Exploring the ‘Shadows’ in the Implementation Processes for National Anti-fraud Strategies at the Local Level: Aims, Ownership, and Impact

This article studies the development and implementation of a nationally drafted strategy for fraud in local government in England. The purpose, relevant to other countries which also face (or evade) problems of policy outreach, is to consider what is required to achieve effective implementation through three broad aspects: aims, ownership, and impact. There is a particular focus on the assessment of the strategy implementation process and what mechanisms translate strategies into effective delivery or what other factors may subvert that delivery. The empirical research draws on the UK government’s 2006 policy review on fraud and the consequential changes between then and 2019, including a number of fraud strategies initiated by the United Kingdom central government. To review implementation in practice, it focuses on the 2011 local government strategy and uses a local-level case study to assess issues concerning aims, ownership, and impact, as well as the effect of other nationally determined policies and agendas. It concludes that without an ‘owned’ strategy implementation process as a whole, national strategies concerning the prevention and policing of fraud in England in the twenty-first century have had—and continue to have—modest impact on practice on the ground at the local level. We find it plausible that this is true elsewhere in the world, not only for crime control but also for other ‘change’ strategies in the public sector. However, testing that proposition is for researchers in other countries. Our aim here is to use this superficially parochial study to raise more universal questions about how policy designers (and academic researchers) need to take better account of circumstances on the ground, the management of strategy implementation and legitimacy, if their strategies are to be more than merely symbolic rhetoric.


Introduction: The Focus of the Article
Towards the end of the twentieth and into the twenty-first century, strategic planning in public sectors in developed countries has 'become ubiquitous' (Poister 2010, p. 246). The traditional approach of a top-down and structured, nationally determined strategy has been predicated on 'getting people to follow precise procedures for the introduction of new policies and strategies, and the use of systems that can ensure that such organizational changes follow a pre-planned sequence of steps' (Andrews et al. 2017, p. 2). The contemporary prevailing perspective from the research into strategy and strategy implementation, however, recognises that such an approach is no longer likely to deliver effective implementation in late modern government. This perspective proposes an evolving or incremental approach where there is devolved responsibility among a number of agencies, the need for review, revision and resourcing, and a means for monitoring, oversight and guidance in recognition of 'a much looser distinction between formulation and implementation' (Andrews et al. 2011, p. 647) 1 .
The use of an emergent or incremental approach to strategy development and implementation which involves more than central government is relevant to a number of developed countries, including England, including the criminal justice sector: 'crime control-and, indeed, the broader task of "policing"-is now portrayed as much more of a joint exercise, involving partnership and information-sharing with a variety of agencies…Importantly, too, decisions about which crimes or offenders to focus on appear to be more strongly influenced by explicit policies, strategies or 'policing plans'-themselves ostensibly the result of careful analyses of crime patterns and trends, as well as taking account of the priorities of central and local government, other agencies and local communities' (Levi and Maguire 2012, p. 195).
The focus of this article is on strategy implementation and on the consequences of a continued reliance on the traditional approach to implementation. Despite the fact that 'strategic management tools and ideas have been brought into play by governments across the world to enhance capacities and performance standards in the face of increasingly challenging economic and political circumstances… research on the dynamics of strategy implementation in the public sector is still in its infancy' (Andrews et al. 2017, p. 1). We argue that implementation is crucial to give meaning to strategy, because 'there is no point in having a strategy if one has no road map for how to get from here to there' (Levi 2013 p. 15). However, for a strategy not to end up as 'one more document to "turn the corner"' (Levi et al. 2003, p. 24) for whatever crime control issue is to be addressed, then working out what is necessary to ensure effective implementation requires consideration of three components. These are '(a) definitions and aims, (b) "ownership" of the problem and (c) how to determine the impact of interventions' (Levi and Maguire 2004, p. 398).
Within a framework of the three components, and assuming that any strategy implementation would be expected to adopt an evolving or incremental approach, this article aims to use UK strategies intended to address fraud as an example of the need for a broader understanding of the antecedents and dynamics of successful strategy implementation 2 . In presenting findings on fraud in England, the local level of government 3 is used as a case study because it is the sector most distant from the national level, and therefore is more likely to be exposed to a greater range of other national policies and agendas which may facilitate or inhibit implementation, including those that may not be taken into account by high-level strategic ambitions that focus solely on fraud. Our focus here is on the strategy implementation process irrespective of whether the policies, when implemented, generate the desired effect on fraud itself.
The findings indicate that national governments' use of the traditional approach to strategies, (lack of) engagement with local level, ownership, resources, and the influence of other policies and agendas pose challenges to successful implementation. Although it is a study based upon only one region in England and upon only one type of crime, the research provides empirical evidence concerning tensions that exist between what a national strategy envisages to be, or should be, done and the responses at a local level to accept, shape, or ignore such strategies in relation to what they see (with their local knowledge) as needing to be done within the obvious (to them) local contexts, priorities, and resource constraints they have. On the assumption that the goals of the strategy are more than merely rhetorical, the article raises questions about the effectiveness of national crime control strategies in terms of what is stated or expected by central government, the general approach to how such strategies should be formulated and put into effect, if practice on the ground is likely to be very different from how central government envisaged. Of course, the political economy and legal components of policy and cross-national policy mobilities can be examined without reference to implementation and impact (outputs and outcomes). However, we hope that this analysis will resonate with criminology and public policy readers and provide perspectives that will facilitate further research on both fraud and non-fraud policy around the world into the strategy implementation process.

