A Guide to School Choice Reforms Reforms Choice School to Guide a Schools) and the US (Charter Schools)
School choice reform will be a key issue at the next General Election, yet the debate so far has focused on the theoretical arguments for and against creating a ‘schools market’ by bringing more independent providers into the state system. The purpose of this report is to learn the lessons of existing school reforms in England (the academies programme), Sweden (free A guide to school choice reforms reforms choice school to guide A schools) and the US (charter schools). We assess the success of reforms in all three countries against seven criteria which we believe a schools market should meet in order to find the right balance between promoting innovation and choice while maintaining accountability and quality control. A guide to school
None of the countries studied have achieved this balance yet, though in each case the introduction of new providers to the choice reforms system has brought benefits, but all of the seven criteria are
met by at least one country. By combining the best aspects of Daisy Meyland-Smith and Natalie Evans Natalie and Meyland-Smith Daisy each system, we argue, it is possible to develop a set of school Xxxxxx xxxxxxx xxxxxxx xxxxxxx choice reforms that will increase diversity and performance while protecting against market failure.
Daisy Meyland-Smith and Natalie Evans
£10.00 ISBN: 978-1-906097-42-4 Policy Exchange Policy
Policy Exchange Clutha House 10 Storey’s Gate London SW1P 3AY
www.policyexchange.org.uk
PX - A GUIDE TO SHCOOL - COVER 02-09.indd 1 20/2/09 10:46:35 A guide to school choice reforms
Daisy Meyland-Smith and Natalie Evans
Policy Exchange is an independent think tank whose mission is to develop and promote new policy ideas which will foster a free society based on strong communities, personal freedom, limited government, national self-confidence and an enterprise culture. Registered charity no: 1096300.
Policy Exchange is committed to an evidence-based approach to policy development. We work in partnership with academics and other experts and commission major studies involving thorough empirical research of alternative policy outcomes.
We believe that the policy experience of other countries offers important lessons for government in the UK. We also believe that government has much to learn from business and the voluntary sector.
Trustees Charles Moore (Chairman of the Board), Theodore Agnew, Richard Briance, Camilla Cavendish, Richard Ehrman, Robin Edwards, George Robinson, Tim Steel, Alice Thomson, Rachel Whetstone, Virginia Fraiser, Andrew Sells, Simon Wolfson. A guide to school choice reforms
About the authors
Daisy Meyland-Smith Natalie Evans Daisy is a research fellow in the Education Natalie Evans is Deputy Director of unit. Her work currently focuses on meth- Policy Exchange, responsible for the out- ods of introducing quality, competition and put and strategic direction of the research diversity into the education sector, particu- team. larly through the judicious use of profit. Prior to joining PX she was Head Prior to joining Policy Exchange, Daisy of Policy at the British Chambers of was a Press and Parliamentary Officer for Commerce. She has also previously been the Conservative group in the National Deputy Director at the Conservative Assembly for Wales. Research Department specialising in wel- She has a first class degree from the fare and economic issues. University of Bath, where she read Politics She has a degree in Social and Political with Economics. Sciences from Cambridge University.
© Policy Exchange 2009
Published by Policy Exchange, Clutha House, 10 Storey’s Gate, London SW1P 3AY www.policyexchange.org.uk
ISBN: 978-1-906097-42-4
Printed by Heron, Dawson and Sawyer Designed by SoapBox, www.soapboxcommunications.co.uk
2 Contents
Executive Summary 4 Introduction 6
1. Academies 9 2. Sweden 23 3. USA 36
Lessons and recommendations 54 Appendix A 63 Appendix B 64 Appendix C 68
www.policyexchange.org.uk • 3 A guide to school choice reforms
Executive Summary
School choice reform will be a key issue at The Ten Key Lessons the next General Election, yet the debate We believe that it is possible to develop so far has focused on the theoretical argu- a programme that adopts the best aspects ments for and against creating a ‘schools of all three systems and could be imple- market’ by bringing more independent mented in this country. We draw ten key providers into the state system. The pur- lessons from our research: pose of this report is to learn the lessons of existing school reforms in England 1. Once established ISFS (Independent (the academies programme), Sweden (free State-Funded School) systems grow stead- schools) and the US (charter schools). ily and reforms are difficult to reverse: In The first three chapters assess the success the US there are now 4,568 charter schools of reforms in all three countries against educating 1,341,687 children. In Sweden, seven criteria which we believe a schools where there are fewer barriers to setting up market should meet in order to find the new schools, approximately 11.9% of chil- right balance between promoting innova- dren are educated in 3,302 free schools and tion, choice and diversity while maintain- pre-schools. ing accountability and quality control. It should be: 2. Most studies of attainment in ISFS show a positive effect 1. Demand-led 2. Easy to Enter 3. A system based on independent 3. Accountable state-funded schools moves naturally 4. Genuinely free towards federation: In Sweden, where 5. Financially consistent and stable there are few barriers to the creation 6. Politically stable of federations, the majority of schools 7. Fair are now run by for-profit companies and the vast majority of new applications None of the countries studied have are for this type of school. In the US achieved this balance yet, though in each there are much stronger barriers but still case the introduction of new providers to as many as 30% of charter schools are the system has brought benefits. The chart involved with management organisations below shows a summary of our findings: in some capacity.
Easy to Genuinely Financially Politically Demand-led Accountable Fair enter free consistent stable
UK X X
USA /* /* X X
Sweden X X X
indicates that the evidence on whether a system passes the given test is mixed * These categories have been given two ratings because of the huge difference between states that have multiple-authorisers and are, therefore, more demand-led and accountable, and states that only allow school districts to authorise.
4 Executive summary
4. ISFS in federations seem to perform Our Recommendations better than one-offs Our first recommendation is that it makes sense to think of school choice 5. Allowing commercial companies to reform as a series of stages rather than set-up ISFS significantly boosts the a ‘big bang’. That way the system can potential for federations to develop: be developed in a coherent fashion For-profit groups are much more likely to rather than reactively in the face of unex- have the scale and ambition necessary to pected difficulties. create multi-school federations. Stage 1) Immediately reform the acad- 6. The authorising/commissioning emies programme by removing barriers process is crucial for the success of ISFS to entry and developing a transparent reform: The best approach, we argue, commissioning process. The DCSF would is that taken by US states which have identify, in a transparent manner, those multiple authorisers. This is because schools it wishes to become academies schools can approach more than one and initiate a public bidding process for authoriser, so reducing the risk to diver- sponsors. Clear criteria for bidders should sity, but authorisers themselves are in be stated, with preference given to those competition and so typically take a more already running successful academies. For- rigorous approach to accountability profit companies should be allowed to bid and oversight. to boost supply.
7. Existing local government providers Stage 2) Transfer the oversight of acad- should not be able to veto provision but emies to a variety of local and regional also should not be prevented from par- authorisers. As the academies programme ticipating in reforms: In all three coun- expands it will become impossible for the tries, most local authorities (or municipali- DCSF, or any national agency, to manage. ties or school districts) have initially been Instead the DCSF should look to approve hostile towards ISFS reforms. a range of ‘authorisers’ such as Local Authorities, elected mayors, universities 8. Accountability is difficult to man- and educational charities. age at a national level: The problem of entirely ignoring local government Stage 3) Introduce a national funding is that it is very difficult to author- formula. A clear and transparent national ise and oversee schools from central funding formula should be introduced government. as recommended in our report School Funding and Social Justice published last 9. Funding needs to be fair and year. All schools would receive per-pupil consistent funding direct from the government.
10. Choice does not necessarily lead Stage 4) Allow the network of authoris- to segregation but admissions policies ers to start commissioning new schools. have to be set carefully: In particu- Authorisers would be able to approve an lar, policies must remove any ‘early bird’ unlimited number of schools from provid- advantage which would favour the better- ers whose educational model had already informed middle classes and should, ide- proved successful. For entirely new pro- ally, incentivise providers to start schools viders an annual cap would be in place to in deprived communities. regulate supply.
www.policyexchange.org.uk • 5 A guide to school choice reforms
Introduction
Arguments over school choice are likely no incentive to improve. Likewise a com- to feature strongly at the next General pletely free market would undoubtedly Election. The current Government’s acad- increase inequality in an already segregated emy programme remains a key plank of school system. The answer is, rather obvi- their education reforms. The Conservative ously, to find the right balance – a model Party and Liberal Democrats have prom- in which alternative providers have a route ised to go much further, allowing parents into a market, so as to keep it dynamic and or charities to set up ‘free schools’ wher- innovative, but one that is designed and ever there is demand, though in both cases regulated to prevent negative outcomes. there remains plenty of detail to fill in. The policymakers who introduced the It is, therefore, an opportune moment to three reforms analysed in this report - the review the performance of the academies academies programme, ‘free schools’ to programme in the UK and independent Sweden, and ‘charter schools’ to most US state-funded schools in the two countries states - have each tried to find this bal- that will feature most in the election ance. None of them has quite managed debates: Sweden and the United States. it, though in each case the introduction of Our goal is to refigure the debate towards new providers to the system has brought evidence rather than theory and, given that significant benefits. In England the acad- all three main parties support alternative emies programme was introduced by the provision in one form or another, discuss Blair government in 2000 to allow new the practicalities of how a schools market sponsors to take over failing schools, but should operate rather than whether we remains tightly controlled by government should have one at all. and difficult for potential providers to Over the past fifty years the argument for access. In Sweden the reforms have led developing a market between state-funded to a much more open system that allows schools has revolved around the ideas of companies and charities to set up schools choice and competition. Supporters have wherever they feel there is demand, with insisted that giving parents the freedom to few conditions to entry. Even so, there are choose provision from a variety of differ- problems with trying to run the reforms ent suppliers, rather than enforcing a state from the centre and the accountability of monopoly, would force standards up over free schools is relatively weak. In the US it time. Opponents have tirelessly fought to can be difficult for providers to gain per- maintain state control by raising fears that mission to set up charters and many states the creeping privatisation of a school sys- impose unnecessary limits and regulations. tem would detract attention from the core Furthermore, charters schools typically duties of a school, benefit the wealthy and receive less funding than public schools. would work to the detriment of the teach- From our analysis of the academies, ing profession. free schools and charters schools we have Continuing these arguments, though, created a typology of seven key tests that doesn’t get us very far. After all, both are a schools market should meet in order to right to some extent. There seems little find the right balance between promoting question that monopoly state provision innovation, choice and diversity while can, without an unrealistically constant maintaining accountability and qual- supply of self-motivated senior manage- ity control. We argue that such a market ment, lead to a complacent system with should be:
6 Introduction
1. Demand-led does not increase segregation in the school Entry to the system and ongoing pres- system. Ideally it should be designed to ence should be based on the level of reduce existing segregation. demand amongst consumers (i.e. parents and pupils) rather than patronage from Each of the three markets we looked at local or central government. met some of these criteria, but none met all of them. Academies were the most 2. Easy to Enter financially stable of the three. They are It is essential that the supply-side not be also politically stable (because contracts restricted by overly complex and expensive are permanent) and fair. They are, how- rules or regulations for entry. If govern- ever, neither demand-led or completely ment (central or local) can arbitrarily block free, and the market is certainly not easy to entry or expansion then a market cannot enter. The Swedish model has almost the function properly. exact opposite benefits and disadvantages, being almost entirely free and demand-led 3. Accountable but lacking financial stability and risking While government should not intervene in segregation. American charter schools are markets arbitrarily, there has to be quality harder to generalise about because laws and control for new entrants to prevent the regulations differ across states, but precisely misuse of public money. This necessarily because of this there is extreme financial and requires subjective judgements, meaning political inconsistency for groups trying to that an organisation at central or local set up inter-state federations of schools. The level will have to effectively commission or extent to which the market is demand-led authorise schools. The same organisation and free depends on the state, but charter will have to hold schools accountable if schools have proved more successful where they fail to perform once established. these features are strong. In the first three chapters we take acad- 4. Genuinely free emies and the two international models Schools in the market must have the in turn, first looking at the background to freedom to innovate while being held reform and its impact on performance, and accountable for their performance. Any then scrutinising them against our seven tests unnecessary limits on freedoms will reduce to help explain the importance of the factors the amount of choice in the market. we have highlighted. In the final chapter we suggest the ten key lessons that can be 5. Financially consistent and stable learnt by comparing the efforts of reformers Per-pupil revenue funding should be consist- in all three countries. We go on to offer a ent across the market and the formula for series of recommendations suggesting how allocating should be kept stable over time. the English schools market could be fur- Funding should be at a high enough level to ther developed in the light of these lessons. support ongoing capital and revenue costs. Obviously this requires some careful balanc- ing (between, for example, accountability 6. Politically stable and freedom) but we believe it is possible to Contracts should be designed to provide design a system combining the best features some protection for providers from chang- of the academies programme with those of es in the political weather. the US and Swedish models. Before we begin our detailed analysis it 7. Fair is worth highlighting three key themes that The market should be designed so that it reoccur throughout this report. The first
www.policyexchange.org.uk • 7 A guide to school choice reforms
is whether independent schools providers can approve and oversee charter schools, should be able to make a profit given that have perhaps found the best balance. funding comes directly from taxpayers. The third theme is ‘scalability’, a Only in Sweden are there no constraints concept that has been underplayed in on this, which explains why their market is both the academies programme and the considerably bigger than those in England Conservative Party’s proposed reforms. or the US. Even in Sweden, though, There has been too much focus on the idea profit remains a contentious issue, with the of parents, community groups, independ- centre-left Social Democrats threatening a ent schools and local businesses running review if and when they return to power. schools and not enough on the potential In the US for-profit companies are allowed for networks of federations run by educa- to operate but usually with local partners, tional companies and charities. We have seriously restricting their ability to grow. observed the same pattern in the run-up In England, despite opposition from the to reform in Sweden and the US, where it ‘Blairite’ faction within the government, was assumed that local co-operatives rather academies are unable to make a profit, than national or regional federations would thereby restricting the market. We argue take advantage of the new opportunities. that there is no basis in evidence to support Yet in both countries federations have this restriction, especially given the role become increasingly important, dominat- private companies play in the provision of ing the Swedish market and representing other public services, and that commercial up to 30% of charter schools. There is organisations should be allowed into the also evidence that schools in independent market as long as appropriate accountabil- federations outperform ‘one-off’ free and ity measures are in place. charter schools. The second key theme is the difficulty This makes complete sense. It is far more of developing an accountability structure efficient to enter a market if you develop a that balances thorough oversight with network of schools that can exploit econo- the freedom to innovate. None of the mies of scale. Running curriculum devel- countries analysed here have found this opment, teacher training and professional balance yet. The academies programme development, leadership programmes and lacks a transparent commissioning proc- behaviour management across networks is ess. Furthermore, as it has grown, it has both economically and educationally effi- become too unwieldy for the Department cient. It achieves what politicians cannot: for Children, Schools and Families (DCSF) scalability of successful practices across to manage – civil servants in Whitehall networks. Having these clusters compete have proven unable to effectively sup- with each other for further business then port sponsors spread across the country. magnifies the levels of efficiency. If a mar- In Sweden it is even harder to support ket is designed so as to only, or pre-dom- schools from the centre because of the lack inantly, include isolated ‘one-off’ entrants of attainment data. In the US many states then it will be only a little more effective require that local school districts ‘author- than monopoly provision, and far more ise’ and oversee charter schools. There is a disruptive as individual schools continually conflict of interest here, as school districts try to reinvent the wheel. It also risks more also run the public school system, and this regular market failure. We strongly believe has held back the development of the mar- that future reforms should be designed ket. Those states, like Ohio and New York, to maximise the potential for successful, which have multiple authorisers (including proven, independent providers to develop universities and educational charities) who federated networks of schools.
8 1
Academies
Background Labour immediately abolished grant-main- While the idea of using markets to provide tained status when they entered government state-funded education has been around in 1997, replacing it with Foundation sta- for over fifty years the argument was only tus. This still gives the governors ownership really joined in the UK in the 1980s, of the school but does not come with extra when the Conservative government seri- money, freedom to select or the same level ously considered introducing a schools of autonomy from the local authority. Blair voucher that parents could ‘spend’ at any and his education advisors, though, quickly school with available places (independent became frustrated by the lack of innovation or state). In the end this was too conten- in the state sector and the apparent accept- tious even for the Thatcher government, ance of permanent institutional failure by although other reforms were introduced some local authorities. After some initial in the 1988 Education Reform Act that experiments with privatising local author- created a quasi-market. The majority of ities and one school (King’s Manor in the annual schools budget was delegated Guildford)2 they settled on effectively resur- directly to schools,1 parents were given recting the CTC idea (only a few CTCs had the right to choose up to six schools by been built by the Conservative Party – the preference and information about school programme was a victim of public sector performance was published in the form of cuts in the 1990s). league tables. In addition, a small number In 2000, schools secretary David of City Technology Colleges (CTCs) - Blunkett (directed by Blair’s education independent state-funded schools spon- guru Andrew Adonis) announced that the sored by businesses - were set up and government would be seeking to replace existing schools were encouraged to go failing schools with new-build ‘city acad- 1 Though it is not and never has ‘grant-maintained’ which gave them emies’ - independent state-funded schools been delegated on per capita freedom from local authority control. sponsored by businesses and charities. The basis, as is sometimes stated 2 BBC News, Private funding Unfortunately, few schools initially applied ‘city’ was soon dropped as the government of schools under scrutiny, 23 for grant-maintained status so the govern- realised there was no reason to restrict the October 1999, see news.bbc. ment introduced an ever-juicier series of programme to urban areas. The first three co.uk/1/hi/education/482840.stm 3 Department for Education carrots (e.g. extra capital funding and the academies opened in 2002 and there were and Skills, Department for ability to select). These actually created 17 open by September 2004. The Five Education and Skills: Five Year Strategy for Children and perverse incentives for failing schools to Year Strategy for Children and Learners Learners, HM Government, go independent and distorted the mar- published in July 2004 set a target of 200 2004, see www.dcsf.gov.uk/ publications/5yearstrategy/docs/ 3 ket. The grant-maintained system was also academies by 2010, which was super- DfES5Yearstrategy.pdf not designed to attract new providers to seded in November 2006 when Tony Blair 4 BBC News, Blair wants anoth- er 200 academies, 30 November run schools and so did not dramatically announced a new target (with no times- 2006, see news.bbc.co.uk/1/hi/ increase innovation. cale) of 400 academies.4 As of January education/6157435.stm
www.policyexchange.org.uk • 9 A guide to school choice reforms
2009 there were 133 open academies, with programme developed into a fully-fledged, a further 80 planned to open in September target-driven, core activity for the DCSF. 2009.5 This had some positive effects (e.g. the Although academies have only been model became more efficient and more running for seven years their role and scalable) and some negative ones (e.g. less purpose has shifted significantly. Initially design freedom and the greater involve- the programme was designed to deal with ment of often hostile local authorities). a small number of badly failing schools in There is no question that Lord Adonis (in deprived areas. The sponsoring organisa- 2005 he was ennobled so that he could tions were given wide-ranging freedoms take up a ministerial post in DCSF) and over curriculum, design of new buildings, his officials realised that a model appro- and staff pay and conditions. At the begin- priate for a small number of schools was ning staff didn’t even need to be qualified being stretched into something it was not teachers. Furthermore, the government designed for. was very explicit that the programme was Between 2004 and 2006 Adonis and designed as an alternative to local author- Blair worked to develop additional reforms ity control. When announcing the original that would free all secondary schools from proposals David Blunkett explained that local authority control, allowing the acad- where a local authority was failing to pro- emies team to focus on replacing failing vide a good quality education: “we have schools in the most deprived areas. Had an obligation to do something for those these reforms gone ahead it would not have children and for the community, because a been necessary to expand the academies school is part of a community.”6 programme since another route to freedom As the programme grew into a corner- for schools would have been available. As stone of Tony Blair’s ‘legacy’ it changed one senior DCSF official explained to a from providing boutique solutions for the recent Blair biographer: “Blair and Adonis most troubled schools into a race to identi- wanted autonomous schools everywhere. fy enough schools to meet the target of 200 Neither wanted local authorities to have and then 400. Less money and time could any real control.”7 At one point officials be spent on each school. To cut costs, seriously considered unilaterally switching and provide a more uniform roll-out, all schools to Foundation status simulta- new academy buildings were incorporated neously, even going as far as asking the into the Building Schools for the Future churches (who run most Voluntary-Aided programme, giving sponsors less control schools) for their support.8 over design. The first academies had been Unfortunately, under intense pres- built in the few authorities receptive to sure from the teacher unions and the
5 DCSF. Currently the website the idea but as the plan expanded more Parliamentary Labour Party, the even- shows 86 academies in feasibil- hostile authorities had to be involved. This tual 2006 Education Act was significantly ity or implementation, including several planned to open in meant concessions, like allowing them to watered down from Blair and Adonis’ 2010 and later. www.stand- ‘co-sponsor’ academies. The number of original plans.9 A new ‘Trust’ status was ards.dfes.gov.uk/academies/ projects/?version=1 sponsors also had to be increased, which developed that was broadly similar to 6 BBC News, Anger at scheme led to unedifying offers of honours to Foundation status but incorporated, from for failing schools, 15 March 2000, see news.bbc.co.uk/1/hi/ potential backers (kicking off the ‘cash academies, the idea of partnering with local education/677996.stm for honours’ scandal) and a shift away charities, institutions or businesses. Trust 7 Seldon A, Blair Unbound, from projects supported by individual mil- schools, however, do not have the same Simon and Schuster, 2007, p 420 lionaires towards multi-academy groups, freedoms over their curriculum and teach-
8 Private information universities, independent schools and even ers’ conditions as academies, nor does the 9 Seldon, pp 419-427 other state schools. In other words the partner have the same level of control as an
10 Academies
academy sponsor. In addition, the school sionally hostile approach from the DCSF, is still regulated by the local authority. will incrementally reduce the difference Hence the academies programme remains between future academies and other types the key mechanism for introducing alter- of schools. The opposition parties have native provision and competition to the been quick to seize on this, offering more English school system. radical plans that are closer to the spirit Though Blair made continuing support of Blair and Adonis’s initial intentions for academies a condition of his retirement than current government policy. The in favour of Gordon Brown,10 there was a Conservative Party have promised to give widespread expectation following the lat- academy-style freedoms to 400 of the top ter’s coronation that the programme was performing schools in the country14 while in jeopardy. It was well known that Brown providing capital funding for 220,000 had always been sceptical about academies places at new ‘academies’ and revenue and that the Treasury had tried to block the funding for any school set up by charities Blair speech in which the target had been or voluntary groups in response to local raised to 400.11 The Brown government demand.15 The Liberal Democrats have was never likely to scrap the programme come up with similar proposals, though completely, but the early signs were not they would not provide any capital fund- encouraging for supporters of academies. ing for new schools. Brown’s closest advisor, Ed Balls, was Thus there seems to be a growing divide made schools secretary, and in his first between the Labour model of alternative speech he reduced the curricular freedoms provision (used only to replace existing of future academies.12 At the 2007 Labour failing schools according to central gov- Party conference Balls reportedly spoke ernment direction) and the opposition about bringing academies back into the model of a looser, demand-led system. We local family of schools – a line used often will return to the issue of further reform by opponents of the programme. There in England in the final chapter. In the has certainly been a significant increase remainder of this chapter we will give a in the involvement of local authorities in brief overview of the relatively scant data the programme since Brown took over on how academies are performing and 10 Private information
(discussed in more detail on p.17). While then look at how well the existing acad- 11 Seldon, p 507
Lord Adonis initially kept his job he has emies programme performs against the 12 Hansard, 10 July 2007: since been moved to the Department seven tests identified in the introduction. Column 1321 of Transport. Two other fervent sup- 13 Garner R, Resignation casts doubt on academies, The porters of academies within the govern- Performance Independent, 8 November 2008, ment – Sir Cyril Taylor, formerly Chair There is not yet enough evidence to draw see www.independent.co.uk/ news/education/education- of the Specialist Schools and Academies firm conclusions about the long-term news/resignation-casts-doubt- Trust and Sir Bruce Liddington, formerly impact of the academies programme, as on-academies-1001068.html 14 As long as they agree to Schools Commissioner – have also left there are only 47 academies with two or partner with a failing school, their jobs, triggering academy sponsor more years of GCSE results. However, it is see Prince R, Tory party confer- ence: Schools to get budget worries that there are no ‘champions’ left clear that the short-term impact is almost freedom under Conservative to fight their cause.13 always positive. Most of the data we do plans, The Daily Telegraph, 30th September 2008 www.tel- The cumulative effect of these changes have comes from the five annual evalu- egraph.co.uk/news/newstopics/ has been to create real uncertainly among ations undertaken for the government politics/conservative/3108378/ Tory-party-conference-Schools- current and potential sponsors. Their con- by PricewaterhouseCoopers (PwC). Their to-get-budget-freedom-under- cern is that the unplanned expansion of final report, which provides trend data on Conservative-plans.html 15 The Conservative Party, the programme under Blair, combined the first 27 academies opened by 2005, was Raising the Bar – Closing the with a increasingly institutional and occa- published in November 2008. The data Gap, November 2007, pp 36-41
