Janus‐Faced Youth Justice Work and the Transformation of Accountability

This article revisits claims about the relationship between &#8216;standardisation&#8217;, &#8216;discretion&#8217; and &#8216;accountability&#8217; in youth justice made in the wake of the Crime and Disorder Act 1998. We argue that less centralisation and less standardisation have transformed accountability, but this is experienced differently according to the place held in the organisational hierarchy. This recognition demands a more nuanced understanding of &#8216;practitioner discretion&#8217;, which can account for differences between managerial and frontline experiences of what we describe as &#8216;janus&#8208;faced youth justice work&#8217;, and a broad definition of the youth justice field and associated actors.

The Howard Journal Vol 55 No 4. December 2016ISSN 2059-1098 'standardisation', 'discretion' and 'accountability' in youth justice made in the wake of the CDA 1998. We argue that changing relationships between central and local government highlight the importance of local bureaucratic hierarchies and occupational cultures in shaping everyday decision making and the prioritisation of tasks -where work remains governed by National Standards (Youth Justice Board 2013), but also in new, less regulated areas of work. This recognition demands a more nuanced understanding of 'practitioner discretion' which can account for differences between 'managerial' and 'frontline' experiences, and a broader definition of the youth justice field and associated actors than recent youth justice scholarship has commonly suggested.

Methodology
This article is based on findings from a study, funded by the Economic and Social Research Council, which explored the implementation of youth justice policy in England through an examination of practitioner sensemaking. The research comprised a study of youth offending services at two case study sites, with the primary data collection undertaken between December 2012 and September 2013, with a variety of youth justice practitioners. The researchers undertook focus groups with practitioners at each site, involving managers, officers and workers (n = 44). The main body of the data collection comprised in-depth qualitative interviews with 71 practitioners (Site A = 31, Site B = 40) ranging from the Heads of Service at each site through to officers, preventions workers and performance managers. The interviews averaged one hour and three people were interviewed twice. We have broadly indicated participants' roles below using job titles and descriptions of the role given in the interviews. For the sake of simplicity, we used the most appropriate of three categories: practitioner, manager, and senior manager. We have indicated the site except where only few participants held identified roles and the argument is not affected.
The sites were originally identified because they were relatively autonomous 'teams' with histories of challenging YJB policies. In the interests of confidentiality the sites will hereafter be referred to as 'Site A' and 'Site B'. The organisation at Site A took a form based on that established by the CDA 1998: a multi-disciplinary team made up of social workers, probation officers, police officers and specialists in other areas, including health and education. The service was divided into pre-and post-court teams, the latter of which dealt with statutory orders while the former dealt with pre-court orders and wider prevention programmes/schemes. In Site B the local authority had subsumed youth justice services into a broader service responsible for young people's offending, employment, and welfare issues more widely. The 'integrated' set-up in Site B meant that youth justice work was largely undertaken by staff trained in youth justice but some of the lower-level youth offending and preventions work was undertaken by practitioners from a range of professional backgrounds, such as youth work or Connexions.

A New 'New Youth Justice'
The 'new youth justice' (Goldson 2000) introduced in England and Wales by the CDA 1998 can be located in a longer tradition of NPM in the public services, in which the founding of the Audit Commission and National Audit Office in 1983 are significant markers. A focus on results and the centralised auditing of 'effectiveness' (in addition to the already established work of the inspectorates) meant that the number of people employed in managerial roles grew, and their reliance on data generated by those working directly with clients to make and justify spending decisions changed existing frontline roles by restricting discretion and producing new demands for information (Clarke and Newman 1997). Indeed, limiting the ability of practitioners to make decisions that might threaten the 'efficient' deployment of resources can be understood as reflective of a move to an approach more concerned with the allocation of resources and system efficacy than meeting needs (Kemshall et al. 1997), and as part of a 'fundamental assault' on professional cultures (McLaughlin, Muncie and Hughes 2001, p.303).