Components and Concepts
The three components generally provide the overall framework: why have a strategy and what is its purpose; how is it to be implemented, by whom, and who has overall responsibility for both the strategy and its implementation; and how to determine the impact of a strategy. The components draw on conceptual issues or perspectives from the literature which identifies two main issues. First, there are 'few studies that actually examine the antecedents of successful strategy implementation in public service 2 Fraud is the loss of income or funds by internal or external perpetrators accessing services, resources or functions through a variety of means, from abuse of trust, deception, misappropriation, or misrepresentation (see Doig 2006 and, for legal provisions, the Fraud Act 2006 and case law). 3 Local government in the UK is 'divided in some areas into county councils (the upper tier) and district councils (the lower tier). A major reform in 1972 established county and district councils throughout England, Wales and Scotland (excluding three Scottish island authorities). Since then, unitary authorities have been established in a number of areas. The creation of unitary authorities has frequently been contingent on Government policy, or local initiative, at a given time, rather than any rationale relating to local economy, geography or identity. The two tiers have distinct functions, though they overlap in some matters. In other areas, "unitary authorities" carry out all local government functions. There are 353 local authorities in England, of which 27 are county councils, 201 are district councils, and 125 are single-tier authorities. Of the latter, 32 are London boroughs and 36 are metropolitan boroughs' (Sandford 2017, p. 4). organizations… with little systematic investigation of any of the major themes relating to implementation success' (Elbanna et al. 2016(Elbanna et al. , p. 1018. Second, there is limited research into the 'alignment across strategy, process, structure, and environment' (Walker 2013, p. 682), while 'none of the articles… found treated strategic planning and management as an iterative process in which, for example, difficulties arising in implementing a strategy may lead to reviewing and perhaps revising it' (Poister et al. 2010, p. 540).
In terms of the first component, the development of a national policy and its translation through a strategy for implementation involves the conventional phases of the process, which begin with 'defining problems and setting agendas and end with implementing and evaluating policy' (Parsons 1995, p. 81). It includes both 'the deployment of resources and development of competences [as] an issue of central importance to successful implementation of strategies' (Johnson and Scholes 1997, p. 411). The practice of policy implementation through a strategy within the dynamic and continuing nature of many crime issues and the complexities of the crime control organisational landscapes makes the traditional, top-down, approach unsuitable. Here an evolving or incremental approach would appear necessary: 'whereas implementation was previously viewed as largely a "top-down" process which followed the policy making "stage", it is now widely recognized that implementation should also be seen as a "bottom-up" process too, whereby street level bureaucrats …both shape public policy through the manner in which they apply it "on the ground", and themselves identify problems with the original policy, which can then be reported back, and eventually lead to policy modification' (Dorey 2005, p. 3).
The second component, that of ownership, has two aspects. The first concerns the best way to manage the implementation process via strategic planning (with 'top-down plans') or strategic control (where devolution is part of the strategy implementation process) (Johnson and Scholes 1997, p. 424). If strategic control is opted for, the second component may require mechanisms to ensure continuing oversight of 'the ways in which the policy is being adapted, and attempting to steer the system if it is deviating too far away from the high level goals for the policy' (Hallsworth 2011, p. 10), including the means by which 'corrective action can be taken to alter strategic direction…in the context of changing circumstances' (McKevitt 1992, p. 35). The second component also concerns who is responsible for the continuing management of the implementation process in a context where 'the translation of policy intent into action is highly complex, involving many people inside and outside government and requiring activity across many fronts' (Norris et al. 2014, p. 23).
The third component concerns effective implementation in assessing strategy impact: 'implementation is widely held to be a critical element of strategy and one which can have a significant impact on performance' (Andrews et al. 2011, p. 663). In the development of a nationally determined strategy, governments need to monitor implementation, such as consistency between the strategy, strategic objectives, and their expression in operational terms, as well as taking account of 'what is happening in the environment' and 'changes that need to occur in organisational culture and operational routines' (see Johnson and Scholes 1997, p. 494). This in turn involves assessment or evaluation-impact-of the strategy to consider: '(1) whether projects are being conducted effectively, in order to learn from and improve project activities; (2) whether the program is making a difference to people, groups, organizations or communities; and (3) to use that evidence of impact to advocate for continued support and/or funding from relevant stakeholders' (Streatfield and Markless 2009, p. 139).

Methods
The research involves a combined approach-publications in relevant areas, insights into official material, and interviews with those involved (see McNabb 2013)-using a case study approach. Here the 'goal of case research is to describe what is happening, why it is happening, and how it happens in an organization…[to] develop concepts, insights, and understandings from patterns they see in their collected data' (see McNabb 2015, p. xix). The literature review and document analysis comprised desk reviews of official and other documentation, as well as public domain sources and academic publications. The North East (NE) region was chosen because it is seen as a coherent but representative grouping of 12 councils; 10 of the 12 have (often long-standing) Labour Party majorities, while two have no one party in control. The population size ranges from 92,000 to over 500,000, and the levels of expenditure range from over £900 million to under £200 million. There are some affluent areas, and others that have been marked by severe decline following the end of heavy industry and coal mining. All were previously the constituent members of the Association of North East Councils and who, under the 2009 Local Democracy, Economic Development and Construction Act, are now members of either the North East or Tees Valley combined authorities. They are covered geographically by three police forces. The choice is based on the pre-existing understanding of, and research into, fraud and fraud control in the region (Doig 2017(Doig , 2018a(Doig , 2018b. These prior relationships facilitated access and also enabled more insightful interpretation of semi-structured interviews (see Pfadenhauer 2009).
Data on fraud and resources were accessed through Freedom of Information requests in 2016 and 2017 from the then Department for Communities and Local Government (DCLG, now the Ministry of Housing, Communities and Local Government [DHCLG]), the police, and local councils. Semi-structured interviews were conducted with seven representatives of three police forces' fraud units or squads, the DCLG, and the Department for Work and Pensions (DWP, the government department responsible for payment of social and welfare benefits). Specifically in relation to practice on the ground, a further five representatives-all senior internal audit staff-from three of the 12 councils were interviewed in 2016 and 2018. The focus of the interviews was the awareness of the strategy developed for local government, any involvement in aspects of the strategy (such as cooperation or partnerships with other agencies), and issues of support. For the internal audit staff, this focus was supplemented by questions on resources, internal adoption or adaptation of the local government anti-fraud strategy, external oversight, and influences that adversely affected implementation.
Those interviewed and their organisations are anonymised. Interviews were recorded by note rather than tape recorder. No direct quotes are used, because the intention was to explore the processes within organisations rather than the role of specific individuals. (All those interviewed were invited to comment on the draft article, and these reactions have been incorporated where convincing.) The study has limitations in terms of 'the size of sample, the snapshot nature of the research, or the restriction to one geographical area of an organisation' (Saunders et al. 2009, p. 538). Nevertheless, the case study approach enables 'the researcher to answer "how" and "why" type questions, while taking into consideration how a phenomenon is influenced by the context within which it is situated' (Baxter and Jack 2008, p. 556). Further 'research in public administration that concentrates on the subject of public policy aims to contribute to a better insight into and a more efficient use of policies: the intention is to improve policy in an instrumental sense' (van Thiel 2014, p. 6).

