A Rigorous Review of The Political Econo

Education Rigorous Literature Review
A rigorous review of the political economy of

education systems in developing countries
Geeta Gandhi Kingdon

Angela Little
Monazza Aslam
Shenila Rawal
Terry Moe
Harry Patrinos
Tara Beteille
Rukmini Banerji
Brent Parton
Shailendra K. Sharma April 2014
This material has been funded by the Department for International Development. The
views expressed do not necessarily reflect the views of the Department for International
Development. The authors are part of the Institute of Education, University of London,
and the Stanford University, California.
This paper can be found on the DFID Research for Development website:
http://r4d.dfid.gov.uk/ and the EPPI-Centre website: http://eppi.ioe.ac.uk/
The EPPI-Centre reference number for this report is 2203.
Kingdon GG, Little A, Aslam M, Rawal S, Moe T, Patrinos H, Beteille T, Banerji R, Parton B
and Sharma SK (2014) A rigorous review of the political economy of education systems in
developing countries. Final Report. Education Rigorous Literature Review. Department for
International Development
© Copyright
Authors of the review hold the copyright for the text of the review. The authors give
permission to display and print the contents of the review for their own non-commercial
use, providing that the materials are not modified, copyright and other proprietary notices
contained in the materials are retained, and the source of the material is cited clearly
following the citation details provided. Otherwise users are not permitted to duplicate,
reproduce, re-publish, distribute, or store material from this review without express
written permission.
i
Contents
Abbreviations .......................................................................................... iii

Executive summary..................................................................................... 1
1. Introduction .......................................................................................... 5
2. Theoretical framework ............................................................................. 7
3. Review methodology ................................................................................ 9
3.1 Methodology ..................................................................................... 9
3.2 Search strategy .................................................................................. 9
3.3 Search criteria .................................................................................. 10
3.4 Testing search strategy for electronic databases ......................................... 10
3.5 Describing the evidence ....................................................................... 10
3.6 Assessing the quality of evidence ............................................................ 11
4. Rigorous review of the evidence ................................................................. 14
4.1 Roles and responsibilities ..................................................................... 14
4.2 Rent-seeking and patronage politics ........................................................ 21
4.3 Educational decision-making and the process of influence .............................. 23
4.4 Implementation issues......................................................................... 30
4.5 Driving forces ................................................................................... 35
4.7 Gaps in the literature ......................................................................... 46
5. Conclusion ........................................................................................... 48
References ............................................................................................. 57
Appendices ............................................................................................. 65
Appendix 1: Authorship of this report ........................................................... 65
Appendix 2: Search Strategy ...................................................................... 66
Appendix 3: Search syntax ........................................................................ 72
Appendix 4: Describing the evidence ............................................................ 76
ii
Abbreviations
ASER Annual Status of Education Report
CMS Community-managed schools
DFID Department for International Development (UK)
EFA Education for All
ERIC Education Resources Information Centre
FTI Fast Track Initiative
GDP Ghana National Association of Teachers
GPE Global Partnership for Education
IBSS International Bibliography of the Social Sciences
INGO International non-governmental organisation
KNUT Kenya National Union of Teachers
NGO Non-governmental organisation
OECD Organisation of Economic Co-operation and Development
PEA Political economy analysis
PETS Public Expenditure Tracking Surveys
PISA Programme for International Student Assessment
QSDS Quantitative Service Delivery Surveys
RCT Randomised controlled trial
RePEc Research Papers in Economics
SADTU South African Democratic Teachers Union
SBM School-based management
SMC School management committee
TIMSS Trends in International Mathematics and Science Study
TTU Tanzania Teachers’ Union
UNESCO United Nations Economic, Scientific and Cultural Organization
iii
Executive summary
Teachers and schools do not exist in isolation of the larger world around them. Frequently,
many of their actions – and the school outcomes that they are accountable for – are
influenced by incentives and constraints operating outside the schooling system. Each of
these factors influences different aspects of education reform, whether policy design,
financing, implementation or evaluation. Given the importance of these power relations in
influencing student outcomes, there is surprisingly little literature to guide us in making
related policy decisions. One reason is that examining these issues in the case of
education may not be amenable to a particular disciplinary lens and is better served
through an inter-disciplinary approach. A key contribution of this review is to pull together
the essential literature from various disciplinary and interdisciplinary traditions and to
provide a conceptual framework in which to situate the analysis of political economy
issues in education research. Another contribution is to carefully review the existing
literature and identify research gaps in it. The review organises the literature along five
key themes and our main findings under each theme are summarised below.
Roles and responsibilities

We found that while a wide variety of stakeholder groups have roles to play, a large
empirical literature shows that teacher unions exert great influence on the shaping of
education policies. Teachers’ bargaining power stems from their ability to influence
electoral outcomes and political fortunes, their disruptive capacity to extract economic
rents, extensive geographic presence, large mobilisation capacity and ability to finance
demonstrations and sustain strikes. Evidently an important reason for teachers’ strong
bargaining power is their political power. The literature highlights that by militating for
higher salaries across the board (irrespective of individual teacher performance), lobbying
against decentralised school management and protecting inefficient and shirking teachers
from dismissal, some teacher unions contribute to educational inefficiency and loss of
local accountability. However, this is not true for all teacher unions, some of which do
engage in constructive work, including actively helping with reform and/or helping with
grievance redressal for teachers who suffer unfair treatment, provision of subsidised
credit, and promotion of professional development of teachers. In any case, negotiation,
fair treatment, constructive dialogue and cooperation with teacher unions is the only
viable route since they are critically important partners with government in delivering
education. By contrast with teachers, parents do not have a collective voice on
educational matters since they are generally not organised. While, in principle, school
accountability can be improved by involving parents in education decision-making, in
practice in many developing countries the level of parental participation and voice in
school management is extremely low. The literature also recognises the importance of
civil society groups (especially in highly politicised countries) and government officers and
international agencies as stakeholders in education whose actions impinge on the
functioning and outcomes of schools.
1
A rigorous review of the political economy of education systems in developing countries
Rent-seeking and patronage politics

Rent-seeking (by different educational stakeholders) and patronage politics are rife in the
educational set-ups within the public school sector in developing countries. The literature
concludes that clientelism, patronage and corruption are the three most intense political
forces that push states to expand access to, rather than improve quality of, education.
The politics of patronage suggests that it is more convenient to expand educational
coverage, e.g. by building more schools or hiring more teachers, than to fix existing
inefficiencies within the system because the former involves spending on political actors
whereas the latter may involve reducing resources allocated to underperforming political
stakeholders. Clientelism drives public teacher employment expansion without demanding
that these employees fulfil the responsibilities of their positions. The literature shows that
rent-seeking and exertion of political influence is also prevalent among teachers in a
patronage-based system where powerful politicians and bureaucrats oblige politically
helpful teachers with transfers of their choice, regardless of school need, which can
negatively impact on the efficiency and equity of teacher deployment. Some quantitative
literature, using stringent empirical techniques, finds that union membership is associated
with significantly reduced pupil achievement in India, Mexico and the USA. Studies from
some countries, especially India, show that teachers influence the school governance
environment not only through their unions’ negotiation with government but also in a
more far-reaching way, through their direct participation in politics, i.e. by themselves
becoming legislators. One quantitative study finds a substantial negative relationship
between teachers’ political connections and student achievement, and finds that low
teacher effort is the channel through which teachers’ political connections reduce student
achievement. The literature emphasises the importance of shoring up teacher
accountability through pupil assessments and other reforms.
Decision-making and the process of influence

A variety of groups influence the educational decision-making process and educational
change. The theoretical literature argues that access to resources affects which groups
will be able to negotiate change, and concludes that groups with low access to all
resources will be in the weakest negotiating position. The literature on the politics of
decentralisation concludes that while decentralisation is a widely advocated reform, many
of its supposed benefits do not accrue in practice because in poor rural areas the local
elite closes up the spaces for wider community representation and participation in school
affairs. This is borne out by the experiences of Mozambique, Zambia, Zimbabwe, Mexico
and Indonesia. At the macro level, research concludes that with a re-structuring of the
governance of education at a global level, international donor agencies and global
education institutions are exerting more influence on education sector decision-making in
many developing countries. A small quantitative literature examines the role of
institutional influence on educational outcomes. It presents evidence that international
differences in student performance are considerably related to institutional as opposed to
resource-level differences between countries. Some of these institutional differences are:
centralised examinations and control mechanisms, school autonomy in personnel and
process decisions, individual teacher autonomy over teaching methods, scrutiny of
students’ achievement, teacher incentive structures and competition from private schools.
Thus the crucial question for education policy is not that of more resources, but of
2
Executive summary
creating an institutional system where all involved people are provided with incentives to
use resources efficiently and to improve student performance.
Implementation issues
Much of the reviewed literature on education has analysed the causes behind policy
implementation gaps and policy failures, and it blames factors such as low state capacity,
poor administration, poor delivery system, poor governance, poor community information,
and corruption/leakages. However, underlying these is likely to be some political economy
constraint, some lack of political will or some vested interest, which hinders the reduction
in corruption or hinders better administration, governance and community information.
This may be because the politicians or bureaucrats that make the policy, or the vested
interests that lobby for that policy, themselves benefit from that policy/corruption. Worse
still, they may even have deliberately chosen to lobby/make/recommend that policy
because it gives them scope for corruption. It has been suggested in the literature that
that is why most education policies are associated with expanding access and providing
inputs to schools, which require expenditure. Some of the evidence analysing the apparent
failure of community-managed schools suggests that participatory programmes are unable
to transform how rural citizens engage with education functionaries. A plausible reason
given for this in India is the grossly unequal power relations between poor rural community
members and highly paid teachers: civil service teachers in the north Indian states of Uttar
Pradesh and Bihar are paid 15–17 times the state per capita income. Some initiatives have
tracked public expenditure to improve transparency in the management of education
resources. The Ugandan Public Expenditure Tracking Surveys (PETS) are credited with
reducing leakages of primary school funds from 74% to 20%, but subsequent studies have
not achieved the impact seen in Uganda, though they are still useful diagnostic tools.
Driving forces
Analysis in a large number of studies indicates that at the national level there are
potential ‘drivers’ or agents of change – some groups and organised interests in civil
society, non-governmental organisations (NGOs), pockets in the mass media, religious
groups, trades unions, and reform-minded elements among the political, bureaucratic and
professional elites. The literature also emphasises the importance of political will as a key
force in driving educational change. However, political will is not sufficient for the
implementation of education reforms, and the literature suggests that analysis of the role
of political will in education reform needs to be pitched at multiple levels. Where national
and local political wills are directed to the same ends in education, they can be mutually
reinforcing. Where myriad local wills are moving against the ends promoted at the
national level they are at best neutralising and at worst undermining. An interesting
strand of the literature considers regime type to be an important political condition that
impacts on educational reform. It measures the effect of regime type (democracy, degree
of openness, etc.) on educational spending and finds robust and significant effects of both
regime type and openness on different types of educational spending, showing that
aggregate public education spending increases (and private education funding decreases)
with a shift towards democracy or openness. Democracy is consistently associated with a
shift in spending from tertiary to primary education. However, the effect of additional
spending on educational outcomes is dependent on the type of democratic institutions in
place.
3
The literature shows there are several other factors that may inhibit or promote
educational reform. Multi-party electoral competition, political knowledge of the
electorate, the extent to which the elite dominates the political arena and the extent of
centralisation of governance can all be powerful forces influencing the provision of basic
educational services in certain contexts. The section on ‘Positive cases of reform’ analyses
examples where benign political economy circumstances were created by change drivers
to achieve good outcomes.
4
1. Introduction
Much hope is pinned on education to yield enhanced productivity, economic growth, social
development and poverty reduction. However, for education to deliver on these
expectations, it must be of sufficient quantity and quality to lead to meaningful learning
among young people, a task known to pose considerable challenges. One of the most
important challenges is to make the educational macro and governance environment more
conducive to reform.
The paths and outcomes of educational policies are overwhelmingly impacted by political
processes and practices. Within this context, there is an overarching need to understand
politics as consisting of ‘all the activities of cooperation, negotiation and conflict in the
use, production and distribution of resources through the interaction of formal and
informal institutions and through the distribution of private and public power’ (Leftwich
2006, p.10). More than 20 studies conducted across the developing world arrive at the
following conclusions: patrimonialism and corruption, and elite capture, is pervasive;
political parties are personalistic, there is limited political will and limited political
demand for extensive reform; commitment to overarching national strategies is weak and
there are very low levels of ‘stateness’ which generate politicised bureaucracies (Leftwich
2006). Thus, the design and implementation of effective and conducive educational (and
other) policies may be significantly influenced by the political economy within which they
are made.
Unfavourable political economy blocks policy reform and its implementation. This research
reviews the literature examining the political aspect of educational decision-making and
the manner in which the politics of the economic resources necessary for policy reform
and its implementation interact. Education reform does not take place in a vacuum, but
under specific constraints and opportunities, many of which are politically driven, shaped
by the interests and incentives facing different stakeholders, the direct and indirect
pressures exerted by these stakeholders, and by formal and informal institutions. Each of
these factors influences different aspects of education reform, whether policy design,
financing, implementation or evaluation.
The ultimate outcome of producing a skilled and knowledgeable population through good
quality education is governed by a ‘value chain’ with stakeholders both making decisions
and operating at various levels within both national and international environments. In
most countries, teacher appointment and management and promotion decisions are also
made at the ministry level. The district level (or equivalent) is usually where education
services are actually delivered and is where key actors such as teachers, headteachers and
other officials usually make decisions (such as teachers deciding how much effort to put
into teaching on a given day) which potentially impact on educational quality. Within this
chain, teacher quality is an intermediate outcome that directly impacts on educational
outcomes. There are numerous opportunities along this entire chain for unfavourable
5
leakages and corrupt behaviours to undermine efforts to achieve the ultimate goals:
delivery of educational services in an equitable, efficient and effective manner1.
Political interests play an important role not only in education, but also in other domains
such as health and infrastructure – as well as at the macro level. Indeed, this has led
researchers to adopt different levels of analysis in their work, depending on the issue at
hand. Macro approaches are typically adopted by researchers who are trying to link
outcomes (economic, health or education outcomes, etc.) to alternative institutional
structures, encompassing variation in type of elections, bureaucratic processes, legal
systems and property rights. In contrast, the political economy of education has primarily
been studied using a micro approach, since specific features of each sector – such as the
political power of teachers – play a crucial role in explaining outcomes. However, this
review focuses not only on studies undertaken in this latter tradition but also looks at the
macro approaches to political economy.
It is hoped that this review will sharpen awareness about the political economy obstacles
to education reform, and showcase how and why in some places, despite the odds, such
obstacles are overcome.
1
Draws from Patrinos and Kagia (2007).
6
2. Theoretical framework
Since the ‘new institutionalism’2 first emerged during the late 1970s and early 1980s, it
has transformed the way political scientists, sociologists and economists understand the
institutions of greatest relevance to their fields, from government agencies to business
firms. Political economy has been at the forefront of these developments, employing
theories – agency theory, theories of collective action and cooperation, rent-seeking
theory, and rational choice theory more generally – that have shed new light on the
origins, structure and performance of both public and private organisations.
The understanding of educational institutions has benefited little from these advances in
surrounding disciplines. This needs to change. Our intention here is to put the theory of
political economy to use in evaluating the research on education systems in developing
nations. Alternative approaches offer somewhat different analytical tools that are more or
less useful depending on what is being studied.
The fundamentals of our analysis are simple and straightforward and form the building
blocks of our initial theoretical framework: who are the relevant actors, what are their
interests, how are their incentives and strategies shaped by the contexts in which they
operate, how do they exercise power in pursuing their ends, and what are the
consequences for students, schools and the larger education system?3
These fundamentals are consistent with the conceptual scheme outlined by Leftwich
(2006) for understanding the politics of development. In its simplest form Leftwich’s
scheme distinguished agents/actors (organisations or individuals) pursuing interests from
institutions (which define ‘rules of the game’) and structural features of the environment
(e.g. natural and human resources, economic, social, cultural and ideological systems).
And, as noted in the Introduction, our approach is also consistent with Leftwich’s
definition of political economy as ‘all the activities of cooperation, conflict and
negotiation involved in decisions about the use, production and distribution of resources’
(Leftwich 2006, p.10).
Our review examines the interests, incentives, strategies, contexts and exercise of power
of key stakeholders in the formulation and implementation of educational decisions. We
focus on two key types of decision: those related to increasing schooling access, and those
related to improving the quality of schooling. Earlier writing has suggested that access-
oriented reforms tend to be easier to implement, since they provide citizens increased
benefits and politicians tangible resources to distribute to their constituencies, such as an
expansion of jobs for teachers, administrators, service personnel, construction workers
and textbook and school equipment manufacturers. Quality-enhancing reforms by contrast
often focus on accountability and cost-effectiveness and threaten the interests of many of
these stakeholders who in turn block their implementation (Grindle 2004).
2
Pertains to the theory that develops the role of institutions, their interactions and influence on
society.
3
For work that applies this approach to education, see, e.g., Chubb and Moe 1990, Kingdon and
Muzammil 2003, 2009, 2013, Moe 2011.
7
In the light of the above, we employ the following five themes and questions to structure
our review of the literature:
1. Roles and responsibilities: Who are the key stakeholders with an interest in the
sector? What are the interests and incentives faced by different players? Has this
varied over time?
2. Rent-seeking and patronage politics: How significant is the extent of rent-seeking
and patronage politics in the education sector, and where is it most prevalent?
What does research tell us about the impact of such behaviour on education reform
and school outcomes?
3. Decision-making and the process of influence: Who are all the participants in the
decision-making process regarding education policies of different types? What is
the identity of all those who exert indirect pressure on the decision-making
process? What are the direct and indirect mechanisms available to different power
groups to exercise their power? What are the implications of this power play for
educational outcomes?
4. Implementation issues: To what extent are policy reforms implemented and what
are the factors that facilitate and impede implementation?
5. Driving forces: What political and economic conditions drive or inhibit education
reform, both in its formulation and implementation?
Towards the end of the review the main findings under each of these themes will be
summarised and conclusions drawn. They will also be used to inform our initial theoretical
framework and to extend and elaborate it. In turn this elaborated theoretical framework
will form the basis of our ‘theory of change’ which we present schematically at the end.
8
3. Review methodology
3.1 Methodology
We followed a series of steps usually adopted in conducting a systematic review, while
acknowledging that conducting a rigorous literature review requires adopting more
flexible standards than those used in a systematic review.
Explicit inclusion/exclusion criteria were used to conduct searches of bibliographical

databases, key journals, and organisational websites. Supplemental keyword searches and
handsearches were also undertaken. Authors and experts were contacted to arrive at a
comprehensive collection of literature covering a wide range of disciplines. For example,
databases such as the International Bibliography of the Social Sciences (IBSS: covering
economics, politics, sociology and anthropology), Science Direct and Web of Knowledge
(covering all sciences and humanities), and ERIC (Education Resources Information Centre)
were used to ensure a broad coverage of all disciplines. In addition to published literature,
sites such as (RePEc), Conference Proceedings Citation Index, etc., were used to search
for working papers, conference papers and PhD and Master’s dissertations. An iterative
procedure was adopted to search for the relevant literature using a number of keywords
and synonyms to ensure the coverage of all theoretical concepts relating to the political
economy of education.
Stringent inclusion and exclusion criteria were agreed for screening the evidence base.
Included studies were characterised on the basis of features such as geographical
region/country (giving some preference to UK Department for International Development
[DFID] priority countries), appropriateness of data collection, and data analysis and study
design (qualitative or quantitative), etc.
3.2 Search strategy

1. Electronic databases: The political economy of education reform covers a range of
disciplines including economics, political science, education and sociology. As the
majority of the searches would take place online, the initial stage of the process
involved identifying electronic databases to conduct systematic searches for
published literature broadly concerned with the topic across all of the
aforementioned relevant disciplines. In addition to identifying databases containing
published literature, other databases were included to capture working papers,
theses, conference proceedings, etc. A full list of searched electronic databases
searched is provided in Table A2.1 (Appendix 2).
2. Organisational websites: In addition to databases, electronic searches were
conducted of the websites of several organisations with global reach, and with
research or technical relevance to the political economy of education reform. A
full list of these organisations’ websites is provided in Appendix 2, Table A2.2.
3. Targeted journal searches: A number of scholarly journals covering each of the
relevant disciplines were identified to conduct searches and ensure capture of
relevant literature in the defined timeframe (post-2000). A full list of included
journals is provided in Appendix 2, Table A2.3.
4. Manual searches: – Once systematic electronic searches of databases,
organisational websites and journals were exhausted, manual searches were
9
conducted by examining the references of identified papers and directly contacting

authors who frequently publish in the area of inquiry.
3.3 Search criteria

To conduct the searches, a set of search terms and keywords (and synonyms) was
developed from practical and theoretical concepts and issues related to the political
economy of education reform. Additionally, a list of countries was developed
corresponding with countries and regions of particular interest to the study, reflecting the
list of priority countries developed by DFID4. A full list of search terms, keywords and
countries is provided in Appendix 2, Table A2.4.
Searches were conducted first using topical search terms in the search title, abstract and
subject heading fields of electronic databases (Steps 1, 2, 3). The same search terms were
used in for organisational websites as well as journal searches.
Next, buckets of specific keywords (education, political, issue-based) were used to form
search strings to allow for focused results in database searches. The first bucket paired
education and political keywords into a search strings for title, abstract and subject
heading fields (Steps 4, 5, 6, 7, 8, 9). The second bucket paired political and issue-based
keywords into search strings for title, abstract and subject heading fields (Steps 10, 11,
12, 13, 14, 15). Where necessary, these searches were prefixed with education, school or
teacher from the education keywords.
In cases where the keyword searches yielded an overwhelming number of results, the
countries from the country list were added to the search strings to focus results within the
scope of the study.
3.4 Testing search strategy for electronic databases

