School Funding and Social Justice

Home , Ellis Guilford School, Policy Exchange, Putteridge High School

SCHOOL COVER HDS.qxp 16/10/08 11:44 Page 1 colFnigadSca Justice Social and Funding School At the moment revenue funding for English schools is a mess; a labyrinth built out of failed policy grants, out-of-date demographics and unnecessary “stability” measures. There is absolutely no need for the system to be this complicated. In this report we have identified three ways to make the system simpler. First we should move to a consistent national funding formula, rather than having a different formula for each local authority. Secondly, arbitrary central government grants should be merged into this formula. Thirdly, the formula should have just three elements for each pupil: an age-weighted base, an area cost adjustment and a premium for those pupils from more School Funding disadvantaged backgrounds.

This last element is crucial. At the moment schools serving and Social Justice deprived communities do, typically, have more money than average. But this money is allocated in an entirely haphazard way, differently for each local authority and often through short-term A Guide to the Pupil Premium grants awarded by politicians eager for positive headlines. Instituting a consistent “pupil premium” would allow schools to plan their budgets around their admissions. Over time schools with large disadvantaged populations would be able to fund programmes to boost attainment, attract the best teachers and eventually middle class parents. Schools in better off areas would have an incentive to spread their nets wider, make their Horner Simon and Freedman Sam admissions more comprehensive and boost their budgets. New school providers entering the market, whether through the academies programme, or future Conservative / Liberal Democrat plans for supply-side reform, would have an incentive to set up in deprived communities, where real choice is needed most.

In this report we have tried to show how a premium could be delivered. We have modelled how this premium would work using real budget data from all English schools and have shown that it would give a significant boost to struggling schools. Sam Freedman and Simon Horner Using our figures – which would see a premium of between £500 and £3000 per pupil depending on the level of deprivation – the Foreword by Julian Le Grand total cost would be £4.6 billion, which could be paid for out of the existing education budget.

£10.00 ISBN: 978-1-906097-35-6

Policy Exchange Clutha House Exchange Policy 10 Storey’s Gate London SW1P 3AY www.policyexchange.org.uk SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 1

School Funding and Social Justice

A Guide to the Pupil Premium

Sam Freedman and Simon Horner

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 Fraser, Andrew Sells. SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 2

About the authors

Sam Freedman Simon Horner Sam is the Head of the Education Unit at Simon was a Research Fellow at Policy Policy Exchange. He achieved a first class Exchange from 2006 to 2008. He studied degree in History from Magdalen College, Philosophy at Kings College London. Oxford. After completing a Masters degree in After graduating in 2005, he became a International History in 2004, Sam joined researcher in the office of the Leader of the the Independent Schools Council as a Opposition for David Cameron MP, before researcher. He left three years later as Head of completing a Masters in Social Policy Research, having also completed a second Research at the London School of Masters degree in Public Policy and Economics. He is now advising the Shadow Management at Birkbeck. Sam joined Policy Treasury Team. Exchange in September 2007 and has written reports on school reform and teacher supply.

Published by Policy Exchange, Clutha House, 10 Storey’s Gate, London SW1P 3AY www.policyexchange.org.uk

ISBN: 978-1-906097-35-6

Printed by Heron, Dawson and Sawyer Designed by Soapbox, www.soapboxcommunications.co.uk

2 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 3

Contents

Acknowledgements 4 Foreword by Professor Julian Le Grand 5 Executive Summary 6 Introduction 8

1 The Theory 10 What is the pupil premium? 10 Where does the idea come from? 11 Doesn’t the current funding system already favour schools in disadvantaged communities? 12 2 Delivery 17 How would a national formula work? 17 How would the pupil premium aspect of the formula work? 19 Further questions and variations 29 3 Cost 33 How much would the pupil premium cost and how could it be paid for? 33 Will the extra money add value? 37 Will it act as an incentive? 41

Conclusion 43

www.policyexchange.org.uk • 3 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 4

Acknowledgements

The authors would like to thank Theodore Agnew for his financial support for this project. They would also like to thank Professor Julian Le Grand for his advice and guidance; Professor Richard Webber for providing MOSAIC statistics; and Davinder Kainth for technical support in building our database of schools. Finally we would like to thank the fifteen headteachers we interviewed for their time and candour.

4 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 5

Foreword By Julian Le Grand

In 1989 I jointly edited a book called policy agenda. As such it needed to be Market Socialism. It was, I believe, one of developed from simply a bright idea to a the first from the left to distinguish clearly concrete policy proposal with the core between the ends and means of attaining principles properly examined and the the progressive agenda. It pointed out that administrative detail fleshed out. This, I attaining progressive ends such as the pro- am delighted to say, is exactly what Sam motion of social justice, the development Freedman and Policy Exchange have done of economic welfare and the preservation in this excellent report. They have shown of individual liberties did not necessarily the complexity of the existing system of require the traditional socialist apparatus school funding (one that probably needs of large state bureaucracies, central plan- some reform in any case) and have illus- ning and top-down command and control. trated how difficult this can make head- Rather, if the conditions were right, the teachers’ decision-making over resources. same ends could be met by methods and They have demonstrated how a pupil pre- mechanisms not normally associated with mium could work, using actual numbers the left, such as user choice, independent and schools, both to simplify funding and providers and quasi-markets. And the book to provide the extra resources for the chil- put forward several ideas showing how this dren of the less well-off that the schools might be done. require if they are properly to meet those One of these ideas was what I rather children’s needs. clumsily called the positively discriminat- They have talked to headteachers to find ing voucher. This was aimed at the prob- out how the extra resources would be spent lem that has haunted governments of all and the likely impact on their behaviour. political persuasions: that of schooling for And not only have they done all that, but the poor and less well off. The idea was they have fully costed the proposals and simple: that parents should choose their shown where the money might come from. schools, that (public) money should follow Overall, this is a study of which any policy the choice, and that part of that money analyst would be proud. should include a premium for children The pupil premium is already supported from poor families or areas. This would by two political parties. It has yet to be give the schools both an incentive to take endorsed by the Government, but I live in such children and the extra resources to hope that this too will come. For, as Sam help with their education. Both social jus- Freedman has demonstrated so effectively tice and educational efficacy would be in this work, it is fundamentally a progres- served. sive idea, one that promotes both social This idea lay fallow for a few years, but justice and social efficiency - an idea that now – successfully retitled the pupil premi- can appeal to all progressives, regardless of um – it has moved on to the progressive nominal political affiliation.

www.policyexchange.org.uk • 5 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 6

Executive Summary

The Theory works against long-term planning and Segregation between well-off and means that funding can never be used deprived communities remains a huge to incentivise behaviour. problem in the English education system. A “pupil premium” would see extra money attached to students from Delivery deprived backgrounds. Schools that For the pupil premium to work it has take large numbers of such students to be allocated through a consistently would be better off, giving them some applied national funding formula. Each additional resources to educate chil- school should receive, for each pupil, a dren from more deprived backgrounds. single base sum (different for primary If these resources were used successful- and secondary), an area cost adjust- ly to boost attainment middle-class ment dependent on the cost of hiring families would start to be attracted to staff in different areas and, if applica- the school, reducing segregation. ble, the pupil premium. The money Addition ally, schools in wealthier areas should be paid straight to schools, might be incentivised to broaden their bypassing local authorities. admissions criteria to attract higher value pupils. Currently local authorities “hold back” up to 20% of the money designated for Both the Conservatives and the Liberal schools by central government. This Democrats have announced their sup- would no longer be possible if the money port for a pupil premium in principle. went straight to schools. Additional In theory it would help to alleviate the money would have to be allocated to risk that their school choice reforms local authorities to cover their expenses – would increase segregation as new but this should be standardised so only providers encouraged to enter the sys- services that have to be provided central- tem would have an incentive to target ly (like transport) are funded. This more deprived areas. should free extra money for schools.

The current government has, so far, not One advantage of the existing system is taken up the idea. This may be because that local factors can be taken into they are already diverting money to account when local authorities reallo- schools in deprived areas. However, this cate their schools’ funding. It is essen- money is allocated in a labyrinthine tial that the mechanism for delivering way. There is no consistency between the pupil premium is nuanced enough schools as each local authority has a dif- to take account of local differences in ferent formula; there is only a weak the context of a national formula. This relationship between changes in pupil rules out using the proportion of pupils population and funding due to unnec- receiving Free School Meals (FSM) as essary “stability” mechanisms imposed an indicator. This has typically been by central government; and the system used as a proxy for deprivation, but is of allocation is so complex that head- too crude a measure as it is based on teachers do not understand the reasons just one income-based variable. We rec- behind budgetary changes – which ommend using a “geodemographic”

6 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 7

Executive summary

classification like ACORN or MOSA- through additional central government IC which analyse individuals postcodes grants: the School Standards Grant, the using 400 variables derived from the School Development Grant and the census and other sources but are rela- Standards Funds. These three together tively simple to understand. The would cover up to £4 billion. The MOSAIC classification is a better pre- remainder of the money could be found dictor of student performance than by scrapping the £550 million other proxies like FSM. Education Maintenance Allowance – a means-tested weekly payment of up to We have developed a model of a pupil £30 for students in post-16 education. premium using MOSAIC, by grouping This has had a minimal impact on par- its 61 “classifications” into six types based ticipation and attainment and will, in on average GCSE performance. In our any case, become defunct once the edu- model the three lowest performing types cation leaving age is raised to 18. would have extra pupil premium money Alternative sources of funding identified attached to them. £500 for the third from within the DCSF budget are the worst, £2000 for the second and £3000 National Challenge programme and the for the third. Under this model 46% of ContactPoint database, both of which students would receive some premium (a have been widely criticised from across close match with the number that do not the political spectrum, and together will get 5 good GCSEs) – as opposed to just cost more than £200 million annually 14% under a FSM model. School-by- over the next four years. school analysis of the impact of this premium on funding indicates a much Interviews with headteachers indicate stronger correlation with actual perform- that the extra pupil premium money ance than a premium based on FSM or would typically be spent on hiring qual- existing allocations. ified teachers – by offering higher salaries if necessary. Another priority would be pastoral care, variously defined Cost and Value as personal tutoring, trips, extra-curricu- The pupil premium would cost an lar activities and counselling. additional £4.6 billion on top of the existing “Dedicated Schools Grant” While academic research suggests that (which provides most of schools annu- spending in these areas should boost al revenue). Given the significant attainment there is still a concern that increase in the Department for increasing the amount of money spent Children, Schools and Families by schools – rather than on local and (DCSF) budget over the past ten years central government priorities – could and the current economic climate we mean less research-led spending, and a think there is a strong argument that greater risk that schools could waste the premium should be funded money on well-sold but unproven tech- through existing expenditure on educa- nologies or programmes. We recom- tion rather than through tax rises or mend that the government should set- other departmental budgets. up an independent body to undertake evidence-based research into educa- A significant amount of this extra money tional interventions to strict interna- could be covered by reassigning money tional standards and disseminate these given to schools, and local authorities, findings to schools.

www.policyexchange.org.uk • 7 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 8

Introduction

At the next election, education will, as attached to pupils from deprived back- ever, be a major battleground. The Liberal grounds, so that schools which take large Democrats and the Conservatives have numbers of such children would be prop- already indicated that they will fight on a erly resourced. An additional benefit platform of reforms that would see the would be that it might incentivise existing principles of the Academies programme schools with predominantly wealthy (non-state providers running independent intakes to broaden their admissions policy, state-funded schools) expanded across the and would certainly give their neighbour- whole of the country. Reform on this scale ing schools located in more disadvantaged requires re-thinking many aspects of the communities the resources to attract mid- school system. Funding is one aspect that dle-class parents. This is especially impor- will need to change dramatically. At the tant if non-state providers have freedom to moment school funding is a disorganised set up schools wherever they want; a pupil mess; the legacy of historical reviews, premium could work to attract them to numerous abandoned policy programmes more deprived communities. and ministerial pronouncements. Per capi- The idea seems simple. However, it ta funding is calculated on a national level, throws up a number of difficult questions. but then re-calculated by every local How, for example, would an individual authority according to a different formula. pupil’s circumstances be quantified? Various grants are unsystematically tacked Means-testing would be far too bureau- onto the main revenue funding. There is a cratic, costly and invasive, but existing need for reform in any case, but it is even measures of school-age disadvantage, like more essential if non-state providers are free school meals, are quite crude. How increasingly going to take over from local much would a “premium” be worth and authorities in running schools. Financial how would it relate to existing deprivation consistency is vital to develop a viable mar- funding? How could this be paid for with- ket; this is why Academies have to be fund- out taking so much money from schools in ed separately by the Department of leafy suburbs that they would be forced to Children, Schools and Families (DCSF). lay off staff? Will schools spend their extra money in ways that will actually boost attainment? “ At the moment school funding is a disorganised mess; In this report we lay out the principles the legacy of historical reviews, numerous abandoned of a new, simple, and consistent national school funding model based on per capita policy programmes and ministerial pronouncements ” payments incorporating a pupil premium. We have designed a mechanism for calcu- lating the premium, based on postcodes, Overhauling the funding system pro- that takes into account a wide range of fac- vides a real opportunity to offer consistent tors without increasing complexity, and financial support to schools that take a lot have calculated realistic costs that would of children from disadvantaged communi- not require tax increases, or raiding other ties. The “pupil premium”, endorsed by departmental budgets. Our financial the Liberal Democrats, Conservatives and breakdowns are based on “Section 52” data a range of think-tanks from all parts of the that all local authorities have to provide to political spectrum, would see more money central government. This gives budgets for

8 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 9

Introduction

every school in the country allowing for a schools in London, but they do offer an school-by-school analysis of the impact of indication of how changes to funding any changes. Alongside the statistical mechanisms would be received on the analysis we include quotes from interviews ground. The report is split into three chap- with fifteen headteachers looking at the ters; the first on the theory of the pupil problems they have with the existing sys- premium, the second on its delivery and tem and how they would spend additional the third on its cost. Each chapter is subdi- “pupil premium” money. We do not pre- vided into a series of questions that, taken tend that these interviews offer a represen- together, show how this exceptionally tative sample, they were small in number important reform could be made to work and heavily skewed towards secondary in practice.

www.policyexchange.org.uk • 9 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 10

