Submission from Ark Education Partnerships Group (Ark EPG) to the Parliamentary IDC inquiry on DFID's work on education: Leaving no one behind

Executive Summary


About this submission

This note provides an outline of evidence regarding the contribution of non-state schools, low-cost private schools and public private partnerships (PPPs) to education in developing countries. It also provides Ark EPG’s recommendations to the enquiry.

Context for

this submission

To deliver access to quality education governments will need to manage and regulate state and non-state providers. This is the simple reality in many developing countries where private school enrolment is high and increasing rapidly.

The performance of the private sector compared with the public sector is sometimes overstated, but it many countries it achieves similar outcomes at a lower cost. There is limited evidence to suggest that private schools are affordable to the poorest children.

The non-state sector is here to stay and, if managed well by government, has the potential to make a major impact on both quality and access.  But the debate should be driven by evidence, not by the ideologies of those on either end of the debate.



Ark EPG supports governments to design and implement PPP policies, where government commissions and funds the non-state sector, with proper oversight of quality. The non-state sector can help governments achieve their public sector objectives to improve access and quality. PPPs can assist in delivering reforms in countries where the public sector struggles to deliver high quality education to every child.

Ark EPG’s mission is to support governments to strengthen their education systems, through smarter accountability systems and effective partnerships with non-state schools. We have first-hand experience supporting governments to shape, design and implement PPPs and stronger accountability systems in South Africa, India (Delhi and Madhya Pradesh), Liberia, Uganda and Kenya.

Ark EPG recommends that PPPs only be implemented hand in hand with better data and accountability, which can help improve outcomes in PPP schools and public schools. All PPPs should be accompanied by a rigorous evaluation to build the evidence base.

Political economy issues

There can be the perception on the one hand that PPPs are a “relinquishment” of government responsibility to provide education to every child. This is false: good PPPs are government-led and are fully part of and embedded within the public sector.

On the other hand, PPPs can be seen as an easy solution to poor education systems, by “outsourcing” education provision to the public sector. PPPs are not easy and getting the public “P” in the PPP right is the key driver of success.  Building state capacity to manage the private sector is challenging and needs much attention. 

What should DFID do?

DFID could support access to better quality and more equitable education by:

  • Placing a strong focus on building up government capacity to hold PPP and non-state schools to account, which will strengthen the whole public system. DFID could leverage UK expertise from Ofsted and the Dept for Education.
  • Investing in the accountability mechanisms, for example value added data, school inspections, learning assessments, that give government far better visibility of school performance across the public, private and PPP sectors.
  • Building up the evidence base on the non-state sector and PPPs, for example through a dedicated research programme like RISE.
  • Avoiding picking winners, and instead funding operators through transparent and evidence-based competition and government PPP funding mechanisms.


1.     The scale of the challenge

Globally, access has increased but many children are still out of school.

In low income countries between 2000 and 2014, primary enrolment increased from 52 to 80 percent. However, in 2014 there were still nearly 61 million primary age children out of school globally. In low income countries, this was nearly 20 million (the same absolute number as in 1973), or 19% of all primary age children.[1]

Unfortunately being in school does not guarantee learning.

About 38 percent of children old enough to have finished primary school have not learned the most basic skills they need to succeed in life.[2] This is unsurprising given the challenges to learning in school; in many countries primary school students receive less than 2.5 hours of teaching per day.[3] By 2030, 69% of children in low-income countries – 264 million – will not have achieved basic primary level skills in literacy and numeracyand the majority of these children will be in school.[4]

Inadequate and poorly-used budgets make the challenge greater

Low income countries spend 46 percent of their education budgets on the top 10 percent most educated students. Across Sub-Saharan Africa, just 0.3 percent of national education budgets is spent on pre-primary education.
On average across low income countries, half of national education budgets – equal to two percent of GDP - are spent on education costs which do not lead to learning[5]. Estimates for the financing gap for education range from 1 to 1.7 trillion USD.[6]

…As does the low share of official development assistance (ODA).