Given the range of disciplines the study covers, an initial search string of keywords was
developed to gauge the both the volume and relevance of results from electronic
databases in the areas relevant to the scope of the study. Appendix 3 summarises the
results of these preliminary searches by (i) describing basic search syntax information for
databases that were tested,5 (ii) identifying the test search strings, and (iii) providing the
results in terms of numbers of hits.
3.5 Describing the evidence

Studies identified as meeting the inclusion criteria were analysed in depth using a
consistent and detailed data extraction methodology. A summary of the evidence is
provided through a systematic map of the body of evidence distinguishing between the
broad types of research using DFID’s (2013) guidelines as a starting point as presented in
Table 4.1 below. Each of the included studies is coded as indicated in Table A4.1 in
Appendix 4.
4
Reflects list of 28 ‘priority countries’ detailed within the 2011–12 Annual report and accounts of
the United Kingdom’s Department for International Development. The report can be found here:
www.gov.uk/government/uploads/system/uploads/attachment_data/file/67353/Annual-report-
accounts-2011-12.pdf
5
Search syntax information is adapted from the search strategy described in Kingdon et al. 2013,
Appendix 2.1: ’Are contract teachers and para-teachers a cost-effective intervention to address
teacher shortage and improve learning outcomes?’
10
3.6 Assessing the quality of evidence

In addition to excluding studies that do not meet the inclusion criteria, this rigorous
review carried out a quality assessment of all included studies by highlighting the
strengths and limitations of the studies individually and then providing an overall weight
of evidence for each of the proposed questions above.
Each individual study was assessed by at least two review members under each of DFID’s
six principles of high quality studies (DFID 2013, p.10). These six principles are:
1. Conceptual framing;
2. Openness and transparency;
3. Appropriateness and rigour;
4. Validity;
5. Reliability;
6. Cogency.
These six principles were applied to each study in a consistent and comprehensive
manner. For example, a hierarchy of evidence was used to evaluate the validity of
quantitative studies ranging from randomised controlled trials (RCTs) (high quality) to less
rigorous methodologies such as simple descriptive statistics that do not allow causal
interpretations (such as comparison of means). The validity of qualitative studies was also
analysed by choosing those that give the wider context, based on factors such as the
extent to which the study employs a methodology that minimises the risk of bias. In
relation to reliability, all studies were judged on the basis of whether their findings are
reproducible at some point in the future in the same place or elsewhere, while the
applicability of the studies was judged on the basis of whether their findings are
appropriately applicable, i.e. whether they can be applied to other low- and middle-
income country contexts. Because we came across studies with varied study designs, our
in-depth review of the studies and data extraction followed a rigorous process to allow
synthesis and comparability. Studies judged to be of low quality were excluded from the
in-depth review. However, pertinent findings from these studies may still form part of the
final review write-up in order to provide a comprehensive overview of the literature in
this field.
One of the key purposes of this review is to provide DFID policy-makers with a clear
guideline on the weight of evidence and overall strength of the evidence in relation to
each of the questions and the assumptions underpinning the literature. We describe the
overall strength of the body of evidence using guidelines provided by DFID (2013) as shown
in Table A4.2 in Appendix 4. The strength of the evidence incorporates an analysis of the
quality, size, context and consistency of the findings for each research question By
following DFID criteria, the review is able to consistently assess the strength of the
evidence supporting each of the themes and assumptions.
A major strength of this review is the authors’ own expertise and significant research
contribution to this theme, which has benefited the entire review process and the final
output. Another key strength is the heterogeneity of the types of publications, disciplines
and research designs that have been incorporated. One of the key limitations of this
rigorous review is drawing meaningful comparisons from studies which use differing
methodologies and examine different counterfactuals. Contextual factors may also have
11
hindered robust comparisons. Therefore, these will be explained since the review itself
will form an important part of the evidence base.
The review process is mapped and documented as indicated in Figure 3.1. Table A4.3A in
Appendix 3 provides details of the 50+ studies that were selected based on the more
stringent search criteria. In addition to these studies, several further reports and pieces of
evidence were included based on feedback from DFID advisors and other experts. These
studies are reported in Table A4.3B in Appendix 4.
12
6
Figure 3.1: Map of the review process
One-stage screening Two-stage screening

Papers identified in ways Papers identified where there
that allow immediate is not immediate screening,
screening, e.g. e.g. electronic searching
handsearching.
485 citations identified

54 duplicates excluded
(automatically by EPPI-Reviewer)
Title and abstract Citations excluded

screening Exclude on date=21
Exclude on relevance=265
Exclude as focus is conflict state solely=4
Exclude on location=91
Exclude on paper type=10
TOTAL Excluded - 383
48 citations (brought forward for

full-text screening)
Acquisition of reports
*48 reports obtained
Full-document
screening Reports excluded
TOTAL excluded - 17
31 studies included +
20 additions from manual searches +
13 studies based on expert feedback
There are some additional studies that

Systematic map and in-
form part of the theoretical framework
depth review
of 64 studies and general discussion but do not form
the core focus of the in-depth review.
6
Please note that some citations were excluded on more than one criterion.
13
4. Rigorous review of the evidence
This section provides an in-depth review of the final set of studies that met the inclusion
criteria. The evidence is presented based on the five themes identified previously, i.e.
roles and responsibilities; rent-seeking and patronage politics; decision-making and the
process of influence; implementation issues; and driving forces.
4.1 Roles and responsibilities

Who are the key stakeholders in the education sector?
What are the interests and incentives faced by different players in this sector?
Has this varied over time?
The political economy of education is influenced by a number of internal and external

actors. The key stakeholders in education systems within a country include inter alia both
bureaucrats and political entities: voters (parents and others), government officials
(Minister of Education, other ministry officials), local government officials (local
authority), school officials/school management (headteacher, governing body and
teachers) and teacher unions.
Teachers as stakeholders
While teachers are perhaps the most crucial input in the educational process, their
interests may sometimes deviate from societal or student interests, making them key
stakeholders in the educational process. Among the stakeholders within a country,
teachers are often also the most organised and vocal group with the power to influence
educational policy and the system in which they operate. A large empirical literature
argues that teacher unions exert great influence in shaping policies and that this effect is
not always positive (Carnoy et al., 2007, Eberts and Stone 1987, Hoxby 1996, Moe 2001).
Béteille (2009) states that unions’ bargaining power stems from their ability to potentially
influence electoral outcomes. First, they can organise an important and large component
of the electorate – teachers – and influence its voting behaviour (Moe 2006, Pratichi
Report 2002, Sharma 2009). Next, being widely dispersed geographically, they can
undertake informal campaigns for (or against) candidates, and thereby influence the
voting behaviour of the average voter (Moe 2006). Militating for higher salaries and
protecting incumbent teachers from new entrants by existing teachers often organised in
unions results in inefficiency within the teaching profession (Hoxby 1996). The political
power of teachers is one explanation behind their large bargaining power. This is the case
in many countries be they developed or developing (for example Moe 2006 for the USA,
and Kingdon and Muzammil 2003, 2013 for India).
Teachers can exert their influence through a number of channels because they vote at a
higher rate, are major contributors to political campaigns, form a dominant presence in
electoral districts, and finally are capable of causing disruption to political leaders
through enlisting their members (Moe 2006). These findings in the USA have been
corroborated by quantitative and qualitative research in Mexico (Santibáñez and Rabling
2006) and India (Kingdon and Muzammil 2003, Pratichi Report 2002).
14
The role of teacher unions and other organised groups in the allocation of educational
transfers in Mexico has been discussed by Fernandez (2011). Mexico has followed a
political logic influenced both by partisan considerations and politicians’ responsiveness to
the influence of organised groups in the education sector. This has meant that the
distribution of federal education grants has followed the interest of organised groups as
opposed to the voices of the electorate. Fernandez argues that the most influential
organised group in the education sector of the country is the teacher unions. These have
great disruptive capacity to extract economic rents, extensive geographic presence, large
mobilisation capacity and ability to finance demonstrations and sustain strikes. Teacher
unions are therefore both an attractive political ally and powerful enemy. Other authors
have also highlighted these factors as a reason why teacher unions have significant roles to
play in educational spending (Hecock 2006). This study highlights the importance of taking
public interest groups into consideration when making decisions regarding public goods
provision. In the Mexican context specifically (and possibly in other Latin American
countries), a change in government is often accompanied by a clearing out of the former
government bureaucrats and their replacement with new appointees know to be
sympathetic to the new government. In the absence of an ‘independent’ and
permanent/continuing government bureaucracy, unions may be more able to press their
interests on a new and inexperienced bureaucracy. This highlights the varying relations
between political and bureaucratic interests across political systems.
Kingdon and Muzammil (2003, 2009) present evidence of significant political penetration
by teachers in India. They highlight the role of teachers in the political process in Uttar
Pradesh to indicate how teachers have become embedded in the political system and the
way teacher associations and unions have actively pursued demands through various
strikes and other forms of actions. While teachers have been successful in demanding
improved pay, job security and service benefits, less progress has been made on broader
improvements in the schooling system such as the promotion of education in general or
improving equity and efficiency in the educational system. According to these studies, two
factors explain the dynamics of the political economy of education in India. The first
stems from the constitutionally guaranteed representation of teachers in the Upper House
of the state legislature, which has led to a culture of political activism among teachers,
many of whom wish to be elected as legislators. Secondly, while teachers in private
‘aided’ schools are government-paid employees, they are also allowed to contest elections
to the Lower House since they are not deemed to hold an ‘office of profit’ in the
government. Therefore, there is substantial representation of teachers in both the Lower
and Upper Houses of parliament. This privileged position – teachers as legislators – has
political consequences for the educational system. While in itself this may not be a
problem, the papers suggest that the issues on which teachers tend to campaign have
been more related to their own personal gains than to broader improvements in the
educational system. Teacher unions have also lobbied extensively for centralised
government management of aided schools to protect themselves from local accountability
in matters of unethical behaviour such as frequent absenteeism and giving private tuition
to their own students. However, while the case study provides good evidence on the
political power of teachers, it is also very specific to the Uttar Pradesh context (or at best
to the six states of India whose legislatures have an Upper House) where teachers are
constitutionally guaranteed representation in the state legislature.
15
Teachers’ political presence in Uttar Pradesh been significant (with on average 17% of the
membership of the Upper House and 6% of the Lower House of the state legislature being
made up of teachers, over the post-independence period). This presence has also been
increasing over time (Kingdon and Muzammil 2013). Arguably, union membership may help
teachers overcome the corruption they face in issues pertaining to their transfers,
promotions and the timely payment of salaries, etc. However, Alvarez et al.’s (2007) study
of teacher unions in Mexico indicates that unions may not either initiate or be supportive
of educational reforms aimed at improving the quality of schooling but they are important
partners for gaining support for such initiatives at the state level.
While there is little published research on teacher unions in African countries that meets
our selection criteria, one paper each from Kenya and South Africa and some reports and
journalistic writing are available. An illustration of teacher union opposition to education
reform in Africa is recorded in Bold et al. (2013). Here a randomised trial study in Kenya
showed that contract teachers significantly raised pupil test scores when tests were
implemented by an NGO but not when implemented by the bureaucratic structures of the
Kenyan government, because of teacher union opposition. The Kenya National Union of
Teachers (KNUT) waged an intense political and legal battle against the contract teacher
programme, including a lawsuit lasting a year, street protests in central Nairobi, and a
two-day national strike, demanding permanent civil service employment and union wage
levels for all contract teachers. Another illustration comes from South Africa where
Zengele (2013) finds that the South African Democratic Teachers Union (SADTU) leaders
exert great influence on teacher appointments, placing union loyalists in promotional
posts irrespective of merit. He goes on to recommend that the Department of Education
must engage the service of employment agencies to handle all the advertising,
shortlisting, interviews and recommendations for appointment processes to avoid all forms
of subjectivity and nepotism. Finally, reports show that the Ghana National Association of
Teachers (GNAT) is effective in pressing its demands before the Ghana Government by
threatening – or actually going on – strike. Some 178,000 GNAT teachers participated in an
eight-day national strike in March 2013 to pursue their demands. An Education
International report on Ghana (Education International 2013) shows that GNAT has been
able to negotiate salary increases above the cost-of-living index each year since 2010.
However, teacher unions are not homogeneous entities in all countries, or equally strong,
or similar in nature and strength. Moreover, they are not ubiquitously perceived as
interfering with school reform programmes by giving higher priority to their own ‘bread
and butter’ issues than to students’ needs. Though not on a developing country, Schleicher
(2011) in a study of Organisation of Economic Co-operation and Development (OECD)
countries finds that ‘the better a country’s education system performs, the more likely
that country is working constructively with its unions and treating its teachers as trusted
professional partners’. He believes it is possible to separate issues of collective bargaining
from professional issues, where ‘teachers and their organisations collaborate with ministry
staff in self-governing bodies to oversee work on entry, discipline, and the professional
development of teachers’. In some countries, teachers’ unions have also developed their
research capacities significantly in recent years.
Some of the literature also finds that teacher unions cooperate with government
education reforms. Languille and Dolan (2012) identify the constructive role of the
16
Tanzania Teachers’ Union (TTU) as one of several potential ‘drivers of change’. They say
that in contrast to other sub-Saharan (and indeed Asian) countries, the TTU is not
considered to be able to block or disrupt government education reform. It is viewed as
capable of playing a facilitating role in the reform process. In a similar vein, Mulkeen
(2010) finds that while teacher unions have been heavily criticised for their advocacy role
for better pay and conditions for teachers, unions are also responsible for engaging
teachers in other activities such as policy analysis and advocacy for improved educational
quality and global education campaigns. Additionally, some unions also provide teachers
with professional development, access to credit and a recourse in the case of unfair
treatment. The study finds that advocacy work was a significant part of teacher union
activities in the eight Anglophone countries it studied (Eritrea, Gambia, Malawi, Uganda,
Zanzibar, Lesotho, Liberia and Zambia), although it took different forms in each country.
Even in India, where teacher unions are perceived as a strong interest group, Kingdon and
Muzammil (2009) report that the secondary teachers’ union in the state of Uttar Pradesh
also works constructively for grievance redressal for teachers and in other ways, for
example, it acted ‘as a watchdog by drawing attention to government irregularities in the
appointment of teachers ... and also raised its concern over malpractices in the
examination system and in the evaluation of answer scripts of students. For instance it
gave the Director of Secondary Education a list of 25 schools and colleges where organised
copying was going on in Board examinations and also named teachers who had been issued
fake identity cards for invigilation and facilitating copying’.
School management, parents, elites and civil society as stakeholders

School governors and school management are also key stakeholders within the educational
system and have a key role in influencing educational outcomes particularly due to their
potential effect on accountability. There are a number of studies in the literature that
identify these stakeholders and investigate the incentives shaping their decision-making.
Alvarez et al. (2007) identifies school governors as critical stakeholders and the
relationship they have with teacher unions will play an important role in shaping education
quality. This study looks at how increased accountability and assessment (student testing,
school ranking and school report cards) leads to improved learning outcomes among
students in Mexico in that these national tests provide constant feedback to stakeholders
and can be used by authorities to design interventions to improve the quality of schooling.
However, as mentioned above, the stronger influence of teacher unions can hinder these
objectives. Context is important here. In Mexico and many other Central and South
America countries there is no tradition of testing for selection of students (as there is in
Africa and Asia). Student assessment has been introduced in the context of testing for
system monitoring, not individual student monitoring/selection. In practice, system
monitoring is perceived by teachers to be teacher-monitoring and hence is resisted. In
systems where children are tested over and over for selection/streaming purposes,
teachers may feel threatened by poor results but they can always (and usually do) pass the
blame for failure on to lack of interest and ability of students and teachers.
Chen (2011) explores the relationship among Indonesian parents, school committees,
schools and government education supervisory bodies from three perspectives:
participation and voice; autonomy; and accountability. Using nationally representative
data on public primary schools in Indonesia, the paper finds that the level of parental
participation and voice in school management is extremely low in Indonesia. Therefore,
17
while one study clearly indicates how powerful unions block reform, another shows a lack
of participation and engagement among other stakeholders as being associated with ill
functioning education systems.
The role of key stakeholders and their ability to influence educational reform based on
their incentives and interest may be lower among disadvantaged groups. Corrales (2005)
argues that the incentives and pressures for educational expansion and improved
educational efficiency are weak and sometimes even perverse for the poorest and
remotely located populations. As a result, states in developing countries are often unable
to achieve sufficient institutional capacity and political accountability to achieve universal
educational coverage. The paper highlights how patronage remains one of the strongest
incentives to expand education but is simultaneously at the root of poor quality and
inefficiency. The author suggests that citizens may be important players in the political
arena pushing for ‘demand-driven’ education, stating that the two most important
ingredients to boost societal demand (income levels and organisation) are often lacking in
developing countries among those who are the last to receive education. However, even
where societal actors may have a strong preference for more education, demand may also
falter if these actors lack capacity to pressure the state.
Little’s (2010a) argument, however, is that in highly politicised societies, such as India,
key stakeholder groups have played critical roles in pushing for reform. For example, civil
society groups in India continue to call on government to do more for elementary
education. Central government continues to appeal to state government and local
government bodies to do more. And all call upon teachers and on parents to support the
education of their children more. Poor parents also look to local, state and national
government bodies to meet the fundamental rights of their children. Meanwhile the
middle classes use private means to look after the educational future of their children.
They use this education to access growing economic opportunities in the modern sector of
the economy linked with the global economy.
The emergence of the private sector and its importance in educational delivery has been
the subject of extensive research and is a topic of another rigorous review commissioned
by DFID contemporaneously with the present review. However, in this review it is
important to highlight one key aspect discussed by Day Ashley (2013) who investigates the
politics of private education in the Indian context by discussing the immense growth of the
private sector (particularly in relation to educating disadvantaged children) and the
recognition of its role by central government (e.g. in Indian government policies such as
the Right to Education Act, the role of private provision).7 This growth of private schooling
has emerged for several reasons including constraints faced by the public schooling
system. However it raises concerns because in some situations this private provision of
education tends to be publicly funded, therefore raising concerns relating to equity (is
there a dual system being promoted and are the correct children being targeted?) as well
as efficiency (are these public funds being utilised effectively through the private
system?). In addition to this, the motives, incentives and interests of these schooling
7
There has also been an increase in the number of associations of private schools in different
Indian states and nationally to defend and champion the interests of private schools. One
prominent such group is the National Independent Schools’ Association.
18
providers then become crucial in having a stake in ensuring the funds reach the rightful
destination.
Government officials as stakeholders

Government officials are key stakeholders with significant responsibility towards the
electorate in delivering education in an equitable, efficient and effective manner. The
political will of government officials, often working in alliance with ‘policy entrepreneurs’
has been identified by Kosack (2009) to be important in relation to the provision of basic
education and Education for All (EFA) using illustrative cases from Ghana and Taiwan.
Pointing out that policies for, and progress on, quality basic education are to be found in
autocratic as well as democratic states, the author asks: ‘under what conditions might
governments be more or less likely to display political will for basic education?’ The first
part of the answer lies in the recognition that governments want to stay in power and to
do so must meet the demands of those who can overthrow them, either through
democratic elections or through military might. The second part is the recognition that
two main groups place demands on governments for basic education and help them to stay
in power. The first group are the poor who cannot meet the direct and opportunity costs
of education themselves and who rely on government and policy entrepreneurs allied with
government to provide it. The second group are employers who, under certain conditions
of labour and wages, will judge that investments by governments in primary education will
lower the wages they must pay for skilled workers (by increasing the supply of skilled
workers).
International actors and agencies as stakeholders

