(Re)moving exclusions: School exclusion reduction in Glasgow and London

School exclusion reduction in Scotland—and especially in the city of Glasgow—has received substantial media and policy attention in recent years. In London in particular, multiple governmental agencies have explicitly expressed a desire to replicate the exclusion reduction which recently occurred in Glasgow, often citing the connection between school exclusion and violence as a key motivating factor. In this paper, after presenting the statistical trends in school exclusions in Scotland, England, Glasgow and London, we mo-bilise original interview data to (1) explain how school exclusion reduction occurred so rapidly in Glasgow between 2007 and 2019, and (2) explore whether a similar reduction in exclusions could occur in contemporary London. We apply a theoretical framework to these issues which derives from Peters’ work on policy coordination, allowing us to compare the conditions in Glasgow and London for well-coordinated pan-city exclusion reduction. Building on previous research which has contrasted national school exclusion policies in Scotland and England, we conclude that policy conditions surrounding school exclusion in the two cities differ substantially. There are substantial barriers to significant exclusion reduction in London, relating to both city-and national-level factors


INTRODUCTION
'Bring down school exclusions, bring down violence' is the striking headline of a recent article on the Mayor of London's website (Mayor of London, n.d.).The banner image features Lib Peck, Director of the London Violence Reduction Unit (LVRU), alongside Maureen McKenna, Glasgow Council's former Director of Education.The contention of the piece is unequivocal: Glasgow has achieved a remarkable reduction in violence over the past 20 years, this was intimately connected with its sharp drop in school exclusions, and so London should follow suit.
The LVRU is dedicating substantial resources to this inter-city emulation, and this call to replicate Glasgow's exclusion reduction has been echoed by a number of other London agencies.In 2020, a London newspaper partnered with the London Community Foundation to launch a £1 million campaign to reduce exclusions in the capital, explicitly inspired by Glasgow's example (London Community Foundation, 2020).Lewisham and Southwark Councils have each produced reports in the last few years (London Borough of Lewisham, 2018;London Borough of Southwark, 2020) highlighting Glasgow as an instructive case study of exclusion reduction-the latter directly stressing 'the relevance of lessons from Glasgow' (London Borough of Southwark, 2020, p. 20).
These empirical claims and political calls touch upon deeply consequential issues in educational research and policy, including the ethics and consequences of school exclusions and the mechanisms through which to best achieve change in schools' practice.While touching on those issues, this paper primarily focuses on the less well-examined question of educational policy transfer between the cities: how might it be possible for London to replicate Glasgow's exclusion reduction-if at all?To address this question, we undertake a comparative analysis of educational policy conditions in the two cities, examining the institutional, political and ideological factors which shape their schools' practice.Our analysis is based on original interview data from relevant policymakers and practitioners

Key insights
What is the main issue that the paper addresses?
The main issue addressed by the paper is school exclusion reduction, and-more specifically-the capacity for policymakers in London to replicate Glasgow's success in rapidly reducing school exclusions.

What are the main insights that the paper provides?
Approaching school exclusion reduction as a policy coordination problem can be a fruitful form of analysis.Effective mobilisation of policy networks, hierarchical power and reframing all made a significant difference to school exclusion reduction in Glasgow.Policy conditions for exclusion reduction in London differ substantially from those in Glasgow in Glasgow and London, as well as being informed by a quantitative analysis of recent exclusion rates.
We conclude that conditions for coordinated exclusion reduction in London today differ significantly from those which were present during Glasgow's recent exclusion drop.In particular, we outline consequential differences in the nature of relevant policy networks, educational governance arrangements and ideological environment in the two cities.These differences mean that coordinating agencies to achieve exclusion reduction across present-day London would be a profoundly complex endeavour, particularly owing to the involvement it would require from national power-holders.Alongside these empirical arguments, we advance two theoretical points: first, that policy coordination concepts (especially Peters, 2013Peters, , 2018) can be of significant value for understanding and comparing educational change in different places; and second, that educational policy transfer can be examined at the city level-and that perhaps the city has been somewhat neglected as a (potential) unit for change in educational policy transfer literature.
In the remainder of this introduction, we lay the ground for comparing education policymaking in Glasgow and London.After first surveying recent literature on exclusions, including research comparing Scottish and English approaches to the issue, we present quantitative data on school exclusion rates in Scotland, England, Glasgow and London.A note on our methodology and conceptual framework then directly precedes our interview data analysis, which focuses on comparing educational policy conditions in Glasgow and London.

