Skip to main content

Designing an education and training system – some principles and issues to bear in mind - Ewart Keep

19 May 2022


Neither research nor the policy formation process in England has tended in recent times to pay much heed to any notion of designing a coherent system of education and training (E&T) provision. The UK government’s approach to the policy process tends to centre on reform by institutional tinkering and a continuance of the great British tradition of incremental accretion of institutions and of ‘muddling through’ (Lindblom, 1959; Norris and Adams, 2017; Keep, 2006). Systems that have developed in this incremental manner often contain tensions between different routes, levels and design principles for funding, governance, etc.

It is also the case that in the past, across the developed world remarkably little E&T research was focused on conceptualising E&T as a system, or on exploring the principles that might inform E&T systems design. In recent years there has been a growing interest, from international bodies such as UNESCO and the OECD, in the question of how the architecture of an E&T system might best be constructed and the implications that different choices of configuration might have for the various potential stakeholders (OECD, 2016). In part, this interest has been driven by the rise of international educational performance league tables and testing regimes (e.g. the OECD’s PISA and PIAAC surveys) and by the resultant ‘battle of the paradigms’ as governments search for the educational model and systems architecture that can deliver optimal national educational outputs/outcomes (Ozga, 2013; Fenwich, Mangez and Ozga, 2014; Fischman et al, 2018).

In England, educational research has traditionally given limited attention to concerns about an ‘education system’ – in part because much academic study focuses on what happens in the classroom and upon pedagogic concerns rather than how education is organised and managed, but also because in the academic community the study of individual sub-systems or marketplaces and what goes on within them has predominated (so study of individual phases or stages of provision such as early years, junior, secondary, further education, higher education, apprenticeship, adult education and lifelong learning). This model of research has to a large extent reflected the reality of the fragmented manner in which activity is funded and managed, and also how policy has approached the reform of different stages or components of the overall education system or marketplace.

Even within research on sub-systems it is relatively rare (in the UK) to think about the design principles for each element – these have, as noted above, developed incrementally over time with little attempt to go back to first principles and construct an overview of how different components and elements of the system (or market) might interact and how they add up and support one another to form a coherent whole. What research on systems thinking, design and governance exists has tended to focus quite heavily upon schools (mainly secondary) and on universities. Further education, adult education and various forms of work-based learning have received more limited and often rather sporadic attention, particularly in recent times.

This short paper does not seek to fill this research gap in any comprehensive manner. To do so would be a massive undertaking and would result in a very lengthy report. Instead, what is attempted here is to construct a ‘think piece’ or speculative essay that outlines a selection of the most important ‘design principles’ or issues that policymakers and other stakeholders would need to address in order to undertake the design or re-design of an educational system or major sub-sections thereof. Its aim is to act as a ‘discussion starter’ and a prompt for further reflection on what we are trying to achieve in education and how best to configure policy design, funding and governance in ways that can best deliver what is desired.


In design thinking, the long-standing dictum is that form follows function (Sullivan, 1896). In other words, values, purposes, mission, and desired outcomes ought to drive the core design principles that will then shape the form of the system. However, within E&T in many instances form cannot simply be ‘read off’ from function or purpose, and in turn choice of form will impact on how and with what effect purpose is delivered. For instance England and New Zealand as two ‘Anglo-Saxon’ model countries both share a commitment to creating and maintaining world class higher education institutions and to be major players in the global overseas student market. The English policy response to these goals is a highly stratified HE marketplace, while New Zealand has opted for a relatively homogeneous and weakly hierarchically differentiated set of universities.

To put it another way, designing and running an E&T system is not simply a series of technocratic decisions (although these do exist and can be important). The process is also very heavily bound up with the ideological beliefs and values of those whose task it is to fashion policy, and plainly these can change over time (see below for further discussion). The purpose, structure and ‘limits of the possible’ in terms of policy instruments, structures and power allocations (Keep, 2009) that can be adopted are essentially shaped and determined by a set of fundamental prior ideological/political choices. Arguably these can be grouped into four ‘bundles’:

  1. The nature of what education is there to achieve/deliver, and for whom (what kinds of education for what proportion of the population and with what intended intellectual and moral outcomes). Thus, in some countries education for active citizenship is an important component of both academic and vocational/technical education (initial and adult), while in other nations one of the core roles of education is to seek the inculcation of unquestioning obedience to the state (and or the dominant religious culture) and a relatively limited and passive model of citizenship. To give another example, in some countries religious and cultural conventions mean that education for male and female students will be different and segregated institutional delivery will be the norm that is required, while in others the values and ideologies of the country will stress a very different model of gender-neutral education aiming at furthering greater equality between the sexes. The remainder of this paper will, with one exception (beliefs about the nature of the ‘skills problem’), have very little to say about this set of issues, and will concentrate on the next two design choices that follow.
  2. The second bundle of fundamental choices concerns the degree to which power is centralised or dispersed/devolved/shared within the E&T system, and if it is shared with whom and what responsibilities rest with the different categories or groupings of stakeholders (e.g. those representing parents, students and employers)? In part, choices in this area depend on underlying conceptions and levels of trust between the different actors. For example, if central government does not trust educational institutions to deliver what is required (perhaps due to a belief in a simple version of principal/agent theory) then it is likely to create a centralised system and to police it via low-trust inspection and punitive performance monitoring regimes.
  3. The third set of fundamental choices centre on the level of marketisation or contestability in funding that is designed into publicly supported delivery mechanisms.
  4. The fourth and final set of decisions to some extent follows on from the third and is to do with the degree to which E&T is to be delivered by individuals’ person investment in learning via courses in private (non-public, non-state) institutions (whether charitable or for profit). Thus in the USA, a significant volume of tertiary education is delivered by a wide variety of private institutions (for overviews of this issue as it plays out globally, see Macpherson, Robertson and Walford, 2014; and Verger, Fontdevila and Zancajo, 2018).

Within each nation state thinking about the optimal choices in each of these bundles is not fixed, far from it. Design principles change over time as ideological beliefs about the desired function and outcomes of education and the correct and optimal way in which to configure an E&T system have changed. For example, in the post-war period England had a very devolved education system where power was heavily dispersed and resided with Local Education Authorities (i.e. local government) which controlled schools, FE colleges and the polytechnics; individual educational institutions; and relatively autonomous intermediary bodies (for example, Her Majesty’s Inspectorate [HMI] and the Universities Grants Committee [UGC]). Central government’s role was limited, there was for example no national curriculum, and civil servants had very little direct managerial responsibilities for how provision was designed or delivered (see Pring, 2012). The last 40 years have witnessed the complete collapse of this model, as power has shifted upwards to central government, intermediary agencies have lost their powers and independence or become regulators in a marketplace, and the role of local government has faded into relative insignificance (Keep, 2006; 2009; 2018). Centralisation and marketisation have, until very recently, been the guiding ideological watchwords that have set the overriding design principles for English E&T.

What follows…..

What follows will try to draw out a few of the most important design choices and to illustrate these general points with some selected examples, drawn from experience in England. Given space constraints, this paper can only scratch the surface of many of these issues.


As noted above, there are a host of issues about the multiple purposes of education and the range of social, cultural and economic policy challenges that the educative process can be expected to address. In some instances policy aims in tackling these will be congruent, in other instances they may be in competition.

The one that will be singled out for attention here centres on vocational education (broadly defined) and concerns the exact nature and causes of and solution to the ‘skills problem’ as it relates to economic performance and outcomes (for the state, for employers and for individuals and the communities in which they live). This has been chosen because the questions that are normally at the top of the English policy agenda – how should the E&T system or marketplace, particularly for vocational education and training (VET) be (re-)configured and what exciting new institutions, programmes, initiatives and qualifications should we develop, are actually at least partially dependent upon the nature of the diagnosis of our ‘skills problem’. How the nature of the skills policy problem is conceived of is fundamental to how you will then design your E&T system to address this problem.

There are three possible elements to any diagnosis of the policy challenge:

  1. Skills supply
  2. Demand for skills from employers
  3. Skills utilisation in the workplace and the productive process

The central policy question is, is the core problem one of skills supply or is it bound up with a broader and much more complex set of interlinked issues that encompass supply, demand and utilisation of skills? If the problem is simply one of inadequate supply of skills, policy is narrower and the state can take the lead, find ways to force up achievement in compulsory education and massify post-compulsory provision (FE and HE). If, on the other hand, it is about potentially deficient underlying levels of demand for skills in the economy, and also about the fact that in many instances the workforce’s skills are being utilised sub-optimally (Keep, Mayhew and Payne, 2006), then a very different policy approach is needed, with more stakeholder engagement and E&T policies making contact and acting in concert with employment, employee relations, innovation and economic development and business improvement policies – as is the case in Scottish, Wales, and the Scandinavian countries. For further discussion of this issue, see Keep, 2017. On the whole, English skills policy has chosen to assume that the problem centres on a deficient supply of skills rather than any other problems (Keep, Mayhew and Payne, 2006; Keep,2006; Keep, 2017), and as a result educational policy and the design of the E&T system has failed to mesh particularly strongly with other areas of government policy.