Finally, the literature identifies officials of international aid agencies with financial
resources and foreign advisors paid for by external agencies as members of a broader elite
that exert some influence over education policy in developing countries. And while they
may not be the most important actors, the interests of the international aid agencies have
influence through the technical lens, the bureaucratic lens and the lens of international
pressure. In Despite the odds: the contentious politics of education reform, Grindle (2004)
applies her enduring concerns with policy elites, policy choice and policy implementation
to a series of education reforms in Latin America during the 1990s. She identifies
‘international linkages’ as an important set of interests and institutions that influence
actions and choices, especially within the agenda setting arena.
The international and comparative education literature contains seminal work looking at
different actors and stakeholders in policy reform. Grindle and Thomas (1991) provide a
comprehensive overview of the role of policy elites and their interests and incentives in
shaping policy and institutional reforms. The authors also offer a number of observations
about the attitudes and behaviours of key decision-makers. The international development
community is made up of individual and groups of actors who bring to their work ideas
about development and the role of education within it. They also bring ‘identities’ which
extend beyond the national and the local host context. Crucially, international
development communities command resources in the form of finance, ideas, and
information and social networks. While some writers (e.g. Chabbot 2003, Mundy 2006,
Samoff 2003) go so far as to speak of an ‘international development regime’ and
international control and world blueprints for education, others speak of a ‘loose
coalition’ of structures, mechanisms and initiatives for education for all (Packer 2007).
19
Many if not most of those that result in action on the ground are located within countries,
while others are driven from outside a country. Undoubtedly the organisations that
command substantial financial resources for education play a dominant role but other
mechanisms play their part, including the ‘set piece’ conferences of Jomtien in 1990 and
Dakar in 2000, United Nations Economic, Scientific and Cultural Organization (UNESCO)
follow-up mechanisms, the UNESCO-led Frameworks and Plans; the Global Monitoring
Report; the Global Partnership for Education (GPE); and campaigns and a large number of
so-called global initiatives for EFA (Little 2011). Softer influence still occurs through
international transfers of educational ideas and practices. For example Sebatane (2000)
writes about the means by which education assessment ideas and practices are
transferred.
Mundy (2006) traces the role of international consensus in formulating EFA policy and the
consequent movement of international finance to support this initiative and how these
developments point towards new forms of multilateralism and global governance in the
21st century. She describes the emergence and development of an international
education-for-development regime in which agencies and development agents, whose
identities extend beyond the national and/or local, play an increasingly important role in
the determination of educational policy. Education she argues is now part of a new
consensus on global development with a broad rapprochement between the neo-liberal
and pro-economic globalisation approaches to development endorsed by the International
Monetary Fund and World Bank. Part of this has involved the establishment of a process to
produce EFA Global Monitoring Reports on an annual basis, a UN task force on gender
equity, the Fast Track Initiative (FTI) for funding EFA plans, and EFA flagship programmes
of UN agencies. New forms of donor coordination have emerged at country level that
have, inter alia, focused around the poverty reduction papers in which education and
other social development goals are integrated with plans for macroeconomic stability,
liberalisation and debt repayment, sector-wide approaches to all stages of education in
which bilateral agencies contribute to a coordinated plan and a sharing of its funding.
There have been more systematic attempts by UNESCO to follow up and support national
EFA plans. New actors and partnerships have emerged within the new regime. New types
of partnership with civil society and private sector organisations have emerged. The
growth of transnational organisations representing coalitions of civil society, sometimes
referred to as international non-governmental organisations (INGOs), is marked. Originally
viewed by the international community as an under-utilised resource in the provision of
educational services, today INGOS have taken on new and unanticipated leadership in
international EFA efforts. INGOs have asserted themselves as advocates and policy
activists.
While the number and networks of international agencies supporting education has
increased, there is scant systematic and high quality research assessing their efficacy and
contributions. Examining the effectiveness of the EFA FTI, Bermingham (2011) concludes
that although overall the FTI made a number of positive contributions to the expansion of
the education systems and provided substantial additional financial support, poor
communications between the international agencies, their competing institutional
interests and consequent delays in implementation caused serious disruptions to the
national policy processes especially in the early stages, so that instead of strengthening,
the FTI may have weakened national education reform processes.
20
In summary, the evidence has shown that government officials, teachers (through their
unions), parents, school-management committees and international aid agencies all have a
role to play and have interests and incentives relating to the education system of a given
country. However, these incentives do not necessarily align and each of these groups are
differentially capable of effecting their interests. Moreover, these stakeholder groups are
likely to exert their influence at different stages of the reform process. The research,
mostly from the USA, India and Mexico, recognises teacher unions as playing an important
role and parents as having the least power in voicing their concerns. Altschuler (2013)
suggests that improvements in accountability of community managed schools in Honduras
and Guatemala can be improved by involving parents in education decision-making. The
rent-seeking and patronage politics of these players, and their potential to be drivers of
change, may impact on school access and quality outcomes, the decision-making process
and educational reform (both in the formulation and implementation stage). These issues
are discussed under the themes below.
4.2 Rent-seeking and patronage politics

How significant is the extent of rent-seeking and patronage politics in the education
sector, and where is it most prevalent (teaching assignments, school construction,
textbooks and so on)?
What can we say about the impact of such behaviour on education reform and school
outcomes?
Rent-seeking (an attempt to gain economic rent by influencing policies or their

implementation) and patronage politics (support provided to specific groups, for instance
rewarding individuals or organisations for their electoral support) are rife in the public
sector in developing countries, including in the education sector. Corrales (2005)
reviewing theoretical and empirical political science literature from developing countries
argues that clientelism, patronage and corruption are the three most intense political
forces that push states to expand education. It has been observed that private schools
tend to invest more in class-based inputs (instructional materials) whereas public schools
invest more in external resources such as wages and procurement. The latter, the author
argues, are typically driven by patronage. Corruption is also seen to go hand-in-hand with
misguided educational investments. Clientelism drives public employment expansion
without demanding that these very employees fulfil the responsibilities of their positions.
Therefore, patronage undermines the legitimacy of the state, magnifies the power of
vested interests, lowers the quality of state provision of services and hinders the impact of
social policy. Patronage can protect spending on salaries; however it does not initiate
expenditures on non-salary items necessary for effective education such as training,
infrastructure maintenance and facilities. Writing about corruption in the provision of
education and healthcare services, Gupta et al. (2000) find that corruption not only
depletes overall resources but can also distort the composition of government
expenditure. The politics of patronage suggest that it is more convenient to expand
educational coverage, e.g. by building more schools or hiring more teachers, than to fix
existing inefficiencies within the system because the former involves spending on political
actors whereas the latter may involve reducing resources allocated to underperforming
political stakeholders.
21
The universality of public education may also be affected by rent-seeking (Gradstein

2003). Due to credit market imperfections, richer households may be able to exert more
political pressure through rent-seeking which results in them being able to secure a larger
share of the educational pie than poorer households. The extent to which such rent-
seeking matters within the educational resource allocation mechanism is in itself a
political decision. Using a theoretical model, the author presents the view that political
bias induced by an unequal income distribution may well generate political support for
social exclusion whereby rent-seeking determines the extent of access to public
education.
Béteille (2009) in her dissertation discusses how rent-seeking and exertion of political
influence is prevalent among teachers in many developing countries. She cites studies that
show that teachers often use their influence in relation to salary increases and issues
relating to recruitment and redeployment. There is some evidence on the prevalence of
discretionary and patronage-based appointments and transfers in public office (Iyer and
Mani 2008, Park and Somanathan 2004, Ramachandran et al. 2005, Sharma 2009, Wade,
1985). As a large evidence base indicates that working conditions and teacher satisfaction
are key to retention and motivation, understanding teacher transfers is important because
transfers alter teachers’ working conditions. However, in this regard there is a dearth of
high quality quantitative and qualitative literature on teachers in developing countries
with only a handful of quality studies investigating this issue (Beteille 2009, Kingdon and
Muzammil 2003, 2013). Sharma’s (2009) work in particular shows how in India many states
do not have stable and transparent transfer policies. A broad qualitative literature cited
by this author suggests that transfers are typically kept discretionary and conducted on
subjective criteria. They allegedly form the bedrock of a patronage-based system where
powerful politicians and bureaucrats oblige politically helpful teachers with transfers of
their choice, regardless of school need. From a school’s perspective, such discretionary
behaviour potentially distorts the overall allocation of teachers to schools and can
potentially negatively impact on the efficiency and equity with which teachers are
deployed8.
Unlike South Asia, Bennell and Akyeampong (2007) believe that poor teacher
accountability is less of a problem in sub-Saharan Africa partly because teachers in this
region are less heavily involved in party politics and because patron–client relations are
not as endemic in Africa as they are in South Asia. They note that while teacher
absenteeism is high in most of the country studies, only a relatively small proportion of
these absences in African countries are categorised as non-authorised.
Citing the African example of Nigeria, Duncan and Williams (2010) say that political
parties are not guided by ideas and programmes but are ‘machines driven by personalities
and patronage’. Thus competition for political power ultimately depends on managing
patronage relationships (p.10). However, within these limitations there are still many
passionate voices calling for a higher quality of electoral democracy.
In a striking example of patronage politics in education in Africa, Zengele (2013) discusses

how the teacher unions have taken over teacher appointment and promotion decisions in
South Africa, in a manner similar to that observed in Mexico. Teacher union leaders were
8
Draws heavily from Béteille (2009).
22
awarded high government posts for supporting the African National Congress in South
Africa’s first democratic elections in 1994. Since then, SADTU representatives have been
given observer status in the panel that appoints teachers but, as the author’s interviews
with union leaders show, when vacancies arise, these representatives pro-actively use
their political muscle (their closeness to senior political leaders) to influence
appointments in favour of union-backed candidates, regardless of merit.
Patronage politics may also partly explain some politicians’ dislike of student assessments.
Alvarez et al. (2007) using Programme for International Student Assessment (PISA)
achievement data from Mexico find that improving teacher accountability through
assessments of student learning outcomes is a cost-effective strategy. However,
governments often do not like assessments that show poor student learning levels, e.g. in
India and Tanzania the education ministries do not recognise the non-governmental
organisation- (NGO-) conducted Annual Status of Education Report (ASER) uwezo student
tests which show lamentable learning levels. The dissemination of assessment findings
leaves the government exposed if students perform poorly; it spotlights what some
consider politically expedient government policies such as increasing teacher–pupil ratios
(through increased teacher appointments) and increasing teacher salaries, i.e.
expenditure on teacher inputs that have little demonstrated benefits for student learning9
but which dispense patronage. Another reason for opposing testing may be that findings of
poor student achievement hurt national pride and self-esteem. Examples of this include
India’s decision not to participate in future PISA tests for at least ten years, after it
ranked 73rd out of 74 countries in the 2011 PISA test. The emergence of non-state
providers and the increasing role of private tuition has also altered the political climate
within which examination reforms are designed and implemented. It has been argued that
the private sector providers do not always support reforms that may be in the students’
best interests. In many countries, private tutors, for instance, may resist examination
reforms designed to introduce continuous assessment on the grounds that this may
‘undercut’ their business which is to help students cram for one-off unseen written
examinations. This is especially an issue when these private education providers are
former public sector officials or school headteachers and teachers. A study by Silova and
Brehm (2013) argues that teachers from Southeast/Central Europe and Southeast Asia who
engage in private tutoring activities have often used this newly created private space to
‘evade and perhaps even defy multiple (neo)liberal regulations permeating their work in
public schools, such as student-centred learning and curriculum standards.
4.3 Educational decision-making and the process of influence

Which groups partake in the decision-making process regarding education policies?
Who exerts indirect pressure on the decision-making process?
What are the mechanisms available to different power groups to exercise their power?
What are the implications of this power play for educational outcomes?
9
See meta-analyses and overviews of developing country studies by Hanushek (2003) and Glewwe
et al. (2011), and see a study on 47 countries (Altinok and Kingdon 2012) and one on India (Kingdon
and Teal 2007) for evidence on the relationship between teacher–pupil ratios and teacher salaries
on the one hand, and student achievement on the other.
23
Groups that influence the educational decision process, and the implications for
educational outcomes
A variety of groups influence the educational decision-making process. Archer (1981)
provides a useful conceptual piece distinguishing three types of educational politics. The
first type is broad educational politics: attempts, conscious and organised to some degree,
to influence the inputs, processes and outputs of education, through legislation, pressure
group action, union action, experimentation, private investment, local transactions,
internal innovation or propaganda. Broad politics can explain educational operations at
any given time and the dynamics of educational change over time, at the systemic level.
The second type is high educational politics: the analysis of interpersonal relations at
government and local government levels. The third type is the politics of aggregation: the
sum of individual decisions – to, for example, go to school, leave school, apply for
university. Using this framework, Archer (1981) describes three types of transaction. The
first is internal initiation by education professionals. Change is introduced from inside the
system by education personnel, possibly in conjunction with pupils or students. This type
includes both small-scale personal initiatives in a particular institution and large-scale
professional action. The second type is external transaction: this involves relations
between internal and external interest groups. This type of transaction is usually
instigated from outside education by groups seeking new or additional services. A third
type is political manipulation by political groups. This is the principal resort of those who
have no other means of gaining satisfaction for their educational demands. This form of
negotiation arises when education receives most of its resources from public sources.
These three forms of negotiation add up to a complicated process of change.
Using the above framework, Archer (1981) argues that access to resources affects which
groups will be able to negotiate change: the greater the concentration of resources, the
fewer the number of parties who will be able to negotiate educational change. At all
times every educational interest group will have a place on the hierarchical distributions
of wealth, power and expertise. She offers the following propositions: groups with low
access to all resources will be in the weakest negotiating position, groups with differential
access to the various resources will be in a stronger negotiating position, groups with high
access to all resources will be in the best negotiating position, and groups in this last
group are most likely to introduce significant educational changes. The crucial relationship
is between the position of the educational interest groups and the availability of the
resources themselves.
The most well-positioned and organised interest group exerting pressure on the
educational decision-making process is usually teacher unions. Kingdon and Teal (2010)
state that conceptually there are two main reasons why teachers may become union
members ( as also indicated by Hoxby 1996). The first is that they maximise the same
objective function as parents, namely student achievement, but have superior information
about the correct input mixes, and union membership provides teachers with a collective
voice to implement these input mixes. This may include, for instance, asking for smaller
class sizes or higher salaries, which helps to attract and retain superior teachers and
which, therefore, helps improve student achievement. The second potential reason for
teachers joining a union is that they have a different objective function than parents or
school management, possibly one in which school policies that directly affect them, such
as teacher salaries, receive greater weight than policies that only indirectly affect them,
24
i.e. membership of a rent-seeking teachers’ union. A rent-seeking union may block reform
of incentives to improve instruction, e.g. by tying salaries to seniority rather than to
performance and by protecting ineffective teachers from dismissal. Under rent-seeking,
unions may also lower student achievement if their pursuit of higher salaries diverts
resources away from other school inputs that raise achievement and if teacher union
strikes disrupt teaching. Finally, since teachers interact with other inputs in order to
produce education, rent-seeking unions could lower the efficiency of the other inputs,
such that more money for schools may not matter (Hoxby 1996). For the above reasons,
the sign of the relationship between teacher union membership and student achievement
could go either positive or negative, and is thus an empirical question.
The literature investigating the implications of vested interests on educational access and
quality has largely emerged through the economics discipline. More recent literature has
focused largely on teacher unions and the potential negative impact on educational
quality as measured by students’ learning outcomes (Hoxby 1996, Woessmann 2003).
Kingdon and Teal (2010), for instance, examine the relationship between teacher
unionisation, student achievement and teachers’ pay, with data from 16 major states of
India. Using stringent empirical techniques (pupil fixed effects regression – where a pupil’s
achievement across different subjects is related to the characteristics of the different
teachers that teach those different subjects), the authors find that union membership
strongly reduces pupil achievement. In addition, union membership is shown to
substantially raise pay. Thus, unions are seen to both raise costs – within a school, a union
member earns a wage premium of 14.9% over non-union members – and reduce student
achievement. Most recently, Lott and Kenny (2013) provide quality evidence from the USA
that indicates that students in states with strong teacher unions have lower proficiency
rates than students in states with weak unions.
Kingdon and Muzammil (2013) explore how teacher politicians and teacher unions
influence school governance by presenting evidence on the political penetration of
teachers, the activities of teacher unions and the stances of teachers’ organisations on
various decentralisation and accountability reform proposals over time in Uttar Pradesh,
India. Using a different dataset to Kingdon and Teal (2010), they ask how student
achievement varies with teachers’ union membership and political connections. Teacher
effort is likely to be greater in governance systems where there is a good system of school
and teacher accountability. However, teachers may not be only passive accepters of that
wider ‘school governance’ environment; they may also consciously shape it to achieve
certain working conditions that determine their effort levels. Teachers may influence that
environment through their organisations (unions’ negotiations with government) and,
possibly in a more far-reaching way, through their direct participation in politics, that is,
as teacher legislators who have a say in education-related legislation. The authors find a
substantial negative relationship between teachers’ union membership and student
achievement, as well as between teachers’ political connections and student
achievement. A student taught by a teacher who is both a union member and politically
connected has about a 0.20 standard deviations lower score than his/her counterpart in
the same school who is taught by a teacher who is neither a union member nor politically
connected. Kingdon and Muzammil (2013) find that low teacher effort is the channel
through which teachers’ political connections reduce student achievement; however, low
25
effort could not be confirmed as the reason for the lower achievement of students taught
by unionised teachers.
In some instances, studies have shown that the negotiation process between unions and
governments can be so prolonged that it significantly undermines the achievement of
educational goals. Mahlangu and Pitsoe (2011) argue that the power struggle between
government and union negotiators in South Africa undermined the process of negotiation
with one or both parties negotiating in bad faith with the result that public education has
declined and poverty has increased and political tactics have been used to obscure the
true problems.
What mechanisms are available to different power groups?

Murillo et al. (2002) state that the impact of unions on student performance depends on
the channels and kinds of political markets in which unions operate, and not necessarily on
the existence of the union itself. They highlight key channels that have been used by
unions in Argentina to influence decision-making. First, union characteristics have an
important effect on teaching days lost to strikes; as teaching days are one of the stronger
explanatory variables in determining student outcomes, one could argue that unions have
a negative impact in this regard. However, unions would counter this argument with the
view that strikes are instrumental in improving teachers’ working conditions and education
budgets and therefore improve learning outcomes. This paper does not confirm or refute
the counter-argument, but it does show that the means used by unions to have their
demands met have strong and negative effects on student learning. The second channel
relates to the tenuring of teachers which has an uncertain net effect on learning because,
while tenured teachers may display higher student outcomes, they are also seen to be
more absent. While unions may be able to increase the share of salaries in the education
budget, there is no strong union effect on increasing public expenditure on education or
on increasing teacher salaries. The authors present evidence that union strength is
positively correlated with lower pupil–teacher ratios and that union participation and job
satisfaction are negatively correlated. An additional channel through which unions can
influence decision-making is discussed by Santibañez and Rabling (2008). The authors find
that the number of teachers increase (despite no equivalent rise in child population
numbers) with every election period which, indicates the union’s influence over elections
and the concessions it is able to obtain from the government in exchange for increasing its
membership. In another paper, Murillo and Ronconi (2004) empirically analyse teacher
strikes and increasing militancy of teachers across the 24 Argentine provinces during the
1990s.
Similarly, Santibáñez and Rabling (2006, 2008), using Mexico as an example, suggest that
union strength alone cannot fully explain the relationship between unions and educational
quality. There are other factors such as union fragmentation and political alignment with
mainstream factions. These factors in particular seem to be associated with student test
scores. Additionally, a union’s influence, when measured by things such as whether
teachers constitute a more highly paid group in the state (similar to professionals),
appears to affect test scores and increases or attenuates the effects of union strength.
State-level factors also play an important role; for example, whether an effective
accountability system has been implemented, whether the state has instituted progressive
26
reforms (such as competitive teacher recruitment processes), etc., all appear to impact
on student outcomes.
Other studies have examined how the mechanism of the power play between and within
teachers and politicians has influenced decision-making within weak accountability
systems. Béteille (2009) argues that rapid school expansion in India and in much of the
developing world has had to contend with limited financial resources and poor
accountability measures. Allocating limited resources across competing uses is itself
difficult, but when these resources are misused and political pressures undermine
educational accountability mechanisms, universal schooling programmes are severely
compromised. The two key accountability problems discussed by this author are
widespread teacher absenteeism and the manipulation of teacher transfers, and how
these are influenced by political factors. Using representative primary data from
government school teachers in seven district-level teacher labour markets in India, the
author finds strategic linkages between teachers and politicians which potentially
complicate policy attempts at influencing teacher accountability. Evidence suggests that
teachers who are politically active are also more likely to be absent. This lends support to
the theory that at least some teachers believe they can get away with absences because
they are protected by powerful connections. The manipulation of transfers suggests
another type of accountability breach because it involves the circumvention of formal
rules. Evidence in this regard suggests that transfers are typically characterised by
informal transactions between teachers on the one hand, and politicians, bureaucrats or
politically connected people on the other. This undermines the ability of the system to
function along professional lines and by official criteria. Patrinos and Kagia (2007) confirm
the existence of these political dynamics among teachers and suggest some solutions as
highlighted in the conclusion of this review. Iyer and Mani (2008) argue that the power
play between politicians and bureaucrats is also a factor generating significant
inefficiencies within the systems in developing countries.
Building-in incentives into the teacher salary structure

High quality studies in recent years have indicated that building incentive mechanisms into
teacher salary structures can help improve student outcomes. For example, Muralidharan
and Sundararaman’s (2011) evidence from a randomised evaluation of a teacher incentive
programme in Andhra Pradesh, India, shows that implementing teacher performance pay
in government-run schools led to significant improvements in student test scores, with no
evidence of any adverse consequences from the programme. Additional school inputs were
also effective in raising test scores, but the teacher incentive programmes were three
times as cost-effective. Duflo, Hanna and Ryan (2010) also find that attendance-related
bonuses to teachers boost both teachers’ school attendance rates and student learning
outcomes in Rajasthan, India. Bruns et al. (2011) assess evidence on two kinds of policies
that can be used to make teachers more accountable: contract teachers and performance-
related pay. The evidence on the short-term impact of contract teachers is fairly
consistent and strong in that it indicates that the use of contract teachers can improve
accountability, resulting in higher teacher effort, which produces equal or better student
learning outcomes than for regular teachers, and at a lower cost per student. However,
these findings are subject to the caveat that learning was assessed on a very narrow basis
and contract teachers were paid very low wages. It is often not clear in these contexts
how many teachers stay in the system and how long they are prepared to stay as fixed-
27
term contract teachers before migrating to more secure state systems. However, generally
speaking, teacher bonus initiatives in developing countries (in contrast to previous US
evidence) suggest that bonus pay incentives can work particularly in some contexts. Bruns
et al. (2011) also seem to indicate that information on accountability has the potential to
be a powerful means of improving school quality in developing countries but this is
strongly dependent on context. However, strong opposition to such reforms and elite
capture may significantly undermine these initiatives as seen in their review of India.
There is also the possibility that when combined with other interventions, there is a
further aggravation of existing inequalities.
The politics of decentralisation in education