The New Labour government elected in 1997 had distanced themselves from Conservative public sector marketisation and their rhetorical emphasis on 'more' law and order, emphasising instead communitarian values and greater attention to the 'causes of crime' (McLaughlin, Muncie and Hughes 2001). However, the commitment to an 'audit culture' remained embedded in their political strategies, albeit in a hybridised 'neobureaucratic' form which combined rule-adherence with the outsourcing of the regulatory functions performed by hierarchical management in traditional bureaucracies (Harrison and Smith 2003). Of central importance to promoting 'effective' and 'consistent' practice was the newly created YJB, a quasi-governmental organisation authorised to draft national standards for youth justice work; approve annual youth justice plans; and, with the Audit Commission and professional inspectorates, monitor the performance of multi-agency youth offending teams (YOTs) (Souhami 2007). While National Standards for processes and procedures that must be followed when 'doing' probation work had been developed from the 1990s, they became more directive following the CDA 1998 and the development of separate standards for workers in the new multi-agency YOTs (Youth Justice Board 2000). As Eadie and Canton (2002) observed: '[s]ignificantly, the language of the Standards for Youth Justice has changed from "should" to "must"' (p.16).
Standardised risk assessment tools, Asset and Onset, were introduced from 2003 (see Baker (2005) for discussion and Case and Haines (2015) for a summary of critiques). The YJB also held funds for evaluative research, and the 'Key Elements of Effective Practice' documents derived from YJB-commissioned research and research reviews formed the basis of new National Qualifications Framework, centring on a Professional Certificate in Effective Practice developed in partnership with the YJB (see Hester 2010). While some argued that there is 'no inherent contradiction between good, consistent service delivery and wide (though not unbounded) discretion' (Eadie and Canton 2002, p.14), post-CDA 1998 attempts to standardise youth justice processes and delineate the youth justice knowledge base were seen by others as an overtly political act and an assault on professional independence. Pitts (2001), for example, represents the 'pursuit of homogeneity' within youth justice as part of a broader 'quest for congruence' between a populist criminal justice agenda and potentially oppositional practice cultures, which amounted to the 'zombification' of youth justice.
However, as Lipsky (1980), in his classic study of street-level bureaucrats argued, 'discretion' can be understood as a necessary element of frontline roles in public services, since even the most robust rule structure requires actors to choose a course of action and choose between rules. In other words care should be taken 'not to confuse the presence of rules with determinacy' (Evans and Harris 2004, p.891). Rules may restrict the permitted options and provide criteria to guide choice, but they do not preclude illegitimate action or ensure choices made reflect those intended by designers. Eadie and Canton's (2002) model engages with this issue by considering 'discretion' alongside the concept of 'accountability'. They propose four practice 'quadrants' with high/low discretion and high/low accountability in each. They suggest that the CDA 1998 and the 2000 National Standards (Youth Justice Board 2000) introduced higher accountability which could be consistent with 'best practice', but only if high accountability is coupled with high discretion. They warn, however, that 'rigid application of the rules' and 'increased standardisation' is the 'wrong strategy' leading to 'constrained practice' (Eadie and Canton 2002, p.24).
Clearly, assessments of the desirability of standardisation of practice can be analysed on their own terms, and the youth justice practice literature has explored how, when, and why, practitioners choose to subvert or 'resist' National Standards, tools and processes (for example , Baker 2005;Canton and Eadie 2008;Hughes 2009;Bateman 2011). Our concern is rather different. The introduction of multi-agency YOTs by the CDA 1998 encouraged a number of empirical studies at the organisational level (for example, Burnett and Appleton 2004;Souhami 2007;Field 2007), but as Phoenix (2016) discusses, recently dominant approaches within youth criminology have tended to flatten 'the complexity (and the specificity) of the social relations that make up the youth penal realm' (p.136). Recent dramatic reductions in the numbers of young people receiving a reprimand, final warning or conviction for the first time ('first time entrants'), numbers of young people sentenced in court and numbers of young people in custody, and a political context in which aspects of the current approach to youth justice work are being questioned (Carlile 2014;Taylor 2016) have again prompted new interest in the local functioning of youth justice organisations (for example, Drake, Fergusson and Briggs 2014;Smith 2014;Byrne and Brooks 2015;Morris 2015). In what follows, we demonstrate how decoupling the issues of: (i) 'standardisation' and 'centralisation'; and (ii) 'managerial' and 'practitioner' discretion, enables analysis of how changing relationships between the central and local bureaucracy in youth justice have impacted on power relations and perceptions of accountability within local organisations.