The Theory

What is the pupil premium? way of identifying the number of poorer At the moment schools are funded on a children in a school) compared to 14.3% broadly per capita basis (with significant nationally.1 Of the 638 schools recently variations). However, some children are identified by the Government as “failing” harder to teach than others – and they because fewer than 30% of their pupils come, disproportionately, from disadvan- achieve 5 A*-C GCSEs including English taged backgrounds. Schools, therefore, and Maths, 85% (542) have above average have an incentive to try to avoid taking numbers of pupils on FSM.2 Academics these children because they will adversely from Bristol University have shown that in affect their performance in examinations a typical local authority in England a child without bringing in any extra funding. from a poor family is half as likely to Schools in wealthier areas are helped in attend a good secondary school as a non- this by the current admissions system poor child.3 which in most cases sees places allocated In British politics this segregation has on the basis of distance from the school. been a preoccupation of the left. Their House prices are higher near successful concern has been entirely justified. schools which means that the children Unfortunately the solutions they have sug- accepted to the school tend to be less chal- gested have typically focussed on directly lenging to teach – which reinforces the controlling the balance of admissions by success of the school. There is absolutely bussing or lottery. As the furore caused by no incentive to widen admissions proce- the recent introduction of an admissions dures to take in those from poorer areas lottery in Brighton has shown this kind of further afield. top-down interference from government is This means that those from poorer both politically and psychologically areas are often segregated into their local unhelpful.4 Parents like to feel that they schools. Some of these are highly success- have some control over where their chil- ful against the odds – often because of dren go to school. Equally parental choice inspirational leadership. But where is important to improving the quality of schools do end up taking considerable schools; without the competition it can

1 Sutton Trust, The Social numbers from disadvantaged back- bring there is often no incentive to Composition of Top grounds their performance often suffers improve. Comprehensive Schools, January 2006. and this can be reinforced by the difficul- The pupil premium, however, in theo- 2 http://news.bbc.co.uk/1/hi/ ty of hiring good staff, who can receive ry, would help to balance admissions while education/7444822.stm the same pay working in a much more still allowing parents to choose the right 3 Simon Burgess and Adam Briggs, School Assignment, pleasant environment. school for their children. It would see School Choice and Social The effects of this segregation are pow- additional money attached, per capita, to Mobility, CMPO Working Paper Series No. 06/157, November erful. The Sutton Trust have shown that children from deprived backgrounds. That 2006. the top 200 comprehensive schools have way schools in wealthier areas would have 4 http://www.timesonline.co.uk/ an average of just 5.6% of children on an incentive to take them on as they tol/news/uk/education/article145 2218.ece “free school meals” (FSM – a common would be properly resourced for the more

10 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 11

The theory

difficult job of providing an education for Democrats have more radical plans that them. Over time, intakes would become would see new suppliers opening up more balanced. At the same time schools schools wherever they saw demand. Under which took a lot of children from disad- the current funding system new suppliers vantaged backgrounds would be a lot rich- might focus on wealthier areas, with easier er. They would be able to use the money, to teach pupils. A pupil premium could for example, to pay top rates to attract the help skew the market so that alternative best teachers or reduce class sizes to make provision is focussed on more disadvan- discipline problems more manageable. taged areas where it is needed most. Over time, if the school used the money well, results would improve and middle class families would be more inclined to Where does the idea come from? apply. Again this would work to balance Both the Conservatives and the Liberal intakes. As this happened the financial dif- Democrats have said that they will go into ferences between schools would start to the next election pledging to introduce a even out. pupil premium. In both cases the aim is to As Julian Le Grand, who first came up make sure that their plans to open up the with the idea in the 1980s, has put it: schools market to new suppliers does not lead to greater segregation. The hope is “Schools that contained a high propor- that the pupil premium will encourage tion of children from poor families these new suppliers to set up schools in would then have more resources per more deprived areas. pupil than those with a low proportion. There has been some dispute over who They would also have better premises had the idea first.6 Of course, as with any and equipment and could attract high- “new” idea in politics, the pupil premium er-quality staff. The outcome would has been around for a lot longer than its either be selective schools, with those current incarnation. It can be traced back that specialised in the education of the to at least 1970 when the US Centre for children of the poor being better the Study of Public Policy recommended equipped and staffed than those that the “Regulated Compensatory Voucher” in specialised in the education of children a review of voucher models.7 They con- of the rich, or, if head teachers or staff cluded that a voucher system could only did not want to engage in such special- work if the inequalities that would be cre- isations, schools that contained a rea- ated by a free market were rectified. So sonable proportion of children for all under their model schools would not be parts of the social spectrum. One way or able to charge top-up fees, would not be another cream-skimming that favoured able to select and poor children would 8 the better off would be reduced or elim- receive a supplementary voucher. 5 Julian Le Grand, The Other 5 Invisible Hand, Princeton inated.” Julian Le Grand introduced the idea to University Press, 2007, p. 148.

the UK in his edited book Market 6 http://timesonline.typepad. The pupil premium is particularly rele- Socialism, written by the Fabian Society’s com/politics/2008/01/george- osborne.html; vant to the current education debate Socialist Philosophy Group, convened fol- http://www.spectator.co.uk/coffe because all three main parties are advocat- lowing the 1983 election to provide intel- ehouse/441406/premium-politics.thtml ing opening up the school system to alter- lectual reinvigoration for the Labour party. 7 Centre for the Study of Public native providers. The Labour party contin- It was presented by Professor Le Grand Policy, Education Vouchers: A Preliminary Report on Financing ues to promote the Academies scheme (who termed it the “positively discriminat- Education to Parents, which sees failing schools handed over to ing voucher”) as the equitable “left-wing” (Cambridge, 1970). sponsors. The Conservatives and Liberal equivalent to the standard free market 8 Ibid. pp. 14-15.

www.policyexchange.org.uk • 11 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 12

School funding and social justice

voucher model proposed by the right at the level of resource before they start failing. It time.9 Le Grand has continued to write also fits better with plans to open up the about the idea in various publications over education market to new providers – as a the last twenty years. mechanism is required to incentivise new Oddly, though, neither the Conservatives schools to set up in disadvantaged commu- nor the Liberal Democrats seem to have nities. got the idea directly from Le Grand So, at the moment, both parties support although politicians on both sides of the very similar schemes, though neither has house were aware of the idea in the 1990s. gone into much detail over how they envis- The pupil premium was introduced to age the premium operating in practice. Liberal Democrat thinking by their 2001 Interestingly Professor Le Grand did rec- Commission for Educational Policy ommend the idea to ministers and advisors chaired by Professor John Howson who in the current Government during his time seems to have been inspired by the funding as a senior policy advisor to Tony Blair model used in the Netherlands; that sees (2003-2005). There have been no public different amounts attached to pupils statements, however, about the pupil pre- according to their backgrounds. Over the mium from the Labour party. last seven years the pupil premium has become embedded in Liberal Democrat policy and was endorsed by the new leader Doesn’t the current funding system Nick Clegg following a supportive analysis already favour schools in disadvantaged of the idea by the CentreForum think-tank communities? in 2007.10 One possible reason that ministers have The Conservatives took the idea from a ignored the pupil premium is that, from 2005 report by Policy Exchange entitled their point of view, additional money is “More Good Schools”. This talked of an already distributed to schools with high “advantage premium” of £5,000 for chil- numbers of pupils from disadvantaged dren at failing schools which they could backgrounds. This is true. A recent report then “take” to other schools (either state or from the Institute for Fiscal Studies suggests independent) who would thus be incen- that, on average, a pupil on FSM will on tivised to accept them.11 This was based on average be “worth” 70% more money than 9 Julian Le Grand and Saul 13 Estrin ed., Market Socialism, the Florida “A+” scheme which saw one who is not. Unfortunately, though, (Clarendon Press, 1989) pp. 198-204. schools receive annual grades based on this extra money does not operate as the

10 Paul Marshall, Tackling their performance in Florida’s Compre- pupil premium would for three reasons: Educational Inequality, hensive Assessment Test (FCAT). If a (CentreForum, 2007). school received two failing grades in any 1. First there is no consistency in the way 11 James O’Shaughnessy and Charlotte Leslie, More Good four year period, the pupils attending that that it is allocated between schools. Schools, (Policy Exchange, school would be eligible for a voucher 2. Secondly, the funding system does not 2005). which could be used at public or private respond quickly to changes in pupil 12 The Florida supreme court 12 declared the scheme unconstitu- school. population, so there is no incentive to tional in 2006. See Greg Forster, Lost Opportunity: an empirical Over the last few years the Conser - recruit pupils from disadvantaged analysis of how vouchers affect- vatives (and Policy Exchange) have moved areas. ed Florida public schools, (Friedman Foundation for away from the advantage premium idea 3. Finally, the system is so complex that Educational Choice, March and towards Le Grand’s original idea. few headteachers understand how their 2008). Using the pupil’s background to determine budget is allocated, making it hard to 13 Luke Sibieta, Haroon Chowdry and Alastair Muriel, the degree of financial report, rather than develop long-term plans, and meaning Level playing field? The implica- school performance, allows for a more uni- that it can have no impact on their tions of school funding, (CFBT/IFS, 2008), p. 7. versal scheme and gives schools a higher admissions decisions.

12 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 13

The theory

1. The first of these problems – the lack of Authorities do now have to make some consistency between schools – has two allowance for deprivation but there is no causes: the redistribution of nationally specified amount that has to be used in this determined revenue allocations by local way. The IFS study published in June 2008 authorities and the preponderance of cen- found that they do not allocate all of their tral government grants that make up a sig- AEN on the basis of deprivation – in fact nificant proportion of the extra funding it is “flattened” across all schools so just received by schools in deprived areas. 40%-50% of the extra funding that local Most of a school’s annual revenue fund- authorities receive for FSM eligible pupils ing is allocated according to a national for- ends up at the schools these pupils attend.17 mula but then reallocated by local authori- In some cases this is because authorities ties who can use their own mechanisms. All have their own priorities about which local authorities receive a Dedicated Schools schools should get funding – occasionally Grant (DSG) from the DCSF which they these are political – but it is not all the fault must spend on their schools (Academies are of authorities. Because of the measures separately funded by the DCSF). The DSG taken by central government to ensure is largely based on historical per pupil year-on-year stability (discussed below) amounts for each authority which are then their hands are often tied. increased by a certain percentage each year. The second reason for inconsistency is Much of the difference between the per that much of the extra money for poorer pupil amount for each authority is due to students comes not from annual revenue previous historical allocations for Additional funding via the DSG but through central Educational Needs (AEN). These were government grants which are often based on the number of children speaking designed to support specific policy out- English as a second language in the author- comes. The IFS study indicates that over ity, the number in certain ethnic groups that 40% of the extra 70% attached to pupils typically underperform and the number of with FSM comes direct from these grants.18 children on FSM. This deprivation funding This is allocated to schools through two represents anything from 5 to 22% of the routes: the School Standards Grant (SSG) full spending grant depending on the local and the Standards Fund. The SSG is a £1.2 authority.14 billion fund (in 2007-8) paid directly to Once they have received the DSG, local schools on the basis of pupil numbers. authorities have some discretion over how Given that the DSG is supposedly calculat- this funding is then allocated to schools – ed on a per capita basis there is no obvious and they all have their own formula. Until reason for this separate fund other than to 2006 there was no requirement for local prevent local authorities from reallocating authorities to use any of this money for the money or holding it back to cover cen- AEN even though the DSG15 was partly tral costs (more on how this “hold back” allocated on the basis of these factors. A works in the next chapter). 14 DCSF, Deprivation Guidance for CSAs and LAs, Treasury review in 2005 found that some The main Standards Fund grant is the http://www.teachernet.gov.uk/m authorities were “unaware that they receive School Development Grant which merges anagement/schoolfunding/ schoolfunding2008to11 funding specifically to meet the costs of a whole range of formerly separate grants 15 It was called the Formula deprivation” and that “local authorities’ (including Excellence in cities, Gifted and Spending Share at the time.

decisions on the balance of funding Talented, Specialist Schools and so on). It 16 DfES/HM Treasury, Child between schools are not leading to depriva- is based primarily on the historical Poverty: Fair Funding for Schools, (London: Treasury, tion funding being accurately or consis- amounts each school got from these earlier 2005), p. 4. tently targeted towards schools in grants – which were disproportionately 17 Sibieta, p. 7. deprived areas.”16 targeted at schools in deprived areas. There 18 Ibid., p. 43.

www.policyexchange.org.uk • 13 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:08 Page 14

School funding and social justice

is then an extra £125,000 added for the between taking in more children from dis- most deprived schools. In 2007-08 around advantaged backgrounds and increasing £1.8 billion was allocated via the School the amount of the pie available to your Development Grant (of which £150 mil- school. As with the DSG annual revenue lion went to local authorities for central grant, the complexity and lack of consis- use).19 tency works against the additional funding operating as an incentive.