School exclusions-Contentions and consequences
School exclusions are a source of national consternation in Britain.In government, civil society and academia, a flurry of publications over the past decade have called into question how, why and with what consequences school students have been removed from mainstream education.
In July 2018, the House of Commons Education Committee published a report with an unambiguous title: 'Forgotten children: alternative provision and the scandal of ever increasing exclusions' (HM Government, 2018b).This 'scandal' prompted the Department for Education (DfE) to order a review, chaired by James Timpson and published in May 2019, to improve practice in exclusions (HM Government, 2019).The Timpson Review highlighted a range of problems in school exclusion practice, but stopped short of directly calling for exclusion reduction, concluding instead that 'there is no optimum rate of exclusions-exclusion rates must be considered in the context in which the decisions to exclude are made' (HM Government, 2019, p. 5; see also Done & Knowler, 2022).Third-sector organisations have been more strident: the legal charity JUSTICE concluded that school exclusions are procedurally unjust in England (JUSTICE, 2019), for instance, and an Institute for Race Relations report (Perera, 2020) suggested that the stark racial disproportionality of exclusions is evidence of a systemically racist education system (see also Harrison, 2020;Joseph, 2020;Joseph-Salisbury, 2020;Kulz, 2017;Parsons, 2008).Alongside this racial disparity, it is well established that exclusions also disproportionally affect students with additional educational needs and students entitled to Free School Meals, raising questions about the role of school exclusion in deepening inequalities (Billingham & Irwin-Rogers, 2022;Black, 2022;Graham et al., 2019).
The urgency surrounding this issue is grounded largely in concerns for the effects that exclusion can have on excluded young people.Excluded students tend to acquire fewer and lower level academic qualifications (see e.g.Children's Commissioner for England, 2019; Gill et al., 2017;HM Government, 2019), and it has also been demonstrated that exclusions have a significantly detrimental, independent effect on mental health (Ford et al., 2018).
Ethnographic research has vividly conveyed this: Irwin-Rogers (2019, p. 607), for instance, describes how one student's exclusion 'ripped into his already damaged sense of self-worth'.
The specific claim emanating from the Mayor of London's office-that a direct connection can be drawn between school exclusions and violence-has also received substantial attention.Studies have concluded that, as well as causing psychological difficulties, exclusions can deepen societal marginalisation (e.g.Briggs, 2010;McCrystal et al., 2007) and heighten vulnerability to criminal exploitation (Commission on Young Lives, 2022; HM Government, 2018a), which can in turn entail increased risk of violence perpetration or victimisation (see Arnez & Condry, 2021;Irwin-Rogers et al., 2020).More generally, links have been established between the experience of exclusion and later criminality, with one recent study finding that, in their sample, permanent exclusion increased probability of custody by 33 percentage points (Cathro et al., 2023).Forty-two per cent of the English and Welsh adult prison population were permanently excluded from school-compared with less than 1% in the general population (Prison Reform Trust, 2022).Research has suggested that exclusion can be a 'critical moment' (McAra & McVie, 2010) or a 'tipping point' (Holt, 2011), which substantially heightens a young person's likelihood of engaging in criminality, including violence.Although none of these studies draw a simple causal link between the two, they do appear to substantiate the view that, among other adversities, school exclusion can play a role in the precipitation of violence.
Alongside this array of analyses, recent media coverage of school exclusions has also frequently referenced an apparent success story: the steep drop in exclusions in the city of Glasgow since 2007.Having previously excluded students 'habitually', the story goes, the city shifted its approach and reduced its exclusions considerably (e.g.Holden, 2021;Whyte, 2021).The comparison with English cities is often made in these pieces: why is exclusion practised so differently on either side of the border?

School exclusions in Scotland and England: Differing policy frameworks, differing rates and divergences in city-level change
As education is a devolved area of policy, Scotland and England have different approaches to exclusion.Although our primary focus in this paper is comparing exclusions at city level in Glasgow and London, practices in the two cities are of course influenced significantly by the national policy frameworks within which they operate.Recent studies comparing these frameworks have stressed the relative punitiveness of English exclusions policy.McCluskey et al. (2019), for instance, argue that policy documents governing English practice are 'much more punitive in tone' than their Scottish equivalents, containing 'no discussion of the value or effectiveness of alternatives to exclusion' (McCluskey et al., 2019, p. 10; see also Parsons, 2008;Tawell & McCluskey, 2022).As Tawell and McCluskey (2022, p. 5) put it, English documents tend to frame exclusion as a 'legitimate sanction', whereas Scottish guidance frames it as 'an undesirable outcome'.This is substantiated by the wording of the most recent guidance for England (HM Government, 2022), which states that exclusion is 'sometimes a necessary part of a functioning system'.There may be a longestablished trend here: respondents quoted by Cole et al. (2019, p. 385) suggested that, ever since the forming of the Coalition Government in 2010, England's education policies appeared to condone rather than condemn 'exclusionary practices'.
These policy contrasts are reflected in the two nations' rates of exclusion in recent years.Since at least 2002/03-the earliest year for which Scottish data is publicly available-permanent exclusion rates in England have been substantially higher than those in Scotland (see Figure 1).While the two countries' rates of all exclusions (including temporary exclusions as well as permanent) are more similar, the English rate has been higher since 2008/2009, and there has been a notable bifurcation since 2012/2013: while the Scottish rate has continued a steady decline, the English rate has been increasing since that year (see Figure 2).It has been suggested that these sustained drops in exclusion across Scotland, particularly since around 2006, were prompted by changes to national exclusion guidance (McCluskey et al., 2019, p. 8) and by an increasingly flexible and inclusive approach to curricula (Cole et al., 2019, p. 386).Respondents in our study also cited the importance of national policy measures beyond education, such as the Getting It Right For Every Child (Scottish Executive, 2006) framework and the Preventing Offending Strategy (Scottish Government, 2015).
Perhaps most striking in this data, however, is Glasgow's exclusion reduction: Glasgow's drop in exclusion rate is significantly steeper than Scotland's national reduction (albeit from a substantially higher starting point), and also differs markedly from London's trend, which is gentle decline followed by plateau in the case of permanent exclusion, and gentle decline followed by an increase when all exclusions are included (see Figures 1 and 2).The difference between Glasgow and Scotland's rates of exclusion is far greater than the difference between England's and London's, which is remarkable given the size of the two cities in relation to their countries: Glasgow makes up around 30% of Scotland's population (1.7 million out of 5.5 million), while London makes up just 16% of England's population (8.9 million out of 55.9 million).The extent and rapidity of Glasgow's exclusion reduction thus warrants academic examination.
F I G U R E 1 Rates of permanent exclusion per 100 students in Scotland, England, Glasgow and London, 2002-2019.After 2010-2011, these figures were collected once every two academic years in Scotland (and in Glasgow).The same data points are included for England and London from 2010 to 2011.We have not included data for 2019/2020 onwards owing to the substantial effects of the COVID-19 pandemic on exclusion rates.Sources: Department for Education (https:// explo re-educa tion-stati stics.servi ce.gov.uk/ find-stati stics/ perma nent-and-fixed -perio d-exclu sions -in-england), Scottish Government (https:// www.gov.scot/ publi catio ns/ schoo l-exclu sion-stati stics/ ) and Glasgow Council (direct correspondence).
There are important caveats here.Exclusions are obviously not the only measure of success in an education system, and there have been significant concerns raised about other aspects of Scottish education in recent years, including its curriculum and exam regime (see Carrell, 2021), and violence within schools (Mackay, 2023).Alongside this, the school exclusions which do occur in Scotland continue to affect particular groups with significant disproportionality, remaining higher for those with additional support needs (×5), and those from deprived communities (×4), for instance (Scottish Government, 2022).There is also danger in over-relying on official statistics-it could be that informal exclusionary practices, such as 'internal exclusions' or illegal 'off-rolling', are occurring without being recorded (see Done, 2022;Done & Knowler, 2021;Power & Taylor, 2020;Whitehouse, 2022), and this issue could be exacerbated by a strong policy push to reduce exclusions.Cole et al. (2019, p. 387) suggest, however, that there is no evidence to indicate that these practices are any more common in Scotland than elsewhere, or that official statistics in Scotland are any less reliable than elsewhere.
Despite these caveats, the statistical trends presented here substantiate the value of examining school exclusion reduction in both Glasgow and London.More specifically, there is a need to compare policy conditions in Glasgow over the past two decades with policy conditions in London now: those London agencies cited at the outset of this paper are calling for London's exclusion reduction today, during the current decade (2020s), to replicate the sharp exclusion reduction which occurred in Glasgow between 2006/2007 and 2018/2019.In the analysis which follows, then, we focus on contrasting educational change in Glasgow between 2006 and 2019 with policy conditions in present-day London.In assessing prospects for educational policy transfer between cities, we are moving away from the focus on national and trans-national entities which predominates in existing education policy transfer literature (see e.g.Ball, 2012;Ball et al., 2017;Forestier & Crossley, 2015;Johnson, 2006).First, we briefly outline our methodology and conceptual framework.