Decentralisation and (Managerial) Discretion
Since the election of the Conservative-Liberal Democrat Coalition government in 2010, the need to promote more localised responses has been emphasised within policy statements (for example, Ministry of Justice 2012). There have been some significant changes to the local criminal justice landscape, such as the introduction of elected Police and Crime Commissioners (House of Commons 2016), a departure from national-level monitoring reflected by the dismantling of the Audit Commission (HM Government 2014) and a failed attempt to dismantle the YJB and move its functions to the Ministry of Justice (see Souhami 2015). While the YJB continues to centrally administer the youth justice system, there is a greater emphasis on practitioner discretion within what the YJB (Youth Justice Board 2011) has recently described as a changing local delivery landscape, 'with more local accountability and decision-making' (p.5).
Souhami (2015) considered how the central bureaucratic organisation, the YJB, functioned and interacted with local bureaucracies as part of her broad project. She describes how the YJB, with comparatively few direct mechanisms for influencing local YOTs, extended their informal influence (firmly framed as 'support') via YOT-appointed 'regional monitors', tasked with 'validating' extensive case-level data collected with the aim of improving practice, and 'practice improvement consultants', with the remit to act directly on that knowledge to shape local practices and improve outcomes (Souhami 2015, p.162). She argues that this went considerably beyond the aggregate-level performance monitoring that the YJB needed to advise the Home Secretary and fulfil its statutory function, resulting instead in 'the extension of central surveillance and intervention' and the 'enlisting [of] YOTs and local authorities into their own self-regulation' (Souhami 2015, p.162).
Despite this, senior managers at both our research sites claimed to be innovating prior to any official departure from a centralised youth justice system following the election of the Coalition government: I think the Youth Justice Board would like to think that they were telling us what to do but actually they threw an awful lot of money at Youth Justice Services across the country . . . . To be fair to them they would try and identify good practice and work up a kind of template for others to follow. (Senior Manager) As Souhami (2015, p.161) acknowledges, the YJB allocated a percentage of YOT funding (about 20% in 2006) but was disinclined to act on the threat to withdraw this, given the likely impact on services. Our research suggests that both sites felt able to act in counter-hegemonic ways by remaining within what Lipsky (1980) describes as the 'rule structure' of the central bureaucracy: improved outcomes from counter-hegemonic approaches still satisfy relevant performance indicators. For example, the introduction of pre-court interventions at both the studied sites reflects a national swing towards increased diversion from court, a change that had been encouraged by revised performance indicators and established before there was a legislative framework in place (Bateman 2014;Smith 2014). 4 Our data similarly suggest that the policy emphasis on decentralised decision making appears to be reflecting and further enabling, rather than driving, change, although we were told that the change of government in 2010 made the exercise of local discretion 'much more acceptable' to the YJB.
