EFA GlobalMonitoring Report

Educ

atio

n fo

r Al

l

0 1 02

Reachingthe marginalizedReachingthe marginalizedReachingthe marginalized

Reaching the marginalized

Reaching the marginalized

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

This Report is an independent publication commissioned by UNESCO on behalf of the international community. It is the product of a collaborative effort involving members of the Report Team and many other people, agencies, institutions and governments.

The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of UNESCO concerning the legal status of any country, territory, city or area, or of its authorities, or concerning the delimitation of its frontiers or boundaries.

The EFA Global Monitoring Report Team is responsible for the choice and the presentation of the facts contained in this book and for the opinionsexpressed therein, which are not necessarily those of UNESCO and do notcommit the Organization. Overall responsibility for the views and opinionsexpressed in the Report is taken by its Director.

Great Clarendon Street, Oxford OX2 6DPOxford University Press is a department of the University of Oxford.It furthers the University’s objective of excellence in research, scholarship, and education by publishingworldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong (China) Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina AustriaBrazil Chile Czech Republic France Greece GuatemalaHungary Italy Japan Poland Portugal Republic of KoreaSingapore Switzerland Thailand Turkey Ukraine Viet NamOxford is a registered trade mark of Oxford UniversityPress in the UK and in certain other countries.

Published jointly by the United Nations Educational,Scientific and Cultural Organization (UNESCO),7, Place de Fontenoy, 75352 Paris 07 SP, France and Oxford University Press, Great Clarendon Street,Oxford OX2 6DP, United Kingdom.

© UNESCO, 2010All rights reservedFirst published 2010Published in 2010 by the United Nations Educational,Scientific and Cultural Organization7, Place de Fontenoy, 75352 Paris 07 SP, France

Graphic design by Sylvaine BaeyensLayout: Sylvaine Baeyens and Tania Hagemeister

Library of Congress Cataloging in Publication DataData availableTypeset by UNESCOOUP ISBN 9780199584987UNESCO ISBN 9789231041297

F O R E W O R D

i

This edition of the EFA Global Monitoring Report 2010, Reaching the marginalized, comes at a timeof great uncertainty. We are still grappling with the far-reaching impact of the global financial andeconomic crisis not only on the world’s banking systems, but on all areas of human development –including education. We are at a crossroads. Either we continue with business as usual and risk undoingthe considerable progress made over the past decade, or we use this crisis as an opportunity to createsustainable systems which promote inclusion and put an end to all forms of marginalization.

The gains achieved since the Education for All and Millennium Development Goals were adopted in 2000are undeniable: great strides have been made towards universal primary education, increasedparticipation in secondary and tertiary education and, in many countries, gender equality. More widely,there have been improvements in overcoming hunger, poverty, and child and maternal mortality.

The global financial crisis could radically change all this. Reaching the marginalized demonstratesthat declining government revenue and rising unemployment now pose a serious threat to progressin all areas of human development. Government budgets are under even greater pressure and fundingfor education is especially vulnerable. So are poor households. Rising poverty levels mean that thechallenge of meeting basic human needs is a daily struggle. Lessons from the past teach us thatchildren are often the first to suffer – as is their chance to go to school.

In response to this crisis, governments urgently need to create mechanisms to protect the poor andvulnerable. They must also seize the opportunity to build societies that combat inequality, so that allmay benefit and prosper. Education is at the front line. Not only do schools teach literacy and laythe groundwork for productive lives, they also play a crucial role in promoting tolerance, peace andunderstanding between peoples, and in fighting discrimination of all kinds. Schools are the place whereindigenous groups can learn to read and write in their mother tongue, where cultural diversity can thriveand where children can try to escape the hardships of conflict and displacement.

This year’s Global Monitoring Report underscores that there is a long way to travel. There are stillat least 72 million children who are missing out on their right to education because of the simple factof where they are born or who their family is. Millions of youths leave school without the skills theyneed to succeed in the workforce and one in six adults is denied the right to literacy.

The 2010 Report is a call to action. We must reach the marginalized. Only inclusive education systemshave the potential to harness the skills needed to build the knowledge societies of the twenty-firstcentury. The international community has a determining role in supporting countries’ efforts to protectand expand their education systems. We must not abandon them at this critical juncture. Promisesto help poor countries out of the crisis must now translate into the financial resources that manygovernments so urgently need.

It is my intention that UNESCO should continue to vigorously advocate for increased investment ineducation. As the lead agency for Education for All, we have a special responsibility to encourage andsupport those most at risk from the present crisis. As we stand at the crossroads, with only five yearsleft to meet our collective commitments, let us have the courage and determination to choose the paththat lets all children, youths and adults fulfil their right to education.

Irina Bokova

Foreword

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

A C K N O W L E D G E M E N T S

i i

This report has been written by the Global Monitoring Report (GMR) team with the support of many peopleand organizations around the world.

The Global Monitoring Report team would like to thank the members of its International Advisory Board for their valuable guidance.

The EFA Report depends greatly on the work and expertise of the UNESCO Institute for Statistics (UIS).We thank its Director, Hendrik van der Pol, along with Claude Apkabie, Saïd Belkachla, Georges Boade,Michael Bruneforth, Talal El Hourani, Friedrich Huebler, Olivier Labé, Weixin Lu, Adriano Miele, Albert Motivans,Juan Cruz Persua, Pascale Ratovondrahona, Ioulia Sementchouk, Anuja Singh, Saïd Ould Voffal, Peter Walletand Yanhong Zhang.

The analysis in the EFA Report is informed by commissioned background papers. We thank all authors whoprepared material for this year’s Report: Abdullahi Haji Abdi, Reda Abou Serie, John Aitcheson, FedericoBlanco Allais, Nadir Altinok, Jane Anthony, Laban Ayiro, Gauthier de Beco, Paul Bennell, Herbert Bergmann,Desmond Bermingham, Sonia Bhalotra, Nicholas Biddle, Lyndsay Bird, Michael Bruneforth, Gwang-Chol Chang,Ian Cheffy, Luis Crouch, Santiago Cueto, Mahamadou Diarra, Janice Dolan, Marie Duru-Bellat, Jude Fransman,Franco Gamboa Rocabado, Malini Ghose, George Godia, R. Govinda, Carola Gruen, Gabriela Guerrero, JialingHan, Ricardo Henriques, George Ingram, Francesca I. Izabel, John Kabutha Mugo, Jean-Francois Kobiané,Ozge Nihan Koseleci, Juan León, Ingrid Lewis, Luis Enrique López, Siobhan Mackay, Cristina Manzanedo,Matthew Martin, Raphaelle Martinez, Geraldo Martins, Juliet McCaffery, Erika Mein, Katharina Michaelowa,Hilaire Mputu, Karen Mundy, Ismael Muñoz, Samir Ranjan Nath, Katy Newell-Jones, Anna Obura, BrendanO’Malley, Zipporah Ongwenyi Nyamauncho, Maciel Pereira, Susan Peters, Marc Pilon, Patrick Quinn, AlanRogers, Sara Ruto, Alan Sanchez-Jimenez, Elisa Seguin, Nidhi Singal, Devi Sridhar, Pablo Stansbery, BrianStreet, Munshi Sulaiman, Tami Tamashiro, Huyen Chi Truong, Yuko Tsujita, Raul Cotera Valdes, ConsueloVélaz de Medrano, Anna Vignoles, Dan Wagner, Peter Wallet and Kazuhiro Yoshida.

We would also like to thank colleagues who helped us find particular information or identify authors forcommissioned background papers, including Izzy Birch, Manuel Contreras, Kate Gooding, Peter Hyll-Larsen,Harounan Kazianga, Louise Meincke, Alemu Melkamnesh, Sophie Qian, Ya Ping Wang and Kai Zhou.

Special mention goes to Kenneth Harttgen, Stephan Klasen and Mark Misselhorn for their work in processingthe database on Deprivation and Marginalization in Education, and to Luc-Charles Gacougnolle for his workon the aid to education database.

We are grateful to the Education Policy and Data Center at the Academy for Educational Developmentfor their work on the EFA global costing exercise, and particularly to Babette Wils and her team. The workbenefited from extensive support and advice from Gwang-Chol Chang in UNESCO. We are also gratefulto Nicola Chanamuto, Caine Rolleston and Jan van Ravens for inputs to specific aspects of the costing work.

We are indebted to several colleagues who reviewed chapters of the report and provided advice and guidance,namely Kwame Akyeampong, Keith Hinchliffe, Jean-Pierre Jarousse and the pôle de Dakar team, Cynthia Lloyd,Paolo de Renzio and Yusuf Sayed, as well as to colleagues who reviewed certain subsections, including ManosAntoninis, Fadila Caillaud, Luis Crouch, Caroline Dyer, David Hulme, Phillippa Lei and Shailen Nandy.

We are indebted to the many colleagues in UNESCO’s Education Sector, particularly Steven Obeegadooand Olav Seim, and to the International Institute for Educational Planning (IIEP), the International Bureau ofEducation (IBE) and the UNESCO Institute for Lifelong Learning (UIL), who all shared their experience with us.UNESCO’s Regional Offices provided advice on country-level activities and commissioned studies. In particular,we would like to express our gratitude to Candy Lugaz and a team from the IIEP Documentation Centre (Corinne Bitoun, Aurore Hagel, Lynne Sergeant, Asunción Valderrama and Aude Zeiler) who, in collaboration

Acknowledgements

A C K N O W L E D G E M E N T S

i i i

with Anton De Grauwe, provided an in-depth review of how marginalization is treated in national education plans. Further thanks are offered to Jean-Mark Bernard, Ulrike Hanemann, Sabine Kube and Clinton Robinson fortheir advice and guidance on literacy issues; to Kenneth Eklindh and Florence Migeon on inclusion; and toJustine Sass and Yong Feng Liu on HIV and AIDS.

We would like to thank Robert Prouty, Kouassi Soman and Mamadou Thiam of the Fast Track Initiative (FTI)Secretariat; Catherine Dom, Steven Lister and Georgina Rawle of the FTI evaluation team; Julia Benn, FredrikEricsson, Valérie Gaveau, Aimée Nichols, Cécile Sangare and Suzanne Steensen of the Development AssistanceCommittee of the Organisation for Economic Co-operation and Development (OECD-DAC); and Dan Coppardand Asma Zubairi with Development Initiatives for their advice on international cooperation and aid to education.

Special thanks to Alexandre Khan and all those who helped make the public consultation for the 2010 Reporta success. Many UNESCO colleagues and partners in non-government organizations went out of their wayto help us gather human interest stories, which are a new feature of the Report this year. They are toonumerous to mention, but we greatly appreciate their time and effort in making materials available to us and,in some cases, speaking to families and children about their daily difficulties and triumphs in getting to school.Equally, we are extremely grateful to those children and their families who took time to explain to us theirviews and experiences of education.

We are grateful to the many colleagues within and outside of UNESCO who have helped with the translationand production of the Report.

The clarity of the Report benefited greatly from the editorial expertise of Rebecca Brite, Andrew Johnstonand Wenda McNevin, as well as from the support of Isabelle Kite, who assisted in developing this year’sbibliography. We also express our appreciation to Jan Worall for preparing the Report’s index.

Nino Muñoz Gomez, Sue Williams and the staff of UNESCO’s Bureau of Public Information provided timelyand energetic assistance in bringing the Report to the attention of global media. A special thanks alsoto Mariso Sanjines.

Any errors or omissions found subsequent to printing will be corrected in the online version at www.efareport.unesco.org

The EFA Global Monitoring Report team

DirectorKevin Watkins

Samer Al-Samarrai, Nicole Bella, Marc Philip Boua Liebnitz, Mariela Buonomo, Stuart Cameron, Alison Clayson, Diederick de Jongh, Anna Haas, Julia Heiss,

François Leclercq, Anaïs Loizillon, Leila Loupis, Patrick Montjourides, Karen Moore, Claudine Mukizwa, Paula Razquin, Pauline Rose, Sophie Schlondorff, Suhad Varin.