METHODOLOGY AND CONCEPTUAL FR AME WORK
This paper draws from Public Health, Youth, and Violence Reduction, a 3 year ESRC-funded project, seeking to develop a stronger evidence base for 'what worked' in Scottish violence reduction, and tracking the evolution of public health approaches to violence reduction in England.The project utilises a range of qualitative and quantitative methods, including interviews with senior stakeholders, community fieldwork and statistical analysis.Data from the wider project includes interview and focus group material from 190 participants across the UK, as well as secondary data analysis, documentary analysis and participant observation.Data collection has focused on the development and movement of violence reduction policies in the UK, including relating to education.
This paper is based on a subsample of interview data dealing directly with the relationship between school exclusions and violence reduction, including participants drawn from politics, civil service, education and the charity sector in both locations of the study.The subsample in this paper was selected for its direct experience of and insight into education policymaking and practice, totalling 23 interviews across both jurisdictions.Semi-structured interviews were conducted owing to the opportunity they present for in-depth qualitative insight into complex social issues, and for flexibly pursuing multiple lines of inquiry.The interviews were conducted primarily online, lasting 35-90 min, focusing on relevant policy and practice change.All interviews were undertaken on a confidential basis with respondents offered anonymity, and were professionally transcribed.
Interview data was thematically analysed (Braun & Clarke, 2006), firstly in its entirety as phase 1 data.The coding resulted in themes such as education, schooling and exclusion.Extracting and reviewing this subset of data highlighted the analytical potential of Peters' policy coordination theory.Thus in a second stage of analysis, this subset of data was coded using concepts from Peters' conceptual framework.The interview data analysis was complemented by a review of academic and policy literature, with particular focus on literature comparing educational systems and practices in England and Scotland.
In addition, quantitative analysis of school exclusion data was undertaken, so that our interview data could be contextualised by statistical trends in exclusion in each jurisdiction.

Conceptual framework: Comparing conditions for coordinated school exclusion reduction
Our comparison of educational policy conditions in Glasgow and London applies a conceptual framework drawn from Peters' (2013Peters' ( , 2018) ) work on policy coordination.Reducing school exclusions throughout a city requires well-coordinated practice across schools, as well as relevant adjacent agencies which support school-age children.Thus, comparing conditions for pan-city exclusion reduction in Glasgow and London necessarily involves contrasting the extent and nature of educational policy coordination in the two cities.Based on empirical studies, Peters ' (2013, 2018) identifies the most significant factors that enable or hamper coordination between governmental bodies.Of particular relevance here, Peters highlights the importance of three interrelated factors: 1. Networks-for different agencies to act in a manner which aligns and complements one another's work, they need to be well connected, in frequent communication and sharing similar ideas, through formal or informal networks (Peters, 2018, p. 5).
Networks can be both 'horizontal'-involving connections across different organisations and individuals at similar levels of seniority-and 'vertical', involving connections between organisations and individuals with different levels of seniority.Networks are often given particular drive and energy by 'boundary-spanning' policy entrepreneurs (see also Mintrom & Norman, 2009 on the importance of policy entrepreneurship for policy change).2. Hierarchy-Peters argues that 'the best candidate for the fundamental instrument producing enhanced coordination would be power' (Peters, 2013, p. 577).Coordination is enhanced when there is a 'hierarchical coordinator' who has a 'repertoire of instruments' at their disposal (Peters, 2013(Peters, , 2018, p. 574, 6), p. 574, 6).Senior decision-makers are well positioned to coordinate policies and practices through the mobilisation of different 'instruments' of power.3. Reframing-policy coordination relies upon different agencies having similar ideas and a common form of problematisation, a similar understanding of the urgency, severity and causes of policy problems (Peters, 2018, p. 6).Thus, for there to be well-coordinated policy change, there needs to be a common 'reframing' of issues-different agencies need to be aligned in how they rethink and reconsider a problem (Peters, 2018; see also Rein & Schön, 1996;Laybourn-Langton et al., 2021).
Mobilising Peters' theory of policy coordination as a conceptual framework enables a novel comparative study of conditions for educational policy transfer at the city level in Glasgow and London.By examining the significance of these three factors within each city, we are able to explore the means through which well-coordinated exclusion reduction occurred in Glasgow, and to assess prospects for a similar reduction in London.In so doing, we build upon existing research which has compared national-level educational policy environments in Scotland and England.