Context: From Concern to Policy
It is well understood that social problems have to be constructed before they are regarded as such, and that fraud is not naturally regarded as a major social problem because, historically, it has not been associated with the 'dangerous classes' and does not require a violent takeover of others' property or person (Levi 2009). Fraud of any kind has never been a high policing priority (Levi and Jones 1985;Levi 1987;HMICFRS 2019). However, the potential risk posed by fraud was highlighted during the early twenty-first century as requiring a national policy response at law enforcement and government levels (see Doig et al. 2001;Home Office 2004). It was listed by the 2003 UK Threat Assessment issued by the National Criminal Intelligence Service (NCIS 2003) as one of the seven most significant (crime) threats facing the United Kingdom. In 2002 a Cabinet Office report blamed emerging types of fraud, such as identity fraud and theft, as 'the key to much financial fraud, for both the public and the private sectors' (Cabinet Office 2002, p. 3). A year later, the Attorney General was emphasising its links to other strategic threats, including terrorism and organised crime (Fraud Advisory Panel 2004, p. ii).
Lacking a strong constituency in the Home Office-the normal origin for criminal policythe issue was taken up by Peter (now Lord) Goldsmith, the Labour Attorney General (AG), who had been convinced of its importance as a policy issue and the need for the justice system to do something about it. In October 2005, the UK Attorney General proposed a national policy review to Cabinet: 'the Fraud Review wasn't at the top of anyone else's agenda, but generally the cabinet was supportive' (Fraud Advisory Panel 2016, p. 3). Led by staff from the AG's office, the review's two reports pointed to a number of problems, from the absence of a strategy on fraud and dwindling police resources to a fragmented approach to fraud in organisations and inadequate application of sanctions. Their recommendations comprised a range of thematic and institutional recommendations that included a comprehensive measurement of fraud; a National Fraud Reporting Centre (later to be misleadingly labelled 'Action Fraud'); a police-based intelligence resource (later established as the National Fraud Intelligence Bureau [NFIB]); a lead police force, the City of London Police (CoLP; which also housed the NFIB); promotion of prevention good practice; and more public/private partnerships. This would be shaped by a national strategy and led by a National Fraud Strategic Authority (NFSA; later renamed the National Fraud Authority [NFA]). The Fraud Review had emphasised that to undertake anti-fraud work at a strategic level, 'any oversight body must have the authority to act to fulfil its functions' (Fraud Review 2006, p. 50), while the strategy was acknowledged as 'the key Fraud Review recommendation' (Attorney General's Office 2007, p. 3).

From Policy to Strategy
Reinforced by an analytical review of literature on the costs and distribution of fraud in the UK (Levi et al. 2007;Levi and Burrows 2008) which was part of the UK government's still ongoing desire to examine the evidence base for serious crime, the NFA published a national fraud strategy in 2008. This proposed a range of general and specific activities to build and share knowledge about fraud, tackle the most serious and harmful fraud threats, and disrupt and punish more fraudsters while improving support to their victims and enhancing the nation's long-term capability to prevent fraud. The continuing pressure, coherence, and consistency lay with the NFA, which supported a Counter Fraud Strategy Forum intended to 'strengthen senior strategic oversight of national fraud enforcement activity' (National Fraud Authority 2010, p. 5). With the Cabinet Office, it was also responsible for a central government Counter Fraud Taskforce (also termed the Fraud, Error and Debt Taskforce). The Taskforce's role was to 'coordinate action across Government and the wider public sector', 'understand the nature and scale of financial loss', and address 'barriers to tackle fraud effectively such as information and intelligence sharing, alongside exploring mechanisms that will embed a counter fraud and error culture across Government' (Cabinet Office 2012, p. 6). This included collaborative working between central and local government; this was important to the general legitimacy of counter-fraud efforts which, like 'organised crime policing', potentially ran the risk of being seen as a central government mission with little relevance to local issues.
In 2011 the NFA had published an overarching national strategy review, Fighting Fraud Together, to 'provide fresh impetus in our fight against fraud'. This review proposed a stronger institutional focus through the National Crime Agency (NCA), whose Economic Crime Command (ECC) would address constraints on police resources by developing 'innovative, partnership solutions working across police forces, the NCA and its Economic Crime Command, other law enforcement organisations and the public, private and voluntary sectors'. It also stated that 'the Government's Fraud, Error and Debt Taskforce will oversee the delivery of measures to tackle fraud against Government… It will also support efforts to tackle fraud in the wider public sector, including the NHS and local government' (NFA 2011, pp. 20, 26; see also Cabinet Office 2012).
In the same year, in conjunction with the Fighting Fraud Locally Oversight Board (FFL Oversight Board; an advisory body largely drawn from local government), the NFA published a local government fraud strategy as 'part of a wider collaboration on counter fraud and [as] the local authority contribution to the national fraud strategy-Fighting Fraud Together (FFL)which encompasses both the public and private sectors response to fraud in the UK' (NFA 2011, p. 7). The local government strategy-generally referred to as the FFL strategy-stated that, 'while individual local authorities have made significant progress in tackling fraud, there has not yet been a concerted and collaborative approach encompassing the whole of local government. Fighting Fraud Locally seeks, for the first time, to bring about a truly inclusive and collaborative approach to tackle fraud across all of local government' (FFL Oversight Board/NFA 2011, p. 13). It proposed that, by 2015, councils were expected to have a more effective fraud response mechanism, radically realigned fraud strategies, governance arrangements driving the delivery of a fraud strategy across local government, and collaborative arrangements with the police and central government departments.

Ownership
In 2013 the NFA was unexpectedly and unilaterally closed, and its functions redistributed. While the government professed that the NFA had been successful in raising awareness of fraud and improving coordination, it believed there should be a single national focus on cutting economic crime as part of the government's approach to serious and organised crime. This, the government determined, should be through the NCA ECC (the government stated that the ECC would be responsible for strategic development and threat analysis). 4 The City of London Police (CoLP) took over the central fraud reporting mechanism, Action Fraud, and the Home Office was stated to be working on raising awareness of fraud, while the NFA's local government work was handed over to the Chartered Institute for Public Finance and Accountancy (CIPFA; a professional association primarily for public sector internal auditors) (https://www.gov.uk/government/organisations/national-fraud-authority; retrieved May 2018 ). The closure of the NFA saw the end of the Counter Fraud Strategy Forum. The Fraud, Error and Debt Taskforce was closed in 2015. The Cabinet Office then set up a Fraud, Error, Debt and Grants Function which included a Centre of Expertise for Counter Fraud and Error Reduction with responsibility for understanding the cross-government picture on fraud (see Cabinet Office 2017, p. 8). The Cabinet Office itself became the policy lead for fraud for central government.