www.policyexchange.org.uk • 11 A guide to school choice reforms
shows that the number of students achiev- behaviour improves. Such changes could, ing 5 A*-C GCSEs at these academies though, explain some of the improvement is increasing, on average, at 8% a year.16 in academic results. There have been no This is four times as fast as the average rate published studies, as yet, looking at wheth- of improvement for English schools and er academies boost attainment for children twice as fast as schools with a similar pro- once other factors, like income and ethnic- file.17 There has also been an average annual ity, have been taken into account. increase of 5% in the number of students While there is nowhere near enough achieving 5 A*-C including English and information yet to draw any firm conclu- Maths, the Government’s preferred meas- sions, the early data suggests that multi- urement. This is five times faster than the academy groups might be outperforming national average and over twice as fast as individual academies, highlighting the key schools with a similar profile.18 Figures for importance of efficiency and scalability results at Key Stage 3 (14 year olds) show that such groups offer (see p.18 below). similar rates of improvement.19 Of those academies in groups that have at Academies have also proved much more least two years of GCSE data, the number popular with parents than their predecessor of children achieving five good GCSEs has schools and are now nearly all oversub- increased 23% since two years before it scribed. The 27 schools in the PwC study became an academy. The equivalent figure have an average of 2.6 applications for every for academies not in groups is 17%.23 available place and just two are undersub- scribed.20 The most oversubscribed has 7.1 The Seven Tests applications for every place (1281 applica- 16 This data excludes 3 three schools from the sample of 27 tions for 180 places). Some of the most 1. Demand-led because they were previously CTCs successful academies opened after 2005 The process of setting up an academy is
17 PricewaterhouseCoopers, have shared similar success. Mossbourne controlled by central and local government Academies Evaluation – 5th in Hackney, for example, is more oversub- and rarely involves anything more than Annual Report, Department for Children Schools and Families, scribed than most grammar schools, with desultory consultation with parents. Thus November 2008, p 211 7.8 applications for every place.21 the system is only very indirectly demand- 18 Ibid Because of this popularity academies led in that it seeks to replace schools that 19 Ibid. p 202 have become more comprehensive, taking are failing (according to a very crude defi- 20 Ibid. pp 40-41 children from a wider range of social cat- nition), which one might assume would be 21 www.mossbourne.hackney. sch.uk/folders/our_academy/ egories when, typically, their predecessor unpopular with local parents. There is no jobs.cfm school took a disproportionate number of other assessment of demand. 22 PricewaterhouseCoopers, children from the lowest social-economic The process of developing a new acade- ibid. p 45-6 groups. Data from the PwC study shows my starts at the DCSF with a list of all the 23 We excluded CTCs as their conversion to academy status a drop in the number of children on Free secondary schools in England considered is purely semantic – they have School Meals (FSM) from 42% in 2002 to to be failing (i.e. those with fewer than had the relevant freedoms since the early 90s. Our definition of a 35% in 2008 (which is still considerably 30% of students achieving 5 A*-C includ- group is a organisation running 22 24 two or more academies with a higher than the national average of 13%). ing English and Maths at GCSE). These stated purpose of providing cen- Opponents of market-based reforms have schools were only publicly identified in tral support for their schools argued that this proves academies are 2008 – as the 638 participants in the 24 This description is based on conversations with current covertly selecting. This is nonsense, as the slightly Orwellian ‘National Challenge’ – and former civil servants at the data is merely showing they are becoming but the list has been used as the starting department more attractive to the middle-classes, which point for academies since the inception 25 The department is now very careful not to use that word fol- leads to more socio-economically balanced of the programme. This definition of fail- lowing the outcry over the han- 25 dling of the National Challenge classrooms and can encourage undera- ure is of course woefully simplistic as it announcement chieving children to flourish as classroom assumes, for example, that an inner-city
12
Academies
school achieving 25% good GCSEs is authority consent they have never sought performing worse than a suburban school to do this. Instead they try to cajole scoring 35%, which is debatable to say the authorities into participating using a vari- least.26 Parents will have more localised ety of carrots (e.g. extra capital funding definitions of success and failure that are through Building Schools for the Future entirely ignored when one metric alone or National Challenge money) and sticks is used to decide the dynamics of a local (e.g. the threat of privatising all or part of market. One hopes that the ‘school report the authority). cards’ that the DCSF is considering intro- These methods do not always work ducing (following a recommendation by against determined ideological opposition. Policy Exchange in 2008) will bring a A number of authorities that contain a con- new level of nuance to school improve- siderable number of National Challenge ment policy, but for now it is based on a schools do not have a single academy, crude and arbitrary cut-off point.27 either running or planned including Hull The next step for the DCSF is to per- (containing seven National Challenge suade the local authorities containing schools), Stoke-on-Trent (with six) and the schools identified as underperform- Northamptonshire (with five) each – see ing to participate. While, under acad- Table 1 for a full list of authorities with emies legislation, the DCSF has the free- no academies opened or planned and how dom to take over a school without local many eligible schools they contain.28
Table 1: Local Authorities with no existing or planned academies
Number of Secondary Number of Percentage of Schools (2008) exclud- National Challenge National Challenge Local Authority ing special schools schools (2008) Schools Hull 16 7 44% Stoke-on-Trent 17 6 35% Northamptonshire 49 5 10% Blackpool 9 4 44% Durham 39 4 10% East Sussex 38 4 11% Kirklees 33 4 12%
Knowsley 11 4 36% 26 The department occasion- ally criticises higher performing Plymouth 18 22% 4 schools that are “coasting” but Sefton 25 4 16% all available sanctions, including replacement by an academy, are Cambridgeshire 44 3 7% targeted at National Challenge schools Hampshire 94 3 3% 27 Lim C and Davies C, Helping Somerset 41 3 7% Schools Succeed: A Framework for English Education, Policy Stockton-on-Tees 17 18% 3 Exchange, 2008, pp 58-72 ; Suffolk 52 3 6% DCSF, A School Report Card: consultation document, 2008, Warwickshire 41 3 7% see publications.teachernet.gov. uk/eOrderingDownload/DCSF- Worcestershire 45 3 7% 01045-2008.pdf
Bournemouth 12 2 17% 28 Figures from the Department for Children Schools and Cornwall 38 5% 2 Families website, see www.dcsf. gov.uk
www.policyexchange.org.uk • 13 A guide to school choice reforms
Dorset 29 2 7% Hartlepool 6 2 33% North Yorkshire 53 2 4% Redcar and Cleveland 11 2 18% Southend 14 2 14% Torbay 10 2 20% East Riding 21 1 5% Gateshead 13 1 8% Halton 8 1 13% Leicestershire 28 1 4% North Somerset 12 1 8% North Tyneside 12 1 8% Northumberland 16 1 6% South Tyneside 9 1 11% Warrington 12 1 8% West Berkshire 15 1 7% York 13 1 8%
Meanwhile other, more co-operative, As the programme grew larger such a authorities have benefited from huge personalised service was no longer feasible. amounts of investment. Parents have no Moreover, as we have seen, when Ed Balls control over this process and cannot force became Education Secretary he insisted the hand of their local authority. They are that local authorities be given a greater role not even made aware when discussions in the process, leading to the introduction between local and central government take of a system of local authority commission- place. The 2006 Education Act introduced ing that has now been running for a couple a new duty for authorities to respond to of years. Since 2007, once an authority parents who asked for a new school in has agreed to allow one of its schools to their area (which would not have to be an be rebuilt as an academy it is given a list academy), but there is no duty to act; the of potential sponsors sounded out by the response can be a simple “no”.29 DCSF. This is still a restrictive process as Once an authority has agreed to participate the DCSF will only discuss sponsorship in the programme a sponsor has to be found. with individuals or organisations it thinks Until 2007 Lord Adonis and his team would are interested (there is no public call for personally offer sponsorship opportunities to bids). Sometimes the authority involved individuals or groups they thought might be will alert local groups and schools to the interested; the local authority involved had possibility of bidding and in the last few 29 Department for Children Schools and Families, Duty little say. This ‘matchmaking’ service worked years there has been a spate of bids from to Respond to Parental well when the programme was in its infancy consortia of local schools and further edu- Representations about the Provision of Schools: A Guide as there were few sponsors prepared to risk cation colleges. for Local Authorities, see www. dcsf.gov.uk/schoolorg/data/ participating in an experimental new scheme. Local authority officers and council- guidance_Documents/duty%20 Furthermore, it was in no sense demand-led lors then debate the merits of the bids to%20respond%20to%20 parental%20representations%20 as the choice of sponsor was not related to before choosing. Again there is no public 2008-03.25.pdf any measure of parental consultation. consultation. The DCSF will not even
14 Academies
release details of schools being considered ing degrees of sincerity. A few authorities, for academy status or potential sponsors such as Sheffield, have balloted parents for that academy until the authority have (though always retaining the final say).34 made their decision and the sponsor has At this point, though, local parents can signed an ‘expression of interest’. This is only reject the new school along with the done to protect the sponsor until they promise of millions in new investment, have decided to commit, but it is unques- or accept. This is especially problematic tionably anti-competitive.30 Authorities do if the sponsor is particularly controversial not even have to retrospectively publish (for example a religious group or a char- reasons for their decisions. Durham is the ity endorsing an experimental pedagogical only authority to publish a (non-binding) approach) school. Most authorities have independent evaluation of bids.31 This not even gone this far, preferring the usual opacity also means that decisions could round of public meetings which tend to be made on sponsors for entirely arbitrary be dominated by a combustible mix of (and potentially political) reasons. supporters and ideologically driven anti- Involving the local authority in this way academy protestors. is a serious structural flaw given that the In short, under the academies model new academy will be competing against parents have virtually no means to instigate authority-run schools. From the start an the development of alternative provision in authority has an incentive to reduce poten- their area. If they are lucky enough to be tial competition by choosing a compli- living in an authority that has decided to ant sponsor. Given that the purpose of participate in the programme they have no the programme is to allow innovative say over the nature of the academy (big or and dynamic new providers to compete small, single sex or mixed, all-through or 30 Anti-academy campaign- with the monopoly system controlled by 11-16/18 etc) or the sponsor. It is a process ers have taken the DCSF and authorities, it seems strange to let the very tightly controlled by government and not Camden council to judicial review over the Swiss Cottage same authorities decide the criteria for at all responsive to demand. academy (co-sponsored by participation. One academy group who Camden and UCL) on the grounds that, by not putting recently expressed interest in bidding for 2. Easy to Enter the sponsorship contract out an academy were sent a list of authority There are a huge number of barriers for to tender, the government have broken EU competition law. The priorities which their bid had to meet. potential alternative providers in the UK government argue that because 35 the sponsors do not make a Improving achievement for children was schools market to overcome. Outside of profit competition law doesn’t listed fourth after (a) reducing health ine- the academies programme the market is apply. See Curtis P, Academy expansion under threat, 5th qualities; (b) improving sexual health and almost impenetrable. Technically authori- November 2008, www.guardian. reducing unplanned teenage pregnancies ties have to initiate a competition open to co.uk/education/2008/nov/05/ camden-school-case and (c) improving provision for children any bidder if they wish to develop a new 31 Smith MJ, Independent 32 with SEN. This is exactly the kind of school, theoretically offering another route Evaluation of Bids From Potential absurd process the academies programme for potential providers. However, there Sponsors of Durham Academies, Department for Children Schools was supposed to circumvent. The 2008 are plenty of methods authorities can use and Families, February 2008 PwC evaluation of the programme noted to avoid doing this. For a start very few 32 Private information tactfully that this method of commis- authorities would even think of opening 33 PwC report, p 12 sioning has led to “a lack of clarity about a new school unless forced to through a 34 Sheffield Telegraph, Third City Academy plan a step closer the respective roles and responsibilities of significant rise in the local population; after vote, 12 December 2008 Local Authorities and Academies”.33 they would rarely if ever add surplus places see: www.sheffieldtelegraph. co.uk/news2/Third-City- Only after a bid has been successful and in an attempt to develop competition for Academy-plan-a.4784672.jp an expression of interest has been signed is existing failing schools. 35 These barriers were the focus of Policy Exchange’s report the deal made public, at which point there Even if additional places are required, Choice? What Choice? pub- is usually a consultation exercise of vary- authorities have a number of alternatives lished in November 2007
www.policyexchange.org.uk • 15
A guide to school choice reforms
to initiating a competition. First, they providers to get access to the market. can opt for an academy – and there Unfortunately, potential sponsors face is evidence of authorities preferring this another intimidating set of barriers. For a route because they can co-sponsor an start they are not allowed to make a profit, academy and choose the other sponsor, thereby excluding all organisations that are effectively keeping the school under their not charities (or do not have a charitable control.36 Secondly, they can expand or wing). It is important to note that there is replace an existing school, which usually no legislative reason why they should not means an exemption from competition. be allowed to make a profit – under the Since September 2006 the DCSF has law they are simply defined as independ- received 100 applications to publish new ent schools.40 There is also no reason why a 36 Sturdy E and Freedman S, school proposals without a competition of commercial company could not bid to run Choice? What Choice? Policy Exchange, 2007, p 30 which 84 were approved – mostly because a school in a local authority run competi- 37 37 Department for Children they were replacement schools. The £45 tion and, if they won, take a profit (though Schools and Families, Outcome of Exemption (s10) Requests billion Building Schools for the Future they would be highly unlikely to win). for New Schools – last programme – which seeks to rebuild or Curiously, for-profit firms are allowed into updated 2 December 2008. HM Government, 2008, see www. renovate all secondary schools in the coun- other parts of the education sector running dcsf.gov.uk/schoolorg/guidance. try by 2020 – has provided authorities nurseries (73% of which are for-profit),41 cfm?id=2 with hundreds of millions of pounds each special schools and pupil referral units for 38 Choice? What Choice? Ibid. p 33 to redevelop their school estate on the basis children unable to participate in main- 39 Private information of projected pupil numbers, meaning that stream education. Academies are different 40 The City Technology (if they get their calculations right) most because each contract is structured so that Colleges, upon which the acad- emies programme is founded, of them can avoid competitions for years the sponsor has to be a charity. This is for were introduced in the 1996 to come. purely political reasons; the government Education Act. Under the provi- sions of this act state funding On the rare occasions when a competi- does not want to be accused of privatising for these independent schools tion does become necessary, authorities education or get caught up in an emo- is reliant upon education being free to all but allows for “other are allowed to bid to run the school them- tive argument over whether organisations conditions and requirements selves. The School’s Adjudicator is then should be allowed to profit from children. in relation to the school as are specified in the agreement“ responsible for deciding the winner but Tony Blair and his advisors were well (c.482 City technology colleges the authority is responsible for the consul- aware when deciding on their educational and city colleges for the technol- ogy of the arts). For examples of tation process - a glaring design flaw. The reforms that there was no logical reason to these funding agreements see www.dcsf.gov.uk/foischeme/ story of the 2006 competition in Haringey, exclude for-profit companies. Indeed, in subPage.cfm?action=collections. which was won by the authority, is given Blair’s first term, when his political capital displayCollection&i_collectio- nID=190 in Policy Exchange’s report Choice? What was high, they were far more adventurous, 38 41 National Day Nurseries Choice? published in 2007. Suffice it to fully privatising a number of failing local Association, “Childcare: a say that even the Adjudicator who awarded authorities and even one school, King’s sustainable mixed economy 42 – achieving the vision for Haringey the contract complained that Manor in Guildford. Moreover, there is children and families”, Policy they had more or less rigged the consul- plenty of evidence that this was working. Paper, September 2007, see www.ndna.org.uk/Resources/ tation. Other educational organisations, A recent analysis of the local authorities NDNA/Advice%20and%20 info/Policy%20Paper%20 like Haberdashers and CfBT, who bid for privatised during Blair’s first term from (September%202007).pdf the contract found the process pointless the CBI shows that they have improved 42 BBC News, The Guildford and expensive and stated that they were significantly faster than other authorities. Experiment, 4th May 1999, see news.bbc.co.uk/1/hi/educa- unlikely to participate in further competi- In the nine privatised authorities, Key tion/specials/private/271718. tions in which the relevant authority was Stage 2 results in English, maths and sci- stm. The school has now been 39 renamed King’s college and is a bidder. ence improved by 25% between 2001 and going from strength to strength: The failure of the competition mecha- 2008. In England as a whole, the average www.kingscollegeguildford. com/?p=about_us nism means that sponsoring an academy level of improvement over the same period 43 Information from the CBI is the only realistic way for alternative was only 14 per cent.43 Of course, other
16 Academies
parts of the public sector have also been pedagogy and then go about raising the opened to private contractors in recent money to sponsor without any prospect of years: prisons, hospitals and welfare-to- future profits. work. The evidence in these sectors is There is some evidence that these blocks also positive. For example, in his highly to entry are causing real problems for influential report on welfare reform David the DCSF in trying to reach their artifi- Freud noted that private brokers operat- cially limited target of 400 academies (the ing programmes through the new deal for 2008 PwC evaluation notes “the planned disabled people have been more successful expansion to 400 academies will create at getting people back to work than volun- challenges for the DCSF in securing suffi- tary or charitable providers.44 cient numbers of high quality sponsors”46). However, in 2000 the academies pro- As of January 2009 there are 133 acad- gramme was considered small-scale and it emies open and, in October 2008, Schools was assumed that enough charitable spon- Minister Jim Knight said that he anticipat- sors would be found to avoid the tricky ed a further 80 opening in 2009 and 100 issue of profit. By the time further reforms in 2010,47 yet according to current DCSF for the 2006 Education Act were being dis- figures just 86 are slated to open in total.48 cussed (which led to the creation of Trust Future expansion has now become wor- schools – see p.10 above) Blair was under ryingly dependent on the sponsorship of much more pressure from the left-wing of higher and further education institutions, the Labour party. Officials and the most local authorities and the church. radical ministers (including Lord Adonis In the past two years, as existing groups and John Hutton) knew that allowing have started to reach capacity and paid profit would provide a significant boost to sponsorship has dropped off following the the market but the politics were considered cash for honours scandal, there has been unworkable. As one Number 10 aide put a sharp shift towards sponsorship by uni- it: “Intellectually, Tony Blair saw that versities and local authorities (often with we were artificially limiting the supply of a compliant ‘lead’ sponsor who is happy schools by not letting private companies to take a back seat) and away from multi- into education to make a profit. But politi- academy groups and individuals. Just 8 cally he realised it was going too far.”45 out of the 133 academies already opened Not only are sponsors barred from mak- were sponsored by universities (6%) com- ing a profit, they also have to raise a£2 pared to 20 of the 86 now in feasibility or million sponsorship fee that they cannot implementation (23%). For local authori- recoup. Some sponsors, particularly multi- ties the figures are 10 out of 133 (7.5%) 49 academy sponsors, have cut individual and 19 out of 86 (22%) respectively. The 44 Freud D, Reducing depend- deals with the DCSF to reduce the cost shift in sponsorship type is illustrated by ency, increasing opportunity: options for the future of welfare- (which is a spectacularly anti-competitive Graph 1. to-work, Department of Work practice), but most have to pay. This lim- This shift – forced by the unnecessary and Pensions, 2007, p. 54 its the range of sponsors to three broad imposition of barriers to entry – risks 45 Seldon, p 421 groups: religious organisations with an changing the entire nature of the acad- 46 PwC, p 12 47 www.theyworkforyou.com/ established philanthropic approach to edu- emies programme. As previously discussed, wrans/?id=2008-10-09a.225415. cation, independent schools and universi- for local authorities to be sponsors com- h&s=speaker%3A11036+section %3Awrans#g225415.q0 ties (who are exempt from paying the fee) pletely distorts the original purpose of the 48 www.standards.dfes.gov.uk/ and a small number of rich individuals or programme: to break their monopoly on academies/projects/?version=1 companies with a specific interest in edu- education provision. Involving universities 49 Under our categories it is possible for an academy to be cation. It would be extremely difficult for a as sponsors is less perverse but still risks sponsored by more than one new organisation to develop an interesting diluting the programme as they cannot type of sponsor (see below)
www.policyexchange.org.uk • 17 A guide to school choice reforms
Graph 1: Change in type of sponsor over time50
45% Up to 2008-9
40% 2009 onwards
35%
30%
25%
20%
15%
10%
5%
0% h
Churc Group Company University Individual Guild/Charity Schools and Local Authority FE Colleges
provide the scalability that organisations academies who could conceivably expand founded for the specific purpose of run- (e.g. the RSA, Edge and CfBT) have made ning schools can (the same goes for inde- it clear that they do not wish to open any pendent schools). more. A significant number of potential Multi-academy groups i.e federations multi-academy sponsors, like SERCO and running more than one school have always the Swedish company Kunskapsskolan,53 been the ideal sponsor for academies as they are put off by the inability to make a profit can provide this scalability and back-office or the need to pay a fee. efficiency. PwC note that multi-academy Others are held back by the restrictions sponsors offer “educational experience of the academy model and specifically the and established infrastructure support, need, in most cases, to replace an exist- 50 Data is taken from the most up-to-date DCSF academies efficiencies and opportunities for shared ing large comprehensive secondary school. spreadsheet. Each project was then codified by sponsor type: resources”, they are “particularly good For example, Steiner have one academy, schools can each have more for leadership succession as leaders can converted from a fee-paying independent than one sponsor and sponsor- type. For example, there were be grown and developed within the wider school in Herefordshire, but they would like 171 separate entries for sponsor network” and they provide “opportunities more. Unfortunately their model requires type in the data from academies already open (133 schools), and for learning and sharing across the com- all-through single form entry schools so 146 entries in the 2009 onwards munity of schools”.51 Nevertheless, there they cannot easily takeover existing state data (86 schools). The latter data set included many more multiple are only five groups with five or more acad- secondary schools. Some of the multi-group sponsor academies. The Guild/ emies opened or planned: United Learning sponsors, notably ARK, have stretched the Charity category includes both livery companies and charities Trust, ARK, OASIS, Harris Federation one-size-fits-all model to its limits (by, that are not primarily academy focused. and British Edutrust. Between them these for example, developing a group of small
51 PwC, p 106 five groups run 35 of the 133 academies schools on the same site), but the need to
52 www.standards.dfes.gov.uk/ currently open and have a further eighteen replace an existing school rather than open academies/projects/?version=1 in feasibility or implementation.52 a new one is a significant restriction. 53 Kunskapsskolan are in nego- Some of these organisations are already Such hurdles are worsened by a lack of tiations to open two academies in Richmond as a global “shop close to operating at full capacity and there standardisation across local authorities. window” for their educational model. They would open many is little sign of new groups appearing. A Sponsors have told us about the differ- more if they could make a profit. number of sponsors operating one or two ent information required of them in the
18 Academies
bidding process. While some authorities tion mark over whether sponsors with no work with selected bidders to develop educational experience should be allowed suitable approaches, others demand large to participate in the programme.56 amounts of material in incredibly short Once an academy is up and running periods of time and the subject matter they are accountable to central govern- covered also varies enormously.54 This ment in the same way as other schools: situation is bad for all parties: local through Ofsted and assessment data (they authorities are liable to lose potentially have to participate in Key Stage 2 tests and good bids when sponsors walk away in GCSEs). The sponsor is also held account- bewilderment, large group sponsors are able through their contract or ‘funding unable to realise the economies of scale agreement’. The DCSF can cancel the that underpin their model and smaller, contract if the academy is put into ‘special diverse pedagogy or parent-run groups are measures’ by Ofsted, if the sponsor is at risk unable to cope with the complexity. of going bankrupt or if the sponsor does These severe supply-side restrictions not fulfil the duties outlined in the agree- for alternative providers in England have ment. If the academy has not infringed caused serious problems for the roll-out on any of these requirements the DCSF of the academies programme and have must give seven years notice if it wishes held back the development of efficient to cancel the contract (as must the spon- scalable models. sor). Of course this responsibility becomes increasingly difficult for the DCSF to 3. Accountable manage with every new academy. While As we have seen, the process of setting up it is possible to have a close relationship an academy is shrouded in secrecy. The with twenty or fifty schools, the systems of public do not get to know about a new accountability become mechanistic and it proposal until a sponsor has been selected. is harder to spot when sponsors might be They are not told who the other bidders getting into difficulty when the number of were (if there were any), or the justification academies far exceeds this. Regrettably, this for why the winner was chosen. Moreover, has become increasingly apparent in recent there are no clear criteria for sponsors months with several academies such as the beyond having to be charitable. There are Richard Rose Central Academy in Carlisle no published guidelines on the extent to and the OASIS academy in Southampton which a sponsor can seek to impose an running into trouble.57 external religious or political agenda on Though academies are commissioned by a school. This lack of clarity has led to local authorities they are not, thereafter, controversy over a number of academies, accountable to them (unless the authority especially those run by the Emmanuel is a co-sponsor), which just shows how 54 Private information Schools Foundation set up by the evan- illogical the 2007 changes to the commis- 55 BBC News, Creationism ‘no gelical Christian philanthropist Sir Peter sioning process are. It is right to protect place in schools’, 11 April 2006, see news.bbc.co.uk/1/hi/educa- Vardy, which have been accused (falsely) academies from local authority regulation tion/4896652.stm 55 of teaching creationism. for the same reason that it is wrong to 56 PwC, p 13
If the process of appointing sponsors was allow authorities to act as commission- 57 http://www.telegraph. less secretive and more demand-led then ers: they are direct competitors. However, co.uk/education/secondarye- ducation/3727195/Emergency- controversies such as this would be far less this does leave academies entirely unac- inspection-launched-at- likely to arise. Currently there is no public countable to their local population. Again, academy-following-gang- fights-and-bullying.html; assessment of the pedagogical approach of parents are dependent on the DCSF and http://www.tes.co.uk/article. individual sponsors (if they have a specific Ofsted’s definition of failure rather than aspx?storycode=6004180; http://news.bbc.co.uk/1/hi/edu- approach at all) and there remains a ques- their own, more nuanced impressions. cation/7856985.stm