The politics that surround decentralisation in education may be especially problematic
when accountability systems are weak, and there is little parental information or
awareness of how to hold schools responsible. In this regard, political economy issues arise
when politicians, for instance, give school management committees (SMCs) a role in
theory but not in practice (i.e. by giving them very few funds or powers) in order to
devalue accountability checks and balances. Several studies have shown how this plays out
in practice. Chen (2011), examining Indonesia’s schooling system, states that there
appears to be more emphasis on top-down and central supervision and monitoring as
compared to school-based management (SBM) and decision-making with the latter having
been found to have had positive effects on learning when measured using test scores. The
author finds that despite frequent interaction between the district and school level, the
lack of rewards and sanctions related to performance leave the system weak. The author
attributes this to the differing levels of power and voice among the actors within the
educational system as well as the interplay between asserting rights and enforcing
responsibilities to ensure the delivery of positive educational outcomes. Increased
autonomy at the school level translates into greater efficiency based on the idea that
those who work at the school level and are involved in the day-to-day operation of the
school have greater knowledge and management control of the needs of the school and
therefore have a better ability to make decisions that are productive in terms of academic
results and efficient spending (Barrera-Osorio et al. 2009). Bruns et al. (2011) also assess
evidence regarding SBM and conclude that these reforms generally take time to produce
visible outcomes. Several studies present evidence that SBM has a positive impact on
repetition rates, failure rates and drop-out rates. Some studies have found SBM to change
the dynamics of the schools but studies that have access to test scores have presented
mixed evidence.
Several authors have examined how decision-making and the process of influence operate
both when decentralisation is an example of a particular type of policy reform and when
the process of decentralisation influences the process of making policy decisions.
Reforming critical educational issues is dependent on context (Buchert 1998). In some
countries, there may be a need to focus on educational quality, while in others on access,
equity and/or efficiency, etc. However, Buchert (1998) states that the implications for
both government and external actors appear to be the same in that ‘critical support must
come from below, and successful and sustained reforms must rely on local rather than
central and on national rather than international initiative and determination’. Donor
supported initiatives for school improvement implemented from above appeared to
undermine autonomous local initiatives in Mozambique, Zambia and Zimbabwe.
28
Decentralisation in Mexico involved transferring the operation of an underfunded

education system to those who deliver and receive the education service with the
consequence that while there is equity in educational distribution, the quality of
education provided remains low and inadequate. This view is corroborated by Alvarez et
al. (2007) who state that school-level autonomy and school governors as well as the
relationship of all actors with teacher unions is critical in facilitating reforms.
Essuman and Akyeampong (2011) claim that while international decision-making and
power play may advocate decentralisation, much of that policy advocacy is not borne out
in practice. In poor rural locales in Ghana, the local elite and relatively more educated
community members tend to take on the role of being the new brokers of decision-making
and, through their actions, close up the spaces for representation and participation by
community members in the affairs of schools. Additionally, ‘social contracts’ based on
principles of reciprocity of roles between the community and schools determine the
extent of community-engagement in that increasingly teachers feel accountable to the
traditional hierarchical educational structure, and not to the community. The paper
argues that the realisation of decentralisation policy in education has to contend with the
realities of local politics of influence in the community, and tap into the positive side of
this influence to improve education service delivery. In a similar vein Mulkeen (2010) cites
several examples: from Lesotho, where community pressure has resulted in schools
employing a local person in preference to a better qualified outsider; from the Gambia
and Uganda, where he reports that headteachers find it difficult to take disciplinary
action against teachers living in the community; and from Uganda where he shows that
teachers working in their district of origin were more likely to be absent, by 3.5
percentage points.
At the international level, decision-making and power play may be undertaken in more
covert ways. Mundy (2006) concludes that recent years have seen the opening up of an
important and active phase in the re-structuring of governance of education at a global
level. It is a phase that will almost certainly involve the redefinition of the appropriate
scale, modes and extent of global action in the field of education. With all its limitations
and diverse interpretations, universal public access to free basic education has now
achieved status and legitimacy as a global public good not realised during the 20th
century.
Woessmann (2003) examines the Trends in International Mathematics and Science Study
(TIMSS) database of more than a quarter of a million children from 39 countries, some of
which are developing countries (such as Iran and Thailand), and presents evidence that
international differences in student performance are markedly related to institutional as
opposed to resource-level differences between countries. Among the many institutional
elements that combine to yield a positive effect on student performance are: centralised
examinations and control mechanisms, school autonomy in personnel and process
decisions, individual teacher influence over teaching methods, teacher unions’ influence
on curriculum scope, scrutiny of students’ achievement, and competition from private
schools. The author finds that student-level estimates reveal that, while the differences in
the incentive structures determined by the institutional features of the education systems
strongly matter for student performance, not all of the individual institutional effects are
particularly large. However the large number of institutional effects combine to yield
29
important consequences for institutional arrangements on students’ educational

achievement. The study reveals that about two-thirds of the variation in average TIMSS
scores at the country level can be accounted for once institutional effects are recognised.
The author therefore suggests that the policy measures which particularly matter for
schooling output do not seem to be simple resource inputs and that spending more money
within an institutional system which does not set suitable incentives does not promise to
improve student performance. What does seem to be the crucial issue for education policy
is not that of more resources but of creating an institutional system where all the people
involved are provided with incentives to use resources efficiently and to improve student
performance.
4.4 Implementation issues

To what extent are policy reforms implemented as per design?
Is this failure of technical design, failure of capacity, or a failure due to political
economy issues?
The literature on education reform and its implementation within the international and
comparative education discipline is wide ranging, and its concerns/focuses have also
varied over time. From the 1960s to the present, educationalists have identified myriad
characteristics and processes that influence the formulation of reform programmes and
their implementation in developing countries (for a review see Little 2008). Most
evaluation studies conducted in the 1960s and 1970s focused on the implementation of
relatively small-scale educational innovations on the ground. They also focused more on
why innovations failed rather than on why they succeeded (e.g. see the review by Hurst
1983). Lewin (1991) identifies six broad approaches to educational policy and
implementation: systems, bureaucratic, scientific, problem solving, diffusionist and
charismatic. Through an analysis of 15 case studies from North America and Europe,
Fullan’s (1989) review recognises the importance of political factors in the
implementation of education policies. It identifies four main determinants of the
implementation of educational innovations – the clarity and complexity of the innovation,
strategies employed (in-service training, resource support, feedback and participation),
the characteristics of the adopting unit (the decision process, organisational climate,
environment-support, demographic factors) and the characteristics of external systems
(design, incentives, evaluation and political complexity). In doing so, it highlights the key
political factors aiding or hindering educational policies.
A second phase of evaluation studies, undertaken from the late 1970s to the late 1980s
focused more on successful educational innovations, designed to impact, variously, on
student achievement, teacher behaviours and commitment. Again, most of the studies
reviewed by Fullan (1989) are drawn from North America. Fullan (1989) offers a
conceptual framework for the process of change, focusing on education policy
implementation.
This snapshot of some of the education evaluation literature highlights a wide range of
factors involved in the implementation of educational innovations and reform that would,
in principle, need to be considered in the development of a theory of change linked with
any education intervention. Within these frameworks, political factors are noted as one of
several that impinge on education policy implementation. And for the most part much of
30
the education evaluation literature from the period to the late 1980s focused more on
national and intra-national factors and processes than on interactions between these and
international factors, including foreign aid.
Problems of education policy implementation also formed the foundation for a programme
of work by political scientists. Grindle’s early edited work (Grindle 1980) focuses on the
characteristics of policies and programmes the impact of political regimes on policy
implementation. Although her early work does not focus on education, she describes how
the implementation of education policy and monetary policy implicates different groups of
actors with different interests. The implementation of monetary policy involves a limited
number of decision-makers in the national capital. The implementation of education
policy by contrast relies on a large number of widely dispersed individual decision-makers.
Where Grindle’s early work (1980) focused on policy implementation during the 1960s and
1970s, the work by Grindle and Thomas (1991), Public choices and policy change: the
political economy of reform in developing countries, focuses on policy formulation and
the scope for the exercise of choice by ‘decision-makers’.
Psacharopoulos’s (1989) study claims that insufficient implementation is the key factor
behind the failure of educational policies in East Africa. The study reviews a number of
educational policy statements from East African countries. An assessment is made of how
successful the policies have been in achieving their original intention. The study concludes
that policy outcomes fail to meet expectations mainly because of insufficient or zero
implementation. This, the author argues, may be either because the designed policies are
not sufficiently ‘concrete’ and are often vague or because they are not based on research-
proven cause and effect relationships (p.193). This is reiterated by Somerset (2011) who
argues that initiatives to achieve universal primary education are unlikely to succeed
unless the tension between access and cost, and its implications for quality, are
recognised and taken into account. If the programme does not incorporate viable plans to
meet the additional costs and prevent quality being compromised, its prospects will
almost certainly be in jeopardy from the outset. Thus, insufficient planning (vaguely
stated policies and a lack of planning for the financial implications of proposed policies)
appears to be one of the key reasons for the failure of educational policies. Another
reason is that the content of the policies is based on empirically unsustained theoretical
relationships between instruments and outcomes. The author calls for the formulation of
concrete, feasible and implementable policies based on documented cause and effect
relationships rather than weak and ill-designed ones that are doomed to fail.
The need for education policies to be formulated based on strong evidence is emphasised
by Kremer (2003) who argues that political factors undermine the usage of randomised
evaluations which are considered a more superior evaluation tool (RCTs are considered the
‘gold standard’ in evaluating/testing the efficacy of medicines in the pharmaceutical
industry, before a medicine is cleared for human use) and may be ‘far cheaper than
pursuing ineffective policies’. One explanation for this political influence is that
politicians who wish to promote/advocate a particular educational programme
systematically mislead swing voters into believing exaggerated estimates of programme
impacts. Advocates block randomised evaluations since they would reveal a programme’s
true impact to voters. The author proposes a complementary explanation in which swing
voters (or policy-makers) are not systematically fooled, but simply have difficulty gauging
31
the quality of evidence. Thus, stakeholders’ incentives and interests as well as the
inability to effectively gauge the quality of evidence may undermine the development of
concrete and implementable educational policies.
More recent work reviewed here has looked at the issues surrounding the implementation
of policy reform across a range of developing countries. Some of this evidence stems from
conclusions drawn from a review of specific programmes. An example of this is the study
by Altschuler (2013) whose analysis of community-managed schools (CMS) in Honduras and
Guatemala indicates that part of the reason for their failure is that while participatory
development programmes are capable of stimulating parents’ individual engagement in
community life, they are unable to effectively transform how rural citizens organise and
engage with the state. States must do more than just pay lip service to increasing citizen
participation and there is a critical need for material support, especially for training, and
for allowing greater organisational autonomy if effective implementation is to be
achieved. In a similar vein, Rawal and Kingdon (2010) suggest that a plausible reason why
community participation in education (through community/parental representation in
school development committees) is ineffective in India is that the power relations
between poor community members and highly paid teachers are grossly unequal. They cite
Kingdon (2010) which shows that civil service teachers in rural north India are paid 10 to
15 times the average per capita income of the community members in the village (also see
Table 5.4 in Dreze and Sen 2013, p.133), and this is due to the Sixth Pay Commission’s pay
recommendations which nearly doubled teacher pay in one go in 2009. Kingdon (2010)
shows that while the Pay Commission, in making its recommendations for pay increases,
took into account the views of teacher unions, it did not take into account the
implications of the massive pay inequality that their recommendations would engender.
A study by Evans et al. (1995) presents six case studies of education reform in Africa, in
Benin, Ghana, Guinea, Mauritius, Mozambique and Uganda. Two main themes emerge from
the comparative analysis of case study evidence: the need for publicly stated education
policies understood and supported by both government and civil society, and the
importance of participation by the diverse groups in society who will be affected by the
policies. Moreover, the authors acknowledge that, in practice, the process of policy design
and implementation is an interactive rather than a discrete process. Because stakeholders’
incentives and interests shape policy reform at all stages, the implementation of a policy
is affected by the various changes that occur at different points even after a particular
policy option is adopted. The authors argue that implementation is hindered and
sometimes blocked even well into the implementation stage when mid-level bureaucrats
and school personnel seek to influence and resist the translation of policies into effective
regulations and actions.
Patrinos and Kagia (2007) note that the failure of effective implementation of educational
investments lies partly in ill-targeted policies and the misallocation of public spending
which is insufficiently focused on quality. These may arise either as a result of poor
capacity or poor governance, though the ultimate outcome is reduced education
effectiveness. Education effectiveness is also reduced when spending decisions are
improperly guided, i.e. when decisions are not based on information, tools and
mechanisms that improve outcomes. Corruption in education is particularly important
because the sector usually accounts for a large share of public expenditures. Corrupt
32
education practices across the world contribute to inefficient use of resources. Many
education stakeholders argue that the Millennium Development Goals for education will
not be achieved without developing and strengthening the instruments needed to control
corruption in education (Transparency International 2005, cited in Patrinos and Kagia,
2007). The authors cite several examples where corrupt practices in education not only
result in a less-than-optimal allocation of limited government resources (for example
through reduced spending on key inputs such as textbooks) but also undermine access,
quality and equity in education.
Educational financing is deemed to be high quality when the systems in place can be
judged to be adequate, efficient and equitable. Internationally there have been several
initiatives used for tracking public expenditure to improve transparency in the
management of education resources. Foremost of these are Public Expenditure Tracking
Surveys (PETS) first developed in Uganda in the 1990s that trace funds through the
different administrative structures managing those funds. The Ugandan PETS study is cited
as best practice in anti-corruption literature due to the success it apparently achieved in
reducing leakages of primary school funds from 74% to 20%. Subsequently, several similar
studies have been commissioned in other countries and while PETS have often proved to
be useful diagnostic tools where they have been implemented, they have not been as
successful in achieving the impact seen in Uganda. The main objective of expenditure
tracking surveys in education, whether they are PETS or Quantitative Service Delivery
Surveys7 (QSDS) or those undertaken by civil society organisations, is to identify key
leakages and bottlenecks in the flow of funds, in order to understand the factors behind
any inefficiency, inadequacy or inequity present in the educational financing system. The
various surveys commissioned across the developing world have achieved this with
differing levels of success. While the pioneering Uganda PETS is cited as a tremendous
success story, later initiatives have not met with the same resounding appreciation.
However, these studies can still provide guidance on potential pitfalls and weaknesses
that can help guide future efforts.
10
A recent example of an expenditure-tracking survey is the PAISA report in India (PAISA,
2012). Given the PAISA report is a very recent publication, the true ‘impact’ of this
initiative on policy and practice cannot be adequately assessed as yet, However, this
citizen-led social audit – albeit with much room for improvement – has put in the public
domain some evidence on the financial management of public education expenditure and,
as such, has opened up a space in a hitherto closed subject, and has attracted media
coverage in India. It has also brought the issue of good financial management,
transparency and accountability to the fore in public discussions. One of the key
recommendations to improve financial planning relates to capacity building of personnel
at both the centre and state levels since low levels of capacity both undermine fiscal
planning and effective implementation of educational policies.
Pedley and Taylor (2009) argue that several factors undermine effective implementation
of educational policies aimed at universal access as envisaged under EFA. Using the
example of Ghana, they argue that the strong influence of the economic and political elite
on both educational policy and the allocation of resources affects educational policy and
10
PAISA literally means money in Hindi. It is the equivalent of ‘cent’ in the USA or ‘pence’ in the
UK.
33
places demands on the educational system that pursues their own interests rather than
those of the poor. Secondly, the use of private schooling by elite families strengthens their
chances of securing access to higher levels of state education which means that the
demands on public resources at higher levels of education arise from this pressure from
the elite. Thirdly, due to increasing primary and secondary education, tertiary education
numbers have also increased greatly. As a result, while governments could be committed
to EFA, elite pressure on limited resources diverts them from primary education and from
pro-poor policies. However, Little (2010b) also notes that in Ghana while there is strong
political will for educational reform, the supply of trained teachers struggles to keep a
pace with the numbers of children enrolled, and, as more and more children complete
primary and junior secondary education the goal posts of what constitutes EFA will shift.
So too will the ideological preferences of political parties in a democratic system. Thus, it
is a combination of factors that appear to hinder and often block effective reform
implementation.
Pedley and Taylor (2009) also claim that donors are operating in a world where it is more
difficult to influence policies. King (2007) presents a discussion on the trade-off between
basic and post-basic education agendas in Kenya showing the complexity of policy-making
and implementation when national priorities are entangled with international agendas.
King argues that while Kenya’s national ‘preoccupation’ has been with the whole of the
education and training system, external donors have often prioritised particular sub-
sectors such as primary schooling. It is inevitable that countries with a high dependence
on external funding potentially end up paying substantial attention to the investment
priorities of their principal development partners to the detriment of national priorities.
Similarly, Kempner and Loureiro (2002) question the cultural neutrality of policies imposed
by major international monetary agencies. The policies of key donors and agencies, they
argue, are based on perceptions that all developing countries are cultural variations of the
same problems and hence are assumed to require the same solutions. Through an analysis
of the failures of the World Bank’s programmes in Brazil and other developing countries,
they call for a greater recognition of the need for solutions in localised contexts. It
appears that it is inappropriate policies that are formulated out of ignorance of the local
culture and context that are most doomed to failure. This view is supported by Kosack
(2009) who reiterates that one of the key implications of his analysis for aid effectiveness
is that donors need to understand the country contexts in which they are working and to
judge whether a government is likely to display the enduring political will to provide basic
education. This likelihood increases when a political entrepreneur of the poor is affiliated
with the government. Thus, it would seem that it is inappropriate design as well as
ineffective implementation that hinder policy reform.
Some of the literature discussed in this section analyses the factors behind policy
implementation gaps and policy failures, and blames factors such as low state capacity,
poor administration, a poor delivery system, poor governance, and corruption leakages for
failing policies and poor outcomes. However, even if the apparent reason for a policy
failure is (say) some leakages/corruption such that the policy is not implemented as
intended, behind that is some political economy constraint, some lack of political will or
some vested interest, which does not want to reduce the corruption. This may be perhaps
because the politicians or bureaucrats that make the policy, or the vested interests that
34
11
lobby for that policy, themselves benefit from the policy . So they have an incentive to
not clean up corruption. Worse still, they may even have deliberately chosen to lobby or
recommend or even make that policy precisely because that policy gives them scope for
corruption. It has been suggested in the literature that that is why most education policies
are to do with expanding access to education, and providing inputs to schools, which
require expenditure.
4.5 Driving forces

What political and economic conditions drive or inhibit education reform, both in the
formulation of reform and in its implementation?
Drivers of educational change