The other major change in terms of possible co-ownership with the NFA at the local level was the abolition of the Audit Commission. While modern external audit at a local level was long-standing, the 1982 Local Government Finance Act established the Audit Commission. It assumed central responsibility for the existing statutory external audit regime of local government, unifying its approach and determining time spent on fraud audits. It collected and published annual fraud and corruption data (published in its annual Protecting the Public Purse reports). It used its powers to set up and lead a national annual data-mining and matching initiative-the National Fraud Initiative (NFI)-from information provided by both public and private sectors. The results were sent as 'risk' matches to be reviewed as potential fraud, overpayments, and errors in public expenditure. The Audit Commission took a continuing and proactive stance on the need to identify and address financial loss and to ensure appropriate governance arrangements (see Public Audit Forum 2001). Its comprehensive performance assessment was intended to ensure that robust anti-fraud procedures were delivered within what it termed the softer characteristics of leadership and standards.
The political decision to close down the Audit Commission, again a decision that was 'unanticipated' (Sandford 2016a, p. 15;Tonkiss and Skelcher 2015), led to concerns about the future ownership of its anti-fraud responsibilities at the local level 'potentially being fragmented across different bodies in an uncoordinated and inconsistent way' (House of Commons Communities and Local Government Committee 2011, Ev93; see also House of Commons Draft Local Audit Bill ad hoc Committee 2012, Ev114). In the event, some of the Audit Commission's work on fraud was transferred to CIPFA, which set up a Counter-Fraud Centre, and the NFI was transferred to the Cabinet Office. Contracts for external audit were bid for and awarded to private sector companies by the Audit Commission and managed through Public Sector Audit Appointments Limited (PSAA), an independent company limited by guarantee incorporated by the Local Government Association (LGA; the representative association for local government).

Resourcing and Support
The abolition of the Audit Commission, as well as of the localism agenda (discussed below), was part of the neo-liberal dismantling of the Big State, but it has had largely unintended consequences for ownership of fraud control. Government itself continued to make available resourcing and support for anti-fraud work, via central government departments and the various bodies to which responsibilities previously undertaken by the NFA and Audit Commission were devolved. Apart from CIPFA, none of these departments or bodies were given any explicit responsibility for, or intended to give any resourcing or support for, the implementation of the FFL strategy.
The Department for Communities and Local Government (DCLG; now the Department for Housing, Communities and Local Government [DHCLG]) has primary central government responsibility for local government. It has a small team which has been primarily working on updating the code of conduct for elected officials at local-level members and supervising the external audit arrangements, but exercised no responsibilities in relation to the FFL strategy. In line with the policy imperatives of the localism agenda, it introduced the Local Government Transparency Code in 2011 (Department for Communities and Local Government 2011). This was a policy document stating what mandatory and optional information councils should provide annually to help citizens hold their councils to account and deliver better value for the money (see Department for Communities and Local Government 2011; Sandford 2016b). For counter-fraud work, the code required information on uses of powers relating to social housing fraud, and numbers, costs, and qualifications of staff. Optional information was proposed in relation to the number and value of cases, and funds recovered. The Department also made £19 million in grant funding available to local authorities to tackle social housing fraud over 4 years from 2011 to 2015. The earmarked fund was split into two sets of £9.5 million for each of two 2-year periods. In 2014 it established a competitive £14 million fund for short-term projects, usually up to 2 years' duration, for innovative ideas on how to address fraud; nearly 60 awards were made.
CIPFA's Counter Fraud Centre took on policy and training responsibilities for public sector fraud and corruption, as well as the publication of checklists and technical tools. The Centre also published the annual tracker report from voluntary returns from councils that was analogous to the Audit Commission's annual Protecting the Public Purse report. Here, however, as in the other areas, CIPFA was given no formal mandatory powers to ensure participation or compliance; it saw its role as a resource, networking hub, and disseminator of good practice. The Cabinet Office developed the NFI data-matching powers which it has seen as a technical exercise. Other bodies with fraud-related roles and responsibilities resulting from the closure of the NFA and the Audit Commission were less involved. The National Crime Agency-whose primary focus is on national and international organised crime-has had virtually no contact with councils in relation to strategic development on fraud. In 2016 Her Majesty's Inspectorate of Constabulary (HMIC) noted that the NCA had not delivered its strategic action plan on public sector fraud-an identified threat area-in response to an earlier HMIC recommendation (HMIC 2016, p. 24). The Home Office provided no support in terms of prevention, but as part of its cross-Government Serious and Organised Crime Strategy undertaken by its Strategic Centre for Organised Crime, the Centre sent out to local councils in December 2016 a voluntary intervention request for a serious and organised crime risk assessment and audit of procurement and other risk areas by individual councils (see Doig and Sproat 2019 forthcoming).
The only direct support for the FLL strategy came when the DHCLG allocated £600,000 to provide strategic support through funding of a network of counter-fraud authorities. In practice, only £100,000 was allocated to the ( Before looking at implementation issues on the ground, it is important to appreciate the changing circumstances in the public sector at that time that would influence the implementation process. First, fiscal austerity for the public sector meant that local government was 'dealing with unprecedented "budget gaps"-that is, a massive shortfall in resources resultant from the combination of funding reductions and cost pressures' (Hastings et al. 2015, p. 602), forcing cuts to the funding of functions and services. Second was the emerging Localism agenda. This was a government-inspired policy imperative to promote local activities, development, and modes of delivery across a range of services and functions, while minimising central government control and increasing local choices, albeit with diminished funds (see Lowndes and Pratchett 2011).