The structure of power

One of the most important decisions about systems design in E&T relates to the dispersion of power within the policy and governance process. Even the briefest study of current arrangements across the developed world reveals a spectrum ranging from countries where central government is massively dominant and constitutes the prime actor and centre of power (e.g. England), to others where power is highly dispersed across different spatial and institutional levels and shared with a range of stakeholders (e.g. Finland).

In essence, the design questions that underlie this state of affairs are:

  • Who to involve?
  • What powers to afford the different actors
  • How to (and how not to) structure the processes through which this involvement occurs.

The list of potential stakeholders is quite lengthy and includes:

  • Students
  • Parents
  • Teachers/instructors/lecturers and the unions and professional bodies that represent them
  • The managers of educational institutions and the organisations that represent them
  • Local/municipal government and local civil society
  • Employers (represented at a range of spatial and sectoral/occupational levels)
  • Trade unions
  • Those representing other stages in the educative process

Choices about those with whom central government is willing to share power and what measure of influence they are to be afforded within the E&T system will have a huge impact on:

  • How institutions and their systems of governance are configured
  • How much buy-in/commitment to decisions and policies can be expected from other actors
  • The range and sources of ideas that can be injected into the policy formation and management processes of the system

The key point is that insofar as other stakeholders are afforded formal rights to be consulted within the decision-making process this may well constrain the power of the state to become the dominant actor and be able to unilaterally impose its own ideas and solutions. Change has to be negotiated with and potentially signed off by the other parties around the table rather than simply dictated from on high by the relevant government minister. The German ‘dual system’ of apprenticeship is a good example of this model in practice. The system is jointly owned by those who contribute to it – employers and the chambers of commerce, the trade unions, and the state (at federal and lander levels). In nations (like England) where such institutional constraints have either never existed or have been removed via the abolition of power-sharing mechanisms and conventions, the central government minister is in effect an ‘absolutist monarch’ pretty much able to decree change as they wish.

As noted in the Introduction, a key determinant of who is involved and in what manner and to what degree power is shared will be the levels of trust that pertain between the actors, particularly between national government and other stakeholders. If trust is absent or weak, then it is unlikely that central government will feel itself to be willing or able to share real power with those it distrusts, and this situation will in turn tend to dictate the kinds of power structures around policy formation, and governance and accountability regimes. For instance, if there is a strong belief that many educational institutions and the staff within them are inclined to be lazy and to deliver sub-optimal performance, then close monitoring by the state of their performance and a high-stakes and perhaps punitive inspection and performance management regime is the likely outcome. If trust is high, government will not feel the need to keep a tight reign on performance monitoring and will feel able to leave performance management and improvement to institutions with an inspection regime that is more heavily focused on formative rather than summative judgements. They will also be more likely to acknowledge professional autonomy and accept that decision making about pedagogy, curriculum and the delivery of education is best devolved to institutions and individual staff. It therefore follows that if there is a desire to shift a system from a centralised settlement towards greater levels of decentralisation, one of the precursors for this being possible is the construction of higher levels of trust between the stakeholders and between them and national government.

The other pillar that can support shared or dispersed models of power and accountability is a recognition by stakeholders (particularly the nation state) that a plurality of interests, voices, and opinions is inevitable, and that disagreement is legitimate and requires to be mediated. In countries that tend towards more centralised and state-heavy settlements central government as the dominant actor often adopts a unitarist view of the world, where dissenting opinions are viewed as illegitimate, unhelpful and need to be marginalised or excluded within policy and governance structures.

It is important to recognise that choices concerning the degree of centralisation or devolution/sharing of decision-making within a nation state do not emerge out of thin air. They generally reflect norms and expectations about the structure of power within civil society that go far beyond the confines of decision-making about education, and often what happens in education is largely a reflection of wider social, economic and political norms. A large body of research literature has sought to analyse the ‘varieties of capitalism’ that exist within different developed nation states and to develop a typology that demonstrates how choice of a given model of economic organisation then impacts the ways in which that state structures its skills system and vice-versa (see Hall and Soskice, 2001; and for a viewpoint that also covers the relationship between these choices and equality and social cohesion, see Green, 2013 and Green and Janmatt, 2006).