As we have seen above once a reform process is in motion many stakeholders exert
influence. But what are the forces that create the imperative for reform in the first place?
These will vary from country to country. In the 1990s in some countries education
reconstruction and reform were conducted within a framework of economic ‘structural
adjustment’. Buchert (1998) presents examples of countries where economic
reorganisation has been largely internally determined and of countries driven to do this
because of their dependence on international aid. In other countries the underlying driver
has been political transformation. In some cases national political transformations have
preceded, in some been an adjunct to, and in others have followed the economic reform
process. The role of international financial institutions, technical and funding agencies
and intergovernmental organisations in national policy-making has varied.
Political drivers for change are of several types. In Ghana the need for reform of the basic
education system stemmed from a political transformation from a military to a democratic
regime and attendant constitutional change. While the reform was supported over many
years by the international community, the fundamental driver was national and political.
In practice the content of the reforms echoed those proposed in the 1970s under a military
regime but lack of economic resources at that time impeded their adoption and
implementation (Little 2010c). In Sri Lanka major reforms of education, in 1972 and 1997,
were driven by political instability. In 1971 a youth insurgency from among disaffected,
unemployed youth compelled the government to promise a radical overhaul of the
education system to make education more relevant to the world of work. In 1994 after
years of civil strife between different groups the government once again committed itself
to an overhaul of the education system (Little, 2011).
In their comparative study of African countries Evans et al. (1995) underline the role of
political imperatives and contexts in ‘triggering’ reviews of education policy between the
1970s and early 1990s. National reviews of education arose in Benin, Mali and Uganda after
a new, often revolutionary and/or newly elected democratic government came to power
(Benin, Mali, Uganda). They also arose at the end of a period of conflict or war
(Mozambique, Uganda), when public dissatisfaction with the condition of education could
no longer be ignored (Benin, Ghana, Guinea), or when macroeconomic adjustment, often
linked with reliance on external development financing, obliged the government to
11
Tanzania is a good example of major over-expansion of a secondary network since 2006 due to
political party/MP interests.
35
reorient its financing and budgetary strategies and practices (Benin, Ghana, Guinea,
Mauritius).
Regime type
Regime type is considered by some authors to be an important political condition that
drives on educational reform. Ansell (2008), for instance, argues that the effectiveness of
education policies will differ depending on regime type with autocracies potentially having
a vested interest in curtailing foreign aid earmarked for education purposes and targeting
funds to preferred political agendas and subsidise higher education for the elite over and
above the interests of the masses. Ansell looks at the relationship between regime type
(openness, democratisation, etc.) on educational spending. The author finds that there
are robust and significant effects of both regime type and openness on different types of
educational spending. Private spending is also strongly negatively affected by democracy.
Democracy is consistently associated with a shift in spending from tertiary to primary
education although the effect of openness is reversed when moving from developing to
developed countries. Hicken and Simmons (2008) finds that while (as several other authors
have found) democracies may spend more on education than non-democracies, the effect
of this additional spending on educational outcomes is dependent on the type of
democratic institutions in place. Bourguignon and Verdier (2005) note that opening up a
developing economy may have unexpected consequences for investments in human capital
through education. The basic argument of this paper is that a ruling capitalist elite has a
political interest in subsidising domestic education as long as the return on financial and
physical assets is positively affected by local human capital endowments. However,
globalisation may reduce the elite’s incentives to fund domestic education. Thus,
macroeconomic policies such as openness may also be significantly linked to the political
economy of education and the associated outcomes. Additionally, the relationship
between education and democracy does not necessarily work in any one specific direction
(Harber 2002). Harber’s study provides specific examples from Africa where education has
not played a significant role in furthering democracy and provides some further examples
of African countries where serious attempts are being made to try to change education
systems in a more democratic direction.
There are several other factors that may either inhibit or promote educational reform.
Multiparty electoral competition may prove to be a powerful force influencing the
provision of basic educational services in certain contexts (Stasavage 2005). Additionally,
an important aspect in relation to education decision-making is political knowledge and
unless the electorate has a basic understanding of the political system and the key
political issues, they will be making choices based on ignorance which can be argued to be
no choice at all (Harber 2002). The extent to which elites dominate the political arena can
also be an important force in determining educational reform. These may differ depending
on when political systems are centralised or decentralised. Archer outlines a number of
important differences between centralised and decentralised systems: in centralised
systems the greater the superimposition and unity among the relevant elites the more
standardised are the educational changes introduced or the existing practices which are
defended. Political manipulation is also usually the dominant process of negotiation in
centralised systems. If solitary governing elites dominate the political arena the measures
introduced reflect its restricted interests. In decentralised systems, on the other hand,
standardised measures are infrequent because those seeking change do not have to
36
accommodate their goals with others, thus diluting their precise requirements in order to
be able to exert greater political pressure.
Others have suggested that education reforms oriented to increasing educational access
are not confined to democratic regimes. Studies of Ghana suggest that it has been
populist, rather than democratic regimes that promoted access (e.g. Kosack 2009, Little
2010b). Equally, it is not inevitable that a regime change towards democracy will
necessarily have benign effects on the education sector. Pherali et al. (2011) suggest that
in Nepal since the Comprehensive Peace Agreement, education has become highly
politicised and an arena for politicians to declare ideological commitments serving their
needs rather than those of the children and without considering implementation
implications, with SMCs increasingly used as vehicles to mobilise voter support. The
greatest strides made in basic education in China in the 20th century – and which formed
the basis for China’s current educational record – occurred during Mao’s mass-oriented but
highly non-democratic regime (Lewin et al. 1994).
The importance of political will

In recent years political will has been seen as a key driving force in both designing and
implementing effective educational reform. In 2010 the United Nations Secretary-General
Ban Ki-moon said: ‘We do not need more pledges. If nations deliver on the financial
commitments they have already made, we can achieve the goals. There is clearly a lack of
political will’ (D’Angelo 2010). The concept and term political will is used most commonly
by the international community to exhort and berate national governments. However it
might equally be used to exhort the international community itself (e.g. Birdsall et al.
2005).
Political will has been defined as ‘a sustained commitment of politicians and

administrators to invest the necessary resources to achieve specific objectives and a
willingness to make and implement policy despite opposition’ (Little 2011). Several studies
stress the importance of political will in driving change. To take one example, Barber
(2013) emphasises the criticality of political will in the large and ambitious Punjab School
Reform programme in Pakistan, where he cites the very strong commitment of the Chief
Minister in the design and implementation of the programme, leading to rapid and
significant improvements in educational outcomes. However, others have expressed some
scepticism about the adequacy of political will as a driver of change. In any case, political
will is ‘endogenous’, i.e. not usually originating from outside, but created by the system
itself. CfBT (2011) highlights how limitations in coherence and consensus at the political
level undermine the sustainability of education reform programmes despite the political
will of those within the education arena and among development partners.
A more nuanced example is from Sri Lanka, which also helps to resolve the apparent
tension between the Barber and CfBT views, and it highlights the importance of
distinguishing between national and local political wills. In a very detailed analysis of the
comprehensive 1997 Education Reforms in Sri Lanka, Little (2011) describes how national-
level political will, manifested through the President’s personal involvement in both policy
formulation and the early stages of implementation, gave the reform process considerable
momentum. It was particularly influential in the reform component judged to have
enjoyed most success – primary education – where good technical design, the involvement
37
of the provincial administrations, wide-scale teacher training and domestic and foreign
financial support also played their parts. The analysis of five components of the reform
underlined the need to analyse political will – both at the national and local level – in
relation to particular reform content. However not all components attracted the same mix
of national and local political will and interest. The reform of junior secondary education
attracted neither national nor local political will at both the formulation and
implementation stage. Little’s analysis extended the arenas in which political will is
exercised to include the sub-national level, from province to school level. Two of the
reforms (school rationalisation and the equitable distribution of school facilities) attracted
myriad political wills at the local level that in turn generated so much resistance that they
impeded reform implementation. Even in primary education that had enjoyed so much
political will at the highest of levels, local political wills influenced the selection of
teacher advisors on non-merit criteria and slowed the pace of implementation.
Conceptually, the findings from this case study resonate with other writings on the politics
of policy formulation and policy implementation, and suggest a more general application.
Political will pervades both policy formulation and policy implementation and is a
necessary but not sufficient ingredient for the implementation of education reforms.
Where policies are not subject to public or parliamentary debate then a culture of
programmatic politics, involving an exchange of votes for public goods of benefit to large
numbers, is muted. The remoteness of the policy formulation process encourages a culture
of patronage politics in which citizens exert their agency in the implementation process
through exchanges of political support for private and/or club goods (e.g. teacher
appointments and transfers, selections of schools for intervention programmes).
The case study suggests that any analysis of the role of political will in education reform
needs to be pitched at multiple levels – from the national-level interests, commitments
and actions of Presidents, Ministers of Education, Ministers of Finance and political parties
to the local-level interests, commitments and actions of citizens, local politicians,
teachers, parents and officials in local and provincial government administrations. Where
national-level and local political wills are directed to the same ends in education, they
can be mutually reinforcing. Where myriad local wills are moving against the ends
promoted at the national level they are at best neutralising and at worst undermining.
Political will is a double-edged sword (Little 2011). This is also apparent from Little’s
(2010a) analysis of India where she argues that the role of political will for and
commitment to elementary education in India has shifted over time and reflects broader
political shifts in the definition of development and in commitments to overcoming social
and economic inequalities.
While political will may be present, the rhetoric is often far removed from reality.
Sørensen (2008) provides the example of Sri Lanka where the government attempted to
combat conflict and foster social cohesion through education. A key strategy within this
was to eliminate discriminatory contents from learning material and to develop a new
curriculum including subjects such as peace, tolerance, and citizenship. The author argues
that although textbooks incorporated elements of citizenship and notions of tolerance, in
reality the hiring of teachers, allocation of resources, distribution of books, etc., were
examples that contradicted the notion of equal citizenship and replaced it with a sense of
partial citizenship due to the influence of politics on education. Political patronage based
38
on ethnicity and socio-economic status is prevalent and recognition and resources can be
seen to be exchanged for votes and other favours. The author highlights the limitations of
merely using a curriculum as an agent of change by stating that as long as students
experience deprivation, discrimination and a system where ‘politics and not policies’
count, then the curriculum is unlikely to induce the notion of equal citizenship with any
credibility.
Positive cases of reform

In much of the above we have identified specific factors and forces that have impeded
education reform. In this section we focus on more positive accounts of reform through
three comparative studies and one detailed case study.
Grindle (2004) considers 39 reform initiatives in the content and structure of basic
education in Latin America (but with a particular focus on Bolivia, Ecuador, Nicaragua and
the state of Minas Gerais, Brazil) between 1977 and 1996. Most of the reforms of the 1950s
– 1970s in Latin America involved the expansion of education in rural areas and of facilities
for poor children in urban squatter settlements. These reforms encountered relatively
little resistance and were generally considered to offer positive cases of reform. Access
reforms ‘provided citizens with increased benefits and politicians with tangible resources
to distribute to their constituencies. They created more jobs for teachers, administrators,
service personnel, construction workers, and textbook and school equipment
manufacturers, they increased the size and power of teachers’ unions and central
bureaucracies. In fact unions were often among the principal advocates for broader access
to public education. Given these characteristics it is not too much to argue that these
reforms were ‘easy’ from a political economy perspective’ (Grindle 2004, p.6). These
positive cases however stand in stark contrast to the quality-enhancing reforms of the
1990s. These ‘involved the potential for lost jobs, and lost control over budgets, people
and decisions. They exposed students, teachers and supervisors to new pressures and
expectations. Teachers’ unions charged that they destroyed long existing rights and career
tracks’ (Grindle 2004, p.6).
Grindle’s (2004) book is probably the most cited source in the political economy of
education. It is a seminal analysis (based on 16 case studies) of how positive education
quality reform took place in Latin America despite strongly mobilised interests opposed to
change. She concludes that whether reforms succeed or not depends on how they are
introduced, designed, approved and implemented. For example, in Mexico President
Salinas waited three years to establish his power and negotiate with the unions, to time
his reforms in a manner optimal for their success. Another factor behind reform was the
successful leaders’ use of their ability to appoint ministers and technical teams in order to
promote their initiatives. Based on a detailed analysis of the design teams in the reforms
that took place in Bolivia, Brazil, Ecuador and Mexico, Grindle shows that while the role of
the reform-design team is often overlooked in political economy discussions (with the
widespread assumption that their work primarily involves the mechanical application of
expertise), in fact such teams balance choices and options in devising a solution and
frequently engage in political conflicts. For example, in Minas Gerais in Brazil, the
backgrounds of the reform team members were similar and it ensured that they came to
agreement relatively quickly on what the education problem was and how to solve it).
Grindle highlights that success also depends on how reform leaders manage their
39
opposition, with negotiation yielding superior outcomes compared to confrontation:

Salinas was able to negotiate with the Mexican teachers’ union instead of antagonising it.
Successful reform also rests on systematically weakening the opposition. Finally Grindle
shows that successful reform efforts are led by people who find opportunities to ‘set the
terms of the debate’ about reform. For example, in Mexico and Bolivia, the Presidents
stressed the importance of reforms for modernising the countries’ economies, implicitly
suggesting that entities opposing education reform were opposed to modernisation,
economic growth and poverty alleviation.
In a comparative study Mehrotra (1998) analyses countries in the ‘developing world’ that
appear to be successful in achieving major improvements in access to education at
relatively early stages of their respective development trajectories12. He suggests that the
countries furthest from EFA have not adopted policies that would increase access, such as
(demand side policies) increasing the physical proximity of schools, mother-tongue
education especially in the first few years and school-feeding programmes, and (supply
side policies) equitable public expenditures by level, low unit costs and adequate
expenditures on materials for teachers and students.
Another influential study describes and analyses successful education reforms in 20

education systems (McKinsey 2010). In this study, education systems are divided into four
categories, those that have improved from poor to fair, fair to good, good to great and
great to excellent, these categories being based on levels of student achievement. All the
developing country examples – Chile, Ghana and the states of Madhya Pradesh in India,
Western Cape in South Africa and Minas Gerais in Brazil – fall into the ‘poor to fair’
category. In some countries several episodes of reform are studied, yielding a set of 34
‘reform journey’ cases. The evidence base for the analyses of reform journeys are
interviews with approximately 200 system leaders and the descriptions of around 600
interventions which they led. Of greatest interest for our review is the analysis of the five
developing country reform journey cases: Chile (2001–2005), Madhya Pradesh (2006+),
Minas Gerais in Brazil (2003+), Western Cape (2003+) and Ghana (2003+). The report
presents 11 examples of interventions common to these five countries’ reform cases.
These are ‘scaffolding and motivation for low skill teachers and principals’ (prescriptive
teaching materials, teacher technical skill building, external coaching, school visits from
officials from the centre, increases in instructional time); minimum quality standards
(outcome targets, assessments, school infrastructure development, textbook provision,
additional funding for low performing schools); and access (increase school places, meet
basic needs of students – meals, clothing, transportation, toilets).
Alongside the analysis of education interventions is a description of political factors that

‘ignited’ some of the reforms. In Chile the transition from the Pinochet government to
stable democracy was considered an important spur for reform. In Ghana the shift to
democratic government and a new constitution provided the framework for education
reform. No significant political or economic events are identified as igniters of reform in
Madhya Pradesh, Minas Gerais or Western Cape. The report also identifies the role of
technical and political leadership in ‘igniting reforms’. In Chile and Minais Gerais new
12
The ten cases were Cuba, Costa Rica and Barbados in the Caribbean and Central America region;
Botswana, Mauritius and Zimbabwe in Africa; Kerala state in India and Sri Lanka in South Asia; and
Malaysia and the Republic of Korea in East and South East Asia.
40
technical and political leaders assumed office at the beginning of the reform period, while
in Ghana and Western Cape only the technical leaders were new (p.97 of the report
incorrectly describes King Abdullah bin Al Hussein as the new political leader of Ghana in
2003). Madhya Pradesh was excluded from this part of the analysis because interviewees
failed to reach agreement on the leadership role. Beyond this, however, little is said
about the broader political economy within which the reforms occurred nor how tensions
between stakeholder groups were managed and resolved. Only one mention of teacher
unions appears in the report and this in relation to Hong Kong.
While this report is presented in an extremely upbeat and persuasive style, its analysis is
flawed in at least one major respect. It focuses only on improving systems. It fails to study
systems whose student performance was stable or in decline. Many of the interventions
mentioned by the leaders and improvers, the political changes associated with the onset
of reforms and the assumption of office by new technical and political leaders are not
uniquely associated with improving systems. These can be found in many other systems,
where less or no success was achieved. Only through a comparison with systems deemed
to have stayed in one place or gone into decline could the authors assert with any
confidence that they have identified the most important reform drivers. The report
acknowledges this when it says: ‘the systems that have been unsuccessful in trying to
improve may carry out the same types of intervention that successful system undertake’,
but it goes on to assert ‘but there appears to be one crucial difference, that they are not
consistent, either in carrying out the critical mass of interventions appropriate to their
performance stage, or in pursuing them with sufficient rigour and discipline’ (McKinsey
2010, p.20). Unfortunately, the report presents no evidence in support of this latter
assertion, which somewhat undermines its instructional value.
A detailed case study of the political factors and driving forces underpinning positive
educational change is provided by Little (1999). This analysis focuses on increases in
education participation among the children of the minority Indian Tamil community
residing in tea and rubber plantations in Sri Lanka. Historically this community had
suffered much lower levels of educational participation than other social and ethnic
groups. During the period 1977 to 1994 access to primary and secondary education rose
dramatically. The analysis considers an array of political, economic and social forces that
drove forward plans to increase educational access. Among the political drivers were the
nationalisation of plantation schools, which removed the control of teachers and schools
from private sector employers interested mainly in maintaining a supply of low price and
unskilled labour, and the passing of legislation on citizenship which enabled a ‘stateless’
community to gain some stake in the future of their children in the country. Other factors
responsible for the increased schooling access were the growth of a labour surplus in the
plantations and a decline in the plantation economy which drove an interest in education
among the parents, but significantly also on the part of plantation owners who preferred
children to be in school rather than unemployed and ‘roaming around’ the plantations. A
growth in foreign aid, much of it focused on the poorest and marginalised groups,
supplemented meagre government resources. The foreign aid was used for the building
and rehabilitation of schools and for a range of ‘quality inputs’ such as increases in the
numbers of teachers, teacher training and upgrading, learning and teaching materials and
supervisory support. Alongside these factors was the ‘will’ and determination on the part
of several groups who wanted to see change on the ground, including education officials
41
and teachers who were themselves of plantation origin and who participated actively in
planning for the development of plantation schools and drove those plans forward at the
school level. Officials (all of them former teachers) and school teachers worked together
as a group for the same ends. The final, but by no means the least, influence was the
broader political and ethnic crisis and the specific position of plantation Tamils within it.
The political crisis and the growing civil war through the 1980s created the conditions in
which ‘windows of opportunity’ opened up. This part of the analysis will come as no
surprise to those who understand the political tapestry of Sri Lanka. But external
audiences are often puzzled by a story of progress among Tamils during this period. How
and why, outsiders ask, did the state promote educational expansion in plantation areas
during a period in which (i) the Sri Lankan state had been accused internationally of
human rights violations against minority Tamils; (ii) there had been open warfare between
Tamil extremists and the Sinhala-dominated state security forces as Tamils repeated their
calls for an independent state of Tamil Eelam; and (iii) thousands of young and educated
Tamils and Sinhalese died. Moreover, they ask, why would a government encourage
investments in plantation people’s welfare when the economic contribution of the
plantations was waning? Surely, the odds were stacked against educational expansion?
Resolution of this conundrum lies in an analysis of the political position of the plantation
Tamil community within the broader conflict, and the strategic actions of political
leaders. The leaders of the two main political parties in Sri Lanka had long understood the
importance of votes from minority constituencies. The vote of the plantation community
was important in the deliverance of the United National Party to power in 1977 and its
maintenance to the mid 1990s. Subsequently it was also important to the opposition
People’s Alliance in the mid 1990s. Much of the support to the two main political parties
was delivered by supporters of the trade union-cum-independent political party, the
Ceylon Workers’ Congress, via their unrivalled leader and political entrepreneur, Mr
Thondaman. Rather than joining the calls for Eelam from sections of the Tamil
communities in the north and east of the island, Mr. Thondaman chose instead to accept
ministerial positions and to promote the interests of the plantation community from within
government. Seizing every window of opportunity to wring concessions from the state, he
promoted increases to the minimum wage, housing, income generation, the resolution of
the citizenship issue – and education. The broader political crisis faced by the President
and his government and the strategic choice exercised by a specific political agency
provide a major part of the explanation for increased access to education among the
plantation community over the 1977–1994 period.
Hoffman (2013) provides an example from Tanzania and suggests that the limited impact
of reform in the country is partly attributed to capacity constraints and to ‘deliberate
design flaws’ (p.24). The author argues that the strong role of the President and party ties
have also debilitated reform. And despite strong political will and commitment to
transparency, poor quality information undermines the reform process. However, the
author notes the emergence of new ‘drivers of reform’ in the form of increased political
competition, internal factions within the ruling party, emergence of civil society
organisations and the ability of the ruling party to transform itself with changing times.
4.6 Evidence map and strength of evidence Table 4.1 provides an evidence map of the
reviewed literature. The five main themes addressed in this review are: roles and
42
responsibilities, rent-seeking and patronage politics; decision-making and the process of

influence; implementation issues; and driving forces. These are depicted along with the
predominant research design used in the analysis and the country context within which the
study was undertaken. The quality of the studies is indicated using a colour-coding system
which represents medium, medium-high and high quality studies. Low quality studies did
not meet the inclusion criteria and have been excluded in the earlier stringent selection
processes adopted, described in section 3.
Broadly speaking, among the studies reviewed, there is only one ‘high’ quality study in
India addressing one of the themes reviewed in this study. A majority of the studies are
medium quality and are clustered around themes 4 and 5: implementation issues (13
studies of medium quality) and decision-making and the process of influence (11 studies of
medium quality), followed by 11 studies of medium quality in the ‘Driving forces’ theme.
The studies cover a broad geographical base and employ a range of techniques in terms of
their research design.
Theme 1, Roles and responsibilities: There are 12 studies that broadly address the first
theme. Of these, half are of reasonably high quality and have been given the medium-high
assessment. Seven of the 12 studies in this theme use quantitative approaches. Finally,
half of the studies reviewed under this theme are on India and there is limited
geographical coverage of this theme in the literature reviewed.
Theme 2, Rent-seeking and patronage politics: Very few studies (eight) appear to be
addressing this theme directly or indirectly in their analyses. However, the quality of the
evidence, when it exists, is reasonably strong.
Theme 3, Decision-making and the process of influence: There are numerous studies
directly or indirectly addressing this theme. Studies employ different types of research
design but are clustered either in the ‘empirical + broadly quantitative’ design or in case
study evidence. The only high quality study reviewed in the analysis, an RCT design,
addresses theme 3 and is based on India. There is a relatively broad geographical coverage
among the studies and while a majority of the studies have been assessed to be of medium
quality, some good quality evidence (medium-high) does exist especially in the ‘empirical
+ broadly quantitative’ and ‘empirical + broadly qualitative’ domains.
Theme 4, Implementation issues in education: There are several studies that delve into
theme 4 and have been assessed to be of a medium quality. However, much of this
evidence is qualitative in nature and there appears to be only one purely quantitative
study addressing this theme albeit a multi-country one. The geographical coverage is
reasonable.
Theme 5, Drivers of educational change: Evidence on the final theme is reasonable in

terms of quantity, is well distributed across the research designs and geographical
coverage is varied.
Strength of evidence: Overall, there is modest evidence with respect to each of the
themes. More robust evidence is warranted and some specific areas of potential future
research are highlighted in the following section.
43
Table 4.1: Evidence map of 64 studies