For more information about the Report,please contact:The DirectorEFA Global Monitoring Report teamc/o UNESCO, 7, place de Fontenoy75352 Paris 07 SP, FranceEmail: efareport@unesco.orgTel.: +33 1 45 68 10 36Fax: +33 1 45 68 56 41www.efareport.unesco.org

Previous EFA Global Monitoring Reports2009. Overcoming inequality: why governance matters2008. Education for All by 2015 – Will we make it?2007. Strong foundations – Early childhood care and education2006. Literacy for life2005. Education for All – The quality imperative

2003/4. Gender and Education for All – The leap to equality2002. Education for All – Is the world on track?

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C O N T E N T S

i v

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii

List of figures, tables and text boxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi

Highlights of the Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Education at risk:the impact of the financial crisis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Double jeopardy: food prices and financial crisis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Expanding ‘fiscal space’: an Education for All priority . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

The international response: missing a human dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Progress towards the EFA goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Early childhood care and education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Universal primary education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Youth and adult skills – expanding opportunities in the new global economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

Youth and adult literacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

The quality of education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

Estimating the cost of achieving Education for All . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

Reaching the marginalized . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

Measuring marginalization in education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

Getting left behind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

Levelling the playing field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

Chapter 3

Chapter 2

Chapter 1

Contents

C O N T E N T S

v

The aid compact: falling short of commitments . . . . . . . . . . . . . . . . . . . . . . . . . 214Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

Aid for education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

Reforming the Fast Track Initiative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

Rising to the EFA challenge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268Overcoming education marginalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

Annex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277The Education for All Development Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

Selected international human rights treaties relevant to the EFA goals . . . . . . . . . . . . . . . . . . . . . . . . . 292

Educational effect of selected social protection programmes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294

Statistical tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296

Aid tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 486

Chapter 5

Chapter 4

Figures1.1: Post-crisis economic growth projections have been revised downwards for all developing regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

1.2: Economic growth matters for education financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1.3: Education financing in sub-Saharan Africa could suffer from slower economic growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1.4: Many countries lack room for manoeuvre in budget management but could use more aid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

1.5: The World Bank has front-loaded concessional International Development Association loans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.1: High levels of child stunting are holding back progress in education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

2.2: Low birth weight sets the scene for lifelong disadvantage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

2.3: Educated mothers have better access to antenatal care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

2.4: Wealth-based gaps in learning begin early and widen over time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

2.5: Children from rich families are more likely to participate in early childhood programmes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

2.6: Children of educated mothers are more likely to attend pre-school programmes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

2.7: Numbers of out-of-school children are declining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

2.8: Missing the target — out-of-school trends projected to 2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

2.9: Different stories — administrative and household measurement of children in school can differ greatly . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

2.10: In Senegal, estimates of children in school by age vary with data sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

2.11: Children in sub-Saharan Africa are the least likely to enter school . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

2.12: A child’s prospects of entering and staying in school vary by country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

2.13: Left behind: out-of-school girls are less likely ever to get into school . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

2.14: Poor and rural children have much less chance of going to school in Burkina Faso and Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

2.15: Most countries improved their primary school enrolment between 1999 and 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

2.16: The relationship between enrolment and gender parity varies across countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

2.17: The gender gap is narrowing, but sometimes because enrolment is declining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

2.18: In Yemen, girls’ enrolment is lowest in the poorest and rural areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

2.19: Pakistan’s primary school attendance is marked by gender, regional and wealth inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

2.20: Most out-of-school children are in poorer countries, but some wealthier countries are underperforming . . . . . . . . . . . . . . . . . . . . . . . . 70

2.21: Children’s precarious pathway from school entry to completion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

2.22: Children who start primary school have varying chances to complete the last grade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

2.23: Many adolescents are out of school, or still in primary school . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

2.24: By age 15, many students in developing countries are nearing the end of their schooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

2.25: Gender inequalities reinforce high levels of youth unemployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

2.26: In most OECD countries, youth face greater risk of unemployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

2.27: In the Islamic Republic of Iran, vocational tracking comes with high dropout rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

2.28: Unemployment increases with level of education, but Iranian women are especially penalized . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

2.29: Adult illiteracy is heavily concentrated in a small group of large-population countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

2.30: In developing countries, illiteracy can affect up from one to three out of four adults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

2.31: The number of adult illiterates is falling despite population growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

2.32: Being so far behind, women have further to travel to reach male literacy rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

2.33: Contrasting experiences in reducing illiteracy and the associated gender gap in four countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

2.34: Within countries, women’s literacy rates are influenced by socio-economic and geographic factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

2.35: At the present rate, regions furthest behind will miss the literacy target for 2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

2.36: There are large gaps in learning achievement across countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C O N T E N T S

v i

List of figures, tables and text boxes

C O N T E N T S

L i s t o f f i g u r e s , t a b l e s a n d t e x t b o x e s

v i i

2.37: There are wide disparities across countries in primary school mathematics performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

2.38: Reading ability in secondary school also varies greatly across countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

2.39: Latin America’s reading league has large performance gaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

2.40: Learning gaps are higher in poor countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

2.41: When schools make a difference — inequality in student performance across schools varies widely in rich countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

2.42: Trained teachers are sometimes in short supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

2.43: National averages can hide large differences in pupil/teacher ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

2.44: The rate at which new teaching posts are created will need to increase if universal primary education is to be achieved by 2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

2.45: Spending on teachers has to rise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

2.46: Many countries need more classrooms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

2.47: Current national spending falls short of the levels needed to achieve basic education goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

2.48: Many countries can mobilize additional domestic resources for basic education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

2.49: Financing gaps are large and unlikely to be eliminated by current donor pledges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

3.1: Measuring education poverty across countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

3.2: Slow progress for Africa’s poorest children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

3.3: Education poverty falls with rising income — but the association varies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

3.4: The education inequality tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

3.5: Pastoralists face extreme education deprivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

3.6: Many countries have large regional disparities in education poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

3.7: Education poverty is high in some of Uganda’s northern districts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

3.8: The Philippines has large wealth gaps in education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

3.9: Children in poor, remote, or conflict-affected regions of the Philippines suffer higher levels of education poverty . . . . . 149

3.10: Wealth and gender widen indigenous education disparities in Latin America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

3.11: The language gap in educational opportunity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

3.12: Poverty, ethnicity and language fuel education marginalization in Turkey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

3.13: The poorest households are more likely to be left behind in education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

3.14: Language often predicts risk of being in the bottom 20% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

3.15: Some regions face acute education deprivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

3.16: Overlapping disadvantages erode education opportunities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

3.17: Small groups, big disadvantages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

3.18: Socio-economic disadvantage in education weighs more heavily in some countries than others . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

3.19: Same country, different worlds of learning achievement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

3.20: Second-generation immigrants in rich countries perform far below native students in science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

3.21: Indigenous Australians perform consistently below the student average in reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

3.22: Mapping global poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

3.23: Patterns of school and work vary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

3.24: In India, scheduled castes and tribes remain disadvantaged at all levels in education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

3.25: Social stigma can undermine test performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

3.26: Poor slum dwellers in Bangladesh depend on non-government education provision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C O N T E N T S

v i i i

3.27: Many of Kenya’s arid districts are left behind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

3.28: Burkina Faso’s children with disabilities face deep but varied levels of disadvantage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

3.29: The Inclusive Education Triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

3.30: Redistribution of public finance benefits the lowest performing districts in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

3.31: Federal government redistribution leaves large gaps in Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

4.1: Africa faces the greatest projected shortfall in total aid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

4.2: Nearly all donors are falling short of their aid pledges for 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222

4.3: Most G8 countries are falling short of their ‘fair share’ in aid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

4.4: Most non-G8 donors also have a long way to go . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

4.5: Spain is on track to achieve its national aid target while Italy is off track . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

4.6: Aid disbursements to education have been on a steadily rising trend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

4.7: After rising in the early part of the decade, aid commitments to basic education are stagnating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

4.8: The lion’s share of aid is committed by a small group of donors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

4.9: The priority given to low-income countries has not changed since Dakar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

4.10: Only a few donors give priority to basic education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

4.11: The extent to which donors use recipients’ financial systems is not related to their quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

4.12: The use of recipient financial management systems varies by donor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

4.13: Conflict-affected poor countries receive a low share of aid to education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241

4.14: Distribution of aid to education among conflict-affected poor countries is uneven . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241

4.15: Spending per primary school child is low in conflict-affected poor countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

4.16: Peacekeeping and reconstruction in the Democratic Republic of the Congo and Liberia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244

4.17: Between 2002 and 2008, the FTI endorsed thirty-six countries’ national plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253

4.18: A small group of countries dominates donor support of the FTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

4.19: There are long delays between allocation and first disbursement from the Catalytic Fund . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

4.20: Full disbursement of Catalytic Fund grants can take years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

4.21: The Global Fund has broad-based donor support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

Tables1.1: Potential revenue loss in sub-Saharan Africa, 2008–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

1.2: Fiscal space in sub-Saharan Africa, selected countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.1: Pre-primary enrolment and gross enrolment ratios by region, 1999 and 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

2.2: Primary enrolment by region, 1999 and 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

2.3: Number and % of children and adolescents of primary, lower secondary or basic education age not enrolled in primary, secondary or higher education, 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

2.4: Enrolment in technical and vocational education (TVE) by region, 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

2.5: Gross enrolment ratios in secondary and tertiary education, 1999 and 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

2.6: Adult (15 and over) illiteracy rates and numbers, by region, 2000–2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

2.7: Results from early grade reading assessments (correct words per minute) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

2.8: Targets for the global costing exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

2.9: 2015 targets for main cost parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

2.10: Costs of achieving Education for All in low-income countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

2.11: Average annual financing gaps in low-income countries, 2008–2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

C O N T E N T S

L i s t o f f i g u r e s , t a b l e s a n d t e x t b o x e s

i x

3.1: Primary net attendance rates in selected regions and districts of Uganda, by gender, 2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

3.2: Selected education indicators, by region, Mexico, 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

3.3: Deprivation and Marginalization in Education, selected data, latest year available . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

3.4: Poverty and early cognitive development by race, United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

3.5: Education indicators by disability status of head of household and wealth, Philippines and Uganda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

3.6: Low and high performing Brazilian states on education and poverty indicators, 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

4.1: Football revenue and school levy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

4.2: Progress on Paris Declaration targets, 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

Text boxes1.1: The Obama rescue plan — protecting education during the economic downturn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

2.1: Early malnutrition leads to long-term educational damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

2.2: Removing cost barriers to maternal and child health services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

2.3: Cash transfer in Nicaragua — overcoming cognitive deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

2.4: Expansion of early childhood education in Chile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

2.5: Children count — but counting children in school is difficult . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

2.6: Benin — on the right track, but tackling marginalization is a priority . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

2.7: Yemen — making progress towards universal primary education and gender parity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

2.8: Pakistan — gender disparities hold back progress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

2.9: Liberia — slipping back in a post-conflict country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

2.10: Turkey — marginalization keeps universal primary education out of reach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

2.11: Private vocational training in Brazil: widespread and successful . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

2.12: Training, skills and youth exclusion in the Islamic Republic of Iran . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

2.13: Morocco — strengthening vocational governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

2.14: Vocational education in Ghana — limited access and poor quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

2.15: Linking skills and employment — Jóvenes programmes in Latin America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

2.16: Entra 21 — tackling marginalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

2.17: Skills and employment in the United Kingdom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

2.18: Singapore’s ‘jewel in the system’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

2.19: Rich countries — poor literacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

2.20: A new generation of literacy statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

2.21: Brazil — ‘making people literate’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

2.22: Gender parity and learning achievement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

2.23: Improving equity in Poland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

2.24: What are early grade reading assessments and what can they be used for? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