There is certainly no simple correlation between taking 2. The stability mechanisms built into the “ Designated Schools Grant (DSG) also in more children from disadvantaged backgrounds and work against the ability of the funding increasing the amount of the pie available to your school ” system to incentivise schools to take more children from disadvantaged backgrounds. This is the second reason why Alongside the main School Develop ment the current system does not operate as a Grant there are a number of additional allo- pupil premium. For a start the DSG is cations made through the Standards Fund. now awarded in three year chunks and the For example, in 2008-2009 £185 million per pupil amount does not change within has been allocated for the “Ethnic Minority those three years. So if a school decided it Achievement” programme and a further wanted to change admissions procedures £300 million has been allocated to support so as to include more children from the “national strategies”.20 These grants are deprived areas they would not be reward- ring-fenced and must be used for their stat- ed for this for up to three years. Moreover, ed purpose in supporting central govern- the current DSG is based entirely on his- ment policies. In recent years several other torical per pupil amounts so there has politically-motivated grants have been been no recalculation of deprivation launched. In 2006 a “personalised learning” whatsoever for the years 2008-2011. The grant was added to the SSG worth £220 DCSF call this the “spend plus” approach million in 2006-7 and £365 million in and acknowledge that “the principal rea- 2007-8 and has subsequently been extended son” for using it is “stability”. They also to 2011. This is primarily distributed acknowledge that “that a continuation of according to the number of pupils with the “spend plus” methodology in the prior low attainment and FSM. In June longer term would make it increasingly 2008 another £400 million was assigned, difficult to explain the differences in over the next three years, to the 638 Dedicated Schools Grant between author- “National Challenge” schools (as we have ities on an objective basis” (i.e. would seen most of these have significant numbers make no sense) so they are reviewing the of pupils in FSM).21 DSG in time for the 2012-2015 alloca- Much of the money that comes through tion.22 This kind of glacial process means all of these allocations ends up at schools in that the relationship between the per

19 DCSF, Section 52 Budget the most deprived areas (though not the pupil amount for each authority and Summary 2007-8 main £1.2 billion annual SSG grant which actual deprivation can be extremely tenu- 20 http://www.teachernet.gov.uk is the only part of schools revenue funding ous. /docbank/index.cfm?id=12227 to be purely per capita). However, it is dif- Furthermore, once the money reaches 21 DCSF, National Challenge: A Toolkit for Local Authorities, ficult for schools to plan on the basis of local authorities, the “Minimum Funding (DCSF, 2008). these funds. They are often short-term and Guarantee” (MFG) kicks in, creating an 22 DCSF, Dedicated Schools Grant: Technical Note for 2008- vulnerable to changes in political priorities. even bigger gap between deprivation fund- 11 Allocation, p.2. There is certainly no simple correlation ing and actual deprivation. The MFG

14 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 15

The theory

means that all schools must receive a set headteachers do not really understand percentage increase in their per pupil fund- how their funding is calculated, they ing even if their local authority formula ignore the relationship between admis- indicates they should receive less money sions and income (except in extreme because they have fewer disadvantaged cases, such as Children with Statemented children. This is the main reason local Special Education Needs who bring with authorities have to “flatten” deprivation them a specific funding allocation). This funding across all of their schools. It also means that additional funding for chil- means schools actually have an incentive to dren from deprived areas does not act as reduce the number of disadvantaged chil- an incentive. dren – because they will continue to Of the fifteen headteachers interviewed receive per pupil funding based on their for this report the majority had a very past pupil population. limited understanding of why their school The three year budgets and the MFG received the money it did. Almost none were introduced after the 2003/2004 had any knowledge about how the origi- school funding “crisis” which saw schools nal DSG was worked out. There seemed claiming, in spring 2003, that they were to be greater awareness of their local not receiving enough funding for the authority formula, but even this was lim- 2004/5 school year, and would have to lay ited and often out-of-date. As the head of off staff. Charles Clarke, then Education a secondary community school in Secretary, blamed the local authorities for London admitted: “to be honest with holding back funding – so introduced a you, we’re more interested in the outcome number of stability measures including the of the formula than how the formula MFG.23 An Audit Commission investiga- actually works, so we wouldn’t spend a lot tion of the “crisis”, though, found that the of time looking at how the formula is problem was entirely unrelated to annual constructed, but clearly how the formula revenue funding, making the reforms is constructed makes a big difference to unnecessary. Instead the problem was with what we get in school”. the short-term central government stan- Others expressed similar weariness : dards funds: “the late announcement of changes to major specific grants (particu- “It would be nice if the minister said larly the Standards Fund)… [left] some that in the next year, the age weighted schools better off than expected and some pupil unit would be this, and this is schools with less funding than anticipat- what they are spending in the LA, and ed.”24 it would be nice if the local authority The current government have in effect wrote and said…the age weighted pupil plumped for stability over flexibility – unit for say a 12 year old will be X. We even though there is no evidence this is don’t get that, what we get is a budget necessary. The MFG means that funding statement, but it’s extremely complicat- can only respond to changes in the com- ed…There is no reason that this needs 25 position of pupil populations at a slow to be as complicated as it is.” 23 Sibieta, p.28.

rate. 24 The Audit Commission, “I find it more and more difficult to Education Funding: the impact and effectiveness of measures 3. The third reason why the current sys- understand …what happens…it does to stabilise school funding, tem does not act as one based on a pupil seem a fairly complicated and laborious (Audit Commission, 2004) p. 2. premium would is the extreme complexi- process, in terms of what happens 25 Headteacher, secondary school, London. ty of the funding model (amply demon- between central government, the council 26 Headteacher, primary school, 26 strated in the previous paragraphs). As and then the schools.” London.

www.policyexchange.org.uk • 15 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 16

School funding and social justice

Under these circumstances no incentive “We were for 5 years what was known effect is possible, because no one appreci- as a ‘beacon school’, now I thought the ates how it works. whole beacon school initiative, sharing The most common complaint with the excellence and expertise with other system itself was that the short-term cen- schools, and having the money to do it, tral government Standards Fund grants was fantastic, so for 5 years we were played havoc with planning. As one head funded to the tune of about £45,000 a told us: year, to work the other schools, which we did locally, nationally and interna- “What tends to happen…is that fund- tionally, but then after 5 years the bea- ing streams become confused…the con initiative went, and so did the streams which are independent and dis- £45,000…we had ongoing proj- creet get amalgamated or hidden with- ects…we’d staffed it, one has to restruc- in other funding streams. It’s quite dif- ture to do things like that.” 28 ficult to unpick that and work out what you’re being funded for and what you’re Yet another head complained of the

27 Headteacher, secondary not being funded for….That’s a really problems caused by recent cessation of the school, London. unhelpful process for us…that funding “Excellence in Cities” stream of the 28 Headteacher, primary school, is attached to jobs…and it’s attached to Standards Fund.29 This inability to plan London. a structure that works very well.” 27 again limits the potential for using the 29 Headteacher, secondary school, North-East. funding system to influence the long-term Another gave us the following example: development of schools.

16 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 17

Delivery

The problems of consistency and com- they would not have access to this money. 30 Section 52 Education Outturn plexity discussed in the last chapter have Secondly, since the supposed value of hav- statement for 2006-2007, Table to be resolved before a pupil premium can ing a separate authority formula is that A England summary (http://www.dfes.gov.uk/localaut operate in practice. This can only be local factors can be taken into account, any horities/section52) achieved by moving to a national funding national formula would have to be suffi- 31 Ibid. formula that bypasses local authorities. By ciently flexible to satisfy heads that their 32 Sibieta, p. 25. removing one of the stages from the exist- local circumstances were being recognised. 33 Schools Forums are local consultative bodies, whose role ing process it would make the system eas- Thirdly, there would need to be a transi- is to scrutinise, and potentially ier for heads to understand – especially if tion from the current formula to the new challenge, each authorities school funding proposals for the one remaining stage was also reduced one which would need to be carefully man- each year. They are a means to to just three elements; a base per capita aged to avoid unbalancing the entire sys- increase the involvement of schools themselves in authority sum, an area cost adjustment and a pupil tem. funding decisions and thereby provide a check on authorities. premium. Authorities are required to con- As local authorities would be prevented sult Schools Forums on the proposed school funding formula from reallocating money according to How would a national formula work? for each year, on particular their formula it would be clear what At the moment local authorities are fund- issues relating to the budget (such as provision for special schools would lose or gain by changing the ed via two sources. They receive a specific educational needs and free mix of pupils over time. The third prob- LA budget worth over £4 billion annual- school meals), and on service contracts agreed by the authori- 31 lem of inflexibility caused by stability ly. As we have seen they also hold back ty. Forums should have 15 mem- mechanisms in the current system could part of the schools budget (£3.65 billion) bers, and are composed of elected schools members (head- be resolved simply by folding the majority for central services. The amount held teachers and governors) repre- senting the schools within each of the Standards Fund money into the back varies from authority to authority authority, with an additional national formula. Without the planning and ranges widely from over 20% to less quota of non-schools members 32 appointed directly by the author- difficulties caused by dramatic shifts in than 5%. The only rule is that authori- ity (which can be no more than central government grants there would be ties cannot increase the amount they hold one-fifth of the total member- ship).Taken from Centre for no need for three-year budgets or the back as a proportion of the schools budg- Public Scrutiny website. Minimum Funding Guarantee. Funding et (unless it is explicitly supported (http://www.cfps.org.uk/scrutiny_ map/subcat.php?mainID=3&subI 33 would be genuinely reactive to the school through the Schools Forum ). This is D=21)

population but would actually be more called the Central Expenditure Limit. 34 Sibieta, p. 25. stable as there would be no dramatic Unfortunately this creates a considerable 35 For example, central administration, access, youth services, changes caused by the end of a political disincentive for authorities to reduce the strategic management and some initiative. amount they hold back as this shift would special educational needs services like education psychologists. There are three potential problems that be irreversible should they wish to 36 For example, admissions, might be caused by a simple, direct nation- increase spending on central services in school forums, pupil referral al formula. The first is that at the moment the future.34 units and other behavioural support, development grant pro- about £3.65 billion of the £32.72 billion Although, for accounting purposes, grammes and other special edu- spent on schools (both DSG and Standards there is a division between what the LA cational needs support. A full breakdown can be found in the 35 Funds) is held back by local authorities to budget is spent on and what the money Section 52 Education Budget 30 36 Statement - Table 1- LEA level provide central services. If they were held back from schools is spent on, it is information (http://www.dfes. entirely bypassed by a national formula difficult to draw clear dividing lines. gov.uk/localauthorities/section52)

www.policyexchange.org.uk • 17 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 18

School funding and social justice

Moreover, the heads we spoke to were ities’ are...not prepared to get rid of the unclear exactly how much of “their” inadequate people.” 40 money was being held back (the exception was the one Academy head we spoke to With one exception there was wide- who knew that he was receiving £375,000 spread agreement that, in principle, as a year that would otherwise have been much money should be delegated directly “held back”) and what exactly it was spent to schools as possible: on. Consistent with the considerable diver- “I’ve always favoured the strongest pos- gence between the “hold back” in different sible level of delegation...to schools and authorities headteachers had widely differ- schools then being held to account.” 41 ing experiences of local authority involvement. Some, in authorities where more is “I’ve convinced that the more that’s del- held back, were critical: egated to us, the better – the more services that we can commission ourselves, “there’s a notion within some of the the better.” 42 services around the local authority that theirs’ are the services that we should be “I’m a great believer in as much money buying, so I think a lot of those depart- being delegated to schools as possible, ments haven’t cottoned on to best- and schools can buy services according value, and the fact that...they should to best value.” 43 be assisting us...in looking for best value.” 37 It is important to note that our sample of heads was heavily biased towards sec- “I think Manchester holds back more ondary schools and that the exception to than other Education authori- this view came from one of the primary ties....Manchester does need to delegate heads we interviewed (“I wouldn’t want the much more funding to its schools, and extra bother of having to commission and if...schools want to buy back services, look for these things”).44 It is likely that we then they should be able to...” 38 would have had similar responses from other primary heads as they do not typical- Others, in authorities where there is ly have the support staff available to sec- more delegation, were more relaxed, ondary heads. If money passed straight to though there were widely divergent views schools and local authorities were unable over whether or not it was worthwhile to hold back funds to provide services

37 Headteacher, primary school, buying back into local authority services: schools would, of course, still be able to London. buy services from local authorities but it 38 Headteacher, secondary “Many of the services are good value for would be entirely on their terms. Primary school, Manchester. money...I could opt out, but...by and schools that did not want to spend time 39 Headteacher, secondary school, North-East. large I think we get a pretty good serv- evaluating services could choose an “all-in” 40 Headteacher, secondary ice and pretty good value for money.” 39 option. This could, potentially, cause prob- school. London. lems for smaller authorities in which there 41 Headteacher, secondary school, East Anglia. “I don’t have to buy [local authority was some demand for a particular service

42 Headteacher, secondary services] and I don’t....they are poor (say payroll management) but not enough school, London. services, and we do our own thing, to make it economically sustainable for the 43 Headteacher, secondary finances, cleaning, maintenance...And I authority to provide the service. There is school, Manchester.

44 Headteacher, primary school, don’t buy in their SEN because it’s use- no reason, though, why in this type of sit- London. less, I buy good consultants, the author- uation smaller authorities could not band

18 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 19

Delivery

together to provide this service to schools authorities do not allocate the deprivation which did not want to opt-out. funding they do get to the schools that actu- Not all the money currently held back, ally need it – partly because their own for- though, could simply be diverted to mulas are not optimal and partly because schools. Much of the funds held back are the Minimum Funding Guarantee forces used to provide services that have to be their hand. provided centrally, such as admissions, transport, home education, pupil referral units and private school fees for children There was scepticism that any formula could accurately with Special Educational Needs. Some of “ pick up small pockets of deprivation or distinguish this is already paid for by the local author- accurately between different kinds of deprivation ity budget but this would certainly have ” to increase if LAs had no access to the DSG budget. It is difficult to estimate an exact figure that would be required but it The other localised factor built into the would not have to increase by the full current DSG, the “Area Cost Adjustment” £3.65 billion currently taken from the (ACA), is as meaningless as the deprivation schools budget. The inconsistency of dif- funding. For a start it is also historical. But ferent levels of hold back between differ- even before it was out-of-date it made little ent authorities could, however, be elimi- sense because it was based on the average nated if school and authority funding was salaries of workers in the authority and not properly separated into two centrally pro- of teachers. This means that there are vided pots. An additional bonus to replac- much larger differences than are justified ing the schools budget “hold back” with a given that we have a national pay agreement higher LA budget would be that authori- for teachers. This is the main reason why ties were no longer actively incentivised to MPs from certain parts of the country are spend more money than was necessary to continually complaining to education hold on to their share (as they are at pres- ministers that their constituency appears to ent through the Central Expenditure have got a bad deal.45 Limit). Any future national funding formula would have to have an ACA, but its role would depend on how teacher salaries How would the pupil premium aspect (which represent about 65% of a school’s of the formula work? budget) are calculated. If they continue to The second delivery problem is making a be calculated through a national pay agree- national formula sufficiently flexible to take ment then the ACA need only be a small local circumstances into account. The the- factor (it would still be necessary because oretical advantage of having a second stage, of the wage differences for non-teaching where funding is re-allocated within the staff). The only exception to this would be local authority, as under the present system, London weighting for teacher salaries

is that local factors which no formula could (which is already accounted for separately). 45 For more detail see the discus- pick up would be acknowledged. If schools were allowed to opt-out of the sion on the ACA between members of the DSG Formula Review 46 It is important to note, in response to national pay agreement then using the Group from the 14th July 2008 this, that in fact the existing system is not average salaries of all workers as a marker (http://www.teachernet.gov.uk/m anagement/schoolfunding/DSG that responsive in relation to local factors. would make much more sense. formulareview/DSGreviewpapers) As we have seen the national DSG formula So it can hardly be said that the existing 46 As recommended in the Policy Exchange report More is purely historical so does not relate to system is adequately nuanced with regards Good Teachers published earlier existing deprivation. Moreover, local to local factors. Nonetheless, it is true that this year.

www.policyexchange.org.uk • 19 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 20