Countries with devolved settlements tend to be those that also have a tradition of embracing the legitimacy of multiple stakeholders within the policy process, with this concept sometimes being embodied in some form of social partnership system (e.g. Germany, and the Nordic countries). In England over the last 40 years central government has very largely abandoned those elements of a relatively limited corporatist or social partnership model that used to pertain, abolishing the tripartite (state, employers and trade unions) Manpower Services Commission (MSC) and gradually scaling back the involvement of trade unions in any form of decision-making about VET. The final acts in this process were the decisions to cease state funding for the trade unions’ collective adult learning organisation (Unionlearn) and to abolish the UK-wide UK Commission for Employment and Skills (UKCES), which were the last remaining substantive elements of a tradition of state-sponsored vehicle for employers and trade unions working together on skills issues.

Choices about the adoption or abandonment of social partnership models reflect the central importance within vocational forms of education of employers. As noted at the outset, if much of the driving force for educational reform is economic performance then there are major questions to be decided about what role employers might be expected to play within the E&T system (Keep, 2019). Thus, a clear settlement around and understanding of the rights, roles and responsibilities (the 3 Rs) of employers is a key component in the design of most successful national VET systems (Keep, 2020). In England, the problem is that the 3Rs remain woefully poorly specified, with no real agreement about what can reasonably be expected of employers and what their role within the skills system is, either in relation to involvement in initial VET or the provision of training to the adult workforce. With the decision by government to cease to fund the Sector Skills Councils and to abolish the UK Commission for Employment and Skills (UKCES), the sectoral architecture for ascertaining employers’ views or mobilising collective action on their part is now weak, fragmented and leaves many sectors and sub-sectors with no institutional framework at all.

Moreover, as noted above, the structure of power between different spatial levels of government in England has shifted over time, away from local government and towards central government and its agencies. This trend is not observable in many other countries, in part because of different traditions and constitutional arrangements regarding the relationship between the central state and smaller (sometimes pre-existing) spatial units of political activity and government. Thus the USA (federal government, state government and various forms of city/municipality/county government), or Germany (federal government and the lander) a relatively dispersed structure of political (and to some extent economic) power is hardwired into constitutional arrangements and cultural expectations. England, lacking a written constitution and with more fluid expectations about the relative roles and responsibilities of different levels of government has witnessed fluctuations over time in the relationship and relative powers of national and local/regional government, and Pring, (2012), Jones (2019), and the recent UK Government’s white paper (HMG, 2022) on ‘levelling up’ provide useful overviews of these shifts.

Another key dimension of system design is temporal:

In some countries systems or sub-systems of education change very slowly and the norm is incremental adjustment (e.g. the German ‘dual system’ of apprenticeship), while by contrast in other jurisdictions change is pretty near constant and often occurs over a relatively short timescale. If a system is characterised by ceaseless change, policy instability, and short institutional and policy life spans, then design of the system and management of policy delivery become a very different proposition from that which is the case in a system where more measured and incremental change is general practice.

The speed of change in turn reflects the way the system that is experiencing it is structured. As noted above, fast change by the centre is only possible in circumstances where the central state is powerful and other actors have little or no right of veto – in other words a country like England (Norris and Adams, 2017). Here even bodies (for example, National Training Organisations [NTOs] and Sector Skills Councils [SSCs]) that have nominally been designated as representing employers have, in reality, been designed, given an explicit remit with boundaries as to what they can and cannot do, and ultimately restructured or abolished by central government acting without any meaningful consultation with employers. In other words, because these organisations were partially funded by the state, they were treated as government-sponsored actors with limited autonomy. By contrast, in countries with social partnership models of systems governance, reform requires negotiation between those proposing the change and those whose interests may be impacted by what is being proposed.

Having looked at the fundamental choice between centralised and dispersed models of power within E&T, we now turn to a second set off decisions that will determine the overall shape and delivery of skills – the choice between market and system.

Markets or systems

A major set of decisions concerns the degree to which educational provision is conceived of and managed as a system, or as a marketplace. Marketisation is often seen by economists as a means to challenge the tendency (as they believe) that publicly funded institutions will often seek to satisfice or ‘coast’ rather than maximise performance and outcomes as the state desires (principal/agent theory). In a market, it is argued, where resources follow student or parental choice, individual institutions will be incentivised to maximise outcomes and provide a high-quality student experience in order to attract students and their associated units of resourcing, and less well-performing institutions will either be forced to improve or will be driven out of the marketplace (see Keep, 2018 for a fuller examination of this model). In other words, marketisation is, at least in part, another reflection of weak trust by the state in the delivery mechanisms and agents that they fund.