2.25: India — remedial reading classes in Uttar Pradesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

2.26: Information used for the global cost estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

2.27: Basic education financing in the Democratic Republic of the Congo and the Sudan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

3.1: Uganda — universal primary education is in sight, but large pockets of marginalization persist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

3.2: The Philippines — leaving the marginalized behind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

3.3: Monitoring gaps and marginalization — Roma in Europe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

3.4: Hausa girls in northern Nigeria — losing out in education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C O N T E N T S

x

3.5: Mali and Zambia — combining child labour and schooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

3.6: Living with stigma — the ‘rat catchers’ of Uttar Pradesh and Bihar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

3.7: Tackling the ethnic divide in Viet Nam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

3.8: Slums in Dhaka — marginalization with rapid urban growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

3.9: China’s hukou system has restricted education opportunities for migrant children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

3.10: Kenya’s pastoralists — ‘we need schools that follow our herds’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

3.11: The human face of conflict in the Philippines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

3.12: Education destruction and reconstruction in Gaza . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

3.13: Prejudice limits educational opportunities for children with disabilities in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

3.14: ‘Tipping points’ in Harlem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188

3.15: Reaching pastoralists in northern Kenya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

3.16: Reaching the most marginalized in Bangladesh through floating schools and programmes for child labourers . . . . . . . . . . . . . 195

3.17: Addressing educational deprivation in northern Ghana through complementary education provision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

3.18: Achieving ‘Excellence in Cities’? A targeted intervention to support deprived urban schools in England (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

3.19: Promoting respect for Aboriginal languages in Australia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

3.20: Roma children’s right to education — using the law to challenge the state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204

3.21: Recent legal challenges to educational marginalization in the United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

3.22: New Zealand’s Ma–ori Renaissance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

3.23: Ethiopia — Productive Safety Net Programme boosts children’s education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

3.24: Redistributive public financing in Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

4.1: Aid and the financial crisis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

4.2: The G8’s disappointing Accountability Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226

4.3: Assessing the total aid contribution to the education sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

4.4: Spain: political will behind increased aid to basic education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

4.5: Education for All and the football World Cup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

4.6: Nicaragua — strengthening management systems through aid alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

4.7: Harmonization and alignment in the United Republic of Tanzania education programme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

4.8: Cash-on-delivery aid raises as many problems as it solves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

4.9: Non-identical donor responses to education systems in Burundi and Rwanda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

4.10: Multidonor trust funds — a promising approach with mixed results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

4.11: Canada’s ‘whole of government’ approach in Afghanistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

4.12: The Education Programme Development Fund . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

4.13: Kenya — FTI support for school fee abolition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

4.14: Mozambique — slow delivery under the Fast Track Initiative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258

4.15: Liberia — an approved plan with no Catalytic Fund support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260

4.16: Private sector initiatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

4.17: A new global fund for education? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266

Reaching the Education for All goals

There has been progress...

The number of children out of school has dropped by33 million worldwide since 1999. South and West Asiamore than halved the number of children out of school – a reduction of 21 million.

Some countries have achieved extraordinary advances.Benin started out in 1999 with one of the world’s lowestnet enrolment ratios but may now be on track foruniversal primary education by 2015.

The share of girls out of school has declined from 58%to 54%, and the gender gap in primary education isnarrowing in many countries.

Between 1985–1994 and 2000–2007, the adult literacyrate increased by 10%, to its current level of 84%.The number of adult female literates has increasedat a faster pace than that of males.

...but much remains to be done:

Malnutrition affects around 175 million young childreneach year and is a health and an education emergency.

There were 72 million children out of school in 2007.Business as usual would leave 56 million childrenout of school in 2015.

Around 54% of children out of school are girls. In sub-Saharan Africa, almost 12 million girls may neverenrol. In Yemen, nearly 80% of girls out of school areunlikely ever to enrol, compared with 36% of boys.

Literacy remains among the most neglected of alleducation goals, with about 759 million adults lackingliteracy skills today. Two-thirds are women.

Millions of children are leaving school without havingacquired basic skills. In some countries in sub-SaharanAfrica, young adults with five years of education hada 40% probability of being illiterate. In the DominicanRepublic, Ecuador and Guatemala, fewer than half ofgrade 3 students had more than very basic reading skills.

Some 1.9 million new teacher posts will be requiredto meet universal primary education by 2015.

Ten years have passed since the internationalcommunity adopted the six Education for Allgoals in Dakar in 2000. The record since thenhas been mixed. While much has beenachieved over the past decade, many of the

world’s poorest countries are not on track to meet the2015 targets. Failure to reach the marginalized hasdenied many people their right to education. With theeffects of the global economic crisis still being felt, thereis a real danger that much of the progress of the pastten years will stall or be reversed. Education is at risk,and countries must develop more inclusive approaches,linked to wider strategies for protecting vulnerablepopulations and overcoming inequality.

Minimizing the impact of the financial crisis on education

The international community needs to identify the threat to education posed by the economiccrisis and the rise in global food prices...

Human development indicators are deteriorating.An estimated 125 million additional people could bepushed into malnutrition in 2009 and 90 millioninto poverty in 2010.

With poverty rising, unemployment growing andremittances diminishing, many poor and vulnerablehouseholds are having to cut back on educationspending or withdraw their children from school.

National budgets in poor countries are under pressure.Sub-Saharan Africa faces a potential loss of aroundUS$4.6 billion annually in financing for education in2009 and 2010, equivalent to a 10% reduction inspending per primary-school pupil.

...and develop an effective response:

Provide up-front, sustained and predictable aid tocounteract revenue losses, protect priority socialspending and support progress in education.

Convene a donor pledging conference in 2010 to closethe Education for All financing gap.

Highlights of the EFA Report 2010

H I G H L I G H T S O F T H E E FA R E P O R T 2 0 1 0

1

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

H I G H L I G H T S O F T H E E FA R E P O R T 2 0 1 0

2

Reaching the marginalized

Governments are failing to address the root causes of marginalization in education. The newDeprivation and Marginalization in Education dataset highlights the level of exclusion in eightycountries...

In twenty-two countries, 30% or more of young adultshave fewer than four years of education, and this risesto 50% or more in eleven sub-Saharan Africancountries.

In twenty-six countries, 20% or more of young adultshave fewer than two years of schooling and, in somecountries, including Burkina Faso and Somalia,the share is 50% or more.

Inequalities often combine to exacerbate the risk ofbeing left behind. In Turkey, 43% of Kurdish-speakinggirls from the poorest households have fewer thantwo years of education, while the national averageis 6%; in Nigeria, 97% of poor Hausa-speaking girlshave fewer than two years of education.

Failure to address inequalities, stigmatization anddiscrimination linked to wealth, gender, ethnicity,language, location and disability is holding backprogress towards Education for All.

...and the need to create inclusive educationsystems:

Increase access and improve affordability for excludedgroups by lowering cost barriers, bringing schoolscloser to marginalized communities and developing‘second-chance’ programmes.

Improve the learning environment by deploying skilledteachers equitably, targeting financial and learningsupport to disadvantaged schools, and providingintercultural and bilingual education.

Expand entitlements and opportunities by enforcinglaws against discrimination, providing social protectionprogrammes and redistributing public finance.

Develop disaggregated data collection systemsto identify marginalized groups and monitor theirprogress.

Meeting the cost of Education for All

The record on aid for education is disappointing…

Overall aid has been increasing, but commitmentsare falling short of the US$50 billion increase pledgedin 2005. Africa faces the greatest projected shortfall,estimated at US$18 billion.

Aid to education has been rising, but commitmentshave recently stagnated. Aid commitments to basiceducation fell by 22% to US$4.3 billion in 2007.

Aid to education is not always reaching those whoneed it most. Some donors continue to give insufficientpriority to basic education. Countries affected byconflict are not receiving enough support, underminingprospects for recovery.

Education lacks a strong multilateral framework foraccelerated progress, suffering from a narrow donorbase and an absence of funding from private sources.

...donors and recipient governments must bothincrease resources available to education andimprove aid governance:

Low-income countries could themselves makeavailable an additional US$7 billion a year – or 0.7%of GDP. Even with this effort, large financing gapswill remain. The Report estimates the financing gapto meet the EFA goals in low-income countries atUS$16 billion annually.

Donors should strengthen efforts to implementthe Paris agenda on aid effectiveness and reviewthe balance of their support for the different levelsof education.

Donors must also scale up aid to countries affectedby conflict, finding innovative ways of providing longer-term, coordinated support.

The international multilateral framework forcooperation in education needs to be strengthenedthrough fundamental reform of the EFA Fast TrackInitiative.

The United Nations should convene an emergencypledging conference in 2010 to mobilize theadditional financing required and to fulfil the Dakarcommitment.

Chapter 1From financial crisis to humandevelopment crisis

The backdrop to the Education for All Global MonitoringReport 2010 is the most severe global economicdownturn since the Great Depression. Education systemsin many of the world’s poorest countries1 areexperiencing the aftershock of a crisis that originatedin the financial systems of the developed world. There isan imminent danger that, after a decade of encouragingadvances, progress towards the Education for All goalswill stall, or even be thrown into reverse, in the face ofrising poverty, slower economic growth and mountingpressure on government budgets. The internationalcommunity needs to act urgently to avert that danger.

Conditions for a concerted push towards the 2015 targetshave deteriorated across the developing world. By 2010,the recession could drive another 90 million people intoextreme poverty. Moreover, many of the worst-affectedcountries are still recovering from high food pricesthat left an additional 175 million malnourished in 2007and 2008. Education systems will not be immune tothe effects of these deteriorating human conditions.The concern is that the increased vulnerability of poorhouseholds and rising child malnutrition will impede

efforts to achieveuniversal primaryeducation andthe widerinternationaldevelopmenttargets setfor 2015.

Insufficientattentionhas been paid

to the consequences of slower economic growth forthe financing of education in the poorest developingcountries. While rich countries nurture the ‘green shoots’of recovery, developing countries face the prospect ofslower growth and diminished revenue collection. In sub-Saharan Africa alone, the potential loss of financingfor education as a result of the global recession willaverage around US$4.6 billion a year in 2009 and 2010 –

double the current level of aid to basic education.Spending per primary school pupil could be as much as10% lower in 2010 than it would have been on pre-crisiseconomic growth projections.

It is easy to lose sight of what is at stake. Ultimately,the world economy will recover from the recession, butthe crisis could create a lost generation of children in theworld’s poorest countries whose life chances will havebeen irreparably damaged by a failure to protect theirright to education. For those individuals and communitiesmost immediately affected, failure to sustain progresswould impose a high price in diminished opportunitiesto escape poverty and vulnerability. But whole countriesalso stand to lose out as weaker progress in educationleads to slower economic growth, reduced job creation,deteriorating public health and a more marginal placein the increasingly knowledge-based global economy.

National budgets have a vital role to play in preventingthe financial crisis from turning into a long-term humandevelopment crisis. Rich countries have put in placelarge-scale fiscal stimulus packages aimed at supportingeconomic recovery and protecting vital social andeconomic infrastructure. Education has been seenas a priority area for public spending, notably underthe American Recovery and Reinvestment Act. Unlikedeveloped countries, most low-income developingcountries lack the capacity to mobilize financing onthe scale required to maintain public spending in priorityareas. They desperately need an increase in concessionaldevelopment assistance to provide breathing spaceto cope with the crisis and maintain spending plansin education and other areas.

The international community has not respondedeffectively to the challenges facing the poorest countries.

1. Throughout the Report, the word ‘countries’ should generally be understoodas meaning ‘countries and territories’.

Overview

OV E R V I E W

3

© X

INH

UA/

Gam

ma/

Eyed

ea P

ress

e

© A

ubre

y W

ade/

PAN

OS

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

OV E R V I E W

4

Rich-country governments and successive summitsof the Group of 20 and Group of 8 have moved financialmountains to stabilize financial systems, but haveprovided an aid molehill for the world’s most vulnerablepeople. Donors have provided some US$2 billion toUS$3 billion annually in new and additional financefor low-income countries as a group, principally throughthe International Monetary Fund, but sub-SaharanAfrica alone faces an estimated revenue shortfall,against pre-crisis projections, of US$80 billion peryear in 2009 and 2010.

