System leadership as depoliticisation: Reconceptualising educational leadership in a new multi-academy trust

System leadership continues to be constructed largely as a desirable, even normative, evolution of educational leadership, with critiques often focusing on implementation rather than principles. This belies its increasingly recognised role in processes of disintermediation, in which the ‘middle tier’ comprising local government is dismantled. In this article, we draw on interview and observation data from our case study research in a new multi-academy trust to argue that system leadership is better understood as a manifestation of, and mechanism for, depoliticisation. We present a reconceptualisation of system leadership in which its primary function is to enable and operationalise three forms of depoliticisation: governmental, societal and discursive. We conclude that system leadership as depoliticisation is a suite of professional practices with linked identities and dispositions that operationalises the state’s project to depoliticise education in England through multi-academisation, or the creation of multi-academy trusts.


Introduction
In this article, we argue that the contemporarily dominant suite of normative entrepreneurial practices known as system leadership is a manifestation of depoliticisation, in which political issues and decisions are removed from the public sphere where they may be debated. Instead, they are 'presented by politicians and policymakers as matters of technical efficiency rather than normative choices' (Clarke, 2012: 298), and so politics as an exercise of power is re-configured as a matter of individual choice, and thereby privatised . Our reconceptualisation of system leadership underpins our conceptual contribution and re-frames normative accounts of system leadership whose main concern is its effectiveness (e.g. Hopkins, 2007), or its usefulness (Kamp, 2018). Through our new framing, the focus shifts instead to the dissonance between two constructs of system leadership. The first is system leadership as a template upon which educational leaders are encouraged through a dominant policy discourse to base their agency, goals and identities (see Courtney and McGinity, 2020). The second is system leadership as a depoliticisatory mechanism whereby such leaders substitute activities concerning and actualising their agency, goals and identities for democratically contestable educational principles.
Our project contributes to a small body of research that focuses on system leadership rather than system leaders. Our data are drawn from interviews with actors constructed as leaders at all senior levels within a single, new multi-academy trust (MAT), and so complements and builds on important research focusing primarily on the MAT chief executive officer (CEO) (Hughes, 2019;Hughes et al., 2020) or on the relationship between system and other leaders (Cousin, 2019). The MAT structure is presently the site and producer of a form of system leadership most privileged through policy in England, which is a nation recognised as a laboratory and international exporter of education reform (Finkelstein and Grubb, 2000;Gunter, 2015b). The conceptualisation and analysis presented here therefore have implications for policy making and policy scholarship in states internationally where system leadership is being considered or enacted, for example in the USA, where charter schools may be brought and managed together under charter management organisations. To realise these contributions, we make a novel claim. This is that, unlike prior education reforms involving leadership (Gunter, 2012;Hall et al., 2013), system leadership does not just enable or facilitate the state's political objectives: instead, it constitutes the political project itself.

System leadership
The term 'system leader' purposively unites diverse roles in the English 'education system', for example National Leader of Education and National Leader of Governance (see Department for Education, 2020). Its growth in education is reflected in other public-service domains (Mangan and Lawrence-Pietroni, 2019; see e.g. Goss, 2015). We illuminate system leadership through focusing on the MAT, which is the legal entity that engages contractually with the Secretary of State for Education to provide education services across numerous sites (i.e. individual academies, with no discrete legal status). The agglomeration of constituent academies within the MAT comprises a 'system' that is 'led', often by a CEO, or executive headteacher/principal (Hughes, 2019).