In a system cooperation between producers is assumed to be an inherent element of its design and operation. However, while ideal type models stress cooperation within systems, the reality is more complex, and all systems are liable to contain potential elements of competition. This is because E&T systems operate with finite levels of public resources and these are liable to be subject to competing claims from component elements within the system as different stakeholders and types of providers vie, for example, for student numbers and their associated funding. Given this reality, the key issue is how such competition is contained and managed.

Markets could be seen as a mechanism for a highly dispersed pattern of power or decision-making as the ‘invisible hand’ of the market displaces actions by the state, and some advocates of educational vouchers-driven models have argued that marketisation and student/parental choice is the means by which the state can step back from too large and direct a role in educational provision. However, as witnessed in England, it is possible to organise and regulate markets in ways that mean that central government continues to hold onto significant powers to direct and control what happens within a marketplace (e.g. for further education) and marketisation does not therefore lead to a devolved pattern of power and diversity in delivery. This is because regulatory and inspection regimes impose a uniform set of officially sanctioned and preferred pedagogic approaches and performance measures backed by very heavy sanctions upon institutions that are deemed to be ‘failing’. In other words, it is entirely possible to combine marketisation with a powerful and direct state influence (see Keep, 2018). Indeed, if the state finds itself unable to countenance the costs and disruption of institutional financial failure and substantial under-performance in learning outcomes then intervention in delivery units has to be ratcheted up if failure is a persistent problem.

To illustrate how centralisation and decentralisation interact with notions of market and system in different countries, see Diagram 1 below:

Diagram 1.

Ewart Keep Blog

The existence of many other considerations and decisions

The foregoing has sketched in a few dimensions of the first order design questions and choices that structure and influence the shape and operation of E&T systems. There are many others that there is not space to cover here, but which will matter. For example, choices around the role of evidence and evaluation within the design and management of the system will also be important. Should it formative or summative? How independent should it be? Who should conduct it? How should its findings feed into the design and management of the system?


Hopefully the foregoing will have begun to illustrate the complexity of designing an E&T system or sub-system. Many of the choices on some of the first order issues outlined above will have impacts on subsequent, subordinate choices – opening up some avenues and closing off others. As a result, systems design ideally requires an holistic approach that acknowledges and understands the interactions between:

  1. The consequences generated by choices on a spectrum of policy options between market and system
  2. The consequences generated by choices on a spectrum between centralisation and devolution
  3. Different institutions in terms of both vertical and horizontal impacts (e.g. local networks or rivalries).
  4. Different levels of provision and flows and learner journeys on pathways from one level to the next
  5. The relationship between education and occupational labour markets, and between education and the economy and labour market at different levels of spatial gradation.

If a system is to operate successfully it should be aiming to be more than just the sum of its parts, and if the required level of congruence between different elements or component parts of the system is not achieved there is a danger that it will operate sub-optimally, with, for example, wasteful duplication of provision and complex and confusing learner journeys. The evidence in recent times in England is that many policy makers struggle to fully grasp how individual elements and sub-systems within E&T function, never mind how the different elements react and interact within the system as a whole. The result has often been unintended consequences and outcomes that have not matched expectations.

In thinking about design the following two intentionally overlapping sets of systems design questions could be used as a starting point to structure discussion and thinking about critical choices that have to be made in order to determine the final shape of resultant system:

Questions to address - 1

  1. What is the nature and the causes of the policy problems/goals that the system is seeking to tackle? What are the aims of the educative process (at different stages and levels)?
  2. What levels of trust and autonomy does government and intermediary/regulatory bodies want to see as the norm?
  3. Who are the stakeholders and what priority will be afforded to their views via what representational mechanisms?
  4. Who sets the priorities and owns the decisions within the system?
  5. How best and by what mechanisms and routines can competing interests and resource needs be mediated?
  6. How centralised or devolved do you want the system to be?
  7. To whom are individual institutions and of any intermediary bodies that fund them to be accountable, and via what mechanisms is this to be delivered? How is institutional governance to be structured and what parties/stakeholders are to be involved and represented within governance arrangements? What underlying levels of trust underpin these arrangements?
  8. How ‘marketised’ and contestable should publicly funded provision be?
  9. By what measures and via what mechanisms is institutional performance to be measured and judged? Who should be involved in setting these and in ensuring that performance is monitored and evaluated?
  1. What, if any, feedback mechanisms need to exist between those who deliver education and those who construct policy at higher levels?
  1. What other areas of policy does E&T need to be linked to?
  2. What is the evidence base and how do we gather, process and use evidence within policy making and the management of the system and its institutions?