School funding and social justice

local authorities do have the power to resource there because you tend to have make decisions based on specific circum- quite a high level of special needs, things stances. For example, an authority where like ADHD, aggression, so as long as most of the educational problems are relat- you can have a formula which is suffi- ed to a specific ethnic minority could focus ciently subtle, then fine. But you might their formula on extra money for that go to another part of London where group, as opposed to one where economic deprivation appears high, and maybe deprivation was more of an issue. you’ve got high density, poor health, but While most of the heads we spoke to they’re all Hindu families, and they’re were in favour of having a national formu- high aspiration and will learn in any la that saw funds delegated directly to school, and if you fund those double schools there was some concern that a cen- then you’re kind of throwing your tral calculation might not be nuanced money away.” 48 enough to pick up on their specific needs. There was scepticism that any formula This is even more of a problem when could accurately pick up small pockets of one considers the need for any formula to deprivation or distinguish accurately be simple. Even if it were possible to between different kinds of deprivation: include so many variables that the formula was adequately nuanced for any eventuali- “The model [would have to be] sophis- ty, this would inevitably reintroduce the ticated enough to take into problem of complexity; headteachers account...small variations in need and would not be able to be follow its workings deprivation...In the kinds of intake that and it would not act as an incentive. we have here, whilst there’s a very There are a huge variety of deprivation deprived community surrounding the indicators currently included in the school you’ve also got a pocket of national and local authority formulas that absolute affluence.” 47 could feasibly be used to construct a pupil premium. At the moment the deprivation “I think it’s [a national funding formu- portion of the current DSG is based pure- la with a pupil premium] a great idea, ly on historical weightings, but it was the problem for me is how you assign, originally based on an analysis by the characteristics and the level of PriceWaterhouseCoopers (PWC) of the weight, because there are some commu- specific costs of certain additional educa- nities who aren’t obvious, one of the tional needs (AEN). They identified two most difficult ones is white working types of additional need: children for class, in one of our Academies that we’re whom English is an additional language about to open, it’s a very tough white (EAL) and social deprivation.49 The first working class area, with low socio-eco- could easily be identified through the nomic deprivation, but high cultural school census. For the second they recom- 47 Headteacher, secondary school, London. deprivation, all of their parents have got mended using Free School Meals (FSM)

48 Headteacher, Academy in jobs, but they might be scaffolders, or as the best proxy indicator. PWC also rec- London. drive white vans, and their aspiration ommended an amount for any child with 49 They actually identified four for their children is ‘school will look EAL or FSM of £1,780 derived from (including SEN and “other learning needs” but these were not after you for 5 years’, it doesn’t really almost 4,000 survey returns from schools. funded through the AEN part of This was based on “met” costs which the the funding formula. matter what you come out with because PriceWaterhouseCoopers, Study you’re going to be a scaffolder like me, or school already had to pay (e.g. learning of Additional Educational Needs – Phase II Final Report, (August, a painter. Those are very difficult chil- support assistants) and “unmet” costs (the 2002). dren to teach, and you do need the additional support schools felt that pupils

20 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 21

Delivery

needed but that they were unable to pro- authorities examined did not use any FSM vide). Unfortunately the “unmet” needs measurement. A third used a simple (and identified by schools came to, on average, crude) per FSM pupil measure.51 £1,020 per child with AEN, which was One alternative, used by 12 authorities, deemed too expensive; so the Govern- is the Index of Multiple Deprivation ment reduced it arbitrarily to £250. They (IMD). This is a combined indicator also reduced the met costs, meaning that developed by the Department of the total additional funding for each AEN Communities and Local Government child was £1,210.50 (DCLG) across seven different domains At the time this was a major step for- (housing, education, health, crime, ward in school funding. Never before had employment, income and living environ- there been any attempt to identify the ment). Previously the IMD was only calcu- additional costs associated with teaching lated at ward level which was not hugely certain groups of children. However, by useful as inter-ward geographies can differ significantly reducing the amount offered dramatically. In 2004 the IMD was for unmet needs, schools with large num- applied to Super Output Areas (SOAs), bers of AEN pupils remained under- which are small areas averaging around resourced. Furthermore, because the for- 1,500 people designed by the Office of mula used just two indicators (EAL and National Statistics, using zone-design soft- FSM), considerable nuance was lost. FSM ware to combine compact shape and rela- is not a comprehensive measure as a signif- tive social homogeneity. There are 34,378 icant number of those eligible do not sign SOAs compared to around 10,000 wards up either because they do not want school at any given time (wards are electoral and food or because of the stigma attached to so boundaries change regularly – one of the receiving free food. In addition FSM is a benefits of SOAs is that the boundaries do relatively crude measure. For a start it is ‘all not change allowing more accurate trend or nothing’ – any formula including FSM analysis).52 Applying IMD at this level assumes that a child whose family income gives it a greater level of accuracy enabling is low enough to receive them (around authorities to use it in a formula. Bury, for £14,000) has additional educational needs example, scores each pupil by their IMD but one whose family is just over the limit and allocates deprivation funding accord- needs no extra support. It is also a purely ingly.53 economic indicator, and as the headteacher IMD is, however, not pupil specific and quoted earlier in this section noted, certain even though SOAs are tightly drawn there communities are culturally rather than are still small pockets of deprivation in pre- financially deprived. dominantly affluent ones and vice versa. FSM is also the predominant indicator And, while IMD takes a few more factors 50 For a sense of the politics in the various formulae used by local into account than just income, it is still involved in this decision see Martin Johnson, Schools authorities to re-allocate their share of the weighted towards economic rather than Budgets – Fair Enough?, (IPPR, DSG. Following the 2005 Treasury review cultural factors. It is also unable to discrim- February 2003). of Local Authority funding for deprivation inate between relatively affluent areas 51 Judith Partington, Indicators of Deprivation for use in School LAs were asked to provide statements because all of the indicators are focused on Funding, about their formula and how it would entirely negative factors. http://www.teachernet.gov.uk/do cbank/index.cfm?id=10254 develop to take account of the criticisms There is another type of indicator that is 52 See made. A technical review of these state- used by just a few authorities at the http://www.neighbourhood.statis tics.gov.uk/ for rankings of SOAs ments gives a sense of the variety of ways in moment which is more promising. These and more detail regarding their which resources can be targeted to disad- are “geodemographic” classifications which construction. vantaged areas. Only 5 out of the 75 use postcodes to provide an extremely fine- 53 Partington, p. 20.

www.policyexchange.org.uk • 21 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 22

School funding and social justice

grain analysis of communities (on average ments (including, for example, Welfare there are about fifteen households in each Borderline or Symbols of Success). For the postcode). They are based on the principle full list of segments see http://www.caci. that similar types of people live in similar co.uk/acorn/acornmap.asp for ACORN types of neighbourhood; that “birds of a and http://www.business-strategies.co.uk feather flock together”. Once certain “seg- for MOSAIC. ments” have been identified as having spe- Just six local authorities use these classi- cific characteristics, cluster analysis tech- fications (all use ACORN) to distribute niques can group all of those individuals, deprivation funding. Most, like Harrow, households or small areas with those char- use it as a factor within a more complex acteristics. Nearly all major companies formula. Wakefield’s system probably now use these classifications to target cus- makes the most comprehensive use of geo- tomers but, as yet, little use has been made demographics: personalised learning fund- of them in the public sector. ing is allocated on the proportion of pupils in ACORN categories 4 and 5; other additional social deprivation funding is allocat- The geodemographic type of an individual pupil was a ed to primary schools with more than 80% “ of pupils on roll in categories 4/5 and sec- better predictor of their GCSE score than IMD ” ondary schools with more than 72% in these categories.56 Recent research suggests that these clas- The two principal commercial classifica- sifications have more predictive power tions are ACORN from CACI and than either FSM or IMD. Professor MOSAIC from Experian. Both use similar Richard Webber (who developed both data sets including: the national census; ACORN and MOSAIC), working with shareholder registers; the electoral roll; the Tim Butler from King’s College London, postal address file (PAF); house price and has measured the predictive ability of vari- Council Tax information; ONS local area ous deprivation measures using a statistical statistics and lifestyle data surveys. Out of method based on deviation from average these both extract around 400 variables, GCSE points score. They found that the though Experian place more emphasis on geodemographic type of an individual the census which provides 54% of MOSA- pupil was a better predictor of their GCSE IC variables compared to 30% for score than IMD. It is also a better predic- ACORN.54 The exact process for choosing tor than whether that pupil was on FSM. variables is not explained in detail This may seem counter-intuitive but actu- (Experian only say that they conduct a ally makes sense because the FSM measure “detailed analysis of societal trends in the has no ability to predict performance United Kingdom”).55 between the majority of pupils who are not Cluster analysis is used to group all the on FSM.57 54 ACORN User’s Guide, (London: CACI, 2006), p. 4; postcodes in the UK into segments out of Webber and Butler went on to investigate MOSAIC Brochure, (Nottingham: Experian, 2004), p. 4. which a comprehensive classification is why the IMD has less predictive power and

55 MOSAIC brochure, p. 4. developed. The ACORN classification has found that when they ordered the MOSA-

56 Partington, p. 23. 56 segments that are grouped up into sev- IC types in descending order of average

57 Richard Webber and Tim enteen broader segments and 5 overarch- GCSE points there was significant misalign- Butler, Classifying pupils by ing categories (Wealthy Achievers; Urban ment between the ranking based on average where they live: how well does this predict variations in their Prosperity; Comfortably Off; Moderate GCSE points and the ranking based on the GCSE performance?, CASA Means and Hard-Pressed). MOSAIC has IMD of the pupils’ Super Output Area (see Working Paper 99, (CASA, 2005), pp. 11-12. 61 categories grouped in 11 broader seg- Table 1). In other words pupils from some

22 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 23

Delivery

MOSAIC types were performing consider- should be based on a geodemographic clas- ably better at GCSE than one would expect sification. There are two clear advantages given their IMD score, while others were over other possible mechanisms. First, it performing proportionately worse. One rea- covers a far wider range of factors because son the authors give for this is that, as the of the vast number of different variables DCSF technical review noted, IMD is built into the classifications. These factors designed to discriminate only between go beyond the economic and encompass deprived areas so misses differences at the ethnicity, lifestyle, culture and proximity higher end of the social scale. However, the to different social groups. This means that authors also point out a more problematic a formula based on geodemographics will issue based on the MOSAIC types that not draw a simple proxy relationship show the greatest misalignment between between financial deprivation and low GCSE performance and IMD rank. Pupils standards. Secondly, although they include from inner-city social housing types and a lot of data, such classifications are rela- types containing predominantly Asian fam- tively easy to understand as they are ilies perform better than their IMD rank designed for a commercial audience, would suggest, while predominantly white whereas the IMD, for example, was pupils from large overspill estates in urban designed for local government. areas tend to perform worse.58 The two most popular commercial Supporting the headteacher quoted ear- options are ACORN and MOSAIC. We lier in this chapter, Webber and Butler have chosen to do our analysis with contrast the attitude of south Asian com- MOSAIC primarily because there is much munities with white working class families more data available for modelling due to living on council estates where material the work already done by Professor standards of living may be higher but Richard Webber in this area.60 “where lack of educational qualifications Assigning a different level of pupil pre- has not traditionally been seen as an obsta- mium to each classification segment would cle to earning good money and where flex- be unnecessarily complex (there are 61 ibility and opportunism are often per- within MOSAIC). So the first step is to ceived to attract greater rewards than dili- reduce the various segments down to a gent consistency of direction.”59 They also smaller number of groups based on educa- contrast inner-city social housing where tional performance. Table 1 shows all of children are surrounded by examples of the MOSAIC classification types with the social role-models and overspill estates average GCSE score for that type and the with “low horizons” where they are very equivalent IMD. We have constructed six few economically successful people groups based on this average GCSE per- around. In other words geodemographic formance of pupils within each segment: analysis can pick up cultural as well as material deprivation which makes it a bet- The first – highest-performing – group

ter predictor of performance than IMD or contains all types with an average 58 Ibid, pp. 14-15.

FSM. GCSE points score of 50 or higher; 59 Ibid. p. 16.

The second all types scoring an average 60 There is a another classifica- of 45 or higher tion called OAC developed by the Office of National Statistics How would the “pupil premium” aspect The third those types above 40 using only census data. If the of the formula work in practice? government where to invest in The fourth types above 35 boosting the predictive power of On the basis of the analysis in the previous The fifth types above 30 OAC by paying for the inclusion of non-census data then this section we believe that the pupil premium And the sixth all types below 30 (the would be a viable, non-commer- aspect of a national funding formula lowest is 24.15) cial alternative.

www.policyexchange.org.uk • 23 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 24

School funding and social justice Small TownSmall Seniors type Mosaic TableTypesScoreMOSAIC GCSE Average1: with Caring Professionals Caring Collar Blue Affluent Retirement Bungalow Connections Global TouristAttendants Families Rung Middle Memories Sepia Enterprise Asian Villages Parochial Subtopia Sprawling Nurseries Fledgling Playgrounds Summer Business Time Small Good Making Families SerenityChildfree ValuesConservative Symphony Pastoral Farmers Hill Upland Colonists Urban New Suburbs Original GuardiansGreenbelt ToClose Retirement Owners New Upscale Elders Spending High Seclusion Semi-Rural TechnologistsHigh Privilege Provincial Nesters Empty Golden Leadership Cultural Chieftains Corporate 20524 25352 11286 23434 17777 16577 13560 16646 20667 10580 12063 17173 12051 Pupils 3943 2774 7287 4877 3086 5335 1412 3850 9027 5138 9075 6439 3648 6226 556 878 437 728 Ave GCSE Ave 41.26 41.55 42.05 43.03 43.22 43.44 43.45 43.48 43.66 43.87 44.47 44.51 44.85 44.85 45.68 46.38 46.66 46.96 47.33 47.42 49.25 49.56 49.73 50.24 50.28 51.10 52.61 53.32 55.34 55.44 56.29 Deprivation 0.00 0.01 0.01 0.00 0.03 0.01 0.00 0.00 0.20 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.0 Source: Richard Webber and Tim Butler, Classifying pupils by where they live, CASA Working Paper 99, 2005. Working99, CASA Paper live, they where Butler,by WebberTim pupils RichardSource: and Classifying Rows Respectable type Mosaic Just Moving In Moving Just City Adventurers City In Military Quarters Military In Dinky Developments Dinky TownTransitionGown University Challenge University Low Income Elderly Income Low Industrial Grit Industrial Burdened Optimists Burdened Settled Minorities Settled TownRefuge Centre Older Right ToRight Older Buy Counter Cultural Mix Cultural Counter South Asian Industry Asian South White VanCultureWhite Metro MulticultureMetro Bedsit Beneficiaries Bedsit Coronation StreetCoronation Rustbelt Resilience Rustbelt Cared For Pensioners For Cared Sharing A Staircase A Sharing Ex-Industrial Legacy Ex-Industrial Old People In Flats In People Old New TownNew Materialism Dignified Dependency Dignified Upper Floor Families Floor Upper TowerLiving Block Low Horizons Low Families On Benefits On Families 13361 27637 10884 11444 10627 12832 24009 12074 16230 18900 14528 25270 23617 13560 Pupils 1870 1162 2017 1796 6096 3613 4766 1120 2801 8555 879 485 559 195 537 580 Ave GCSE Ave 41.22 41.20 40.89 38.98 38.77 38.52 37.88 37.28 37.02 36.79 36.07 35.72 35.38 35.13 34.99 34.26 32.98 32.80 32.07 30.96 30.95 29.52 29.39 29.14 28.10 27.74 26.67 25.87 25.81 24.15 Deprivation 0.00 0.08 0.10 0.00 0.20 0.14 0.12 0.04 0.01 0.01 0.16 0.08 0.20 0.29 0.23 0.03 0.62 0.21 0.10 0.08 0.06 0.75 0.20 0.17 0.09 0.46 0.37 0.61 0.48 0.50