Overviews of system leadership's development (e.g. Cousin, 2019;Greany and Higham, 2018;Kamp, 2018;Pont et al., 2008) indicate its international, meta-discursive appeal and facilitative conditions. Through a brief conceptual mapping, we reveal these and also the epistemological positions and associated knowledge claims underpinning system leadership. We differentiate analyses of system leadership through the lenses of functionalist and socially critical knowledge-production domains, following Raffo and Gunter (2008). Whilst not all systemleadership research aligns with this binarised categorisation, it is helpful to think heuristically with (and sometimes against) it. Functionalism is an instrumentalist epistemological disposition and policy approach (i.e. delivery focused; therefore often normative). It constructs complex social phenomena as being susceptible to improvement through removing dysfunctions at individual, institutional or societal level. Socially critical scholarship may be as normative as functionalist, but the basis of its recommendations lies in the amelioration of the effects of power on social actors' experiences, agency and subjectivities -we exemplify this approach in this article. Socially critical research rejects the 'what-works' functionalist assertion that atheoretical or even anti-theoretical research is more credible and useful, and locates knowledge claims in an assumptive architecture whose acknowledgement and elucidation are key to trustworthiness . This conceptual framework enables us to map system leadership and explore its embodiment in new professional roles and identities. Functionalist constructions of system leadership have driven its development through policy and related research. They come from three imbricated sources. The first is the state, or para-statal bodies acting on its behalf in producing policy (e.g. National College for School Leadership: see Gunter, 2012), or supra-statal institutions (e.g. the Organisation for Economic Co-operation and Development (OECD)). The second is from research commissioned for and often published by such bodies (e.g. Hargreaves et al., 2008;Pont et al., 2008). The third source is researchers who reproduce functionalist framings (e.g. Greany and Higham, 2018;Hopkins, 2007;Kamp, 2018). Functionalist constructions of system leadership reify the ontological category of system leader from diverse professional roles in different locations. As an organisational category constructed as concrete a priori, definitions may be unconvincing or circular: 'System leaders' are those head teachers who are willing to shoulder system leadership roles, who care about and work for the success of other schools as well as their own. (Hopkins, 2008: 22) Here, system leadership is a geste philanthropique, an uber-leadership trait in, ironically, heroic individuals. Elsewhere, Hopkins (2007) depicts system leadership as a social responsibility that school leaders ought morally to evince. This exemplifies how functionalism constructs system leadership as a universally normative school-improvement mechanism, through overt promotion (Hargreaves et al., 2008;Hopkins, 2007Hopkins, , 2008Pont et al., 2008) or through focusing on improving, rather than rejecting it (Greany and Higham, 2018;Kamp, 2018). Functionalist scholarship proposes system leadership as a correlative indicator of an education system's health, or value for money (see Kamp, 2018).
Socially critical researchers understand the 'objectives' above as discursive collocations invoked to legitimate certain policy agendas and activities and to superordinate certain actors (Gunter, 2016). The label 'system leader' not only renames actors already in professional roles, but also enables the creation of future, more corporatised roles through discursive inurement.
Adopting 'system leader' as a descriptor implies no taxonomic distinction between a specialist leader of education (see Department for Education, 2019), a headteacher with two schools, or a MAT CEO remunerated with a salary of hundreds of thousands of pounds, thereby immunising the category from connotations of conflicts of interest, lack of transparency and/or higher costs that appear disproportionately in academies (Bubb et al., 2019;Williams, 2017).
Socially critical researchers see the removal of the 'middle tier' of a given education systemdisintermediation -as key to contemporary school-improvement policy. Disintermediation is effected, and somewhat mitigated, through replacing local authorities with 'system leaders', who embody a construction of 'autonomy' that is more discursively than materially observable (Salokangas and Ainscow, 2018), or that consists in what Greany and Higham (2018: 35) call 'coercive autonomy', characterised by increased responsibility rather than freedom, along with more intense accountability. System leaders accepting that label are de facto policy ambassadors, and may be rewarded through empire enlargement. Thus, the most visible, yet most politically undesirable effect of normal market functioning -school closure -is re-configured as an indicator of the acquiring system leader's success (Courtney, 2017).
System leadership may be seen as recognition that inter-school competition has had negative consequences and as an attempt to shift the responsibility for that structural deficiency onto school leaders in return for higher status and salaries: this, of course, is a new verse in an old tune (Gunter, 2012).