www.policyexchange.org.uk • 19 A guide to school choice reforms
4. Genuine freedom developed their own pay structure and Academies have considerably greater ARK are looking into developing one as autonomy than other state-funded schools. well. Many academies have used their However, they are still more constrained freedoms to increase the length of the than (most) charter schools in the US and school day, and most of them pay extra to Swedish free schools, particularly since the compensate for this. Government has chipped away at some The first few sponsors also had consid- of their initial freedoms over the last few erable freedoms over the design of their years. As PwC put it, “the reining back of academy. They were allocated a specified the flexibilities originally granted to acad- amount of money and then allowed to emies is an important policy shift”.58 employ their own project managers and Initially they had complete freedom architects under the ‘Design and Build’ over the curriculum (those set up before framework. In mid-2006, however, acad- 2007 still do). However, in his first emy construction was integrated with speech as Education Secretary Ed Balls the Building Schools for the Future pro- announced that new academies would gramme and control was handed to the have to follow the National Curriculum in Partnership for Schools (PfS) quango. As English, Maths, Science and Information the DCSF put it at the time: “academies Technology.59 This is not a problem for will now be included in local authori- most sponsors, who do this anyway,60 but ties’ estate planning [which] will allow it is for a few such as Steiner who have a more integrated implementation of their radically different pedagogy which does not strategic vision for secondary education fit at all with the National Curriculum.61 provision across the local authority” – Sponsors also have freedom over teacher another example of the illogical fudging of pay and conditions for new staff but usu- academy independence.63 Sponsors were ally have to transfer existing staff to the informed that they would now “have a 58 PwC, p 12 new school with their previous terms and limited role during the academy construc- 59 Hansard, Col 1322, 10 July 62 2007 conditions. They also have to offer the tion, but will be informed of progress and 64 60 The biggest problem with Teacher Pension Scheme which restricts will be consulted when required.” New the curriculum is not the subject their freedom to significantly alter the salary builds are now assigned a project manager content but the vast array of add-ons – like financial literacy structure (for example, by paying a higher by PfS and architects have to be selected – which ordinary schools have to do but from which academies salary to young staff in return for a reduced from an approved list. While some of the are exempt. pension contribution). Interestingly few bigger multi-academy sponsors have been 61 Steiner would also rather not sponsors have made extensive use of their able to use their muscle to stretch the rules use GCSEs in their academy, but have to do so for assess- freedoms over pay to-date (except with on project management, most have not. As ment purposes. headteachers who are paid considerably one sponsor put it: “The BSF programme 62 The Department’s policy is more than other state school principals), is locking the door on any genuine diver- that Academy projects which involve the closure of an exist- which may be because the majority of their sity for decades to come. It is completely ing school or the merger of 65 two or more schools should be staff are still holdovers from the previous top-down.” conducted on the basis that the management and thus have to be paid TUPE (Transfer of Undertakings, Protection of Employment) according to the national pay agreement. 5. Financially consistent and stable Regulations apply, unless there There are signs, though, that some of the Each academy receives a General Annual are exceptional circumstances which render this inappropriate. multi-academy sponsors are increasingly Grant (GAG) made up of the following www.standards.dfes.gov.uk/ seeing this freedom as way to differentiate elements: (a) school budget share which academies/faq/?version=1 themselves and attract high-quality staff. is equivalent, per-pupil, to the amount 63 Choice? What Choice? p 21 The Harris Federation pays staff an annual that other secondary schools receive in 64 Guidance to Academy Sponsors on PfS, DCSF, 2006 ‘Harris allowance’ on top of standard pay that local authority; (b) local authority 65 Choice? What Choice? p 21 scales, the United Learning Trust has ‘hold back’ grant which is equivalent to
20 Academies
the top-slice authorities take from most US who, in most cases, have no funding schools (usually around 10%) to pay for until pupils arrive at the school. central services; (c) some Standards Fund The level of financial support for acad- and Standards Grant funding that is paid emies means that there is little risk of to all maintained schools from central gov- sponsors running into difficulties unless ernment; (d) additional money to cover they are spectacularly badly managed. To higher insurance and VAT costs (as inde- date only one sponsor has negotiated to pendent schools academies are liable for end their contract, the building company VAT, unlike maintained schools); and (e) Amey who set up the Unity City Academy extra start-up funding in the first year(s) in Middlesbrough in 2002 (the first acad- of operation to cover new equipment and emy to open), who claim this has nothing over-staffing if the academy is new and to do with finances, although they have therefore opening with fewer pupils than not given any other explanation.68 the planned total.66 All of this adds up to an amount 6. Politically stable greater than state secondary schools in Academy contracts do not come up for their authority because of item (b) the automatic renewal, meaning that sponsors hold back grant.67 For a reasonably sized are protected to some extent from changes secondary school this means an extra in the political climate. Furthermore, it £500,000 or so a year. While, technically, means that a future government could this money is supposed to cover services not change the contractual basis for exist- that other schools would receive from ing academies. The current government their authority (like SEN support), there have actually tried to retroactively alter is usually enough left over to top-up staff contracts, but if sponsors refuse the DCSF salaries and to pay for other amenities. have no legal recourse. If a future govern- Some multi-academy sponsors get their ment was extremely hostile to academies schools to pay some money annually to a they could legislate them out of existence central back-office to provide support on altogether. That said, this would be a risky curriculum, human resources, IT and so course of action, especially if (like gram- on, but it is rarely as much as the author- mar schools) they continue to be very ity ‘hold-back’ (and is usually better value popular with parents. The security of the for money). contracts makes incrementally weakening Academies also receive funding to cover existing academies very hard. maintenance (exactly how much depends However, as we have seen, new acad- on their building contract and whether it is emies have fewer and fewer freedoms each funded through PFI) and, crucially, imple- year and there is a real risk that, because mentation funding that covers transition the programme’s expansion is tightly con- management and staffing (new headteach- trolled by central government and is not ers are often hired up to a year before the demand-led, academy status will cease to new academy opens to develop new strate- mean anything over time. gies on curriculum, behaviour etc.) The 66 www.lgfl.net/lgfl/leas/ islington/web/Schools%20 amount of funding available to support 7. Fair Circular/2007-05- implementation decreased after Tony Blair One of the central arguments of oppo- 21/2iiapd1responsedfesredsg. doc increased the academy target from 200 to nents to academies has always been that 67 Freedman S and Horner 400 (the Treasury would only support a they will increase segregation in the school S, School Funding and Social cost neutral proposal). Nonetheless, it system because sponsors will be eager to Justice, Policy Exchange. 2008, p 30 still gives sponsors a huge advantage over manipulate their intake to boost results. 68 news.bbc.co.uk/1/hi/educa- not-for-profit providers in Sweden and the However, academies have to follow the tion/7663484.stm
www.policyexchange.org.uk • 21 A guide to school choice reforms
same admissions code as all other schools The only concern raised by PwC is that (since 2007) and have never been able to in 2007 the academies in their study per- select academically.69 Moreover, the avail- manently excluded a higher percentage of able data shows that there is absolutely no their population than the national average evidence of ‘cream-skimming’. of 0.22% (PwC do not give the percent- In the 27 academies looked at by PwC age excluded from academies). However, the percentage of children eligible for free there were considerable differences between school meals (FSM – a standard proxy academies, with six excluding far more than for disadvantaged students) had decreased the national average and the rest much from 42% to 35% since they took over closer to the average. In addition, as the from the predecessor school. But this is authors of the report explain, “they may still much higher than the national aver- be applying low threshold approaches to age of 13%. More significantly it is still misbehaviour in their early days in order a higher percentage than the average for to establish a culture of excellent behaviour their catchment areas. Indeed, the total and achieve rapid improvement in learning number of students on FSM at these acad- and attainment: if so, this may well be a emies has increased from 6,581 to 7,995 positive approach in the longer term, but as the academies are typically bigger and with the short term consequences of higher fuller than predecessor schools. The aver- than average rates of exclusions”.73 This is age number in each academy has increased certainly the main explanation given by from 313 to 333 while the average number academy sponsors, who also cite a higher
69 The government created in local schools with overlapping intakes initial number of students who have previ- unnecessary confusion by allow- has decreased from 257 to 241. So acad- ously been excluded from other schools in ing academies (and, in fact, all specialist schools – 90% of emies have actually had the effect of reduc- the area. In any case the government have state comprehensives are now ing the number of children from poorer recently altered the funding agreements for specialist schools) to select 10% of their intake on aptitude families in other local schools – the exact new academies so that they have to make in their specialism. We are not 70 aware of any academies which opposite of cream-skimming. Moreover, a payment to the local authority if they use this power and it should the average pupil intake for academies exclude a child (assuming that the authority be scrapped for avoidance of doubt. have lower scores in their Key Stage 2 is prepared to make an equivalent payment
70 PwC, pp 41-43 assessments from primary schools than the other way if the academy accept a child 71 74 71 Ibid, p 51 those going to other local schools. They excluded from another local school). This
72 Ibid, p 53 also have many more children with special means that from now on it will be in the 73 Ibid, p 62 education needs than other local schools financial interest of academies to avoid 74 Ibid, p 89 (33% to 23%).72 exclusion unless absolutely necessary.
22 2
Sweden
The 1992 Swedish school reforms have be a contentious issue in the run up become something of a totem for the to a General Election and the potential 75 For example, Pollard S, A centre-right in the UK. Pretty much every implementation of Conservative Party and Class Act: World Lessons for UK Education, Adam Smith Institute, centre-right think-tank has published Liberal Democrat reforms, it is worth 2001; School Choice For All, something on Sweden over the past eight exploring the reforms in some detail. Reform, 2004; Hockley T and 75 Nieto D, Hands up for School years. This has built to a crescendo in the Choice, Policy Exchange, 2005 past 18 months with two more think-tank Background 76 Hlavec M, Open Access for reports,76 a whole series of newspaper arti- In 1992 the Conservative government in UK Schools: What Britain can Learn from Swedish School 77 cles and the decision by the Conservative Sweden introduced legislation to allow pro- Reform, Adam Smith Institute, Party and the Liberal Democrats to use ‘the viders who wished to set up independent 2007; Cowan N, Swedish Lessons: How Schools With Swedish model’ as a shorthand for describ- (or ‘free’) schools to receive state funding More Freedom Can Deliver ing their programmes of reform (even as long as they did not select academi- Better Education, Civitas, 2008 77 For example: Freedman S, though, as we shall see, there are significant cally. Initially free schools received 85% They’ve given us Ulrika, saunas differences).78 Even so, the introduction of of the per-pupil funding that went to state and Ikea. Now, socialist Sweden is about to give us David the reforms, their history and impacts are schools, although this was decreased to Cameron’s BIG idea for schools, not well understood in the UK. The com- 75% in 1995. The assumption was that Daily Mail, 12 July 2008, see www.dailymail.co.uk/news/arti- mentary from the centre-right has focused free schools would be run more efficiently cle-1034603/They-8217-ve-given- on the theoretical underpinnings of the that state schools so could survive on less Ulrika-saunas-Ikea-Now-socialist- Sweden-David-Camerons-BIG- reform and selective statistical evidence of money, reducing costs for the whole system idea-schools.html; Garner R, Free its impact on performance.79 and (unlike state schools) they were not School: Conservatives eye the Swedish model, Independent, Those on the Left, meanwhile, have expected to admit children with serious spe- 1 December 2008, see www. 83 independent.co.uk/news/educa- focused solely on evidence that segrega- cial needs. They were, however, allowed to tion/education-news/free-school- tion in Swedish schools has increased over request top-up fees from parents.84 conservatives-eye-the-swedish- model-1042734.html the past few years.80 In recent months Much of the impetus for reform came 78 Gove M, We need a Swedish the Government have weighed into the from parents in rural areas who wished to education system, Independent, debate, disingenuously pretending that the run state schools threatened with closure 3 December 2008, see www. independent.co.uk/opinion/ Swedish reforms are a dangerous new idea by their local authority, and from organisa- commentators/michael-gove- recently discovered by the Conservative tions with radical pedagogies like Waldorf we-need-a-swedish-education- system-1048755.html 81 Party. As the introduction to the Steiner and Montessori. It was expected 79 An partial exception should be Government White Paper that led to the that these groups would dominate the made for Cowan (2008) which is considerably more detailed than 2006 Education Act makes clear, Sweden new independent sector and at first they any other contribution to the dis- was a direct inspiration to Tony Blair and did. However, in the mid-90s there was cussion so far. Andrew Adonis and, as we saw in the pre- a significant shift towards ‘general profile’ 80 See, for example, Toynbee P, Beware the lesson of the vious chapter, their original conception of schools whose curriculum approach was Tory wolf in liberal clothing, The Trust Schools was quite close in spirit to broadly similar to public schools but inno- Guardian, 8 April 2008, see www. guardian.co.uk/commentis- 82 the 1992 Swedish reforms. vative enough to attract parents. Some of free/2008/apr/08/conservatives. Given that it seems the role played these schools were founded by teachers guardiancolumnists and any statement on the Swedish sys- by the market reforms in Sweden will and headteachers from the state sector tem from the NUT.
www.policyexchange.org.uk • 23 A guide to school choice reforms
who were unhappy with the restrictions the last five or so years than has been a placed on them while others were set up significant increase in the number of free by entrepreneurs attracted by the ability to gymnasium from 200 to 300 (50%) since 81 See, for example, Knight make profits.85 According to the Swedish 2003; the number of pupils has more than J, The Tories still stand for privilege, Times Education Schools Inspectorate 10% of free schools doubled in the same time. The Swedish Supplement, 24th October are ‘confessional’ and 6% are run by high school curriculum is designed around 2008, see www.tes.co.uk/article. aspx?storycode=6004229 Steiner Waldorf. The vast majority of the seventeen programmes, only two of which 86 82 Department for Education remainder are general profile schools. are academic. This means it is relatively and Skills, Higher Standards, Better Schools for All, HM In 1994 the Social Democrats (the cen- easy to set up a small high school to teach Government, 2005, p 4 tre-left party in Sweden) returned to power. just one or two of these programmes. The 83 Böhlmark A and Lindahl They were opposed to the fledgling free market at this level has grown so quickly M, Does School Privatization Improve Educational school movement but were constrained by that some are worried it is overheating and Achievement? Evidence from their popularity with parents and support that many of the schools will have to close Sweden’s Voucher Reform. IZA 90 Discussion Paper No. 3691, for the schools by their coalition partners, in the next few years. September 2008, p 4 the Green Party.87 Rather than eliminate free For-profit operators now dominate the 84 Apparently few did. Interview schools altogether they passed a law in 1996 free school market, especially at the com- with Maria Rankka (President, Timbro, 25 November 2008) which meant that the National Education pulsory and high school level. According to 85 Böhlmark A and Lindahl Agency, the body which approved free the Swedish schools agency website, 1,671 M, p 6 schools, had to take representations from licenses have been awarded to run schools 86 Telephone conversation with the Swedish Schools local authorities when deciding whether or at this level (more than are actually open Inspectorate (December 2008) not to allow a new school. This has had lit- at the moment as operators have two years 87 The Green Party’s support is tle effect in practice as free school operators to open after their application has been due to their fondness for special 91 pedagogies, such as Waldorf have tended to seek local political support approved). Of these exactly two-thirds Steiner, that emphasise play and before making a bid (for reasons explored (1,114) have been awarded to for-profit discovery over formal learning. 88 88 Interview with Anders Morin below – see p.29). They also banned top- operators. In the last few years for-profit (Senior policy manager, wel- up fees but raised per-pupil funding for free operators have submitted the vast majority fare policy at Confederation of Swedish Enterprise, 25 schools to 100% of that received by state of new applications. At the pre-school level November 2008) schools in the local authority. fewer schools (25% or 595 schools) are 89 Data kindly provided by The legislation has been left more or for-profit but this percentage has grown Skolverket. less unchanged since then, and there are considerably in the last few years (from 90 Interviews with Anders 92 Hvarfner (CEO, Vittra, 26 now 3,302 free schools, or 21.5% of all 18.4% in 2003). There is an ongoing November 2008) and Bertil Östberg (State Secretary to Swedish schools and pre-schools. 209,410 process of consolidation as they buy up Minister for Education Jan children (approximately 12% of all chil- smaller groups or one-off schools.93 Björklund, 26 November 2008) dren) attended free schools in 2007. The The core principles of reform have now 91 www.skolverket.se tables in Appendix A show the number been accepted by all political parties (bar 92 Data kindly provided by Skolverket and percentage of free schools and students the Communists) but debate continues
93 For full tables of school num- attending free schools at each of the three on whether school operators should be bers see Appendix A levels of Swedish education: Forskola (pre- allowed to make a profit or, more specifi- 94 Interviews with Anders Morin (Senior policy manager, school), Grundskola (compulsory schools cally, whether they should be allowed to pay welfare policy at Confederation for ages 7-15), and Gymnasium (high dividends to shareholders, especially foreign of Swedish Enterprise, 25 94 November 2008) and Carl- schools for ages 16-18). It is notable that shareholders. While profit margins are rel- Gustaf Stawström (Permanent free schools are typically much smaller than atively low for companies running schools Secretary, Swedish Association 95 on Independent Schools, 25 state schools (and these are, in turn, much at the moment (at around 2% on average ) November 2008) smaller than their English counterparts).89 they are growing as operators become larger 95 Renstig M, Women’s Business The average size of a compulsory-level free and develop further economies of scale. Research Institute, Utveckling och lönsamhet för privata school is 132 pupils (207 for state schools) Some of the bigger companies are now utförare Presentation given to and for gymnasium 188 pupils (638 for quite profitable. Alongside standard ideo- the Confederation of Swedish Enterprise, 2008 state schools). It is also notable that in logical (and illogical) concerns about state
24 Sweden
money ending up in the hands of investors forming. There are national published tests there are more specific worries that schools in all subjects at the end of the ninth grade are maximising profits by turning away (15 year olds), but these are not standard- students with special needs or learning dif- ised (i.e. marked externally). Standardised ficulties.96 There are also concerns, albeit tests in Maths, English and Swedish are with no supporting evidence, that some also taken at the end of the ninth grade, unscrupulous operators are entering the although it is only since 2004 that these market for just a short period before getting tests been taken by all students rather than out with quick profits. just a sample. In 2007 the average non- At the moment the Conservative Party standardised grade point average (GPA) (in coalition with the centre-right Liberal for pupils in free schools was considerably Party) are in power, but before they lost higher than in state schools (227.2 ver- the election the Social Democrats had set sus 207.5).99 One explanation offered for up an enquiry into for-profit free schools, this difference has been that free schools and they are likely to resurrect this when- have a more favourable pupil profile. ever they return to office, with potential Unfortunately, there have been no studies outcomes including limits on dividend pay- that attempt to compare scores while con- outs. That said, the free school sector is now trolling for factors such as parental income 96 Skolfront, Skolor sager nej, SVT2, 23 October 2008 so dominated by for-profit providers that it and education. Similarly in high school 97 Interview with Bertil Östberg is unlikely this could be achieved without tests, where free schools outperform state (State Secretary to Minister for fatally undermining the whole system. schools by 14.3 grade points out of 20 to Education Jan Björklund, 26 November 2008) The current government, while continu- 14.0, the same caveat applies.100 98 Interview with Per ing to support free schools, have moved A number of studies have attempted Ledin (President and CEO, Kunskapsskolan, 25 November their attention towards increasing central to measure the relationship between the 2008) oversight of all Swedish schools. They are number of free schools in a given local 99 Skolinspektionen / Swedish planning to introduce national testing in authority and performance, on the assump- School Inspection Agency, 2008, see www3.skolverket.se/ primary schools, expand their very mini- tion that increased competition should friskola03/friskola.aspx mal national curriculum and boost the drive up overall results. Two studies based 100 Ibid. capacity and scope of the national inspec- on data from 1997/8, when the reforms 101 Ahlin Å, Does School 97 Competition Matter? Effects tion agency. If this sounds familiar it is had only been active for a few years, sug- of Large-Scale School Choice because they are using England as a model gested a significant boost to performance Reform on Student Performance, Uppsala University, 2003; for some of their changes. As we shall see in local authorities with more free schools Björklund A et al, The Market these changes do threaten some of the free for all students. Åsa Ahlin found that a ten Comes to Education in Sweden: An Evaluation of Sweden’s schools’ very broad freedoms but most of percent increase in the number of children Surprising School Reforms, the free schools – especially the general attending free schools led to a six percen- 2004, see www.nek.uu.se/ StaffPages/Publ/P374.pdf, p 112 profile ones – are supportive of the changes tile increase in performance in standard- 102 Ahlin found an increase 101 as they will provide (they believe) an unbi- ised ninth-grade maths tests. This result of 0.17 standard deviations in ased illustration of their value to parents.98 was more or less replicated by Mikael Maths test scores associated with a 10% increase in free 102 Sandtrom and Fredrik Bergstrom. school enrolment, Sandström Performance and Bergström found a increase Ahlin found no impact on standardised of 0.19 standard deviations. One of the reasons that the current Swedish Swedish and English tests (Sandstrom and Sandström M and Bergström F, School Vouchers in Practice: government are so keen to introduce more Bergstrom did not test for this). Sandstrom Competition Won’t Hurt You!, national testing is that there is very little and Bergstrom also found a significant Working Paper No.528, The Research Institute of Industrial data available at the moment that can be positive effect for the non-standardised Economics, 2002. Dietrichson used to evaluate the relative performance ninth-grade GPA – something that Ahlin J, Introducing Competition in Public Services – The Case of individual schools. At the moment there did not test for. The GPA covers all the of the Swedish Compulsory are no primary level tests so it impossible subjects taken by ninth grade students but School, School of Economics and Management at Lund to say how free schools at this level are per- because it is not standardised there is a risk University, 2005, p 33
www.policyexchange.org.uk • 25 A guide to school choice reforms
that results will be affected by differences In a follow-up study Böhlmark and in grading between types of schools and Lindahl found that the impact of free municipalities.103 Using data from 1998- school enrolment at ninth-grade level wears 2001 Björklund et al found a significant off over time and that it has no significant effect on GPA (a four percentile increase effect on high school GPA (equivalent to for a ten percent increase in the number A-levels).107 Their results are questionable of children attending frees schools). They for a number of reasons. First, in the sec- argue that this is more significant than test ond year of high school (which lasts three 103 There is some dispute over data because it can be measured across the years) they find a 2.5% percent boost to whether free schools inflate non-standardised grades com- entire Swedish population whereas stand- GPA from a 10% increase in ninth-grade pared to state schools. Christina ardised test data was (until 2004) only col- free school enrolment, which is higher and Magnus Wikström claim 104 they do, by about 15% at high lected for a small sample of students. than the impact they find on ninth-grade school level when compared More recently these results have GPA for the same students. This effect with performance achieved on the externally marked Swedish been challenged by Anders Böhlmark then seems to disappear in the third year Scholastic Assessment Test. and Mikael Lindahl, who found a much of high school when no positive effect is Böhlmark and Lindahl, however, find that free schools deflate smaller (though still significant) impact found. The authors can find no convinc- grades at ninth-grade level 108 compared to state schools. As on GPA of one percentile increase for a ing explanation for this. Moreover, the the latter study is looking at 10% increase in free school enrolment.105 authors have again used the ninth-grade compulsory schools and the former at post-compulsory high The main reason for the difference seems share as a measure of free school enrol- schools they could both be right, to be the time period under investigation ment. If the share in all compulsory grades though that would be counter- intuitive. See Wikström C and (Böhlmark and Lindahl’s dataset goes had been used it is likely that they would Wikström M, Grade inflation and up to 2003) and, more importantly, the have found a significant impact on high- school competition: an empirical analysis based on the Swedish measure of free school share in a munici- school performance. upper secondary schools, pality used. Previous studies looked at Despite the generally positive picture Economics of Education Review, Vol. 24, No. 3, 2005, pp 309- the number of free school students in all painted by these analyses, many policy- 322; Böhlmark A and Lindahl M, The Impact of School Choice on compulsory grades (1-9) in a municipality makers in Sweden are concerned that free Pupil Achievement, Segregation and correlated this against performance schools have not had a greater impact on and Costs: Swedish Evidence, IZA Discussion Paper No. 2786, data at ninth-grade level. Böhlmark and performance. There is considerable unhap- May 2007, p 29 Lindahl only looked at the share of stu- piness that Sweden is currently ranked 104 Björklund et al, p 116 dents in free schools in the ninth grade. lower in international comparative stud- 105 Böhlmark and Lindahl, They argue that including earlier grades ies like TIMSS and PISA than in previ- 2007, pp 25-26 109 106 Ibid, p 32 (which doubles the impact to two percen- ous years (though still outperforming 110 107 Böhlmark and Lindahl 2008, tile points in their study) is inaccurate: England in PISA). There are a number pp 17-18 “we find it difficult to believe that there of possible explanations for this. First, 108 Ibid, pp 20-22 would be such a large causal effect from the number of children in Sweden has 109 Highlights From the Trends the amount of private schooling in early been increasing for almost the entire time in International Mathematics and Science Study (TIMSS) 2003, grades....on the grades received by those that the free school reforms have been National Center for Education leaving grade 9 in the same year.”106 One in place.111 This has meant that few state Statistics, 2004, see nces. ed.gov/pubs2005/2005005.pdf, could counter-argue that a higher level of schools have had to close as a result of com- p 7; and www.pisa.oecd.org/ dataoecd/31/0/39704446.xls free school share at earlier grades could petition. Effectively, free schools have been
110 And there are any number indicate that the students taking the absorbing new demand in many munici- of methodological flaws in ninth grade exams are more likely to have palities. Barring any major immigration, international studies. See, for example, Smithers A, England’s benefitted from being at a free school high school populations are expected to Education: what can be learned throughout their education. Either way fall for the next few years (see Graph 2) by comparing countries? Sutton Trust, 2004 all analyses to date have found evidence and many schools - both state and free -
111 Lim C and Davies C, of a statistically positive impact at ninth- will have to close. As competition increases Helping Schools Succeed: grade level. None has suggested a negative its impact on performance is likely to Lessons from Abroad, Policy Exchange, p 75 impact on attainment. become significantly more pronounced.