Since 2002, the share of education in total aid has fallen from 13 to 10 percent, and only increased in cash terms from $5 billion to $12 billion. By contrast, the share of total ODA for the health sector has increased from 15 to 18 percent, a cash terms increase from $7 billion to $21 billion.

More spending alone won’t solve the learning crisis.

Whilst resources for education are low, we know from hundreds of studies that just providing “more of the same” - more time in school and more traditional inputs - will not produce the needed acceleration in learning. All manner of reforms have failed when they have not been embedded within an effective system that supports and values high performance.[7] In addition, more places will be needed, particularly at pre-primary and secondary level. In Uganda, for example, 3.6 million extra secondary places will be needed by 2032.


2.     Evidence on the role and impact of non-state schools

Private school enrolment is increasing – but most students are not educated privately.

World Bank figures show that in the thirty years between 1984 and 2014 in UN classified ‘least developed countries’, primary enrolment in private institutions as a share of total enrolment more than doubled from 6.8 to 14 percent:

  • In Pakistan around 37 percent children in primary school are enrolled in private institutions.[8]
  • In Lagos, Nigeria it is estimated that 70 percent of primary age children are enrolled in private schools.[9]
  • For low-income countries in Sub-Saharan Africa, the figure for private primary enrolment is lower - 11.4 percent. However, this is likely an underestimate, due to the difficulty of tracking enrolments in often informal, unregistered private schools.[10]

There is strong evidence to suggest teaching is better in non-state schools

The 2014 DFID rigorous review of studies on non-state schools found strong evidence to suggest that teaching is better in private schools than in state schools. This was in terms of both higher levels of teacher presence and actual teaching activity, as well as using teaching approaches that are more likely to lead to improved learning outcomes.[11]

There is moderate evidence to suggest learning outcomes are better in non-state schools

The DFID rigorous review found moderate evidence to show that learning outcomes are better in non-state schools compared to government schools. However, this comparison tends to be based on worryingly low levels of overall achievement across the entire education system and, therefore, any relative advantage associated with the non-state sector may still not be sufficiently large to alleviate quality concerns.[12]

…though the evidence is inconclusive, and weak about the effect on the poorest.

Better learning outcomes for privately educated children are partly the result of ‘selection’ effects – students coming from higher-income backgrounds. In Kenya, private school students perform 0.6 standard deviations better in reading assessments than their government school counterparts, but this drops to 0.1 SD when controlled for wealth.[13] Further, there is weak evidence on the extent to which private schools are accessible for the poorest, both in terms of geography and families’ ability to pay.[14]

The evidence on costs is unclear but suggests non-state schools cost dramatically less.

Although the cost to parents is typically higher at non-state schools, the total cost including government spending is typically much higher at government schools. Public schools in Kenya or India can cost up to 2.5 times as much as equivalent non-state schools[15][16]. This is usually due to non-state schools paying lower market-wage salaries rather than higher civil service rate salaries.

Parents often perceive private schools to be better.

There is moderate evidence to show that parents perceive private schools to be better than public schools. However, the basis for these perceptions is often a focus on inputs, such as English language instruction, and not indicators of actual student performance.


3.     Evidence on public-private partnerships

An insufficient number of rigorous studies on PPPs are available…

Overall, there are an insufficient number of rigorous studies to show the impact of public-private partnerships for education. This is in part due to the early stage and limited number of PPP reforms, and partly the difficulty of creating experimental evaluations, which can properly control for selection effects and therefore isolate the actual effect of a PPP.[17] This dearth of rigorous research needs to be urgently addressed.

The quantity and quality of evidence relating to different types of PPPs is promising but mixed.

The number of studies on contract-management style PPPs is small but provide suggestive evidence of positive effects of contract-management PPP schools on enrolment and learning outcomes. The evidence base for voucher and subsidy-based PPPs is larger. The evidence for subsidy-based PPPs is suggestive of a weakly-positive impact on increased enrolments, reaching the poorest and improving learning outcomes. There is some positive evidence to show that voucher-PPPs can be more cost effective, as they generate the same or better learning outcomes as public schools, but at less cost.[18]

There is some evidence that PPPs can work – but it is context-specific

As with the evidence on non-state schools, the key challenge in evaluating different types of PPPs lies in overcoming ‘selection’ effects. As such a rigorous review of recent evidence on PPPs found only seven high or high/medium quality studies[19]. Whilst, on balance, the evidence was positive, these seven studies covered three different types PPP schemes across just eight countries. So, whilst the evidence on different types of PPPs continues to grow it has not yet reached the stage whereby definitive conclusions can be reached on a wide scale. Greater investment in evidence is much needed.