The importance of agentic senior managers to local youth justice is also illuminated by responses to a question we commonly asked participants from all levels of both services -whether they believed there was a coherent vision driving recent developments in their service. Responses were varied across sites and between sites, but senior managers were often believed to have been responsible for significant changes. Some themes stood out. First, as noted, a desire to avoid 'unnecessary' criminalisation and, instead, provide 'supportive' pre-court interventions, was commonly identified as an overarching rationale for service provision at both sites (see Kelly and Armitage (2015) for discussion). Second, in addition to these new, locallydeveloped interventions, senior management teams had recently altered service structures (radically, in the case of the integrated service) and introduced wholly new areas of work. 5 Practitioner experiences at Site A and Site B were very different, but changes at both sites were commonly understood to relate to the specific ways in which local senior management teams adapted to the significant budget cuts that had affected all local authority services following the financial crash of 2008 and the so-called 'austerity measures' introduced by the Coalition government (for example, Innes and Tetlow 2015): I think it [the move to an integrated service] was primarily money driven and then maybe secondary driven by better outcomes for young people. (Practitioner, Site B) When I started there was three times as many people sat upstairs there really was. And people have left due to retirement and you know illness and things like that and then the Council obviously tried to make cutbacks so people took early retirement. (Practitioner, Site A) Nobody can say why it is we're not advertising the job until October other than to potentially save money. (Practitioner, Site A) Our data, therefore, suggest that 'managerial discretion' is not only encouraged in the current political context, but seems to have been more discretely in operation prior to the 'official' push towards greater decentralisation by the YJB. This appears to reflect studies which have discussed tendencies towards the local restructuring of youth offending services (for example , Fielder, Hart and Shaw 2008;Youth Justice Board 2015) and more recent discussions about the importance of occupational cultures in shaping the interpretation of policy (Byrne and Brooks 2015;Field 2015;Morris 2015;Souhami 2015). However, the assumption that increased managerial discretion means increased (frontline) practitioner discretion, or that managerial discretion is unconstrained under decentralised conditions both require greater scrutiny.

Embedded Managerialism
Together, strategic decisions to retrench, reorganise and develop nonstatutory pre-court interventions have brought considerable changes to the front line. Fewer young people processed by the courts means less statutory youth justice work, although those in the remaining cohort have more complex needs (see also Carlile 2014), while new pre-court interventions and new areas of work with children and young people not in trouble with the law require associated practitioners to become familiar with new assessments and other forms of 'paperwork'. Practitioner experiences differed at the research sites. At Site A, practitioners were divided into 'post-court' and 'pre-court' teams, but some workers in the new integrated service at Site B managed both informal, non-statutory interventions and court orders (see also Kelly and Armitage 2015). In addition to these changes however, we were also told of considerable continuities, where job titles or routes into YOT services might have changed, but much of the work undertaken remained very similar.
We came to understand contemporary youth justice work at the front line as 'janus-faced', involving 'looking up' to managers, but also 'looking across' to the children and families who made up a case manager's or worker's caseload. These empirical observations are not new: the difficulties associated with balancing 'relational' and 'instrumental' aspects of professional practice have been explored in relation to youth justice (for example, Prior and Mason 2010;Drake, Fergusson and Briggs 2014;Morris 2015) and social and probation work carried out in non-youth justice settings (for example, Broadhurst et al. 2010;Mawby and Worrall 2013). We were, nonetheless, surprised to learn that the majority of our interview participants claimed that they spent between 50% and 80% of their time doing 'paperwork', with time available for direct work with children and young people subsequently under pressure: Careworks is the base rate so everything revolves round it. So all the reports, all the assessments, every telephone call we make has to be logged on there. So telephone calls, emails, every conversation, every meeting with a young person, every meeting with, you know, their social worker, everything is recorded on Careworks . . . . When I first started . . . Careworks didn't exist and now it is one of the over-riding things that we spend most of our time looking at. (Practitioner, Site B) I think ten years ago we had more time to spend with young people and families. I think now with the increase of more paperwork, the lack of staff and the lack of resources, changes in legislation which then introduced more paperwork with more government incentives that you've got to do more for. I would say now that 70%, at least a minimum of 70%, is paperwork and data inputting where before at least there was a good balance. (Practitioner, Site A) While some accepted this aspect of their working lives, others expressed real frustration or a sense of weariness. Often this was because practitioners felt there was an inverse relationship between time available for young people and time spent on paperwork: I prefer to be out, you know, I'd like to go do an assessment, I'd like to spend three hours with a young person but I think 'I've spent three hours, it's going to take six hours, seven hours to put that on the system', even just