26 Sweden
Graph 2: Number of children of each age in Sweden112
140,000
120,000
100,000
Number 80,000
60,000
40,000
20,000
0 012345678910 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Age
Secondly, although the reforms have been market is still growing. In the application in place since 1992, only in the last few years period 2008-9, there were 441 applica- have the benefits of scalability really started tions to open new schools (59 pre-schools, to kick in.113 The free school groups have 137 compulsory schools, 237 high schools now reached a size at which they can col- and 8 special schools), despite the falling lect comparative performance data; translate number of children.114 Polls also show that best practice from one school to another; free schools are more popular. Appendix provide additional training and professional B contains data taken from polling by development for their own teachers; and the Confederation of Swedish Businesses develop their own internal leadership pro- in 2006 of parents, students, and teach- grammes. They have also now had the ers. They show that free school parents experience to fine-tune their curriculum and are happier with their choice than state teaching methodologies. Thus it is likely school parents in every single category. In that, at least for the bigger groups (which a number of categories like “the school is are nearly all for-profit), performance will good at taking care of skilled and talented start to improve over the next few years. students” and “as a parent I can influence Graph 3 shows how this is happening in how the rules in school are followed” free Kunskapsskolan schools. For the first years school parents are three times more satis- 112 Figures from Statistics of operation, performance has been steadily fied than their counterparts in the state Sweden, see www.scb.se/tem- above average, albeit not dramatically. Last sector. Looking at the overall measure of plates/tableOrChart____159285. asp year, results improved dramatically and are satisfaction, 9 out of 10 free school parents 113 Even groups already in forecast (on the basis of internal assessment) give their school a positive grade compared existence will have taken time to reach scale, but it must also be to do so again this year. with 6 out of 10 state school parents. understood that not all groups Of course another important way of The difference in student satisfaction is and companies in the market- place came into being at the measuring the success of free schools is less clear cut (perhaps for the same reasons birth of the reform. For example their popularity. The expansion of the free their parents are happy – discipline is tougher Vittra was founded in 1993, while Kunskapsskolan is a new- schools market from zero to around twelve and they are expected to work harder). comer, started in 1999. percent of the available pupils in the last Nevertheless free school students still rate 114 Data from the Swedish fifteen years is testament to a growing their school higher in 21 out of 25 categories School Ministry: www3.skolver- ket.se/friskola03/friskola.aspx (9 popularity amongst parents. Moreover, the including big leads for questions measuring February 2009)
www.policyexchange.org.uk • 27 A guide to school choice reforms
Graph 3: Kunskapsskolan average grades, compared with the other schools in their school district and the national average
Points Scored
320
National Average Kunskapsskolan Municipality Average Kunskapsskolan Average Projected Average
240
120 2004 2005 2006 2007 2008 2009 Year
a positive work environment. Interestingly, esting tasks, feel that they are expanding despite free school students being happier their professional competence and believe
112 Figures from Statistics with their school, under almost all of the that they are learning new things. There are Sweden, see www.scb.se/tem- measures both sets of students rate their no categories in which state school teachers plates/tableOrChart____159285. asp schools more or less the same overall (any are more content to a statistically signifi- 113 Even groups already in difference less than 5% is not statistically cant level. Interestingly the pollsters did existence will have taken time to reach scale, but it must also be significant) and are equally happy with their not ask a question about salary and there understand that not all groups choices. The authors of the report speculated is no available evidence on differences in and companies in the market- place came into being at the that this is because free school students are working conditions but any problems with birth of the reform. For example more likely to have been actively involved in pay or conditions would surely have been Vittra was founded in 1993, while Kunskapsskolan is a new- choosing their school so will be more aware picked up in the questions on camaraderie comer, started in 1999. that alternatives are available to them. and feeling appreciated.115 114 Data from the Swedish School Ministry: www3.skolver- Crucially free school teachers are more It is also worth noting a more subtle ket.se/friskola03/friskola.aspx (9 positive in 11 out of 13 categories. This is effect of the reforms. Swedish education February 2009) important as the main opponents of mar- has traditionally been highly progressive. 115 All polling taken from Laurent B, Friskolorna är bäst i ket reforms in every country in which they Very few schools expect their pupils to klassen, Swedish Confederation are mooted are the teacher unions, pre- wear uniforms, for example, and reading of Business, 2006, translated for the authors by Jonathan Giertta. sumably because of the belief that working is not taught until children are seven years
116 Which suggests the stand- conditions will be harmed and they will old. The curriculum is minimal, there are ard union argument in England lose some of the control they enjoy in a very few national tests and parents rely on that teacher status is low because of a lack of autonomy centralised system. Nonetheless, teachers teacher assessment. The status of teachers is probably not correct. They have much more autonomy in at free schools take more pride and enjoy- has always been very low (even lower than Sweden. ment from their work, receive more inter- in England).116
28 Sweden
As discussed earlier, the current prepared to follow the (very slim) national Conservative government are looking to curriculum; be able to show demand for change this by expanding the national cur- their schools in the municipality proposed; riculum, increasing tests and introducing a show that they have the financial capacity tougher inspection regime. It is question- to set up the schools; show that they will able whether they would be able to do this not be narrowly faith-based (schools can if free schools had not been introduced. be run by religious organisations but the While there are no studies of parental curriculum offered must be secular) and demand (an important gap in the research), so on. Since 1996 the municipalities (local anecdotal evidence suggests that parents authorities) affected have been able to have been attracted to more rigorous educa- make a submission regarding the impact of tional models. The Internationella Engelska the proposed school on local state schools, Skolan offer a traditional English education and the central agency have had to take (tough behaviour management included) this into account in their judgement. and are probably the most popular free In practice, even though going through school group in Sweden. Even groups like this process can be time-consuming and Kunskapsskolan, which look progressive bureaucratic, it is not overly restrictive. Very to the English eye as their curriculum is few, if any, applications have been blocked based on personalised learning, are far more because of a municipality submission (per- attainment-oriented than Swedish state haps because prospective operators usually schools and have been pushing for more spend some time warming up local politi- national testing. In other words it seems as cians before making their application).117 if Swedish parents have different views on Some are rejected due to problems with what is best for their children than genera- their application. About 36% were rejected tions of Swedish educationalists, and entre- last year, and anecdotal evidence suggests preneurs have been able to take advantage slightly more are now being rejected than of this. This, in turn, has had a significant in the past.118 However, there seems to be impact on the whole nature of the educa- a general consensus that the process is fair tion policy debate. and a good application will nearly always succeed. Certainly the established groups The Seven Tests have no problems getting permission to set up new schools.119 However, there are Demand-Led question marks over whether this process The Swedish system is the most demand- accurately filters out poor applications, led of the three examined in this report. given the ease with which a system based Unlike the academies programme entry on paper applications can be ‘gamed’ by 117 Interview with Anders to the market is decided by the provider operators who know what the inspectorate Morin (Senior policy manager, rather than initiated by the government. are looking for. welfare policy at Confederation of Swedish Enterprise, 25 Unlike America there are no set caps on Naturally, potential market share in November 2008) the number of entrants. There is not, how- any given municipality is the driving force 118 Information provided by the Skolverket; Interview with Carl- ever, an entirely free market. If a provider behind the majority of new applications, Gustaf Stawström (Permanent wishes to set up a school they have to apply which come from for-profit companies. Secretary, Swedish Association on Independent Schools, 25 to the school inspectorate (the Swedish This is usually determined by surveying November 2008) equivalent of Ofsted). Until October 2008 demand for free schools and competition 119 Interviews with Per they applied to the Skolverket or School from other groups, as well as the size and Ledin (President and CEO, Kunskapsskolan, 25 November Board (roughly equivalent to the QCA). demographic composition of the pupil 2008) and Claus Forum (Manager of Education, The inspectorate have a list of criteria that population. Alongside market demand the Baggium Vocational Schools, 26 applicants have to meet. They must: be other key issue for providers before sub- November 2008)
www.policyexchange.org.uk • 29 A guide to school choice reforms
This map of Baggium Vocational Schools (left) shows the typical concentration of free schools in the more poopulated south, while Kunskapsskolan (right) do not even show the north of the country!
Ӧrnskӧldsvik Falun Sundsvall Uppsala Enkӧping Vӓsterӓs
Gӓvle Norrkӧping Nykӧping Falun Uppsala
Karlstad Vӓsterӓs Stockholm Arjӓng Ӧrebro Gӧteborg Gamleby Norrkӧping Sӧdertӓlje Boras Uddevalla Nykvarn Trollhӓttan Linkӧping Orust Aneby Borӓs Jӧnkӧping Göteborg Vӓxjӧ Varberg Kalmar Halmstad Helsingborg Helsingborg Landskrona Lund Karlskrona Malmö Ystad Malmӧ Kristianstad Trelleborg
mitting an application is the politics of Easy to Enter municipality involved. An unfriendly local There are very few restrictions on the type administration cannot prevent a school of organisation that can enter the market being built but it can be extremely obstruc- as long as they are able to fulfil the crite- tive. As we will see (p.33) they may set ria required for a successful application. artificially low per-pupil funding, they can The inclusion of for-profit companies has also obstruct planning permission for new been crucial to ensuring the swift expan- or converted school buildings and appeal sion of the market and has also meant against approved applications as a delaying that the majority of new providers offer a tactic (though such appeals rarely suc- general education rather than special peda- ceed).120 The relative performance of state gogies or faith-based programmes. Mikael schools is not an important factor for the Sandström, a senior advisor to the Swedish larger school providers, but it may be for Prime Minister and author of one of the one-off free schools set up by teachers or free school performance studies discussed 120 Böhlmark and Lindahl 121 (2008), p 7-8; Interviews with parent co-operatives. earlier, has argued on a number of occasions Per Ledin (President and CEO, Kunskapsskolan, 25 November These factors have meant that the mar- that their system would not function prop- 2008) and Anders Hvarfner ket is heavily focused on more populated erly without profit, as good schools would (CEO, Vittra, 26 November 2008) and more conservative towns and cities in not have the incentive to expand and access 121 Ibid. Southern Sweden, especially Stockholm would remain limited: “The profit motive 122 Garber R, You’re not going far enough, Swedish expert tell where between 20% and 30% of children is good for making schools less selective… Cameron, The Independent, 1 attend free schools depending on the age If a school cannot make a profit, it is not December 2008, see www.inde- pendent.co.uk/news/education/ group. In northern areas, where voters tend a smart system…It allows them to expand education-news/youre-not-going- to be more left-wing and pupils are more and they can take more pupils in, definitely far-enough-swedish-expert-tells- 122 cameron-1042733.html dispersed, the market is less developed. from more disadvantaged parents.”
30 Sweden
The costs of entering the system are also the cycle was only moved from six to low. Although providers do not get capital three year inspections in 2008.125 The cur- upfront with which to build the school rent government is proposing to push the there are far fewer restrictions on the type of system further in this direction by tough- building that can be used than in England. ening up inspections and increasing the Typically operators lease buildings rather scope of the national curriculum while than buy them outright and cover their bringing more regular national testing so rent from their annual per pupil allocation as to provide more data for inspectors to (which, in most municipalities, includes work with.126 money for capital costs). Some money Until these changes are implemented will then be spent re-fitting the hired there is probably not enough information building which could have been anything made available about schools, free or state, from an office block to a warehouse to an for the market to operate optimally. At the astronomy lab in a previous life. Even the moment, the primary mechanism for ena- most ambitious groups, though, will only bling parents and students to make deci- spend a fraction of the funds spent on new sions about schools are ‘recruitment fairs’. school buildings in England. To take one There is an obvious problem that schools example, Kunskapsskolan, an established which spend a lot on marketing or, at high chain for whom design is a crucial aspect school level, offer exciting looking courses of the curriculum offering, spend about of little actual worth, may attract busi- £1 million on a school for 350 students ness without offering a valuable service. compared to between £30 and £40 million While this kind of market failure might be for a new academy for 1,500 students.123 acceptable when making small purchases The disproportionate costs of setting up a at a supermarket it is not when a child’s new school in England is one reason why life chances are involved. One municipal- the government has to control the acad- ity, Nacka, has tried to solve this problem emies programme so tightly and ration by forcing parents to make a compulsory the money to replace only failing schools choice aided by a brochure with informa- rather than allowing new school entrants. tion about all schools.127 While this process theoretically puts Accountable all parents in the same position, it can- Once open, free schools are accountable to not deal with the basic problem: a lack of the School Inspectorate and are inspected data on which to make a decision. While every three years.124 Municipalities often a list of compulsory age secondary schools run their own annual inspections but with ninth-grade results are available on do not have the formal power to impose the school board website, we could find sanctions. Very few schools have been no information at individual school level closed down and there are relatively wide- about primary or high schools. This is a 123 Private information spread concerns that the national inspec- real problem and may explain why free 124 Interview with Claus tion regime is too light with insufficient schools have not boosted performance Forum (Manager of Education, Baggium Vocational Schools, 26 emphasis on attainment; so much so that by as much has been originally hoped – November 2008) the more successful chains have set up especially at high school level. It is worth 125 Lim C and Davies C, their own reporting mechanisms to qual- emphasising, though, that this is a prob- Helping Schools Succeed: Lessons from Abroad, Policy ity control their schools. Even this level of lem for all schools - not just free schools. Exchange, pp 84-85 accountability is relatively new to Sweden In fact, free schools are presently more 126 Interview with Bertil Östberg (there is no Swedish word for account- accountable than state schools because the (State Secretary to Minister for Education Jan Björklund, 26 ability). Until 2003 there was no regular School Inspectorate can withdraw funding November 2008) programme of national inspections and from the former if they are not performing 127 Cowan, pp 35-36
www.policyexchange.org.uk • 31 A guide to school choice reforms
well but has no formal powers over the lat- one of the few countries in the developed ter (they can only request that the school world not to have centralised pay bargain- develop an improvement plan).128 ing for teachers. The relevant changes were pushed through, with support from Genuinely Free the teaching unions, in the mid-90s when All schools in Sweden – free and state – Sweden was recovering from a severe have considerable pedagogical freedom. recession triggered by a collapse in the The national curriculum, followed in banking system and local government every school, is a very slim document would only discuss increasing pay in the of 17 pages that sets outcomes but does context of moving to a individualised pay not specify content or pedagogy. This scheme.133 Severe shortages in the number freedom is a key driver of the success of good qualified teachers (due to the low of the Swedish system as school groups status of teaching as a profession) meant with general profiles still develop innova- that the primary effect of these reforms tive ways of delivering the curriculum. was to increase starting wages for new Kunskapsskolan, who run 32 schools, use teachers.134 However, some free school a personalised learning model in which operators have used this freedom to all students work at their own pace with develop performance-based pay models. one daily fixed lesson, occasional lectures Kunskapsskolan, for example, allocates and 15 minutes each week with a personal salary budgets to its schools depending tutor.129 The school group Vittra chose to on test results, financial management and timetable subjects into three-hour blocks the results of an annual survey of pupils to allow for multi-stage lessons and deep and parents. Teachers can also negotiate absorption in topics through case- and individual pay rises with their headteach- problem- based learning.130 Baggium, a ers.135 Teachers remain unionised in all group of vocational schools at high school schools so it is not easy to remove bad level, have set up workshop-based schools teachers. Even so, because free schools are in industrial areas and then developed new entrants rather than replacements 128 Lim and Davies, p 85 their own apprenticeship programmes for failing institutions they do not have 129 Interview with Per 131 Ledin (President and CEO, with local businesses. the problem of taking over schools full of Kunskapsskolan, 25 November 2008) As illustrated in the previous section failing staff. It is nevertheless problematic
130 Interview with Anders this level of freedom can be problematic for operators who have to deal with poor Hvarfner (CEO, Vittra, 26 in the absence of suitable accountability teachers and puts greater emphasis on November 2008) structures. The fact that schools do not successful recruitment policies. 131 Interview with Claus Forum (Manager of Education, have to participate in national testing Finally, as we have seen, free school Baggium Vocational Schools, 26 before 15 is seen by many as a beneficial operators have almost complete control November 2008) freedom, but it means there is not enough over the design of their buildings. Because 132 Interview with Göran Fant (Chairman, Swedish Federation available information with which to assess there are no major upfront capital costs of Waldorf Schools, 25 November 2008) the various curriculum offerings. While there is no need for the complex, bureau-
133 Lim and Davies, p 86; Strath some of the groups using more progressive cratic and centralised procurement struc- A, Teacher Policy Reforms in methods like Waldorf Steiner are worried tures we have in England, though the Sweden: The case of individual- ised pay, UNESCO, 2004 that any new tests will not be appropriate downside is that there tend to be fewer 132 134 Ibid. for their students, most of the more suc- specialist facilities for practical science, 135 Freedman S, Lipson B cessful general profile schools are happy art and music. and Hargreaves D, More Good Teachers, Policy Exchange, that they will have unbiased information 2008, p 64; Interview with Per to compare themselves to competitors. Financially Consistent and Stable Ledin (President and CEO, Kunskapsskolan, 25 November All schools in Sweden have more or Per-pupil costs for free schools are decided 2008) less complete freedom over pay, making it at the municipal (local) level, which means
32 Sweden
that the system is not financially consist- the voucher is pupil-led and pupils can ent. Even though the variation is not near- attend schools in different municipalities ly as dramatic as in the US, it has skewed to the one in which they live, there are the market away from certain areas. The discrepancies in the funding of students existing law states that free schools must attending the same school. In one munici- be given the same as state schools in pality, Lund, where the education voucher the same municipality. Establishing how is set at a high level, this has become a sub- much state schools get per pupil, though, ject of political debate as residents feel they is neither an easy nor entirely objective are subsiding neighbouring areas.139 exercise. Capital costs pose a significant This is a bigger problem at high school problem in that state schools may not than compulsory level because different require annual payments from the munici- rates are set in each municipality for each pality if they already own their building, of the seventeen programmes available to yet some municipalities include these costs students. Two of these programmes are in the free school ‘voucher’ while others academic (natural or social sciences) and do not. The current government have set the other fifteen are vocational. While up an enquiry to look at this problem and the costs for the academic programmes is likely to recommend a change to the do not vary that much between munici- law requiring all municipalities to include palities, funding for vocational courses these costs, but until this happens the is often used as a lever for local politi- 136 Interview with Carl- market will remain skewed towards those cians to solve labour market shortages. Gustaf Stawström (Permanent Secretary, Swedish Association 136 that already do. The difference between For example, current funding levels in and on Independent Schools, 25 including capital costs or not is about 10% around Stockholm are 30-40% higher for November 2008) 137 Interviews with Per to 20% of the value of a voucher, which plumbing courses than students training Ledin (President and CEO, translates for the majority of providers, to become electricians while trainee car Kunskapsskolan, 25 November 2008) and Anders Hvarfner who are limited liability companies, into mechanics in Göteborg are funded between (CEO, Vittra, 26 November 2008) the difference between making a profit SEK 70,000 and 120,000 depending on 138 Spending on public serv- and going bankrupt. Unsurprisingly it is their home municipality. This creates some ices, in general, is far more devolved in Sweden than in the major factor for groups when they are bewildering budgetary complications for England and the majority of deciding where to open new schools.137 school leaders, especially for schools offer- taxes are collected at the local level. To avoid inequalities Although capital costs make up the ing a range of programmes (say plumbing between the value of services provided between rich and poor bulk of the difference between authorities, and car mechanics) to children from a municipalities these local taxes which could be solved through simple range of municipalities. Operators running are nationally equalised, so, for example, a significant proportion changes in legislation, there is a wider issue a number of free schools typically have to of Stockholm’s taxes are distrib- that is more intractable. The principle that cross subsidise, leading to equity issues.140 uted elsewhere. Nevertheless the principle that these are local municipalities should decide how much Interestingly, this cross-subsidisation is taxes and that, therefore, spend- to spend on local schools means that the prevented in the English system: academy ing should be directed at a local level is an important one. market is not funded consistently across groups are not allowed to move money 139 Interviews with Per 138 Sweden. Some municipalities consider between schools, despite variations of up Ledin (President and CEO, Kunskapsskolan, 25 November education a greater priority than others (as to £1400 per child per year between the 2008) and Bertil Östberg opposed to, for example, local transport different local authorities.141 (State Secretary to Minister for Education Jan Björklund, 26 infrastructure or healthcare) and set their These issues would be relatively easy to November 2008) voucher accordingly higher. This variation solve if funding was set through a national 140 Interview with Claus causes two problems. Firstly, free school formula. A potentially more serious prob- Forum (Manager of Education, Baggium Vocational Schools, 26 operators are more attracted to areas that lem, highlighted by critics of reform before November 2008) place higher importance on education, they were enacted, is that the whole educa- 141 Interview with Dan Moynihan (CEO, Harris which typically contain better educated, tion system could become more expensive Federation of South London middle-class families. Secondly, because because of the inefficiencies created by allow- Schools, 11 December 2008)
www.policyexchange.org.uk • 33 A guide to school choice reforms
ing new entrants to provide surplus places though in the case of Björklund et al it for students that could have been placed in was not statistically significant. Böhlmark local state schools. This is one of the main and Lindahl’s results (using slightly more arguments used by the UK government recent data – which goes up to 2003) indi- against those who have suggested introduc- cate a two percent increase in costs for a ing a Swedish-style system to England. In a ten percent increase in private school share, lecture to the Cass Business School in 2008 which, as they note “compared with most Education Secretary Ed Balls, referring to estimates of resource effects in the litera- the Swedish model, insisted that “this alter- ture...seems like a very high return” given native approach has huge costs – because of the positive impact on attainment.146 the huge expense of creating all the surplus places that this model relies upon – as well Politically Stable as the new buildings”.142 At a national level the free school system has The evidence from Sweden indicates survived for sixteen years, and for all but four that there is no basis whatsoever for this of these years the left-wing Social Democrats assertion (Mr. Balls provided no evidence have been in power. There is widespread feel- in his speech). Of the five studies we have ing that free schools have become so embed- been able to find investigating the impact ded in the system, and they are so popular of free schools on costs only one shows with the parents that use them, that it would a significant effect. This study by Anna be impossible for a future government to Fransson and Irene Wennemo, funded scrap the reforms. As we have seen the Social by the Swedish Trade Unions Congress, Democrats might seek to curb the powers found that in 2001 a one percent increase of free schools to make profit but given that in the number of free schools increased the majority of free schools are now operated costs by 250 krona (about £20) for each by commercial organisations it is hard to see student in the municipality.143 However, how this could be done without undermin- these findings have been pretty thoroughly ing the whole system. debunked by other researchers who have We have also seen that municipalities pointed out that Fransson and Wennemo’s hostile to free schools have some tools at results do not show a causal link as they their disposal to prevent entry in the mar- simply measure the relationship between ket, including the right to appeal and the the number of free schools and cost. As ability to set the value of the voucher. Free we have seen free school operators aim for school operators have typically worked 142 Ed Balls, Guaranteeing a more expensive areas so costs in munici- hard to develop good relations with local Good School For All: how incen- tives and intervention must work palities with more free schools would be politicians before opening, but because the together, Lecture to the Cass higher whether they were there or not. municipalities have no formal control over Business Schools, Thursday 13 November 2008, p 15 It is more useful to ask if an increase in free schools a change in the local adminis-
143 The authors did not find free schools leads directly to an increase tration does not lead to forced closures (as any effect from 1995-2000. in costs. Björklund et al, Böhlmark and is the case in the United States). Anna Fransson and Irene Wennemo, Valfrihetens pris – En Lindahl, and Dietrichson have all tried analys av grundskolan 1995- 2001, Landorganisationen i to measure this effect and have found it Fair Sverige, 2003; summarised in to be either nonexistent (Dietrichson) or Free schools are not allowed to select aca- Dietrichson, p 40 minimal.144 In their own study the Swedish demically or charge top-up fees (though they 144 Dietrichson, p 40; Björklund et al, pp 109-111; Böhlmark and School Board confirmed that there was no were from 1992-1996). They are obliged to Lindahl, 2007, pp 28-41 clear, causal effect.145 Björklund et al and take pupils from anywhere in the country on 145 Summarised in Dietrichson, Böhlmark and Lindahl did find a positive a ‘first come first served’ basis. In practice this pp 41-42
146 Böhlmark and Lindahl relationship between an increase in free means that most schools have ‘open lists’. (2007), p 40 schools over time and municipality costs, Technically it would be possible for parents
34 Sweden
to put their children’s name on the list from grants 3.3% more likely than Swedish born birth, but this does not seem to happen (as it parents.150 This effect is then exacerbated by does at fee-paying schools in the UK). Most the fact that most free schools are set up in parents add their names to several lists a year urban areas that have higher proportions of or two before their child is due to start or educated and immigrant families. Björklund move school, meaning that popular schools et al also found that if the sample is lim- can still fill their spaces relatively early. This ited to general profile schools then the effect causes some problems for operators as they amongst immigrant parents disappears and cannot be sure what percentage of their list amongst educated parents is reduced to just will accept an offer – in areas with a mature 1.4%, suggesting that “much of the relation- school market the figure can be as a low as ship between private school enrollment and 20%.147 Offers are only made in the June student background comes from schools before the school year starts which leaves with a particular profile, be it a subject, peda- little time for schools caught with few accept- gogical, confessional or ethnic profile.”151 ances. If parents have not added their name The rights and wrongs of this kind of to any list in time to win a place then the ‘aspirational segregation’ are more ambigu- municipality is responsible for finding them ous than for sorting based on income a place at a local state school. (which occurs in school systems with sig- All of the evidence suggests that, unsur- nificant numbers of fee-paying schools). prisingly, this system favours those parents After all, aspiration is freely available to who are prepared to spend more time mak- anyone and one could argue that allowing 147 Interviews with Per ing an active choice.148 The studies that have highly aspirational families from poorer Ledin (President and CEO, Kunskapsskolan, 25th November looked at the question of whether an increase communities to educate their children at 2008) and Anders Hvarfner in the number of free schools increases segre- schools based in wealthier areas is preferable (CEO, Vittra, 26 November 2008) gation in the school system agree that it does, to an admissions system based on catch- 148 None of our interviewees disagreed that this was the case. but in quite a specific way. There is no evi- ment areas that indirectly segregates on 149 Sandström and Bergström, 152 dence that parents who earn more are more income (via house prices). Nevertheless, p 28; Böhlmark and Lindahl, likely to choose free schools (usually termed on balance, we would argue that future pp 40-41; Björklund et al, pp 107-109 a ‘sorting effect’ in the literature). However, market reforms should be designed to avoid 150 Björklund et al do not sepa- parents with higher levels of education or segregation of any kind, given that sorting rate first and second genera- tion immigrants. Böhlmark and who are second generation immigrants are by parental education could pass low aspi- Lindahl do and find the sorting more likely to go for free schools.149 This ration on from one generation to the next. effect is limited to the second generation. Böhlmark and effect is relatively small – Björklund et al sug- In Sweden’s case, avoiding this would be Lindahl, p 41 gest that, in any given municipality, parents relatively easy to do by changing the admis- 151 Björklund et al, p 108 with university degrees are 4.5% more likely sions system and allocating extra funding 152 Many of our interviewees made this case including Maria to choose a free school that parents with just for pupils from deprived areas to encourage Rankka, Bertil Östberg, and a compulsory school education and immi- free school operators to set up there. Anders Hvarfner.