Third were the development of, and changes relating to, the policing of fraud. After the closure of the NFA, the National Crime Agency (NCA)-which was created in 2013 to replace the Serious Organised Crime Agency-led on the government's organised crime strategy. Its Economic Crime Command (ECC) took responsibility for serious, organised fraud, including but not exclusively involving organised crime groups. In a separate development, a draft policing fraud strategy was developed in 2015 for the police and fraud by the City of London Police's Economic Crime Directorate, which also placed the emphasis on organised crime or criminal networks, but left responsibility for 'local' fraud to local force discretion. A 2018 review noted that in practice, 'fraud is not prioritised strategically at the local level, neither in police and crime plans nor at an operational level by senior leadership teams' (Police Foundation 2018, p. 37). In other words, despite being designated as the national 'lead force', it was not willing (or arguably not able) to step in and investigate the fraud that local police were unwilling to resource. HMICFRS, the successor agency to the HMIC, were critical of the ambiguities in responsibility for fraud policing ( 2019), though they did not specifically comment on the policing of fraud against central or local government. 5 Fourth was the government's policy announcement in 2010 to revamp the payment of welfare and other benefits through the introduction of the universal credit (UC) benefit delivered by the Department of Work and Pensions (DWP). This included a proposal to incorporate the payment of housing benefit-which had been administered by local government-within the UC and a plan for a Single Fraud Investigation Service within the DWP. This involved the merger of council fraud investigators and DWP's existing local 5 For example, the HMICFRS report noted: "the current model of local investigations supported by national functions is the right one. And we have, as ever, found examples of some excellent work that is being done to tackle fraudsters and support their victims, particularly at a local level. But the police need a much more coordinated national approach with clear roles and responsibilities, clear operating procedures and a commitment to provide resources for the long term." It also noted: "in only two of the forces we inspected was fraud made an explicit priority. Some forces included fraud within other priorities of tackling economic crime or protecting vulnerable people. Fraud was often said to not score highly enough to be considered a priority when compared with other crimes…." ( 2019, pp. 4, 9). service staff, resulting in the potential transfer of some 1800 council investigators who had been funded and incentivised to administer housing benefit, including fraud investigations. This resource also provided councils with an in-house investigative capacity which, the Audit Commission estimated, was dealing with up to 40% of non-benefit fraud by 2010-2011 (see Doig 2014). This development severely reduced the capacity of councils to investigate fraud.
In summary, on the likely impact of the implementation of the FFL strategy at the local level, the intended aims are affected at the macro level by changing central-local relations. The central government's localism and austerity agendas intersected with the low priority and level of resource given to the policing of fraud in the declining expenditure on policing. This coincided with (a) the diminution of internal council resources following the roll-out of UC and (b) the absence of central guidance relating to controls and anti-fraud culture following the abolition of the Audit Commission. The financial austerity process, not linked directly to the strategy or allocated for specific objectives or uniformly, meant that while some councils could take forward innovative or creative responses, others' responses were driven by pragmatism and local circumstances. In particular, the abolition of the NFA and the diffusion of the various responsibilities both of it and of the Audit Commission eliminated central ownership, oversight, or generalised pressure to advance the FFL strategy implementation across all councils. It also removed the means to review and revise the strategy to take into account the impact of the other agendas and the availability of resources which even previously had been uncoordinated and rarely linked to supporting the strategy.

Reviewing 12 Councils' Published Approaches to Fraud
The FFL strategy had over 30 suggested anti-fraud and anti-corruption actions in its voluntary checklist. In terms of their delivery by the 12 councils, many councils appeared to undertake many-but not all-of them as discrete actions or activities rather than as part of an overall strategy implementation. Most of the councils published their own (and accessible) anti-fraud and anti-corruption strategy or policy. Most also referenced the two main CIPFA publications, the Code of Practice on Managing the Risk of Fraud and Corruption and the accompanying Guidance Notes. A lesser number had a related investigation or response plan. All referenced the NFI and all encouraged reporting of concerns, benefit fraud, and so on, by the public; most but not all appeared to have a published whistleblowing policy. Overall it would appear that while a small number of councils mentioned the 2011 FFL strategy, none used the strategy or the checklist as an overall framework, and none reported on how the council had delivered or was delivering the strategy. Only one mentioned ensuring that the whole council was aware of the council's approach to fraud and corruption; none provided any information on training or organisational culture. Nor was it evident what steps the central government had taken to encourage or require the councils to report back on their progress.
Such variations and disparities are present in the transparency code information (as of the end of 2018). Of the 12 councils, five appeared to make no online reference to counter-fraud work. Of the seven who did, one provided information for 2017, while two were undated, and the rest were dated before 2017. Only two provided the optional information. Some councils did and some did not differentiate between absolute and full-time equivalent (FTE), so the number of fraud-related staff was difficult to calculate. However, the figures would suggest fewer than 40 staff in the region overall (and a maximum of fewer than 20 FTEs), most with counter-fraud qualifications. The total reported amount spent on counter-fraud work by the seven councils was under £1 million (on approximately £3 billion annual expenditure). Reported investigated cases totalled over 3700 annually (but the range was from 11 to 1465; the median was approximately 450). Only two councils made public (and perhaps only two had calculated) the costs of the fraud (£1.5 million between them); only one council reported funds recovered (over £155,000 and a further £1.3 million saved by discontinued false applications for the purchase of council-owned houses).

Local Circumstances and the FFL Strategy: Arrangements on the Ground
Three councils were interviewed to determine how far they now reflected the expectations of the 2011 FFL strategy in terms of a more effective fraud response mechanism, radically realigned fraud strategies, governance arrangements driving the delivery of a fraud strategy across local government, and collaborative arrangements with the police and central government departments. In the three councils interviewed about practice, work on fraud continues to lie primarily with the internal audit function. All of these draw on the CIPFA material; one uses the CIPFA Fraud Tracker and the FFCL Oversight Board action plan for training and to identify current and future risks. All have moderate internal audit fraud capacity with a primary focus on internal procedures and controls, but localism and other agendas and the financial constraints had a significant influence on whether their practices and procedures would reflect the intent of the FFL strategy'.