A ‘smoke and mirrors’ reporting system has led toexaggerated accounts of the international aid directedto low-income countries. Much of the reported supportprovided to the poorest countries is in fact repackagedor reprogrammed aid. The World Bank has increasedassistance principally through early disbursementof existing concessional loans. While such innovativeapproaches to funding are welcome, the danger isthat they will create future financing deficits – andthey are no substitute for real resource transfers.

In 2010, the international community will gather ata Millennium Development Goals summit to reviewprogress and assess prospects for achieving the targetsset. Those prospects hinge critically on early action toaddress the threats facing many of the world’s poorestcountries as a result of the global economic downturn.Education is a priority area. Any slowdown in the rate ofprogress towards the education goals will have adverselong-term consequences for economic growth, povertyreduction and advances in public health. Earlyinvestment is critical.

This Report estimates it will cost US$16 billion a yearto achieve universal primary education and widerEducation for All goals by 2015. This price tag appearsconsiderable, unless measured against the scale ofresources mobilized to rescue ailing financial institutions.It represents about 2% of the amount mobilized torescue just four major banks in the United Kingdomand the United States. Of course, governments pointout that securing the financial assets and balance sheetsof banks represents an investment. But the same is trueof international aid for education, which is an investmentin poverty reduction, shared prosperity and a moreequitable pattern of globalization.

The urgent international measures required include:

increased concessional financial support throughbilateral aid and the World Bank’s InternationalDevelopment Association (IDA), with a commitmentto increase IDA replenishment from US$42 billionto US$60 billion;

a review of the implications of the global economicdownturn for the financing of development targetsin advance of the 2010 Millennium DevelopmentGoals summit;

an emergency pledging conference during 2010to mobilize additional aid for education;

budget monitoring to pick up early warning signs offiscal adjustments that threaten education financing,with UNESCO coordinating an internationalprogramme to these ends;

revision of the IMF’s loan conditions to ensureconsistency with national poverty reduction andEducation for All priorities.

© A

ubre

y W

ade/

PAN

OS

© J

eroe

n Oe

rlem

ans/

PAN

OS

OV E R V I E W

5

Chapter 2Monitoring progress towardsthe Education for All goals

The goals adopted in 2000 at the World EducationForum in Dakar remain the benchmark for assessingprogress towards Education for All. Much has beenachieved: some of the world’s poorest countries haveregistered advances on many fronts, demonstrating thatnational leadership and good policies make a difference.But the world is unequivocally off track for the Dakargoals and the battle to achieve universal primaryeducation by 2015 is being lost.

Changing this picture will require a far stronger focuson inequality and the most marginalized groups insociety. Gender remains a priority area because of thepersistence of institutionalized disadvantage for younggirls and women. Strategies aimed at equalizingopportunity in education will also have to addressdisadvantages rooted in poverty and social discrimination.The monitoring evidence points clearly to the need fora greater sense of urgency on the part of governmentsand donors. With less than five years to the target date,the window of opportunity for putting in place theinvestment and policies needed to bring the educationgoals within reach is closing.

Early childhood

Early childhood care and education is the bedrock ofEducation for All. Good nutrition, effective health careand access to good pre-school facilities can mitigatesocial disadvantage and lead to improved learningachievement. Yet early childhood provision continuesto be marked by neglect.

That neglect starts early. Around a third of all childrenin developing countries, or 175 million annually, enterprimary school having experienced malnutrition thatirreparably damages their cognitive development. Unsafepregnancy and childbirth take a devastating toll. Birthasphyxia leaves around 1 million children a year withlearning difficulties and other disabilities. Maternal iodinedeficiency poses a risk of mental impairment for around38 million children a year. These are problems rooted inpoverty, gender inequality, and the failure of child andmaternal health services to provide affordable accessto decent care. Abolishing user fees for these servicesis an immediate priority. More broadly, it is importantthat policy-makers develop more integrated approachesto education on the one side, and to child and maternalhealth care provision on the other.

Participation in early childhood care and educationprogrammes remains uneven. Coverage levels areespecially low in South and West Asia, and sub-SaharanAfrica. Children from the poorest households potentiallyhave the most to gain from good early childhood careand education. Yet they are the least likely to have access.In Egypt, children from the wealthiest households aretwenty-eight times more likely to be in pre-school thanchildren from the poorest households. Such outcomespoint to the importance of barriers linked to cost andlocation. Yet several successful programmes, suchas Chile Crece Contigo, demonstrate that targetedinvestment can break down social disparities.

Universal primary education

Overall progress towards universal primary educationin the past decade has been encouraging. In 2007, some72 million children were out of school – a 28% declinefrom the start of the decade. Since 1999, enrolment ratesin sub-Saharan Africa have been increasing five timesas fast as during the 1990s, with countries includingBenin, Ethiopia, Mozambique and the United Republic ofTanzania registering rapid advances. In addition, genderdisparities in primary school have been narrowing.

© A

lfred

o Ca

liz/P

ANOS

© P

atri

ck L

e Fl

och/

Expl

orer

/Eye

dea

Illus

trat

ion

Nevertheless, the world is not on track to meet theuniversal primary education goal. Current trends willleave some 56 million children out of school in 2015 – and there are worrying indications that the rate ofprogress towards universal primary education isslowing. Two-thirds of the total decline in out-of-schoolnumbers since the Dakar conference took place from2002 to 2004. Regional progress has also been uneven.Out-of-school numbers have fallen far more rapidlyin South Asia, driven by rapid advances in India, thanin sub-Saharan Africa. Most of the countries that areoff track for achieving universal primary education by2015 are low-income countries that, having startedfrom a low base, are either increasing enrolmentsimpressively but too slowly (as in Burkina Faso andthe Niger) or stagnating (as in Eritrea and Liberia).Countries affected by conflict figure prominently inthis group. More surprisingly, higher-income countriessuch as the Philippines and Turkey are in danger offailing to achieve the target, largely because of deeplyentrenched national inequalities.

Deep-rooted inequalities are a major barrier to universalprimary education. Disparities linked to wealth, gender, ethnicity, language and location are holding back progress in many countries. While gender gaps are narrowing, theyremain very large in much of South and West Asia andsub-Saharan Africa. In twenty-eight countries, there arestill fewer than nine girls in school for every ten boys.Closing the gender divide will require a sustained effortto change attitudes that diminish the value of girls’education, along with practical policies that createincentives for greater equity. Poverty exacerbates thegender divide. In Pakistan there is no discernable gendergap for the wealthiest urban households, but only one-third of girls from the poorest households are in school.

Enrolment is just one measure of overall progresstowards universal primary education. While enrolmentrates are rising, millions of children enter primaryschool only to drop out before completing a full primarycycle. Some 28 million pupils in sub-Saharan Africadrop out each year. In South and West Asia, 13% ofchildren entering school drop out in the first grade.

01 OV E R V I E W

Moreover, current approaches to monitoring andassessment may be putting a positive gloss on underlyingproblems, for three reasons:

Data reported by governments may systematicallyunderstate real out-of-school numbers for primaryschool age children. Household survey data indicatethat total out-of-school numbers may be as much asone-third higher than those reported by governments.

Reporting conventions render invisible the 71 millionchildren of lower secondary school age who are outof school.

Current monitoring tools do not provide an integratedway of measuring the three things that count inprogress towards universal primary education:entering school at an appropriate age, progressingsmoothly through the system and completing school.Chapter 2 sets out the case for a more comprehensiveapproach based on the net cohort completion rate.

Adult skills and learning

The global economic crisis has pushed youth and adultskills and learning – goal 3 of the Dakar Framework forAction – to the centre of the Education for All agenda.With youth unemployment rising, governmentsincreasingly see skills development as a vital componentof overall strategies to combat marginalization. Morebroadly, there is recognition that, in an increasinglyknowledge-based global economy, the premium on skillsas a driver of employment, productivity and economicgrowth is rising.

Countries vary enormously in the coverage andeffectiveness of technical and vocational education.In Germany and Japan, vocational education has playeda vital role in smoothing the transition from school towork and in combating youth unemployment. Vocationaleducation in East Asia was an integral part of industrialdevelopment strategies that fostered rapid growth,employment creation, and higher levels of skills andwages. It is increasingly recognized that one shot ateducation is not enough. For youth and young adultswho emerge from school lacking basic learning skills,vocational training can help provide a second chance.Experience from Latin America and the United Statesdemonstrates that technical and vocational training canextend opportunities to marginalized young people whodropped out of school, including the chance to re-entereducation systems.

While there are many examples of good practice,the overall record of technical and vocational education

6

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

20

© L

ouis

e Gu

bb/C

orbi

s

OV E R V I E W

7

is open to question. Many national programmes sufferfrom a combination of underinvestment, poor qualityand weak links to employment markets. Governmentsin the Middle East have invested heavily in vocationaleducation with little to show for it in the way of jobs.In sub-Saharan Africa, vocational education largelybypasses the informal sector (where most marginalizedyoung people work), and is shunned by parents andpupils. Vocational programmes in India reach only about3% of rural youth and there is little evidence that theyenhance employment prospects. The image of technicaland vocational provision as a form of second-classeducation that provides limited benefits for employmentremains largely intact.

Changing that image will require far-reaching reforms.Successful vocational education systems typically providea strong link between the world of school and the worldof work, requiring active engagement by the privatesector. One of the features of Brazil’s model, for example,is that the country’s employers’ federation is a majorprovider, with high-quality training geared towardsareas characterized by labour market shortages. Thecurriculum and approaches to teaching also matter.Too often, vocational education focuses on narrowtechnical abilities rather than broader, more flexible‘learning to learn’ skills. Several countries – includingAustralia and the Republic of Korea – are addressing thisproblem and the associated poor reputation of traditionalprogrammes, by allowing for greater fluidity betweenvocational training and academic education.

Adult literacy

Literacy is a vital asset and key component of skillsdevelopment. Yet adult literacy remains one of themost neglected of the Education for All goals. There arecurrently some 759 million illiterate youths and adultsin the world. Reflecting the legacy of gender disparitiesin education, two-thirds of this number are women.While gender gaps are narrowing, they remain very large.Except in East Asia – principally China – progresstowards the target of halving illiteracy has been painfullyslow. On current trends, the world will be less thanhalfway towards this goal by 2015. India alone will havea shortfall of some 81 million literate people.

There have been some encouraging developments inrecent years. Several countries with large numbers ofilliterate adults are increasing investment in nationalliteracy programmes. The Literate Brazil Programme,which started in 2003, is an example: it has reached8 million learners. India is reconfiguring and expandingits national literacy programme to focus more strongly onwomen, low-caste groups and minorities. Burkina Faso’s

national education strategyhas scaled up investmentin literacy from 1% to 7%of the education budget.Governments and donorsneed to learn from emergingmodels of good practice andact with greater resolve inprioritizing literacy withinwider education strategies.

Education quality

The ultimate measure of anyeducation system is not howmany children are in school,but what – and how well –they learn. There is growingevidence that the world is moving more quickly to getchildren into school than to improve the quality of theeducation offered.

Learning achievement deficits are evident at manylevels. International assessment exercises pointconsistently towards severe global disparities. The2007 Trends in International Mathematics and ScienceStudy (TIMSS) found that average students in severaldeveloping countries, including Ghana, Indonesia andMorocco, performed below the poorest-performingstudents in countries such as Japan and the Republicof Korea. Inequalities within countries, linked tohousehold disadvantage and the learning environment,are also marked. The problem is not just one ofrelative achievement. Absolute levels of learning aredesperately low in many countries. Evidence fromSouth and West Asia and from sub-Saharan Africasuggests that many children are failing to master basicliteracy and numeracy skills, even when they completea full cycle of primary education.

