Enacting autonomy reform in schools: The re-shaping of roles and relationships under Local Schools, Local Decisions

Local Schools, Local Decisions (LSLD) was a package of school autonomy reforms operating in the state of New South Wales, Australia from 2012 to 2020. The set of reforms centred on the devolution of additional powers and responsibilities to school principals, namely enhanced capacity to manage staffing and financial functions in response to local conditions. Using a conceptual lens of policy enactment, we analyse interview data gathered from 31 teachers and school leaders on how these reform areas were understood and enacted at the school level. Our findings highlight the tensions in enacting devolutionary reform in schools. While the centrality of the school principal’s role was emphasised, including in relation to contested levels of principal discretion, the enactment of devolved powers and responsibilities also produced a fracturing of staff relationships within schools, notably between principals and teaching staff. This finding is understood within a context of heightened workload and unclear expectations which attended the policy’s introduction. We contribute to the school autonomy literature through: (a) the inclusion of teachers’ voices, a stakeholder perspective often missing in the autonomy literature, enabling the impact of the reforms on interpersonal, relational dynamics to come to the fore; and (b) exploring implications for future reform suggested by the fate of LSLD. In doing so, this article deepens knowledge on the enactment of autonomy reforms in schools, drawing implications for understanding school autonomy reform around the globe.


Introduction
It has been 10 years since the introduction of the Local Schools, Local Decisions reforms in New South Wales (NSW) public schools, a policy representing "potentially the greatest change to school governance and decision-making in more than 160 years of public education in NSW" (Dinham, 2012, p. 13). This article offers a critical reflection on the reforms, examining their 'implementation' and felt impact on teachers and school leaders at a time when neoliberal policies have advanced an agenda for enhanced school autonomy and local choice. Portrayed as a panacea for existing centralised 'one-size-fits-all' procedures and processes that were perceived as inhibiting local decision-making, the reforms, introduced in 2012, intended to afford principals and schools greater authority and flexibility to manage staffing and financial responsibilities in response to local complexities (NSW DEC, 2011). It was expected that enhancing school (principal) autonomy would also increase student achievement, deemed critical to overcome the declining performance of public schools, as measured by the OECD's PISA 1 tests, and match benchmarked 'high performing' education systems overseas (Piccoli, 2011;Productivity Commission, 2012).
This article aims to advance knowledge about the enactment of devolutionary policy in schools, examining principals' and teachers' roles as policy actors and the implications for in-school relations at a time of changed governance arrangements in schooling. Understanding the interpretation and translation of the policy by schoollevel actors, as well as implications of the policy for school leaders' and teachers' work and relationships, is important in an international context where autonomy reform is increasingly common (Keddie, 2016), yet under-studied. Using the LSLD reforms as a case study, the following research questions are examined in this article: 1. How has devolutionary policy been enacted in NSW public schools? 2. What impacts on teachers' and school leaders' work and relationships are evident in the enactment of this policy?
Drawing on data collected from an interview-based study of NSW teachers and school leaders five years into the LSLD reforms, this article contributes insights on how school autonomy is enacted, and the contested, sometimes divisive responsibilities generated for teachers and school leaders (mostly principals), which are compounded by heightened workloads and weakened centralised support. Our inclusion of and close focus on both teacher and school leader perspectives enables findings beyond those of contemporaneous departmental evaluations, highlighting not only how the role of the principal has changed, but the impact of this changing role upon relationships with teaching staff, and implications for work and workload.
The article proceeds as follows. First, it explains the particular policy context of NSW, Australia, then provides a literature review of decentralisation in schools and introduces the conceptual framing of the article using the policy enactment approach. We next outline the study's method, before providing an analysis and discussion of LSLD according to each dimension of the policy, situated within existing literature on devolution and autonomy in schooling, as well as other departmental evaluations of LSLD. We conclude with a discussion of policy implications arising from enhanced decision-making power.

The policy context: Local Schools, Local Decisions
Increasing school autonomy has been a major policy priority in Australia, where education is a constitutional responsibility of the states and territories. Funding, governance and operational arrangements vary across states and territories. There is also variation in education systems with different degrees and forms of devolution nation-wide. While devolutionary initiatives have been a feature of, for example, the education system in the state of Victoria for some time (Blackmore, 2004), NSW has moved to a more autonomous public schooling model only in the last decade. Although devolutionary ideas were first introduced in NSW under the Greiner Government's review of school administration in the late 1980s to decentralise some education functions, most initiatives of the government's Schools Renewal Strategy were not implemented .
Some 20 years later, the NSW Department of Education launched the LSLD reforms. This was introduced in tandem with the Empowering Local Schools initiative of the Federal Labor Government which aimed to facilitate greater autonomy for government and non-government schools (Productivity Commission, 2012). LSLD aimed to give NSW government schools more authority to make local decisions about how best to meet the needs of their students. The LSLD reform package encompassed 37 different initiatives across five key reform areas-making decisions, managing resources, staffing schools, working locally, and reducing red tape. Due to the phased rollout of the reform, each of these reform areas was progressively implemented from 2012 and were fully in operation by the end of 2018 in all NSW public schools.
The reforms focused on allowing principals to make decisions based on the needs of their school community, with the view that individual schools are 'best placed' to understand, and respond to, local complexities (NSW DEC, 2011, p. 3). The policy emphasised how rules and processes at the time made it hard for principals and teachers to respond quickly to issues, with the Minister for Education commenting that: "For too long public schools have had their capacity to adapt to meet the needs of their students stymied by bureaucratic red tape and overcentralised command and control" (Piccoli, 2011). In relation to staffing in particular, the 'one size fits all' staffing formula was viewed as failing to consider that "[e]very school is different, with different challenges" (NSW DEC, 2011, p. 7). The reforms would also allow schools to manage more than 70% of the state's public school education budget, a substantial increase from 10% in 2013 (NSW DOE, 2018, p. 12).
In tandem with LSLD, a new model facilitating funding through the National Education Reform Agreement 2 was introduced in NSW public schools in 2014. This new 'needs-based' Resource Allocation Model was deemed a 'simpler, more transparent model' that provided targeted (individual student) funding and equity loadings for identified areas of disadvantage on top of a base school allocation (NSW DOE, 2018, p. 10). In 2018, NSW public schools received approximately $851 million in equity funding through this model (CESE, 2020, p. 27). Importantly, this funding model and the funding received through it is different to the LSLD reforms themselves, although as will be seen below, they are sometimes conflated.
Somewhat surprisingly, given political commitment to enhancing decision-making power for schools, the LSLD reform was subsequently removed after a decade of operation in NSW public schools. A final report produced by the NSW Department of Education identified problems with the policy's implementation and its failure to achieve improvements in key performance indicators, like student outcomes (CESE, 2020). This article therefore presents a timely academic analysis of the LSLD reform, drawing on, but distinct from, Department analyses, and in the context of the popularity of autonomy models in education systems around the globe. In our analysis, we draw on both principal and teacher voices, as well as school leaders such as Assistant/Deputy Principals and Head Teachers. This enables us to provide a rounded critical analysis of autonomy reforms and contribute insights on key stakeholder voices to the academic literature on autonomy models in school education.