Comparing conditions for coordinated school exclusion reduction in Glasgow and London
In this section, we mobilise interview data to analyse the sharp decline in Glasgow school exclusions between 2006 and 2019, and to assess the potential for a similarly well-coordinated reduction to occur in today's London: could recent exclusion reduction policy and practice from Glasgow be successfully transferred into present-day London?For each city, we examine the three key enabling factors for policy coordination identified by Peters (2013Peters ( , 2018)): networks, hierarchy and reframing.

NET WORKS Glasgow
For Peters (2018, p. 5), well-functioning policy networks can help to prevent significant divergences in thinking and practice between organisations, and can-in a more positive sense-drive diverse agencies to ensure their work is aligned and complementary.From our interviews, it is apparent that both national-and city-level policy networks played a significant role in the coordinated exclusion reduction which occurred in Glasgow in the 2000s this whole range of voices were saying the same thing and repeating the same message and backing each other up in all sorts of ways … There was a period of time in which we just used each other's names and we said the same thing on the platform again and again This influential multi-agency national movement was about reshaping perspectives on young people, including their behaviour in education.During the same period, there were particularly tight-knit networks within Glasgow-respondents described it as having 'a citywide theory of change' in which 'everybody was on the same page'; and a 'public collective which essentially convened health, education and the police'.Driven in part by the Scottish Violence Reduction Unit's focus on preventing violence, which focused particularly on Glasgow at this time, individuals and organisations across public services recognised 'we're part of it all'.This quote, from a Glasgow headteacher, was echoed by other respondents from within Glasgow schools.Inspired especially by connection with professionals from policing and justice, there was a sense of responsibility among educationalists to contribute to the wider mission of supporting young people's safety and wellbeing-feeding into the drive to reduce exclusions.There was thus substantial 'horizontal coordination' (Peters, 2018, p. 2) between relevant policy actors both nationally and across Glasgow.
This horizontal coordination was matched by a vertical form, described by the academic as 'three tiers of voices': a 'top tier' of those in 'respected posts' as senior power-holders; a 'middle tier' of influential professionals and campaigners with a wide reach; and then, thirdly, 'a whole lot of other people that will be much less known, but who are bringing practice change'.This vertical alignment was important for bringing about change, in line with Peters' (2013Peters' ( , 2018) ) findings, which demonstrate the need for both bottom-up and top-down networking.This multi-directional connection and collaboration-which gained particular momentum within education and adjacent sectors such as Early Years and justice-appears to have been key to policy movements on violence reduction and adverse childhood experiences (ACEs), among others, which cumulatively created conducive conditions for school exclusion reduction in Glasgow.These policy networks effectively produced 'a movement with enough force' (Berglund et al., 2022, p. 311) to bring about substantial change in how young people were perceived and responded to, including in schools (see Zeedyk, 2021).
Particularly influential individuals within these networks mobilised a sense of crisis and urgency.Across the interviews, there was a tendency for specific names to re-appear, such as John Carnochan and Karyn McCluskey (senior figures in the Scottish Violence Reduction unit between 2005 and 2018).They acted as 'boundary-spanning' policy entrepreneurs (Peters, 2013, p. 579) who traversed multiple realms with the public sector, connecting different organisations with similar ideas.Their influence within policy networks galvanised a drive for change: the state of children and young people's lives, including their educational experiences, began to be seen not just as a condition, but as a problem-a prerequisite for any policy change (Kingdon, 1993, p. 42).This sense of crisis was widespread within Glasgow in the 2000s and 2010s: a grassroots community leader told us that 'young people were struggling … completely fell through the net'; a police officer that 'schools had just lost their way'; a teacher that 'we just kept on doing the same thing without getting results'; and, lastly, a senior figure from council said that 'exclusions were just ridiculous'.