Low learning achievement stems from many factors.Schools in many developing countries are in a poorstate and teachers are in short supply. By 2015, thepoorest countries need to recruit some 1.9 millionadditional primary school teachers, including 1.2 millionin sub-Saharan Africa, to create a good learningenvironment for all children. More equitable teacherdeployment is also vital: all too often, the poorestregions and most disadvantaged schools have thefewest and least-qualified teachers. Several countries,including Brazil and Mexico, have introducedprogrammes targeting schools serving disadvantagedcommunities. Governments can also raise standardsby spotting problems early, using constant monitoringand early-grade reading assessments.

© R

EUTE

RS/

Zohr

a B

ense

mra

The Education for All financing gap

Achieving the Education for All goals in low-incomecountries will require a major increase in financing.These countries themselves can do a great deal tomobilize more resources for education. But in theabsence of a step increase in aid, efforts to accelerateprogress in basic education will be held back by a largefinancing gap.

This Report provides a detailed assessment of the costsassociated with achieving some of the core Educationfor All goals. Covering forty-six low-income countries,the assessment includes estimates for improvedcoverage in early childhood programmes, universalprimary education and adult literacy. Unlike previousglobal costing exercises, it includes a provision forreaching the most marginalized. That provision isimportant because it costs more to extend opportunitiesto children disadvantaged by poverty, gender, ethnicity,language and remoteness. Among the central findingsand recommendations of the Report:

Low-income developing countries could makeavailable an additional US$7 billion a year – or 0.7%of GDP – by raising more domestic resources andmaking national budgeting more equitable.

Even with an increased domestic resource mobilizationeffort, there will be a global Education for All financinggap of around US$16 billion annually –1.5% of the GDP – for the forty-six low-income countries covered. Sub-Saharan Africa accounts for around two-thirdsof the global financing gap, or US$11 billion.

Current aid to basic education for the forty-sixcountries – around US$2.7 billion – falls short ofwhat is required to close the gap. Even if donors acton their commitments to increase aid, the financinggap will remain significant at around US$11 billion.

An emergency pledging conference should beconvened in 2010 to mobilize the additional financingrequired to fulfil the commitment made at Dakar.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

OV E R V I E W

8

Chapter 3Marginalization in education

Governments across the world constantly reaffirm theircommitment to equal opportunity in education. Underinternational human rights conventions they are obligatedto act on that commitment. Yet most governments aresystematically failing to address extreme and persistenteducation disadvantages that leave large sections ofsociety marginalized. These disadvantages are rooted indeeply ingrained social, economic and political processes,and unequal power relationships – and they are sustainedby political indifference.

Marginalization in education matters at many levels.Having the opportunity for a meaningful education isa basic human right. It is also a condition for advancingsocial justice. People who are left behind in educationface the prospect of diminished life chances in manyother areas, including employment, health andparticipation in the political processes that affect theirlives. Moreover, restricted opportunity in education isone of the most powerful mechanisms for transmittingpoverty across generations.

Extreme deprivation in education is a particularly strikingcase of what the economist and philosopher Amartya Senhas described as ‘remediable injustices’. The Reportlooks at the scale of the injustice, examines its underlyingcauses and identifies policy remedies. The key messageto emerge is that failure to place inclusive educationat the centre of the Education for All agenda is holdingback progress towards the goals adopted at Dakar.Governments have to do far more to extend opportunitiesto hard-to-reach groups such as ethnic minorities, poorhouseholds in slums and remote rural areas, thoseaffected by armed conflict and children with disabilities.

Measuring marginalization: a new data tool

Measuring marginalization in education is inherentlydifficult. There are no established cross-countrybenchmarks comparable to those used in assessingextreme income poverty. National data are oftennot detailed enough to enable marginalized groupsto be identified. An underlying problem is that manygovernments attach little weight to improving dataavailability relating to some of the most disadvantagedsections of society – child labourers, people living ininformal settlements and individuals with disabilities –and to remote regions. This year’s Report includesa new tool, the Deprivation and Marginalization inEducation (DME) data set, which provides a window

© C

hret

ien

Eric

/Gam

ma/

Eyed

ea P

ress

e

OV E R V I E W

9

on the scale of marginalization within countries andon the social composition of marginalized groups.

Despite the progress of the past decade, absolutedeprivation in education remains at extraordinarily highlevels. On any global scale, having fewer than four yearsof education, the minimum required for basic literacy, isan indicator of extreme disadvantage. The DME data setestablishes this as a benchmark for ‘education poverty’,with less than two years in school as an indicator for‘extreme education poverty’. Findings from sixty-threedeveloping countries include the following:

Education poverty. In twenty-two countries, 30% ormore of those aged 17 to 22 have fewer than four yearsof education, and the share rises to 50% or more ineleven countries of sub-Saharan Africa.

Extreme education poverty. In twenty-six countries,20% or more of those aged 17 to 22 have fewerthan two years of schooling and, in some countries,including Burkina Faso and Somalia, the share is 50% or more.

These averages mask extreme inequalities linked towealth and gender. In the Philippines, education povertyrates among the poor are four times the nationalaverage. In some countries, high levels of marginalizationamong poor females account for a significant share ofeducation poverty. Just under half of poor rural femalesaged 17 to 22 in Egypt have fewer than four years ofeducation and in Morocco the rate is 88%. Socialinequalities also explain some striking cross-countrydifferences. With a per capita income comparableto Viet Nam’s, Pakistan has over three times the levelof education poverty – a reflection of disparities linkedto wealth, gender and region.

The factors leading to marginalization in education donot operate in isolation. Wealth and gender intersect withlanguage, ethnicity, region and rural-urban differences tocreate mutually reinforcing disadvantages. Detailed DMEdata for those aged 17 to 22 help identify groups facingparticularly extreme restrictions on education opportunityand highlight the scale of national inequalities.

Cross-country analysis reveals complex patterns of marginalization. Some social groups face almost universal disadvantage. Pastoralists in sub-Saharan Africa are anexample. In Uganda, which has made strong progresstowards universal primary education, Karamajongpastoralists average less than one year in education.Many countries also show large disparities linked tolanguage. In Guatemala, average years in school range from 6.7 for Spanish speakers to 1.8 for speakers of Q’eqchi’.

The DME data set makes it possible to look beyondabsolute deprivation to identify some of the keycharacteristics of those who are being left behind.Using surveys, it identifies people found in the bottom20% of the national distribution in terms of years inschool. The results highlight the powerful influenceof social circumstances, over which children haveno control, in determining life chances. They also drawattention to unacceptable levels of inequality:

The wealth divide means that being born into apoor household doubles the risk of being in the bottom20% in countries ranging from India to the Philippinesand Viet Nam.

Regional divides mean that living in areas such asrural Upper Egypt, northern Cameroon and easternTurkey increases significantly the risk of falling intothe bottom 20%.

Gender, poverty, language and culture often combineto produce an extremely heightened risk of being leftfar behind. In Turkey, 43% of Kurdish-speaking girlsfrom the poorest households have fewer than twoyears of education, while the national average is 6%;in Nigeria, some 97% of poor Hausa-speaking girlshave fewer than two years of education.

Time spent in school is just one dimension ofmarginalization. There are also marked gaps in learningachievement linked to socio-economic status. Children ofparents in the wealthiest fourth of the population in Braziland Mexico score 25% to 30% higher in mathematicstest scores, on average, than children of parents in thepoorest fourth. Having a home language that is differentfrom the official language of instruction is also associatedwith lower test scores: in Turkey, Turkish speakers are30% less likely than non-Turkish speakers to scorebelow a minimum benchmark in mathematics.

© G

iaco

mo

Piro

zzi/P

ANOS

01 OV E R V I E W

1 0

Marginalization in education affects all countries.While absolute average achievement levels are higherin the developed world, extreme relative deprivation isa widespread concern. In the European Union, 15% ofyoung people aged 18 to 24 leave school with only lowersecondary school education, and the figure rises to 30%in Spain. Household wealth has a significant bearingon education achievement. In England (United Kingdom),pupils receiving free school meals – an importantindicator for social deprivation – score 29% lower, onaverage, than the national average for mathematics.

Evidence from the United States highlights the powerfulinfluence of wealth and race. African-Americans aretwice as likely to be out of school as white Americans,and young adults from poor households are three timesas likely to be out of school as those from wealthyhomes. International learning assessments illustratethe extent of national disparities. On the TIMSS scalefor mathematics, the United States ranks ninth outof forty-eight countries, but schools with highconcentrations of poverty score thirteen places lower.The bottom 10% of performers in the United Statesscore 25 places below the national average and belowthe average for Thailand and Tunisia.

Measuring marginalization is not an end in itself. Itshould be seen as a means of developing policies anddesigning targeted interventions that can translatecommitment to Education for All into meaningful action.The starting point is for governments to set targets forreducing inequalities and narrowing the gap betweenmarginalized groups and the rest of society. Monitoringprogress towards these targets using disaggregateddata could help both to provide an evidence base forthe development of targeted policies and to increasethe visibility of the marginalized.

National equity targets in education should be seenas an integral element of Education for All goals. Theycould include time-bound commitments to work towardshalving gaps in school attendance between, say, thewealthiest and poorest households, the best performingand worst performing regions, boys and girls, and ethnicor linguistic minorities and the rest of the population.Data of the type provided in the DME data set providea tool for monitoring, auditing and evaluating progresstowards equity targets.2

Marginalization in education is driven by social inequalities

Marginalization in education is the product of a toxiccocktail of inherited disadvantage, deeply ingrainedsocial processes, unfair economic arrangementsand bad policies.

Being born into poverty is one of the strongest factorsleading to marginalization in education. Some 1.4 billionof the world’s people survive on less than US$1.25 a day.Many are parents struggling to keep their children inschool. Household surveys point to parental inability to afford education as a major factor behind non-attendance.

Household poverty goes hand in hand with vulnerability.Even a small economic shock caused by drought,unemployment or sickness, for example, can forceparents into coping strategies that damage children’swelfare. Girls are often the first to feel the effects. InPakistan and Uganda, climate-related shocks resultin far more girls being taken out of school than boys.Child labour is another corollary of poverty that hurtseducation. There are an estimated 166 million childlabourers in the world. Many of these children arelocked in a losing battle to combine work with education.In Mali, around half of all children aged 7 to 14 reportthat they are working. With labour activities taking upan average of thirty-seven hours a week, most of thesechildren do not attend school.

Language and ethnicity lead to marginalization ineducation through complex channels. Poverty is animportant part of the equation. In Ecuador andGuatemala, malnutrition rates among indigenouschildren are twice the level for non-indigenous children.Other factors powerfully reinforce the effects of socialdeprivation. One reason that many linguistic and ethnicminority children perform poorly in school is that they areoften taught in a language they struggle to understand.Around 221 million children speak a different language at

2. The DME data set is one of many tools available to support such approaches.It is available online at: http://www.unesco.org/en/efareport/dme.

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

20

© A

bbie

Tra

yler

-Sm

ith/P

ANOS

OV E R V I E W

1 1

home from the language of instruction in school, limitingtheir ability to develop foundations for later learning.

At the same time, language policy in education raisescomplex issues and potential tensions between groupidentity on the one hand, and social and economicaspirations on the other. Parents in many countriesexpress a strong preference for their children to learnin the official language, principally because this is seenas a route to enhanced prospects for social mobility.

Stigmatization is a potent source of marginalizationthat children bring with them to the classroom. FromAboriginals in Australia to the indigenous people of LatinAmerica, failure to provide home language instructionhas often been part of a wider process of culturalsubordination and social discrimination. Caste systemsin South Asia also disadvantage many children.Research from India is instructive. It shows that childrenfrom low-caste households score at far lower levelswhen their caste is publicly announced than whenit is unannounced – an outcome that underlinesthe debilitating effects of stigma on self-confidence.