Literature review: Decentralising power and responsibility
School autonomy or 'devolution' is a current trend in school governance worldwide, involving transferring of decision-making responsibility to a lower level (Keddie, 2016). The imperative to shift more responsibility and decision-making authority to schools under LSLD characterises the policy as a devolutionary reform which seeks to enhance the self-management of public schools. Like other Western liberal democracies, school autonomy and devolution have been a feature of education reform in Australia in recent decades amidst the popularisation of corporate managerialist and economic rationalist ideologies (Lingard et al., 2002). Centralised arrangements have been increasingly eschewed for their inflexibility and perceived barrier to achieving quality and equity, as schools have little control or capacity to respond to their local context (Gobby, 2013). School autonomy implies that devolving responsibilities from large bureaucratic organisations to smaller entities and 'empowering' those who work in schools variously produces more efficient, effective, and democratic organisations and outcomes (Gobby, 2013;Lingard et al., 2002). Indeed, the World Bank concludes that most countries whose students perform well give their schools substantial authority to shape local education provision and determine the allocation and management of resources (Bruns et al., 2011).
Others, in contrast, emphasise a lack of empirical evidence which connects devolution to improved student learning and outcomes, and argue that it leads to further consequences, such as managerialisation of the school principal's role and expanded marketisation of school education (e.g. Holloway & Keddie, 2020;Keddie et al., 2018). An emerging consequence of decentralising reforms has also been the increased workload outcomes of heightened 'accountability' and teacher 'responsibility' combined with greater work intensification (Brennan, 2009;Dinham, 2013). Studies on devolutionary-style policies have revealed how teachers have experienced an increased and unmanageable workload under these reforms, resembling a 'tsunami of paperwork' . With responsibilities increasingly devolved to the local school level, there is an accompanying increase in accountability requirements, which, in turn, elevate workload pressures (Brennan, 2009;Dinham, 2013). But despite these concerns about the mixed findings related to efficacy and effects, there appears to be a continued political commitment to decentralisation and school autonomy (Gobby, 2013). In NSW, it is notable that while the LSLD reform has been removed in name, it has been replaced with the 'School Success Model' which the Department describes as 'building on' (NSW DOE, n.d., p. 1), rather than entirely revoking all changes made under LSLD. We discuss the School Success Model further in the final section of this article.
Scholars have also observed how the policy rhetoric of decentralising education systems and empowering local schools has been accompanied by a simultaneous recentralisation of administrative power at the national level, through reforms such as the development of national curricula and standardised testing (Robinson, 2015). Neoliberal policy agendas in education have seen teachers become increasingly "constrained in their own freedom and agency" and closely scrutinised and monitored due to these recentralising forces (Edwards-Groves et al., 2010, p. 46 cited in Hickey et al., 2021. Indeed, Hickey et al. (2021) observe how a peculiar feature of current devolutionary reform agendas is the emphasis on responding to 'local', unique needs of schools, while accountability measures 'steer' the work of schools towards centrally mandated requirements. Within this context of decentralising power and responsibility, this article will examine the enactment of LSLD as a devolutionary reform framed as empowering local schools with greater decisionmaking authority, and examine its implications for re-shaping in-school roles and relationships.