Council A has an internal audit team of 17.4 FTEs whose audit plan allows for nearly 20% of auditors' time devoted to fraud prevention and protection work. Council B has 10 audit staff but, with audit responsibilities for a number of other bodies, has about 6.5 FTEs to undertake council work; the audit plan does not include allocated time for fraud work, although staff provide training in fraud awareness and expertise for complex cases. Council C has six FTE audit staff, with about 5% of the annual audit plan allocated for fraud work. Two of the councils were significantly disadvantaged by the Single Fraud Investigation Service. With many council fraud staff transferring over, councils in seeking to achieve government-imposed austerity savings targets were unwilling to prioritise the development of any new or replacement anti-fraud capacity. Since 2005, Council B's anti-fraud capacity had fallen from 20 investigators to four in 2015 and to none in 2016, after which it appointed a corporate fraud investigator who retired. Some 2 years later, a new post is being filled with a focus on social housing fraud; proposals for a corporate fraud team, proposed in 2016, are still being slowly developed.
For both Councils B and C, in line with the devolution and decentralisation of the provision of functions and services, much of the operational responsibility rested with specific departments. For Council B, for example, social services have their own audit capacity (although this is more about validating claims than pursuing fraud), while the council's parking services deal with disabled blue badge misuse, with its enforcement teams focusing on sanctions under the Traffic Act and the various changes brought in by the 2013 Disabled Persons' Parking Badges Act. For Council C, flatter arrangements mean that departmental managers are responsible for investigations, including NFI referrals. The revenue department has two retained investigator staff dealing primarily with council tax discounts and other council tax fraud. Additionally, Council C uses two private sector companies to provide data-matching facilities, checking council data against commercial credit data linked to the same addresses. The companies will also send out letters to potential offenders to clarify their situation and alert them to the consequences. On the other hand, the arrangements and the absence of central fraud reporting mean that internal audit does not have an overall view of the control of and responses to fraud.
Only one of the three councils established a corporate presence, and that was only possible because of retained staff and resources which provided an existing capacity but at the outset did not require additional council funding. Council A had lost 12 of 16 staff to DWP but was initially able to use housing benefit grant subsidy balances to retain four investigators in a corporate fraud team with a manager and administrative support; this is now funded directly from the council's general revenue. The team was given responsibility to consolidate existing fraud work, including the installation of centralised case management and data-warehousing and data-matching software, supplemented by information from staff following internal training and network management, the audit plan, a review of all customer complaints, NFI referrals, disciplinary issues, etc. All potential fraud work is referred through the corporate fraud team unless the investigation is suitable to be undertaken by departmental managers. It undertakes some 800 cases annually, resulting in actual savings of nearly £800,000 (up from £200,000 in 2016), nearly three times the cost of the team. 6 In terms of internal arrangements, fraud is primarily seen as an audit compliance function, and is usually addressed as part of the reporting arrangements on audit and risk to the council audit committee. Fraud and corruption policies are compliance-focused; council fraud strategies in terms of organisational culture and wider governance arrangements do not appear to be developed at the council level. In terms of collaboration and relations with other organisations, anti-fraud work for all three councils is supported by regular meetings with the DWP, because they share a common 'fraud stock'-benefits and council tax fraud-while Council A's position has allowed it to lead on a social tenancy group among neighbouring councils. Using its software, the council has proposed a data-sharing hub involving four adjacent councils, including Council B, and local housing associations. There is limited contact with the police outside operational matters, although Council A, as a consequence of the DCLG funding, locates a member of staff for half a day once a week in the police headquarters under a formal protocol to share data on areas of mutual concern. There has been little contact with the DHCLG and the Home Office, and none with the NCA EEC. Support for anti-fraud work from the current external audit regime for the three councils depends on the specific external audit person, although none of the current contracts with the three councils concerned include days for fraud (with one council noting that major fraud reported to the external auditor for an assurance review of council control arrangements could involve the external auditor seeking additional funding from the PSAA, which would be charged back as an additional cost to the council and would potentially be a disincentive to investigate or report on fraud).

Component Three: Determining the Impact
The case study research indicates that none of the interviewed councils have implemented the full intent and expectations of the 2011 FFL strategy, with only one able to develop a number of its recommendations because of available internal resource and management priorities.
However, it was clear that fraud remains a technical issue to be resolved in many councils by a particular department; wider governance arrangements, external collaboration, and other priorities have diluted or constrained effective implementation. The implications for national oversight and assessment of impact remain unexamined, however, because the NFA and Audit Commission, as owners of, or responsible for, the FFL strategy, were not replaced by any body with the means to review and revise the FFL strategy in the light of practice on the ground, or to evaluate implementation or impact. 7 Given the top-down nature of the strategy, despite local government input into its drafting, and its brief ownership by the NFA, the abolition of the NFA also removed the means for ensuring alignment across strategy, process, structure, and environment. Additionally, it ended its role as a locus of responsibility for ensuring that the aims of the strategy were maintained in the face of changing circumstances and other policy agendas that may facilitate or inhibit the choices of those involved in effective implementation. The research suggests that none of the purposes of the FFL strategy-a more efficient fraud response mechanism, radically realigned fraud strategies, governance arrangements driving the delivery of a fraud strategy across local government, and collaborative arrangements with the police and central government departments-have been implemented. It therefore remains hypothetical what effects the strategy has had, or would have had, on actual levels of fraud.
Indeed, many of the issues adversely affecting implementation, and underlining the consequences of not adopting an emergent approach to the strategy implementation process, were revealed when the 2016 strategy was being redrafted by the advisory FFCL Oversight Board in conjunction with CIPFA. The preparation received, for the first time, substantial feedback on implementation and problems identified with compliance with the intent of the original strategy. These, reflecting the countervailing influences of other agendas, included local authorities reporting 'that they were still encountering barriers to tackling fraud effectively, including incentives, information sharing and powers', and that capacity to tackle fraud and corruption was likely to be reduced or had already been reduced as a result of austerity-related funding reductions. Staff and expertise had been lost and were unlikely to be recovered, and there were limited budgets for training new staff. Furthermore, reflecting organisational priorities, 'senior managers were finding it difficult to dedicate sufficient time to demonstrate their support for counter fraud activities', and counter-fraud work had a low profile and its benefits were not 'fully appreciated' (FFCL Oversight Board/CIPFA 2016, pp. 6, 15).