if you met them down the town or something and got talking but you have to input it. (Practitioner, Site A) While a number of workers found aspects of the 'paperwork' (particularly assessments) supported their work with young people, and a few workers believed that the recording systems could be usefully used to 'self-audit' their work, a sense of frustration was also bred of a view that 'the paperwork' was more about monitoring them as individual practitioners rather than their 'practice' and was based on distrust. While a number of workers identified work with young people as the 'real work' and as more important, they felt the service (and for some, themselves as workers) were held to account for the other aspect of their role: We're constantly told if it's not written down it didn't happen, well in my opinion it did happen -I was there. I've done something meaningful . . . just because I haven't recorded it in a specific place does that mean it's any less valuable? You can get practitioners who spend 90% of the time in front of a computer and haven't got the time or the patience or the energy to be investing in the young person to make the changes. And on paper they look fantastic, but in reality does the paper reflect the reality of a situation? (Practitioner, Site A) I'm not into it for ticking the boxes. I know I've got to do it but my prime focus with every case, and I will justify why certain boxes aren't ticked, is to work with young people. (Practitioner, Site B)

Individualised Perceptions of Liability
Analysing the full range of responses about the usefulness of 'paperwork' (especially when this overly-broad term was disaggregated into discrete tasks) 6 and its connections to the 'face-to-face work' is too complex for this article. In many ways, our research reflects earlier evaluations of standardised assessment tools (for example, Baker 2005; Wilson and Hinks 2011) in that it shows practitioners report using these in different ways -from strict adherence to effective dismissal of available guidance. However, while there was some flexibility in the way tools were used, there was little space to refuse to use them at all. Even practitioners who felt 'paperwork' made little difference to their work with young people tended to acknowledged the importance of an 'audit trail', and few overall could imagine a situation in which record-keeping was reduced. In a phrase we heard particularly at Site A -'if it isn't recorded, it didn't happen'.
We were told that records were important to allow the transfer of cases, for example in the case of staff illness. Some of those interviewed, however, saw the 'paperwork' (and particularly the recording of 'contacts') primarily as a means of protecting themselves if something were to 'go wrong': It's a back covering exercise. I don't know how helpful it is to the young person . . . . It's about what'll happen to me because I haven't recorded it or I haven't told that person. It's not really because that person will then get a beating from their boyfriend although it should be. It's about 'well it says here in your risk management plan that you would do that regardless of whether someone's been beaten up', but that will be used obviously. It'll be like 'well this is what's happened but you failed to manage your plan, you failed to do this, you failed to do that'. (Practitioner, Site A) However, as this manager acknowledged, some incidents may not be preventable, especially if frontline workers struggle to recognise or respond quickly to unexpected crises: The red flags are the things that you can predict. So there are some things that you'll never be able to predict, it'll just happen . . . I mean I think a lot of the time we're very good at assessing risk but sometimes it's a question of certain sets of circumstances which happen all at once and it just pushes that person over the edge, whatever that means. (Manager, Site B) A more generalised professional anxiety was also visible in the integrated service where staff had been required to take on youth justice work with limited training (and sometimes amongst those responsible for 'gatekeeping' their work): I'm constantly on edge to be 'have I done it all right?' 'cause I don't want to be the one that gets kicked out because I've done it wrong. (Practitioner, Site B) More experienced practitioners also shared concerns that some of the peer support mechanisms which had previously facilitated 'on-the-job' training and supplemented supervision of challenging cases had been eroded by the rapidly-changing organisational contexts discussed above: Now that the team is significantly smaller . . . sometimes there's maybe one other person in the office so there's not as many opportunities to get a cross section of opinions. I think everybody is still really supportive if somebody's struggling.

(Practitioner, Site A)
Gatekeeping [that is, checking the quality of 'paperwork' and offering feedback] that is anything but face to face is really, really difficult, and I have learnt to stamp my feet a bit . . . . Sometimes it's about me moving and sometimes it's about the other person moving because, yeah, I can give you feedback on the phone but then we're talking about which line of the report and anyway that's not really what, it's not about proof-reading, it needs to be a discussion. So I think there's been practical difficulties [when we are] no longer in the same building and some concerns about how you maintain the level of expertise as time goes on. (Manager, Site B) These changes were not absolute: we were told of continuing peer support, successful supervision by senior practitioners and opportunities for group discussion of cases at both sites, as well as developments such as 'practice improvement' posts intended to identify potential issues and support frontline workers. Nonetheless, the central importance of collaborative and supportive peer networks for initial training and ongoing development was widely emphasised by staff at all levels, which parallel systems of quality assurance -and particularly depersonalised forms of 'gatekeeping' -could not adequately replace.