Theme
Research 1. Roles and 2. Rent-seeking and 3. Decision-making and 4. Implementation issues 5. Driving forces
design responsibilities patronage politics the process of influence
RCT India (39)
EMP-Quant India (20, 24, 25) Mexico (2) Multi-country (51) Multi-country (56) Multi-country (28)
Mexico (15, 19) Multi-country (3) Mexico (2) Mexico (19)
Indonesia (11) (9) Africa (50)
India (23, 59) Indonesia (11)
Ghana (57) India (20, 24, 25)
Kenya (54) Argentina (40,41)
Mexico (45, 46)
Multi-country (42)
EMP-Qual India (5) India (47) India (6) Multi-country (7) Sri Lanka (35)
Tanzania (60) (14) Multi-country (7) Ghana (14) Tanzania (58)
South Africa (64) Ghana (13) Zimbabwe (55)
Nepal (63)
TP (12, 38) (4) (4, 38) Kenya (22, 48) (4, 17)
Nigeria (56) South Africa (36) India (37) Sri Lanka (49)
Ghana (43) Brazil (21)
East Africa (44) Pakistan (52)
MX India (42) Multi-country (52) Chile, Colombia, Mexico, Benin, Guinea, Mauritius, Sri Lanka (30, 33)
Taiwan, Ghana (26) South Africa, Mozambique, Mozambique, Uganda (14) Ghana (32)
Zambia, Zimbabwe (10) Ghana (26, 32)
Multi-country (62) India (34)
Taiwan (26) Multi-country (61)
44
Theme
Research 1. Roles and 2. Rent-seeking and 3. Decision-making and 4. Implementation issues 5. Driving forces
design responsibilities patronage politics the process of influence
IE Honduras (1) Honduras (1)

Guatemala (1) Guatemala (1)
THE (16) (8, 29)

Multi-country (28)
LR Multi-country (27)
(31)
Research design: RCT = randomised controlled trial; EMP-Quant = quantitative analysis of observational data; EMP-Qual= qualitative analysis of observational
data; TP = think piece; MX= mixed methods (see Tables A4.3A and A4.3B for details); IE = impact evaluation; THE = theoretical; LR = literature review.
Study code: 1, 2, 3…64 (please see Tables A4.3A and A4.3B for details).
Quality of individual studies: Medium-low, Medium, Medium-High, High.
*Some studies do not relate to specific countries and therefore only the study number appears in parenthesis.
45
4.7 Gaps in the literature

The literature on the political economy of education is under-developed in geographical
scope, robustness of methods utilised and theoretical richness. Moreover, the array of
theoretical frameworks developed since the emergence of the ‘new institutionalism’ of
the 1980s – principal agent theory, agency theory and the theory of repeated games – has
not been well utilised in the reviewed literature.
This review makes clear that large parts of the world, especially most countries of Africa
and South-east Asia, remain virtually untouched by research on the ways in which
political-economy forces affect their education-sector decisions, processes and outcomes
in areas as diverse as planning, budgeting, curriculum, law, regulation, fees, textbooks,
salaries, etc. As highlighted by many of the authors of the reviewed studies, where
research does exist, the findings are very context specific and non-generalisable, which
underlines the need for countries and regions to have their own locally relevant analysis.
Among published research the quality of the research varies, with few rigorous studies.
Kremer (2003) and Muralidharan and Sundararaman (2011) stress the importance of
ensuring that the quality of political economy analysis (PEA) is more easily distinguishable
by policy-makers. They suggest this can be done by a certification organisation that would
help policy-makers identify credible studies to ensure that reform is informed by strong
theory and evidence, to improve design and implementation. However, PEA is not easily
amenable to randomised experiments. Political economy constraints and variables are not
random and usually occur selectively, largely depending on context. It would, therefore,
not be possible to assign such variables to control and treatment groups for randomised
evaluations. More rigorous research, however, is possible by using good quality panel data
and quasi-experimental methods and combining a more mixed-methods approach in
researching the political economy of education.
The review has indicated that teachers are politically active in several countries.
However, it was difficult to infer the extent of political involvement of teachers in
different countries let alone the impact this political involvement could potentially have
on educational outcomes. DFID could do well to commission studies that aim to gather
information on the level of political involvement of teachers in DFID priority countries.
Additionally, simple descriptive studies that aim to identify teacher salaries (at various
education levels) and average per capita GDP in different countries would allow a deeper
understanding of whether teachers lobby for higher salaries or for other outcomes. If
teacher salaries are less than average per capita GDP (of equivalent persons in the
country), then it could be argued that teachers could potentially be lobbying to better the
lot of underpaid teachers. If, however, the average teacher salaries are found to be
several times more than the per capita GDP in a country, teachers’ political engagement is
likely to be to achieve other (perhaps less desirable) objectives.
Mixed results of the effect of teacher unions on student outcomes (or on teacher absence)
call for a better understanding of the internal structure and dynamics of unions. There is a
need for systematic analysis of the role of teacher unions and teachers’ participation in
politics in different countries, and to understand these issues: to what extent and why
these are perceived as constructive in some countries/regions but apparently negative in
others; whether the relationship between teachers and politicians operates mainly through
unions; to what extent middlemen – the private intermediaries and government clerks –
46
are an important phenomenon in teacher labour markets; and how patronage politics and
teachers’ political power ultimately affect student learning across the income
distribution. In order to do this, not only do we need better measures to capture political
relationships, but we also need credible measures of student performance in elementary
grades, measures that are free from manipulation and can be compared across settings.
Our review has also indicated that there is insufficient evidence specifically addressing
where rent-seeking is most prevalent and why, i.e. whether it is in teacher assignments,
school construction, textbooks and so on. This is a significant gap in the literature which
future research could aim to address.
From a more theoretical standpoint, Leftwich (2006) calls for a more intensive focus on
the political dynamics of change. He highlights that although many of the studies in his
report contain richness of detail, not many employ conceptual/theoretical tools to analyse
political practices and trace their pathological relations with economic
activities/institutions. Studies have tended to employ a very general understanding of
‘political economy’. We echo Leftwich’s view that future work needs to develop more
conceptual clarity and more nuanced political theories about change and particularly
about how alternative structural, historical and institutional conditions determine varied
possibilities and constraints within which actors in different polities have to work thereby
generating differing developmental paths.
47
5. Conclusion
Unfavourable political economy blocks educational reform. This review confirms that
education reform takes place under circumstances that in many cases are politically
driven, and shaped by the interests and incentives facing different stakeholders, as well as
by formal and informal institutions. Insights from the literature urge consideration of the
interests, actions and choices of a wide range of actors, working in a wide range of
institutions across a number of interacting stages, in the process of education policy
reform – from agenda setting, to programme design, to adoption, to implementation to
institutionalisation and sustainability.
We have reviewed both the macro political economy literature which tries to link
education outcomes to alternative institutional structures encompassing variation in types
of election and bureaucratic processes, and also the micro political economy of education
literature which uses a sectoral approach, since specific features of the sector– such as
the political power of teachers – play a crucial role in explaining outcomes. We have put
the theory of political economy to use in evaluating the research on education systems in
developing countries. Focusing on developing countries, we examined the relationship
between politics and education outcomes from multiple disciplinary lenses, including
economics, political science, sociology and education.
To avoid the pitfalls of simple meta-analyses (which give equal weight to all studies, good
and bad), stringent inclusion criteria were applied when selecting the papers and articles
to be included, and the included literature was then graded by quality, based on DFID’s six
principles of high quality studies. We carried out extensive searches through bibliographic
databases, journals, organisational websites and consultation with experts to arrive at a
comprehensive collection of the quantitative and qualitative literature on the political
economy of education in developing countries. A hierarchy of evidence was used to
evaluate the validity of quantitative studies ranging from RCTs (high quality) to less
rigorous methodologies such as simple descriptive statistics that do not allow causal
interpretations (such as comparison of means). The validity of qualitative studies was also
analysed by choosing those that give the wider context, employ a methodology that
minimises the risk of bias, and whose findings are reproducible.
The review found that the theoretical themes of the literature focus on the effects of
regime type (e.g. democracy), degree of openness, the role of competing parties, and
concentration of resources. In addition, we discovered a growing literature on the role
played by vested interest groups, such as teachers’ unions, which has been crucial in
furthering our understanding of how power is exercised by different players.
We return to our theoretical framework. At the outset we suggested its fundamentals

revolved around actors, their interests, the shaping of their incentives and strategies by
contexts, the means by which actors exercised power in pursuing their ends, and the
consequences for students, schools, and the larger education system13. Our review of the
13
For work that applies this approach to education, see, e.g., Chubb and Moe, 1990; Moe, 2011;
Kingdon and Muzammil, 2003, 2009, 2012.
48
5. Conclusion
literature provides evidence for an extended and more elaborated formulation of

theoretical propositions as follows.
Because we are interested here in the prospects for educational change, we put special
emphasis on the key theoretical importance of vested interests. Vested interests are
universal. They arise in all government institutions, and in all countries of the world,
because certain people and groups benefit from the institutions’ operation: clients who
receive services, employees who occupy institutional jobs, administrators or politicians
who control the money, businesses that reap revenues as contractors, and so on.
As universal forces, vested interests are likely to be especially important to the politics of
change for three main reasons. First, they have strong incentives to resist any reforms that
alter, reduce, or eliminate their benefits, or render them uncertain—which most major
reforms would do. Second, they have strong incentives to become organised for political
action, and to invest in political power, in order to protect their institutions from change.
And third, they typically have incentives to do these things even if the existing institutions
are performing badly, and thus are in desperate need of reform—because, as job-holders
or administrators or contractors, their benefits depend on the continuation of existing
arrangements, not on performance. When major reforms are proposed in any nation, in
any realm of policy, vested interests are likely to be the prime source of resistance.14
We should emphasise that not all reforms are good, and vested interests are not
necessarily a negative force preventing positive change from happening. Sometimes,
politicians may promote reforms that are self-serving or unwise, and vested interests may
be doing society a service by resisting such reforms. Also, the mass constituencies that
receive governmental services are vested interests too; but unlike job-holders or
administrators or contractors, they may see virtually no benefit from existing
arrangements and demand major reforms that improve performance. The problem,
however, is that they are a large, diffuse group, and are unlikely to wield the kind of
power that other vested interests can wield, unless they form alliances with powerful
political and/or policy entrepreneurs who can promote their interests Our point, then, is
not that all vested interests always behave in ways that are bad for society. It is that, as
interests that are intensely affected by existing arrangements and very often resistant to
change, they are of key theoretical importance – and we should be focusing on them if we
want to understand the political dynamics of reform.
The application to education is straightforward. Every school system gives rise to various
kinds of vested interests. As we have seen, the most obvious are children and parents,
who are the direct recipients of services; but as a large, diffuse group, the barriers to
effective collective action are high for them, and they are at a disadvantage in the larger
politics of education. Other vested interests are likely to be much better organised and
politically much more powerful – and these interests will typically have a deep stake in
existing arrangements, and thus in resisting change, even if performance is very bad.
14
For a more extensive treatment of the logic underpinning a theory of vested interests, see Terry
M. Moe (2013) Vested interests, theory, and the political dynamics of American education.
Stanford University. The application in that paper is to the American context, but the logic is quite
general and can readily be adapted to any country.
49
The most notable of vested interests are teachers. They are likely to be quite numerous,
and intensely concerned about protecting their jobs (and incomes, careers, security, etc.)
from the threats entailed by major reforms. In many countries – not all – they may also
belong to teachers’ unions, and thus have a potentially powerful means of bringing those
interests to bear in the politics of education. When that happens, they can be expected to
resist reforms that radically change the existing system, require higher standards, or put
accountability pressures on them to perform. On the other hand, they can also be
expected to support reforms that work to their advantage – most notably, expansions of
spending and access that will produce more teaching jobs, and possibly higher pay and
benefits. The nature of the proposed reform is key to its sustained implementation.
The other major category of vested interests includes ‘officials’ of various kinds – national
politicians, local politicians, and national and local bureaucrats who have administrative
roles in managing the education system. The bureaucrats have vested interests in their
jobs, in the money they control, and in their autonomy to allocate money and make
education decisions as they see fit. They will resist reforms that threaten their jobs and
sources of power – as school choice and genuine decentralisation tend to do, for example –
and support reforms that do the opposite, such as those that expand the size of the
system. Politicians are more complicated. Because the education system is a giant
reservoir of jobs and money, politicians are in a position to use those resources for
patronage, clientelism, and their own profit. This is particularly true in very poor
developing nations, where the entire political culture may be fraught with corruption,
nepotism and rent-seeking. Many politicians in these settings, then, may see themselves as
having a deep vested interest in maintaining and expanding education systems for reasons
that have nothing to do with providing a quality education for children, and everything to
do with propping up their own political power and security.
Politicians are complicated, however, because – depending on the larger political system
and its institutions – they may also have incentives to be good, popular leaders of their
people, and thus to pursue reforms designed to bring about effective schools and the
efficient expenditure of money. It might seem that this ‘good leader’ role – and thus the
political will to promote high quality education reform – would be more likely to emerge
the more democratic a system’s institutions, but this is not necessarily so. Democratic
politicians are often driven to engage in clientelism, to use public money for payoffs to
powerful interest groups, and so on, in order to buy votes and stay in office, and these
incentives may induce them to plunder educational resources and refuse to engage in
productive reform. Democracy may also induce leaders to support reforms that expand
educational spending and access – thus pleasing ordinary people and vested interests at
the same time – while refusing to carry out structural reforms that are necessary (but
opposed by vested interests) to ensure that the schools are actually effective and the
money well spent.
Nevertheless, institutions do matter: they structure incentives – and thus behaviour. Even
though politicians in developing countries will often be faced with strong incentives to use
education systems toward their own material ends, their incentives will also be shaped by
the larger political system – and thus by the type of regime in which they are embedded.
The specifics of political regimes vary considerably from country to country, and we
cannot say in the abstract how they affect the incentives of particular leaders. We can
50
5. Conclusion
simply say that, in any given context, what we expect of leaders will be a combination of
knights and knaves– and that, in the absence of ‘correcting’ incentives from the larger
political system, politicians are likely to see their educational systems in opportunistic
terms, and to resist major reforms that aim to bring improvement.
Finally, we should emphasise that the set of actors is not homogeneous in its approach to
reform. A nation may be fortunate enough to have national political leaders with the
political will to enact real education reform, but these reforms may be eviscerated at
lower levels as local political and bureaucratic officials, together with teachers’ unions,
use their power in the educational trenches to make the reality of reform a total failure.
Similarly, the teachers’ unions may see professional training for their members as a path
to higher pay and status (and performance), yet politicians may insist on using teacher
jobs for patronage and nepotism, ensuring that the movement toward greater
professionalism goes nowhere. What matters is how the whole system and all of its various
actors, together, operate to produce outcomes. And almost all systems are inherently
stacked against successful reform: because there will be vested interests – somewhere, at
some level – that are against change, and use their power to obstruct major change and its
successful implementation.
How, then, can major reform of the education system be brought about? It would be
comforting to think that there is a simple formula that can be followed, but this just is not
so. Given the theory we have outlined – which, we think, captures the essence of the
problem – the starting point of any serious analysis should be: most reforms of any
consequence will tend to fail, or be watered down or distorted, because they will be
resisted by powerful vested interests, including most political leaders. There are not any
real solutions. The vested interests are not going to go away. Their power is not going to
go away. Politicians cannot be counted upon to have the necessary political will. The best
reformers can do, typically, is to pursue strategies that generate a modicum of progress
against difficult odds. How can they do that? The theory points to various avenues that are
worth discussing. We will discuss four here – three that are not very promising and one
that is.
The three that are not very promising are fairly obvious. (i) Reformers can try to reduce
the power of the vested interests – by, for example, outlawing teachers’ unions (or
outlawing strikes and other union weapons), or shifting authority from ‘problematic’
public officials to others who are more trustworthy. But if these vested interests are
genuinely powerful, then they will use their power to prevent anyone from taking their
power away. And such efforts are unlikely to work. (ii) Reformers can accept the power of
vested interests, and cooperate with them to arrive at potential agreements about
reform. But the problem is that major reform is threatening to their interests, and they
will demand steep payments and compromises in return. The resulting reforms are likely
to be weak (and expensive). Once they have been adopted, moreover, the vested interests
will take action on the ground over time to weaken the reforms still further. (iii)
Reformers can try to mobilise parents and citizens to give massive political weight to their
interests in high quality education. This may work in some contexts – in some villages or
localities, or in some party systems in reasonably democratic systems. But in general, this
strategy is likely to prove disappointing, because the great majority of citizens – especially
in developing nations, where poverty is endemic – are simply not interested in becoming
51
politically active, certainly not on an ongoing basis. They work, they have difficult lives,
the collective action problems are enormous. In most countries most of the time, they
cannot be counted upon to drive education reform unless they form the constituency of
powerful political entrepreneurs and can work in concert with them.
There is a fourth strategy, however, that shows much more promise. This one involves
taking advantage of unusual windows of opportunity to make specific kinds of structural
changes in the education system. A window of opportunity can open up, for example,
when a nation undergoes a political transition – say, from authoritarian to democratic
government, or from one party to another, or from one leader to another – and the new
leadership (for its own reasons) finds it advantageous to pursue serious education reforms.
This, again, is not the norm. But it can happen in the midst of changing incentive
conditions. A window of opportunity can also open up when outside funders – the World
Bank, say – offer money and expertise for education reform, and induce national leaders
to go along. Here too, there are no guarantees, as there may be strong vested interests at
work to siphon off money and undermine reform or interests that strongly oppose the
reforms promoted by the outside funder. But the intervention of vigilant, strong outsiders
who are committed to reform, and who are not embedded in a country’s vested interests,
can be a driver of changes that would not happen otherwise.
If these windows of opportunity open up, what structural changes would then work to
bring about genuine reform? The answer is to be found in formal rules designed to try to
eliminate the discretionary, individualistic decision-making on which the existing vested
interests thrive. These formal rules might require such things as transparency in the
expenditure of money, in hiring and firing, and in all aspects of the organisation of
schools; a formal merit system that disallows nepotism, patronage, and other such
approaches to school personnel policy; high formal standards for teacher certification;
formal examinations of students to provide a basis for assessing school performance;
serious performance-based evaluations of teachers, including monitoring of absenteeism;
and so on. These are precisely the kinds of rules that teachers’ unions and other vested
interests tend to oppose. But with an appropriate window of opportunity, leaders who
strike while the iron is hot may be able to put a new formal structure in place. And if they
can do this, it will not only shape behaviour and outcomes within the education system,
but it will also generate its own vested interests – new interests with a stake in the new
system. National and local bureaucrats, for example, will now have jobs making and
administering tests, running the merit system, evaluating teacher performance,
monitoring absenteeism, and the son on – and their jobs and vested interests will become
tied to these new formal structures. Teachers hired under a merit system will probably
see it as beneficial to them – and will have an interest in supporting it. And so on.
This strategy is not a silver bullet. Windows of opportunity are rare. New structures will
face resistance (although less than usual – as that is what makes a window of opportunity
what it is). And once they are in place, there will still be resistance from vested interests,
politicians will still have incentives to use education’s resources for their own ends, and
the larger culture of corruption, patronage, and clientelism may still persist and continue
to threaten the operation of the new structures. But this is what ‘good reform’ looks like
in a difficult setting. Once it is in place and operating, it has a decent chance of becoming
more entrenched and gaining greater support. One would hope that long-term economic
52
5. Conclusion
growth, for example, or long-term movement toward a healthier political system can be
achieved through education reform and can actually bring major change, and work to the
great benefit of children and their societies.
We capture much of the above in the theory of change presented below in Figure 5.1. In
column 1 we nest our concerns with actors, incentives, disincentives and strategies within
a discussion of the underlying drivers of or imperatives for reform. While the precise
nature of these will vary from country to country our review of the literature has
generated examples likely to be found in several settings (e.g. political instability,
constitutional change, economic policy shifts). We also nest our understanding of the
political economy of education within the underlying social, political, economic and
educational structures of the country within which reforms are promoted (and resisted).
The examples here include the structure of the education system, the political regime,
constitutional and legal frameworks and the nature of recent reform experiences. Both
the imperatives and underlying structures for reform generate the need and political
legitimacy for reform. In column 2 we identify a range of actors with vested interests in
reform. We divide these into internal and external actors to distinguish those whose
identities and spheres of influence are largely internal to a country and those whose
identities are less rooted in particular countries and who work across countries, including
the country in question. In column 3 we identify incentives that promote reforms and
threats that generate resistance to reforms, both of which will have more or less salience
for the different actors identified in column 2. The balance between incentives and
threats will lead (or not) to policy decisions (column 4), which in turn vary in terms of
content, clarity, complexity, and strength of intrinsic technical design. The timing of
policy decisions and the ability to take advantage of windows of opportunity are critical
here. While in general policy decisions lead to policy implementation we express the
relationship between decision and implementation reciprocally in recognition of the fact
that implementation sometimes precedes formal decision and that initial implementation
frequently leads to adjustments in the policy itself. During the implementation process
actors employ a range of strategies to promote and resist implementation (column 5). It
should be emphasised that these same strategies may also be employed prior to policy
decisions being made and formalised. Policy implementation not only influences
adjustments to decisions but also brings to the fore vested interests that may have been
silent during the process of policy formulation. For example local politicians and
administrators may participate little in the dialogue and negotiation surrounding policy
decisions; they may become vocal and called upon to act by teachers and parents once
policies manifest themselves for implementation at the local level. Policy implementation
may also lead to shifts in the constellation of incentives that promote and resist reform
and changes in the strategies employed to implement reform.
In principle, all of these actions have implications for the access and quality of learning
experiences of students and the characteristics of schools and/or education systems
(column 6).
53
Figure 5.1: Theory of change
54
5. Conclusion
The review found that research on the political economy of education is confined mainly
to developed countries, Latin America and South Asia. There is much less systematic work
on this topic in Africa and East Asia. Additionally, the political economy of education
literature is characterised by theoretical, descriptive, correlational and qualitative work.
Work examining the causal effect of political interests and power play on educational
outcomes is virtually non-existent, the empirical challenge being compounded by the
problem of selection and simultaneity between educational outcomes and political and
institutional processes. The major findings of the review are summarized below.
 Among all stakeholder groups, teachers and their organisations have great political
power because of their ability to influence electoral outcomes and political
fortunes. By militating for higher salaries not tied to performance, lobbying against
decentralised school management, and protecting inefficient and shirking teachers
from dismissal, some teacher unions cause educational inefficiency, though others
are milder and work constructively to improve the welfare of teachers. By contrast
parents do not have a collective voice on educational matters, since they are not
organised. Government and international agencies are recognised as the other
major stakeholders in the education sector.
 Rent-seeking and patronage politics are rife in educational set-ups in developing
countries. The politics of patronage suggests that it is more convenient to expand
educational coverage, e.g. by building more schools or hiring more teachers, than
to fix existing inefficiencies within the system because the former involves
spending on political actors whereas the latter may involve reducing resources
allocated to underperforming political stakeholders. Some literature using stringent
empirical techniques finds that teachers’ union membership and political
connections are both associated with significantly reduced pupil achievement in
India, and that teachers there also influence the school governance environment
through their direct participation in politics, i.e. by themselves becoming
legislators. This highlights the importance of shoring up teacher accountability
through pupil assessments and other reforms.
 A variety of groups influences the educational decision-making process and
educational change. The literature concludes that the supposed benefits of
decentralisation do not accrue in practice because in poor rural areas the local
elite captures all the space for participation in school affairs. At the macro level,
international donor agencies and global education institutions are exerting more
influence on education decision-making in many developing countries. There is
some evidence that international differences in student performance are
considerably related to institutional as opposed to resource-level differences
between countries.
 Much of the reviewed literature on education analyses the causes behind
implementation failures, and it blames factors such as low state capacity, poor
administration, poor delivery system, poor community information, and
corruption/leakages. However, underlying these are likely to be some political
economy constraints, some lack of political will or some vested interests. This may
be because the politicians/bureaucrats/vested interests that lobby for a policy may
themselves benefit from the policy/corruption. Evidence suggests that
participatory programmes (such as CMS) often fail and a plausible reason suggested
55
is the prevalence of grossly unequal power relations between poor rural community
members and well paid teachers. While the Ugandan PETS is credited with reducing
leakages of primary school funds from 74% to 20%, subsequent studies have not
achieved the impact seen in Uganda, though they are still useful diagnostic tools.
 Analysis in a large number of studies indicates that at the national level there are
potential drivers or agents of change. The literature also emphasises the
importance of political will as a key force in driving educational change. However,
political will alone is insufficient for driving change. An empirical strand of the
literature finds robust and significant effects of both regime type and openness on
different types of educational spending, showing that aggregate public education
spending increases with a shift towards democracy or openness. Democracy is also
consistently associated with a shift in spending from tertiary to primary education.
However, the effect of additional spending on educational outcomes is dependent
on the type of democratic institutions in place.
 The literature shows there are several other factors that inhibit or promote
educational reform. Multi-party electoral competition, political knowledge of the
electorate, the extent to which the elite dominates the political arena and the
extent of centralisation of governance can all be powerful forces influencing the
provision of basic educational services in certain contexts. The section on ‘Positive
cases of reform’ analyses examples from Sri Lanka and some Latin American
countries where benign political economy circumstances were created by change
drivers to achieve good outcomes. The cases illustrate the many factors that
ultimately converged to create an environment conducive for education reform.
 The literature shows that decentralisation policy in education has to bear in mind
the realities of local politics of influence in the community, and tap into the
positive side of this influence to improve education service delivery. It also
indicates that the crucial question for education policy is not that of more
resources, but of creating an institutional system where all involved people are
provided with incentives to use resources efficiently and to improve student
performance.
DFID would do well to address each of our five questions for the countries where it works.
To be effective in-country, DFID advisors need to know how to position themselves in
relation to the various constellations of stakeholder groups, and to understand the recent
history of policy reform processes in education and others sectors in that country. Our
findings provide a guide to the types of answers that might arise. The evidence we have
provided is overwhelmingly country-based. In the absence of comparative studies
addressing similar questions and employing similar designs it would be incautious to
translate findings from one context to another. Modestly costed country-based research
programmes could generate evidence on the political economy of education reform of
enormous strategic value to DFID advisors working in-country.
56
References
Altinok N, Kingdon G (2012) New evidence on class size effects: a pupil fixed effects
approach. Oxford Bulletin of Economics and Statistics 74(2): 203–234.
Altschuler D (2013) How patronage politics undermines parental participation and