www.policyexchange.org.uk • 35 A guide to school choice reforms 3
USA
In recent years the US has been some- ate the US with an extreme laissez-faire what eclipsed as the primary internation- attitude towards public services that has led al comparator for UK school reform by to considerable social division as opposed Sweden.153 There are several reasons for to Sweden’s socialistic welfare state and a this. For a start, the Swedish reforms happy, homogenous society. It is, therefore, are much easier to follow. There is one in the interests of those trying to push school national programme which has changed reform in England to refer to the Swedish little since 1996 and, because the Swedish example when counteracting negative con- system currently produces so little data, notations of the word ‘markets’. In fact, as there are only a small number of evidence- discussed previously, the Swedish govern- based analyses of the performance effects ment were able to introduce a wide-ranging of the reforms (though, as we saw in the and extremely liberal reform without much previous chapter, these are not necessar- fuss whereas reformers in the US have ily that easy to interpret). In the US the had to fight a state-by-state battle against picture is much more complicated. First, entrenched supporters of the status quo in each state is responsible for the structure state education. Furthermore, the Swedish of their school system so there are fifty- free school system is dominated by for-profit one (the states plus Washington D.C.) providers whereas in the States they control systems. Secondly, there are two different only a small proportion of the market.155 153 This was not always the case. Most international school types of reforms that are often discussed This shift in focus away from the US is choice reports before 2005 gave space to both countries. See, for interchangeably and sometimes operate unfortunate as the variation in state reforms example, Pollard (2001); Nieto simultaneously in the same state: charter offers a fertile test-bed for researchers who and Hockley (2005) schools and vouchers. The former are wish to search for the common grounds of 154 For a digest of all currently operating voucher schemes see: independent state-funded schools set up success. In addition, there is far more data Lips D, School Choice: Policy by non-governmental bodies. Voucher available due to a national accountabil- Developments and National Participation Estimates in 2007- programmes, on the other hand, typically ity framework even more comprehensive 2008, Heritage Foundation, allow children (usually a select group of than our own (but far better designed). January 2008, pp 2-6. For a bal- anced overview of performance deprived children but occasionally all of Moreover, US reformers have succeeded studies of voucher schemes see Gill B et al., Rhetoric vs. them) to spend a designated amount on in areas where the Swedish model is poor, Reality: what we know and what any non-state schooling, which usually specifically accountability and fairness, and we need to know about charter schools, RAND, 2007, pp 79-94 involves co-opting much of the capacity of less well in areas where the Swedes seem
155 For more on the gap the existing fee-paying sector. In this chap- to have got it right, such as specifically between perceptions of the US ter we will focus entirely on charter schools supply-side entry and political consistency, and Sweden and the reality of school reform see Baggesen for the sake of simplicity and because they making the models complementary. Klitgaard M, School Vouchers are more comparable with academies and and Political Institutions: A 154 Comparative Analysis of the free schools. Background United States and Sweden, University of Southern Denmark, The other reason for the concentration Ironically, although school choice had 2007 on Sweden is politics. Most people associ- been a rallying cry of the right since
36 USA
Milton Friedman introduced the concept has thus been a relatively broad coali- of vouchers in 1955, the ‘charter school’ tion with many Democratic politicians, was introduced to America in 1988 by especially Bill Clinton,162 supporting Albert Shanker, who was the head of their development as well as Black and the AFT union (American Federation Hispanic groups that have seen them as of Teachers).156 Though hardly a typi- a potentially key driver of social mobil- cal union leader (he also spearheaded ity.163 It is notable that Barack Obama moves to increase public accountability of has chosen an Education Secretary, Arne schools – something vociferously opposed Duncan, who is known to be a friend of by other unions) Shanker was still a man charter schools.164 They are, however, of the left and saw charters as way of still opposed by many on the left and counteracting calls from the right for both main teacher unions argue for a private school vouchers with a choice pro- return to Shanker’s original concept.165 157 gramme driven by teachers and parents. Over the past eighteen years the char- 156 Though it wasn’t his idea. His vision was of groups of teachers being ter school movement has expanded rap- See Kolderie T, Ray Budde and the Origins of the Charter given a ‘charter’ to run a state school for idly. As of 2008 authorisers have empow- Concept, www.educationevolv- up to five years (at which point it could ered 4,568 schools in forty states plus ing.org/pdf/Ray_Budde.pdf 157 Kahlenberg RD, “The be renewed) that would enable them to Washington D.C., educating 1,341,687 Charter School Idea Turns 20: try out new and innovative ideas. This children (see Table 3 for a breakdown A History of Evolution and Role Reversals”, Education Week, original conception of chartering can still by state and Table 2 for annual increases March 26 2008 166 be seen in the legislative framework for since 1992). The exact status of these 158 Ibid. charters in Minnesota, which was the first schools depends on which state they are 159 For a state-by-state state to pass a charter law in 1991 with in. As Jack Buckley and Mark Schneider guide to charter laws see www.edreform.com/index. 158 Shanker’s guidance. Here the majority note: “almost every aspect of the legisla- cfm?fuseAction=cLaw of a charter school’s directors must be tion governing charter schools, such as 160 See, for example, the hugely influential book by John licensed teachers (though this provision the length of charters, who can apply Chubb and Terry Moe: Politics, can now be waived).159 for charters, and the like, can and do Markets and America’s Schools, Brookings Institute, 1990 Initially the school choice movement vary widely from state to state, so there 161 Also Kahlenberg RD, Tough on the right of American politics ignored are exceptions to almost every general Liberal: Albert Shanker and the charters, preferring to concentrate on statement describing charter schools.”167 battles over schools, unions, race and democracy, Columbia more direct voucher schemes that would Nevertheless there are some universals. University Press, 2007 allow parents to send their children All charter schools are publically fund- 162 www.ed.gov/ 160 PressReleases/04-2000/wh- to existing private schools. However, ed. Charters are a contract between an 0428a.html they soon realised that although Shanker authoriser (nearly always a public body 163 www.bampac.org/issue_ may have intended a charter to be held like a school district, state education advocacy.asp by teachers, laws could be extended to board or university) and the school oper- 164 http://features.csmonitor. com/littlebillclinton/2008/12/16/ allow anyone to hold a charter - from ator. In return for signing this contract what-arne-duncan-means-to- disaffected parents to community groups the operator is given greater autonomy charter-schools to corporations. The right embraced over the administration of the school 165 www.aft.org/topics/char- ters/; www.nea.org/charter/ charters as a mechanism for increasing and its curriculum. If it does not use index.html parental choice, forcing monopolistic these freedoms to improve performance 166 www.edreform.com/_ upload/CER_charter_numbers. state systems to compete for students. then the charter may not be continued pdf
Shanker was left bemoaning the hijack- when it comes up for renewal, usu- 167 Buckley J and Schneider ing of his scheme for teacher co-oper- ally every five years. Crucially parents M, Charter Schools: Hope or Hype?, Princeton University atives by precisely the same politicians actively choose to send their children to Press, 2007, p 3 and corporations that had previously charters, whereas in most parts of the US 168 Zimmer R and Buddin been endorsing vouchers.161 Because pupils are assigned based on residential R, Making Sense of Charter Schools. RAND Corporation 168 of its origins, the charter movement location alone. Occasional Paper, 2006, p 1
www.policyexchange.org.uk • 37 A guide to school choice reforms
Table 2: Annual Growth of the Charter School Movement169
Year Operational schools by Year Opened Number of Operating Charter Schools
1992 1 1 1993 29 30 1994 49 79 1995 121 200 1996 146 346 1997 208 554 1998 361 915 1999 397 1312 2000 356 1668 2001 362 2030 2002 334 2364 2003 298 2662 2004 451 3113 2005 423 3536 2006 376 3912 2007 361 4273 2008 284 4557
Graph 4: Annual Growth of the Charter School Movement
5000 Operational schools by Year Opened
4500 Number of Operating Charter Schools
4000
3500
3000
2500
2000
1500
1000
500
0
1992 1994 1996 1998 2000 2002 2004 2006 2008 169 Data from www.edreform. com/charter_directory/data1.cf m?CFID=3853032&CFTOKEN= 44663510
38 USA
Table 3: State by state Charter School Data170
Operating in Opening in Total Total Closed Total State 2007-8 2008-9 Operating Since 2002 Enrollment Alaska 25 1 26 5 5198 Arizona 477 29 506 97 119516 Arkansas 18 7 25 6 6750 California 698 65 763 100 252569 Colorado 140 7 147 10 53249 Connecticut 19 2 21 5 3932 Delaware 19 2 21 2 8740 D.C. 75 14 89 17 25385 Florida 347 37 384 7 108382 Georgia 65 17 82 7 40807 Hawaii 29 3 32 0 7317 Idaho 30 2 32 1 10492 Illinois 63 11 74 9 27683 Indiana 41 9 50 2 12631 Iowa 10 0 10 0 1462 Kansas 33 6 39 10 3361 Louisiana 54 12 66 10 23634 Maryland 30 4 34 2 7301 Massachusetts 62 2 64 6 23905 Michigan 243 6 249 27 93892 Minnesota 145 13 158 28 28371 Mississippi 1 0 1 0 367 Missouri 37 2 39 5 13125 Nevada 21 3 26 7 7295 New Hampshire 11 0 11 2 1212 New Jersey 58 6 64 20 17986 New Mexico 66 4 70 3 11426 New York 99 21 118 10 32602 North Carolina 102 2 104 33 30445 Ohio 285 14 299 38 94171 Oklahoma 15 0 15 1 4770 Oregon 81 12 93 8 13612 Pennsylvania 130 3 133 12 61823 Rhode Island 11 0 11 0 2894 South Carolina 29 7 36 10 8705 Tennessee 12 2 14 1 2585 Texas 320 13 333 35 108541 Utah 60 8 68 1 23233 Virginia 3 1 4 3 275 Wisconsin 246 8 254 22 41799 Wyoming 3 0 3 0 244 170 Data from: www.edreform. TOTAL 4213 355 4568 562 1341687 com/_upload/CER_charter_num- bers.pdf
www.policyexchange.org.uk • 39 A guide to school choice reforms
Beyond these shared features, the (CERU) at the University of Arizona have shape of the market in individual states for a number of years produced a survey of is dependent on a huge range of variables. for-profit EMOs. Last year they found 453 For a start, rules on funding and freedoms charters were run by 50 EMOs, or 10% of differ widely (explored below on pages 50 the market, although they acknowledge that, and 51) and, perhaps more fundamentally especially for smaller and recent start-up there are differences in rules governing EMOs it is difficult to ensure comprehensive who is allowed to run charters and who coverage.171 This market has quadrupled may authorise them. since 1998-1999 but has plateaued since The majority of charters are founded 2003-2004 as a number of bigger companies and managed by smalls groups of parents, have closed or consolidated (see Graph 5). teachers, education administrators or com- There is no equivalent annual survey for the munity organisers. Unfortunately there seem more recent phenomenon of CMOs but to have been no studies breaking down the researchers working for the National Charter proportion of each of these groups within School Project (NCSP) at the University the overall number of charters. An increasing of Washington Bothell have identified 251 number, though, are run by corporations charters run by 46 CMOs, or 5.5% of the known as EMOs (Education Management market.172 This is almost certainly an under- Organisations) if they make a profit and estimate as it is difficult to track down small- CMOs (Charter Management Organisations) er CMOs or find all of the schools affiliated if they do not. It is extremely difficult to work with franchise-style operators. For example, out exactly how much of the market is held KIPP (Knowledge is Power Programme) by these groups, but there have been some are the biggest CMO with 66 open schools, concerted attempts to estimate numbers. The according to their website, but only 20 are Commercialism in Education Research Unit recorded in the NCSP database.173
Graph 5: Total number of EMOs and Number of States with EMOs over time174
60 Total number of EMOs
Number of States with EMOs 50
40
30
171 Miron G et al., Profiles of For-Profit Education Management Organizations: 20 2007-2008, University of Arizona, 2008. A further 80 state schools are managed by these EMOs under direct contract. 10