London
Relevant policy networks in London appear to be more diffuse and less close-knit than in Glasgow.As one Scottish civil servant put it, policy circles in Scotland are 'really tight', but within England you tend to find 'layers within layers', and there are particular coherence challenges in London-its size and complexity were described as 'momentous'.
London respondents spoke about shared energy behind the drive for school exclusion reduction, but a relative lack of well-coordinated messaging and collaboration across networks.One LVRU staff member, for instance, described 'growing momentum … collective momentum' behind making schools more inclusive and nurturing places, particularly among local authorities, but there were doubts among some respondents about the potency of this networking-'cross-borough discussion' is limited, and there is a lack of 'citywide theory of change'.Greater coherence is needed: 'how do we all start to say the same thing and work towards the same thing?'Where in Glasgow, and Scotland more broadly, it appears that tight-knit networks of policy actors really were 'saying the same thing' around nurture and educational inclusion, networks of relevant London agencies are struggling to achieve that degree of collective alignment.
The capacity for relevant policy networks in London to cohere around a common vision for change appears to be undermined by a key barrier to coordination identified by Peters: competition.Where agencies are in competition, there is a tendency for them to 'want to defend their budgets, personnel and policies, and fear that coordination with other organisations will endanger their "turf"' (Peters, 2018, pp. 5, 9).These issues were evident from a range of our London respondents.One education charity manager described 'stakeholders [in the education landscape] vying for their position from agendas that are not compatible with one another', while a LVRU staff member said: 'Everybody has an answer.Everybody wants to tell you that they can build Rome in a day … that I did it, I'm the one who did it'.Funders such as the LVRU and the London Community Foundation have dedicated substantial resources to exclusion reduction initiatives-as mentioned in the Introduction-but multiple London respondents suggested that competitive grant-making processes can heighten organisations' sense of 'turf'.Although there is grant-giving within Scottish education, through the Scottish Attainment Challenge for instance (see Scottish Government, 2023), which may create competition between service providers, our Scottish respondents did not highlight tensions over 'turf' as a significant factor affecting the overall aim of collectively supporting inclusion.
The need for vertical policy coordination in London was clear from our data: as an education charity manager said, 'there needs to be a top-down direction as well' if coordinated exclusion reduction is to be achieved.The capacity for vertical coordination in London is undermined, however by pan-London policymakers' lack of power over education, as discussed at length in the following section, in which we focus on issues of hierarchy and power in the Glasgow and London education system.

Glasgow
As the mention of a 'top tier' in Glasgow policy networks suggests, there was a key ingredient which helped to enable coordinated policy change: power.Peters highlights the central role played by 'hierarchical coordinators' who have a 'repertoire of instruments' at their disposal to enable coordination (Peters, 2013(Peters, , 2018, pp. 574, 6), pp. 574, 6).In Glasgow-unlike London-substantial educational power sits at city level: all the city's state schools are run by Glasgow Council.It is perhaps unsurprising, then, that the most important 'hierarchical coordinator' driving exclusion reduction was Glasgow Council's then-Director of Education, Maureen McKenna.It is noticeable that the year McKenna took up this role, 2007, is the same year that exclusions began to plummet in the city, and more acutely than the decline across Scotland (see Figures 1 and 2).A Glasgow headteacher cited a range of 'instruments' which McKenna deployed to achieve change.She oversaw the city's educational data, and used this to influence practice: as one Glasgow teacher said, 'she showed us a lot of statistics about what happens to children and the data for long term outcomes'.She used her formal authority to make sure exclusions were reduced-another teacher interviewee stated frankly: 'there wasn't really a choice, it was very much dictated by authority, you can't exclude'.In addition, new bureaucratic processes and 'paperwork' were brought in which 'made it very difficult to exclude a child'.
Alongside this multi-faceted mobilisation of 'hard' forms of power, it appears evident that McKenna was also able to adeptly deploy soft power, skilfully influencing the ideas and practices of headteachers: McKenna was a schoolteacher … she knows how to speak to school headteachers … she knows the challenges that they will put forth and she knows how to address it in a way that's not condescending.(Education charity manager) In policymaking, as Lowndes and McCaughie (2013, p. 545) put it, 'personality and passion and individual qualities matter'.It is perhaps little wonder that the LVRU so frequently references McKenna's work specifically when talking about exclusion reduction in Glasgow, as it appears justified for her to be personally credited with a considerable role.