Livelihoods and location are often strongly linked withsocial disadvantage in education. One reason pastoralistsin South Asia and sub-Saharan Africa register such highlevels of deprivation in education is that their livelihoodsinvolve children travelling long distances. Immobileschool infrastructure is ill equipped to respond to theneeds of highly mobile groups and the schooling providedis often irrelevant to their lives. Slums are also focalpoints for education deprivation. This is partly becauseof poverty and partly because many governments failto provide slum dwellers with the legal rights requiredto establish an entitlement to education.

Conflict is a potent source of marginalization ineducation. Worldwide, around 14 million children aged 5to 17 have been forcibly displaced by conflict, often withincountries or across borders, into education systemslacking the most rudimentary education facilities. Lesseasy to measure than the impact on school attendanceare the effects on learning of trauma associated witharmed conflict. In 2008 and 2009, Israeli military actionsin Gaza led to the deaths of 164 students and 12 teachers,and severely damaged or destroyed 280 schools andkindergartens. In an area where 69% of adolescents werealready reported as experiencing post-traumatic stressbefore the latest episode of violence, many childrenreturned to school carrying with them the effects ofanxiety and emotional shock. An investigation into themilitary actions submitted to the United Nations GeneralAssembly concluded that both Israeli and Palestinianauthorities had targeted civilian populations.

Some sections of society face problems rooted in publicperceptions and official neglect. Children living withdisabilities suffer from social attitudes that stigmatize,restrict opportunity and lower self-esteem. Theseattitudes are frequently reinforced within the classroom,where teachers often lack the training and resourcesneeded to deliver a decent education. Children living withHIV and AIDS, and those who have been orphaned by thedisease or are living with affected household members,also face distinctive pressures. Some of these pressuresoriginate in economic hardship and the need to providecare. Others can be traced to practices rooted in socialdiscrimination and to the effects of loss experienced byAIDS orphans. Evidence from many countries suggeststhat education planners are not responding effectivelyto these problems.

Reaching and teaching the marginalized

There is no single formula or blueprint for overcomingmarginalization in education. Policies need to addressunderlying causes such as social inequality, genderdisparities, ethnic and linguistic disadvantages, and gapsbetween geographic areas. In each of these areas,equalizing opportunity involves redressing unequal powerrelationships. The inequalities that the marginalized facestart in early childhood and continue through school ageyears. They are deeply engrained and highly resistant tochange. Yet progress is possible with sustained politicalcommitment to social justice, equal opportunity andbasic rights.

© A

mi V

itale

/PAN

OS

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

OV E R V I E W

1 2

This Report identifies three broad sets of policies that canmake a difference. They can be thought of as three pointsin an inclusive-education triangle:

Accessibility and affordability. Removing school feesis necessary to make education more affordable forthe poorest, but is not sufficient to remove cost barriers.Governments also need to lower indirect costs associatedwith uniforms, textbooks and informal fees. Financialstipend programmes for identifiably marginalized groups– such as those developed in Bangladesh, Cambodia andViet Nam – can help provide incentives for education andenhance affordability. Bringing schools closer tomarginalized communities is also vital, especially forgender equity – a point demonstrated by the sharpdecline in out-of-school numbers in Ethiopia. Moreflexible approaches to providing education andmultigrade teaching in remote areas could bringeducation within reach of some of the world’s mostmarginalized children. Non-government organizationsoften play an important role in extending access to hard-to-reach populations, including child labourers, out-of-school adolescents and children with disabilities.In Bangladesh, one non-government organization hasdeveloped a system of ‘floating schools’ in order to reachthe Bede (River Gypsy) community, whose livelihooddepends on their moving about on boats. The provisionof non-government organizations is most successfulwhen it is integrated into national systems, allowingchildren to continue their studies in formal schoolingor to gain meaningful employment.

The learning environment. Getting marginalized childreninto school is just a first step. Ensuring that they receivea good education poses significant policy challenges.Targeted financial support and programmes to facilitateimproved learning in schools in the most disadvantagedregions can make a difference, as can programmes thatdraw well-qualified teachers to the schools facing thegreatest deprivation. Language policy is another key area.Reforms in Bolivia have emphasized the important roleof intercultural and bilingual education in providingethnic and linguistic minority children with good-qualityschooling, and in overcoming social stigmatization.Ensuring that children with disabilities enjoy opportunitiesfor learning in an inclusive environment requires changesin attitude, backed by investment in teacher training andlearning equipment. The Convention on the Rights ofPersons with Disabilities provides a framework fordelivery that should serve as a guide to public policy.

Entitlements and opportunities. Many of the measuresneeded to overcome marginalization in education operateat the interface between education policy and widerstrategies for change.

Legal provisions can play a role in overcomingdiscrimination and realizing the right to education.Some marginalized groups, such as the Roma inEurope, have successfully challenged the legalityof policies that result in institutionalized segregation.Legal provisions are likely to prove most effectivewhen backed by social and political mobilization on the part of marginalized people – New Zealand’s Ma-ori language movement and Bolivia’s indigenousmovements are cases in point.

Social protection is a critical pathway to mitigatingthe vulnerability that comes with poverty. Conditionalcash transfer programmes in Latin America, forexample, have a strong track record in improvingschool attendance and progression. Several countriesin sub-Saharan Africa are also investing in socialprotection programmes. One large-scale exampleis the Productive Safety Net Programme in Ethiopia,which provides guaranteed employment forcommunities affected by drought, with positiveeducational effects. Increased investment in suchprogrammes can enhance equity and accelerateprogress towards the Education for All goals.However, equity and cost-effectiveness considerationsrequire detailed attention to the design ofinterventions, targeting and levels of support.

Redistributive public spending is one of the keys toexpanded entitlements and opportunities. Becausemarginalization in education is associated with poverty,

© H

andi

cap

Inte

rnat

iona

l

OV E R V I E W

1 3

Chapter 4International aid

The Dakar Framework for Action includes a pledgeby donors that ‘no countries seriously committed toeducation for all will be thwarted in their achievementof this goal by a lack of resources’. That pledge hasnot been honoured. The collective failure of donorsto mobilize aid on the scale required is holding backprogress in the world’s poorest countries. With theglobal financial crisis adding to pressure on nationalbudgets, it is vital for donors to deliver on the Dakarpromise. While primary responsibility for educationfinancing rests with developing country governments,the poorest countries lack the resources to achievethe 2015 goals without a major increase in aid – andan improvement in aid effectiveness.

Aid to education is inevitably influenced by overalldevelopment assistance levels. In 2005, donors pledgedto increase aid by US$50 billion by 2010, with half theincrease going to Africa. After two years of decline, aidflows rose sharply in 2008. However, planned increasesfall well short of the levels promised in 2005. Currentlyprogrammed aid for Africa points to a potential shortfallagainst pledges of US$18 billion in aid spendingrequired by 2010.

Donors have a mixed record in delivering on the promisesmade in 2005. Some countries, including Ireland, theNetherlands, Spain and Sweden, have exceeded their ‘fair share’ of the commitment. Others have been movingin the wrong direction. Italy has cut its aid-to-GNI ratio(from an already low level) and Japan and the UnitedStates fall well short of their ‘fair share’. The free riding of bad performers on the commitment of goodperformers has become a pervasive problem that couldworsen as governments respond to fiscal pressures.Some donors – such as the United Kingdom – haveundertaken a commitment to maintain, in real terms,planned increases in aid spending. With many low-income countries facing crisis-related budget pressures,this is an approach other donors should consider.

Levels of aid to education remain a source of concern.Overall disbursements of development assistance forbasic education have been on a rising trend, reachingUS$4.1 billion in 2007. However, commitments stagnatedfrom 2004 and fell by around one-fifth in 2007. Oneunderlying problem is the narrow base of donor supportfor education: overall flows are dominated by a smallgroup of countries. Another problem is the skewing of aid towards post-primary levels. Three major donors –France, Germany and Japan – commit over half their

the regions most affected often have the least capacityto mobilize resources. Most countries have someredistributive element in public finance, but typicallyit is underdeveloped. The programme of federalgovernment transfers in Brazil is an example of anattempt to narrow large state-level financing gapsin education, with some positive effects.

Overcoming marginalization in education is an imperativefor human rights and social justice. It is also the key toaccelerated progress towards the Education for All goalsset at Dakar. No government seriously committed to thegoals can afford to ignore the deep social disparities thatare stalling progress in education. Nor can it ignore thewider consequences of marginalization in education forsocial cohesion and future prosperity. That is why thisReport stresses the urgency of all countries developingstrategies for more inclusive education linked to widerstrategies for overcoming poverty, social discriminationand extreme inequality.

© C

risp

in H

ughe

s/PA

NOS

© F

ranc

çois

Per

ri

01 OV E R V I E W

1 4

education aid to post-primary levels, with a large shareof what is allocated to higher education being spent indomestic institutions. While aid to post-primary educationis justified, several donors need to review both theirpriorities and their aid modalities.

With pressure on aid budgets mounting, it is crucial fordonors and recipients to strengthen aid effectiveness.There is evidence of progress – but there is also a greatdeal of room for improvement. Aid flows are oftenunpredictable: in 2007, less than half of scheduled aidarrived on time. Use of national public financialmanagement systems is growing, but there is worryingevidence that many donors continue to operate outsidethese systems, thereby adding to transaction costs.

The education sector has hadlimited success in tapping intonew sources of innovativefinancing. Several majorinternational companies andphilanthropic institutions supportinitiatives in education, but theoverall impact has been dilutedby the absence of crediblemultilateral delivery mechanismsof the type developed in globalhealth initiatives. Advocates foreducation must seize opportunitiesto generate new sources offinance. This Report provides anexample: it calls for a small (0.4%)

‘Better Future’ levy on the commercial marketingrevenue of the major European football leagues, withthe 2010 World Cup providing a launch pad. The initiativecould mobilize US$48 million annually and finance qualityeducation for around half a million children a year.

The international donor community has not respondedeffectively to the problems of low-income countriesaffected by conflict. These countries account for one-thirdof out-of-school children, but less than one-fifth of aid toeducation. Moreover, aid flows are dominated by a smallgroup of conflict-affected states – notably Afghanistanand Pakistan – while a far larger group is neglected.

While conflict and post-conflict environments confrontdonors with immense challenges, current approachesare leading to lost opportunities for rebuilding educationsystems. Education receives less than 2% ofhumanitarian aid, including in countries such as theDemocratic Republic of the Congo. And countries suchas Burundi and Liberia have received insufficient supportfor education reconstruction. Overly rigid application ofrules on aid governance and reporting has hampered thedevelopment of more effective and flexible responses.

The Fast Track Initiative

The Fast Track Initiative (FTI) was hailed at its inceptionas a ‘historic step’ towards delivering Education for All,establishing a multilateral framework for strengtheningnational education plans and galvanizing the financingrequired to achieve universal primary education, amongother goals. While the initiative has registered someimportant achievements, overall performance has fallenshort of expectations – and comprehensive reform isan urgent priority. The international community urgentlyneeds a multilateral architecture fit for the purposeof accelerating progress towards the 2015 targets.

© R

EUTE

RS/

Ahm

ad M

asoo

d

© U

NES

CO/S

amer

Al-

Sam

arra

i

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

20

OV E R V I E W

1 5

The reform process starts by setting an appropriatescale of ambition, identifying areas in which a multilateralframework can add value to current efforts and settingout an agenda for governance reform to give developingcountries a stronger voice.

Insufficient clarity over the FTI’s remit should not beallowed to obscure its weak performance. There is nocredible evidence to support the argument that theinitiative has spurred an increase in bilateral aiddirected through country programmes. The FTI’s mainfinancing mechanism, the Catalytic Fund, has madelimited financial transfers with high transaction costs.While cumulative donor commitments had reachedUS$1.2 billion by March 2009, disbursementsamounted to just US$491 million. Several countrieswhose FTI plans were endorsed between 2002 and 2004have yet to receive their full allocation. Disbursementproblems have been compounded by the stringentapplication of World Bank rules, in some cases forcinggovernments and bilateral donors to adopt practices thatweaken donor coordination and undermine nationalownership.