24 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 25

Delivery

Some of those types in the sixth group – that nine times as much money would that would receive the highest level of need to be spent on black children to get pupil premium – have lower IMD scores them to the same average level of attain- (i.e. deprivation is not as severe) than some ment as whites (American reseachers tend of those in higher groups; which illustrates to use race as a proxy for deprivation the problem of using indicators that are rather than class).62 primarily economic rather than social or Setting the premium, even for the most cultural. deprived children, at nine times the base Once the MOSAIC segments have been per pupil cost (around £40,000) is not, of grouped according to average GCSE per- course, remotely feasible. Nor, even if there formance, financial values need to be was, would there be enough teachers to assigned to each group. Ideally these values provide all the one-on-one tutoring this would be calculated according to the cost would pay for. However, we believe it is of bringing attainment for children from fair to argue that the premium should be a disadvantaged backgrounds up to the high as possible. Taking into account the national average. However, as current economic conditions we have put PriceWaterhouseCoopers found, when as much money into our theoretical premi- deriving the original value for the um as feasible without assuming tax Additional Education Needs (AEN) in increases or shifts from non-educational 2002, this is more or less impossible. They spending to pay for it. surveyed almost 4,000 schools asking for Working from these assumptions we the costs of teaching children with AEN assigned £500 for children in Group 4 of and while there was some agreement about our MOSAIC groupings; £2,000 to the existing costs (extra teaching assistants; Group 5 and £3,000 to Group 6 (there is teacher time; pastoral support) there was a more on the total cost in the next chap- huge variation in the “unmet need” report- ter). The reason for spreading the money ed by schools i.e. the needs that schools across all of the bottom three groups is were not able to support for lack of that spending it all on the most deprived resources. School estimates ranged from is too crude – it would mean schools with very little to just under £27,000 per pupil. only slightly different kinds of population Estimates were arbitrarily capped at getting very different amounts of money.63 £1,800 by PWC meaning that the mean of This is one of the main problems with all the school estimates was £590 rather using just FSM – 14.4% of children are than £1,020 as it would have been if left eligible for FSM but 46% would be given 61 Johnson, p. 5. uncapped. This was reduced to £250 in extra support through our pupil premi- 62 Ludger Woessmann and Paul Peterson (editors), Schools and another arbitrary change made by the um. This is roughly the same number that the Equal Opportunity Problem, Government because of cost pressures.61 do not achieve 5 A*-C grades at GCSE; (MIT Press, 2007). The schools making the highest esti- and of course there is a strong correlation 63 There are relatively similar number of pupils in groups 4,5 mates were probably, though, the most between the two groups. By grading the and 6 (14.5%, 15.5% and 16.1% of all pupils respectively) so the accurate. Because schools have a relative- amounts across the three worst perform- amounts allocated to each group ly small impact on pupil attainment com- ing MOSAIC groups we still ensure that could be tweaked proportionally without affecting the overall cost pared to parents and peer groups a very the most deprived get the most money (for example, £750 for Group 4 large amount of money would have to be but that all of the groups that currently and £1,750 for Group 5). It would, of course, though affect spent to get all students from the most are below average get some extra support. the amount going to each deprived social groups to the national The maximum amount for Group 6 school. We decided to focus disproportionately more funding on average. There have been no studies look- pupils - £3,000 – is roughly equivalent to the poorest groups so as to ing at how much would be needed in the the uncapped version of the PWC AEN make a significant funding difference to schools in the most UK but American researchers have found allowance adjusted for inflation. deprived areas.

www.policyexchange.org.uk • 25 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 26

School funding and social justice

Table 2: Luton Secondary Schools Funding

School Name % Students Achieving Existing MOSAIC FSM 5 A*-C including Funding Premium Premium English and Maths

Cardinal Newman Catholic School 60% £4,249 £4,493 £4,089 Icknield High School 53% £4,285 £4,402 £4,078 Challney High School for Boys 50% £5,222 £5,724 £5,795 and Community College Denbigh High School 43% £4,922 £5,538 £5,576 Lealands High School 43% £4,480 £4,760 £4,474 Putteridge High School 42% £4,722 £5,071 £4,684 Stopsley High School 41% £4,352 £4,207 £4,449 Lea Manor High School 40% £4,766 £5,490 £4,872 Challney High School for Girls 37% £4,933 £5,567 £5,501 Ashcroft High School 37% £4,444 £4,973 £4,459 Halyard High School £5,044 £5,917 £5,201 Now Barnfield West Academy South Luton High School £4,800 £5,719 £4,842 Now Barnfield South Academy

Amounts are annual per pupil funding Existing funding taken from 2006/7 Section 52 returns - and includes current deprivation, SEN and Standards Fund money GCSE results taken from official DCSF tables for 5 A*-C including English and Maths for 2007 FSM premium based on £5,500 per child on FSM MOSAIC premium based on year-11 MOSAIC data from 2005 (we have assumed all other years have the same proportion of students in each MOSAIC group for our estimate)

To see how this would impact on school existing funding for SEN pupils with state- budgets in practice we added the pupil pre- ments, but without existing deprivation mium to existing per pupil funding for all funding, School Standards Grant and 22,309 schools in England.64 The above Standards Fund money – the overall cost tables show three per pupil values for each would be the same at the MOSAIC pupil secondary school in four example authori- premium). ties. The first is the actual per pupil sum The first table gives these three values for 2006/7, including existing deprivation for the secondary schools in Luton – a fair- funding passed on by the local authority, ly average authority. The first thing to note the School Standards Grant and the is that using the MOSAIC premium all 64 We used data matching the 2005 pupil census for years 6 Standards Fund money allocated to each schools bar one get more money than they (for primary) and year 11 (for school; the second is a hypothetical per do at the moment, while using the cruder secondary) with MOSAIC codes provided by Professor Richard pupil sum including the pupil premium FSM premium three schools are worse off Webber. Full school data was based on MOSAIC outlined above and another two get roughly the same as unavailable so we weighted the total number of pupils according (including existing funding for SEN pupils they currently do. More important though to the numbers in each MOSAIC with statements, but without the existing is the relationship between performance group for the two years we did have. Data was unavailable or deprivation funding, School Standards and funding. The two schools that would corrupted for about one sixth of Grant or Standards Fund money); and the have got the most additional funding schools. For these schools we kept existing costs (including third is a hypothetical sum based on a sin- under the MOSAIC premium, but not standards funding) constant when measuring the total cost of gle FSM measure (with £5,500 attached to much more under FSM, were Halyard introducing a premium. each pupil with FSM and again including High School and South Luton High

26 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 27

Delivery

Table 3: Nottingham Secondary Schools Funding

School Name % Students Achieving Existing MOSAIC FSM 5 A*-C including Funding Premium Premium English and Maths

Trinity Catholic School 73% £4,455 £4,734 £4,232 Fernwood School 68% £4,307 £4,146 £4,170 The Nottingham Bluecoat 57% £4,718 £5,266 £5,258 School and Technology College Greenwood Dale School 51% £4,457 £5,672 £5,025 Farnborough School Technology College 32% £4,474 £5,318 £4,820 Ellis Guilford School and Sports College 30% £4,672 £6,068 £4,954 The Nottingham Emmanuel School 30% £5,959 £5,502 £5,946 Henry Mellish Comprehensive School 27% £4,968 £5,986 £5,061 Big Wood School 25% £4,709 £5,942 £4,353 Fairham Community College 22% £6,074 £6,393 £5,292 Top Valley School 22% £4,383 £5,341 £4,327 William Sharp School 20% £4,733 £6,343 £5,130 The River Leen School 14% £4,625 £6,044 £5,206 Haywood School 10% £6,713 £6,525 £7,100 Elliott Durham School 7% £5,706 £6,429 £5,604 Hadden Park High School 7% £4,778 £6,462 £5,357

School. Both became Academies last year the MOSAIC premium even more clearly. because their performance was so bad. The Table 3 gives values for the secondary same is true for Ashcroft High School – the schools in Nottingham – which has some next (equal) worst performing school in of the worst performing schools in the Luton and Lea Manor (the next worst). country. Again there is a strong correlation This illustrates clearly that low parental between the value of the MOSAIC premi- income is not necessarily a good fit with um and poor performance. The two low performance. Taking culture and schools that have fewer than 10% of stu- neighbourhood into account is far more dents achieving 5 A*-C including English accurate. Incidentally if the differences and Maths at GCSE would both get far between the funding models don’t seem more money under a MOSAIC premium significant – consider that Lea Manor has than an FSM one. Both these schools are 1110 pupils – the MOSAIC premium situated near extensive estates which have would bring an extra £800,000 a year, serious crime problems and where young which would pay for around 25 extra people often have few positive role-models. teachers. The fact that many parents earn a little The next three tables are more extreme more than the cut off for FSM is not par- examples that illustrate the superiority of ticularly relevant. MOSAIC picks up these

www.policyexchange.org.uk • 27 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 28

School funding and social justice

Table 4: Tower Hamlets Secondary Schools Funding

School Name % Students Achieving Existing MOSAIC FSM 5 A*-C including Funding Premium Premium English and Maths

Oaklands School 56% £6,849 £7,466 £8,068 Morpeth School 48% £6,510 £7,019 £7,143 Mulberry School for Girls 48% £6,304 £7,611 £8,134 Central Foundation Girls' School 47% £6,304 £7,180 £7,679 Swanlea School 42% £6,335 £7,310 £8,220 Stepney Green School 36% £6,550 £7,397 £8,399 Bishop Challoner Catholic 35% £5,791 £6,616 £6,740 Collegiate Girls School George Green's School 34% £6,979 £7,390 £8,350 Sir John Cass Foundation and Redcoat 33% £6,335 £7,302 £8,339 Church of England Secondary School Langdon Park Community School 32% £6,686 £7,611 £9,637 Raine's Foundation School 32% £5,854 £6,450 £6,828 Bow School 30% £6,876 £7,338 £9,901 Bethnal Green Technology College 26% £6,610 £7,182 £7,767 St Paul's Way Community School 20% £6,805 £7,699 £9,502

Bishop Challoner Catholic Collegiate Boys School not included due to lack of MOSAIC data

differences and would see a lot more receives considerably more funding as a money go to these schools. The third lot more children take FSM (and much of worst performing school, Haywood, the existing funding model is based on bucks the trend as the local authority FSM). Under a pure FSM model Tower already allocate it a huge amount of Hamlets schools receive vastly more Standards Fund money – but it is due to money than those in Barking. Bow close in 2009 in any case. Otherwise the School would receive almost £10,000 per lowest performing schools all get much pupil (£2,500 more per pupil than they more money under MOSAIC than either get at the moment).65 Under the MOSA- of the two other methods. IC premium schools in both authorities The final two tables offer perhaps the would be better off but by similar

65 If you were wondering why starkest contrast between the funding amounts. Again, this is because a lot of the MOSAIC premium seems to models. Table 4 gives values for secondary children in Barking live on white working be more generous than the FSM one in Nottingham and Luton schools in Tower Hamlets and Table 5 for class estates with low aspiration but while costing the same overall those in Barking and Dagenham. In terms incomes above the relatively low level this is the reason – huge amounts of money would be of performance the two sets of schools are needed to qualify for FSM. In Tower transferred to a few very (finan- relatively similar – below average, but not Hamlets far more children quality for cially) poor areas – especially in inner London. as bad as Nottingham, yet Tower Hamlets FSM but the majority are Bangladeshi –

28 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 29

Delivery

Table 5: Barking and Dagenham Secondary Schools Funding

School Name % Students Achieving Existing MOSAIC FSM 5 A*-C including Funding Premium Premium English and Maths

Barking Abbey Comprehensive 58% £6,445 £6,392 £6,126 School and Sports College All Saints Catholic School 55% £5,570 £6,217 £5,501 and Technology College Robert Clack School 50% £5,795 £6,869 £6,463 The Sydney Russell School 37% £5,423 £6,609 £6,152 Dagenham Park Community School 33% £6,393 £7,015 £7,139 Eastbury Comprehensive School 30% £6,010 £6,933 £6,996 Eastbrook Comprehensive School 28% £4,995 £5,845 £5,004 The Warren Comprehensive School 28% £5,369 £6,226 £6,402

Jo Richardson Community College not included due to lack of MOSAIC data

and Bangladeshi families tend to have existing age-weighted funding and adding high aspirations for their children.66 This in pupil-led special educational needs is one reason why Tower Hamlets is one of funding (which is typically tied to a spe- the fastest improving authorities in the cific pupil statement of needs and would country despite refusing central govern- need to paid regardless of any other pre- ment and business help in the form of mium). The age-weighted base funding is Academies. Again the actual performance calculated differently for each local of schools correlates far better with the authority – because of deprivation fund- MOSAIC premium than either existing ing and the area cost adjustment in the funding or a premium based on FSM. Dedicated Schools Grant – and is in turn based on an out-of-date central funding formula. If we were to move to a National Further questions and variations Funding Formula, as recommended The figures contained in Tables 2-5 are above, there would be a national per pupil based on a number of assumptions; so it is base with a standard area cost adjustment worth going over a few potential variables (ACA). However, we decided not to try to and questions to be resolved that could incorporate this because it is impossible have significant impact on the amounts to calculate an ACA without knowing

involved. what will happen to the national pay 66 See, for example, Geoff Dench, agreement. We believe that the Academies Kate Gavron and Michael Young, The New East End, (Profile Books, 1. The age-weighted base per pupil value: programme has blown a hole through the 2006).