Depoliticised times
Over the last 25 years, depoliticisation has provided a useful lens through which researchers have explored the relationship between education and neoliberalism (e.g. Jayasuriya, 2015;Vincent et al., 1996). For Clarke (2012), contemporary neo-liberalism in education disavows its political nature . . . by reframing political issues in economic terms through processes of commodification and by assuming and promoting a broad consensus in relation to this economising agenda -in each case, backgrounding the struggle over values central to both policy and politics . . . [and] undermin[ing] the democratic potential of education . . . . (Clarke, 2012: 298) The state's strategy is to remove democratic, i.e. political engagement through depoliticisation. Wilkins (2016) agrees: 'At the heart of these reforms [to governance] is a ruthless depoliticisation which disguises anti-democratic measures' (Wilkins, 2016: 147). These examples focus on what Wood and Flinders (2014) call 'mode[s] of statecraft ' (2014: 152). They propose additionally three imbricated and interdependent forms: first, governmental depoliticisation, where decisions, functions and activities previously undertaken by governments are instead delegated to arms-length bodies and subjected to bureaucratic and/or technical control. The second is societal depoliticisation, where social issues are moved from the public to the private sphere, and so become matters of individual choice. The third form is discursive depoliticisation, where language shifts issues to 'the "realm of necessity" in which "things just happen" and contingency is absent ' (2014: 165). We use this tri-partite mapping in our analysis, since we agree with Hay (2014) that it is 'broadly persuasive, analytically elegant and eminently useful' (2014: 299). Wood and Flinders' (2014) formulation has been taken up by Gunter (2015a) to explain how Ofsted and the (now defunct) National College for School Leadership exemplify and operationalise governmental depoliticisation; how school choice exemplifies societal depoliticisation; and how targets help create the templated thinking characteristic of discursive depoliticisation. More recently, Gunter (2019a) has adapted this framework to elucidate 'depoliticised privatism' (2019a: 85), which she argues is about the "private" in privatization or how the public has had to learn very quickly to insure against risks that previously were pooled and protected with fellow humans through the state and taxation . . . [and] have had to learn to adapt and be resilient. (Gunter, 2019a: 87) Depoliticisation has been used less extensively in educational leadership. It has been identified predominantly as a context for and producer of diverse professional practices deemed 'educational leadership' (see e.g. Blackmore, 2011;Courtney, 2018;Gunter, 2015a). Gunter (2019b) goes further in arguing that 'those who provide services in the form of schools (trustees, owners, chief executives, principals . . . )' (2019b: 165) operationalise depoliticisation in return for the delegated power immanent in the project and for the status it brings. They are more than simply co-located in a depoliticisatory structure: they are 'integral to the movement' (Gunter 2019b: 165).
In summary, to date, the field has focused on depoliticisation as a technology of governance in education as in other public services. What has not been a focus of scholarly attention is the way in which educational leadership has been marshalled in a new form, or technology -as system leadership -to give form, purpose and context for the project of depoliticisation, and how that is rendered through the accounts and experiences of 'system leaders'.

The case study
This article reports conclusions and analysis from the 'Multi-Academisation and its Leadership' project, a qualitative case study examining how key actors involved in the establishment of a new MAT in England understand its purposes and their own role as system leaders in creating it. The object of our study is MAT system leadership, which we interpret as practice and identities reported or claimed by professionals in positions of authority throughout a MAT. Our interpretation (i.e. that system leadership is not just the property or activities of the executive headteacher/ principal, or CEO) is not intended as support for a distributed model of leadership, but instead enables methodologically our scrutiny of discursively privileged constructions of system leadership as intra-MAT, for example: Working closely with system leaders in the MAT, in partnership with those working in teaching school alliances and beyond, we can demonstrate that the most effective practice in any community can be shared so that new and better practice is introduced into the schools we want to improve quickly. (Department for Education, 2016: 5, emphasis added) We undertook two semi-structured interviews with the MAT CEO, David, and one with each of the post-holders named in Table 2. We also observed a key board meeting soon after the MAT came into legal effect: this meeting focused on purposes and identity. Data were analysed simultaneously by both researchers, adapting a method described by Elliott et al. (2013): the same transcripts were read and annotated individually in the same room in extracts of one to three pages at a time, followed by a discussion and mutually agreed interpretation, which we noted in memos. This process was facilitated by our shared epistemological framing and attention to discourse: we adhere to the tenets of critical policy scholarship, in which understanding of a policy is located in its historical, political and ideological context (Grace, 1995), and questions of power and equity are foregrounded in the analysis. Our analysis produced a series of themes that spoke to depoliticisation, which we mapped against Wood and Flinders' (2014) model.