One thing that is clear is the importance of the role of government as the ‘steward’ or duty bearer of education

Evidence suggests that education systems where the government commissions, regulates and funds education, but allows a diverse set of operators to provide education, have the potential to deliver better student performance.[20] There is moderately strong evidence that despite flaws and challenges, basic government quality assurance and certain forms of state regulation can help to facilitate effective collaborations with the private sector and improve coordination between government and the non-state sector.[21]

This can be difficult for states to do effectively.

There is moderate evidence that governments often lack the knowledge, capacity and legitimacy to implement frameworks for collaboration with, and regulation of, non-state schools.[22] There is great need to address this.


4.     Political economy and the role of government in non-state education provision

The political economy of non-state provision is complex

The issues regarding the role of non-state provision in education are many, varied and divisive: the ideological, practical and economic cases for and against the role of non-state actors in service provision; human-rights approaches to ensure equity; the influence of the profit-motive; protection of workers’ rights; fidelity to the national curriculum; and political capital deriving from providing education.

The debate too often focuses on the wrong issues

Issues such as the extent to which non-state schools may use scripted lesson plans, or employ qualified teachers, can be distracting from the fundamental objective of how states can get more children into better quality schools. The role of government should be to fund and quality-assure education for all. The non-state sector is a significant provider of education and governments need to be supported to define how to work with them to achieve their public sector education objectives.

Including opposition to private education on human rights grounds.

It is challenging to make a robust argument opposing private education on human rights grounds. Article 26 of the UN Declaration of Human Rights states that parents have a prior right to choose the kind of education that shall be given to their children. Lant Prichett notes that this appears to be a mandate for choice and PPPs (e.g. vouchers), as the only way to achieve both free and compulsory schooling and to give parents choice is to allow money to follow the student.

PPPs are a policy option with great potential as they can highlight systemic issues relevant to government schools.

The challenges to implementing an effective PPP are in fact the same challenges to delivering effectively through government schools – what objectives are set for individual schools, how is performance monitored, and what are the consequences for good and bad performance. The need to write all of this down in a contract with a non-state provider can make explicit some of the weaknesses in the system of management and oversight of government schools.

Some governments are taking active steps to better- define their role.

Some governments in developing countries are actively seeking to harness the non-state sector to help achieve their objective of improving public education. They are defining their stewardship role as a more effective commissioner, funder, regulator and quality assurer of education, and in turn are creating the environment required to encourage higher quality non-state school operators into their education systems. Engaging the non-state sector in education is not an easy way out for governments – it is not a simple matter of “outsourcing”.

Ministries of Education in Liberia, Kenya, Uganda, South Africa, Brazil, Ivory Coast, Pakistan, India and other geographies have requested that Ark EPG helps them define policies to enable them to work more effectively with the non-state sector, both through PPPs and oversight of school performance. The next section provides three case studies of countries where Ark EPG is working with government to help facilitate the development of effective, publically-led PPP.


5.     Case studies of government partnerships with the non-state sector

LiberiaPartnership Schools for Liberia

Ark EPG is the policy advisor to the Ministry of Education for Partnership Schools for Liberia (PSL). The Government of Liberia (GoL) has taken a bold step to improve their public education system. In starting a small, rigorously evaluated contract-management PPP pilot (93 public schools managed by eight non-state school operators including BRAC, Omega, Bridge, Rising Academies), GoL is testing a mechanism designed to dramatically improve the quality of students’ learning outcomes. Through this, GoL has demonstrated some of the key aspects of an effective, government-led PPP: competitive selection of partner school operators; focusing the intervention on those schools most in need of improvement; establishing performance-based contractual arrangements; developing a monitoring and accountability framework to ensure schools are delivering. A randomised control trial, led by Justin Sandefur, is underway.