London
In direct contrast, London respondents discussed the lack of pan-London powers over education.There was recognition among interviewees that the Mayor's Office (including the LVRU) can initiate 'programmes to do ground-up work', but they lack power to initiate 'topdown' change: the Mayor has no powers over schooling in the city.As one staff member from the LVRU put it, Mayoral bodies can 'advise' and 'advocate', and 'build relationships with head teachers or councils', but cannot 'dictate'.A London headteacher said that there was no pan-London agency which had significant influence over their running of the school, and a Director for Children's Services highlighted that no school nor multi-academy trust is held to account for its exclusion rates by a pan-London agency.
Power is diffuse in London education, rather than concentrated in an agency or agencies which could effectively coordinate practice, either across the city or even within individual borough local authorities.Seventy-three percent of London's secondary schools are now academies, which are not run by local authorities (Local Government Inform, 2022).Many of these academy schools are part of multi-academy trusts, which span multiple boroughs, or even multiple cities, thus affecting local coordination between schools.One education charity manager described the stark difference between the English and Scottish education systems in this regard: It's a lot easier to have that cohesive feel to education when you just have schools maintained by a Local Authority [as in Scotland] … Where it is far more fragmented in England … you've got so many different types of schools, free schools, academies, studio schools.And not just in terms of all of the different types of schools, but where the funding comes from.This sentiment was echoed by the London headteacher, who suggested that the number of different agencies running schools in London may be in the thousands.The fragmentation of English education in contrast to the Scottish system has long been observed: Arnott and Menter (2007, p. 262) wrote of the 'increasingly fragmented schooling system' in England, tying this to 'marketisation and privatisation', and suggested, by comparison, that the Scottish system is 'relatively homogeneous'.The relative diffusion of power within London education is a significant barrier to effective policy coordination.
Education in London is shaped substantially by national agencies which hold significant shaping power over schools' practice across the country.Alongside the relative punitiveness of national exclusions policy guidance in England, discussed in the Introduction, our respondents also highlighted the effects of more specific messages and methods deriving from two particularly powerful bodies: the DfE and the Office for Standards in Education (Ofsted).Two of our respondents who had interacted with the DfE on educational policy suggested that 'preventing school exclusions didn't sit well with [them]', and that recent governments' approaches to education 'do not support inclusion and nurture'.This is particularly consequential for London (and other cities) given the relative lack of educational power in the city as compared with the DfE-as one education charity manager put it, every English school 'has to follow very particular messaging from the DfE'.The power of Ofsted as the school inspectorate was emphasised by many of our respondents-it was described as the body with most potential to be 'the agency for change' in education.The London headteacher we interviewed, however, stressed that current Ofsted practice did not encourage inclusion: 'the latest Ofsted framework doesn't give you much credit for the type of [inclusive practice] that we've invested in'.This tallies with the views of English Local Authority officials interviewed by Cole et al. (2019, p. 386), who 'saw changes in school inspection criteria as curtailing approaches that had previously promoted inclusive practice'.
Intense national pressure on particular educational performance measures may also disincentivise inclusive practice in English schools.The London headteacher described the situation for headteachers as 'quite football manager-ish', because poor exam results can lead to rapid sacking, incentivising headteachers to focus on short-term outcomes, which is 'at odds with being as inclusive as you possibly can be'.This view is corroborated by English local authority interviewees in Cole et al.'s (2019, p. 386) research, who highlighted that one particularly high-status measure introduced by the DfE to compare schools' performance-'Progress 8' scores-makes it 'more difficult for teachers to respond to cognitive, social and emotional needs of at-risk children'.Peters (2018, p. 5) argues that pressure on individual organisational results can hamper coordinated action: 'performance management has had a particularly negative effect on coordination', because where agencies are set high-status targets for their individual operations, they 'will tend to ignore collective goals'.Comparative studies have also substantiated this point, suggesting that performance management is more intense in England than in Scotland.Arnott and Menter (2007, p. 260-1) found that 'the performativity agenda is far less developed in Scotland than in England … market-based forms of accountability have shaped [practice] in [English] schools much more explicitly than has been the case in Scotland'.In similar terms, Grek and Ozga (2009, p. 949) argue that the influence of academic performance data is more strongly mediated by 'cultural and historical considerations' in Scotland; education has a particular kind of place and status in Scottish society which has kept it relatively well protected from the pressures of performance measurement and audit culture.
It would appear that the policy direction of national power-holders, particularly when combined with the relative lack of educational policy power within London, presents a substantial barrier to coordinated exclusion reduction in London.The extent of influence that national power-holders have over London-level practice may be evidenced by the close mirroring of London and England's rates of exclusion, as shown in Figures 1 and 2.

Glasgow
Policy networks and hierarchy are both closely tied to a third key factor influencing policy coordination-the extent to which relevant agencies share similar assumptions, ideas and priorities.For well-coordinated policy change to occur, there needs to be a collective reframing of pertinent issues: aligned change in practice is precipitated by common shifts in thinking (Peters, 2018, p. 6).The issue of school exclusion appears to have been effectively reframed in Glasgow, as part of a wider move to reconsider children and young people's lives.Through policy networks and more top-down measures by power-holders, energised by a sense of crisis, there seems to have been a shift in how children and young people were viewed, and thus in how problems such as violence and exclusion could be addressed: All the conversations that were going on between people … [involved] sharing ideas, and spreading ideas in different ways … thinking about violence, and being involved in a different approach.(Scottish academic) Three specific ideas were especially influential in reframing relevant issues: the principle of nurture, the importance of ACEs, and the need for 'upstream' early intervention, particularly to prevent violence.
According to an interviewee from the LVRU, 'a core reason for why [exclusion reduction in Glasgow] worked so well … is the nurturing city that they aspired to be'.This was substantiated by our Scottish interviews: we heard from a senior manager in Glasgow Council, for instance, about 'a mountain of training on nurturing principles' being delivered in schools, prompting the whole workforce to consider 'what did I do wrong for that child to act out?Not the other way round'.Behaviour in Schools research (Scottish Government, 2012) from the period of rapid exclusion reduction in Scotland echoes this, finding that 'approaches to promote positive behaviour had improved over time'-these approaches including nurture principles.Other key approaches mentioned in the research, such as restorative practice, were similarly grounded in a shift away from emphasis on punishment and sanction.
Adverse childhood experiences was the second key concept to have made a considerable difference to thinking and practice in Glasgow education during the 2000s and 2010sone Glasgow headteacher said that their staff had received regular training about ACEs for many years.By centring attention on the influence of difficult early experiences, the ACEs framework encourages all practitioners to take a more empathetic and relational approach to work with young people-and it can thus discourage punitive practices such as school exclusion.An ex-headteacher emphasised the impact of Glasgow policymakers mobilising the ACEs model: 'they rooted that as … the starting point … foundational'.The wide-ranging influence of the ACE-aware nation campaign, which spread the message of the importance of ACEs to both public service professionals and to the public, was also noted by participants (see also Zeedyk, 2021).
The third and final idea mentioned frequently by our interviewees was 'upstream' working-focusing on inclusion and early intervention, rather than over-relying on enforcement and punishment to solve issues once they have presented.This principle was strongly embedded across the Scottish public sector from the early 2000s, particularly through the Getting It Right For Every Child framework (Scottish Executive, 2006), which placed children's wellbeing and societal inclusion at the centre of all agencies' planning and decision-making, including schools, encouraging comprehensive inter-agency collaboration.At the same time, the Scottish Violence Reduction Unit advanced early support and addressing the root causes of violence as a core part of the public health approach, particularly 14693518, 0, Downloaded from https://bera-journals.onlinelibrary.wiley.com/doi/10.1002/berj.3925 by Test, Wiley Online Library on [03/11/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License through their work in Glasgow.An education charity manager described that the work done in Glasgow 'started very, very upstream … they were really coordinated about what that actually looked like … that's where they put their efforts'.Supported by a broader national focus on Early Years, this work ran in parallel with a kind of biographical reframing: rather than centring attention on the misdemeanours of young people as adolescents or young adults, agencies intentionally shifted the spotlight to early childhood.
The coherence and potency of these ideas were no-doubt enhanced by the networking and mobilisation of formal power discussed above, and by the influence of individual policy entrepreneurs (see Peters, 2013, p. 580), but may have also been supported by another factor: the impetus to be different from England.Grek and Ozga (2009, p. 948-9) suggest that, since the start of devolved SNP government in 2007, there has been a strong incentive for Scottish power-holders to signal positions that are different to the direction of policy in England, particularly regarding institutions which are central to Scotland's national identity, including education.