our figures hold the base per pupil figure agreement (Academies can set their own 67 See Sam Freedman, Briar constant across all three columns. This pay).67 The logical extension is to allow all Lipson and David Hargreaves, More Good Teachers, (Policy “base” figure was calculated by taking the schools to opt-out of the agreement, in Exchange, 2008).

www.policyexchange.org.uk • 29 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 30

School funding and social justice

which case average local pay rates become It might also be necessary to gradate a much more important and the ACA primary premium more gently as they needs to be more comprehensive. In any have smaller annual intakes so annual case the national base / ACA element of variations in pupil population could cre- the national formula would need to be ate greater instability than at secondary phased in over a number of years if it led level. The principal though is the same – to dramatic changes from what schools typically the schools that have the weakest currently received. Our intention, with Key Stage 2 scores would benefit most these figures, is only to show how the from a MOSAIC premium and a pure pupil premium aspect of any formula FSM premium would mean too much could work. variation between schools of relatively It is also worth remembering that, at the similar intake in terms of attainment. moment, the money passed on to schools Incidentally, it is worth noting that it depends to some extent on how much makes sense for pupils to be reassessed for their local authority holds back which dif- funding purposes when moving from pri- fers quite dramatically (from 5%-22%). If mary to secondary as their parents situa- local authority budgets were fully separat- tion may have altered. ed from schools budgets and consistent across the country then these differences 3. Sixth-form pupils: we have assumed that would be eliminated. This would likely A-level pupils would be eligible for a pre- mean that the per pupil age-weighted base mium even though they are funded sepa- for schools in some authorities would rately via an allocation from the Learning increase and for others it would fall. It and Skills Council (LSC). The LSC alloca- should be possible to set the local authori- tion includes an amount for deprivation – ty budget at a low enough level that the calculated using a similar range of factors considerable majority of schools would be to the DCSF formula; so in our calcula- better off – by making sure that it covered tions A-level pupils are getting counted only those areas that had to be dealt with twice for the purposes of deprivation (we centrally (such as admissions, transport have taken this into account when calcu- and pupil referral units). lating overall costs). Sixth-form pupils make up just 11% (many students move 2. Primary vs. Secondary: we have to sixth-form or FE colleges at sixteen) of assumed that the premium would be paid secondary school pupils - and a much equally for children at primary and sec- higher proportion are from well-off fami- ondary schools. Given that the base rate is lies than their younger counterparts – so it lower for primary schoolchildren (they does not make a huge difference. Not are taught in bigger classrooms and with- applying the premium above 16 would out specialist subject equipment) it could make some sense though as by definition be argued that they should receive a pro- if young people have made it A-level stan- portionately lower premium; which dard they are doing relatively well (and would mean that a higher premium for very few sixth-formers take vocational secondary-age pupils would be affordable. qualifications in school sixth forms at the Equally one could argue that primary is moment), so perhaps money should be currently under-resourced given its key targeted earlier in the school cycle. This role in securing the life chances of its approach would save some money that charges. There is relatively little a second- could be used to make the 5-16 premium ary school can do if the children they a little more generous (or cut overall admit cannot read and write. costs).

30 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 31

Delivery

If, however, the premium was to be extend- some kinds of severe SEN have no proxy ed to 18 then does the national funding for- (they occur randomly). The rest is allocated mula need to be extended to educational pro- on the basis of deprivation. Here low birth vision outside of schools? The LSC is in the weight and incidence of income support process of being disbanded with the money were used rather than FSM – as they were for 16-18 due to be passed to local authorities stronger proxies for SEN.69 Of course, as so it is a propitious time to re-think funding with all other parts of the DSG, the current for this age group. If A-level funding was to allocations are out-of-date as they are simply be taken into a national schools funding for- based on past allocations. mula, though, what would happen to 16-18 year olds in other forms of education? In a recent article Professor Alison Wolf notes “ If A-level funding was to be taken into a national that most colleges recruit actively and are not schools funding formula, though, what would happen to oversubscribed in the way the best schools 16-18 year olds in other forms of education? are so there is less need for an incentive.68 She ” has also argued that a premium might not be suitable for further education or apprentice- ships as it could act as an incentive against If we were to move to a national funding accepting adult learners as providers would model then money for statements could be focus on 16-18 provision. directly assigned from the centre according However, not applying the premium to the needs of individual children – avoid- above 16 might provide a disincentive to ing the need to incorporate severe SEN in schools to encourage pupils to stay on. the main national formula. In fact this To cope with the disincentive effect against would help solve an ongoing problem with adults, the scheme could be extended to the statementing process: local authorities’ cover them as well – although, of course conflict of interest (they decide whether a this would raise its costs. pupil needs a statement and then have to pay the costs of any statements awarded). If 4. Special Educational Needs: as mentioned statements were awarded by an independent above we have included pupil-led SEN panel and were then paid for out of a dedi- money in all three columns of tables 2-5. cated pot by the national agency this con- This money is typically assigned by local flict could be resolved.70 There is, of course, authorities to schools as funding attached to an accompanying risk that the number of statements of special needs; in other words statements would skyrocket along with costs for children that have an identifiable and seri- – the system would have to be designed so ous condition that requires additional fund- no-one involved had a perverse incentive to ing. We believe that this money would con- artificially inflate the number of statements tinue to be paid regardless of any premium as awarded. most types of statemented SEN do not have There remains a separate issue of how 68 Alison Wolf, Premia Post-16: Not so Easy?, Economic Affairs, a strong proxy relationship with deprivation. to deal with less serious SEN within a June 2008, pp.82-84.

At the moment authorities do not receive national funding formula. There is no 69 DCSF, DSG Formula Review a separate pot of money to cover statements. separate allocation for this in the current Papers, High Cost Pupils – Special Education Needs, 14th Instead part of the DSG is allocated on the DSG and it is difficult to see why one July 2008.

basis of “high-cost” SEN incidence (esti- would be necessary given the strength of 70 This pot could also be used mated at 3.4% of the school population). the relationship between low-level SEN to pay for places at the succes- sors to Pupil Referral Units Unfortunately this is calculated in a very (primarily behavioural problems) and social (PRUs) for children excluded rudimentary manner. Some is allocated deprivation. Using MOSAIC is helpful in from mainstream education. Most (though not all) pupils at purely on the basis of student numbers as this respect. As we have seen MOSAIC is PRUs have some kind of SEN.

www.policyexchange.org.uk • 31 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 32

School funding and social justice

good at identifying communities with low saved by making local authority budgets aspiration (even if they are not particularly consistent to school budgets would also financially disadvantaged). It is precisely reduce the number of overall losers, especial- these communities that produce large num- ly if this was weighted towards wealthier bers of children with behavioural problems, areas through an area cost adjustment. which is one of the main reasons why per- Even if these measures were introduced formance is so low at their schools. there would still be some serious losers. Though it is impossible to know how many 5. Losers: introducing a premium inevitably without modelling the steps mentioned means that the budgets of some schools in above we estimate it would be in the region wealthier areas will decrease. Under the of 1,000-2,000 schools or between 5%- model we’ve used many more schools would 10%. For these schools a transition fund win than lose (by a ratio of almost 2:1) but it could be set up to manage a reduction in is unlikely that this government or any future budgets over a number of years. This would government would want to reduce the budg- cost in the region of £250 million to start et of any school, with immediate effect, to with (based on a minimum average of the point where they had to lay off staff. £150,000 per school). Assuming a pupil There are a number of ways the number of premium were not introduced before the “losers” could be reduced. Increasing the age- next funding settlement in 2012, this fund weighted base at primary level (by reducing could be paid for through an inflationary the premium) as discussed above would limit rise in school budgets as long as the figures the number of serious losers. Adding money we have used for our modelling were stable.

A National Formula – what organisation should deliver it?

A National Funding Formula would have to be delivered via an arms-length National Funding Agency so as to avoid unhelpful politicisation. The DCSF already provide direct funding for Academies, but at the moment this represents fewer than 100 out of 22,000 schools. National funding would require a separate bureaucracy to develop the formula, keep it up-to-date and resolve dis- putes between schools over individual pupils. If the DCSF tried to handle this centrally it would overwhelm its strategic functions while drawing ministers into specific controversial cases that would undoubtedly arise. England already has experience of a National Funding Agency – between 1994 and 1999 a Funding Agency for Schools (FAS) was set up to calculate and pay grants to grant-maintained schools, to monitor their financial position and to administer a borrowing regime for the schools. Though there were numerous problems with grant-maintained school model the FAS worked well. In 1998 it allocated over £2 billion to over a thousand grant-maintained schools with a staff of 326 and operating costs of £12.4 million.71 Once a National Funding Agency had been established it could take on additional responsibili- ties. For example it could also fund statements of Special Educational Needs – with decisions on whether to grant the statements made either by local authorities or independent local assessment boards. If the pupil premium were to be accompanied by new legislation that allowed schools to be built by non-state providers regardless of local authority support – as is currently promised by the Conservatives and Liberal Democrats – then a Funding Agency could also take responsibility for new school applications. Central government could set a range of criteria that would have to met by any new entrant and the Funding Agency could decide whether new applicants fulfilled these criteria. 71 http://www.archive.official- Again this would help the DCSF from becoming embroiled in controversies over individual schools. documents.co.uk/document/cm41 /4157/dee02.htm

32 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 33

Cost

It is axiomatic that policy proposals should raised by ending existing central govern- have realistic cost implications. We saw in ment grants sitting outside the Dedicated the previous chapter that a pupil premium Schools Grant. Moving to a straightforward could hoover up considerable funds and national formula with just three elements – still not achieve the objective of parity a base age-weighted amount, an area cost between deprived and wealthy communi- adjustment, and a pupil premium – with ties – the difference in performance is sim- no additional major sources of funding ply too great. We argued that the premium would massively simplify the system. The should be set as high as possible without premium could then act as a clear incentive requiring additional taxation or cuts in for headteachers to avoid cream skimming. other departmental budgets. The amounts There is no good reason for continuing given for children from the most disadvan- with significant funding streams outside taged communities, which we used to the main national formula. The School demonstrate how the premium would work Standards Grant would be entirely redun- in the previous chapter, were roughly dant if all school revenue was paid directly equivalent to the uncapped cost figures for to schools. Its only purpose at the moment Additional Educational Needs produced by is to allow central government to bypass PriceWaterhouseCoopers in their 2001 local authorities. There is also no good rea- analysis. son to continue with the Standards Fund which is based on a historical amalgam of defunct policy grants. It typically diverts How much would the pupil premium funds to schools in deprived areas but the cost and how will it be paid for? pupil premium would do the same in a far Using figures from the 2006-2007 “Section more logical and consistent way. The same 52” returns from local authorities we calcu- is true for other politically motivated pots late that the total cost of per capita funding of money such as the “personalised learn- delegated by authorities (and the Learning ing” grant and the “national challenge” and Skills Council) to schools was £26.4 funds. These kind of short-term funding billion. This figure excludes money spent interventions increase the complexity of the centrally by local authorities on services for funding system and confuse headteachers. schools and money delegated to schools via As we saw in the first chapter headteachers the School Standards Grant and the find this kind of funding extremely difficult Standards Fund. If our model of a pupil to manage; a sudden influx of funds for a 72 The full cost with MOSAIC was £31.2 billion but this dou- premium, using MOSAIC, had been oper- specific initiative (like the beacon schools ble-counts deprivation funding ational in 2006-7 the amount delegated scheme) will often lead to a restructuring of for A-level students as LSC deprivation funding is included as directly to schools would have been £31 bil- staff that then becomes unaffordable when well as the pupil premium. We lion; an increase of £4.6 billion.72 the money stops. It was changes in assume deprivation funding accounts for around £200 million We believe this can be paid for by reas- Standards Fund allocations that led to the (11% of the total £1.75 billion signing funds from within the education financial “crisis” of 2003 for which the gov- LSC funding for A-level pupils) as no official figures are avail- budget. Much of the extra money could be ernment blamed local authorities. able.

www.policyexchange.org.uk • 33 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 34

School funding and social justice

Table 6: Standards Funds Allocation

Standards Fund Grant* (millions) 2007-8

Ethnic Minority Achievement 178.6 Targeted Improvement Grant 12 Extended Schools 64.5 Targeted Support for Primary Strategy 141.4 Targeted Support for Secondary Strategy 106.3 Primary Strategy: Communication Language and Literacy 2.6 School Improvement Partners 12.6 Music Services 59.6 Music at Key Stage 2 23 Education Health Partnerships 11.6 Playing for Success 9.5 Regional School Travel Advisors 0.2 London Challenge 0.2 Choice Advisors 5.3 School Workforce Data Collection Pilot 1.8 School Intervention Grant 15 Flexible 14-19 Partnership Funding 25.9 Walking to School Initiatives 2.6 General Duty on Sustainable Travel 4 Total 676.7

* Excluding School Development Grant and Capital Funding (Source: DCSF)

Using the latest available figures from or wished to use the money for a different 2007-8 – the SSG cost £1.17 billion, the set of policy-oriented grants - the SSG, the School Development Grant (which makes personalised learning grant and the School up most of the Standards Fund) cost £1.84 Development Grant alone add up to £3.36 billion, other Standards Fund grants (not billion. including school meals grants) came to In addition to this there are further sav- £676 million and the personalised learning ings that could be made within the educa- grant was £356 million, giving a total of tion budget to fund a pupil premium. One £4.04 billion.73 It is unlikely that any gov- of the most significant new lines in the ernment would want to allocate all of this budget since 2002 has been the Education money directly to schools as the “other Maintenance Allowance – a means-tested Standards Fund grants” cover a consider- payment of up to £30 a week for young able range of political priorities – for exam- people in post-compulsory education

73 DCSF, Section 52 Budget ple in 2007-8 £250 million was spent on between the ages of 16-19, introduced to Summary 2007-8, (lines 1.0.2; grants to support the national primary and boost post-16 participation and attain- 1.0.4; 1.0.6; 1.0.7; 1.6.1; 1.6.2) secondary strategies. However, it is likely ment.74 In 2007-8 it cost £550 million.75 To 74 http://www.direct.gov.uk/en/ EducationAndLearning/14To19/ that a number of these grants could be date the impact on post-16 participation MoneyToLearn/EMA/DG_066951 rolled into a pupil premium (see Table 6 and attainment has been marginal at best. 75 http://www.dcsf.gov.uk/abou for a full list). Even if a future government A recent analysis published by the Institute tus/reports/pdfs/7493-DCSF- AnnualReport.pdf, p.89. did not wish to cease any of these grants – for Fiscal Studies (IFS) suggests that there