accountability: community-managed schools in Honduras and Guatemala. Comparative
Education Review 57(1): 117–144.
Alvarez J, Garcia Moreno V, Patrinos HA (2007) Institutional effects as determinants of

learning outcomes: exploring state variations in Mexico. Well-Being and Social Policy 3(1):
47–68.
Ansell BW (2008) Traders, teachers, and tyrants: democracy, globalization, and public
investment in education. International Organization 62(2): 289–322.
Archer M (1981) Educational politics: a model for their analysis. In: Broadfoot P, Brock C,
Tulasiewicz W (eds) Politics and educational change: an international survey. London:
Croom Helm.
Barber M (2013) The good news from Pakistan: how a revolutionary new approach to
education reform in Punjab shows the way forward for Pakistan and development aid
everywhere. March 2013. London: Reform.
Barrerra-Osorio, F, Fasih, T, Patrinos, H.A, Santibáñez, L (2009), Decentralised decision-

making in schools: the theory and evidence on school-based management, The World
Bank, Washington D.C., USA.
Bennell P, Akyeampong K (2007) Teacher motivation in sub-Saharan Africa and South Asia.
London: DFID.
Bermingham D (2011) The politics of global education policy: the formation of the
Education for All – Fast Track Initiative (FTI). Journal of Educational Policy 26(4): 557–570.
Béteille T (2009) Absenteeism, transfers and patronage: the political economy of teacher
labor markets in India. Dissertation, Stanford University, California.
Birdsall N, Levine R, Ibrahim A (2005), Toward universal primary education: investments,

incentives and institutions, UN Millennium Project Task Force on Education and Gender
Equality, London: Earthscan
Bold T, Kimenyi M, Mwabu G, Ng’ang’a A, Sandefur J (2013) Scaling up what works:

experimental evidence on external validity in Kenyan education. Working Paper.
Washington, DC: Centre for Global Development.
Bourguignon F, Verdier T (2005) The political economy of education and development in

an open economy. Review of International Economics 13(3): 529–548.
57
Bruns B, Filmer D, Patrinos HA (2011) Making schools work: new evidence on

accountability reforms. Washington, DC: World Bank
Buchert L (ed.) (1998) Education reform in the South in the 1990s. Paris: UNESCO.
Carnoy M, Luschei T, Marshall JH, Naranjo B, Sorto A (2007), Improving Panamas and Costa
Rica’s education systems for the 21st century economy: a comparative study, Stanford:
Stanford University School of Education.
CfBT (2011) Recovery in crisis: the politics of education reform in Zimbabwe. Reading:
CfBt Education Trust.
Chabbott C (2003) Constructing education for development: international organisations

and Education for All. London: Taylor and Francis.
Chen D (2011) School-based management, school decision-making and education outcomes

in Indonesian primary schools. Policy Research Working Paper Series 5809. Washington,
DC: The World Bank.
Chubb JE, Moe T (1990). Politics, Markets and America's Schools. Washington, D.C.: The
Brookings Institution.
Corrales J (2005) The state is not enough: the politics of expanding and improving schools
in developing countries. Working Paper. Project on Universal Primary and Secondary
Education. Amherst, Massachusetts, USA.
D’Angelo C (2010), Development: political will the missing link for MDGs, Friday 19 March,
Inter Press Service, http://www.globalissues.org/news/2010/03/19/4908
Day Ashley L (2013) The shifting politics of the private in education: debates and
developments in researching private school outreach in India. Comparative Education
49(2): 206–225.
DFID, (February 2013), How to note when assessing the strength of evidence, A DFID
practice paper, UK.
Dreze J, Sen A (2013) An uncertain glory: India and its contradictions. London: Allen Lane.
Duflo, E, Hanna, R, Ryan, SP, Incentives work: getting teachers to come to school, MIT:
Department of Economics, http://economics.mit.edu/files/5995
Duncan A, Williams G (2010) Making development assistance more effective by using

political economy analysis: what has been done and what have we learned? Paper
presented at: Carnegie Endowment for International Peace/USAID/DAI Workshop on:
‘Advancing Integration of the Political and the Economic in Development Assistance:
Sharing UK and US Experiences’, Washington, DC, June 2010.
Eberts R W, Stone JA (1987) Teachers’ Unions and the Productivity of Public Schools
Economic Inquiry 24: 631-44.
58
References
Education International (2013) Study on trends in freedom of association and collective

bargaining in the education sector since the financial crisis 2008-2013, Nora Wintour,
report commissioned by Education International.
Essuman A, Akyeampong K (2011) Decentralisation policy and practice in Ghana: the

promise and reality of community participation in education in rural communities. Journal
of Educational Policy 26(4): 513–527.
Evans DR, Sack R, Shaw CP (1995) Formulating education policy: lessons and experiences
from sub-Saharan Africa. Paris: DAE/UNESCO.
Fernandez MA (2011) Teachers, protests and parties: the political economy of educational
transfers in Mexico. Paper presented at: Research seminar, Center for US-Mexican Studies,
University of California, San Diego, 19 October 2011.
Fullan MG (1989) Implementing educational change: what we know. PHREE Background

Paper Series, Document number PHREE/89/18, Education and Employment Division, The
World Bank.
Glewwe P, Hanushek E, Humpage S, Ravina R (2011) School resources and educational

outcomes in developing countries: a review of the literature from 1990 to 2010. NBER
Working Paper 17554. Cambridge, MA: National Bureau of Economic Research.
Gradstein M (2003) The political economy of public spending on education, inequality, and
growth. World Bank Policy Research Working Paper No. 3162. Washington, DC: World
Bank.
Grindle M (ed.) (1980) Politics and policy implementation in the Third World. Princeton,
NJ: Princeton University Press.
Grindle M (2004) Despite the odds: the contentious politics of education reform.
Princeton, NJ: Princeton University Press.
Grindle M, Thomas J (eds) (1991) Public choices and policy change: the political economy
of reform in developing countries. Baltimore, MD: Johns Hopkins University Press.
Gupta S, Davoodi HR, Tiongson E (2000) Corruption and the provision of health care and
education services. Washington, DC: International Monetary Fund.
Hanushek EA (2003) The failure of input-based schooling policies. Economic Journal

113(485): F64–F98.
Harber C (2002) Education, democracy and poverty reduction in Africa. Comparative

Education 38(3): 267–276.
Hecock RD (2006) Electoral competition, globalization, and subnational education

spending in Mexico, 1999–2004. American Journal of Political Science 50(4): 950–961.
Hicken A, Simmons JW (2008) The personal vote and the efficacy of education spending.
American Journal of Political Science 52(1): 109–124.
59
Hoffman BD (2013) Political economy of Tanzania. Washington, DC: Georgetown

University, Centre for Democracy and Civil Society.
Hoxby C (1996) How teachers' unions affect education production. Quarterly Journal of
Economics 111(3): 671–718.
Hurst P (1983) Implementing Educational Change: a Critical Review of the Literature. EDC
Occasional Papers no 5. London: University of London Institute of Education.
Iyer L, Mani A (2008) Traveling agents: political change and bureaucratic turnover in
India. Draft.
Kempner K, Loureiro A. (2002) The global politics of education: Brazil and the World Bank.
Higher Education 43(3): 331–354.
King K (2007) Balancing basic and post-basic education in Kenya: national versus
international policy agendas. International Journal of Educational Development 27(4):
358–370.
Kingdon G (2010) The impact of the sixth pay commission on teacher salaries: assessing
equity and efficiency effects, RECOUP Working paper No. 29, University of Cambridge.
Kingdon K, Aslam M, Rawal S, Das S (2013) Are contract teachers and para-teachers a cost-
effective intervention to address teacher shortage and improve learning outcomes?
London: EPPI-Centre, Social Science Research Unit, Institute of Education, University of
London.
Kingdon G, Muzammil M (2003) The political economy of education in India, teacher

politics in Uttar Pradesh. Oxford: Oxford University Press.
Kingdon G, Muzammil M (2009) A political economy of education in India: the case of Uttar
Pradesh. Oxford Development Studies 37(2): 123–144.
Kingdon G, Muzammil M (2013) The school governance environment in Uttar Pradesh:

implications for teacher accountability and effort. Journal of Development Studies 49(2)
251–269.
Kingdon G, Teal F (2007) Does performance related pay for teachers improve student
achievement? Some evidence from India. Economics of Education Review 26(4): 473–486.
Kingdon G, Teal F (2010) Teacher unions, teacher pay and student performance in India: a
pupil fixed effects approach. Journal of Development Economics 91(2): 278–288.
Kosack S (2009) Realising Education for All; defining and using political will to invest in
primary education. Comparative Education 45(4): 495–523.
Kremer M (2003) Randomized evaluations of educational programs in developing countries:

some lessons. American Economic Review 93(2): 102–106.
Languille S, Dolan J (2012) Political economy analysis of the education sector in Tanzania.
Report for DFID.
60
References
Leftwich A (2006) Drivers of change: refining the analytical framework: conceptual and
theoretical issues. York: University of York, Department of Politics.
Levacic R (2009) Teacher incentives and performance: an application of principal-agent

theory. Oxford Development Studies 37(1): 33–46.
Lewin KM (1991) Chapter 1 Introduction in Lewin K M, with Stuart J S (eds) Educational

Innovation in Developing Countries; Case Studies of Change Makers. MacMillan. 325pp.
ISBN (limp-back) ISBN 0-333-49094-0 (Hard-back) (2nd edition, first edition 1991)
Keith ML, Wang YJ, Chang HQ, Wang L, Yong LJ, Jiaqi WZK (1994) Implementing basic
education in China: progress and prospects in rich, poor and national minority areas. IIEP
Research Report No. 101, UNESCO: International Institute for Educational Planning, Paris
Little A (1999) Labouring to learn: towards a political economy of plantations, people and
education in Sri Lanka. London: Macmillan.
Little A (2008) EFA politics, policies and progress. CREATE Research Monograph No. 13.
London: University of London.
Little A (2010a) Access to elementary education in India: policies, politics and progress.
CREATE Research Monograph No. 44. London: University of London.
Little A (2010b) Access to basic education in Ghana: politics, policies and progress.
CREATE Research Monograph No. 42. London: University of London.
Little A (2010c) The politics, policies and progress of basic education in Sri Lanka. CREATE
Research Monograph No. 38. London: University of London.
Little AW (2011) Education policy reform in Sri Lanka: the double-edged sword of political
will. Journal of Education Policy 26(4): 499–512.
Lott J, Kenny LW (2013) State teacher union strength and student achievement. Economics
of Education Review 35: 93–103.
Mahlangu VP, Pitsoe VJ (2011) Power struggle between government and the teacher unions
in South Africa. Journal of Emerging Trends in Educational Research and Policy Studies
2(5): 365–371.
McKinsey (2010) How the world’s most improved school systems keep getting better.
November 2010. London: McKinsey & Company.
Mehrotra S (1998), Education for all: policy lessons from high-achieving countries.
International Review of Education, Volume 44, Issue 5/6: pp. 461-484.
Moe TM (2001) Private vouchers: politics and evidence. In: Wang MC, Walberg HJ (eds)
Education in cities: school choice versus best systems. Mahwah, NJ: Lawrence Erlbaum.
Moe TM (2006) Political control and the power of the agent. Journal of Law, Economics
and Organization 22(1): 1–29.
61
Moe TM (2011) Special interest: teachers unions and America’s public schools.
Washington, DC: Brookings Institution Press.
Moe TM (2013) Vested interests, theory, and the political dynamics of American education
Stanford University, USA.
Mukundan MV, Bray M (2004) The decentralisation of education in Kerala State, India:
rhetoric and reality. International Review of Education 50(3/4): 223–243.
Mulkeen A (2010) Teachers in Anglophone Africa: issues in teacher supply, training, and
management. Washington, DC: World Bank.
Mundy K (2006) Education for All and the New Development Compact. Review of Education
52: 23–48.
Muralidharan K, Sundararaman V (2011) Teacher performance pay: experimental evidence

from India. Journal of Political Economy 119(1): 39–77.
Murillo MV, Ronconi L (2004) Teachers' strikes in Argentina: partisan alignments and
public-sector labor relations. Studies in Comparative International Development 39(1): 77–
98.
Murillo MV, Tomassi M, Ronconi L, Sanguinetti J (2002) The economic effects of unions in
Latin America: teachers' unions and education in Argentina. Research Department
Publications. Washington, DC: Inter-American Development Bank, Research Department
No. 3156.
Packer S (2007), International EFA architecture lessons and prospects: a preliminary

assessment, Background paper prepared for the Education For All Global Monitoring Report
2008, UNESCO.
PAISA (2012) Do schools get their money? PAISA 2012. New Delhi: Accountability Initiative
and ASER, India.
Park N, Somanathan R (2004), Patronage in public administration: presidential

connections, position assignments and the performance of Korean public prosecutors,
1992-2000. Indian Statistical Institute, Planning Unit, New Delhi Discussion Papers 04-03,
Indian Statistical Institute, New Delhi, India.
Patrinos H, Kagia R (2007) Maximizing the performance of education systems: the case of
teacher absenteeism. In: Campos, JE, Pradhan S (eds) The many faces of corruption.
Washington, DC: World Bank.
Pedley D, Taylor D (2009) Politics and policy in education: aligning and harmonising with
what country priorities. In: A safari towards aid effectiveness. June 2009.
www.norrag.org/fr/publications/norrag-news/online-version/a-safari-towards-aid-
effectiveness.html (accessed 21 January 2014).
Pherali T, Smith A, Vaux T (2011) A political economy analysis (PEA) of education in

Nepal. May 2011. London: DFID.
The Pratichi Report (2002). Pratichi India Trust.
62
References
Psacharopoulos G (1989) Why educational reforms fail: a comparative analysis.

International Review of Education 35(2): 179–195.
Ramachandran V, Madhumita P, Sharada J, Shekar S, Sharma J (2005) Teacher Motivation

in India, DFID, UK.
Rawal S, Kingdon G (2010), Akin to my teacher: does caste, religious or gender distance
between student and teacher matter? Some evidence from India, DoQSS Working Paper 10-
18, Department of Quantitative Social Science Institute of Education, University of London
Santibañez L, Rabling B (2006) Muscle, timing and priorities: teacher unions and education
quality in Mexico.
Santibañez L, Rabling B (2008) Conflict and power: the teachers' union and education
quality in Mexico. Well-Being and Social Policy 3(2): 21–40.
Samoff J (2003) No teacher guide, no textbooks, no chairs: contending with crisis in

African education, in R. F. Arnove and C. A. Torres (Eds.) Comparative Education: the
dialectic of the global and the local (second edition), Lanham, MD: Rowman and
Littlefield.
Schleicher A (2011) Building a High-Quality Teaching Profession Lessons from around the
World, OECD publishing.
Sebatane M (2000) International transfers of assessment: recent trends and strategies.