Conceptual approach: Policy 'enactment'
Over the last two decades, policymaking has become an 'epidemic' of global proportions (Levin, 1998cited in Braun et al., 2010. Education policymaking has been appropriated by the central state in its determination to modernise education provision and 'raise standards', resulting in demands placed on schools and teachers to implement multiple, sometimes contradictory, policies and be held increasingly accountable for their actions (Ball, 1997). Within the last decade, scholars have turned their focus to understanding how schools actually deal with these policy demands-how they interpret policy texts and translate these into practices in real material conditions and with varying resources; and how they are enacted (Ball et al., 2012, original emphasis). This approach attempts to redress rationalist policy analyses which generally take policy as a finished object crafted at higher levels of bureaucratic structures and reduces schools to decontextualized subjects that must simply 'implement' policy determined elsewhere (Ozga, 2000, p. 42 cited in Ball et al., 2012. This article uses the idea of 'policy enactment' (Braun et al., 2010) to understand how the LSLD reforms are interpreted and translated by diverse policy actors in the school environment, rather than simply implemented. In comparison to policy implementation which takes an 'outcome-driven' perspective, in policy enactment, teachers and principals are viewed as agents of translating policy into practice (Armstrong, 2003cited in Braun et al., 2010. This more critical approach foregrounds the messy reality of engagement with policy and examines how those charged with policy 'implementation' engage with, respond to, and enact policy under complex conditions (Ball et al., 2012). This involves processes of interpretation and translation of policy texts, where interpretation includes reading and making sense of a policy, while translation involves 'enacting' policy through various initiatives like meetings, school plans, classroom lessons or school websites (Ball et al., 2012). As part of this approach, Ball et al. (2012) also observe the critical role played by various policy actors in the enactment of policy, including: • Narrators: interpret, select and enforce meanings; mainly done by school leaders • Entrepreneurs: advocate and integrate policy • Outsiders: support policy through partnership and monitoring • Transactors: account, report and monitor policy • Enthusiasts: advocates in policy work • Translators: producers of texts, artefacts • Critics: critique policy; mainly union representatives • Receivers: cope with and/or depend upon policies; mainly junior teachers Surprisingly few studies have sought to understand the enactment of devolutionary policy and its effects, particularly for teachers, despite the global trend of decentralising decision-making authority. Gobby (2013) is one study that mobilises Ball et al.'s approach to policy enactment. Gobby found that the devolutionary Independent Schools Program (IPS) in Western Australia was designed to empower principals, and principals indeed took up flexibilities offered by the program. Principals also, however, felt burdened by an excess of administrative tasks being devolved to schools, shifting their work from driving educational improvement to managing risks. This had the effect of constraining their capacity to innovate and problem-solve, thus frustrating the program's goals of added flexibility and autonomy (Gobby, 2013). More recently, Gobby et al. (2018) examine school autonomy under the IPS initiative in both Queensland and Western Australia, highlighting how experiences of autonomy are mediated by school type (primary and secondary) as well as individual principal disposition, concluding that "the enactment of competing responsibilities relies on the capacity to draw on the assemblage of professional discourses and practices through which school leaders exercise their autonomy" (Gobby et al., 2018, p. 170).
Another recent article which engages with policy enactment, but specifically focuses on the experiences of teachers, is Wilkins et al. (2021), who argue that the kind of professionalism characteristic of autonomous schools is 'neo-performative', shifting understandings of equity and social justice through 'audit culture'. Other studies have engaged with ideas around policy enactment or doing policy but without specifically tying this conceptualisation to the work of Ball and colleagues. For instance, Forsey's (2004) early study of the work of one principal inspired by devolutionary ideals (but without these, at the time, being formalised into state policy), is revealing of how take-up of policy ideas can vary according to individual actors, with this principal reflecting something of the policy 'entrepreneur'.

Method and data sets
To examine the enactment of devolutionary policy in NSW public schools, this article draws upon data from a study of teacher workload conducted over 2017, in partnership with the NSW Teachers' Federation 3 (NSWTF) and carried out with full academic independence. This ethically-approved, interview-based study focused on the issue of teacher workload in which questions relating directly to the LSLD reforms were asked, given the researchers' prior interview work which suggested a link between devolution and increased work requirements in schools (see Fitzgerald et al., 2019). In the study reported in this article, however, we were interested in explicitly addressing the reform itself, rather than the general experience of an overall policy climate as reported in Fitzgerald et al. (2019). This involved conducting interviews with 31 teachers and school leaders comprising: Principals (n = 3), Assistant/Deputy Principals (n = 4) and Head Teachers (which in Australia is a head of faculty) (n = 6) and classroom teachers (n = 18). In the data we report below, these participants are quoted according to their position (T = Teacher, P = Principal, AP/ DP = Assistant Principal or Deputy Principal, HT = Head Teacher), followed by a number. Given we are only reporting the views of three principals out of a dataset of 31 interviews, it may seem curious that the principal becomes a point of focus in the arguments that follow. However, this is due to commentary on this role offered by other interviewees, for whom this role appeared to take on a heightened, yet disconnected kind of significance in the reform period studied.
Participants were recruited via the NSWTF, which sent an email to randomly selected schools to fulfil the quota of desired schools across participant role, school type, school geography and ICSEA 4 value. The 31 participants were selected out of the 99 potential respondents who volunteered to participate, with this number considered to both ensure sufficient spread across these contextual variables, and yet remain manageable for a detailed qualitative analysis. This modest sample size allows discernment of overall trends regarding the enactment of devolutionary policy, despite not being fully representative of all NSW schools or school staff. Interviews were approximately 60 min in length, conducted via telephone, and recorded with the explicit permission of the participants via signing of a consent form. The questions we report on in this article directly focused on the five aspects of the reform. For each of these areas, participants were asked two questions: 1. What changes, if any, have you observed (or if you are a principal, implemented or experienced) in relation to: a. How resources are managed in your school? b. How staffing is handled in your school? c. How your school is working with other local schools, businesses and community? d. Reducing red tape in your school? e. How making decisions is handled in your school? 2. What do you think the impact of these changes, or lack of change, has been?
The data gathered in response to these questions were extracted from the overall data set and grouped together for each of the five aspects of reform, before being subjected to a thematic analysis process as described by Ezzy (2003). As such, themes were able to emerge from the data whilst still remaining in context in relation to the particular area of the reform that we had asked participants to speak to. Once grouped, data were read carefully by the second author of the article, and then re-read and annotated with emerging codes that were eventually grouped into key themes constituting "the argument, or central story" (Ezzy, 2003, p. 87) with which our discussion of results below is organised. The insights thus gathered depict the multiple ways in which policy reform was enacted in schools over a 5-year period since the introduction of LSLD in 2012.

Reported change in and impact of the reform areas
In this section, we present our interview findings from these two questions related to LSLD, structured according to the five areas of the reform. We place these data into conversation with the Department's own final evaluation of the policy (CESE, 2020) and associated studies documented in the evaluation report, where relevant. The discussion is further supplemented by broader research on devolutionary reform and change in schools.