172 We are very grateful to Robin Lake at the National Charter School Project for shar- 0 ing this data with us.
173 www.kipp.org/09/schools/ list.cfm 1998-1999 1999-2000 2000-2001 2001-2002 2002-2003 2003-2004 2004-2005 2005-2006 2006-2007 2007-2008 174 Data from: Miron G et al., 2008. Op. Cit. p 5
40 USA
Thus at least 15.5% of the market is focus their efforts on states with friendlier controlled by EMOs and CMOs and this legislation like Arizona, Florida, Michigan, may be a significant underestimate, espe- Indiana, California, New York and Ohio cially if one is a little looser about one’s (see Appendic C for percentages of charters definition of management. Adam Lowe run by EMOs). and Margaret Lin estimated, using data One of the most important aspects of 175 Lower A and Lin M, collected by the National Association of legislation in this context specifies who Steadying the Three-Legged Stool: Authorizers, Charter Charter School Authorizers in 2004-2005, is allowed to authorise new charters and Schools, and Education Service that 30% of charters receive substantial whether they are limited in the number they Providers, NACSA, 2006, p 1 management services from contractors, a can award. The market is much stronger, 176 Harvey J and Britt D ed., Quantity Counts: The Growth of percentage that will almost certainly have for both managed and one-off charters, in the Charter School Management risen in the past few years.175 The market in states where more than one different kind of Organizations, National Charter School Research Project, 2007, managed charters is growing very quickly, organisation is allowed to authorise. Where pp 11-17; Wilson SF, Learning especially CMOs, as educational philan- only one type of body is allowed to author- on the Job: When Business Takes on Public Schools, thropists like Reed Hastings and Bill Gates ise, the job usually falls to local school dis- Harvard University Press, 2006, are increasingly focusing their efforts on tricts who have an active incentive to avoid especially pp 39-78 177 Arizona, California, scaling up successful models across federa- competition as they are in charge of existing Colorado, Delaware, Virginia and tions. If the first ten years of the charter state schools. This is clearly analogous to Wisconsin school movement to 2001 saw organic the problems of local authority involve- 178 Hawaii, Iowa, Mississippi and Tennessee. Although growth as new legislation led to a ‘land ment with academies in England, although Hawaii’s case is more complex rush’ of new schools opening by a wide districts are typically much smaller and have than it appears at first: they have entered into an extensive 181 variety of individuals and organisations, even less strategic capacity. partnership with Edison, who now run 35 elementary, middle more recent history has seen the develop- There are around 800 different authoris- and high schools within the pub- ment of a ‘managed growth’ strategy based ers across the US and over 90% are school lic school system. See: www. 176 edisonlearning.com/our_offer- on federations. districts, many of which are very small and ings/achievement_services/ However, the development of EMOs, 50% of which only authorise one school.182 partnership_profiles.be and to a lesser extent CMOs, has been The other 10% is made up of state educa- 179 www.edreform.com/index.c fm?fuseAction=cLaw&stateID=2 frustrated by disparities in legislation in tion agencies, state authorizing boards, 0&altCol=2 different states over who can manage char- universities, not-for-profit foundations and 180 www.edreform.com/index. ters. Just six states allow EMOs to hold one Mayor’s office (in Indianapolis). These cfm?fuseAction=cLaw&stateID= 8&altCol=2 177 charters directly and four do not allow types of authorisers tend to be bigger and 181 For example, in California for-profits to be involved at all.178 The rest more proactive in finding school operators 31% of districts serve fewer than 500 pupils and will typically run do not allow EMOs to hold charters but and then supporting and challenging them only one school while 1% serve do allow them to partner with local found- (a small number of very large school dis- more than 50,000 pupils (includ- ing one district for the city of ers (even within this category, a wide range tricts such as the ones managing education Los Angeles) and are responsi- of conditions prevail). Alongside different in Chicago and Los Angeles could also be ble for 21% of the total number of pupils. www.edsource.org/ 183 rules on per-pupil funding and freedoms placed in this bracket). To take a few data_districtsize06-07.html such as teacher certification and pay, some examples of this: 182 U.S. Department of Education, Supporting Charter states have restrictive rules on governance School Excellence Through which pose a major problem for manage- The State University of New York Quality Authorizing, 2007, p 2; Rotherham AJ, “Smart Charter ment organisations trying to maintain (SUNY), who are allowed to award School Caps: A Third Way on a stable relationship with a school. For up to 75 charters in the city and are Charter School Growth”, in Lake RJ ed., Hopes, Fears and example, in South Carolina the govern- one of the most successful authoris- Reality: A Balanced Look at ing body of the charter is elected annu- ers in the country, work closely with American Charter Schools in 2007, Center on Reinventing 179 ally by staff and parents and in Kansas a not-for-profit called the New York Public Education, p 67 ultimate control of governance remains Center for Charter School Excellence 183 U.S. Department of with the school district.180 Unsurprisingly, who are responsible for finding and Education, 2007; Gau R, Trends in Charter School Authorizing, most of the management organisations supplying applicants who fit with Thomas Fordham Institute, 2006
www.policyexchange.org.uk • 41 A guide to school choice reforms
SUNY’s mission. This can be loosely impact on academic achievement from described as providing intensive aca- a huge variety of organisations, only a demic schools to some of the poorest few of which use high-quality research neighbourhoods in New York.184 methods and/or are genuinely impartial. Bart Peterson, the Mayor of The excellent National Charter School Indianapolis, has used his office to Research Project website lists 77 major attract some of the bigger management achievement studies since 2001. Of these organisations to his city, launched a 77 studies, 20 are rated as having a poor programme to recruit and train new research design and no studies are consid- entrepreneurs to run charters (called ered to have a uniformly excellent design, the Mind Trust) and developed a while just three are considered to be in financing scheme which allows tax- the next highest category: excellent/very exempt interest rates for the acquisi- good.188 The number of studies available tion, construction and renovation of a and the variation in their quality means school facility.185 that it is possible to make pretty much any Central Michigan University, one of argument about the success or failure of the most ‘hands-on’ authorisers in the charters and cherry-pick studies to fit. In 184 U.S. Department of Education, 2007, pp 14 and 18; country and one of the few to receive spite of this, it is possible to draw some www.newyorkcharters.org per-pupil funding for every charter they broad conclusions about performance and 185 U.S. Department of award, charters 58 schools. They helped the impact of different types of legislation Education, 2007, p 17; Skinner D, “Indianapolis Mayor Bart to form the Michigan Public Education and management. Peterson: The Peyton Manning of Charter Schools”, Education Facilities Authority that gives charters Only two studies have tried to take a Next, Summer 2007, Vol. 7, access to the state’s tax-exempt financ- national perspective because of the dif- No. 3 ing status and its credit rating, and ficulty of collecting data and the question- 186 U.S. Department of Education, 2007, p 37; they also developed the Authorisers able value of mixing up states with vastly www.cmucso.org/modules. Oversight Information System (AOIS) different charter laws. Both of these studies php?name=Pages&sp_id=246 that allows them to keep track of stu- were published in 2004 and, as they came 187 Lake RJ, “Should Charter Schools be More Different Than dent attainment and charter compli- to entirely opposite conclusions, sparked Alike?”, in Lake RJ ed., Hopes, ance. The system is now used by other off a nationwide debate about the efficacy Fears and Reality: A Balanced Look at American Charter authorisers. They have subsequently set of charters. The American Federation of Schools in 2008, National Charter School Research up a series of bodies to support other Teachers published the first study which Project, 2008, p xii authorisers in the Great Lakes area.186 showed that, in 2003, students in charter 188 www.crpe.org/cs/crpe/view/ schools on free or reduced price lunches projects/1?page=yes&id=3&p arent= Some have argued that authorisers can be performed significantly worse than those
189 The authors found that too proactive by developing accountability in state schools by the equivalent of around charter school students per- frameworks and compliance rules that are half a year of schooling.189 Shortly after- formed six scale points worse in grade four maths, seven scale just as stultifying as government bureaucra- wards Harvard Professor Caroline Hoxby points in grade four reading, 187 seven scale points in grade cies. Nevertheless, at the moment these (who moved to Stanford in 2007) released eight maths and four scale authorisers seem to be the most successful. a study showing that, nationally, in 2003, points in grade eight reading. Only the first three are statisti- Certainly they seem to attract EMOs and charters slightly outperformed their nearest cally significant. Achievement CMOs, which explains why Michigan, state equivalent by 5% in grade four read- data came from the National Assessment of Educational Indiana and New York have the above ing and 3% in grade four (9-10 year olds) Progress, a standardised average percentages of managed charters. maths and considerably outperformed them annual test run by government- sponsored research institute in states with large numbers of charters (up on a representative sample Performance to 35% in Washington DC). Hoxby used of students. Nelson FH et al., Charter School Achievement on The terrain of charter school perform- a much larger dataset containing informa- the 2003 National Assessment of Educational Progress, American ance studies is vast and treacherous. There tion about state standardised scores for all Federation of Teachers, 2004 are literally hundreds of reports on their fourth graders whereas the AFT researchers
42 USA
used data from the National Assessment of in Idaho).194 Again, the trend is broadly Educational Progress sample, which only positive. None of the studies in either looked at 3% of charter school pupils.190 meta-analysis, however, show particularly Hoxby’s data is therefore more robust but, large differences in either a positive or neg- unfortunately, both studies suffer from the ative direction, so any findings could still same major flaw: neither take into account be due to methodological differences.195 prior achievement as both use only one year Before moving on to discuss differences of data. Even though both studies use some between types of charter school it is worth kind of demographic proxy to compare highlighting one of the state studies fea- schools (free school meals for the AFT and tured in both meta-analyses as it is the only schools in the same area for Hoxby) neither one to employ the ‘gold standard’ of edu- can tell us whether the differences in average cational research: a genuinely randomised scores are attributable to the schools or to experimental design (as opposed to simply the prior ability of the students.191 trying to control for demographic vari- At the state-level there are a considerable ables). This study by Caroline Hoxby and number of studies that do look at achieve- Sonali Murarka from the National Bureau ment over time. The National Alliance of Economic Research, looked at students for Public Charter Schools (NAPCS), in who enter lotteries for New York City a 2007 meta-analysis, identified 40 pub- charter schools and then compared the lished since 2001, of which 19 look at performance of those who were successful value-added to individual students and 21 in winning a place and those who were looked at school- or grade-wide changes not.196 The ‘lotteried-in’ students were over time.192 Of these studies, 21 found found to gain 3.8 scale more points a year that overall gains in charter schools were in maths and 1.6 more in reading for every larger than in state schools, 10 found high- year they were in a charter school.197 This er overall gains for specific types of charter difference is not large but the strength of 190 Hoxby C, A Straightforward school (such as at-risk schools in Texas or the research method makes the findings Comparison of Charter Schools elementary schools in Arizona), five found more robust. Hoxby has undertaken a sim- and Regular Public Schools in the United States, Harvard comparable results and four found that ilar study in Chicago with similar results, University, 2004 charter schools performed worse than state but the researchers only looked at the 191 Gill et al., 2007, p 98 schools (two of these studies come from nine schools run by the Chicago Charter 192 National Alliance for Public 193 Charter Schools, Charter School North Carolina). This broadly positive Schools Foundation so it is more difficult Achievement: What We Know, overview is supported in another meta- to extrapolate wider significance.198 October 2007, p 9 analysis by the RAND corporation, which Of course, as the NAPCS researchers 193 Ibid. focused more specifically on 13 of the put it: “asking about the quality of ‘charter 194 Gill et al., 2007, pp 106-108 methodologically strongest longitudinal schools’ as a group is a bit like asking about 195 Ibid. p 107 studies (RAND’s sample correlates more the quality of ‘new restaurants’ or ‘American 196 As pretty much all the charter schools in NYC are or less exactly to the NCSRP’s top-rated cars’ – any overall generalization will mask oversubscribed this study avoids 199 a potential difficulty of lottery- studies). Out of the 13 analysed studies, the great diversity within.” While it is studies: that charter schools four show overall gains in charter schools important to study the total benefits or oth- requiring lotteries might be bet- (two in Texas, and one each in New York erwise of charters, it is equally important to ter than average. 197 Hoxby C and Murarka S, and Florida), two show overall gains in ask why some charters do better than oth- New York City’s Charter Schools specific areas (Wisconsin in maths and an ers. After all, one of the key benefits of inde- Overall Report, National Bureau of Economic Research, 2007 anonymous large urban school district in pendent state-funded schools is supposed to 198 Hoxby c and Rockoff J, behaviour), five show comparable results be that they encourage innovation that can The Impact of Charter Schools (two in Arizona, two in California and one then be adopted by others. Unfortunately, on Student Achievement, www. innovations.harvard.edu/show- in Michigan) and two show slightly nega- relatively few studies have tried to iden- doc.html?id=4992, May 2004 tive results (one in North Carolina and one tify differences between types of charters 199 NAPCS, 2007, p 3
www.policyexchange.org.uk • 43 A guide to school choice reforms
(though there is far more research on this California, where a fifth of the charters are than for academies or Swedish free schools). ‘non-classroom’, a recent study has shown Nevertheless it is possible to draw out some that they underperform classroom-based broad themes from the available research: schools even though their demographic make-up suggests they should be doing Age much better (students who attend them Charters improve with age. A study in are more likely to be middle-class and Texas by Hanushek et al found that charter white). They also, perhaps unsurprisingly, schools in their first year had a negative have a much higher rate of student move- effect on reading and maths scores while ment between schools.204 This is supported those in the fourth-year of operation or by an Ohio research project which found older had positive effects on maths scores.200 that so-called ‘cyber charter schools’ lag These results were corroborated by another behind state schools whereas other charters Texan study which found that students perform at a comparable level.205 spending three years in established char- ters made larger gains than those in new Pedagogy ones.201 Further supporting evidence comes There has been no analysis of differenc- from Florida, where Tim Sass found that es in performance between pedagogical students performed worse in first-year char- models. The only pedagogical typology ter schools, but those in their fifth year of that we are aware of was constructed by operation were performing better (by 10% Dick Carpenter, Associate Professor at the of the average annual achievement gain).202 University of Colorado, and was based This is an important finding for two reasons on 87% of the charters set-up by 2002 in because it suggests that aggregate results for Arizona, California, Florida, Michigan and charter schools in state or national studies Texas (1,182 schools in total). He found 200 Cited in Gill et al., 2007, pp 99-100 will improve over time and it also offers a that 29% of the schools had a ‘progres-
201 Ibid, p 99 possible explanation for some of the more sive’ pedagogy compared to 23% ‘tradi-
202 Sass TR, “Charter Schools negative longitudinal analyses. As Gill et al tional’ and 29.5% ‘general’ i.e. similar to and Student Achievement in explain, the Michigan study, which featured normal state schools. The remainder were Florida”, Education Finance and Policy Vol. 1(1), Winter, 2006 in both the NAPCS and RAND meta- vocational (12.3%) or ‘alternative deliv- 206 203 Gill et al., 2007, p 108 analyses, only looked at charter schools ery’ such as cyber-schools (6.2%). It is 204 Crane E and Edwards B, in their first two years of operation. This extremely likely that the proportion of tra- California’s Charter Schools: Measuring Their Performance, might explain why it found they were only ditional and general schools has increased Edsource, 2007, p 18 comparable to state schools, even though since 2002 as CMO managed schools, the 205 Hassell B, Summary of Michigan’s charters are generally consid- fastest growing sector, are almost all in District and Charter School Performance in the Ohio 8: ered to be some of the best in the coun- these categories. This could have a signifi- 2006-07, Thomas B. Fordham try. Furthermore, the studies from North cant impact on future state studies if there Foundation, 2007 Carolina were undertaken when the charter is a significant difference in attainment 206 Carpenter DM, Playing to Type? Mapping the Charter school system in that state was less than between these types of schools and pro- School Landscape, Thomas. B. Fordham Foundation, 2005, p 3 five years old, which may go some way to gressive schools. Interestingly, the percent- 203 207 Ibid. 47% of children in explaining their unusually poor results. age of poor and minority students is higher general schools were on free or in general and traditional schools than reduced price lunches compared 207 to 40% in traditional and 36% Classroom Setting in progressive ones. The only clue we in progressive. Similarly, 52% Non-classroom charters, where a signifi- have so far is from Hoxby and Murarka’s of children in general schools were from minorities, compared cant proportion of instruction is deliv- research in New York, which suggests to 48% in traditional and 43% in ered through home schooling or distance there is a correlation between a longer progressive learning, perform considerably worse than school year and higher achievement.208 208 Hoxby and Murarka, 2007, p v those based in traditional classrooms. In Typically longer school years are part of
44 USA
an intensive academic approach that will found that in the first three years of opera- include a traditional pedagogy. tion Edison schools’ performance is similar to the comparison groups, but that they Management then pull away in the fourth and fifth year While there is very little data on types of – fitting with the hypothesis than charters school there have been some studies look- improve over time.212 ing at differences in management – spe- cifically whether EMO/CMO run charters Deprived Communities perform better than those run by com- One of the most interesting developments munity groups, parents and teachers. A over the past few years has been the growth 2003 study of charters in ten states by the of academically intensive schools targeted Brown Center at the Brookings Institute at extremely deprived urban communities. found that students in EMO-run schools This has been driven by CMOs whose main improved significantly more from 2000 focus is typically to find the most effective to 2002 than other charters (which in programmes for children from the poorest turn improved faster than state schools).209 families with the least social capital. This The Edsource study in California that trend began with the spectacular success of showed classroom-based charters outper- KIPP, which has been chronicled in charter forming non-classroom based ones also school research, popular news programmes found EMO/CMO run schools signifi- such as 60 Minutes and in generalist cantly outperforming other types of char- books (for example, Malcolm Gladwell’s ter. After controlling for demographics the Outliers).213 The core elements of the KIPP authors found that 55.9% of students at programme are: a long school day and a EMO/CMO charters achieved proficiency long school year, meaning children spend in 2006 English tests (covering all age 60% more time in school than at a state groups) compared to 49.5% in other types school; parental contracts; tough discipline; of charter. The figures for Maths tests are and a focus on key subjects. Not only has 54.6% for EMO/CMOs and 46.4% for this approach been disseminated through its 209 Brown Center on Education Policy, How Well Are American 210 other charters. own franchise programme (incorporating Students Learning?, Brookings In his 2006 book on the brief history of 66 schools), other smaller CMO federations Institute, 2003, pp 34-35 EMOs, Steven F. Wilson reviews all the such as Achievement First, who operate fif- 210 Crane and Edwards, 2007, pp 14-15 studies available to that point on achieve- teen schools in New York and Connecticut, 211 The other groups did not ment in the biggest six EMOs and KIPP (a have implemented the same strategy.214 Paul provide adequate data for Wilson to make a judgement. It CMO) relative to state schools. While he Hill compares these schools, which he calls is also worth noting that KIPPs finds much of the statistical analysis to be ‘new college-prep charters’, to the urban data was based on just three of their schools, though further of relatively low quality, Wilson argues that Catholic high schools that offered social schools in the network have there is fairly good evidence that Edison (84 mobility to Irish and Italian immigrants in made similar gains subsequent- ly. Wilson (2006), pp 312-313 th schools), National Heritage Academies (57 the 20 century. According to Hill, their (for his full review pp 284-316). schools) and KIPP (66 schools) outperform key shared characteristics are: a demanding 212 Gill B et al., Inspiration, 211 Perspiration and Time: comparative state schools. An analysis of intellectual climate; a ‘centripetal’ curricu- Operations and Achievement Edison schools published by RAND after lum that focuses on mastery of key subjects; in Edison Schools, RAND Wilson had completed his analysis pro- close attention to the progress of individual Corporation, 2005, pp xxiii-xxix 213 Gladwell M, Outliers: The vided more robust support for the largest students, including frequent testing; coordi- Story of Success, Allen Lane, EMO – indicating that school-wide average nation among teachers; a strongly managed 2008, pp 250-269 proficiency rates in maths increased 17% school climate (i.e. tough and consistent 214 Whitman D, Sweating the Small Stuff: Inner-City Schools between 2002-2004 versus 13% in state discipline); overt biculturalism “so that stu- and the New Paternalism, schools serving comparable populations and dents learn the behavioural codes associ- Thomas B. Fordham Foundation, 2008, pp 96-121 and pp 152- 11% versus 9% in reading. RAND also ated with higher education and professional 191
www.policyexchange.org.uk • 45 A guide to school choice reforms
work” – such as the use of KIPP’s SLANT be harder to come by during the current mnemonic;215 and validation of the school’s economic troubles. Perhaps more seri- effectiveness by connecting students with ously, it is hard to retain teachers at such successful graduates.216 schools because their workload is usually In his book on this new trend, Sweating considerably higher than those in other the Small Stuff, journalist David Whitman kinds. As more and more develop it will estimates that there are 200 of these inten- become more difficult to find teachers of sively academic schools nationwide with the quality necessary to support such ven- more being founded every year.217 His book tures. Nevertheless college-prep charters focuses on six of the best known groups are probably the single most promising and includes data showing that they are avenue for reformers. uniformly and significantly outperforming Alongside the evidence regarding aca- local schools. This data is not longitudinal demic achievement, charters are becoming and we await more comprehensive analy- increasingly popular with the public. In sis of this phenomenon. Nevertheless the 2000 the national Phi Delta Kappa poll meteoric rise of these schools is strong on educational issues found that charters evidence of their effectiveness. To take – described as a school under a “contract one example, the American Indian Public that frees them from many of the state Charter School (AIPCS) in Oakland, which regulations imposed on public schools and has now developed into a small federation permits them to operate independently” – of schools, was run from 1996 to 2000 as a found that 42% supported their expansion multicultural school for Native Americans. with 47% in opposition. By 2007 the same By 2000 it was recording 436 out of question produced 60% in support against 1000 in California’s Academic Performance 35% opposed (see Table 4). The few avail- Index (API) tests – one of the worst scores able polls of parents with children at char- in a city not known for great schools. It ters also indicate high levels of support. was then taken over by a new headteacher, A 2000 survey sponsored by the Arizona Ben Chavis, who introduced all of Paul State Board for Charter Schools found Hill’s characteristics listed above. By 2006 that 61% of charter school parents rated the school was one of the best schools in their school A or A+, compared to 38% of 215 Sit-up. Listen. Ask and 219 answer questions. Nod your the state scoring 920 on the API and beat- state school parents. The figure for char- head so people know you are ing the best fee-paying middle school in ter school parents improved the following listening and understanding. 220 Track the speaker by keeping Oakland, Piedmont Hills – all this despite year to an A or A+ rating of 64% and your eyes on whoever is talking. 90% of children at AIPCS being eligible for a Zogby poll of charter school parents in 216 Hill PT, “Equal Opportunity: Preparing Urban Youth for free or reduced price lunches, compared to New York found 42% gave their school an College”, in Lake RJ ed., 2008, 0% at Piedmont Hills.218 A grade compared to 21% who give their pp 27-28 The ongoing development of these types child’s previous state school an A. The poll 217 Whitman, 2008, p xv of schools is not without problems. Many also found that 51% of parents gave their 218 Ibid, pp 68-95 of the CMOs driving the process are heav- child’s charter school an A for instruction 219 www.ncsc.info/newsletter/ 221 June_2002/debate.htm ily dependent on philanthropy which may and 28% a B. 220 www.asbcs.state.az.us/ PSS.asp Table 4: National Support for Charter Schools Over Time (%)222 221 www.manhattan-institute. org/html/cr_37.htm 2000 2002 2005 2006 2007 222 Data from: Larson G, Familiarity Breeds Content: As Favour 42 22 49 53 60 the Charter Movement Grows, So does Public Approval, Oppose 47 42 41 34 35 (National Alliance for Public Don’t Know 11 13 10 13 5 Charters Schools, February 2008), p5
46 USA
Graph 6: Teacher autonomy over various aspects of school life in charter and traditional schools223
70% Charter Traditional 60%
50%
40%
30%
20%
10%
0% Performance Establishing Determining Evaluating Hiring new Setting How budget standards curriculum PD content teachers FT teachers discipline policy is spent
As in Sweden, teachers feel they have have an inbuilt customer base because they more professional autonomy at charter replace failing schools. Unlike Sweden, schools. Figures from the U.S. Education however, there is no largely objective, set Department’s School and Staffing Survey application process for new schools to go last undertaken in 2003-4 show that more through. Each state has different rules charter school teachers consider them- about who is allowed to authorise new selves to have influence on curriculum, schools and charters have developed over- school practice and policy than state school whelmingly in states that allow a range of teachers (see Graph 6). On this basis, bodies to authorise rather than just existing charters are arguably not quite as distant school districts. States with multiple spon- from Albert Shanker’s initial vision as the sors have, on average, nearly eight times teacher unions make out. as many charters as states with a single sponsoring authority and multiple author- Seven Tests isers are the most important variable when accounting for differences in the numbers Demand-led of charter schools across states.224 One of the few things that pretty much Another block to a genuine demand-led all charter school laws across the US share market are caps on the number of charters is that they are demand-led, in the sense allowable at any given time. Only fifteen that the initiative for the development states have no limitation on new charters of new schools comes from non-gov- while others have quite low limits. For ernment organisations. Although most example, the Mississippi legislation allows 223 Christensen J and Lake RJ, “The National Charter School states allow existing state schools to con- only 15 charters; in Rhode Island numbers Landscape in 2007”, in Lake RJ vert to charter status, the vast majority are limited to 20; and in Tennessee to 50.225 ed., 2007, p 11 of charters are new schools that rely on Much of the debate on charter schools in 224 Buckley J and Kuscova S, The Effect of Institutional unmet demand to fill places. Financing these states focuses on whether to increase Variation on Policy Outcomes: The Case of Charter Schools is largely (though not entirely) per pupil, the cap once its ceiling is reached. In New in the States, National Center so if parents cannot be persuaded to sign York the debate lasted years before the cap for the Study of Privatization in Education: Occasional Paper, up the charter has to close. was increased from 100 to 200, meanwhile 2003, #79
In this sense the US is similar to Sweden a backlog of 12,000 students built up on 225 www.edreform.com/index. but distinct from the UK where academies existing charter school waiting lists. As cfm?fuseAction=cLaw
www.policyexchange.org.uk • 47 A guide to school choice reforms