34 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 35

Cost

may have been an increase of just 2% in the pared for in their previous schooling. number of female participants post-16 and Moreover the pupil premium is consider- no increase at all in the number of male ably easier to administer, while the EMA participants.76 In terms of attainment, the requires a complex means-test that adds IFS analysis suggests that it may not have another layer of complexity to the strained increased average performance at all for education bureaucracy. This year comput- female students and by just one quarter of er problems have left numerous teenagers one A-level grade for male students.77 without the payments they had been Figures obtained by the Conservative party promised.81 suggest that just 400 extra children on FSM There are a number of other existing have stayed on to do A-levels since 2004.78 costs that could be cut to provide funds for So the EMA is, in effect, a massive dead- a pupil premium. The National Challenge weight cost – providing payment to 46% of programme for “failing” schools that was learners,79 the vast majority of whom would launched this summer to widespread dis- have been in post-16 education in any case. may in the education community will cost Once new government legislation to make £400 million over the next three years.82 As 16-18 education or training compulsory numerous commentators have pointed comes into force in 2013 the entire cost of out, it makes little sense to focus this kind 76 Haroon Chowdry, Lorraine the EMA will effectively become dead- of resource and attention on a list of Dearden and Carl Emmerson, weight. As young people will have to partic- schools derived from one variable that can Education Maintenance Allowance: Evaluation with ipate anyway, it can have no positive incen- change dramatically from year to year Administrative Data, Institute for tive effect. (whether fewer than 30% of their pupils Fiscal Studies, November 2007, p. 3. The IFS researchers used The only possible remaining argument receive 5 A*-C including English and two different methods to analyze the impact of the EMA. The first for the EMA is social justice – that young Maths). Is a school where 31% of children saw participation and attainment people from poorer backgrounds deserve achieve 5 A*-C so much more successful in pilot local authorities which began receiving EMA in 1999 to be supported from 16 rather than at 18. than one where just 29% of children make compared with authorities desig- This is a pretty weak argument given that the cut? Furthermore the interventions on nated as controls when the IFS conducted their original analysis the vast majority of such young people live which money will be spent may well be of the EMA pilots. The second at home – and their parents receive child counterproductive. £140 million will be compares authorities which began receiving EMAs in 1999 benefit until they are 19 (or leave full-time spent on National Challenge Advisors, and 2000 with all other authori- education). In fact the government origi- Leadership programmes and additional ties in England. We quote figures calculated using the first of nally intended to pay for the EMA by end- learning support, all of which cut across these methods. ing child benefit for over-16s but were existing programmes (School Improvement 77 Ibid., p.6. warned off by child poverty campaigners.80 Partners, the National College of School 78 http://www.guardian.co.uk/ education/2008/aug/04/tories. Continuing the EMA and the post-16 Leadership and the National Strategies / allowance?gusrc=rss&feed=edu child benefit while making post-GCSE Making Good Progress / Every Child pro- cation education compulsory is entirely illogical. grammes) and are likely to cause confusion 79 http://www.theyworkforyou. com/wrans/?id=2008-04- There is a far better social justice case for and bureaucratic overload. The rest of the 29b.201685.h&s=%22education the pupil premium – improving the money will be spent on Trusts and +maintenance+allowance%22 80 http://news.bbc.co.uk/1/low/ resources for schools in deprived areas and Academies – but this spending would be education/2116635.stm

reducing cream-skimming would signifi- unnecessary if the structural reforms pro- 81 http://www.timesonline.co.uk/ cantly boost attainment before 16. This in posed by the Conservatives and Liberal tol/life_and_style/education/articl e4656751.ece itself which would lead to much greater Democrats came into force alongside the 82 £200 million of this is “new” rises in post-16 participation than the pupil premium. money announced in the 2008 EMA has managed, and it could be done Another significant, and entirely unnec- budget the other £200 million will be diverted from elsewhere by raising aspiration rather than forcing, essary, drain on education resources in the in the education budget. DCSF, future is the ContactPoint database, which National Challenge: A Toolkit for or bribing, young people to take courses Schools and Local Authorities, that they have not been adequately pre- will hold records of all the children in the June 2008, p.17.

www.policyexchange.org.uk • 35 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 36

School funding and social justice

country so that practitioners in health, likely. Given this, the original cost esti- education and social services will be able to mates seem ludicrously low. As Terri highlight concerns about potential at-risk Dowty, from the charity Action on Rights children. It was set up in the light of the for children (ARCH), noted in her evi- Victoria Climbie case, despite the fact it dence to the House of Lords: would have made no difference to this tragedy. It has been widely criticised by “The Government has already award- children’s charities and school groups for ed the contract for construction of potentially placing young people at risk ContactPoint to CapGemini and set- because so many people will have access to up costs are estimated as £224m. We the database. It has also been criticised for doubt this figure, particularly given logistical naivety as the very children who the frequency of cost overruns on gov- are most at risk are the least likely to have ernment IT projects. To give just one up-to-date information available for a example: in the first year of HM database. The House of Lords select com- Revenue and Customs’ ‘Aspire’ pro- mittee report on the database, published in gramme, the cost of the 10-year con- July 2007, synthesised many of these contract with CapGemini reached £539m cerns: against an estimated spend of £385m…. The annual running costs “We do not consider that the of ContactPoint are set at £41m. This Government have demonstrated conclu- amounts to around £270,000 per local sively that a universal database is a pro- authority to cover the costs of staffing, portionate response to the problem being training, system maintenance and addressed, or that the additional bene- administration. We are concerned that fits of a universal approach justify the the figure of £41m is a substantial additional costs and risks, as compared underestimate.” 85 with a selective approach which would not include a child in the database The DCSF budget has allocated £88 unless or until the child’s needs for spe- million, £86 million, and £85 million to cialist or targeted services became ContactPoint over the next three years in apparent…. On current estimates, capital and revenue costs. An honest re- ContactPoint will hold data on 11 mil- costing of the programme would produce lion children, and there may be over figures far higher. Scrapping the database 300,000 users of the scheme. The enor- now would save at least £100 million a mous size of the database and the huge year. number of probable users inevitably There are other ways of saving money increase the risks of accidental or inad- within the education budget – a number of vertent breaches of security, and of existing education quangos could be deliberate misuse of the data (e.g. dis- scrapped or merged, for example. But the closure of an address with malign pupil premium could be paid for by the 83 House of Lords, Merits of Statutory Instruments intent), which would be likely to bring savings outlined above. For the purpose of Committee, 27th report 2006-7 83 session, Draft Children Act 2004 the whole scheme into disrepute.” the table below we have assumed that all Information Database (England) existing Standards Fund grants would be Regulations 2007, July 2007. ContactPoint has already run into trou- diverted to the premium except for those 84 http://www.guardian.co.uk/ education/2008/sep/02/schools. ble. It was supposed to be launched in currently targeted on the national strate- children1 Spring 2008, and again in October 2008 gies as it is unlikely that any future govern- 85 http://www.publications.pa but the pilot scheme has now been delayed ment would wish to scrap these without rliament.uk/pa/ld200607/ldselect 84 /ldmerit/146/14608.htm until January 2009. Further delays are putting anything in their place.

36 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 37

Cost

their pupils. If the government could Table 7 spend it better through, for example, Standards Fund grants, then it would not Cost of Pupil Premium £4.6 billion make sense to implement all of these

School Development Grant £1.84 billion reforms. School Standards Grant £1.17 billion The evidence of the benefits of increased Personalised Learning Grant £356 million spending on education over the past ten Other Standards Fund Grants £426 million years suggests that the Government have Education Maintenance £550 million not been particularly successful at allocat- Allowance ing resources. Figures published by the National Challenge (annual) £133 million Office of National Statistics (ONS) last ContactPoint £100 million year show that school productivity Total £4.6 billion (defined by GCSE results and the number of children in schools) had increased by only 1% between 1996 and 2006 despite Of course, should any future govern- an 83% increase in spending from £27 bil- ment wish to raise additional tax revenue lion to £50 billion (which includes money to support the pupil premium, then it spent on other areas of education as well as could either be paid for without recourse schools).87 Since 1999 productivity has to existing funds or the premium could be actually been falling. set at higher levels. The Liberal Democrats Some of the money ploughed into edu- have said that they would raise £2.5 bil- cation by the current Government was lion for the premium; £1.5 billion by tak- never likely to have a significant impact on ing wealthier families out of the tax credit productivity, but was, nevertheless, neces- system and a further billion by “savings sary. For example, increased capital expen- from duplication and waste in West - diture on a crumbling infrastructure and minster”.86 The first half of this package is increases in teacher salaries that had fallen realistic but it is questionable whether too low. However, as we have seen, much another tax rise should be used to fund the of the new money has gone to schools in premium when the education budget has the form of grants oriented to a specific more than doubled in the last ten years. policy, with the aim of improving attain- Given the current economic climate it is ment. As the ONS figures show, the gains more likely that any tax increases from a in attainment associated have not been future government would be needed to proportionate to the increases in spending. fund other targeted tax cuts or reductions Furthermore, it is arguable that the ONS in borrowing. are being generous in using official government attainment statistics. Analysis by academics at the University of Durham sug- 86 http://www.guardian.co.uk/ education/2007/nov/20/schools. Will the money add value? gests that children of the same ability get uk2 There is a clear social justice case for the higher GCSE results now than they did in 87 http://www.timesonline.co.uk/ tol/life_and_style/education/articl pupil premium. The funding system we 1996. Using their own “Yellis” test as a e2388507.ece; http://www.statis have outlined would be far more consistent control they show that between 1996- tics.gov.uk/articles/nojournal/Edu cation_productivity_2007_execs and logical than the current one; and it 2007 the average GCSE grade achieved by umm.pdf

would redistribute money to those schools candidates of the same ability has risen by 88 Robert Coe and Peter which look after the most deprived chil- almost two-thirds of a GCSE grade.88 If Tymms, “Summary of Research on Changes in Educational dren. It will only have a positive impact, this analysis is correct then productivity Standards in the UK”, in Mike Harris, Education Briefing Book though, if schools spend the money on has actually fallen on average over the last 2008, (Institute of Directions, resources that will boost the attainment of ten years. August 2008), pp. 97-98.

www.policyexchange.org.uk • 37 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 38

School funding and social justice

The key question is whether additional teaching assistant, and the combination unsupervised spending by schools would of those two would give sufficient flexi- produce even worse results. The evidence bility within the Maths team, to be able from our interviews with headteachers was to target all the difficulties that we’ve encouraging, though the sample was not got.” 91 large enough to draw definitive conclu- sions. It was clear that spending priorities “I think we can push the workforce per- would differ widely between different formance thing much further, and have schools indicating that there are few, if any, qualified support staff, and have excel- uniform problems that additional central lent teachers in shortage subjects….Pay government programmes could solve. The your Maths teacher £60,000 a year, let range of projects on which schools wished them teach large classes, but in technol- to spend money suggested that headteach- ogy rich areas, supported by three or ers had given considerable thought to solu- four good maths support staff.” 92 tions that would be applicable to the particular problems of their communities. Headteachers talked about making the One thing on which all our interviewees most of existing flexibilities within the agreed was the primary importance of good national pay agreement – especially the quality teachers. Most heads said that their new Advanced Skills Teachers for excellent priority for extra spending would be either teachers who wish to stay in the classroom. to hire more staff or better staff. As one put But there was also agreement from the it when asked what she would spend extra majority of interviewees that additional money on: “it would be teaching, people flexibility would help them attract the best make the difference, you invest in staff, you teachers, especially if they had more money only raise achievement and give your chil- than schools in wealthier areas: dren a wonderful education if you’ve got the right people doing it…”89 Others gave simi- “I think that heads are still, as they’ve lar answers: “It would be hiring additional always been, creative in the way they use qualified teachers. It wouldn’t be hiring recruitment and retention [allowances] many unqualified staff, because in my view, but certainly a freeing up of that area when your needs are as complex as the ones would be a good idea. It does seem odd, we have to provide for; qualified, well- for example, that schools have to come up trained staff, even though they’re more with a recruitment and retention policy expensive, they contribute greater.”90 which more or less treats everybody the There was cautious acknowledgement same when the whole purpose of the pol- that to hire the best staff it would be nec- icy is to recognise that some people are essary to offer higher salaries: harder to recruit than others, so to get a head of maths may cost more than to get “Currently my school is doing extremely a head of history.” 93 89 Headteacher, primary school, London. well in terms of 5 A*-Cs, but not very

90 Headteacher, secondary well if you include English and Maths, “The key [to resolving teacher shortages] school, London. English is going brilliantly, the issue is is the workforce reform element, teach- 91 Headteacher, secondary school, North-East. Maths. I’ve got a Maths team that ers don’t like it, the unions despise it,

92 Headteacher, secondary works very hard, I just don’t have the but I think there is a lot of mileage school, London. depth. With that sort of money, I think there, in terms of the para-professional 93 Headteacher, secondary I would be looking to appoint, in addi- thing, we still have to recruit good school, London. tion to my current staff, an advanced mathematicians and good scientists, if 94 Headteacher, secondary 94 school, London. skills teacher, and probably a high level they are available.”