Limited disbursement is not the only FTI weakness.The estimation of financing gaps has been characterizedby inconsistency and systematic underestimation, withFTI plans reflecting what donors may be willing to financerather than what developing countries need to meet the2015 targets.

Governance is another concern. While the FTI is widelypresented as a partnership, it is for practical purposes a‘donor club’. Developing countries are under-representedat all levels and have a weak voice in financing decisions.The FTI also effectively excludes from funding thosecountries most in need of a multilateral financingmechanism, since most conflict-affected countries havebeen viewed as not meeting the standards for accessingCatalytic Fund support.

The FTI experience contrasts strongly with multilateralinitiatives in health. To take the most notable example,the Global Fund to Fight AIDS, Tuberculosis and Malariahas succeeded in mobilizing and delivering additionalresources through a broad donor base. One of thestrengths of the Global Fund, in contrast with the FTI,has been the creation of innovative financing windows forphilanthropic donations. Governance arrangements differmarkedly from those of the FTI. The Global Fund is anindependent organization, staffed by a strong secretariat,and developing countries have a strong voice at all levels.It has delivered significant results in terms of impact,including in countries with weak capacity: it haddisbursed US$7 billion by 2008 and supplied antiretroviral

drugs to 2 million people. Notwithstanding some obviousdifferences and the problems associated with verticalinitiatives geared towards specific diseases, there areimportant lessons to be drawn for FTI reform.

There are several key ingredients for more effectivemultilateralism in education. Some of those ingredientscan be found in the principles underpinning the FTI,such as the commitment to back national planning andstrategies for achieving the Education for All goals withincreased aid. However, it is also important to establisha level of ambition commensurate with the challengeahead. The remit for the FTI should be clearly focused onclosing the Education for All financing gap, with a strongcommitment to the development of quality education andequity. Provision should be made for attracting supportfrom philanthropic foundations. And developing countriesshould have a far greater voice in governance. Butperhaps the most important ingredient for a moredynamic multilateral architecture, and the ingredientmost conspicuous by its absence to date, is high-levelpolitical leadership.

© U

NES

CO/E

rnes

to B

enav

ides

01

02

Education f

or

All

Glo

bal M

onitori

ng R

eport

Chapter 1

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

1 6

The financial crisis: global impact

© X

INH

UA/

Gam

ma/

Eyed

ea P

ress

e

Overcoming disadvantage:a boy does homework in Freetown,Sierra Leone

0102Education for All Global Monitoring Report

Education at risk: the impactof the financial crisis

1 7

© A

ubre

y W

ade/

PAN

OS

Children queue for food in Pakistan:rising prices hit the poor hardest

© J

eroe

n Oe

rlem

ans/

PAN

OS

The global financial crisis hasprovided a stark reminder of therealities of global interdependence.With the aftershock now reachingmany of the world’s poorestcountries, poverty levels are rising,malnutrition is worsening andeducation budgets are comingunder pressure. Some of theworld’s most vulnerablehouseholds are feeling the effectsof a crisis that originated in thebanking systems of the rich world.It is too early to assess exactlywhat the financial crisis will meanfor progress towards the EFA goals.But this year’s Report starts bylooking at the early warning signs.It then assesses the internationalresponse to the crisis andconsiders what can be done to avoid major setbacks.

Introduction .................................................................. 19

Double jeopardy: food prices and financial crisis ............................................. 21

Expanding ‘fiscal space’: an Education for All priority ................ 28

The international response: missing a human dimension .............. 32

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

1 8

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

I n t r o d u c t i o n

1 9

Introduction

The backdrop for this edition of the Educationfor All Global Monitoring Report is the deepesteconomic downturn since the Great Depression.While several financial indicators have improved inrecent months, fuelling optimism that the ‘greenshoots’ of recovery are taking root in the developedworld, many developing countries stand on thebrink of a human development crisis driven byrecession and rising poverty.

For the Education for All goals adopted in Dakarat the World Education Forum, 2010 will be a makeor break year. The past decade has witnessedremarkable progress on many fronts. The numberof children not in school has been falling, gendergaps are narrowing and more children arecompleting a basic education. Some of the world’spoorest countries have demonstrated thatuniversal primary schooling and wider educationgoals set for 2015 are attainable. With just fiveyears to go to the target date, the challenge is toconsolidate these gains and accelerate progressin countries that are off track. The danger is thatthe aftershock of the financial crisis will slow,stall or even reverse the hard-won gains of thepast decade.

Such an outcome would be indefensible. Childrenliving in the urban slums and rural villages of theworld’s poorest countries played no part in thereckless banking practices and regulatory failuresthat caused the economic crisis. Yet they standto suffer for the gambling that took place on WallStreet and other financial centres by losing theirchance for an education that could lift them outof poverty. The guiding principle for internationalaction should be a commitment to ensure thatthe developing world’s children do not pay forthe excesses of the rich world’s bankers.

Policy-makers need to recognize what is at stake.Developments in education – unlike indicators forstock markets, economic growth and the stabilityof financial systems – take place beyond the glareof media attention and public scrutiny, and aretypically reported after the event. Following adecade of broad-based progress, governmentsmight assume that underlying trends will remainpositive. But reversals in education can happen,as the experience of the 1990s demonstrated,and they have far-reaching consequences.

Depriving children and youth of opportunities forlearning has damaging implications for progressin other areas, including economic growth, povertyreduction, employment creation, health anddemocracy. If the financial crisis is allowed tocreate a lost generation in education, this willsound the death knell for the MillenniumDevelopment Goals, the international targets setfor 2015 – and it will call into question the futureof multilateral cooperation on development.

First, avoiding that prospect requires actionon two levels. National governments need tostrengthen their focus on fairness in publicspending to protect poor and vulnerable peoplefrom the impact of the economic crisis. Second,the world’s richest countries need to support low-income countries by providing concessionalfinancing. Without this lifeline, large-scale andmostly irreversible human development setbacksare inevitable. Education systems will sustainsevere damage – and children marginalizedby poverty, gender and ethnicity stand to bearthe brunt.

At successive summit meetings, political leadersof the Group of Twenty (G20) and the Group ofEight (G8) have helped stave off a deeper economiccrisis by increasing global liquidity, stabilizingfinancial systems and unlocking credit markets.Unfortunately, little has been done to protecthundreds of millions of the world’s mostvulnerable people from the impact of a crisisthey had no part in creating. The world’s richestcountries have moved a financial mountain tobail out their banking systems, but have mobilizedan aid molehill for the world’s poor.

Progress since Dakar has been driven partlyby stronger policies in education, but also byaccelerated economic growth and povertyreduction. Now, just five years before the 2015Education for All target date, policy-makers areoperating in a far more hostile environment.The financial crisis and steep food price riseshave created ‘perfect storm’ conditions for amajor setback. Slower economic growth couldtrap another 90 million people in poverty in 2010 –and more children face the threat of malnutrition.Meanwhile, national education budgets arecoming under intense pressure. In the absenceof an effective international response, low-incomecountries in particular will find it difficult toprotect spending on education, let alone to scaleup investment.

The danger is thatthe aftershock ofthe financial crisiscould reversethe hard-won gainsof the past decade

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

2 0

This chapter has five core messages:

The economic slowdown has far-reachingconsequences for education financing in thepoorest countries. Slower growth and decliningrevenue are jeopardizing public spending plansin education. For sub-Saharan Africa, theresources available for education could fall byUS$4.6 billion a year on average in 2009 and2010, or more than twice the current amountof aid to basic education in the region. Spendingper primary school pupil could be as much as10% lower in 2010 because of the effects of therecession. This potentially damaging outcomeunderlines the importance of real time budgetmonitoring, with a focus on adjustments to 2009budgets and spending outcomes, and theformulation of 2010 budgets.

Increased international aid would help reducebudget pressures. Governments in the world’spoorest countries urgently need an increase indevelopment assistance to offset revenue losses,sustain high-priority social spending andundertake the countercyclical investmentrequired to create the conditions for recovery.New evidence set out in this chapter shows thatlow-income countries in sub-Saharan Africa havea limited ability to shield public spending fromthe effects of the downturn, but a significantcapacity to productively absorb increased aid.In addition, a temporary moratorium on officialdebt payments would reduce pressure ongovernment budgets, potentially releasingresources for spending in areas such aseducation and health. Such a moratorium wouldbe in the spirit of the fiscal stimulus packagesdeployed in developed countries, attenuating theimpact of the global crisis on economic growthand poverty reduction efforts. The cost of thedebt moratorium for forty-nine low-incomecountries would amount to around US$26 billionfor 2009 and 2010 combined.

The international response to the financial crisishas failed to address major human developmentconcerns. Global summits and domestic policiesin rich countries have played a crucial role instabilizing financial systems and establishingthe foundations for early recovery. By contrast,the response to the crisis unfolding in theworld’s poorest countries has been markedby systemic indifference. ‘Smoke and mirrors’financial reporting has produced large headlinenumbers for financial transfers while obscuring

the modest level of real resources mobilized.Sub-Saharan Africa stands to lose someUS$160 billion in government revenue in2009–2010 as a result of slower growth andreduced revenue. Best estimates of theinternational response for low-income countriessuggest that additional concessional finance forthe period will amount to no more thanUS$6 billion to US$8 billion.

Education for All financing gaps should be closedunder a human development recovery plan.Governments, aid donors and financialinstitutions urgently need to assess the financinggaps for achieving the Millennium DevelopmentGoals. Making available the resources requiredto close these gaps should be part of thecoordinated international response to the globalfinancial crisis. A major new financial costingexercise carried out for this Report (discussed indetail in Chapter 2) puts the Education for Allfinancing gap at around US$16 billion. Thatheadline figure appears large in absolute terms,but has to be placed in context. It represents lessthan 2% of the financial rescue package puttogether by governments in just two countries –the United Kingdom and the United States – forfour commercial banks and is equal to a smallfraction of the wider financial systems bail-out.

International action must be taken before the2010 Millennium Development Goal summit.The impact of the financial crisis and newevidence on the scale of financing gaps demandan effective international response. With aMillennium Development Goal summit plannedfor 2010, the United Nations Secretary-Generalshould convene a high-level meeting of donorsand governments of low-income countries toreassess the external financing required toachieve the Education for All goals.

This chapter is divided into three parts. Part 1looks at the mechanisms through which thefinancial crisis and the food crisis are hurtingeducation systems. Part 2 examines ‘fiscal space’,the room for manoeuvre that governments haveto protect public spending in education and otherareas from the effects of the global economicdownturn. Part 3 critically reviews the internationalresponse to the crisis, highlighting in particularthe failure of the current G20 framework.

Governments in the world’s

poorest countriesurgently need an increase indevelopmentassistance to

offset revenuelosses

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

D o u b l e j e o p a r d y : f o o d p r i c e s a n d f i n a n c i a l c r i s i s

2 1

Double jeopardy: foodprices and financial crisis

‘We were hearing that there was no work and thefactory would be shut down. It all happened quitefast actually. Although there was much talk aboutthe factory shutting down, the authorities did notreally tell us anything until almost the last week.’

Anwarul Islam,a Bangladeshi migrant labourer in Jordan

‘Since I lost my job sometimes we eat only once or twice a day. I don’t know what to do. We are justcamping in front of the factory gates, waiting forthe company to pay us.’

Kry Chamnan,garment worker in Cambodia, February 2009

‘My factory retrenched 150 workers including me.I’m in deep trouble thinking about how to live with my two children.’