Our project was a single-site case study of Tonbury and Swain MAT, which is located in a coastal county of England, Lintshire (counties, towns, schools and people are pseudonymised). Details concerning the institutional composition of the MAT are provided in Table 1.
In Table 2, we summarise the research participants and their sometimes multiple roles within the MAT. Stakeholders beyond the case study MAT were not included (e.g. local authority representatives).

Tonbury and Swain: A story of multi-academisation
David led the academisation of Oak Manor in 2011 from a school that 'was an ex-grant maintained, so it had a mindset previously' (David). The CEO is here positioning Oak Manor as agentic and entrepreneurial -we interpret such statements to mean that it was he who had these qualities, but that he is unwilling either to own them or to tell us that he owns them. David's data reveal the school's long-standing problematic relationship with the local authority (LA): 'it went grantmaintained because the LA had taken away its sixth form' before David's arrival. Having converted from grant-maintained to foundation status under New Labour (1997-2010) meant that, for David, academising 'wasn't a big jump for the school . . . because it had been running its own payroll, seeing itself as the employer of the staff, owner of the land, etc.' David locates a key purpose of academisation in enabling schools to become 'more empowered to do what they think is right for their community' -this has persisted in his thinking about multiacademisation. His plan to academise in tandem with the neighbouring school, Skelton High, was initially thwarted by that school's poor Ofsted outcome, which led Lintshire 'to get rid of the head and . . . the chair of governors . . . which meant that the process was scuppered' -for the moment. Skelton High staff were relieved: Nicola reports 'a big, big fear' that Oak Manor were 'just going to come and take us over'.
In an ultimately fruitless attempt at union, a third local school 'added another bit to the narrative' (David). Academised against its staff's will, its fate exemplifies the many 'what-ifs' and near misses of Tonbury and Swain's establishment:  [The school's head and I] had a conversation about it being sponsored by us and I said 'yes fine' . . . because I'd rather all the money stayed in the town, rather than any of your money get sliced off to go and support an HQ elsewhere . . . so he said to me: 'you will become my boss and not somebody else' and I said 'well if you want to see it like that, but we work collaboratively and it would be a strong peer accountability, but yes technically it will be like that. (David) As mentioned, this relationship came to naught: 'the school's governors chose to go with a national chain' (David). This extract is nonetheless significant in demonstrating how David narrates himself as an unwilling leader: his taking the CEO role is inescapable, yet almost inadvertent, and he will do leadership differently once in post. However, the MAT is created finally only thanks to David's ambition and perseverance; there is nothing inadvertent about it.
Then, Rushton Green Special School 'was under threat with regard to its numbers . . . and the head there is a sort of entrepreneurial type of guy . . . and he said, "can I come and join you?"' (David). Ben's account foregrounds Lintshire's role in this, by aiming to reduce Rushton Green's pupil numbers by 16%. Halsby Junior's head spoke about push and pull factors in multi-academisation: There are changes in the educational landscape . . . issues around funding [and] with the performance of local authorities . . . and then being approached more directly about the formation of a local offer . . . This would enable me to address all of those concerns, safeguard our school and then actually not have to join a national chain. (Lucy) Meanwhile, Skelton High came back into the picture through forced academisation: its inclusion as one of the four constituent academies in the Tonbury and Swain MAT was endorsed and arranged by the regional schools commissioner (RSC), but by now welcomed by Nicola, following multiple interim heads within a year: 'it became very clear that this was not some sort of hostile takeover and it was not something that was being done to us'. David's importance throughout multi-academisation and beyond is recognised by most participants: To be mentored, coached, steered from somebody like David is an opportunity most people don't get.
However, there were several near misses: David failed to persuade the governors of a local further education college to join the MAT; Tonbury and Swain's proposal to create a free school failed to win Regional Schools Commissioner (RSC) support; and the head of the infant school that shares a site with Halsby Junior preferred to federate elsewhere. The MAT is consequently more tenuous and its membership more a product of chance and luck than participants often stated. It was very hard won, and this has produced contradictions. For instance, its claimed unique selling point and the foundation of its creation myth is localism, yet its schools are geographically located in Tonbury and Swain, meaning that many more 'local' schools do not belong to it. It is a structure in search of a purpose, which we elucidate below.