of 2007, 10,000 children in Illinois and EMOs (set-up in the mid-1990s) ran into 16,000 in Massachusetts were on charter trouble was that: school waiting lists, as politicians and lob- byists discussed increasing the cap.226 “When statutory caps…and the difficulty In a recent article Andrew Rotherham of securing customer boards [local partners] argued, in our view accurately, that made management contracts difficult to “statutory caps as a policy approach obtain, the organizations had to look far are too blunt an instrument to address and wide for viable projects. Supporting a quality”.227 He suggests an alternative sys- smattering of client schools in distant states tem of ‘smart caps’ that would avoid the proved enormously costly…EMOs had to problems of unlimited and uncontrolled modify their standard curricula to align expansion while still allowing demand to them with the standards in each state in be met. The two key features of such a which they operated and then develop and system would be: the removal of any caps deploy unique test preparation programs… for federations or franchises that have Customized designs diminished the poten- proven their educational model to be suc- tial for scaling the business, increased cessful in state tests; and an annual allow- corporate staff, and delayed the financial ance of new schools without a proven breakeven point.”230 track record to promote innovation and diversity. States should provide additional Researchers at the National Charter School support to this group of schools in the Research Project confirm that this remains form of start-up funding and ‘charter a problem for the more recent wave of school incubators’ - resource centres pro- CMOs, many of whom: viding technical assistance and adminis- trative support to founders (good exam- “admitted that their organizations started ples include the AppleTree Institute in out spending too little time determining Washington and the Colorado Incubator the circumstances under which they should for Charter Schools).228 The states clos- pursue a new school opportunity or turn it est to this model at the moment are down. Most…initially…did not ‘target’ Ohio, where the cap ceiling has been met districts or states, but responded to oppor- but charters meeting state performance tunities as they presented themselves.”231 targets are exempt and may open new schools for each school meeting targets, Apart from direct limitations on the size and Arkansas, which has a cap of 24 but of markets the biggest problem for poten- also exempts successful existing charters. tial suppliers is caused by rules regarding Even so, neither of these states have the governance. Just six states allow for-profit vital second step of allowing a small EMOs to hold charters directly. In most number of new schools to be incubated states for-profits must have local not-for- 226 Rotherham, 2007, pp 65-66 each year regardless of the total cap.229 profit partners who are responsible for 227 Ibid, p 66 governing the school involved. Finding a 228 Ibid, pp 70-71; dlc.org/ Easy to Enter suitable partner means another significant ndol_ci.cfm?kaid=139&subid=27 3&contentid=251957 Needless to say, the artificial limitations of restriction on the number of available 229 www.edreform.com/index.cf single authorisers and caps can cause frus- opportunities, mitigating against proper m?fuseAction=cLaw&stateID=33 &altCol=2; www.edreform.com/ tration for suppliers unable to gain access planning. Even when partners can be index.cfm?fuseAction=cLaw&sta to the market. It is especially problematic found, the complex relationship between teID=13&altCol=2 for EMOs and CMOs looking to develop management organisation, local governing 230 Wilson, 2006, pp 191-192 federations. As Steven Wilson explains, body and authoriser can produce signifi- 231 Harvey and Britt ed., 2007, p 24 one of the reasons that the first wave of cant tensions. As Wilson notes:
48 USA
“This structure gives rise to immense prob- below) and the applications process can be lems. Companies, their client boards [i.e. extremely complex and time-consuming. local partners], and principals vie for con- This is why authorisers (e.g. the Mayor of trol over the new school…even when the Indianapolis and the State University of local board is willing to delegate much of New York) who work closely with charter its authority for day-to-day management to incubators have proved more successful the contractor, state regulators and charter than others. authorisers often prevent it….When boards planned a stark delegation of responsibility Accountable to their contractors, regulators insisted on Unlike in Sweden, where the absence of inserting language throughout the contract data creates a real accountability deficit, that rendered the EMO’s powers ambigu- there is plenty of information about the ous by subjecting them to ‘the board’s ulti- performance of individual charter schools. mate oversight’.”232 They have to participate in whatever bat- tery of standardised tests their state uses Wilson goes on to record the CEO of to comply with the federal No Child Left Mosaica, one of the larger EMOs, describ- Behind Act. Like all schools, charters are ing avoiding conflicts with ‘customer expected to meet their Annual Yearly boards’ as the “single biggest problem we’ve Progress (AYP) target on these tests and had” and the CEO of SABIS, another large the information on their performance is EMO, questioning the logic of authorizing available to parents and legislators. In addi- EMO-affiliated charters, premised on the tion, charters have to be explicitly renewed EMO’s school design and performance by authorisers (usually every five years) so record, and then permitting local boards accountability is built into the process. to unilaterally terminate that arrangement Nevertheless, there are serious questions and continue to operate the charter.233 over the ability of some authorisers to keep The NCSRP researchers concur, in the their charters accountable. In their study of context of CMOs, arguing that “reducing authoriser decisions to approve or revoke the cost of finding good partners is essen- charters Bryan Hassell and Megan Batdorff tial to the financial viability of MOs” and argue that most are well-founded.235 There that “problematic partners are expensive have, however, been a number of high to maintain and often damage an MO’s profile cases of extremely lax oversight, like reputation”.234 the unforeseen collapse of sixty charters run For non-EMO/CMO founders (e.g. by the California Charter Academy (CCA) parent, teacher or community-run schools) group, many of which were authorised by the barriers to entry are less complex as different school districts from the one in they will either be located in an area sup- which they were located (a practice that portive of charters or not, though they will has since been stopped).236 Research com- 232 Wilson, 2006, p 223 233 Ibid, pp 226-227 still be affected by caps and the inability missioned by the Thomas B. Fordham 234 Harvey and Britt ed., 2007, in some states to apply to more than one Foundation found that 44% of authorisers p 31 authoriser. While these founders typically practice limited oversight of their schools 235 Hassell B and Batdorff M, have full operational control, they may and this was more likely to occur among High Stakes: Findings From a National Study of Life-or-Death choose to contract out some managerial school district and county based authorisers Decisions by Charter School and administrative functions. For smaller rather than independent chartering boards Authorizers, Public Impact, 2004 236 Fiscal Crisis and community groups who are interested in and university authorisers. They also found Management Assistance Team, setting up one-off charters, the biggest that four-fifths of authorisers would use Extraordinary Audit of the California Charter Academy, problem is one of resources. Funding additional staff to monitor academic per- www.cde.ca.gov/sp/cs/ac/docu- is rarely available for new projects (see formance and two-thirds would like addi- ments/csccaaudit.pdf
www.policyexchange.org.uk • 49 A guide to school choice reforms
tional technology to monitor accountabil- critical oversight if introduced without her ity.237 Robin Lake, Director of the NCSRP, other suggestions); authoriser report cards, has argued that this lack of oversight could aggregating results for all their schools and be the “achilles heel of the charter school publishing them; independent perform- movement” as “the existence and persist- ance audits; specific legislatively mandated ence of low-performing charter schools is performance goals, perhaps using stand- the primary contributor to uneven charter ardised No Child Left Behind data; and school quality. In the end, if the charter the development of markets for authorisers school movement fails to prove itself as in states that limit authorising to a single a viable source of higher quality public agency or type of agencies.241 schools, bad authorizing and oversight will probably be a major reason.”238 Genuinely free There is plenty of good practice on The extent to which charters have to con- this issue, especially amongst the large form to various state regulations varies authorisers who frequently offer far more significantly between states. With EMO/ nuanced and productive accountability CMO run schools there is usually the structures than any national government additional complexity of whether the man- could provide. The US Department of agement organisations or the client board Education have produced a checklist of hold specific powers. Nevertheless, the practices established by the best authoris- majority of charters have significant free- ers: they use information and technology dom. There is no national curriculum to to streamline compliance with regulations follow, though curricula are prescribed to and performance reviews using manage- some extent by the content and nature of ment systems (like the one developed by state standardised tests, and there are typi- Central Michigan University); they use cally few restrictions on school design or site visits strategically – some authorisers, buildings. Indeed, many charters are set- like Ferris State University have developed up in church halls, office blocks and spare their own small inspectorate teams; they rooms of existing district schools. know when to intervene early by keeping a Freedom over teacher tenure and pay close eye on, for example, board minutes; are more complex. Rules regarding pay dif- and they base decisions on solid evidence fer widely between states, perhaps reflect- rather than conjecture or parental pres- ing charters’ muddled origins as an ini- sure.239 Moreover, the best authorisers tiative by teacher unions. Six states require 237 Gau, 2006, p 12 spend more time on the initial application charter teachers to be included in collective 238 Lake RJ, Holding Charter process, incubating good ideas, spending bargaining with local district schoolteach- Authorizers Accountable: Why It Is Important and How It Might time getting to know applicants and clari- ers and a further seven states assume col- Be Done, NCSRP White Paper fying expectations, so as to avoid future lective bargaining but may allow some Series No. 1, February 2006, p 1 240 239 US Department of problems. waivers or negotiated freedoms in individ- Education, 2007, pp 35-49 Robin Lake has suggested a number ual charter agreements.242 In contrast, the 240 Holland M and Rainey of ways in which state legislatures could states containing the majority of schools L, Finding a Balance: How Application and Authorization ensure that more authorisers adhered to allow charter founders to bargain directly Policies Impact School Supply, these practices. These include: increas- with their teachers. Just 10% of charters NCSRP Working Paper 2008-4 ing transparency by forcing mandatory are in states which require collective bar- 241 Lake, 2006, pp 5-8 disclosure of formal policies on approval, gaining, with 54% operating in states with 242 The Center for Education Reform oversight and revocation; allowing charters no requirements and the rest operating in 243 DeArmond M, Gross B and to anonymously rate authorisers and even states which either require teachers to opt- Goldhaber D, Look Familiar? 243 Charters and Teachers, in Lake switch to different authorisers (although out of bargaining (4%) or opt-in (32%). R ed., 2007, p 45 Lake acknowledges this could lead to less Analysis of the 1999-2000 School and
50 USA
Staffing Survey (SASS) indicates that start- Financially consistent and stable up charters and those without collec- Differences in funding, both between pub- tive bargaining are much more likely to lic and charter schools and between differ- use non-traditional salary schedules and ent states, are one the biggest headaches merit pay than conversions or schools in for the charter movement. Much of the areas with collective bargaining. Around problem comes down to the overwhelming 45% of start-ups use their own schedules complexity of the school funding system and/or merit pay.244 Those authorised by which operates at a federal, state and local independent or not-for-profit authoris- level and consists of numerous different rev- ers are also significantly more likely to enue, special programme and capital grants. use merit pay than those authorised by This complexity, and the localised nature of school districts. This freedom is considered each individual school’s funding means that important by management organisations per-pupil funding is not consistent, seri- (particularly EMOs) as they tend to have ously affecting the operational capability of specific management structures that differ charters and distorting the market. from traditional state schools and most use The majority of state laws require that detailed performance evaluations attached charters receive 100% per-pupil fund- to bonuses or salary increases.245 ing. Some states such as New York In many states, charters are able to hire and Pennsylvania offer far less (70% and a certain percentage (in some cases 100%) 70%-82% respectively) but this typically of non-certificated staff who have not been relates to just federal and state revenue through teacher training courses. Some funding. In a 2005 analysis of sixteen groups, like the American Indian Public states plus Washington D.C., researchers Charter Schools discussed on p.46, use commissioned by the Thomas B. Fordham this freedom as a good way to attract high- Foundation found that charters receive on quality graduates who do not want to go average 21.7% less funding than public through teacher training. According to schools (about $450,000 less a year for the 2003-2004 SASS 78% of state school a 250 pupil school). The largest gap – teachers earned their undergraduate degree 39.5% - was found in North Carolina.249 from an education programme compared Most of this gap is due to the inability of to 63% of those in charters, and 41% of charters to access local revenue and facili- state school teachers held a Masters degree ties finance, usually funded by local taxes 244 Ibid, p 45-47 from an education programme compared and administered by school districts (no 245 Wilson, 2006, p 133 to 26% in charters.246 Finally, and per- states in the sample offered full access to 246 Lake R, in Robin Lake ed., 2007, p 9 haps most importantly, very few charters local or facilities funding).250 There are 247 Robelen EW, “Efforts to are unionised, potentially as few as 2%, significant differences between states in Unionize Charters Grow, But which means that it is far easier to remove the extent to which they use local taxes to Results So Far Appear Modest”, Education Week, March 22 2006 247 underperforming teachers. Most char- support schools ranging from 64% of all 248 Wilson, 2006, p 134 ters, especially EMOs and CMOs, consid- funding in Illinois to 15% in New Mexico 249 Speakman S and Hassell er this to be one of their most important and, notably, the three states with the low- B, Charter School Funding: Inequity’s Next Frontier, Thomas advantages over traditional schools, who est percentage of school funding coming B. Fordham Institute, 2005, p 6 find it even harder to remove poor teach- from local taxes also have the lowest gap 250 Ibid, p 13 251 ers than their counterparts in England. As between public and charter schools. 251 Ibid. Alan Olkes, Head of Human Resources The funding gap is almost certainly the 252 See, for example, Hutton P, Eckerling W and Balczarek for Chancellor Beacon Academies, one biggest single problem for charter school A, Short-Changed Charters: of the bigger EMOS, put it: “One of founders, especially when it comes to find- How Funding Disparities Hurt 252 Colorado’s Charter Schools, our key advantages is that we can hire and ing and supporting facilities. It is largely Colorado League of Charter fire people.”248 intractable as there is only so much a state Schools, 2008
www.policyexchange.org.uk • 51 A guide to school choice reforms
government can do to force the hand of of each district (of which there are 13,500 local districts. In California, for exam- in total). As the CEO of Charter Schools ple, districts are now required to provide USA, a large EMO, has put it “the contin- charters with facilities to accommodate ued political undercurrent requires charter students in “conditions reasonably equiva- operators to spend an enormous amount lent to” public schools in the district, yet of time defending [themselves]…while still many districts (who are also competitors having to operate schools.”254 of the charters) have gone to the courts As with free school operators in Sweden, to argue against this requirement, while most EMO/CMOs now spend a consid- some EMOs have been reluctant to push erable amount of time developing com- the issue to avoid upsetting their authoris- munity support before applying to open ing district. Other support in the form of a school, but this is both costly and time- state ‘lease aid’ is subject to annual appro- consuming. Furthermore, even if charter priation and therefore cannot be budg- founders are successful in winning sup- eted for (the same is true in Minnesota). port for initial proposals, those author- Only Washington D.C. grants charters a ised by directly elected institutions are per-pupil capital allowance equivalent to vulnerable to wider changes in political public schools (averaged over five years) momentum. The Mayor of Indianapolis, and this is only possible because there is who has politicised his role more than any no distinction between state and local gov- other authoriser, has had to work harder ernment.253 Needless to say the differences than others to institutionalise practices between states also cause huge problems and win support for charters so that the for groups attempting to develop national next mayor is forced to carry on with the federations. Responses to these problems programme.255 Of course, the direct politi- are likely to be piecemeal at best unless cisation of charters can work both ways, individual states or the national govern- with poor schools kept open despite the ment were to reduce the extent to which concerns of authorisers, because they do schools are funded locally (by introducing not want to risk upsetting influential com- something like the Local Management of munity groups or parents.256 As discussed Schools implemented in English schools earlier, this occasionally weak account- from 1988), admittedly an extremely ability is one of the biggest threats to the unlikely scenario. charter movement.257 Even after taking these issues into Politically stable account, the main problem created by the Charter school initiatives in most states direct politicisation of authorising remains are less politically stable than Swedish free the extreme instability it causes within schools or academies because the majority individual districts. Steven Wilson relates of authorisers (school districts) are directly the story of Edison winning an initial vote elected and charters are typically short- from the San Francisco board of education term contracts with a built-in timescale for to take over one of the worst schools in renewal. Even for charters that are author- the city in 1998, only to find new board ised by non-elected bodies like universities members elected two years later who were or private not-for-profits, districts still have viscerally opposed to the idea of profit- 253 Wilson, 2006, pp 203-204 huge influence over funding, facilities and making schools. Despite Edison’s invest- 254 Harvey and Britt, 2007, p 20 planning. Not only does the charter move- ment in new facilities, improved scores and 255 U.S. Department of Education, 2007, p 55 ment have a permanent battle to maintain almost unanimous support from parents 256 Ibid, pp 55-56 or develop strong legislation in each state, and teachers, the new board members 257 Lake, 2006 it must also respond to the local politics ended up forcing a revocation of the char-
52 Lessons from North America
ter. Edison were allowed to transfer their account in three states (Delaware, New charter to the state board but only on con- Hampshire and Rhode Island), none of dition that they paid $300,000 costs, paid which have many charter schools.260 ongoing rent to the San Francisco board In practice charters have been pre- and agreed not to manage any additional dominantly focused on the poorest com- schools in the district.258 The role of profit munities as this is where the demand for remains the single most contentious politi- higher-quality state-funded education is cal issue, due to the largely illogical yet to be found. There is evidence that oppo- intuitive dislike many people have towards sition to charters is much lower in poorer any relationship between commerce and areas and Black/Hispanic voters are more children. Wilson also relates the story of likely to support charters than the public a KIPP director who, when approaching in general.261 Many EMOs attempting to new school district boards, is regularly operate in single authoriser states have asked “are you for-profit?” and when he found that the only districts prepared replies that they are not is told “Oh Good. to countenance for-profit are the most Now I can talk to you.”259 desperate. CMOs are even more focused on the poorest communities as they are Fair philanthropic institutions. This means One of the original concerns of charter that charter schools, on average, have far school opponents was that they would lead poorer intakes than public schools (in 258 Wilson, 2006, pp 272-281 to an increase in segregation as allowing contrast to Sweden where for-profits are 259 Ibid, p 98 260 The Center for Education parental choice might result in cream- able to open in most of the country and Reform; National Center for skimming. As it stands, the vast majority admissions are done on a first-come-first- Education Statistics, nces. ed.gov/ccd/tables/2009305_02. of charter schools laws were designed to served basis). A recent survey of charters asp give all parents an equal chance of winning found that 52% of students are from 261 Larson, 2008, p 5 a place if they applied. Only five states ethnic minorities, compared with 44% in 262 The Center for Education do not mandate that over-subscription the entire public sector, and 54% are low- Reform, Annual Survey of America’s Charter Schools to charters must be settled by lottery income (as defined by uptake of free or 2008, p 12; National Center for rather than academic ability or interview. reduced-price lunches) compared to 41% Education Statistics, see nces. ed.gov/pubs2007/ruraled/tables/ 262 Academic ability can only be taken into in the public school sector. table1_8.asp
www.policyexchange.org.uk • 53 A guide to school choice reforms
Lessons and recommendations
In the last three chapters we have provided The Ten Key Lessons of School Reform information on the politics and perform- ance of independent state-funded school 1. Once established ISFS (Independent (hereafter known as ISFS) reforms in State-Funded School) systems grow stead- England, Sweden and the United States. ily and reforms are difficult to reverse We have looked at the background to In the US and Sweden the number of char- reform, evidence on academic achievement ter schools and free schools have grown and seven ‘tests’ designed to identify which steadily every year since reforms have programmes have been best at solving spe- been introduced. In neither country was cific policy problems. As Table 5 shows, no there an explosion of interest after the country has designed a system that resolves first reforms but sizable markets developed all of these difficulties. But because at least as initial problems were ironed out and one country has found a solution to every potential entrants were assured about the problem, we believe it is possible to develop longevity of the reforms. In the US there a programme that adopts the best aspects of are now 4,568 charter schools educating all three systems and could be implemented 1,341,687 children – or 2.7% of the total. in this country. There are cultural and polit- In Sweden, where there are fewer barriers ical differences between the three countries to setting up new schools, approximately that prevent some ideas from being directly 11.9% of children are educated in 3,302 transferable. Even so, many of the issues, free schools (including pre-schools). In such as tensions between local and central this respect the academies programme in government and inequalities between rich England is somewhat different as its expan- and poor, are faced by reformers in each sion has been driven by government targets nation. In this chapter we start by drawing rather than demand. Nevertheless, as with together what we believe to be the ten key charter and free schools, academies enjoy lessons from the previous chapters and then the broad support of all mainstream politi- go on to give our recommendations for cal parties due to their popularity with reform based on these lessons. parents and the success of major alternative
Table 5: How each country fares in our seven tests
Easy to Genuinely Financially Politically Demand-led Accountable Fair enter free consistent stable
UK X X
USA /* /* X X
Sweden X X X
indicates that the evidence on whether a system passes the given test is mixed * These categories have been given two ratings because of the huge difference between states that have multiple-authorisers and are, therefore, more demand-led and accountable, and states that only allow school districts to authorise.
54 Lessons and recommendations
providers like ARK. In all three countries, operating within the same system. Far more however, the political right are typically research in all three countries is needed to more supportive of the process than the help explain what works and what doesn’t. left, who have to accommodate the usu- However, it is possible to draw out some ally hostile views of affiliated trade unions, broad themes. On the basis of the success even though over time unions find them- of the Harris and ARK groups in the UK, selves with members in ISFS and have to the charitable CMOs in the US and the moderate their hostility accordingly. The popularity of more traditional groups such left are particularly hostile to commercial as the International English Schools in providers and, in the US and Sweden, Sweden, we believe that providers offering centre-left politicians have moved away an academic education with a focus on core from attacking the general principle and subjects and strong discipline, especially to towards attacking the profit motive. children from deprived communities, may be the most successful. These providers 2. Most studies of attainment in ISFS also counter the low expectations often show a positive effect: held for children from such communities. ‘Evidence’ can be cherry-picked to support We hope that researchers in each country any line on the benefits or otherwise of will focus on this aspect of reform over ISFS reform. In this report we have tried the next few years. We also believe there to include at least some mention of every is strong evidence, especially from the US, available major study in the US, Sweden that federations perform better than one- and England. In short, the overall pic- off schools – perhaps, in part, because they ture tends towards the positive, especially are more likely to offer more traditional when research methodology is taken into curricula based on core subjects. This account. In this country there is too little theme is explored in more detail in the achievement data available to draw strong next three sections. conclusions about long-term average trends. Nevertheless the data from the recent PwC 3. A system based on independent state- evaluation suggests academies are improv- funded schools moves naturally towards ing much faster than other local schools. federation: Longitudinal datasets with proper controls In all three countries the initial intention will be needed to reach firmer conclusions. of reformers was not to create federations. In Sweden, the lack of national tests before Those behind the academies programme 15 circumscribes the kind of analysis pos- assumed – based on the experience of City sible but all three main studies looking at Technology Colleges – that local sponsors the impact of an increase in free schools on from the business community would come the performance of any given municipal- forward to sponsor one-off schools. As the ity are positive. In the US just four out of programme has developed some of the the 40 longitudinal studies assessing the most important sponsors, like ARK and impact of charter schools on achievement Harris, have taken on a number of schools are negative, while 31 found some kind of and created branded networks but that gain – which is particularly notable given has come about organically and, because that charters receive, on average, 20% less of this, unnecessary barriers remain to this funding than public schools. way of working, such as the sponsorship While analyses of the overall impact of fee for every new academy and the inability ISFS are important justifications for pursu- to move money between schools. ing this policy, they can hide significant In the U.S. and Sweden reformers disparities in the performance of ISFS assumed that schools would be run by
www.policyexchange.org.uk • 55 A guide to school choice reforms
local community groups or teacher/parent for-profit federations is indicative – at the co-operatives. Indeed, the very first ISFS very least – of the benefits of economies in both countries conformed to expecta- of scale. The two studies from the US on tions as parents took control of rural this topic that compare federations to one- schools in Sweden and teachers opened off charter schools both strongly favour the first charters in Minnesota. However, federations, despite the barriers to their in subsequent years federations (typically development. Although more evidence is using general or traditional academic-style still needed in this area, the performance curricula) have become more prominent. of CMOs like KIPP and AIPCS seems In Sweden, where there are few barriers especially strong. to federation, the majority of schools are now run by for-profit companies and the 5. Allowing commercial companies to set- vast majority of new applications are for up ISFS significantly boosts the potential this type of school. In the US there are for federation much stronger barriers: for-profit provid- The question of profit is one the most ers can only run schools directly in six emotive aspects of ISFS and has been the states; state laws differ widely meaning focus of discontent on the left of American scalability is compromised and funding is and Swedish politics. There seems to be extremely variable between states and even little basis for this, or for the queasiness school districts. Nevertheless, as many as over potential profit-making schools in the 30% of charter schools are involved with UK, beyond an intuitive dislike of the idea management organisations in some capac- that money could be made from educating ity. The growth in the number and size of children. The most common argument charitable CMOs over the past few years, used in opposition - that it takes money with many of them focusing on provid- away from schools to line the pockets of ing ‘college-prep’ academic education to shareholders - simply ignores economic deprived communities, is a particularly reality. If it were true it would be an argu- interesting development. ment for nationalising all industries. We do not believe that the inclusion of for- 4. ISFS in federations seem to perform profit firms in any reform is essential to better than one-offs: making reform work – and reformers may In the UK and Sweden there have been consider the additional opposition it cre- no statistical comparisons of perform- ates not worth the trouble. However, for- ance between federations and one-offs. profit groups are much more likely to have Nevertheless the initial data on academies the scale and ambition necessary to create suggests that multi-academy groups are multi-school federations. Larger charities outperforming sponsors with one or two with a national or regional focus, such as schools. Furthermore, in their qualita- Harris and ARK in the UK or KIPP and tive analysis PwC emphasise the signifi- the other CMOs in the US, are also able cant benefits of this kind of sponsorship, to do this, but such groups are relatively including the potential for developing few and far between, especially during an school leaders in-house, the scalability of economic downturn. Parent or teacher co- curricula and economies of scale. To this operatives and community groups are, by we would add the potential for developing their nature, focused on one locality. a coherent brand which conveys informa- In Sweden, where there are no restric- tion about quality to parents and the abil- tions on profit, there are far more national ity to set up multi-school teacher training multi-school federations, and ISFS repre- programmes. In Sweden the dominance of sent a far higher proportion of the total