38 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 39

Cost

“Ideally the performance management more effective pathways for individu- system would be more flexible than now, als…our maths department want to and we would be able to offer staff more experiment with bigger class sizes to give of a financial incentive, but one has to greater flexibility to take smaller groups realise this can be very divisive as well, of children off….but…all of that costs we have to be very sensitive about it.” 95 money.” 98

In “More Good Teachers”, a report pub- “I think we have to be flexible in the lished by Policy Exchange in July 2008, it way we use our resources. For exam- was recommended that schools should be ple…for year 7 intake, we try to inden- able to opt out of the national pay agreement tify a group of kids who are particular- (at the moment only Academies are ly low attainers, poor literacy, poor exempt).96 We argued that this would benefit numeracy and you can relate entirely to schools in deprived areas who need to pay poor social class background. We put more to compensate staff for a tougher work- these kids into much smaller groups, we ing environment; especially if, thanks to a then tailor curriculum to meet their pupil premium, they were better off than needs, and it’s more like a primary cur- other schools. Our thinking is explained in riculum than a secondary curriculum, more detail in the full report, but the inter- with a smaller number of teachers views undertaken for this project suggest a working with those kids…” 99 significant number of headteachers would make good use of any additional freedoms. For others pastoral support meant pro- Another theme that ran through a num- grammes to increase the aspirations of ber of interviews was pastoral support. their pupils: This, though, meant something different to each of the headteachers we spoke to, “It’s raising money to do all the other dependent on the kind of community they enrichment things…when they say they were serving. For some it primarily can’t afford to go away for the week, involved additional academic support: paying for them…and if they want to learn the trombone then we’ve got a “Pastoral care is a slightly dated term, it trombone for you to learn…or they can implies a counselling role…which clearly each go to football training or athletics has to be done by schools…But I’mtalking training….all those things that the par- about a slightly more focussed academic ents won’t do, all the extra things.” 100 coaching…I guess the best way that I can describe it is what students benefit from at “We do a lot of residential trips and that university, where they have a personal sort of thing…trying to get funding for tutor, who really challenges and pushes these things is difficult so you give stu- 95 Headteacher, secondary school, London. them, and knows them extremely well as a dents the advantage of these things, in 96 Sam Freedman, Lipson and learner…the quality of that is dependent extreme cases we had 23 who went across Hargreaves, pp. 61-65. on the time that is spent having those deep to China, because we’ve got a link over in 97 Headteacher, secondary school, London. conversations about learning, and that’s Beijing and these kids learned 98 Headteacher, secondary what makes the difference to young people Mandarin and that sort of thing, and school, Cambridgeshire.

in schools, [but] both structurally and we managed to get a lot of sponsorship, so 99 Headteacher, secondary financially we find that difficult.” 97 we only paid a third of the £900 it cost. school, London. Managing that, a lot of work goes on, 100 Headteacher, secondary school, London.

“Additional funding should be geared outside school hours just to set up these 101 Headteacher, secondary to…increasing flexibility, personalising trips and trying to raise funding.” 101 school, London.

www.policyexchange.org.uk • 39 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:05 Page 40

School funding and social justice

Finally, there were some schools whose central or state programmes (in the US).104 local communities were so deprived, finan- But, smaller scale studies show that specif- cially and culturally, that the headteachers ic, costly, interventions can have a signifi- saw an urgent need for more basic pastoral cant impact on different types of schools. support. One headteacher of a federation This is not particularly surprising. As of schools explained: was evident from our interviews, and indeed our general experience of English “We have an exclusion unit that we run schools, different schools require different here, and we take students from the kinds of programmes to boost results. Take other school two miles away….very dif- class sizes, for example. Over the past few ficult school, all boys, mainly decades there has been a stream of studies Caribbean, and you will find our unit is showing class size has only a small, or in full of their students…the kids are terrif- some cases non-existent, impact on per- ically disadvantaged, with vast numbers formance.105 This seems counter-intuitive of Special Needs, poor behaviour, and until one considers the individual circum- there’s a dozen children who are elec- stances of each school. In some cases it tronically tagged. We put them in deten- could lead to a dilution of quality – if there tion and they pull out their plastic cred- are only a few highly able teachers and it card which says I am under curfew, I fewer children are given access to them. A 102 Headteacher, secondary need to be home at 4 o’clock….So I guess number of headteachers we interviewed school, London. there’s an example of a school with a spoke of increasing class sizes to give more 103 Headteacher, secondary school, London. large number of SEN kids, and there, children access to the best teachers, if they 106 104 Eric Hanushek, The Failure your resource priority would need to be could afford support staff to assist. Where of Input-Based Schooling Policies, The Economic Journal, not just classroom support, it would need teacher quality was not such an issue other 113 (February 2003), pp. 64-98; to be the pastoral side.” 102 headteachers did want to reduce class sizes Eric Hanushek, “School Resources”, in Handbook of the – especially for their weakest students who Economics of Education, ed. “Basically the money would go on sup- needed the most support. Bearing this in Eric Hanushek and Finis Welch, (Elsevier, 2006). port staff, the school is always under a mind, a government programme to system-

105 See, for example, Nye, B., lot of pressure, but so are the pastoral atically reduce class sizes would have mixed L. V. Hedges, and S. offices…It comes back to this, a tremen- results. Alternatively, giving the money Konstantopoulos, The effects of small classes on achievement: dous amount of work goes into follow- directly to schools allows them to spend it The results of the Tennessee class-size experiment. American ing up the pastoral problems which go appropriately. Educational Research Journal with the students…We don’t get much The same is true for other forms of 37:123–51 (2000) and Akerhielm, Karen, 1995. "Does class size support, because we’re not priority in expenditure; extended services, ICT matter?," Economics of the borough, we get somebody who resources and longer school days all work Education Review, Elsevier, vol. 14(3), pages 229-241, comes in once every two weeks…the stu- well in the right circumstances but uni- September. dents we have that have an attendance formly applied will provide only meagre 106 The UK’s pre-eminent problem need much more chasing up returns on investment. The one form of expert on class sizes, Peter 103 Blatchford, has made a similar than that.” expenditure which guarantees better argument – that there are specific instances in which reducing attainment is teacher quality (though not class sizes is beneficial, for The solutions proposed in our inter- necessarily as measured by teacher test example with those struggling with literacy. He argues it should views closely matched those suggested by scores). We discuss national programmes be carefully targeted, which is academic research. American academic to improve quality extensively in “More best done within schools. Blatchford, P., Russell, A., Eric Hanushek has established that there is Good Teachers”, but it was encouraging Bassett, P., Brown, P., and no easily identifiable relationship between that nearly all of our interviewees men- Martin, C., The effect of class size on the teaching of pupils overall financial inputs and outcomes in tioned the importance of using funding to aged 7-11 years, School the school system, and this has been con- boost the numbers of high-quality teachers Effectiveness and Improvement, 18, 2, June, pp. 147-172. firmed by numerous studies of specific in their schools.

40 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:19 Page 41

Cost

Despite this correlation between aca- Setting up this kind of institution and demic research and the responses of our widely disseminating its results would not interviewees we remained concerned that in itself guarantee that schools would opt increasing direct funding to schools could for proven programmes, but it would at exacerbate an existing tendency for schools least prevent conscientious headteachers to buy into new technology or new cur- from being hoodwinked. It would also offer riculum models which do not have an ade- schools, or groups of schools, that devel- quate research base. It is hardly the case, in oped their programmes the opportunity to England, that Government spending is have their work accredited. Finally if schools determined by evidence-based research decided not to use interventions that were (see, for example, the mass expenditure on proven to work, this could be taken into whiteboards before research-backed cur- account by Ofsted if the school was failing. riculum uses had been developed). However, it is at least theoretically possible that expenditure controlled by central gov- Will it act as an incentive? ernment could be tied to the results of high One of the proposed benefits of the pupil quality research programmes. The more premium is that it will act as an incentive money that goes direct to schools the for schools to stop cream-skimming chil- greater the risk that it will be spent on well- dren from wealthier areas as they will lose sold but unproven schemes. out financially. We believe there is a social To help counter this the government justice and effectiveness case for the pupil should set-up an independent body to premium regardless of whether or not it undertake evidence-based research into edu- acts as a direct influence on the behaviour cational interventions to strict international of headteachers. However, such an influ- standards. It is, frankly, astonishing that we ence would be an important additional do not already have such an institution – benefit, especially in the context of a system indeed the new Institute for Effective which allowed non-state providers to devel- Education at York University is the first op and run new schools with state money. body of any kind to offer this kind of research in the UK. It is true that the nature of the evidence-base required to “prove” the If schools decided not to use interventions that were value of an educational intervention is con- “ proven to work, this could be taken into account by Ofsted tested but as Professor Robert Slavin, the if the school was failing Director of the York Institute has written: ”

“The idea of evidence-based reform is simple to grasp and almost impossible Whether this would happen as the result with which to argue. Use what works. of the changes we suggest is extremely diffi- How could anyone disagree with that? cult to predict. There are no specific inter- Yet the difficulties inherent in agreeing national examples – the Netherlands does on what works are daunting, and there use a pupil premium type funding model are powerful pressures to keep the system but its school system is so different from as it is. Still, other fields, such as medi- ours that is impossible to draw useful com- cine, agriculture, and technology, have parisons. However, what we do know from instituted evidence-based practices, and international evidence is that both current they had to overcome similar pressures. and new school providers do generally respond to changes in the financial incen- There is no fundamental reason that 107 http://www.cddre.org/sol education cannot do the same.” 107 tive structures that schools face . In Sweden, utions/reform.html

www.policyexchange.org.uk • 41 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 13:13 Page 42

School funding and social justice

the possibility of using state funding to sup- Headteachers at school with relatively port non-state schools has led to the open- disadvantaged populations certainly ing of many new independent schools, thought that extra resources would allow innovative in form and popular with par- them to attract parents that would not con- ents. Various voucher schemes in the United sider them at the moment. As one put it: States have led to improved performance in both public and private schools.108 “Quite a large number of parents send When we asked headteachers of schools their kids to local primary schools but with relatively wealthy populations if the when they get to secondary, they send pupil premium would affect their behaviour them to the independent sector and…if with regards to admissions, the standard we were able to improve to such an response was to deny that any cream-skim- extent that we had the faith and trust of ming happens at the moment – which is a wider middle-class group then I unsurprising if inaccurate. In any case most would be happy for that, in as much as headteachers we spoke to were not in control I would really like this school to repre- of their own admissions system at the sent the community it serves, in its moment, which would not be the case if the entirety, not just a proportion of it and Academies / free school programme was I think if that was the case then there expanded. In terms of new providers, there is would be a reduction in the funding as yet no real market so again it is impossible according to the formula you described, to predict how they might respond to a but you can’t have it both ways.” 109 financial incentive to build in deprived areas. What is clear is that schools which cur- In any case, even if there was no incentive rently have relatively disadvantaged popula- effect, the pupil premium would reduce seg- tions would (if they used their new resources regation simply through providing extra well) attract more middle-class parents, resources for schools with many disadvan- which would have the same impact as a taged pupils. As the Julian Le Grand quote direct incentive on schools with wealthier we used in the first chapter put it: populations. This has already started to happen when previously failing schools have “Schools that contained a high propor- been replaced by well-resourced Academies tion of children from poor families with expensive new buildings. To take two would…have more resources per pupil examples, Haberdashers Hatchams Academy than those with a low proportion. They in Lewisham and Mossbourne in Hackney would also have better premises and have both seen significant influxes of middle- equipment and could attract higher- class parents in recent years as results and quality staff. The outcome would either behaviour have improved. Both schools have be selective schools, with those that spe- used a “banding” admissions system to create cialised in the education of the children genuinely comprehensive intakes. This sees of the poor being better equipped and equal numbers taken from each quarter of staffed than those that specialised in the the ability group which acts to widen the education of children of the rich, or, if school catchment area to cover a wider range head teachers or staff did not want to of communities. This has led to criticism, engage in such specialisations, schools

108 For a review of the evi- but only because there are no other high-per- that contained a reasonable proportion dence, see Le Grand (2007), forming school in these authorities. If this of children for all parts of the social pp.69-73. happened more widely – as a result of a pupil spectrum. One way or another cream- 109 Headteacher, secondary school, London. premium – then it would help to reduce seg- skimming that favoured the better off 110 Le Grand (2007), p. 148. regation in the education system. would be reduced or eliminated.” 110

42 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 43

Conclusion

At the moment revenue funding for ets. New school providers entering the English schools is a mess; a labyrinth built market, whether through the Academies out of failed policy grants, out-of-date programme, or future Conservative / demographics and unnecessary “stability” Liberal Democrat plans for supply-side measures. Not a single headteacher we reform, would have an incentive to set up spoke to could explain how their budget in deprived communities, where real was allocated. Even amongst policy-mak- choice is needed most. ers there is very little understanding. In In this report we have tried to show how these circumstances it becomes impossible a premium could be delivered. We recom- to use funding as a way to drive school mend using “geodemographic” classifica- improvement: if no one knows where the tions based on individual pupil’s postcodes money comes from it cannot act as an rather than measures based purely on incentive. income, like Free School Meals, as these There is absolutely no need for the sys- take into account cultural aspiration, tem to be this complicated. In this report which can be just as important as financial we have identified three ways to make the circumstances. We have modelled how this system simpler. First we should move to a premium would work using real budget consistent national funding formula, data from all English schools and have rather than having a different formula for shown that it would give a significant each local authority. Secondly, most of the boost to struggling schools. Using our fig- arbitrary central government grants should ures – which would see a premium of be merged into this formula. Thirdly, the between £500 and £3000 per pupil formula should have just three elements depending on the level of deprivation – the for each pupil: an age-weighted base, an total cost would be £4.6 billion, but this area cost adjustment and a premium for could be paid for out of the existing educa- those pupils from more disadvantaged tion budget. backgrounds. A wholesale shift in the funding mech- This last element is crucial. At the anism for schools is obviously a radical moment schools serving deprived commu- step. As the amount of money going nities do, typically, have more money than directly to schools would be significantly average. But this money is allocated in an increased relatively few individual schools entirely haphazard way, differently for each should have to lose out (rather the excess local authority and often through short- would be directed at schools in deprived term grants awarded by politicians eager areas). Nevertheless the transition would for positive headlines. Instituting a consis- need to be handled sensitively so that tent “pupil premium” would allow schools essential services currently provided by to plan their budgets around their admis- local authorities could continue sions. Over time schools with large disad- unharmed, while releasing as much money vantaged populations would be able to as possible for schools. At the moment the fund programmes to boost attainment, Government are reviewing funding for attract the best teachers and eventually 2012-2015. It should be possible to move middle class parents. Schools in better off to a national formula in 2012 with the full areas would have an incentive to spread system in place by 2015, and should be their nets wider, make their admissions supported by any Government that wishes more comprehensive and boost their budg- to call itself progressive.

www.policyexchange.org.uk • 43 SHCOOL TEXT HDS AMENDED.qxp 17/10/08 12:09 Page 44 SCHOOL COVER HDS.qxp 16/10/08 11:44 Page 1 colFnigadSca Justice Social and Funding School At the moment revenue funding for English schools is a mess; a labyrinth built out of failed policy grants, out-of-date demographics and unnecessary “stability” measures. There is absolutely no need for the system to be this complicated. In this report we have identified three ways to make the system simpler. First we should move to a consistent national funding formula, rather than having a different formula for each local authority. Secondly, arbitrary central government grants should be merged into this formula. Thirdly, the formula should have just three elements for each pupil: an age-weighted base, an area cost adjustment and a premium for those pupils from more School Funding disadvantaged backgrounds.

£10.00 ISBN: 978-1-906097-35-6

Policy Exchange Clutha House Exchange Policy 10 Storey’s Gate London SW1P 3AY