Managing resources
The 'managing resources' reform area was intended to enable more flexible and responsive decision-making at the local level through implementation of the RAM and affording schools capacity to manage a much larger share of the overall education budget. The new funding model replaced centrally-run programs targeting different student needs that often required schools to write separate plans, reports, and budgets for each program. Schools instead had one budget, guided by a single school plan.
A sense of greater autonomy was confirmed by participants working in leadership roles. One AP/DP described how things were now: More managed from a school perspective, [with the] school given way more of its allocation to expend itself, as opposed to [having] more tied grants that we're specifically told we have to spend on specific things. We have greater autonomy and resourcing (AP/DP-3).
For some, this greater responsibility for managing school finances was positive, with one principal taking on the role of policy entrepreneur in commenting that the change had allowed for the hiring of "amazing educators who can bring about and lead and support improvements to teaching and learning" (P-3). This reflects the positive associations with autonomy reform noted in existing research (see e.g. Keddie et al., 2018), where enhanced 'freedoms' can be used to address school needs. Department data support these participants' comments, showing that funding was primarily used to implement specific literacy and numeracy interventions, employ additional staff, and provide release time for teachers to engage collaboratively on curriculum design and mentoring (NSW DOE, 2018, p. 24).
However, another theme from our interviews was that although there were more resources to handle locally, there was also "a lot more requirement for accountability, without enough information" (HT-4). This suggested that the reform also involved undertaking a challenging process of narration or sensemaking in enacting these new responsibilities, given the limited perceived information with which to make financial decisions and report on outcomes of those decisions as transactors. The relationship between devolutionary reform and heightened accountability is well-established (e.g. Gobby, 2013;Gobby et al., 2018), however these data indicate not only an accountability requirement associated with the policy, but a notable lack of clarity about what this involved. These concerns reflect criticisms reported by the NSW Auditor-General (just before the LSLD reform was replaced), which found that principals were poorly supported to manage additional accountability for funding, received inconsistent advice on how to spend the money, and were given complicated, ineffective ways to account for it (Crawford, 2020).
What is new in our data is that this lack of clarity around how resources are to be 'managed' could create "mistrust or bewilderment or dismay" (HT-4), and could be "unsettling" for other staff (HT-4), suggesting the adoption of a critic role by some teachers. Some participants, such as HT-1 and T-13, described seeing less money at faculty level, rather than more, due to the introduction of a global budget and feeling responsible for a wider range of services than previously, meaning it felt like resources were spread more thinly as they were rendered receivers of allocation decisions being made differently and according to different principals' personal professional priorities . It also, however, suggests that the 'fracturing' of relationships under autonomy reform may not only be within the wider educational community (Holloway & Keddie, 2020), but within the individual school. Teacher-participants indicated a sense of frustration, with a described opacity around resource management, controlled by senior executive and invisible to classroom teachers: "It's very opaque how resources are managed at our school" (T-13).
In the view of this participant, money could be given or withheld and used "as a tool to put pressure" on particular staff members at the discretion of the principal. T-14 similarly commented on resources being managed "a little too secretively" at their school. Another participant expressed similar sentiments but with heightened empathy, seeing it as a lack of skills for resource management on the part of principals: I mean it's a big job on any principal about having to manage all the money and all that sort of stuff now. But again, I don't think they probably have the right skills to be able to manage it wisely (T-8).
These comments extend arguments we have made elsewhere, observing an apparent disconnect between principals and the needs of their staff in the context of devolutionary policy  and highlighting how autonomy reform can serve to intensify the divisions between these roles. It also reflects research which notes the administrative impost that tends to be associated with autonomy reform more generally (Gobby, 2013). How individual leaders experienced and responded to these pressures appeared to be seen as key; according to AP/ DP2, "it just comes down to management styles I think by the principal" to make sure allocations are equitable and reasonable. But as T-8 noted, this was a big job for principals. In sum, evidence points to lack of clarity about what accountability reporting should actually report on, a repositioning within schools of who has input to financial decision-making, and associated fracturing of relations between principals and teachers, with the individual figure of the principal taking on particular significance.