56 Lessons and recommendations
number of schools. However, the lack of ers come closest to managing this. This is effective quality-control means that there is because schools can approach more than anecdotal evidence that some less impres- one authoriser, so reducing the risk to diver- sive groups have gained access to the mar- sity, but authorisers themselves are in com- ket, perhaps with a quick profit in mind. petition so typically take a more rigorous Because the availability of profit increases approach to accountability and oversight. the incentive to set-up schools it also Such a system would work even better if increases the need for strong accountabil- authorisers were themselves accountable to ity – both in the commissioning process government and parents for the perform- and in continuing oversight. Of course, if ance of the schools they oversee. for-profit schools are not offering any value then parents will soon desert – forcing it to 7. Existing local government providers close anyway – but education is too impor- should not be able to veto provision but tant to risk even brief market failure. also should not be prevented from par- ticipating in reforms: 6. The authorising process is crucial for In all three countries the traditional monop- the success of ISFS reform oly school provider has been local govern- Authorising is an often overlooked aspect ment and most local authorities (or munici- of ISFS reform, perhaps because the idea palities or school districts) have been hos- of a market in schools is suggestive of an tile, at least initially, towards ISFS reforms. organically evolving sector. However, given In the UK this hostility has manifested itself that significant amounts of state funding in the refusal of many authorities to partici- are involved, someone has to decide who pate in the programme despite the govern- has access to this market. This decision ment’s enticing variety of sticks and carrots can be made at different levels of national to encourage involvement. Alternatively, or local government or remitted to non- local authorities display a grudging accept- governmental bodies. The authorisation can ance of academies on condition that they involve the application of minimum stand- can co-sponsor them – retaining significant ards (in the knowledge that it is impossible control. In the US, relatively few school to make the process entirely objective) or districts (7-8%) have been prepared to a more rigorous commissioning approach authorise charters and often after lengthy designed to favour specific types of schools political battles. Where charters have been (which is subjective in any case). The benefit authorised by other bodies the districts in of the ‘minimum standards’ approach is which they are set up typically refuse to that it increases diversity and innovation in offer fair funding and can be extremely hos- the market, while the benefit of the ‘com- tile unless significant groundwork is done missioning’ approach is that it reduces the in advance to soften up local opposition. In potential for market failure. The Swedish Sweden municipalities play no direct role system tends towards the former while, in in the authorising process, although they most states, the US tends towards the lat- are allowed to file an objection which the ter. In contrast the academies programme national authoriser has to take into account, authorising mechanism is entirely lacking in but again those whose support is not won transparency and depends on the available in advance can make life difficult by with- local sponsors and the whims of the relevant holding suitable buildings or failing to offer local authority. equivalent funding. In practice the best option is probably to Given this pattern of hostility it would balance the two approaches and we would not be sensible to introduce any reform argue US states that have multiple authoris- that gives local government an effective (de
www.policyexchange.org.uk • 57 A guide to school choice reforms
facto or de jure) veto over new provision. 9. Funding needs to be fair and consistent That does not mean that local government In England academies are funded cen- should never participate in the authorising trally according to the per pupil revenue process. Some authorities, districts and that schools in the same authority would municipalities have embraced the potential receive, plus start-up funding and the of ISFS, especially those in larger urban money authorities would receive to pay for areas (like Stockholm or Chicago) and/ central services. As academies also receive or with a enhanced strategic capacity (like free new buildings there are no problems the privatised authorities in Hackney and with capital costs. This works acceptably at Stoke). Where possible, local government the moment but if ISFS reform were to be should play a significant role in the devel- broadened a transparent national funding opment of local markets. formula would be preferable to hundreds of bespoke agreements. In the U.S. fund- 8. Accountability is difficult to manage ing for charters is handled by a chaotic mix at a national level of federal, state and local agencies lead- The problem of entirely ignoring local gov- ing to confusion and disparity. Charters ernment is that it is difficult to authorise receive on average 20% less than public and oversee schools from central govern- schools, primarily because they do not ment. Many of the difficulties faced by receive equivalent funding for facilities, the academies programme stem from the which seriously limits their ability to suc- DCSF trying to manage it centrally. This ceed. In Sweden funding is controlled by approach has led to a chaotic commission- municipalities and there have been prob- ing process in which local authorities and lems over how to fairly calculate a per pupil sponsors are confused about their role and equivalence which takes leasing costs into has engendered a total lack of transpar- account. There are also problems caused by ency. Moreover, if an academy fails – as differences in the proportion of local taxes has happened on a number of occasions – allocated to education in different munici- the DCSF feel compelled to get involved, palities. In both the US and Sweden, which can mean throwing money at the providers with some mobility (i.e. com- problem. Trying to manage hundreds of mercial companies and national/regional schools in this way will become increas- charities) have focused their attention on ingly implausible. A new national quango states, localities or municipalities with would be unlikely to manage much bet- higher rates of funding – thereby skewing ter. The experience of Sweden here is not the market. Consistent and transparent encouraging. In a country a fifth of the national formulas, which are nuanced to size of England, the schools’ agency has take into account local differences, are the not proven itself able to manage quali- only way to resolve this difficulty. ty-control or long-term accountability as effectively as one would hope. In fact, at 10. Choice does not necessarily lead to the end of 2008 responsibility for manag- segregation but admissions policies have ing free schools passed to the relatively new to be set carefully: schools’ inspection agency (their equiva- In all three countries a central concern for lent of Ofsted). Their oversight is likely to opponents of ISFS reform has been that be more extensive as it will link up with choice may lead to further segregation of more regular inspections and with new the education system on the basis of class high-stakes tests. Nevertheless it simply and income as wealthier professional fami- cannot have the nuance of a more local lies will have better access to information. and focused authoriser. In a sense this is more of an issue in the
58 Lessons and recommendations
US and Sweden where state school places academies programme – which embraces are typically allocated by government. In some important aspects of reform, like England parents already have a choice so independent provision, but also bears little the issue of providing useful information resemblance to a proper market. It makes to parents is crucial anyway. Nevertheless, sense, therefore, to start by adapting the providing a wider range of choices could academies programme, making the proc- theoretically lead to further segregation. ess of entry for providers easier and more There are two ways of mitigating this. transparent. Once proper commissioning, First, to make sure that parents are not funding and accountability mechanisms penalised for being slower off the mark. are developed then reforms should be wid- In Sweden admissions are done on a ened to allow existing academy providers first-come-first-served basis which does and new providers to set up new schools give a clear advantage to middle-class directly in response to demand. This could parents. In the US oversubscription is usu- be done using smart caps to regulate provi- ally dealt with by lottery which removes sion without diminishing the potential for the ‘early-bird’ advantage and in England diversity and innovation. Our four-stage academies have to abide by the admissions process of reform would look like this: code which gives a variety of criteria for dealing with oversubscription including 1. The academies programme should be the use of catchment areas and fair band- immediately reformed by removing bar- ing. The second way is to more actively riers to entry and developing a transpar- incentivise providers to set-up in more ent commissioning process. deprived communities. This has happened The DCSF should identify those schools it to some extent in the US by default as wishes to turn into academies (regardless of school districts in deprived urban areas whether the relevant local authority agrees) have been more prepared to try radical new and give its criteria for doing so, which approaches. In England the government should be more nuanced than just failing has limited academy sponsors to replacing to reach the 30% good GCSEs target (per- poorly performing schools (by their defini- haps making use of the school report cards tion) which are usually in deprived areas. being introduced by the government)263 An approach that was genuinely demand- and should include some measure of pub- led might require incentives for providers. lic demand in the area. They would then invite public bids, in manageable waves, Our Recommendations from any interested providers. There should Our first recommendation is that it makes be clear criteria for bidders with prefer- sense to think of school choice reform ence given to groups who are successfully as a series of stages rather than a ‘big running other academies, which would bang’. That way the system can be devel- encourage federation, and those who have oped in a coherent fashion rather than prior educational experience. The winner reactively in the face of unexpected dif- would be subject to proper public con- ficulties. It also should reduce the risk of sultation. Contracts should be simplified 263 Following a recommenda- initial market failures, making the reforms and freedoms over curriculum and teacher tion from Policy Exchange in easier to defend politically. We should certification should be restored. Lim C and Davies C, Helping Schools Succeed: A Framework be able to predict what problems might As the commissioning process would for English Education, Policy arise in advance as we have the examples be clear and transparent there would be Exchange, 2008, pp 58-72 ; DCSF, A School Report Card: of Sweden and the US to study. Unlike no need for providers to pay sponsorship consultation document, 2008, the Swedes and Americans in the early- fees (which act, at the moment, as a crude see, publications.teachernet.gov. uk/eOrderingDownload/DCSF- 90s, we already have a hybrid model –the quality-control barrier) so groups without 01045-2008.pdf
www.policyexchange.org.uk • 59 A guide to school choice reforms
access to capital would be able to bid. We manage. This is not an argument for sim- would also allow commercial organisa- ply handing control to local authorities, tions to bid, which would not require many of whom remain hostile to the prin- any change in legislation as for-profit ciple of ISFS. Instead the DCSF should companies are already allowed to run state look to approve multiple local and regional schools. There is simply no logical reason authorisers. Local authorities would be able to exclude them as long as they can show to apply to authorise but so would other in a bidding process that they would be devolved local and regional governmental more effective than other alternatives and institutions such as Regional Development they are held accountable for their per- Authorities (RDAs) and elected mayors, formance on an ongoing basis. Dropping and even non-governmental organisations, the sponsorship requirement and allow- specifically universities and educational ing commercial organisations to bid and charities. A key advantage of including make a profit would increase the number universities is that it would help develop of providers participating in competitions, closer links between secondary and higher thereby boosting standards. education, boosting the drive for wider While we think it is sensible for the participation. They have also proved to DCSF to actively intervene to replace be some of the best authorisers in the US. the worst performing secondary schools Some of the authorisers would have a local we would also widen the scope of the focus while others might look to author- academies programme. First, to include ise across regions or even nationally. The poorly-performing primaries, who are cur- DCSF would retain its ability to authorise rently excluded for no obvious reason, as a last resort for areas as yet uncovered by and secondly to allow the governing bod- any other body. ies of successful schools to convert to Existing academies would have to agree academy status. This would entail setting to be transferred to another authoriser up a trust that could then also bid for as their contracts state that the DCSF is other schools. The Conservative Party have responsible for oversight. However, in the already moved in this direction by propos- future, schools considered to be perform- ing to give the top 400 secondary schools ing at a very poor level (a group that should the same freedoms as academies in return continue to rapidly diminish in number) for federating with a poorly performing would be identified by the DCSF criteria, school.264 That said, we don’t see why this go through a centralised bidding process should be limited to 400 schools or why and then be assigned to an authoriser of the the initial quid pro quo is necessary. There winning bidders choice. High-performing should be a wider entitlement with fewer schools that wished to become academies strings attached in the expectation that would apply directly to an authoriser of some high-performing school trusts would their choice. The DCSF would be left in go on to build federations. charge of existing academies that wished to remain under its control and a small 2. The DCSF should transfer the over- number of new academies without access
264 Prince R, Tory party confer- sight of academies to a variety of local to any other authorisers. Local authorities ence: Schools to get budget and regional authorisers. would retain control (regardless of whether freedom under Conservative plans, The Daily Telegraph, As the academies programme expands – they chose to become authorisers) of chil- 30th September 2008 www.tel- which would happen faster if primaries dren’s services and any aspect of education egraph.co.uk/news/newstopics/ politics/conservative/3108378/ and high-performing schools were allowed that required local co-ordination such as Tory-party-conference-Schools- to-get-budget-freedom-under- to participate – it will become impossible diplomas, information about admissions, Conservative-plans.html for the DCSF, or any national agency, to exclusions and transport.
60 Lessons and recommendations
The network of authorisers would then authority activities would be funded sepa- be responsible for monitoring the per- rately. The per-pupil amount would con- formance of their academies and re-com- sist of three elements: a base element (dif- missioning any school that was failing or ferent for secondary and primary schools), whose provider pulled out. This would an area cost adjustment dependent on the allow for a far more dynamic and flex- cost of hiring staff in different areas, and, ible process of accountability than is cur- if applicable, a ‘pupil premium’ – addi- rently possible via brief Ofsted inspections. tional funding for pupils coming from Ofsted would only have to inspect the deprived communities. This would not authorisers not their individual schools. only represent a fairer and more consistent Authorisers would be funded through a way of allocating funding, and especially combination of a dispersal of DCSF and funding for disadvantaged students, but it Ofsted funds and, if necessary, a small levy would also actively skew a market towards (no more than 0.5%) on their schools. the most deprived areas of the country, thus mitigating the potential for increased 3. A national funding formula should be choice to lead to increased segregation. introduced. An unintended outcome of the expansion 4. Once a network of authorisers is estab- of the academies programme has been lished they should be allowed to start the development of a twin-track fund- commissioning new schools. ing system whereby academies are funded Once a network of authorisers and a centrally but all other schools are funded national funding formula were set-up, the by their local authorities (admittedly, the right balance of accountability and incen- extent to which local authorities have the tives would be in place to allow authoris- power to specify individual school’s fund- ers to start commissioning new genuinely ing is somewhat constrained). This is not demand-led ISFS to compete with existing sustainable as the academies programme schools. Providers, who could be exist- continues to expand. It has already led to ing charitable or for-profit federations, the perception of disparity and unfairness parent or teacher co-operatives, or com- which is, to some extent, justified. The munity charities, would be able to make twin-track system means that authorities their case to an authoriser of their choice. can ‘hold back’ money from ‘their’ schools Authorisers would be able to approve an but not from academies, meaning that the unlimited number of schools from provid- latter are better funded. It also means that ers whose educational model had already money allocated by the DCSF for ‘disad- proved successful. For entirely new pro- vantage’ is not passed on to schools in any viders an annual cap would be in place, coherent way. Furthermore, as the market perhaps five or ten for each authoriser, develops, unwarranted disparities between to regulate supply. This would encourage authorities based on out-of-date calcula- the demand-driven growth of successful tions could skew providers towards certain federations while not excluding those with areas of the country, as has happened in new ideas. It would also give authorisers Sweden and the US.265 the chance to intensively ‘incubate’ schools A clear and transparent national fund- run by less experienced new providers. All ing formula should be introduced as rec- new schools should have to abide by the ommended in our report School Funding existing admissions code while retaining and Social Justice published in 2008. All their own admissions policy. schools would receive per-pupil funding A serious issue, which has caused prob- 265 Freedman S and Horner S, School Funding and Social direct from the government and local lems in Sweden and the US, is how to Justice, Policy Exchange, 2008
www.policyexchange.org.uk • 61 A guide to school choice reforms
support the expansion of the school estate federations with up-front capital or leasing which demand-led choice reform requires. costs while providing temporary accommo- Because it is focused solely on replac- dation for new providers in custom-built ing existing schools, the academies pro- ‘incubators’. A number of cheap, small gramme only requires that existing build- buildings could be developed in each local ings are handed over or, more typically, authority as part of the Building Schools rebuilt. Setting up a new school means for the Future programme that could be either funding a new development or, if leant out to providers to try new concepts planning regulations are relaxed, leasing so as to avoid wasting capital. If they were buildings designed for a different purpose successful they could then be given access to (e.g. office space, community centres or capital to develop a new school; if they were warehouses). This extra cost is the main not successful the space would be leant to a reason why charter schools in the US are so new provider. much worse off than public schools. The end goal of the reforms would There are three options for resolving be a system that is built on the this problem (a fourth would be to leave it following principles: unresolved, as in the US, and expect ISFS to make efficiency savings to cover the cost Allowing for demand-led expansion of of leasing). The first is for the government successful groups, emphasising federa- to give a large chunk of up-front capital tion and educational experience; funding to each new school. This is less The promotion of innovative new than ideal for three reasons: it would be ideas; extremely expensive, it would arbitrarily A commissioning system that is open limit the number of new schools that could to any provider, large or small, for- be set-up and untested providers could profit or charitable; waste the money on expensive experimen- Careful quality controls to avoid mul- tal designs that don’t work. The second tiple market failures; is to do what happens in most Swedish A combination of autonomy and municipalities and provide annual leas- strong accountability for providers; ing costs. This is preferable as it does not Incentives for providers to set-up in require massive up-front expenditure and communities where achievement is is less risky. Nevertheless it would mean an currently lowest. additional annual cost to the system which would have to come from elsewhere in the A school system based on these principles education budget (or increased spending). would combine the best aspects of choice- It is difficult to estimate exactly what this based reforms in Sweden, the US and this cost would be without knowing how many country to date. These reforms, which have schools would set up (or how many existing already boosted achievement in all three schools would close down as the result of countries, have the potential to transform competition, freeing up land that could be the lives of millions of children. sold or offered to new ISFS). The third pos- sibility could involve supporting successful
62 Appendix A Swedish School Numbers
Förskola 2003 2004 2005 2006 2007 Number of schools State 6616 (75.3%) 6576 (75.2%) 6769 (75%) 7076 (75.5) 7324 (75.4%) Free 2175 (24.7%) 2170 (24.8%) 2252 (25%) 2301 (24.5%) 2392 (24.6%) Total 8791 8746 9021 9377 9716
Number of students State 293075 303107 315481 329371 344790 (83.3%) (83.3%) (83.3%) (83.1%) (82.7%) Free 58647 60938 63473 66860 72151 (16.7%) (16.7%) (16.7%) (16.9%) (17.3%) Total 351722 364045 378954 396231 416941
Grundskola 2003 2004 2005 2006 2007 Number of schools State 4571 (89.5%) 4476 (88.8%) 4387 (88.4%) 4312 (87.9) 4262 (87.5%) Free 538 (10.5%) 565 (11.2%) 576 (11.6%) 596 (12.1%) 610 (12.5%) Total 5109 5041 4963 4908 4872
Number of students State 997180 979387 952273 919312 881637 (89.5%) (93.6%) (93%) (92.3%) (91.6%) Free 60045 (5.7%) 67054 (6.4%) 71451 (7%) 76145 (7.7%) 80712 (8.4%) Total 1057225 1046441 1023724 995457 962349
Gymnasium 2003 2004 2005 2006 2007 Number of schools State 517 (72.1%) 515 (68.1%) 516 (67.6%) 523 (65.8) 518 (63.3%) Free 200 (27.9%) 241 (31.9%) 247 (32.37%) 272 (34.2%) 300 (3.7%) Total 717 756 763 795 818
Number of students State 295320 299424 306183 311124 319540 (91.8%) (89.7%) (88.1%) (86.6%) (85%) Free 26323 (8.2%) 34504 (10.3%) 41530 (11.9%) 48291 (13.4%) 56547 (15%) Total 321643 333928 347713 359415 376087
www.policyexchange.org.uk • 63 A guide to school choice reforms
Appendix B Opinion data from the Swedish Conference of Business
PARENTS (984 participants)
Free schools Municipal Schools Positive Negative Positive Negative
My children have involved teachers 88% 1% 65% 6%
The school helps make students take 90% 1% 62% 7% responsibility for their studies
The school is good at giving each student the 78% 4% 40% 17% support that he/she needs
The school helps strengthen the students’ self 81% 3% 41% 16% esteem and confidence
I know how the school deals with bullying, 73% 10% 58% 17% should it occur
How pleased or displeased are you with the 79% 3% 47% 14% parenting role of the school?
The school listens to me as a parent 87% 2% 63% 11%
The school is good at utilizing co-operation 75% 5% 37% 20% with parents
As a parent the school encourages me to get 84% 2% 51% 14% involved in my child’s education
As a parent I continuously receive information 88% 1% 72% 9% on how the student develops
I am pleased with how I, as a parent, am 93% 1% 79% 4% treated during contact with the school
As a parent I can influence how the rules and 47% 12% 19% 32% norms in school are followed
How pleased or displeased are you with 75% 4% 41% 13% parents’ influence in the school?
The school makes sure students get the 87% 1% 58% 12% knowledge they need for further studies
The school gives the best prerequisites for 77% 4% 38% 19% learning
The school is good at encouraging 46% 6% 13% 22% entrepreneurialism
64 Appendix B
The school enables the students to work from 76% 5% 41% 15% their own curiosity and interests
The school is good at taking care of skilled and 67% 3% 24% 16% talented students
The school is good at incorporating students in 48% 9% 28% 20% need of special support
The school does a good job on following up on 80% 3% 56% 11% the results of the students
The teachers are competent and 86% 0% 61% 5% knowledgeable
The school has good equipment and means for 68% 6% 35% 22% learning
How pleased or displeased are you overall with 85% 2% 50% 14% learning in the school
I have confidence in the school’s teachers 88% 1% 70% 6%
I have confidence in the management of the 86% 4% 54% 13% school
I am pleased with the school’s demands on the 86% 3% 47% 22% students with regard to order
I clearly see that the school works for increased respect and sense of community between 85% 3% 46% 19% students and adults
What overall grade would you give your child’s 91% 2% 63% 7% school?
STUDENTS (1008 participants)
Free schools Municipal Schools Positive Negative Positive Negative
The teachers are interested in the views and 81% 4% 65% 9% opinions of the student
The teachers listen and react to the students’ 79% 5% 67% 7% opinions
The teachers co-operate in teaching so that 66% 11% 60% 13% separate subjects create a whole
The teachers are involved in their task 86% 2% 80% 3%
I get judged fairly in grades and reports 82% 5% 75% 7%
www.policyexchange.org.uk • 65 A guide to school choice reforms
Our teachers immediately act on bullying or rule 72% 5% 58% 10% violations
The teachers seem to enjoy themselves in my 91% 2% 83% 3% school
Overall, how would you grade teachers in your 82% 3% 74% 5% school?
There are good common rooms to be in 55% 20% 32% 41% outside of education
Students can influence how the school 65% 9% 48% 20% environment can be improved
I can talk to teachers in school if I’m sad, 84% 5% 77% 6% concerned or have other issues or problems
Students can influence how we work in school 62% 10% 50% 13%
My school is orderly 68% 7% 55% 12%
We are good at being on time for class 53% 21% 45% 26%
I have peace and quiet to study in school 74% 9% 59% 13%
Overall, how would you grade the learning 79% 4% 65% 5% environment in your school?
In my school students take significant 76% 6% 65% 7% responsibility for their learning and results
Education takes place in suitable size groups 80% 3% 63% 10% for the lesson aims
I know where I am and what I need to do to 79% 4% 79% 4% reach my goals
I get the help I need to reach my goals 84% 3% 81% 6%
Through school I get in touch with the outside world, for example business, universities and 51% 16% 46% 23% colleges.
The school is good at encouraging 67% 9% 56% 12% entrepreneurialism
I would choose this school again if I had the 74% 15% 78% 10% choice over again
I would choose the same program or major 70% 17% 76% 11%
Overall, what grade would you give the 82% 3% 83% 4% education that you partake in, in your school?
66 Appendix B
STAFF (712 participants) Free schools Municipal Schools Positive Negative Positive Negative
How do you see your chances of...
receiving more interesting tasks 83% 13% 75% 19%
expanding your professional experience 79% 21% 66% 32%
learning new things 79% 19% 66% 33%
increasing your competence 57% 42% 47% 52%
enjoying your work 72% 28% 59% 38%
feeling positive about your future 45% 53% 365% 61%
feeling pride 95% 4% 87% 11%
planning your own work 91% 9% 83% 17%
working at a controlled pace 85% 15% 89% 10%
taking your own initiative 76% 23% 72% 27%
testing new ideas 84% 16% 88% 12%
experiencing camaraderie 74% 25% 69% 29%
feeling appreciated 79% 20% 73% 26%
www.policyexchange.org.uk • 67 A guide to school choice reforms
Appendix C Number of charters run by EMOs and as a percentage of charter schools in each state (CMOs not included)266
State Charters Operating Charters run by EMOs Percentage For-Profit
Michigan 249 144 57.83 Wyoming 3 1 33.33 Ohio 299 77 25.75 Missouri 39 10 25.64 Virginia 4 1 25.00 Arizona 506 82 16.21 Nevada 26 3 11.54 Florida 384 43 11.20 Iowa 10 1 10.00 Georgia 82 8 9.76 New York 118 11 9.32 Colorado 147 13 8.84 Pennsylvania 133 11 8.27 Indiana 50 4 8.00 Illinois 74 5 6.76 Idaho 32 2 6.25 North Carolina 104 6 5.77 D.C. 89 5 5.62 Arkansas 25 1 4.00 Texas 333 13 3.90 Massachusetts 64 2 3.13 Louisiana 66 2 3.03 Maryland 34 1 2.94 Minnesota 158 4 2.53 California 763 17 2.23 Wisconsin 254 4 1.57 Oregon 93 1 1.08 Alaska 26 0 0.00 Connecticut 21 0 0.00 Delaware 21 0 0.00 Hawaii 32 0 0.00 Kansas 39 0 0.00 Mississippi 1 0 0.00 New Hampshire 11 0 0.00 New Jersey 64 0 0.00 New Mexico 70 0 0.00 Oklahoma 15 0 0.00 Rhode Island 11 0 0.00 South Carolina 36 0 0.00 Tennessee 14 0 0.00 Utah 68 0 0.00
266 Data from: Data from: Miron TOTAL 4568 472 G et al., 2008. Op. Cit.
68 School choice reform will be a key issue at the next General Election, yet the debate so far has focused on the theoretical arguments for and against creating a ‘schools market’ by bringing more independent providers into the state system. The purpose of this report is to learn the lessons of existing school reforms in England (the academies programme), Sweden (free A guide to school choice reforms reforms choice school to guide A schools) and the US (charter schools). We assess the success of reforms in all three countries against seven criteria which we believe a schools market should meet in order to find the right balance between promoting innovation and choice while maintaining accountability and quality control. A guide to school
None of the countries studied have achieved this balance yet, though in each case the introduction of new providers to the choice reforms system has brought benefits, but all of the seven criteria are
met by at least one country. By combining the best aspects of Daisy Meyland-Smith and Natalie Evans Natalie and Meyland-Smith Daisy each system, we argue, it is possible to develop a set of school Xxxxxx xxxxxxx xxxxxxx xxxxxxx choice reforms that will increase diversity and performance while protecting against market failure.
Daisy Meyland-Smith and Natalie Evans
£10.00 ISBN: 978-1-906097-42-4 Policy Exchange Policy
Policy Exchange Clutha House 10 Storey’s Gate London SW1P 3AY
www.policyexchange.org.uk
PX - A GUIDE TO SHCOOL - COVER 02-09.indd 1 20/2/09 10:46:35