London
This widespread ideological reframing in Glasgow was closely tied to the strength of its policy networks, and to the sense of crisis surrounding its young people in the early 2000s, which galvanised the urgent drive for change within those networks.Our London respondents suggested that there is not such a sense of youth crisis in the English capital today within education circles, and that this may undermine the strength of collective drive towards the reframing of policy issues such as exclusion.
As reasons for this, respondents cited London's relatively high rates of educational achievement and its relatively low rates of exclusion compared with national averages (see Figures 1 and 2).In our interview with a London headteacher, the tension within this story of educational success became apparent: The narrative is that London [education] is sorted … I don't think the case for change is that clear.I mean, I say that and knowing how many children have been killed on the streets of London this year, so maybe I take that back, but certainly in the [education] sector, it doesn't feel like it's particularly in the conversation.
There may be a sense of crisis in London around violence, but less so around education, and the potential connections between the two are controversial.While some respondents highlighted associations between exclusion and violence, others were keen to stress that there is not a simple causal link between them.In Peters' terms (2018, p. 6), different relevant agencies 'have different ideas about good policy and the ways in which to address problems' in relation to exclusion and violence, which can undermine coordination between them.This is in contrast to Glasgow, where it appears that relevant policy agents and networks drew together issues of violence and school exclusion under a broader shift in thinking about young people, and that many educationalists saw themselves as playing a key role in the drive towards violence reduction.
In addition, London respondents cited various powerful stakeholders who may be resistant to substantial reframing of relevant issues around inclusion and exclusions-including headteachers. 1As one charity leader stated, headteachers can perceive measures to rethink punishments and to reduce exclusions as 'taking powers away from them to handle challenging young people'.The London headteacher suggested that strict educational philosophies retain significant power at the highest levels, pushed by particularly influential school leaders, government advisers and multi-academy trusts: 'those [strict] ideas are still very, very strong, particularly in the very influential corner of the system which is close to government, your sort of Tom Bennett, Katharine Birbalsingh, Harris [Federation], and so on'.One VRU staff member suggested that a number of schools within London were aligned with this more exclusionary philosophy.
More broadly, an education charity manager suggested that a deep-rooted ideological change was needed, away from a focus on 'aspiration and individual achievement', which currently has an 'incredibly significant primacy' within the DfE and Ofsted.Arguably, relevant hierarchical coordinators currently achieve greater coordination in their encouragement of educational individualism than they do in the measures they promote to enable inclusion or exclusion reduction across schools (see also Cole et al., 2019).Arnott and Menter (2007, p. 259) cite an analysis of voter attitudes which may suggest that this reflects public opinion: the poll concluded that there was a 'a greater belief that education should be inclusive and can bring about wider improvements in society' in Scotland than in England.This may suggest that efforts to reframe these issues would need to be directed at the public as well as at powerholders, if there is to be a consequentially widespread shift in thinking about the fundamental purposes of education in England.
The ideological discussion about exclusion in London is made all the more urgent, and all the more contentious, by debates about institutional racism in the city's schools.One charity manager highlighted the tension between those who are eager to address racism in education and criminal justice, and those powerholders who are not: I do think there's a massive connection between exclusions and criminality and race … the government, they don't: the Sewell Report [2021 report of the British government's Commission on Race and Ethnic Disparities] came out not too long ago, and basically said, there's no systemic racism, institutional racism.
It is well established that school exclusions in England disproportionately affect students from particular ethnic heritages-those from Black Caribbean, Gypsy, Roma and Traveller backgrounds in particular (see e.g.Billingham & Irwin-Rogers, 2022;Graham et al., 2019).A number of our respondents argued that this is due to racism in the education system, echoing research which has made similar claims (e.g.see Joseph-Salisbury, 2020;Kulz, 2017;Parsons, 2008), and that this problem may be getting worse.An education charity manager directly contrasted this with the situation in Glasgow: I think the movement in Glasgow and Scotland was about 'us'.It was about we're all in this, this affects everybody.I feel like London, it's more about 'us and them', and I feel like that othering of children from different ethnic groups and with special educational needs … all our educational policies are about moving all those people out.
It is clear that reframing the issue of school exclusion in London would be a profoundly contentious endeavour.Our research seems to reinforce an important point about policy transfer made by Jones and Newburn (2021, p. 25), that it is far more difficult to effectively transfer policies from other jurisdictions when those policies cannot be approached in a technocratic or politically neutral manner.Substantially reducing school exclusions in London, following Glasgow's lead, could never be a 'technical' matter-it will always have profound normative and ideological implications, and will thus be 'substantively challenging' for many powerholders, which can compromise the 'portability' of policies (Jones and Newburn, 2021).