Staffing schools
The 'staffing schools' reform area aimed to provide greater support to schools to improve teacher quality and provide greater flexibility over staffing mix. In relation to this reform area, a small number of participants (teachers) explain having seen no real change to staffing processes during the past five years; an equally small number report seeing positive changes due to enhanced flexibility, although this is arguably confounded with changes to funding amounts rather than funding mechanisms. One enthusiast principal participant commented that through "flexible funding" they had been able "to employ somebody…to lead programs as well as take teachers off class for additional…time together to plan, to research, to design and deliver, effective teaching and learning programs. It's allowed time for teachers to observe one another" (P-3).
There was, however, some concern that additional employees tended to be in leadership: "our Gonski money 5 has all gone on executive sort of positions" (T-13), which weren't always felt to relieve pressures in the day-to-day operations of schools or in the classroom. This comment is supported by data from the Department's interim evaluation of LSLD showing the large increase (331%) in the number of Assistant Principals employed, the very large increase (7480%) in the number of Instructional Leaders, and the large increase (94%) in Business Managers over the 2012-17 period (NSW DOE, 2018, p. 33). Hence, while the policy enabled enhanced power over staffing for schools, this was not necessarily perceived positively by participants, particularly teachers.
A more negative view was collectively expressed in relation to the enhancing of merit selection opportunities for hiring staff in schools (under LSLD this increased to every second teacher appointment). One participant, positioned in this explanation as a relatively powerless receiver, felt that the merit selection system had all but destroyed the 'transfer' system, which had previously allocated most staff centrally at Department level: "the priority list is non-existent now; my priority date now is almost I think just about when I was born [a very long time]" (AP/DP2). Indeed, HT-2 commented that most positions, in their view, now seemed to be appointed via merit selection, despite being intended only for every second appointment: "I see most go to merit based" (HT-2). While some agreed with merit selection in principle, they felt in practice it did not always ensure best fit for a school: "you can get people who are very good at selling themselves…and can get jobs that way but they may not be necessarily the best person to fill the job" (HT-2). A more concerning and frequent position was taken in relation to how the system could, it was perceived, be manipulated by principals to select particular staff they liked, rather than being legitimately merit-based. AP/DP-2 described merit selection as "a bit of a crock" because: I've seen the jobs how they've been advertised and absolutely loaded up with all these requirements [skills] that they want…this teacher to have, and I'm thinking 'no teacher could possibly do all that' so they're obviously loading it up for a particular person.
In the words of T-13: Local Schools, Local Decisions [leaves schools] wide open for nepotism. It's jobs for your pals and that definitely is evident in our school. Also you drive out people that you don't like and…if you do like somebody you just engineer everything so that you give them all the opportunities…All the power is in one person's hands [who] can dispense favours and that's what happens -there is favouritism, and somebody…is given all the opportunities and therefore they're able to move up the ladder.
As such, for this participant, "equity goes out the door with Local Schools, Local Decisions" (T-13). This explanation suggests that an entrepreneurial enactment of staffing affordances by principals was not always appreciated by staff, and indeed responded to quite negatively by this critic. T-2 concurred: "If the Principal likes you, you're right…[A]dversely if he [sic] doesn't like you then you're in trouble… They've got many staff scared" (T-2). Again, there are concerns here around the shifting role of the principal and how additional 'powers' were perceived to be (mis) used by principals as individuals.
Relatedly, interviewees commonly reported an apparent rise in temporary positions, which are always selected locally unlike the permanent positions just discussed. This created a sense of growing precarity within the profession, as described by T-4: The biggest impacts I'm seeing right now is a huge increase in the number of temporary teachers in the school…I was the first temporary teacher they'd had in a while. Shortly after that a second temporary teacher arrived, we now have four temporary teachers and that's not good for the kids because the four of us are going at the end of the year and that's not good for the continuity of the kids.
Another participant commented on a similar apparent rise in temporary staff while the number of permanent appointments stagnated: "there's not much shift in permanent, but there's a lot of positions available that are casual and temp contracts" (AP/DP-3). Indeed, perceptions actually match reality, with growth in temporary positions relative to permanent ones , and intense frustration on the part of teachers in temporary roles . Here and in previous research , teachers have connected this rise explicitly with LSLD, given the enhanced power over staffing and an associated increase in the general sense of control which principals feel they have over these matters contributing to a shift in school relationships. As T-4 notes above, this had impacts for students, but also for the staff themselves, again positioned as receivers: "the temporary work is quite insecure, so even though…this is my fifth year at the school, always in the back of my mind there's that, 'what will happen next year'?" (T-12). In the words of another, it was "pretty hard on staff to not know if they had a job and then all of a sudden be told there is a job after you were told there wasn't" (T-3), creating "lots of stressed teachers and I think a bit of a breakdown of trust" (AP/DP-2).
A counterweight to these concerns about increased temporary roles and 'nepotism' in permanent appointments were concerns relating to principal and executive workload due to their increased powers. One principal participant felt they "spend half of my life on staffing at the moment, so I can have enough staff in the right areas to cover curriculum" (P-1), highlighting how principals were not always entrepreneurs but also sometimes critics. The merit selection process for permanent staff was described as quite time consuming: [H]aving to do the merit based selection instead of just getting appointments… who can start straight away [is] burdensome workload because you have to convene a panel to set job adverts to do the panel culling, [read] the CV and then [conduct] the actual interviews. And it drags out a process that was fixed in 24 hours before devolution. It can now turn a 24 hour appointment into… 6-8 week[s of] prolonged extra workload (T-10).
As such, once again we note a disjuncture between principals and other teaching staff in the school around the issue of staffing, seemingly caused-or at the least intensified-by the nature of this autonomy policy. With principals having enhanced power over enacting what they believe is best for their school, this can interact in complex ways with the needs of executive, teaching staff, and students. Additionally, and as we have previously argued , although the principal's role in managing staffing has grown, principals are not always seen to have the resources to undertake it satisfactorily. We extend that analysis through our argument here, which, in drawing on the voices of teachers, allows this clash in perspective to become clear from both sides.

Working locally
The 'working locally' reform area endeavoured to strengthen schools' consultation with local communities and enhance the sharing of resources to better meet local needs and support student learning. Again, teacher participants-on the whole-did not report seeing much change due to often being receivers of policy decisions: "I wouldn't say there's any major changes in that area" (HT-3); "I don't think that there has been any change in that" (T-1); "haven't really noticed much" (T-7). This suggests some limitation to the reform in its aim of enhancing local connections with community, and may reflect the research literature on school autonomy which tends to indicate only some enhanced local capacity alongside a strong focus on accountability (e.g. Gobby, 2013). Where participants did see changes they perceived them positively, enthusing that they were "very much more aware of ourselves as a community school" (AP/DP-4), seeing 'vast' improvements in relationships with community "in the last few years" (HT-4). As T-10 explained: There is a huge increase in networking and connecting and working together as a community of schools, not just as individual schools, and in one way that's very positive because it means if we've got certain resources and access to funding and another school in our community of schools does as well we can pool them together and get even more and we can share that. So there's been some positive effects in schools networking and linking up as a community of schools.
During the two-year pilot of LSLD on 47 schools in NSW prior to its full introduction in 2012, principals reported already having a diversity of mechanisms for community engagement (NSW DEC, 2012). Some schools had pre-existing boards or councils that included community members, which played a role in considering school needs and in supporting the recommendations of the principal and staff to undertake certain initiatives. In more cases, parents and citizens groups shaped pilot initiatives and monitored achievements of the school (NSW DEC, 2012). Interestingly, in our data, although some participants saw positive changes in working locally, this was sometimes qualified as not having been specifically within the last five years-as HT-1 commented: "I think schools are opening up to have better relationships with the broader community. I think that started a while ago though". This reflects the complex nature of policy; rather than being introduced and implemented, it interacts with local complexity including, in this case, where such efforts are already underway. This suggests that, with respect to working locally, principals are again acting as narrators in their interpretation of this reform area by relying on existing and established mechanisms to facilitate community engagement and do not feel administratively burdened by interpreting and enacting new process requirements, as in the case of devolved staffing and financial responsibilities.
These findings also reflect insights from the Department's evaluation, which found that 63% of principals disagreed that LSLD positively impacted the way schools consulted with parents and the school community to inform their local decision-making (CESE, 2020, p. 35). These results largely occurred because school staff did not feel LSLD contained concrete advice about mechanisms for increasing community engagement, beyond the initiatives already deployed by schools (CESE, 2020, p. 13). The level and nature of community consultation therefore often depended on principal's interest and capacity for community engagement (Crawford, 2020).