System leadership as depoliticisation
We argue in these next sections that system leadership is a mechanism for and manifestation of depoliticisation. System leadership is invoked to advance the politico-structural project of multiacademisation, or the legal, discursive, cultural and organisational processes whereby a group of schools is removed from the local authority and privatised through adopting a MAT structure in a contractual arrangement with the state. Calling these practices 'system leadership' privileges the motives and actions of key individual actors, constructed as 'transformational', over multiacademisation as a state-sponsored political project.
We have structured the reporting of our analysis around Wood and Flinders' (2014) three-part conceptualisation of depoliticisation, comprising governmental, societal and discursive forms.

System leadership as governmental depoliticisation
This is where decisions previously undertaken by governments are delegated to arms-length bodies and subjected to bureaucratic control.
The MAT exemplifies governmental depoliticisation, whose delegatory features are processes that themselves require bureaucratised management. This role is played by the RSC -a new layer in the bureaucracy that supports and enables multi-academisation, and which is neither part of nor accountable to local government. Both the RSC and the MAT itself represent a formalised and intentional recalibration of relationships and responsibility in the planning and delivery of localised education provision. Tonbury and Swain leaders note the political dimension to the RSC's bureaucratic oversight: [At the RSC offices] There's a big map on the board and you've got on one map, you've got . . . all the MATs . . . dotted all over the place, you've got them in various parts of the country and then you've got Lintshire and there's nothing. So, politically, they wanted a dot in Lintshire, so you can see from their agenda they were wanting to make it work as well. (Ben, Rushton Green Special School) Our data reveal that the MAT, bureaucratically overseen by the RSC, is the devolved, or armslength structure that takes on responsibility for making decisions concerning education and delivering them -this illuminates governmental depoliticisation in education provision and governance, as evidenced in David's account, 'I've been a puppet of the RSC to bring about national policy'. Ben's data reveal how he understands himself as a system leader, rather than the state, to be responsible for system reform: I've always been a believer that you can either go outside the system, talk about, debate it and so forth, or you get stuck into the system and try to change if from within.
Like all MATs in a depoliticised landscape, Tonbury and Swain MAT needs to operate bureaucratically in order to fulfil its functions as a para-statal body. The MAT bureaucracy developed by David is centralised, with no substantive powers concerning governance delegated to local governing boards, or committees. Yes, it is a bit of a centralising model, but I think there's a difference between centralising the administration and bureaucracy of the MAT; I think that's slightly different from denying people voice with regard to accountability. (David, MAT CEO) These data show a misalignment between the way in which David frames the MAT and the way in which he has structured it, along with his recognition of the implications of that structuring. David reported that the MAT was purposively established according to democratic and consultative principles: these also informed his decision to academise Oak Manor as a first step to multiacademisation: My position is, whilst there is the agenda of academisation, using it for those purposes of schools being more empowered to do what they think is right for their community, it gave an opportunity for that. (David) We note that the agent in the data extract above is 'schools', implying shared decision-making. However, David's enstructured retention of authority at trust level, rather than any substantive powers being delegated to its constituent academies, means in practice that the CEO makes or strongly influences most decisions. David embodies the MAT, which is the site of bureaucratic control, following Wood and Flinders' (2014) model, even despite David's desire for local empowerment. However, this produces contortions: I think just because you haven't necessarily delegated authority, have you necessarily denied the democracy of decision-making and accountability? (David) David demonstrates here a particular form of Bourdieuian misrecognition (Bourdieu, 2000): he is able to articulate a case for using multi-academisation to increase democracy and empower local professionals, but makes structural decisions that undermine those objectives. David is aware of these tensions: To what extent am I a victim of the regime, whereas I think I'm being clever and taking advantage of it to achieve what I think is right, and I think what it offers you is a critique, how foolish is the guy? And he thinks that he's doing right, but actually he's just caught up in a system where however much he knows his chains, he's actually perpetuating them. (David, MAT CEO) David has structured out the possibility of substantive, meaningful dissent. Concerning vision and governance, for example, the MAT's scheme of delegation permits local governing boards only to appoint and remove chairs of local governing boards; ensure that local vision aligns with the trust's; and approve and review academy action plans and self-evaluations. System leadership here is not distributed, in the sense of power shifting and hence difference being evident: When David's met with the staff and put forward his vision, he's said categorically, you talk to one of these guys [Skelton High's headship team] and it's the same as talking to me. (Nicola) This close alignment is predictable when thinking with governmental depoliticisation, where tight bureaucratic control is an integral feature. Owing to his role in the centralised structure and also to his charisma, David as CEO is fundamental to the MAT's identity, or brand, and to its strategic activities. System leadership is less likely in this iteration but is invoked and deployed to operationalise governmental depoliticisation, along with David's embodied performances as leader. These performances are vital to securing buy-in to multi-academisation from the wider leadership team, as Sarah demonstrates: So, academisation isn't necessarily politically a model that I would have gone 'yeah, I love academisation', but actually, practically, working in it and now we have the opportunities that academisation has brought for us to come together as a range of schools is magic.
We note that multi-academisation, for Sarah, promises a transformation that is magical, reinforcing multi-academisation as a preternaturally and axiomatically beneficial strategy best conceived and operationalised at leader/institutional level, rather than a significant structural re-organisation of provision best planned at (local) government level: You heard Lucy's input on STEM [science, technology, engineering and maths] yesterday -we just wouldn't have had access to that level of expertise and high expectation so closely with such reinforced drive from Lucy for the secondary schools to gain from that experience and expertise in primary without being in that model. (Sarah) Such assertions simultaneously and amnesically overlook the diverse ways in which schools have collaborated over the years (Armstrong and Ainscow, 2018; Keddie, 2014) -indeed, collective professional amnesia appears to be fundamental to making new education reforms work (Gunter, forthcoming), as does derision of the past: Importantly, the contribution expected from board members is business focused, enabling corporatised MAT system leadership: I can help them with that commercial focus and that business acumen and the private sector lens of how to run an organisation as complex as a multi-academy trust. (Paul) I'm an employer. I have difficulties getting employees with the right attitude, the right capabilities. As a consultant, I've done over 150 projects . . . so I've seen what schools are and are not producing, so it was just an opportunity to perhaps add some value. (Roger) Concerns have been raised regarding the extent to which multi-academisation represents or permits 'empire-building' (Courtney, 2017): in a governmentally depoliticised landscape, the MAT takes on responsibility from the state for addressing this, and its administrators and governors, labelled system leaders, grapple with the tensions: Our field notes from that board meeting indicate that no solution was reached to that question of creating a distinctive identity and purpose for the MAT that might supersede or distract from the fundamental issue of privatised merger and acquisition. Indeed, the message that the MAT has a higher purpose than empire building is slower to reach trustees: There's appetite [to expand] and I can tell you from my own perspective of the civil service . . . it's all about empire building, because in the past, and continuing today, remuneration is in line with size. (Roger) In summary, intra-MAT system leadership is marshalled through the advocacy of the MAT CEO as system leader in order to bring about multi-academisation. The MAT represents a parastatal body that does politics on behalf of (and with support from) the state, through making and upholding arrangements for education structures and provision. Further, we note that the good faith and intentions of the protagonists are irrelevant: any role-holder can be replaced by another with different intentions, but the structures will persist.

System leadership as societal depoliticisation
This is where social issues are moved from the political to the private sphere, becoming matters of individual choice and personal decision-making. The data reveal societal depoliticisation in the language used as well as in the themes and issues it describes. For example, recounting the conversations leading up to the formal creation of the MAT, Ben recalls: We wanted to be masters of our own destiny. Oak Manor had got to a point where the regional schools commissioner was kind of talking to them saying 'we would like you to be sponsoring and supporting more schools locally', but they didn't want to be perceived as this sole entity that's just taking over schools either, so me and David had a meeting and had a long chat about lots of things. I'll never forget the phrase he said to me; he turned round and he said: 'you're not gonna leave me at the altar at this point are you?' One of the key constituents of societal depoliticisation is the extent to which the delegation of decision-making concomitantly aligns with the 'decline of issues as salient matters in societal debate' (Wood and Flinders, 2014: 161). In much the same way, the decision to wed (or to jilt) is a matter of private choice; its invocation by one of the executive principals to represent the tensions inherent in making such important decisions about wholesale educational restructuring behind closed doors is significant. In lieu of a process of democratic consultation, the final decision, whose consequences have impacts beyond those presented by these key protagonists, is taken in private.
David was right to be concerned: the present make-up of the MAT is not the one originally planned because some early protagonists changed their mind or could not win their governors' support. This produced a tension that perhaps underpins the present leaders' inability to frame the MAT's distinctiveness and unique purpose. Tonbury and Swain MAT is framed by its leaders as a structure that is almost inevitable, because it represents the best possible answer to the 'problem' of insufficient localism that they decided was the key differentiator from former arrangements under the local authority. But nearby education providers exercised their personal agency in refusing, changing profoundly the most important regional structural education reform in several generations: this, we argue, illuminates societal depoliticisation. A differently constituted MAT nearly happened, one with equal or better claim to the unique selling point of localism that the present iteration is attempting to claim. It was prevented only by individual choices and private decisionmaking at multiple levels (in the private offices of the executive principals, other local and community leaders, the RSC and the Department for Education). As disintermediation implies, absent entirely was the local authority, and significantly marginalised were the local communities for whose benefit so often an idealised construction of localism was invoked. In examples of societal depoliticisation, both the content and location of decision-making matter because ultimately it is about 'shifting issues off the agenda of public deliberation' (Wood and Flinders, 2014: 152).
In order to achieve this shift away from the public and into the private sphere, societal depoliticisation commonly manifests through appeals to personal conviction (rather than a political mandate) about what is 'right': To consider creating the first home-grown multi-academy trust in the county was quite a challenge . . . We knew it was the right thing to do, but how would we strategically go about it? (Ben, emphasis added) Societal depoliticisation also commonly appeals to the family. Its invocation or deployment represents a form of privatisation, where a family makes closed and private decisions in a closed, private unit . We saw this in our interview with Lucy, in which she framed decisions, expertise and motivations through a family, rather than professional, lens: This conviction cannot be counted on to exist or matter beyond this instantiation. Where the public is deprivileged, there is no automatic expectation of debating the advantages and disadvantages of these arrangements in a politically open and democratic forum or manner, with a reasonable expectation of, or even possibility of, wider change.

System leadership as discursive depoliticisation
This form draws attention to the ways in which language shifts issues to 'the "realm of necessity", in which "things just happen and contingency is absent"' (Wood and Flinders, 2014: 165) . Here, they are constructed as mere elements of fate, although this whimsy may conceal purpose. Our data reveal multiple instances where system leaders attempt to change social reality through manipulating the language used to describe it.
For example, the concept and language of 'system leadership' itself is intended to function discursively, by conjuring an illusory distributed power structure. For instance, Ben and Lucy's formal titles were enhanced on multi-academisation, from head/principal (of their school) to executive principal (of the MAT). Unlike typically, the inclusion of the word 'executive' in their job title does not mean that the holder has the CEO role. Its use is justified because of Ben and Lucy's role on the MAT executive board, though they are neither trustees nor members. In fact, through multi-academisation, Ben and Lucy have arguably experienced a diminution in substantive autonomy, recompensed by the new title, although each sees it differently. Ben stresses that, despite David's role as CEO, he, Lucy and Ben are in fact equals. Lucy comments on how she has lost autonomy but gained 'responsibility'. The MAT's centralised organisational structure belies these interpretations, so the deployment of 'executive' allows each of these heads to claim hierarchical and strategic oversight of the MAT exceeding their previous, school-focused domains, yet real power resides in the CEO. David supports this interpretation: It's been important to have these school leaders to feel a sense of parity and esteem and all be on the executive and be called executive head, because my role within that . . . call it a sort of contradiction, but there is an element of first amongst equals, but the reality is I am the accounting officer of the whole lot.
The whimsy of a job-title change conceals the purpose of disguising how power works. In a further example of discursive depoliticisation, David explained how he and his team use language to capitalise on localism to build the MAT's brand and justify its existence: It's about changing the language: rather than talking about the youngsters in terms of the school they attend, we talk about the youngsters in terms of the town where they live. So we are all responsible for the education of the youngsters in this town. (David) Here, through their use of language, the MAT's leaders are staking a territorial claim to all pupils in the town as a matter of necessity and without debate. This is the MAT as colonial presence, planting its branded flag in a land constructed as available because it is public and so is not yet owned. Individual families' decisions about which school to attend and hence how to identify are not sought: families' choices are paramount, except where they are wrong. The family is invoked to serve the interests of the state, acting through its delegated authority, the MAT. Unsuspecting inhabitants are thereby converted semantically into subjects: this is concealed behind claims to be taking responsibility for them out of altruism.
In discursive depoliticisation, language is used to trivialise political issues and impede substantive debate and scrutiny. In our data, Lucy described the process and motivation for academising and joining the MAT in the following way: The landscape enabled us to be part of something really rather fabulous, to perhaps extend our offer elsewhere in our town. (Lucy) We note first the way in which the meat of the sentence -the acknowledgement of empire building as an objective -is tucked away behind the debate-stopping word 'fabulous'. But there is more here: systems and structures in education are being realigned and reimagined fundamentally. However, this process is not being conducted in the political arena, but instead it is depoliticised, its success dependent upon system leaders such as Lucy. Her role is not to win the argument, since the state's national multi-academisation agenda will 'just happen' anyway (Wood and Flinders, 2014: 165) -is fated to happen -yet won't happen without system leaders to imagine, operationalise and want, as Lucy says, 'to be part of' it. Gunter (2012) identified the way in which school leaders were vital to New Labour's reform agenda. We argue that this present phenomenon goes further in responsibilising leaders for creating new structures that exceed the limits of the unitary school: unlike in the New Labour government, this responsibilisation is not just facilitative of or even integral to the state's reform agenda, but constitutive.

Conclusion
In this article, we have argued for the first time that system leadership is a mechanism for depoliticisation and have provided empirical support and analysis through the case of leadership in a MAT. Multi-academisation represents a significant political intervention into the restructuring of compulsory education provision in England, yet the associated risks are high. Our data show that system leaders are taking on responsibility not only for this restructuring, but also for devising retrospectively a rationale and set of values for it, and for accepting the risks involved. The ambiguities and gaps in what should be policy are delegated to these system leaders. Just as under New Labour (Gunter, 2012) and before (Grace, 1995), contemporary educational leaders may be agentic believers in and effective operationalisers of reform, here, multi-academisation as a structural project. They may even think, like David, that their version will somehow be more democratic and/or have more noble objectives than other attempts. It is certainly true that our research participants were largely pleased with the MAT and with multi-academisation. However, the main issue regarding MATs is both the structure -it is contractualised, privatised and re-bureaucratised in a less transparent and accountable manner, with little possibility that high motives either interrupt or outlast this structure -and the way in which professionals as MAT system leaders are responsibilised for creating the structure and making it work. This constitutes a significant privatisation.
What is required is repeated public failures of the present model, or, should this prove too challenging a prospect for those working and learning in schools, sufficient failure to be represented as such meaningfully to the public (just as comprehensive education did not fail, but was successfully represented as failing to justify its erosion). This is because such failures of depoliticisation may in turn prompt the demos to demand more control over the issues that affect them. Flinders and Wood (2015) argue that there are forces inherent within what they call hyperdemocracy and hyper-depoliticisation that ultimately produce the other. They suggest that increasing the power of unaccountable, para-statal institutions and removing the mechanisms by which citizens may participate in decision-making processes erodes support for depoliticisation and builds afresh the case for democratic forms of governance and of public life. Evidence of such failure is certainly building (see e.g. Adams and Barr, 2018;Sodha, 2018), but has not yet provoked a re-politicisation of education policy enactment. Politics globally has become particularly challenging and consequential, and so, concomitantly, polities are persisting with depoliticisation as a protective strategy and with system leadership as its mechanism in public services. This means that socially critical analyses of system leadership are required to shift the focus from a 'what works' agenda that may enable depoliticisation and corporatisation, to one that re-focuses the field on the implications of system leadership for a public education whose principles are democratically contestable.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by the University of Manchester's School of Environment, Education and Development.