Six months into the pilot, GoL has more visibility on PSL schools under non-state operation than it has ever had on its government-run public schools.

Western Cape, South AfricaCollaboration Schools

In January 2016 the provincial government of the Western Cape in South Africa launched ‘Collaboration Schools’. Ark EPG supported the design of the initiative and continues to be the policy advisor to the Western Cape Education Department. Collaboration Schools is a policy framework through which seven government schools have been entrusted to five non-state school operators under a contract management PPP. Alongside establishing the policy and contractual frameworks for Collaboration Schools, the Western Cape government has begun to improve its school data management and school quality assurance mechanisms, which will apply to schools of all status, not just Collaboration Schools. This is a good example of how establishing a PPP to improve quality in specific schools has positive spill-over effects, which help to strengthen the whole system.

UgandaPPP for Universal Secondary Education

The government of Uganda (GoU) launched a PPP for secondary education in 2007, under a student-subsidy model. More than 400,000 secondary school students are funded through the PPP. This was initially designed to improve access and in this regard has been largely effective[23]. However, the PPP has not demonstrated similar gains in quality or equity. GoU is now considering how best to redesign the PPP and deploy the non-state sector to help improve quality. In partnership with DFID Uganda, Ark EPG, funded by DFID, is working with GoU to help them review the policy and recommend how this could be strengthened to improve quality and equity.

Uganda – value added data

School performance is generally measured through absolute not progress scores. This does not account for the socio-economic status of the students enrolled, nor prior attainment. It does not give government sufficient visibility of school performance in the public or private sectors to be able to hold them accountable. Ark EPG supported the introduction of a value-added measure for secondary schools in Uganda. This will help the government hold PPP schools accountable for their performance.



6.     Recommendations for DFID

The need for more rigorous research is clear

Globally, 3 percent of international spending in education goes to research compared to 21 percent in health.[24] DFID could fund and support the conduct of more rigorous, experimentally-designed studies into the impact of the non-state schools and public-private partnerships in a variety of contexts. The current rigorous evidence-base is too small to be able to draw generalisable conclusions to inform policy across different development contexts. A research initiative similar to RISE, dedicated to the non-state sector – would be a very helpful way for DFID to show neutral, outcomes orientated leadership.

...particularly on learning outcomes.

The development of a rigorous, internationally comparable measurement of learning outcomes would be a substantial step forward. This would help by allowing programmes and interventions to be evaluated fairly and accurate assessments of impact – and cost-effectiveness – to be made.

DFID should work with governments to define a better ‘reality’ for their non-state engagement...

Substantial non-state education provision is a reality in nearly all developing country contexts, and will continue to be so. As a key development partner in education, DFID should invest in governments and focus on system-strengthening – reinforcing the notion of government stewardship of education. For example, supporting governments by assisting with the development of improved policy frameworks and regulatory mechanisms for non-state providers. DFID should focus its efforts on ensuring that the public “P” in the PPP is strong and capable more than investing in individual school operators.

…which should include the development of enabling environments…

DFID should support governments to strengthen the enabling environments around non-state operators, including: developing innovative approaches to increasing school autonomies; improving government’s commissioning processes for non-state schools; strengthening government accountability systems, including value-added data and school inspection/quality assurance. DFID is well-positioned to access the best of the English education sector to help on these issues:

  • DFID could leverage the capacity of Ofsted for school accountability and school inspections system strengthening, for example the partnership between DFID, Ark EPG and Ofsted in Madhya Pradesh, India.
  • DFID could leverage the capacity of the Department for Education to strengthen PPPs and accountability systems, for example the DFE’s support to Ark EPG’s value added data project in Uganda, or expertise from the UK Academy Programme in PPP school commissioning and financing.