DISCUSSION
Through our interview data, we were able to gain substantial insight into how exclusions were reduced in Glasgow in the 2000s and 2010s, and the complexities affecting potential exclusion reduction in London.In particular, our findings support the importance of networks, hierarchy and reframing for achieving well-coordinated policy change, as has been suggested by Peters (2013Peters ( , 2018)).In Glasgow, coordinated exclusion reduction was enabled by tight-knit policy networks (which were energised by boundary-spanning policy entrepreneurs and by a shared sense of crisis), effective city-level hierarchical coordinators and a widespread reconceptualisation of relevant policy issues.These three factors were closely entwined and mutually reinforcing-it appears that they were all necessary, and none of them alone would have been sufficient.
Prospects for exclusion reduction in London are affected by the nature of these same three phenomena in the capital: its policy networks; the priorities and practices of the powerholding agencies affecting its schools; and the potential for reframing relevant policy issues among city and national-level stakeholders.From our analysis, it would appear that there are some networks mobilising for the reduction of exclusions; some powerholders pursuing measures to achieve the same; and some progress being made by those seeking to change the nature of the policy conversation around exclusion.Policy conditions in London differ substantially from those which enabled sharp exclusion reduction in Glasgow, however.Our interviewees pointed to a number of challenges which would need to be overcome if coordinated exclusion reduction is to be achieved in London.Of particular importance were competition between different agencies working in relevant policy spaces; the fragmentation of the city's education system; the need for better incentivisation of inclusion by Ofsted and the DfE; and particular challenges to reframing the issue of school exclusion in London.Some respondents suggested that the differences in scale, demography and governance between London and Glasgow are simply too great for fruitful comparison of policy conditions-perhaps echoing Travers' (2003) contention that London is an 'ungovernable city'.On a more hopeful note, an LVRU staff member captured what it would take for the city to see substantial, well-coordinated exclusion reduction: if you've got from the government, from the Mayor of London's office saying that, yes, we want to bring everyone together and this is what our aim is and everybody, whoever is doing it, what are you doing towards fixing this [exclusions] challenge, tell us, and create that unity, then I think it's possible It is clear that this could not happen through an attempt to directly replicate Glasgow's exclusion reduction-if school exclusion reduction policies are to be successfully transferred from Glasgow, they must be subject to substantial 'modification, mutation and reconfiguration' (Jones & Newburn, 2021, p. 13).In particular, given the relative lack of educational policymaking power within London, school practice change within the city is far more reliant on exerting upward influence-attempting to influence national policy.The style and content of policy reframing must also be adapted to London's context: the ideological environment surrounding education London is substantially different to that in Glasgow, reflecting contrasts in the cities' history, demography and size, as well as national differences in educational philosophy-it appears that Scottish and English policymakers have quite divergent perspectives on what schooling is ultimately for.

CONCLUSION
Alongside presenting an empirical comparison of educational policy conditions in London and Glasgow, this paper has touched upon wider issues in education practice and research.Firstly, we have sought to demonstrate that theories and concepts drawn from policy studies-and policy coordination literature in particular-can provide helpful tools for analysing educational policymaking.Similarly, although it could only be a minor feature, our works suggest that the study of social movements (see e.g.Berglund et al., 2022;Laybourn-Langton et al., 2021) can be fruitfully applied to educational change.Secondly, our analysis of citylevel conditions for policy emulation is of clear relevance to the educational policy transfer literature (see e.g.Ball, 2012;Ball et al., 2017;Forestier & Crossley, 2015;Johnson, 2006).These studies tend to focus on the nation-state or trans-national entities such as 'global policy communities' (Ball et al., 2017) as units of change-our analysis here may point to the importance of cities as other potential vessels for policy transfer (as well as pointing to potential barriers to inter-city transfer).
Related to this, there is the issue of where educational power sits.At least since the abolition of the Inner London Education Authority in 1990 (see Reynolds, 1991), it would seem that power over its schools has been draining from London.Although the 'London Challenge' schools improvement initiative achieved considerable success in the early 2000s, the fragmentation of school governance in the capital has grown substantially since then, and central government was heavily involved in both catalysing and steering the London Challenge (see Ogden, 2008).While Glasgow's local authority is able to 'hold' all of its schools, there is no agency in London with a comparable role.If the twenty-first century is to be 'a century of cities' (see Barber, 2003), it is notable how varying cities' powers can be, even within Britain, over arguably the most consequential institutions for any city's future: its schools.

A C K N O W L E D G E M E N T S
With thanks to Alistair Fraser, Tim Newburn, Keir Irwin-Rogers, Gillean McCluskey and our three anonymous reviewers for helpful comments on earlier drafts of this paper.

F U N D I N G I N F O R M AT I O N
The research for this paper was supported by a grant from the ESRC-project reference ES/T005793/1.

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors have no conflicts of interest to declare.

D ATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.

E T H I C S S TAT E M E N T
The study has been approved by the College of Social Sciences Research Ethics Committee at the University of Glasgow according to their guidelines for social scientific research.

F
Rates of all cases of exclusion per 100 students inScotland, England, Glasgow and London,  2002/2003-2018/2019.After 2010-2011, these figures were collected once every two academic years in Scotland (and in Glasgow).The same data points are included for England and London from 2010 to 2011.We have not included data for 2019/2020 onwards owing to the substantial effects of the COVID-19 pandemic on exclusion rates.Sources: Department for Education (https:// explo re-educa tion-stati stics.servi ce.gov.uk/ findstati stics/ perma nent-and-fixed -perio d-exclu sions -in-england) and Scottish Government (https:// www.gov.scot/ publi catio ns/ schoo l-exclu sion-stati stics/ ).
14693518, 0, Downloaded from https://bera-journals.onlinelibrary.wiley.com/doi/10.1002/berj.3925 by Test, Wiley Online Library on [03/11/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License and 2010s.An academic who was involved in these networks at the time stressed the coherence of wide-ranging policy 'voices' across Scotland: