From assisted places to free schools: subsidizing private schools for the Northern English middle classes

This paper examines how the English educational state has consistently acted to support private schooling in areas where fee-paying schools would be otherwise financially unviable. Educational data on private school participation since the 2008 financial crisis reveals the stark regional divides between London and the South-East of England and the rest of the country. This analysis of contemporary trends is framed within a historical understanding of the spatial dualism of the English middle class in relation to education. The paper traces the policy lin-eages of the spatial logic of state subsidies for elite models of schooling in northern England, noting the continuity between Direct Grant grammar schools, the Assisted Places scheme and the recent conversion of private schools into state-funded academy or free schools. A review of applications from private schools to become free schools highlights, how differentiated local class structures affect the viability of elite education without state support. I think nothing but good can arise from an extension of the Board’s supervision and subsidy to Secondary Schools providing education for the poorer amongst the so-called professional classes


Introduction
In the context of rising global inequality and political ruptures, the last decade has seen renewed interest in sociological analyses of elite education (Maxwell and Aggleton 2015a; Van Zanten, Ball, and Darchy-Koechlin 2015). As well as more classic ethnographic modes of sociological analysis of elite formation through schooling (Khan 2011), the analysis of elite education has also taken a geographical turn (Koh and Kenway 2016;Kenway et al. 2017;Waters and Brooks 2015;Larsson and Hultqvist 2018). However, these geographical analyses have not tended to examine how schools relate to specific broader regional geographies of place and regional geographies of class and economic inequality. Burdett (1990, 1989) explored the regional geography of private school participation and the effect of state subsidies to private schools over the 1980s, but this work has not been updated. This paper examines the geography of private school participation in England since the 2008 financial crisis as a means of exploring how the resilience of private schooling relates to regional inequality and uneven geography of the English middle class. It also explores the reappearance of state support for forms of elite education in areas where fee-paying education ceases to be viable, looking at the example of private schools that have become state-funded academy or free schools. The economic implications of coronavirus for private schools could, according to one consultant, 'leave the North of England a desert when it comes to private schools' (Speck 2020). Reviewing both the geography of private school participation and the experience of private schools that have recently integrated with the state sector (Gamsu 2020) is thus timely.
These contemporary trends are set in relation to a deeper historical argument about the geography of wealth, class and education and the distinctions between the South-East of England and the rest of the country. England has long been heavily economically divided. The long-term structuring of the economy has concentrated economic, political and cultural power in London and the South-East of England with these patterns set in motion in the late 19th century (Robson 1986). This regional inequality was sharply reinforced by the politics of the 1980s (Martin 1988) and again since the 2008 crisis (Hutton and Lee 2012), with Brexit revealing the long-term consequences of both historical and more recent, austerity-driven regional economic inequality (Hazeldine 2017). Sociological analyses of elites and the geography of class formation have highlighted how the geographies of higher education are intertwined with these processes (Wakeling and Savage 2015;Cunningham and Savage 2015). However, sociological analyses of schooling and the analysis of elite education have largely not responded to these macro-regional and historical inequalities in English society. A notable exception to is the work of Green et al. (2018, 26-28) who show how the use of private schooling is concentrated in the top percentiles of the earnings distribution. This paper seeks to extend these analyses by focusing on the geography of private school participation across England and embedding an analysis of contemporary trends with an understanding of the historical geographies of class and the role of the state in supporting elite forms of education. Specifically, I draw on the work of economic historian Bill Rubinstein (1986Rubinstein ( , 1977b to argue that the downward trend in private school participation in the North of England since the crisis and the stability and growth of these schools in London and the South-East represents a long-term historical continuity. These historical arguments will be explored below in the first section of the literature review.
Using data allowing a regional and local analysis of private school use across England, the paper tracks changes in the geography of private school participation from 2007 to 2018. This is first examined at a regional scale, to examine how private schools were affected by the 2008 financial crisis. These regional scales of analysis show the persistence of the South-East/London divide. I then examine more local patterns within two of the three Northern regions of England, the North-West and the North-East of England, which includes the post-industrial cities of Liverpool, Manchester and Newcastle. Declining private school participation in many of the local authority areas that make up these regions is underpinned by patterns of school closures, mergers and conversion to state-funded academies or free schools. The latter pattern of private fee-paying schools becoming state schools forms the final part of the analysis.
From its very inception, the system of fee-paying private or 'independent' schools (also known as public schools) has received significant state support (Walford 1987;Boden, Kenway, and James 2020). The logic of state support for schools for the middle class that would otherwise struggle to afford selective private schooling has a long history which is examined in the second part of the literature review below. There is thus a particular policy lineage which connects this element of contemporary neoliberal free school policy to earlier English educational reforms which have also offered support to private schools (Walford 1987(Walford , 2014. However, what is important in the context of the analysis of the geography of private schooling participation is that this pattern of state support for more marginal private schools has been consistently focused on schools in the North of England Burdett 1990, 1989). In the final section of analysis, I examine how the state has stepped in to provide support for private schools in areas where these schools would cease to be financially viable. In doing so, I connect the recent integration of private schools to a longer argument about how the state has historically acted to support private schooling (Boden, Kenway, and James 2020). The spatial implications of this tendency of the state to support certain types of private or quasi-private schooling are explored here.
This section examines how the free school and academies programme has allowed the conversion of thirty-one private schools into free schools or academies. I focus on sixteen of these schools which are the larger, more traditional model of academic secondary schools providing education to the age of eighteen as opposed to small specialist or faith private schools. These schools are also concentrated in North-West England, the region which has seen the largest fall in private school population since the 2008 financial crisis. Using the application documents provided by these schools to justify their transition into the state sector, I examine how schools referred both to the shrinking numbers of parents able to afford private schooling and how the loss of earlier forms of state funding affected their financial position. In the literature review below, I first examine the historical geography of fee-paying elite schools and how they are tied into geographies of class, drawing on historical analyses of education, class and geography. A section outlining the theorization of class and elite schooling follows. I then describe in greater detail the history of earlier education policies which have funded fee-paying selective schools offering access to middle-class families who otherwise could not afford this education.

The uneven regional geography of private schooling and historical geographies of class and elite schooling
To understand the contemporary geography of private school participation in England, it is necessary to understand the historical geography of elite education in the UK and England in particular. During the late 19th century, alongside the reformed universities of Oxford and Cambridge, the public schools created an elite, white and initially exclusively male (Delamont 1989) national culture, in which the British middle class could also gain social prestige (Anderson 2006(Anderson , 2007. These reforms were 'consciously intended to produce a public service bourgeoisie, or urban gentry, a new stratum of men blending elements of the aristocracy and the upper middle class' (Anderson 2006, 51-52). The dominant 'gentlemanly' 'cultural ideal' of the Anglo-British imperial elite was in Dodd's (2014, 31) words, 'sited in certain institutions which underwent transformation, served "national" not local needs, gained authority to define themselves and others and inculcated appropriate (male) behaviour defining its function in and to the national culture' . In this light the geography of public schools and elite educational culture was not something which had a distinctive regional geography by intention. It was intended that this elite educational culture would be nationally and imperially spread across the country and beyond. However, as we will see this culture was most closely associated with a dominant set of class fractions concentrated on the South-East of England with implications for the geography of schooling.
In the government Commission on Secondary Education of 1868, Bryce (1868, 585, 589) described Rossall, a still-extant Anglican boys boarding school in North-West England, as 'the only representative in the county of what is vulgarly called the "public school" system' and 'the only school of its class in the north of England' . The boarding school model of elite education was rarer in northern England from the early days of modern secondary schooling. Whilst elements of the public school curriculum were present across day secondary schools in Northern cities, the growing concentration of wealth and power in London gave this 'national' process of class formation and the elite educational culture it involved a distinct skew towards the London and the South-East (Anderson 2007, 266). The relative scarcity of boarding school education in the North of England, the spatial political economy of the UK and the differentiated class structure that it created, contributed towards a regional skew in the patterns of elite schooling.
This analysis of the geography of the Victorian elite and middle class and how it was reflected in relation to education was developed by the economic historian, Bill Rubinstein. In a series of articles, Rubinstein (1977a, 617-618;1986, 199-200) analysed the geography of middle-class incomes over the 19th and early 20th century, showing how the London and south-eastern English middle classes were wealthier than their peers in the industrial counties of the North of England, the Midlands and South Wales. This gap increased over the 19th century and precisely during the period when the system of fee-paying, selective secondary schooling for the middle classes and the elite became established (Honey 1977). The coincidence of these two historical trends is important as what became known as the 'public' schools were fee-paying and the survival and expansion of these elite schools would have been dependent on the capacity of parents to afford fees. The concentration of higher salaries amongst the middle classes of the South-East of England during the precise period when elite forms of secondary schooling were being created is historically and theoretically important. It provides us with a tentative working hypothesis for understanding the current spatial distribution of elite education within England. Framing more recent trends in the geography of private schooling allows us to explore how contemporary educational inequalities are shaped by the intertwined histories of the school system and historical geographies of class and the economy. Rubinstein (1986, 200) also showed how the British elite of the late 19th and early 20th century tended to come from the South-East of England. In a less clearly corroborated argument he argued that this spatial dualism in the late Victorian middle classes and elite was linked to education. He noted the concentration of the famous public schools in South-East England and the fact that Oxford and Cambridge did not accept students who were not of the Anglican faith, as dissenting, non-Anglican religious traditions were stronger in the provincial, industrial centres (Rubinstein 1977b, 113-114;1986, 200). There is some empirical support for this (Roach 1959, 138) but as Anderson (1995, 31) later noted, Rubinstein's theorization was 'not fully demonstrated empirically' . Building on Rubinstein's hypothesis about the spatial dualism between the middle classes and elites of London and the South-East and the education system is timely given renewed sociological interest in the distinctiveness of elite class formation in London and the South-East and the role of education within this (Cunningham and Savage 2015;Wakeling and Savage 2015). This paper will seek in part to examine this theorization of the link between England's, political economy, the geography of class and the contemporary geography of the private school system.

Theorizing class and private schooling as a form of elite education
Before examining this further, I will briefly describe the theorization of class that underpins this analysis and how this relates to elite education and hierarchies of schooling. As noted above the history of private schooling in the UK is deeply intertwined with the expansion of the middle classes during the late 19th and early 20th century. Rubinstein's (1986) analysis showed how only a small number of public schools were dominated by super-rich aristocrats and major commercial or industrial plutocrats of this period. He argued that only Eton, and to a lesser extent Stowe, Westminster, Winchester, Marlborough College and Wellington College were schools with a significant elite presence (Rubinstein 1986, 173). Delamont's (1989, 47-49) study of two elite girls' private schools in Scotland from the late 1960s through till the mid-1970s described their intake as being primarily recruited from senior managerial and professional families. Maxwell and Aggleton (2015b) have noted the increasing presence of international students from very wealthy families. Affordability has certainly declined since the 1980s for many median and top earners (Green et al. 2017, 19;CEBR 2019) and only for the highest paid professional occupations (doctors and executives) are fees likely to be easily affordable (CEBR 2019, 14-16). Nevertheless, the use of private schools by professional and managerial occupational groups appears from survey data to be relatively stable (Green et al. 2017, 29-31). Accessing data on the actual social class intake of private schools is not possible so we are left with case study data and proxy data from larger surveys to estimate private school intakes. These analyses suggest that private school is a comfortable option for the upper middle-classsenior, highly-paid professional and managerial occupations, alongside extremely wealthy national and international elites. For lower-paid professional groups that may otherwise be considered middle class (teachers, nurses, IT professionals), funding private education is likely to be a much more demanding undertaking.
Disaggregating the intake of private schools also involves considering the diversity and heterogeneity of private schools and hierarchies of schooling more broadly. We need to consider what elite schooling means. Gaztambide-Fernández (2009) defines an 'elite' boarding school as one that is independent (in its governance and curriculum), academically elite in terms of attainment, historically, geographically and demographically elite. Families using private schools may be primarily concentrated in the higher percentiles of the income distribution, but the number of schools that have an intake drawn exclusively from extremely wealthy and/or powerful backgrounds probably continues to be small. Demography alone does not determine institutional status and there have always been distinctive hierarchies and intake differences amongst particular private schools (Power, Edwards, and Wigfall 2003, 13-15). Not all UK private schools are necessarily socially elite in terms of their intake or hold the cultural and economic capital that garners elite status for a school. Private schooling in the UK includes schools for those with disabilities and small religious schools for minority faiths. These are generally far from being elite schools. However, many more schools rely on offering or claiming to offer an education that is 'superior' in some way from the state-maintained system; holding or claiming 'elite' status is central to these schools' survival.
The distinction between holding an elite status and aspiring to one is important. Institutional mimicry of educational practices of institutions seen as holding greater prestige by dint of wealth, age, symbolic or cultural influence by institutions lower down the pecking order has been a common educational phenomenon (Veblen 2007, 239-241;Lowe 1982;Banks 1955, 220). There have always been intermediate sets of both private and high-performing state schools that have blurred the boundaries of elite schooling usually associated with the older 'public' schools and this continues to be the case (Leinster-Mackay 1981;Gamsu 2018;Gamsu and Donnelly 2020). Defining what makes schools 'elite' is a relational process with fuzzy boundaries that change depending on the scale and dimension of inequality in question. This also means we need to reflect on how certain schools can maintain a position of being locally elite, operating a dominant position relative to other local schools and colleges, without necessarily being nationally 'elite' relative to the more socially exclusive and/or academically successful elite private or state schools (Gamsu 2017, 215). This paradox is perhaps best expressed by Anderson (2007, 273) who argued that the public schools were 'the nationally standardized system of middle class education. ' These nuances lie behind the use of the term elite and middle-class schooling. This reflects the fuzzy realities of private schools are in many cases locally elite in terms of attainment and have a largely (upper) middle class intake without the presence of elites that are concentrated on London and the South-East (Cunningham and Savage 2015). This relational construction of elite schooling is key to understanding the position of the private schools discussed below that have converted to state providers over the last decade, many of which had previously been in a similarly liminal position of partial autonomy from/reliance on the state.

From direct grant grammar schools and assisted places to free school conversion: state subsidy for private schooling
This paper also examines how the state has intervened in the private school system in ways which effectively subsidises selective and traditionally 'academic' forms of middle-class secondary schooling in areas where they would otherwise be financially unviable. Recent analyses have shown how elected politicians and members of the legal elite have consistently used the structures of state power embedded in parliament and the legal system to protect the financial advantages of private schools provided by their 'charitable' status (Lowe 2020;Boden, Kenway, and James 2020). Certain schools that were in a more marginal financial position than wealthier and sometimes more socially exclusive private schools have relied on more direct forms of financial support from the state. The Direct Grant status permitted to certain English grammar schools until 1976 and the Assisted Places scheme both had a distinctive geography, skewed towards supporting largely middle-class schools in the North of England. The Direct Grant (DG) referred to a form of direct state funding which developed gradually as some older, endowed secondary grammar schools found their endowments inadequate to deal with rising student numbers (Partington 1967, 4-5). These grants gradually become more formalised under state regulations from 1902 onwards and by the inter-war period, DG status referred to a situation where central government provided grant funding for a proportion of free places whilst most students at the schools paid fees. DG status also allowed the school to remain autonomous from local government. This autonomy meant that these DG grammar schools sat between the 'public' schools, which were largely or exclusively fee-paying, and state-maintained secondary schooling which was funded and managed directly by local government. The original intentions of this type of arrangement were expressed concisely by Robert Morant, permanent secretary of the Board of Education: I think nothing but good can arise from an extension of the Board's supervision and subsidy to Secondary Schools providing education for the poorer amongst the so-called professional classes (Partington 1967, 25) This blunt description of the classed nature of this scheme suggests a particular dimension of DG status -to provide a selective, academic education similar to that found in the public schools to middle-class families that otherwise may struggle to afford it. After this status was abolished by a Labour government in 1976, with the 116 of the 174 DG schools joining the private sector rather than becoming state comprehensive schools. In response, the 1979 Conservative manifesto committed the party to creating the Assisted Places Scheme: The Direct Grant schools, abolished by Labour, gave wider opportunities for bright children from modest backgrounds. The Direct Grant principle will therefore be restored with an Assisted Places Sch0eme. (Dale 2000, 279) This portrayal of DG schools as advantageous to the working classes was a sleight of hand; 49.9% of free places at direct grant grammars went to students from professional or managerial backgrounds, with only 10.3% from manual occupational backgrounds (Donnison 1970b, 201). Over 50 years after Morant's comments, little had changed in the policy's class bias. This model of funding and relative autonomy was thus geared towards protecting and strengthening locally (and sometimes nationally) elite, selective forms of schooling that tended to educate the middle classes.
The Assisted Places (AP) scheme continued this class logic of supporting largely middle-class families to attend forms of private schooling. The beneficiaries of the AP scheme were not generally from 'unambiguously working class' backgrounds, with researchers referring to the 'submerged middle class' to describe the less affluent and less senior, but still highly educated, professional parents of most AP students (Douse 1985, 215-216;Fitz, Edwards, and Whitty 1986, 185-186). Some evidence also suggests there was a distinctive geography to the AP scheme. The DG grammars had tended to be concentrated in urban areas and were particularly geographically concentrated in the cities and conurbations of North-West England (Donnison 1970a, 58). Private schools in North-West England received a disproportionate number of assisted places relative to the region's population in the 1980s (Bradford and Burdett 1990, 41).This was true across Northern regions, with the AP scheme effectively offsetting some of the geographical polarisation caused by private schools in the South-East growing faster than those in the North (Bradford and Burdett 1989, 51). In their more detailed study, Edwards et al. (1989, 80-85) found that geography of uptake of APs at a local authority level was more complex, but they did not show percentages making it hard to draw a clear conclusion. However, they did find that it was consistently the former DG schools taking higher numbers of AP students relative to the older, wealthier and generally more prestigious public schools. Taken together what these studies suggest is a long-term North-South divide in private schooling, some evidence suggesting a regional bias to the AP scheme and clear evidence showing that DG schools were more likely to offer AP.
There has been a persistent educational geography of elite and middle-class social reproduction, with private schooling smaller and weaker outside of the South-East of England and particularly in parts of the North. There has also always been a group of quasi-independent schools reliant on state financing. In some provincial cities, like Greater Manchester, Merseyside and Birmingham, the schools that offered large numbers of Assisted Places formed a large proportion of the private schools in these areas. The AP scheme, like the DG schools before, primarily offered places to middle-class students as opposed to those from manual working-class backgrounds (Edwards et al. 1989, 161-164). The state has thus consistently acted to support forms of quasi-independent, locally elite schooling in areas where these schools would otherwise struggle to recruit enough fee-paying students. In what follows I first examine recent regional trends in private school participation before looking at more localised patterns in private school participation at local authority level in the North West and North East of England. Finally I explore how the conversion of private schools into free schools and academies represents the continuation of the DG and AP schemes and their tendency to sponsor forms of 'academic' schooling for the middle classes where these schools would otherwise struggle if forced to rely on fee income alone.

Data and methods
Three sources of data are used here. The graphs draw on the Schools Pupils and Characteristics datasets which are published annually by the Department for Education (DfE 2018) and show the number of students of primary and secondary school age in independent school compared to those in state-maintained provision. I aggregated data for England over multiple years to produce the time series found below, the years 2007-2018 were chosen to examine how trends in participation were affected by the 2008 crisis and any geographical variation therein. This aggregated data was plotted in graph form using R with tools from the tidyverse packages. I first explore regional trends (Figures 1 and 2) before examining the geography of private school participation at the lower spatial scale of the local authority (Figures 3-6). Local authorities refer to administrative areas that normally have a full range of local government functions. They vary from being large rural areas with sparse populations to densely populated London boroughs; in 2018 the median size for a local authority in England was 279,665 (Author's calculation. Population Estimates Unit 2019). It is worth noting that this data only allows us to examine the numbers of students who attend schools in these areas. It is not possible to tell whether students in private schools in these areas live elsewhere. As a result these figures only allow us to assess the health of schools' intakes. I use the term 'private school participation' to refer to both the percentage of private school students of the total school population of each area and the absolute number of private school students. Analysing both figures is important as private schools may decline in percentage terms, making private school an experience for a smaller proportion of the school age cohort, whilst numbers remain static, allowing schools to continue.
For the final section of analysis which examines the conversion of private fee-paying schools to academy and free schools, I draw on another DfE (2019) dataset, Get Schools Information. This provides a record of school openings and closures in England. For table      one this is the primary source of data but I have cross-referenced this with local media reports and school websites to try and limit errors and absences in the DfE dataset.
This final section also draws on a review of the application documents of schools applying to become state-funded academies to explore the reasons behind these conversions. This section examines how falls in private school participation in certain areas is linked to the conversion of these schools. It also explores how these school intakes prior to conversion affected the decision to apply for full state-funding. Actual data on the socio-economic intake of private schools is not available as private schools are not obliged to report data which could be used as a proxy for social class. This limits our capacity to test Rubinstein's analysis of the differentiation of the middle classes in relation to education. Using the application documents alongside local news articles, we can still make more limited assessments about local and regional distinctions in the nature of the private school intakes.

The regional geography of private schools since the crisis
The regional graphs shown in Figures 1 and 2 reveal the stark regional disparity between London, the South-East and the rest of the country when it comes to private schooling. The South-West and the East of England are in a slightly intermediate position, as is Outer London with its lower density of private schools making its percentage of students in private schools closer to the national average than inner London. Comparing the absolute numbers and the percentages here is important, as only Inner London has seen both rising numbers and a rising share of the school population. All other regions have seen percentage falls with most regions peaking during 2007-2009. However, in absolute terms London, the South-East, the East of England and the West Midlands have seen student numbers increase or remain steady over the last 15 years. All other regions have seen student numbers fall by 4000 students or more. Whilst a decline of this magnitude may not seem large, it is important to consider that even large private secondary schools with sixth forms will have little more than 1000 students and most are considerably smaller. Declines of this scale suggests school closures, mergers or conversion to the state sector. To examine these localized indications of private school numbers it is necessary to look at this trend data at a lower geographical scale. As we will explore, these localized effects have been more prominent in regions outside of London and the South-East. We will examine more localized trends in the North-West and the North-East to explore how these regional trends reflect shifts in local school systems and may suggest distinctive patterns of class formation. Tables for other regions broken down to local authority level are provided in the appendix/supplementary material available online.

Localised trends in the North-East: a small sector shrinks further
The North-East has the smallest private school sector in the country in both absolute and percentage terms underlining the deep regional divides in private schooling (Figures 3  and 4). However, these regional differences hide considerable intra-regional variation. Within the North-East, Newcastle stands out as being the local authority with the highest proportion of private school students as a percentage and absolutely. However, this has been falling consistently since 2009. During this period, the head of Newcastle's Royal Grammar School claimed that the school's future was ensured by senior public professionals in the universities and health service who had not seen their incomes fall substantially during the crisis (Warrell 2014). Whilst highly-paid public sector professionals are undoubtedly present in private schools elsewhere in the country, this reliance on the local public sector elite suggests a distinct differentiation in the intake of private schools in the North compared to those in the South-East. This apparent reliance on public sector professionals may have been exacerbated by the collapse of the Newcastle-based bank Northern Rock during the financial crisis (Lee 2009;Dawley et al. 2014;Champion and Townsend 2012). These macro-scale regional economic changes may have affected schools in the North-East and elsewhere.
Elsewhere in the North-East, we can clearly see the weak position of private schooling across the region. In Darlington, North Tyneside and Sunderland we see downward trends from 2008 or 2009 followed by dramatic downward jumps. These sharp drops in student numbers and percentages reflect school closures and, in each of these three local authorities, the conversion of private sector schools to state-funded free schools. We will return to the policy of free school conversion below expanding on these examples, but as noted above in the case of Newcastle, the relatively weak position and small scale of affluent middle-class families appears to have played a role. Only Stockton and Northumberland appear to have relatively stable private school populations and only in Gateshead do we see rising numbers of pupils in both absolute and relative terms. This anomaly is due to the growing intake of seven private Jewish schools in Gateshead are Jewish private schools, four of which opened between 2009 and 2019). The post-industrial city of Middlesborough has virtually no private school students and three other boroughs, Redcarr and Cleveland, Hartlepool and South Tyneside are not shown here as they effectively had no or only very small providers for the whole period.

North-West: a larger more stable sector but considerable losses as private schools go state-funded
North-West England saw a very large decline in the private sector with student numbers shrinking by nearly 10,000 -in percentage terms a drop of nearly a full percentage point (Figure 1). Like the North-East, several boroughs (Warrington, Knowsley, Halton and Wigan) were excluded here with almost no private school students. It is worth re-iterating this point -the economic basis for significant private school provision of any sort, let alone the elite forms of 'public school' education commonly associated with private schooling, simply does not exist in these post-industrial towns. There is no material basis, no critical mass of affluent parents, for the existence of private schools. The raw, descriptive contrast between these areas and local authorities in the South-East and London (See supplementary material) underlines the fundamental economic underpinning of the uneven distribution of private school provision.
Elsewhere in the North-West, the graphs show a sector that is substantially larger and generally more stable than in the North-East. Local authorities like Blackburn, Oldham and Sefton saw declining pupil numbers in the years following the 2008 crisis followed by an apparent stabilization at a lower level. Blackburn, Blackpool, Cheshire, Liverpool and the Wirral all saw sharp declines in a single year which in all cases except Blackpool are due to private schools converting to become academies or free schools. Trafford, Manchester, Stockport and Bolton all appear to have relatively stable private school populations though the percentage figures for these areas show a declining percentage of the school population using the private sector overall. Only Salford has a growing private school population which, as in Gateshead, reflects the existence of private Jewish secondary schools. In Salford in 2019 there were 23 private schools of which 21 were Jewish private schools (11 have opened since 2010). Lancashire also appears to have a slight growth in its private school population though this is down to the re-location of a school from Blackpool just over the local authority border into the neighbouring town of Lytham when two private schools merged. It is also worth emphasizing that whilst there might be regional trends and patterns to private school participation in post-industrial towns and cities that tend to be concentrated in the North of England, the reality of private school provision, even in the North is not consistent across regions. Instead there is a more archipelagic patterning of private school provision with rural counties, towns and cities having varying histories and contemporary capacities for sustaining fee-paying education. It is perhaps more accurate to think of the clear regional clustering of areas of high private school use with pockets of local authorities that contradict regional patterns.

Free school conversion: subsidizing elite educational culture for the provincial Middle class
Behind these aggregate trends of decline at local authority level lie individual school mergers and closures and, of particular interest here, conversion to academy or free school. Table 1 provides lists all the schools that have converted to free school or academy status. In the North-West where the decline in private school participation has been sharpest, there have been several private school mergers and ten private schools have converted to an academy or free school. Three of these, Liverpool College (2012, 5), Chetwynde School (2012, 15) and Queen Elizabeth's Grammar School Blackburn (2013, 18), all cited the economic situation as having affected parental ability to pay fees. A further six converter schools across the country mentioned the effects of the financial crisis and/or a falling number of families able to afford fees in their application to become a free school. Interestingly, Queen Elizabeth's Grammar School Blackburn (2013, 25-26), which had just 456 students at the time of their application, noted that 'under the Direct Grant and Assisted Places schemes, the school was at capacity with some 1200 pupils' . Similarly Liverpool College (2012, 10) noted that numbers reached 'a historic height of 1100' under the AP scheme and fell to 728 in the year before conversion. In West Yorkshire, Batley Grammar School was one of the first wave of Free Schools and its conversion application to convert provided data on school rolls from 1988 onwards. A former DG school, Batley Grammar School (2010, 20-21), had over 50% of their students on the AP scheme by 1998. Numbers declined over the 2000s as the scheme was run down, and the school noted in its application the 'number of people able to afford fees in such a deprived area has always been limited' (Ibid). This suggests continuity with Bradford and Burdett's (1990, 41;1989, 51) earlier analysis that suggested the AP scheme had disproportionately gone to private schools in the North and, in the North-West, particularly to those independent schools which had previously been DG-funded.  This suggests a particular continuity within education policy: free school and academy conversion of private schools, where it involves larger secondary schools, is effectively state-sponsorship of private schools which are otherwise likely to face financial difficulty. This process of reform was first fostered by New Labour (Morrisson 2007;MacLeod 2006), with then education minister, Andrew Adonis, deliberately framing it as a recreation of DG schools. Conversions from private to academy/free school accelerated under the coalition government after 2010. Both the AP scheme and DG status provided state subsidy of a particular form of secondary education associated with middle-class patterns of social reproduction, an 'academic' curriculum and forms of selection on entry to and within the school. On a smaller scale, academy and free school conversions are playing a similar supportive role to some institutions within the private sector, though this remains a minority phenomenon as only 32 schools have converted. As can be seen from Table 1, many of the conversions also involve small specialist or religious schools so the continuation of the DG/ AP policy trajectory only applies to the larger secondary schools. Of the larger schools (>400 pupils), ten out of the sixteen held DG status at some point. Liverpool College explicitly makes this link to the DG policy: The Governors and Leadership have carefully considered the opportunities presented through the Academy proposition, which has been described as a modern and forward-looking interpretation of the highly successful Direct Grant initiative (Liverpool College 2012: 12) The academy-free school conversion represents a continuation of a long-term state intervention to financially support independent schools where they are not financially viable. These liminal independent schools highlight how the state actively intervenes to support schools that are intended to cater for the middle classes in areas where there is an absence of elite wealth and where the middle class is perhaps smaller or more financially vulnerable.
It is also notable that there is a relative absence of schools in London and the South-East converting to free schools or academies and those that have are small specialist or religious providers. In cities like Bradford, Liverpool and Wolverhampton the conversion of these schools means the private sector has shrunk substantially. It has disappeared completely in towns like Darlington or Barrow-in-Furness. The conversion documents of Chetwynde School (Barrow), Royal Wolverhampton School (Wolverhampton), Batley Grammar (Batley) and Polam Hall School (Darlington) support their cases with reference to their higher than average local attainment. These are institutions which preserve the academic culture of the private sector in areas which are post-industrial and do not have large affluent local populations. This is a policy that, intentionally or otherwise, acts to preserve schools in areas where the middle classes appear to be too small or financially weak to support such provision.
As noted above, parental social class data specific to individual schools is hard to access, but converting schools frequently made reference to the demographics of their intakes. As Batley Grammar's application implied, the financial vulnerability of these schools may reflect the restrained financial circumstances and smaller size of the provincial middle-class in certain areas, particularly in post-industrial towns and cities. Sunderland's Grindon Hall School was another earlier converter and the Head's description of the intake further reinforces the geographical and social stratification of intakes in the private sector: Grindon Hall was not, says Gray, 'a posh private school' . A typical parent might own a minibus taxi business depending on public sector work which, in the northeast, is contracting sharply as cuts bite. The school's roll dropped from 342 in 2006 to 241 in 2011 […] 'I spent the last year wondering who was going to be the next parent to come and say they can't afford the fees' . (Tighe 2013) Elsewhere in the North-East, King's Priory School in North Tyneside was an early converter and two Newcastle girls private secondary schools merged in 2014 (Nicholls 2014). Private school mergers of private schools took place across the North of England as well as in some Midlands towns (Peel 2015, 54-55). In Blackpool, the merger of two local private schools (Rogers 2011) also took place in the context of falling rolls. Shortly afterwards, Blackpool's only other private school, Langdale Preparatory School, became a state-funded free school. Langdale School claimed in its free school application: Given the socio-economics of the area, Langdale already has a very diverse pupil base, very different from most other independent schools in the country. The parents of Langdale specifically, […] are not the 'sharp-elbowed, pushy middle-class parents' […] quite the contrary. They are exactly the working/lower middle-class parents, struggling to make ends meet, starved of access to sufficient high-quality education for their children, that the policy is designed to support. (Langdale Free School 2012, 74) Whilst this is not accompanied by quantitative data, this is strikingly similar to the account of the Sunderland head. From this perspective, the policy of allowing free school/ academy conversion has a geographical class logic as the independent to free school conversions in the North result partly from falling student numbers and the weaker financial position of the (lower) middle class. There is not necessarily a deliberate geographical skew of this policy towards private schools in the North of England and the evidence does not show conclusively that this is the case. But the policy has an implicit spatial logic. The state is stepping in to preserve institutions which, in the context of a much smaller private sector, are rarer in parts of northern England. Within several of the cases discussed here, integration has led to a substantial decrease in private school participation in areas with very small private school populations in the first place. The state is effectively subsidising forms of middle-class education, and schools that are locally elite, where the schools would otherwise face closure. The relationship between differentiated local class structures and private school closures, mergers or conversions is complex and further research is needed to examine this relationship in depth.

Conclusion
This paper has revealed a distinct geography of private school provision, which reflects an unequal distribution of economic wealth and creates a distinctive geography of social reproduction and education. The uneven geography of private school participation with its skew towards the South-East appears to have deepened since the 2008 financial crisis and may deepen further in the coming years. In a way that is often overlooked in academic and policy discussions, school provision is underpinned by the uneven spatial distribution of economic capital. The regional polarization of the economy shapes the geographies of class that underpin the viability of private schooling. Falling student numbers, school closures, mergers and conversion to the state sector suggest that traditional regional patterns of class and schooling have been remarkably persistent over the last 150 years. It still remains to explore the geographies of social class intake of private schools, but the data presented here provides tentative further evidence for Rubinstein's hypothesis regarding the spatial dualism of the English elite and middle classes and the schools they use. If we think of the 'norm' of the public school -small class sizes, an academic curriculum, high-end sports and cultural facilities, this is not an institution which has been evenly geographically distributed. The intention may have been to create a national elite educational culture, but the unequal geographies of class and economy mean that this has never truly been the case.
This uneven geography of private schooling also underlines how the state is an active agent in the maintenance of this model of schooling. There is clear continuity in state support for fee-paying schools that are otherwise unviable, first through the DG scheme, then through the AP scheme and most recently through academy/free school conversion. It has been argued that contemporary reforms to English education and the broader expansion of academies as the primary mode of state school governance represents a reversion to the patchwork of providers favoured by the state in the late 19th century (Ball 2012, 99-102;Exley and Ball 2013, 29). However, in the case of these schools that have always existed on the margins of the fee-paying sector, effectively requiring state funding to survive, the logic of the state has been remarkably consistent. This is not only a classed logic of state power in supporting schools that primarily serve the middle classes, there is also an implicit spatial logic of state action here. This works in two senses, firstly in preserving traditional forms of 'academic' schooling in areas where the state sector otherwise tends to have low academic results (e.g. Darlington, Barrow-in-Furness). There is a certain irony to free school and academy conversions here, as removing fees is likely to reinforce the local polarization of school populations as more affluent students move into the newly fee-free schools, an outcome partially corroborated by Allen and Higham (2018) as well as in a new report (Private Schools Policy Reform, Forthcoming). Secondly there is a regional bias to these conversions which suggests geographies of class are almost inevitably involved. Where private schools convert in concentrated ways across regions, or where the conversion of a school means the private school sector disappears completely in a local area, this conclusion seems legitimate, although we would need more granular intake data to be able to show this conclusively.
Within the sociology of education, this paper also makes the case for integrating history and geography in how we understand education. Uncovering how power is exercised by the state in fragmented, complex neoliberal systems of education requires combining an understanding of the local exercise of power at a local level with a historical understanding of state intervention and the geographies of these interventions. Thinking about macro-regional patterning of the economy and populations also allows us to link our understanding of class and education to the spatial political economy of capitalism. This paper has concentrated on class, but analyses of the historical geographies of schooling and the economy are equally applicable to the analysis of race and schooling (Maurer 2018) and the implications of gendered historical geographies of labour (McDowell and Massey 1984) on schooling remain equally under-explored. Tying these macro scales of historical and spatial analysis to the minutiae and context of the local, allows us to see how contemporary educational inequality is shaped by deep historical geographies of educational power.

Disclosure statement
No potential conflict of interest was reported by the author.