Reducing red tape
The 'reducing red tape' reform area emphasised reducing the administrative burden to enable schools to prioritise a focus on teaching and learning. For participants, this aspect of LSLD was the one that generated the most consensus, and most antagonism, and a clear role of critic across participant categories, with a resounding 'no' offered in response to the question of whether red tape had reduced during the LSLD reform: "Reduction in red tape? No (laughing)" (HT-3); "No, no, there's been no reduction in red tape, absolutely not, absolutely not" (P3); "What reduction?…It's increased" (T-2); "We are more red tape bound than ever" (T-4); "the changes that have…been made here are all around having that accountability as a principal, and evidence of what we're doing, and if anything, that's more workload and red tape, not less" (P-2). Or as HT-5 summarised, "everywhere we go there are forms and forms and forms to fill out" (HT-5). Together this evidence supports findings noted elsewhere that devolutionary reform tends to not only impact principals in terms of workload, but also teachers McGrath-Champ et al., 2019). It also emphasises the importance of attending to the lived experience of policy, and how educational environments can be reshaped as specific reform efforts are undertaken-what Ball (1993) refers to as the 'second order' effects of policy.
Participant comments align with findings from the Department's evaluation, where 90% of principals did not agree that LSLD had simplified administrative processes (CESE, 2020, p. 12). The LSLD reforms came with many tools and processes around school management and financial management, which principals and administrative staff described as cumbersome, time-consuming and complicated (CESE, 2020). Principals thus reported difficulty in fulfilling their role as educational leader due to the large proportion of time spent on administration. Problems also compounded as LSLD was introduced in parallel with multiple other large-scale reforms and policies in the state, each with their own set of tools and departmental expectations, and a lack of communication as to why these tools were important for schools (CESE, 2020, p. 32).
The effects of these changes, for participants, were: More pressure on schools and more responsibility and more blame and I don't know that we're really able to make local decisions beyond…playing around a little bit with staffing entitlements and funding. There's a lot more red tape. There's a lot more accountability that has come with it [which then] takes time away from what we're measured against. It's a vicious cycle. We're busy justifying decisions and spending and then we're asked why we haven't improved (P-3).
The accountability function of 'paperwork' under devolution is clearly highlighted here (Gobby, 2013). P-2 articulated that red tape "increases people's stress and impacts their personal life as well, and it also makes friction between me and staff", recalling again the impact of devolution on staff relationships, noted already in these findings and in the literature (see e.g. McGrath-Champ et al., 2019;Stacey et al., 2022). T-16 summarised thus: The impact has been quite devastating really because it's taking time out of your teaching, taking time out of your programming, it's taking time out of your training. So the red tape or the bureaucracy has really had a significant impact on teachers and certainly executive staff.
Only one participant (out of 31 interviewed) thought there had been some reduction to workload, and this was due to a school-level change around the format of reports, an aspect that was not specifically part of the 'reducing red tape' reform area.
While the 'reducing red tape' reform area intended to reduce the administrative burden on schools to enable them to focus on teaching and learning, school leaders and teachers were active critics of the unintended consequences of enhanced local responsibility and accountability that increased workloads and required 'evidence' of decision-making, impinging on teachers' core role as educators and principals' role as educational leaders. While both teachers and school leaders were forthright in their criticism of this area of reform, it did not unite them. Instead, participants' perceived increase in pressure and stress negatively affected relationships between teachers and school leaders, by creating an environment of heightened demand for all.

Making decisions
The 'making decisions' reform area was intended to enable school leaders to better respond directly to the learning needs of their students, with the opportunity to develop responsive and evidence-based local solutions. This included potential for enhanced decision-making around the recruitment of staff, budgetary decisions about centrally-provided funding, as well as maintenance and planning, within state and national policy guidelines. There was very little focus on greater authority in relation to curriculum.
Comments from participants focused on the increasing importance of the role of the principal in relation to 'making decisions', with the principal cast as entrepreneur: as expressed by an AP/DP, "it's about, I'm the principal and I'm directing you to duties as I see fit" (AP/DP-1). Again, we see here a fracturing of school staff relationships, within school executive and particularly between executive roles and classroom teachers. Another participant explained how, in their view: The principal and deputy [principal] have become more dictatorial…I think it was supposed to be the other way around, that those decisions were supposed to be spread out with teachers and the executive, but with this one basically the principal is like 'my word is it and this is the way I want it' and she doesn't like anybody disagreeing with that (T-5).
This participant also noted, however, that they had a different experience of these same reforms at another school. This echoes again the theme about the importance of the perceived individual qualities of the principal that suffused the data. Further, it reflects research which highlights the complex balancing of competing responsibilities which principals must enact under autonomy . As HT-3 put it from their perspective: "the decision making, I think has deteriorated, but that's mainly as a result of the principal, rather than anything else" (HT-3); for AP/ DP-3, it "comes back to the leader of the school…It depends on how they want to run it. They can either do that collaboratively at times, or make the decisions themselves. So, it does usually depend on the leader" (AP/DP-3). For many teacher participants, decision-making was therefore felt to occur without much staff consultation, as principals took on the role of narrator and entrepreneur. 'Making decisions' invests principals' with a new level of authority, interpreted as enabling power to reside with the school principal, rather than encouraging school-wide consultation and devolved decision-making, at the discretion of the principal.
The negative critic comments above largely come from participants in executive roles such as faculty Head Teacher or Assistant/Deputy Principal. Responses from principals, however, largely focused on the difficulty of making decisionsthe range of policies and stakeholders that needed to be consulted. This labour was, perhaps, not always visible to other school staff. In contrast, results from a 2019 principal survey run by the Department found most principals felt more empowered to make local decisions, with LSLD enabling them to better meet the diverse learning needs of their students (CESE, 2020, p. 30). The data we have presented adds nuance to this initial insight of principals feeling more empowered to make local decisions by highlighting the work that is nevertheless involved in doing so, as well as the impacts on interpersonal dynamics that such a shift in making decisions can evoke.