Destandardisation and (Practitioner) Discretion
Recent changes to the National Standards for Youth Justice Services (Youth Justice Board 2013) suggest a partial relaxing of standardisation (see Drake, Fergusson and Briggs (2014) for discussion). Increased flexibility is reflected, for example, in changes to requirements relating to compliance and breach. Section 8.15 of the most recent revision requires that 'any decisions not to refer the matter to court or panel under breach proceedings, where this is otherwise warranted by a pattern of non-compliance, are approved by a manager and properly recorded' (Youth Justice Board 2013, p.30). This contrasts with the more rigid 2010 standards which outlined clear rules and timescales for when practitioners must initiate breach proceedings (Youth Justice Board 2010, pp.64-5) and stated in Section 8.7, that YOT managers should stay breach only 'in exceptional circumstances', with a requirement to 'record this fully with justifying reasons on the child or young person's file' (Youth Justice Board 2010, p.62). As with pre-court interventions, this appears to formalise existing decisions: frontline practitioners at both sites gave examples of where they had sought exceptions to breach proceedings (as below), and one site had revised their local policies. Greater tolerance for variation is also reflected in the area of risk assessment: controversial scoring has also been removed from the new AssetPlus tool and practitioners will be able to use their 'professional judgement to contribute to the final likelihood of reoffending rating' (Youth Justice Board 2014, p.14).
What remain in place are most of the deadlines for the completion of tasks set out within National Standards. This is significant given many frontline practitioners told us about the range of performance management measures in place to secure compliance with these deadlines and to track progress against other locally determined performance indicators. Careworks automatically flags when work governed by National Standards is overdue (for example, initial meetings are required to take place within three days of a court hearing when a court order is made, and plans resulting from assessments must be completed within 15 days of the initiation of an assessment), and managers were able to track the progress of individual staff members against these prescribed deadlines. Managers at both sites also 'audited' samples of cases to check that all required documentation was adequately completed in a timely manner.
There was considerable pressure to meet deadlines and follow procedures governed by National Standards, although it was apparent that there was some space for discretion here too. Sometimes this was due to individual prioritisation of tasks: one of the practitioners we interviewed, for example, described how it 'boiled her piss' to see colleagues take young people engaged in education out of the classroom during school hours in order to meet the deadlines prescribed by National Standards, and a number of staff told us that they refused to prioritise deadlines. Most commonly, however, this followed an application to a manager to 'authorise' missed deadlines for the completion of 'paperwork' or a decision not to comply with National Standards in other areas, such as compliance and breach: If there's an issue where we need to discuss that because we need to veer away from National Standards, or we need to make a decision on lack of engagement or breach, then I'll have those conversations with the case manager, we'll reach a decision which is mutual I think, based on what they're saying, based on what I think ought to happen, but that's balanced out against any perceived risk. So if there's a risk to anybody then it's going to influence our decision. (Manager, Site A) If it's a statutory appointment and they don't attend, yes, that's what happens: you send the warning. But we always write in the warning 'if you've got a valid reason for that absence you need to provide evidence within seven days' and then there would be that option of discussing it with the manager, do we think that the fact that they'd been knocked off their bike was acceptable. If that's the case we'll write to them and say 'this counts as an acceptable absence'. It gets wiped. (Practitioner, Site B) The new pre-court (that is, diversion) work was governed by fewer prescribed standards. One manager we spoke to felt this could be 'hugely innovative and free', and allow staff to focus on 'building a relationship' with young people rather than 'locking horns'. This was, indeed, reflected by some staff assessments of the new interventions, although we also found some evidence of 'system-creep', where workers imported aspects of postcourt work to the new pre-court services (see Kelly and Armitage 2015). One aspect of this was apparent at Site A, where we were told that a mirror system for monitoring the timely completion of paperwork had been introduced within pre-court services: Pre-court we have a lot less of the deadlines, structures etc. in place that post court have. The quality standards, inverted commas, and the ones that we do have are the ones we've decided to import from our colleagues. (Manager, Site A) Fascinatingly, the voluntarily adopted target given as an example appeared to be unachievable, since it was reportedly never met ('at the moment we're not even getting close'). Meanwhile at Site B, a practitioner told us of a 'joke' within the service following a communication from a senior manager which claimed that the organisation no longer worked to targets but also emphasised the importance of 'criteria' being met.