Assessment in Educational Principles, Policy and Practice 7(3): 401–416.
Sharma R (2009) The internal dynamic. In: Sharma R, Ramachandran V (eds) The
elementary education system in India: exploring institutional structures, processes and
dynamics. New Delhi: Routledge.
Silova I, Brehm W (2013) The shifting boundaries of teacher professionalism: education

privatisation in the post-socialist education space. In: Seddon T, Steiner-Khamsi G, van
Zanten (eds) A world year book of education. London: Routledge, pp. 55–74.
Somerset A (2011) Access, cost and quality: tensions in the development of primary
education in Kenya. Journal of Educational Policy 26(4): 483–499.
Sørensen BR (2008) The politics of citizenship and difference in Sri Lankan schools.
Anthropology and Education Quarterly 39(4): 423–443.
Stasavage D (2005) Democracy and education spending in Africa. American Journal of

Political Science 49(2): 343–358.
Transparency International (2005) Stealing the Future: Corruption in the Classroom

Berlin.
Wade (1985) The Market for Public Office. Why the Indian State is Not Better at
Development. World Development.
63
Woessmann L (2003) Schooling resources, educational institutions and student

performance: the international evidence. Oxford Bulletin of Economics and Statistics
65(2): 117–170.
Zengele T (2013) The unionization of the teaching profession and its effects on the South
African education system: teacher unionism in South Africa. Journal of Social
Science 35(3): 231–239.
64
Appendices
Appendix 1: Authorship of this report
Researchers
Geeta Gandhi Kingdon (Lead Investigator), Institute of Education, University of London
Angela Little, Institute of Education, University of London
Monazza Aslam, Institute of Education, University of London
Shenila Rawal, Institute of Education, University of London
Terry Moe, Stanford University, California
Advisors
Harry Patrinos, The World Bank, Washington, DC
Tara Beteille, The World Bank, Washington, DC
Rukmini Banerji, Pratham Education Foundation, New Delhi
Research assistants
Brent Parton, The World Bank, Washington, DC
Shailendra K. Sharma, Pratham Education Foundation, New Delhi
Contact details
Geeta Gandhi Kingdon
Institute of Education
University of London
20 Bedford Way
London WC1H0AL
E-mail: g.kingdon@ioe.ac.uk
Tel: + 44 (0) 7612 6000
65
Appendix 2: Search Strategy
Table A2.1: Electronic databases

Name
1. EBSCO
a. EconLit
b. Education Full Text
2. EconPapers
3. JSTOR
4. ProQuest
a. ERIC
b. International Bibliography of the Social Sciences
c. PAIS International
d. ProQuest Dissertation
e. Worldwide Political Science Abstracts
5. Science Direct
6. Social Science Research Network
7. Web of Knowledge
Web of Science
8. Google Scholar
9. Research Papers in Economics
Table A2.2: Organisational websites

Organisational websites
1. Centre for International Development at Harvard University
www.hks.harvard.edu/centers/cid/publications
2. Consortium for Research on Educational Access, Transitions and Equity (CREATE)
www.create-rpc.org/database/
3. International Labour Organization (ILO)
www.ilo.org/sector/lang--en/index.htm
3. United Nations Development Programme (UNDP)
www.undp.org/content/undp/en/home/librarypage.html
4. United Nations Educational, Scientific and Cultural Organization (UNESCO)
www.unesco.org/new/en/unesco/resources/publications/
5. World Bank
http://publications.worldbank.org
6. Department for International Development, UK (DFID)
www.gov.uk/government/organisations/department-for-international-development
66
Table A2.3: Scholarly journals

Journals
International and Comparative Education
Comparative Education Review
Comparative Education
International Journal of Educational Development
International Review of Education
Prospects
OECD Education Focus
Journal of Education Policy
Oxford Review of Education
Education, Knowledge and Economy
Review of Education, Pedagogy and Cultural Studies
Peabody Journal of Education
Educational Policy
Journal of Educational Change
Journal of Education and Practice
Globalisation, Societies and Education
British Journal of Sociology of Education
Sociology of Education
Educational Researcher
Educational Evaluation and Policy Analysis
American Educational Research Journal
Economics, Political Science and Sociology
Political Research Quarterly
Political Science Quarterly
The American Political Science Review
Political Research Quarterly
Journal of Political Economy
67
Review of International Political Economy
Review of African Political Economy
American Sociological Review
International Journal of Social Economics
68
Table A2.4: Search terms, keywords and countries

Search terms
“political economy of education”
“teacher(s) unions/organizations”
“block(ing)/resist(ing)/oppose(ing) reform in education”
“enable (ing)/advance (ing)/partner (ing) reform in education”
“implement (ing) reform in education”
“teacher(s) strike”
“politicization (politics) of education”
“politicization (politics) of schools”
“politicization (politics) of teachers”
“education and principal agent theory”
“teachers and principal agent theory”
“rent-seeking and education”
Keywords
Education Political Issues-based (prefix)
education, political economy governance (school,
education)
schools, schooling politics, political transfers (teacher)
teacher (s) politicisation reassignments (teacher)
educator (s) policymaking, policy- funding (school, education)
making
students policy formulation (de) (re) centralization
(education)
school boards policy implementation accountability (school,
education, teacher)
school management decision-making, dismissal (teacher)
committees decision-makers
reform testing, (student, school,
education)
resist (ing) reform finance (school, education)
block (ing) reform assessment (school,
education, teacher, student)
union(s), unionism efficiency (education)
rent-seeking equity (education, school)
principal agent theory evaluation (school, teacher)
strike (s) achievement (student, school)
politicians outcomes (education)
opposing reform quality (teacher, school,
education)
conflict parents, parental involvement
corruption community, community
involvement
special interests inspection (school)
patronage aid, foreign aid (education)
patron client donor (s) (education)
policy, reform agenda privatisation (education,
school)
legislation bureaucracy (education,
school)
legislators NGOs (education)
enable (ing) reform practice (education, teaching)
advance (ing) reform curricula (reform)
69
implement (ing) reform management (education,

school)
partner (ing) reform recruitment (teacher,
principal)
politicians vouchers (school, education)
parties (political)
party concentration
voter (s)
Countries
Afghanistan
Bangladesh
Burma
Democratic Republic of the Congo
Ethiopia
Ghana
India
Kenya
Kyrgyzstan
Liberia
Malawi
Mozambique
Nepal
Nigeria
Palestine (West Bank,
Gaza)
Pakistan
Rwanda
Sierra Leone
Somalia
South Africa
Sudan
South Sudan
Tajikistan
Tanzania
Uganda
Yemen
Zambia
Zimbabwe
70
Table A2.5: Search strategy sequence

Search String Groupings
1. Topical search terms TITLE
2. Topical search terms ABSTRACT
3. Topical search terms SUBJECT
4. Education AND Political TITLE
5. Education AND Political ABSTRACT
6. Education AND Political SUBJECT
7. Education AND Political AND Country TITLE
8. Education AND Political AND Country ABSTRACT
9. Education AND Political AND Country SUBJECT
10. Political AND Issue-based TITLE
11. Political AND Issue-based ABSTRACT
12. Political AND Issue-based SUBJECT
13. Political AND Issue-based AND Country TITLE
14. Political AND Issue-based AND Country ABSTRACT
15. Political AND Issue-based AND Country SUBJECT
71
Appendix 3: Search syntax
1. EBSCO
Search syntax
 Boolean operators can be used within and between fields – AND, OR, NOT
 Searches are run separately in title, abstract and subject – TI, AB, SU. Searches can
 Wildcards are entered as ? (1 missing character) or # (0 or 1 missing characters).

also be run separate by database, to allow individual searches of thesauruses.
 Proximity searches are implemented using Nx (within x words of each other, any
 Parentheses ( ) are used to group terms. If search for (x) and (y), the database first
order) and Wx (within x words of each other, in order specified in search).
 Exact phrases can be searched for using quotation marks “x”. Quotation marks as
finds x, then within those result, finds y.
in the search string below will not be recognised when copied and pasted. They
 Punctuation internal to phrases does not affect searches e.g. “decision making”
must be entered manually in the search window.
will find “decision-making.”
Initial search
 Search mode set to ‘Boolean/Phrase’
 Searches are run separately in title, abstract and subject – TI, AB, SU.


Results limited to 2000–13.
No filter by Source type was applied (e.g. Academic Journals and Books).
EconLit
TI (("education" OR "schooling" OR "teacher" OR "educator")) AND TI (("political economy"
OR "politics" OR "political"))
112 hits
AB (("education" OR "schooling" OR "teacher" OR "educator")) W3 AB (("political economy"

92 hits
SU (("education" OR "schooling" OR "teacher" OR "educator")) AND SU (("political economy"

124 hits
Education Full Text

TI (("education" OR "schooling" OR "teacher" OR "educator")) AND TI (("political economy"
485 hits
AB (("education" OR "schooling" OR "teacher" OR "educator")) W3 AB (("political economy"

298 hits
SU (("education" OR "schooling" OR "teacher" OR "educator")) AND SU (("political economy"

111 hits
72
2. EconPapers
Search syntax
 Search fields: the available search fields are author, keywords and title, JEL code,
 Exact phrases: these can be searched for using quotation marks. These need to be
and free text.
manually typed into the engine; copy and paste will not input quotation marks,
 Combining terms: the standard Boolean operators AND, OR, NOT are supported.
which the search engine recognises.
 The default setting is AND. Parentheses are also supported.

 Proximity searches: the operator NEAR is supported.
 Wildcards: the * wildcard is supported. This indicates any string of characters,
including none.
Initial search
 Search in Keywords & Title among working papers and articles and books &
 Date restrictions are not possible.

chapters
("education" OR "schooling" OR "teacher" OR "educator") AND ("political economy" OR

"politics" OR "political")
439 hits
3. JSTOR
Search syntax
 Search fields: JSTOR is non-bibliographic, and does not contain any thesaurus. It
 Exact phrases: quotation marks are used to define exact phrases, and brackets to
also contains abstracts for only 10 percent of articles contained on the database.
 Combining terms: standard Boolean operators apply – AND, OR, NOT. The default
delimit search fields.
 Proximity searches: are implemented using the tilde symbol, with ~x denoting
operator is AND.
‘within xX words of each other, in any order’”. However, such searches can only be
implemented across single terms, not in terms using Boolean operators, so cannot
 Plurals: adding & at the end of a word specifically searches for both singular and
be applied here.
plural forms at the same time. This includes cases where plural and singular are
 Wildcards: ? searches for a single character, * searches for multiple characters, #

spelled entirely differently.
finds all variations on a given word, e.g. operate# finds operator, operating,
operation, and so on. However, only four wildcards can be included in any given
search.
Initial
search


Do not restrict by languages, or by type of publication

Results restricted by date: 2000–13.
Restricted to journals in the following fields: economics, education, political
 Used one initial string, searching separately in title and abstract:

science, public policy and administration, sociology
(ti:(("education" OR "schooling" OR "teacher&" OR "educator&") AND ("political economy"

OR "politics" OR "political")) AND ab:(("education" OR "schooling" OR "teacher&" OR
73
"educator&") AND ("political economy" OR "politics" OR "political"))) AND disc:(economics-

discipline OR education-discipline OR political science-discipline OR public policy-
discipline OR sociology-discipline)
147 hits
4. ProQuest
Search syntax
 Databases: Proquest is a platform with access to a number of databases. The
search strategy proposes initially searching ERIC, IBSS, PAIS International, ProQuest
 Thesaurus terms are database specific, so each database must be searched

Dissertation, and Worldwide Political Science Abstract.
 Separate codes with commas can search two fields at once, e.g. TI,AB (education)
individually.
 Descriptors are referred to as subject terms, and can be searched using the field
will search both title and abstract for education.
 Exact phrases: exact terms are specified using “x”, and brackets can also be used.
DE.
 Quotation marks must be typed directly into the search engine, the versions
appearing below are not recognised. Punctuation marks inside quotation marks are
 Combining terms: standard Boolean search terms apply – AND, OR, NOT. These can
ignored.
 Proximity searches: can be implemented as within searches, which ignore word

be applied across or within fields.
order, using W/x. Can also be implemented as pre-searches, which retain word
 Wildcards: standard wildcard characters can also be used, with * for any number of
order, using P/x.
characters, and ? for one character only.
Initial search


For the purpose of the test search, ERIC and IBSS are searched.

Search in title, abstract and subject, and separately in thesaurus,
Restrictions by date: 2000–13.
ERIC
TI (("education" OR "schooling" OR "teacher" OR "educator") AND ("political economy" OR
"politics" OR "political")) AND YR(>=2000)
538 hits
AB (("education" OR "schooling" OR "teacher" OR "educator") P/3 ("political economy" OR

407 hits
SU (("education" OR "schooling" OR "teacher" OR "educator") AND ("political economy" OR

7027 hits
Given the large amount of hits from the preceding search string process, a simplified,
targeted search of pre-identified search terms was employed. The results are as follows:
TI (political economy of education)

22
74
AB(political economy PRE/3 education)

47 hits
SU (political economy of education)

5 hits
SU ("Politics of Education") AND ("political economy")

38 relevant hits
International Bibliography of the Social Sciences

A high volume of results returned from the following search strings:
TI (("education" OR "schooling" OR "teacher" OR "educator") AND ("political economy" OR

AB (("education" OR "schooling" OR "teacher" OR "educator") P/3 ("political economy" OR

SU (("education" OR "schooling" OR "teacher" OR "educator") AND ("political economy" OR

In turn, the test process employed a targeted search of pre-identified search terms. The
results are as follows:
TI (“political economy of education”)

11 hits
AB (“political economy of education”)

1 hit
SU (political economy PRE/3 education)

22 results
A secondary “unbounded” search using these search terms was then employed:
TI (political economy of education)

25 hits
AB (political economy of education)

371 hits
SU (political economy of education)

103 hits
75
Appendix 4: Describing the evidence
Table A4.1: Describing studies

Code Types of research design
RCT Randomised controlled trial
QE Quasi- experimental
IE Impact evaluation
EMP-Quant Quantitative analysis of observational data
EMP-Qual Qualitative analysis of observational data
CS Case study (like books, case studies may include several
designs, most usually EMP-Quant and EMP-Qual)
ETH Ethnography
SR Systematic review
THE Theoretical
TP Think piece
BK Entire book (may include several designs)
LR Literature review
Table A4.2: Assessing the strength of evidence

Quality of the body Definition
of evidence
Robust evidence Many/the large majority of single studies reviewed have been assessed as
being of a high quality, demonstrating adherence to the principles of
rigour, validity and reliability.
Modest evidence Of the single studies reviewed, approximately equal numbers are of a
high, moderate and low quality, as assessed according to the principles of
rigour, validity and reliability.
Insufficient evidence Many/the large majority of single studies reviewed have been assessed as
being of low quality, showing significant deficiencies in adherence to the
principles of rigour, validity and reliability.
Source: adapted from DFID (2013), p. 16.
76
Table A4.3A: Summary of studies reviewed (from initial searches)

Paper Quality
# Author (date) Research design Country rankings
1 Altschuler (2013) IE Honduras/Guatemala Medium-high
2 Alvarez et al. (2007) EMP-Quant Mexico Medium-high
3 Ansell (2008) EMP-Quant Multi-country (113) Medium-high
4 Archer (1981) THE Medium
5 Day Ashley (2013) EMP-Qual India Medium
6 Béteille (2009) EMP-Qual India Medium-high
7 Bermingham (2011) EMP-Qual Multi-country Medium
8 Bourguignon and Verdier (2005) THE Medium
9 Bruns et al. (2011) EMP-Quant Medium-high
Chile, Colombia, Mexico, Mozambique, South Africa, Zambia
10 Buchert (1998) BK and Zimbabwe Medium
11 Chen (2011) EMP-Quant Indonesia Medium-high
12 Corrales (2005) TP Medium
13 Essuman and Akyeampong (2011) CS EMP-Qual Ghana Medium
14 Evans et al. (1995) CS EMP-Qual Benin, Ghana, Guinea, Mauritius, Mozambique and Uganda Medium
15 Fernandez (2011) EMP-Quant Mexico Medium-high
16 Gradstein (2003) THE Medium
17 Harber (2002) TP Medium
18 Hicken and Simmons (2008) EMP-Quant Multi-country (40) Medium-high
19 Hecock (2006) EMP-Quant Mexico Medium-high
20 Iyer and Mani (2008) EMP-Quant India Medium-high
21 Kempner and Loureiro (2002) TP Brazil Medium
22 King (2007) TP Kenya Medium
23 Kingdon and Muzammil (2009) EMP-Quant India – Uttar Pradesh Medium
24 Kingdon and Muzammil (2013) EMP-Quant and EMP-Qual India – Uttar Pradesh Medium-high
25 Kingdon and Teal (2010) EMP-Quant India Medium-high
CS EMP-Qual and EMP-
26 Kosack (2009) Quant; also THE Ghana/Taiwan Medium
77
27 Kremer (2003) TP/LR Multi-country Medium

28 Leftwich (2006) THE Multi-country Medium-high
29 Levacic (2009) THE Medium
BK/EMP-Qual and EMP-
30 Little (1999) Quant Sri Lanka Medium
31 Little (2008) LR Medium
32 Little (2010b) EMP-Qual and EMP-Quant Ghana Medium
33 Little (2010c) EMP-Qual and EMP-Quant Sri Lanka Medium
34 Little (2010a) EMP-Qual and EMP-Quant India Medium
35 Little (2011) EMP-Qual Sri Lanka Medium
36 Mahlangu and Pitsoe (2011) TP South Africa Medium
37 Mukundan and Bray (2004) TP India – Kerala Medium
38 Mundy (2006) TP Medium
Muralidharan and
39 Sundararaman(2011) RCT India – Andhra Pradesh High
40 Murillo and Ronconi (2004) EMP-Quant Argentina Medium
41 Murillo et al. (2002) EMP-Quant Argentina Medium
42 Patrinos and Kagia (2007) BK Multi-country Medium
43 Pedley and Taylor (2009) TP Ghana Medium
44 Psacharopoulos (1989) TP East Africa Medium
45 Santibáñez and Rabling (2006) EMP-Quant Mexico Medium
46 Santibáñez and Rabling (2008) EMP-Quant Mexico Medium
47 Sharma (2009) EMP-Qual India Medium
48 Somerset (2011) EMP-Qual and EMP-Quant Kenya Medium
49 Sørensen (2008) TP Sri Lanka Medium
50 Stasavage (2005) EMP-Quant Several African countries Medium-high
51 Woessmann (2003) EMP-Quant Multi-country Medium-high
78
Table A4.3B: Summary of studies reviewed (from additional searches – expert advice)
Paper Quality
# Author (date) Research design Country rankings
52 Barber (2013) TP Pakistan Medium-low
53 Bennell and Akyeampong (2007) MX Multi-country Medium
54 Bold et al. (2013) EMP-Quant Kenya High
55 CfBT (2011) EMP-Qual Zimbabwe Medium
56 Duncan and Williams (2010) TP Nigeria Medium-low
57 Education International (2013) EMP-Quant Ghana Medium
58 Hoffman (2013) EMP-Qual Tanzania Medium
59 Kingdon and Muzammil (2009) EMP-Quant India Medium
60 Languille and Dolan (2012) EMP-Qual Tanzania Medium
61 McKinsey (2010) MX Multi-country Medium
62 Mulkeen (2010) MX Multi-country Medium
63 Pherali et al. (2011) EMP-Qual Nepal Medium
64 Zengele (2013) EMP-Qual South Africa Medium
79
This is material has been funded by the Department for International Development. However the
views expressed do not necessarily reflect the Department’s official policies.
The report was designed in April 2014 by Philip Rose, EPPI-Centre, Social Science Research Unit,
Institute of Education, University of London
Tel: +44 (0)20 7612 6397 http://eppi.ioe.ac.uk http://www.ioe.ac.uk/ssru
Kingdon GG, Little A, Aslam M, Rawal S, Moe T, Patrinos H, Beteille T, Banerji R, Parton B and Sharma
SK (2014) A rigorous review of the political economy of education systems in developing countries.
Final Report. Education Rigorous Literature Review. Department for International Development
Cover image © UN Photo/Fardin Waezi
View publication stats

A Rigorous Review of The Political Econo

Uploaded by

Copyright:

Available Formats

A Rigorous Review of The Political Econo

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Rigorous Review of The Political Econo

Uploaded by

Copyright:

Available Formats

Education Rigorous Literature Review

A rigorous review of the political economy of

Geeta Gandhi Kingdon

The EPPI-Centre reference number for this report is 2203.

Abbreviations .......................................................................................... iii

ASER Annual Status of Education Report

CMS Community-managed schools

DFID Department for International Development (UK)

EFA Education for All

ERIC Education Resources Information Centre

FTI Fast Track Initiative

GDP Ghana National Association of Teachers

GPE Global Partnership for Education

IBSS International Bibliography of the Social Sciences

INGO International non-governmental organisation

KNUT Kenya National Union of Teachers

NGO Non-governmental organisation

OECD Organisation of Economic Co-operation and Development

PEA Political economy analysis

PETS Public Expenditure Tracking Surveys

PISA Programme for International Student Assessment

QSDS Quantitative Service Delivery Surveys

RCT Randomised controlled trial

RePEc Research Papers in Economics

SADTU South African Democratic Teachers Union

SBM School-based management

SMC School management committee

TIMSS Trends in International Mathematics and Science Study

TTU Tanzania Teachers’ Union

UNESCO United Nations Economic, Scientific and Cultural Organization

Roles and responsibilities

Rent-seeking and patronage politics

Decision-making and the process of influence

Explicit inclusion/exclusion criteria were used to conduct searches of bibliographical

3.2 Search strategy

conducted by examining the references of identified papers and directly contacting

3.3 Search criteria

3.4 Testing search strategy for electronic databases

3.5 Describing the evidence

3.6 Assessing the quality of evidence

One-stage screening Two-stage screening

485 citations identified

Title and abstract Citations excluded

48 citations (brought forward for

*48 reports obtained

There are some additional studies that

4.1 Roles and responsibilities

The political economy of education is influenced by a number of internal and external

School management, parents, elites and civil society as stakeholders

Government officials as stakeholders

International actors and agencies as stakeholders

4.2 Rent-seeking and patronage politics

Rent-seeking (an attempt to gain economic rent by influencing policies or their

The universality of public education may also be affected by rent-seeking (Gradstein

In a striking example of patronage politics in education in Africa, Zengele (2013) discusses

4.3 Educational decision-making and the process of influence

What mechanisms are available to different power groups?

Building-in incentives into the teacher salary structure

The politics of decentralisation in education