Discussion of themes arising across the reform areas
This article has considered the enactment of devolutionary policy in NSW public schools, examining principals' and teachers' roles as policy actors in a time when autonomy reform is increasingly common. By analysing the interpretation and translation of the five main reform areas of the LSLD policy by school-level actors through interview data, we have explored how the reform has been enacted through principals' and teachers' adoption of a range of policy actor roles (Research Question 1), and its impacts: elevating the importance of the principal, straining in-school staff relationships, and producing what seem to be unintended knock-on effects for workload (Research Question 2). In this section, we offer an integrated discussion of the data presented above, to more fully address these research questions. In doing so, we argue that the inclusion of teachers' voices offers a novel contribution to the school autonomy literature, highlighting relational tensions and revealing the stark experiences of principals and teachers in differing policy actor roles when schools are afforded enhanced powers.
The first research question this article addressed was how devolutionary policy has been enacted in NSW schools. The overriding finding from the analysis presented above, and clear across almost every reform area, is that it is principals who most actively engaged with this reform, taking on, in Ball et al.'s (2011) terms, the roles of policy narrator, enthusiast, entrepreneur and/or transactor. All these roles entail positive dispositions of engagement and agency, with principals seeming cognisant of particular forms of agency as leaders in the policy enactment process, as opposed to passive recipients (i.e. receivers), variously interpreting, championing and facilitating policy at their school. Additionally, principals' understanding of policy and what they want to achieve from it can change how the reform appears to, and is experienced by, teachers (reflecting Bergmark and Hansson's (2020) finding that enacting policy is a process fraught with fragility and instability). Significantly, while LSLD is framed as empowering schools, with more authority and flexibility in decision-making around key areas such as staffing and finances, principals' interpretation of the policy sees principals themselves more empowered to take on these new responsibilities. Such empowerment impacted in-school relations. Principals' enhanced discretionary power was not always favourably experienced by school staff, with teachers feeling principals failed to engage school staff in decision-making, and interpreted the policy in a way that suited their leadership style such that reform initiatives could be enacted differently, and often not favourably, at each school. This additionally reflected what seemed to be an enhanced focus on the principal as an individual; rather than focusing critique on the policy settings of which they are a part, teacher participants frequently commented on the personal qualities of principals, suggesting an environment in which individuals may feel pitted against one another rather than united in common cause.

3
In contrast to principals, the enactment of almost all the reform areas of this policy placed teachers in less clearly active roles, with a tendency to find themselves positioned, in relation to Ball et al.'s (2011) typology, as receivers or critics. In our data, teachers did not, or were not able to, take on more active roles of entrepreneur or narrator, or even enthusiast. Often teachers felt they had little knowledge or oversight of policy decisions as principals narrated policy directives, rendering teachers mere recipients. While Ball et al., (2012, p. 3) emphasise that "[p]olicy is done by and done to teachers', they are actors and subjects, subject to and objects of policy", it is clear that principals' enactment of devolutionary power limits the potential with which teachers can shape policy processes within schools. Indeed, that teachers frequently saw no change and generally did not seem familiar with the specific (five) reform areas of LSLD also evidences their exclusion from this policy shift.
Yet we argue that the experiences of teachers are relevant to a full understanding of how autonomy reform impacts upon schools. While autonomy reform may ostensibly target principals, it affects teachers too. Responses to the enactment of the 'Staffing Schools' reform area highlight this particularly clearly. Enhanced power of schools (principals) to directly hire staff worked in tandem with an increased number of temporary positions in schools, creating stress for teachers in these precarious positions and feelings of conflict with the principals who were understood to have placed them there . This 'teacher perspective' is a core contribution of this article, in a field dominated by the voices of principals (e.g. Gobby, 2013;Holloway & Keddie, 2019Keddie et al., 2018). Including a fuller range of staff roles can serve to highlight how schools as professional environments change through the interaction between policy and those who enact it at local levels. In the analysis we have presented above, relationships between principals and teachers fractured as teachers, excluded from changes to which they were nevertheless subject, were positioned as receivers of policy. This positioning meant that policy changes could be seen as primarily flowing from and attributable to the principal, who became a key target of teacher critique. This supports our previous analyses which suggest that the workload burden on principals via autonomy reform can compromise their support of, and work with, teachers . On the flip side, it would seem that autonomy reform can create fertile ground for teachers to feel a heightened mistrust and suspicion of principals.
This finding addresses our second research question, regarding impacts on principals' and teachers' work evident through enactment of the policy. Heightened tensions between principals and other school staff were clearly apparent in our findings, reflected both in the vastly different enactment roles across four of the five reform areas, and the abundant substantive comments which depict disparity of views and positions regarding these. In only one of the five reform areas, 'Reducing Red Tape', were teachers and principals agreed -and here as critics. Despite the policy enactment role similarity, this is equally as dire as the tensions generated by the other reform areas. Aspects of the reform relating to prescriptive administrative, technical and process-based changes were perceived as contributing very substantially to workload, with principals in particular reported as having to work hard to understand and undertake what was being asked of them. We argue that workload increase for principals was an unintended consequence of the reform, echoing Gobby's (2013) findings around the enactment of IPS policy in Western Australia where, although principals took up flexibilities offered by the program, they also experienced a lack of support, excessive administrative burden, limits on their capacity to innovate, and distraction from their role as educational leaders.