A Transformation of Accountability
When asked about the distinction that many professionals felt existed between meeting the needs of young people and meeting the needs of the service to have its performance measured and to be auditable, one senior manager explained it this way: The way I try and get it across in this service, and in common with many other services in the country, is we needed to get people to record better than they ever did do for monitoring purposes, for inspection purposes, if it isn't recorded it didn't happen purposes. We needed to evidence it. And we needed to improve the quality of data that we got to be able to manage the service, to know what we were doing in terms of reoffending and all that stuff. So we banged on and on and on and on and on and on and on about feeding the beast, and not surprisingly, . . . the distinction between the proper work and the admin stuff . . . , the wall just got bigger and bigger and bigger between the two, and stronger and thicker and so on. (Senior Manager) Indeed, we were also told that a reduction in centralised reporting requirements (Youth Justice Board 2011) had caused few noticeable changes in demands for information at the local level: What the YJB used to give us was a set of counting rules and a huge number of tables that we submitted to them on virtually everything . . . . We've all gone to local measures and they marketed it as reducing the burden on us but in fact we still need to be recording all of the same things and because our management boards and team management teams etc., have been used to seeing this information they still wanted to see the information so we were still having to produce it all locally. (Manager, Site B) This supports the recent speculation by Drake, Fergusson and Briggs (2014) that: Far from advancing in lock step, increased localisation is in principle capable of reducing practitioner discretion. It has the potential to generate small-scale centralism, whereby newly empowered local managers replace the paraphernalia of centralist controls with their own closely monitored strictures. (p.26) Thus, while the introduction of local controls appears contrary to the current direction of travel of the YJB (which is arguably moving back from 'must' to 'should' after relaxing some requirements in the most recent iteration of National Standards), there is less contradiction when managerial accounts (and frontline accounts of managers) are considered. When considering the benefits of recording decisions and decision makers to local bureaucracies, the 'janus-faced' nature of all youth justice work, and not only frontline work, becomes clearly apparent. Youth justice services were understood by many to be competing for recognition locally and nationally, with senior managers keen to secure favourable assessments not only by the YJB, but by local political actors such as local councillors, the local authority executive and the magistracy and judiciary. We were also told that favourable assessments by HM Inspectorate of Probation (HMIP) were particularly important, not only in providing some form of external validation, but as a means of securing the local reputation of the service, its political position within the local authority and thus its continued funding position: I mean, the holy book is the HMIP criteria, which I don't know if you've had a look through those at all for the inspection? And while they're a pain if you're being blunt, they do actually give you a very clear idea of what's regarded as good practice . . . . We could think we're effective in terms of outcomes for young people as much as we like but if HMIP come and say 'no, you're performing dreadfully' then [the senior managers] are out of a job. (Senior Manager) [Head of Service] is really good with that stuff and has [raised our profile] with the YJB and with [local authority] and our Management Boards certainly. Politically to me it looks . . . the height of stupidity to start cutting away something that's showing as working and has been seen as good practice within an organisation because you need to make cuts. (Manager) Could we be taken to court for not following the law? I don't know. I've never really thought about it in those terms. I've thought more in terms of actually if we don't have good relationships with courts, don't have good relationships with police, don't have good relationships with YJB, don't impress the Inspectors, then outcomes for young people are likely to be worse. (Senior Manager) If future funding (and their own jobs) are believed to depend on being able to demonstrate continual service improvement to local political masters in the context of declining resources across all children's services, it is counter-intuitive for senior managers to reduce performance information, or to run the risk of a poor inspection or serious case review after informally overseeing a reduction in standards around recording or timekeeping. However, managers at all levels also acknowledged that the weight of information demands in youth justice risk undermining the 'real work' of the service (in the memorable words of one participant -'you don't make a pig fatter by weighing it').