In keeping with the top-down nature of the original FFL strategy, and despite such views from the ground, the revised strategy proposed that councils not only would continue the 2011 themes, but would also address questions of leadership and risk assessment, as well as using resources, collaboration, and technology more effectively. It also expected councils to respond to further recommendations reflecting two other national agendas-the Serious and Organised Crime Strategy and the first UK Anti-Corruption Plan. The strategy also proposed that progress was to be monitored and an annual report issued 'setting out what has been achieved and what remains to be done, so that local authorities and other stakeholders have clear visibility of how the strategy has improved outcomes' (FFCL Oversight Board/CIPFA 2016, 26); this has not been delivered to date. In 2018, as its commitment to and capacity for antifraud work declined, CIPFA handed responsibility for the local government strategy and support of the FFCL Oversight Board to Cifas, a not-for-profit fraud prevention membership organisation, whose members are primarily financial services and corporate businesses, though it has broadened its base in recent years.

Analysis: Lessons for Managing the Strategy Implementation Process
The research allows the three components to be further addressed in order to establish the primary factors influencing the strategy implementation process. On the first component, the proposal for a strategy on the basis of identified policy concerns with appropriate institutional ownership was addressed and then implemented. On the other hand, the development of both the national and local strategies very much reflected national government's approach to both policy and policy delivery via a strategy. This saw a relatively structured top-down strategy drafted for local implementation. The 2006 Fraud Review was essentially a policy exercise as 'a matter of technical administration'-'correct policies to be found by expert analysis' (Burch and Wood 1990, pp. 40, 41). This reflected a top-down approach: an 'underlying view that policy formulation and delivery are separate and distinct activities' and that, after policymakers 'have control over creating the policy', it is 'transmitted for others to execute faithfully' (Hallsworth and Rutter 2011, p. 17). In this case, strategic management of the implementation process became the responsibility of a central agency, the NFA, which saw its role to orchestrate and sustain the government's response to fraud, launching 'programmes to identify the strategic challenges in key areas of the fraud problem and build joined-up programmes to tackle these' (National Fraud Strategic Authority 2008, p. 14), with regular programmes of activity, identifiable outputs, and updates. Insofar as the programme implicitly recognised an incremental strategy, then that was seen as part of the role of the NFA.
In terms of the second component, the NFA took ownership of the implementation of both national and local strategies, where the latter would have been supplemented by the monitoring and information-gathering roles performed by the Audit Commission. The abolition of the NFA, along with the Audit Commission which could have provided ownership, monitoring, and evaluation at the local level, ensured that there was no oversight (or overview) of the implementation process, no means for identifying resource limitations, promoting inter-agency relations, receiving information to review or revise the strategy during implementation, or evaluating the impact of the strategy. As a deliberate or top-down strategy, the absence of an owner, and one capable of moving to an emergent approach, meant the strategy continued unchanged in the face of significant changing circumstances which dictated which parts were seen as relevant or could be resourced at a local level. In terms of the third component, the abolition of the strategy owner would have the greatest significance on impact; as the HMICFRS reported in 2019, the closure of the NFA and distribution of roles and responsibilities meant that 'the notion of a single national fraud strategy such as Fighting Fraud Together was largely forgotten' (HMICFRS 2019, pp. 30-31).
What is interesting about such findings is that they identify two related issues that are reflected generically at the local level and in other areas of the public sector. These are that strategies have to be evolving or adaptable rather than purely deliberate and top-down, and that the strategic implementation process needs to be 'owned' by a body with the authority or the means to ensure appropriate guidance, resources, and support, as well as feedback to review and adapt the strategy to changing circumstances. We have been able to validate our findings from similar responses to policy or strategy implementation in relation to fraud and related issues in areas other than local government, and at both the national and local levels.
Almost 12 months after the 2015 national police fraud strategy was announced, the policing review body, Her Majesty's Inspectorate of Constabulary (HMIC), was reporting that in nearly all forces there was 'an absence of strategic leadership', less than half of forces regularly considered 'the impact of fraud in their strategic risk assessments' and nearly two-thirds of forces failed to provide a nominated senior officer to oversee their force's approach to fraud (HMIC 2015, p. 67). In 2019 the police inspectorate (now Her Majesty's Inspectorate of Constabulary and Fire & Rescue Services [HMICFRS]) reported that, in relation to the strategy, the 'document was intended to provide guidance and support to police forces. However, [they] did not find any evidence of general awareness of the document, or that it had ever been formally adopted by police forces' (HMICFRS 2019, pp. 30-31). For the three police forces in the North East, strategy proposals to develop partnerships, shared information, and coordinated or cooperative approaches have not been achieved in any meaningful way. This is not because of a reluctance to do so, but because of an absence of central or national support, the absence of resources, and the priority given to other local policing objectives (Doig 2018b).
At the national level, where there has been less institutional upheaval, fewer difficulties arising from the London-centrism of government perspectives, and a lesser influence of competing agendas, the implementation process of the 2008 national fraud strategy has not functioned well. As noted above, the Cabinet Office had become the policy lead for fraud for central government (excepting the Home Office Joint Fraud Taskforce, now in its second phase), but that did not include responsibility for the 2008 national strategy which, as the HMICFRS noted later in relation to the police, has also been largely forgotten at the national level. Thus, 8 years after the publication of the 2008 national strategy, the NAO reported that levels of fraud against central government were unknown, that in the absence of incentives or resources there was poor reporting, with gaps and inconsistencies in information collection, and that departments were relied on to manage fraud but had mixed capacity and capability 'to understand and address fraud risks'. There was an absence of 'mechanisms for holding departments to account for performance' and there were limited means to evaluate 'success'. In particular, it noted that the Cabinet Office, which also had no mandated resources or executive authority, did 'not have strong levers to direct actions as even "mandates" have to be negotiated and agreed' and that 'in the absence of firm levers, it has tried to improve aspects of governmental capability through influence, and promote collaboration between departments' (National Audit Office 2016, pp. 6, 37, 35).