Implications for policy
This impact on workload was also related to the nature of the reform in question. The act of 'doing' a devolution-style policy in a market-driven system resulted in confusion and differing local experiences between schools as processes are simultaneously decentralised but centrally-determined, and performative accountability mechanisms are heightened (Gavin and McGrath-Champ, 2017). This reflects Ball et al.'s (2012) contention that policy processes are subject to different interpretations and recontextualisations, perhaps especially the case with autonomy reform. For instance, in enacting the 'Managing Resources' reform area, principals were both narrators and to some extent critics -while they felt greater capacity to handle resources locally, less centralised support was provided and principals felt a lack of clarity of what to report on. This reflects policy enactment as an "ambiguous, messy process" (Maguire et al., 2015, p. 485). To some extent, principals were also critics over new merit selection requirements enacted as part of the 'Staffing Schools' reform area, feeling this process contributed to increased workload and perceived as more time-consuming compared with previous staffing processes. As critics, principals expressed frustration regarding the lack of guidance about implementation of some policy initiatives and pressures applied by policy accountability measures. The findings from our participant interviews, supported with evidence from Auditor-General and departmental reports (CESE, 2020, p. 10;Crawford, 2020), shows the lack of communication and consultation with key stakeholders impacted by these reforms, as well as omission of built-in mechanisms for policy review to iteratively evaluate implementation outcomes.
Based on the data and analysis presented in this article, we suggest several implications for policy, some of which are echoed in the Department's evaluation. It is clear that devolutionary policy is subject to differing enactment at the school level as policy actors (namely principals) are empowered to narrate and be entrepreneurial in their interpretation and translation of the policy within the school context. However, confusion can also arise in this interpretation whereby principals attempt to decipher new flexibility around staffing and financial management, within a context of unclear accountability requirements. To an extent this reflects Lingard's (1996) notion of principals playing an important role in a 'steering at a distance' policy trajectory where they construct their own interpretations in relation to policy implementation while being steered towards accountability. We argue, however that LSLD lacked sufficient guidance on reporting. This was echoed in departmental evaluations which emphasised the complicated and unsuitable accounting and finance processes decentralised to schools without adequate support, producing a lack of capacity-building at the local level (CESE, 2020, p. 12). Thus, evaluation processes are necessary to improve policy and program effectiveness and to ensure accountability for outcomes, alongside the provision of meaningful, effective support for schools. It is important for the Department to take a greater role in providing support to schools to make these local decisions, in order to free schools up to focus on educational leadership and student outcomes, priorities hindered through increased workload demands placed on both principals and teachers. This involves ensuring that processes and systems are fit for purpose before wider-scale implementation, and that school staff are provided with targeted training. We would also emphasise that it seems important to consider the full policy context when implementing policy (Stacey et al., Under Review), given the range of competing demands that principals, in particular, reported managing. Even if policies appear logical individually, the process of implementing a reform initiative can have unintended problems within a complex, multi-layered policy context (Stacey et al., Under Review).

Conclusion
The rollback of LSLD was announced in early 2020, with the NSW Education Minister stating that improvements to LSLD would aim to 'strike the right balance' between autonomy, accountability and support for schools (Henebery, 2020a). Commenting on policy deficiencies, Minister Mitchell acknowledged the government could not keep track of school funding decisions since "we have a policy that totally devolves decision making to each school" (Bolton, 2020). LSLD has since been replaced with the School Success Model, which increases control over schools to ensure greater accountability in how money is spent by schools, tied to particular outcomes (Henebery, 2020b). This is alongside new targets focused on educational attainment, school attendance and student learning outcomes that trigger departmental interventions if not met (Baker, 2020) in response to findings from the Department's evaluation that there had been no substantial improvement, and in fact some decline, across key learning outcomes over the life of LSLD (CESE, 2020, p. 10). While we welcome an emphasis on providing support and clear expectations to schools, we also caution that a context of apparently declining student results may provide fertile ground for a stronger focus on measurable achievement metrics associated with performative forms of accountability (Wilkins et al., 2021), and thus may not address the issues identified in this article around adversarial staff relationships and workload. Whether these matters will also be an issue under the School Success Model, as the successor of LSLD, is yet to be seen. Research into this new devolutionary policy will be needed.
Overall, and alongside other criticisms of devolutionary reform (see Hickey et al., 2021), we question the efficacy of recent decentralisation efforts. While autonomy in pedagogy and curriculum may be desirable (OECD, 2011), increased powers over matters such as administration, resources and staffing appear problematic: exacerbating work demands, inequalities and complex power dynamics within schools. We acknowledge that this study was conducted prior to the full introduction of LSLD, and that its qualitative insights are not generalisable to the whole population of teachers and principals who were impacted. However, it is the qualitative nature of the study that has enabled detailed insight into the experience of this reform, indicating a shifting landscape of interpersonal relations in NSW schools. The lessons learnt from LSLD, and our analysis of it in this article, are relevant not only for future iterations of devolutionary reform in Australia, but also relevant in an international context where autonomy reform is increasingly common (Keddie, 2016).