Conclusion
Within social work, where there is no central bureaucratic equivalent of the YJB, the critique of managerialism is primarily formulated with reference to local management as well as outside regulatory bodies such as the Health and Care Professions Council. This has encouraged greater analysis of tensions within occupational hierarchies. Rogowski (2012), for example, forcefully suggests that the social work profession has been degraded into 'a so-called profession with managers now dominating what practitioners do' (p.921). The problematic concept of 'professionalism' is still more difficult when applied to youth justice due to its multi-agency basis, diversity of access routes and lack of professional registration (for example, Hester 2010). When understood as a synonym for agentic practitioner, however, our data point to a need to analyse shifts in 'the balance of power, not the elimination of professional power' (Evans and Harris 2004, p.892) within local bureaucracies.
Our fieldwork occurred at a time when decisions taken by youth justice professionals, in the context of shifting economic and political conditions, had encouraged rapid and fairly radical changes to the organisation of local youth justice services. This was particularly apparent at Site B, which had moved to an integrated model of provision; however, the management team at Site A had also restructured the service in the face of significant budget cuts. The importance of 'professional judgment' and 'discretion' as a means for determining responses to young people in trouble with the law had also been re-emphasised within revised National Standards (Youth Justice Board 2013) and within the revised standardised assessment tool, AssetPlus (Youth Justice Board 2014). We have argued, however, that it is crucial to distinguish between 'managerial' and 'practitioner' discretion when considering the impact of standardised processes and nationally-specified requirements on decision making. Despite expressed managerial support and dedicated local initiatives such as practice improvement officers or 'surgeries' staffed by experienced practitioners, a number of frontline workers interviewed for this study struggled to envisage a situation where they could 'say no' and challenge rules or standard operating procedures within the service. When this did occur, (middle) managerial support was commonly reported as a necessary condition for the exercise of 'practitioner discretion'. We were also told that reporting requirements had not been reduced by decentralisation since information demands introduced by the YJB now proved useful to local managers and/or those further up the hierarchy.
In this article, we have argued that the day-to-day realities of youth justice work and the outcomes of such work are shaped by 'centralisation', 'standardisation' or 'discretion' but not solely, exclusively, or in an overdetermined fashion. The political realm does play a determining role in youth justice practice via its transformation of 'accountability', but, as the data above demonstrate, agentic practitioners at different levels within and across youth justice agencies and organisation also have a part to play (see also Phoenix 2016). The data also demonstrate that other non-youth justice agencies and actors also have influence, particularly when it comes to shifting the locus of control. Less centralisation and less standardisation have certainly brought local authorities and local government much closer in shaping youth justice work 'on the ground', if only because of the effect it has on transforming 'accountability'. Practitioners remain accountable for their youth justice practice and work. Managers, however, become ever more accountable for the outcomes of their teams -and to a wider range of actors and organisations, albeit at the local political level, than recent youth justice scholarship has commonly discussed. 7 Notes 1 There was no single definition of 'practice' articulated by practitioners in this study.
Some believed practice encompassed all roles performed, while others felt it mainly related to the face-to-face work. 2 Hereafter 'practitioners', but practitioners covers probation officers, social workers, youth workers, health workers, teachers, police, and others who had come up through the original YOTs. The interviewees worked across all levels, including service leads, senior managers, officers, and sessional workers. 3 'Contacts' could include verbal exchanges, meetings, phone calls, text messages, or emails. 4 The Youth Crime Action Plan 2008 (HM Government 2008) pledged to reduce the numbers of young people entering the youth justice system. The controversial police 'offences brought to justice' target was revised to place more emphasis on serious crimes in 2008 and removed in 2010. 5 Such expansion seems likely to increase. At the time of writing, Kate Morris, Director of Operations at the YJB, was encouraging YOTs to consider 'opportunities for alternative funding streams' in the face of new budget cuts (Youth Justice Board 2016). 6 While this article focuses on the distinction made between 'paperwork' and the 'real work' made by some of our sample, the project as a whole considered the multiple roles performed by youth justice workers.