What does this say about strategy implementation? It may be argued that the 2011 FFL strategy has failed to achieve much so far. It could be described as a 'precarious success' which has delivered 'on the edge of failure. Policies do exhibit small achievements, but departures from goals and levels of opposition outweigh small levels of support. They often amount to a pyrrhic victory for policymakers. Initially government does fulfil some of its policy making goals, but the costs of doing so become such that short-term success cannot be sustained' (McConnell 2010, p. 355). A practitioner review of subsequent developments asked whether 'a list of initiatives-with no explicit mechanisms to foster cooperation or to monitor and measure progress and outcomes-ever add up to any kind of strategy…without the continuity of purpose and stability of institutions and budgets to build real capacity and resilience over the long term' (FAP 2016, p. 8). Reflecting the findings from the research, given the cumulative impact of various national and local agendas and reorganisations, one identifiable variable is the degree of ownership of the strategic implementation process. Central government has 'a residual function of overseeing how a policy is being realised… A policy is not just made and then executed; it is made and constantly re-made by multiple players throughout the system. Central government has a continuing role in this re-making; the alternative to 'top-down' policy making is not simply relinquishing all responsibility for public outcomes. Too much faith in self-organisation too soon may lead to problems: there is still a role for an overarching perspective and a capacity for steering… The point of system stewardship is that, when choosing an intervention (whatever it may be), policy makers should be thinking in terms of overseeing an overall system, rather than in terms of launching another stand-alone initiative that tries to ignore or supplant all its predecessors' (Hallsworth 2011, pp. 28-29).
This article proposes that understanding the strategy implementation process has been a neglected but crucial component of judging the effective implementation of any national government strategy: 'the translation of policy intent into action is highly complex, involving many people inside and outside government and requiring activity across many fronts. Policymakers often have their eyes on a moving target and their actions can be thrown off course by external events or by the limits of central government control' (Norris et al. 2014, p. 23). This is not an issue unique to the UK, or to fraud; existing research emphasises not only that 'one important issue to emerge from this stream of work on strategic management is that strategy implementation matters' (Walker 2013, p. 683), but also that 'more research is needed to generate a comprehensive understanding of strategic management in the public sector' , p. 539). Friström et al. (2009 echo this lesson when they show how different interpretations of welfare benefit fraud among Swedish agencies in different areas led to an unmanageable increase in reports to prosecutors that failed to include the necessary legal constituents for crimes to be proven. Such policy shifts always need to be managed and communicated to different actors and refreshed over time if they are to lead to implementation as intended. We would argue that it is important and valuable to study this 'lost-in-translation' component of anti-fraud initiatives, and that the themes of this research should be tested in many other anti-fraud measures at an EU level (e.g. OLAF and the European Public Prosecutor's Office) and at a national level elsewhere.

Conclusion
The interplay between national strategies and initiatives, local level priorities, an emphasis on budget reductions, and an absence of oversight and ownership remains a continuing dynamic in the UK in relation to fraud. Both the rhetoric and the recommendations of the 2011 FFL strategy may still be valuable and may be practised in some parts of England, but the research has detected little evidence regionally of the concerted and collaborative approach, sufficient conditions for comprehensive and continuing delivery, and of ownership to guide and monitor the process. Formal ownership alone does not guarantee implementation if the levers of influence and power are weak, but the absence of ownership has had the unintended effect of removing oversight of, and action against, clashes between reforms and other agendas at central/local levels. It has also meant continuing mismatches between the potential for fraud at the local level and the institutional arrangements or resources to oversee, support, or guide them. Despite what we regard as the best intentions of the fraud strategies and other initiatives, whether individually or collectively, the resources, priorities, and institutional arrangements on the ground are in the short-term unlikely to deliver them.
Governance is a complicated business, and formal devolution in Northern Ireland, Scotland, and Wales has made it more complicated. Yet the UK government remains very Londoncentric, and this applies also to its relationship with English regions. With the political pressures for constant action, the effort required for implementation-including creating and sustaining its legitimacy-may be intermittent or not even contemplated, even where, as in this case, there may be a genuine desire to see the strategy adopted. This inattention is more likely to happen if politicians and civil servants do not regard it as a key priority. Both the localism agenda and financial stringency/austerity as well as the control environment in local authorities have affected and will continue to affect the fraud dynamic. The level of resource, the institutional configurations, the procedures for inter-agency working, the roles and locations of external support, supervision, and monitoring, are still competing for resolution. Culture, prevention, and organisations' internal investigative capacity remain contested spaces. Local policing in the UK and in many other countries is focused on issues other than fraud against government (or indeed any kind of fraud, except perhaps where it concerns 'organised crime' or bears strongly on a loosely defined category of 'vulnerable individuals'), while the modest resources of central fraud policing are focused on vulnerable individuals and forms of fraud that evoke more popular anxiety and sympathy. Although there is ad hoc partnership working, joint working and information sharing among councils, and between councils and national government agencies such as the DWP, there is an identifiable dis-connectedness-even dysfunction-at a strategic level. This has led to a failure to ensure that the aims achieve all the intended effective and sustainable outcomes, or even outputs. We hope that the UK government's Economic Crime Strategy Board and Economic Crime Delivery Board will take note, though the UK's public-private sector Economic Crime Plan (HM Government and UK Finance 2019) is of a very different order of ambition and complexity from the local fraud against government examined here.
It is not rare that government, inter-government, or non-government strategies result in implementation failure in any country. It is important to be realistic and to see such implementation as a matter of degree rather than a binary total success or total failure. We are not arguing that local government fraud reduction is a key policy goal. Our key observation is that the implementation process needs serious management-which requires both time and cultural awareness-and is very far from being automatic. Implementation failure is not the only source of policy failure: the policy may be mistaken (theory failure) and/or there may be difficulties in demonstrating impact objectively (measurement failure). Some policies may only be intended to have political, symbolic effects (though this is very rarely declared), and some may be based on folk theories that persist despite the absence of evidence of impact on the phenomena they are supposed to reduce (Halliday 2018). 8 However, especially at a time of fiscal crisis, there are good reasons for states to seek to reduce fraud against government, and there are no obvious improper or political bias reasons to inhibit more energetic responses (though there might be some countries where corrupt individual or party finance reasons might generate official 'tolerance' of tax or procurement fraud, or other elite malefactions). The evidence regarding anti-fraud implementation does not exist in most European systems outside of Scandinavia, but different governmental and bureaucratic traditions may produce different responses from those observed here in the highly London-centric UK government. However, the challenges of putting strategies into effect are relevant to any country seeking a 'strategic' approach to fraud that requires leverage at a local level. As the poet T.S. Eliot might have put it, between the policy and its implementation lies the shadow. It is the task of competent government and research to shine a light on shady places if the welfare objectives of policy are to be attained; if not, observers are entitled to conclude that reform efforts have not been serious.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.