: reflecting and contributing to the national vision, while providing locally suitable sustainable solutions.
Endnotes 1 See Stone's (2004) traditional view of policy transfer which described a 'free market' for policy, where supply and demand actors exchange and select policies based on expected excellence of performance. This market view accentuates successful transfers, emphasizing replicability based on perceived similarities, or generic criteria, as part of the process of deriving 'best practice' or 'lessons learnt' to be repeated where similar conditions are believed to occur. Underlying this is the suggestion of the decay of ideas as successful 'exemplar' cities transfer policy and ideas to 'lesser' cities in an imperfect process of reproduction (Healey, 2010). 2 The key lesson from the literature on international eco-towns is the significance of the local context -responding to local municipalities' existing governance structures, funding models that can reflect existing land ownership patterns and sustained local leadership creating characterful places with a strong identity. This is achieved not top down by 'imposing an unwanted burden' but by involving the whole community in deciding on the kind of place in which they want to live and physically and spatially providing settlements that work within the local infrastructure network (HCA, 2009, p. 16).
A key priority of the Northern European eco-schemes was to equip local organizations and authorities with the skills to deliver better quality housing more rapidly than in the UK (PRP et al., 2008). 3 Interviewees of varying seniority were identified to cover a breadth of disciplines and viewpoints. Selection targeted those who actively provided strategic and practical advice to schemes, for example the project managers for Bicester and Whitehill Borden eco-town projects who made the procurement decisions, the civil servant who drafted the policy documents and guidance, or the design advisor who ran and selected participants for the review panels. Access was aided by the author's professional experience working on the fringes of the initial eco-towns policy development, and by the opportunity to talk with and observe colleagues working on eco-towns. I do not claim to have taken either a particularly objective or an immersive ethnographic approach. However, in endeavouring to construct a theoretical framework and interpretation of the activities and techniques that were part of the shared day-to-day practice, interviewees have commented that the interviews provided an unusual opportunity to reflect on the nature of the process in addition to recording its trajectory and outcomes. The research provided space to articulate what they had learnt, and how they have subsequently extracted value from the experience; with the interview process itself considered another effective mechanism to translate and communicate policy.