… supporting rigorously evaluated PPP pilots…

DFID is well positioned to support the establishment of system-strengthening PPP programmes and pilots. These could include pilot funding for government capacity building, including data management and analysis, school quality assurance (inspection) and accountability frameworks, and for the development of a range of non-state school operators who are able contribute to the government’s public sector objectives. DFID should ensure that every PPP pilot is underpinned by core principles including: multi- not single-operator commissioning, focus on equity and quality, a rigorous impact evaluation, affordability and sustainability.

…and market-shaping interventions to improve supply.

DFID could also address the supply-side of the public-private spectrum. For example, DFID could consider:

  • Analysing what has and has not worked from programmes like DEEPEN in Nigeria or IDP’s Rising Schools in Ghana, to drive the design of new programmes that improve the performance of local non-state schools;
  • Creating ‘challenge funds’ for non-state school operators, who are able to have system level impact, for example through a PPP. These could be run internationally or nationally (e.g. through the Government of Liberia’s Partnership Schools for Liberia operator fund or an operator fund linked to a new Ugandan or Ghanaian PPP). DFID should avoid “picking winners” and should instead invest in operators through transparent and evidence-based competitions.
  • Supporting the global and national networks for school operators, for example the Global Schools Forum.
  • Consider incentivising high quality education provision through innovative financing mechanisms (results-based financing or development impact bonds).
  • Improving CDC’s education performance by ensuring that its invests in multi-operator funds that deliver better educational outcomes for the poorest children.

DFID should consider its position on for-profit education provision

DFID should consider that the UK government’s domestic education policy does not allow for-profit education provision. For-profit provision is illegal in some countries (e.g. India) and ideologically challenging in most. There is a case to be made for the benefit of for-profit provision in developing countries, for example to attract more talent, innovation and capital into the market. However, the political economy challenges related to for profit education are challenging and make it a very difficult issue to navigate. The disadvantages of DFID making direct grants to, or direct investments in, individual for-profit providers may outweigh the advantages.

The above would help refocus the debate away from ‘public vs. private’…

DFID should help to shift the global conversation away from the ideological dichotomy of the public versus the private sectors. Instead, DFID should foster more conducive debate around how the public sector can work more effectively with the private sector, and vice-versa. DFID should develop a formal position on the role of the non-state sector, including principles for responsible and effective partnerships between states and non-state actors. 






World Bank databank, accessed 19.03.17

[2] UNESCO Global Education Monitoring Report 2016

[3] ‘The Learning Generation’, The International Commission on Financing Global Education Opportunity, 2016

[4] Ibid

[5] Ibid

[6] D. Capital, ‘The Impact of Private Investment in Education in Sub-Saharan Africa’, 2015; and ‘The Learning Generation’ 2016.

[7] RISE Vision Document 2: Ambitious Learning Goals Need Audacious New Approaches

[8] World Bank databank

[9] DEEPEN 2015

[10] World Bank databank

[11] Day-Ashley et al., ‘The role and impact of private schools in developing countries’, DFID 2014. A

[12] Aslam et al., ‘Public-Private Partnerships (PPPs) in Education in Developing Countries: A Rigorous Review of the Evidence’, Ark, 2017.

[13] Justin Sandefur & Dev Patel, Center for Global Development, 2017.

[14] Day-Ashley et al., 2014.

[15] Sandefur, CGD, 2017.

[16] Pritchett and Aiyar, 2015, Value Subtraction in Public Sector Production: Accounting versus Economic Cost of Primary Schooling in India

[17] Ibid

[18] Shakeel et al., ‘The Participant Effects of Private School Vouchers across the Globe: A Meta- Analytic and Systematic Review’, University of Arkansas, 2016.

[19] Aslam et al., ‘Public-Private Partnerships (PPPs) in Education in Developing Countries: A Rigorous Review of the Evidence’, Ark, 2017.

[20] Patrinos et al., ‘The role and impact of Public-Private Partnerships in Education’, World Bank, 2009.

[21] Wales, Day-Ashley, ‘The Impact of Non-State Schools in Developing Countries: A synthesis of the evidence from two rigorous reviews’, DFID, 2015

[22] Day-Ashley et al., 2014.

[23] Barrera-Osorio et. al, 2015.

[24] Burnett, 2017, ‘Unfinished Business in Global Education’, NORRAG