Questions to address - 2

  1. Who is responsible for delivering what, and how much can the state trust them to deliver what is required?
  2. Who pays for what?
  3. Who represents employers?
  4. Who represents individuals?
  5. Who represents wider society?
  6. What is the balance between education and training?
  7. What is the balance between resources devoted to throughput and to longer-term capacity building within delivery institutions?


Fenwick, T. Mangez, E. and Ozga, J. 2014. World Yearbook of Education. Governing Knowledge: Comparison, Knowledge-based Technologies and Expertise in the Regulation of Education. London. Routledge

Fischman, G. E., Topper, A. M., Silova, I., Goebel, J., and Holloway, J. L. 2018. ‘Examining the influence of international large-scale assessments on national education policies’, Journal of Education Policy, DOI: 10.1080/02680939.2018.1460493

Green, A. 2013. Education and State Formation: Europe, East Asia and the USA, 2nd edition, Basingstoke: Palgrave.

Green, A., and Janmaat, G. 2006. Education, Equality and Social Cohesion – A Comparative Analysis, Basingstoke: Palgrave.

Hall, P. A. and Soskice, D. (eds.) 2001. Varieties of Capitalism – The Institutional Foundations of Comparative Advantage, Oxford: Oxford University Press.

HM Government. 2022. Levelling Up the United Kingdom, CP 604, London: HMSO

Jones, M. 1999. New Institutional Spaces: Training and Enterprise and the Remaking of Economic Governance, London: Routledge.

Keep, E. 2006. ‘State Control of the English VET System – Playing with the Biggest Trainset in the World’, Journal of Vocational Education and Training, Vol. 58, No. 1, 47-64.

Keep, E. 2009. ‘The limits of the possible: shaping the learning and skills landscape through a shared policy narrative’, SKOPE Research Paper No. 86, Cardiff: Cardiff University, SKOPE.

Keep, E. 2017. ‘Current Challenges: Policy Lessons and Implications’ in C. Warhurst, K. Mayhew, D. Finegold, and J. Buchanan (eds.), The Oxford Handbook of Skills and Training, Oxford: Oxford University Press, 671-691.

Keep, E. 2018. ‘Scripting the Future – exploring potential strategic leadership responses to the marketization of English FE and vocational provision’ FETL Monograph, Tetbury: Further Education Trust for Leadership.

Keep, E. 2020. ‘Employers, the ghost at the feast’, Journal of Education and Work, Vol. 33, Issues 7-8, 500-507.

Keep, E., Mayhew, K. and Payne, J. 2006. ‘From Skills Revolution to Productivity Miracle – Not As Easy As It Sounds?’, Oxford Review of Economic Policy, Vol. 22, No. 4, 539-559.

Lindblom, C. E.. 1959. ‘The Science of “Muddling Through”’, Public Administration Review, Vol. 19, No. 2, 79-88.

Macpherson, I. Robertson, S. and Walford, G (eds.). 2014. Education, Privatisation and Social Justice: case studies from Africa, South Asia and South East Asia. St Petersburg: Symposium Books.

Norris, E., and Adam, R. 2017. All Change – Why Britain is so prone to policy reinvention and what can be done about it, London: Institute for Government.

OECD. 2016. Governing Education in a Complex World (Education and Research Innovation Series), Paris: OECD.

Ozga, J. 2013. ‘Accountability as a policy technology: accounting for educational performance in Europe’, International Review of Administrative Sciences, Vol. 79, No. 2, 292-309.

Pring, R. 2012. ‘Sixty years on: The Changing Role of Government’, British Journal of Educational Studies, Vol. 60, No. 1, 29-39.

Sullivan, L. H. 1896. ‘The Tall Office Building Artistically Considered’, Lippincott’s Magazine, March.

Verger, A., Fontdevila, C., and Zancajo, A. 2017. ‘Multiple paths towards education privatization in the globalizing world: a cultural political economy review’, Journal of Education Policy, Vol. 32, No. 6, 757-787.

The views expressed in Think Further publications do not necessarily reflect those of AoC or NCFE.