Lalitha, a 35-year-old worker in Sri Lanka

‘You think about your children when you lose yourjob. That’s the first thing that came into my mind –when school starts, how am I going to buy theuniform, the exercise books and all that. The food,you know how expensive that is now…The childrendepend on me, I’m a single mother.’

Kenia Valle, Managua, Nicaragua

These four voices provide a reminder that, inan increasingly interdependent world, economicshocks travel rapidly across borders (Emmett,2009). Faced with a daily barrage of reportingon the state of the global economy and recoveryprospects for rich countries, it is easy to losesight of the human costs of the global downturnfor those who live away from the media spotlight.The recession, sparked by reckless gamblingon Wall Street and the regulatory failures in richcountries, is leaving its mark on people living inslums and remote villages in the world’s poorestcountries. The effects on education systems arecomplex and varied, but overwhelminglydestructive.

Economic slowdown threatenseducation financing

The financial crisis is being transmitted toeducation systems through various channels.The degree to which countries are integrated intointernational trade and financial markets, thestructure of employment, patterns of import andexport, and pre-existing poverty levels all play apart in determining who is affected and for howlong (McCord and Vandemoortele, 2009; te Veldeet al., 2009). For low-income countries, trade isthe primary transmission mechanism from worldmarkets to the national economy, with exportersof minerals and primary commodities hit by acombination of lower prices and falling demand(IMF, 2009b, 2009e).

Deteriorating prospects for economic growth havefar-reaching implications for education financing.Since the onset of the crisis, growth forecasts havebeen revised downwards on a regular basis. Alldeveloping regions are affected. With a pre-crisisgrowth forecast of over 5%, sub-Saharan Africanow faces the prospect of growing at less than 2%,which is below the rate of population increase.Latin America is projected to face an economiccontraction in 2009 (Figure 1.1).

Slower growth and declining export and importactivity have adverse consequences for governmentrevenue and hence for public spending (IMF, 2009b,2009d). Budgetary pressure is evident in data onfiscal balances. Sub-Saharan Africa is movingfrom a fiscal surplus in 2008 to a projected 2009deficit equal to about 6% of gross domestic product(IMF, 2009e). The combined effect of slowereconomic growth and lower levels of revenuecollection will translate into losses equivalent toabout US$80 billion in 2009 and the same in 2010(Table 1.1). This is revenue that could have beenused for investment in areas ranging fromeconomic infrastructure to health and education.

The importance of economic growth for educationfinancing is not widely recognized. Rising wealth isnot automatically associated with improvement ineducation – and many countries with low averageincomes have registered extraordinary progress.But increasing national income does createfinancing conditions conducive to higher publicspending on education. Economic growth expandsthe resources available to governments throughtaxation. Moreover, the share of national incomecollected in government revenue tends to rise as

In an increasinglyinterdependentworld, economicshocks travelrapidly acrossborders

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

2 2

poverty falls, and economic growth is an importantcondition for sustained poverty reduction.

The experience of sub-Saharan Africa is instructive.During the 1990s, economic stagnation and highlevels of external debt undermined governments’capacity to finance education, with per capitaspending declining in many countries. That picture

has changed dramatically, with public spending onprimary education rising by 29% over the periodfrom 2000 to 2005 (Figure 1.2). This turnaroundwas instrumental in reducing the numbers ofchildren out of school and strengtheningeducation infrastructure. Around three-quartersof the increase was directly attributable toeconomic growth, with the balance accountedfor by increased revenue collection and budgetredistribution in favour of the education sector.

What does the economic slowdown mean foreducation financing in sub-Saharan Africatowards 2015? The answer will depend on theduration of the slowdown, the pace of recovery,governments’ approach to budget adjustmentsand the response of international donors. Thereare many uncertainties in each area. Nevertheless,governments have to draw up public spendingplans in an uncertain environment. One wayof capturing the potential threat to educationfinancing is to consider a scenario that holdsthe share of expenditure invested in educationconstant, with adjustments for reduced economic growth and lower revenue-to-GDP ratios (Figure 1.3).

-4

-2

0

2

4

6

8

10

20042003 2005 2006 2007 2008 2009

GDP

annu

al g

row

th ra

te (%

)

Advanced economiesJuly 2009

Emerging anddeveloping economies

07/09

01/09

11/08

10/0804/08

Figure 1.1: Post-crisis economic growth projections have been revised downwards for all developing regionsReal GDP growth projections since April 2008, selected regions, 2003–2009

Note: Regions shown are those used by UNICEF, which differ to some extent from the Education for All regions.Source: IMF (2009f).

24.7 20.8 21.6

378 402 427376 322 347

1.4 79.8 80.3

1.4 16.4 26.70.0 63.4 53.6

Revenues, excluding grants (%GDP) average, April 2009

Pre-crisis government revenue projection1

Post-crisis government revenue projection2

Potential revenue loss associated with economic crisis

Slower economic growthDecreased revenue-to-GDP ratio

Table 1.1: Potential revenue loss in sub-Saharan Africa, 2008–2010

Notes: These estimates are based on weighted and aggregated single country gross domestic productprojections. Countries were weighted using GDP based on the purchasing power parity share of the region. ‘Pre-crisis projections’ are for April 2008 and ‘post-crisis projections’ for April 2009. Excludes Somalia andZimbabwe.1. Based on April 2008 growth projections and 2008 revenue-to-GDP ratios.2. Based on April 2009 growth projections and adjusted revenue-to-GDP ratios.Sources: IMF (2008, 2009e, 2009g).

2008Constant PPP 2006 $US billions unless specified 2009 2010

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

D o u b l e j e o p a r d y : f o o d p r i c e s a n d f i n a n c i a l c r i s i s

2 3

-4

-2

0

2

4

6

8

10

12

2007 2008 2009

April 2008

July 2009

Sub-Saharan Africa

GDP

grow

th ra

te (%

)

-4

-2

0

2

4

6

8

10

12

2007 2008 2009

April 2008

July 2009

Developing Asia

GDP

grow

th ra

te (%

)

-4

-2

0

2

4

6

8

10

12

2007 2008 2009

April 2008

July 2009

Middle East

GDP

grow

th ra

te (%

)

-4

-2

0

2

4

6

8

10

12

April 2008

July 2009

Latin America and the Caribbean

GDP

grow

th ra

te (%

)

2007 2008 2009

Figure 1.2: Economic growth matters for education financingPrimary education expenditure in sub-Saharan Africa between 2000 and 2005,growth decomposition

-10

-5

0

5

10

15

20

25

30

Expl

aine

d va

riatio

n (%

)

Totalvariation

in spending

Change due to GDPgrowth

Change dueto revenuecollection

Change due to education

budgetallocation

Change dueto primaryeducation

budgetallocation

Sources: Pôle de Dakar (2002, 2004, 2005, 2007); UIS database.

Figure 1.3: Education financing in sub-Saharan Africa could sufferfrom slower economic growthEstimated forgone income for education due to the crisis in 2009 and 2010

-6

-5

-4

-3

-2

-1

0

1

Pre-crisis projections

Cons

tant

PPP

200

6 US

$ bi

llion

s

2009 2010

Primary education spending

Total education spending

Notes: The shares of GDP devoted to education and primary education have beenconsidered constant and correspond to median shares in 2007 for both data. Forgone incomeis the difference between education spending estimated with pre-crisis projections and thatcalculated with the most recent post-crisis projections.Sources: IMF (2008, 2009g); UIS database.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

2 4

In terms of the potential resources forgone foreducation spending, such a scenario would result in:

an average loss over 2009 and 2010 ofUS$4.6 billion per year, compared with estimatedaid disbursements to the region for basiceducation of US$2 billion;

a cumulative loss to 2013 of about US$30 billion;

a loss in 2010 of US$13 per pupil for primaryschool spending – equivalent to about 10% ofcurrent spending per pupil.

These figures provide only an estimate of onepossible scenario. They do not chart an inevitablecourse. Even so, the magnitude of the potentialeconomic growth effect serves to illustrate thebudget pressures many countries face. And thesenew pressures have to be seen against thebackdrop of an already large external financinggap – averaging around US$16 billion a year –for the Education for All goals in low-incomecountries (see Chapter 2).

Wider human impact

The economic downturn in the poorest countrieshas had direct consequences for vulnerablehouseholds. For people surviving below or justabove the poverty line, it has meant less securelivelihoods. Income from remittances is falling.Employment prospects are diminishing in manycountries. And the downturn has followed hard onthe heels of a steep rise in international food prices,with higher levels of poverty superimposed ondeteriorating nutrition indicators.

The combination of global food crisis and financialcrisis has worsened the environment for achievingthe Education for All goals. From 2003 to 2008, cornand wheat prices roughly doubled and rice pricestripled. Domestic price rises have not tracked thoseof international prices, but food price inflationreached over 17% in sub-Saharan Africa during2008, rising to 80% in Ethiopia (Lustig, 2009; vonBraun, 2008). In other regions, many countriesrecorded inflation rates in excess of 10%. Becausepoor households spend a large share of theirbudgets on food, price rises hit them particularlyhard (World Bank, 2008a, 2008e). Many have had tocope either by diverting spending from other areasor by going hungry. Meanwhile, governments havefaced rising food import bills and budget costs fornutrition programmes. Although food prices have

started to fall, they remain high by historicalstandards. At the end of 2008, domestic staplefood prices across a large group of developingcountries averaged 24% higher than two yearsearlier (FAO, 2009).

The lethal cocktail of high food prices and economicrecession has left a deep imprint on the lives ofmillions of vulnerable people. According to the Foodand Agricultural Organization of the United Nations(FAO), the number of malnourished people in theworld increased by 75 million in 2007 and by100 million in 2008, reaching a global level of justover 1 billion (FAO, 2008). Recent FAO projectionsfor 2009 indicate the financial crisis could push125 million additional people into malnutrition(Headey et al., 2009). In some regions, droughthas exacerbated underlying food security pressureassociated with higher prices. For example, inEthiopia, 12 million people are in immediateneed of food and other assistance.

Poverty levels continue to fall, principally as aresult of strong economic growth in China andIndia, but the rate of decline has slowed markedly.According to the World Bank, the downturn willleave an additional 75 million people below theUS$1.25 poverty threshold in 2010 and anadditional 91 million below the US$2 threshold(Chen and Ravallion, 2009).

Rising malnutrition and deteriorating prospects forpoverty reduction have far-reaching consequencesfor education. Hunger undermines cognitivedevelopment, causing irreversible losses inopportunities for learning. There are often long timelags between the advent of malnutrition and dataon stunting. But increased malnutrition among pre-school and primary school age children hasbeen reported from several countries, includingGuatemala (von Braun, 2008). Rising food priceshave also had wider consequences for the placeof education spending in household budgets.In Bangladesh, about a third of poor householdsreport cutting spending on education to cope withrising food prices (Raihan, 2009). In Ghana andZambia, poor households report eating fewer andless nutritious meals, and reducing expenditureon health and education (FAO, 2009). Governmentbudgets have also been affected. In September2009, Kenya announced plans to delay financingof free education for 8.3 million primary schoolchildren and 1.4 million secondary school children,prompting school administrators to press for atemporary restoration of user fees. The government

Risingmalnutrition and

deterioratingprospects for

poverty reductionhave far-reaching

consequences for education

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

D o u b l e j e o p a r d y : f o o d p r i c e s a n d f i n a n c i a l c r i s i s

2 5

claimed costs associated with emergency feedingprogrammes forced the delay. More equitableavenues could have been explored, however.

Diminished prospects for reducing poverty willseverely damage efforts to accelerate progresstowards the Education for All goals. More povertymeans parents have less to spend on children’seducation. Household poverty also pushes children out of school and into employment. Counteracting the impact of rising poverty and deteriorating nutritionwill require strengthening of social protectionprogrammes – an issue discussed in Chapter 3.

Sub-Saharan Africa, which has the furthest to travel to achieve universal primary schooling, faces some of the starkest poverty-related threatsto education. The region’s recent progress has been encouraging, driven by strong economicgrowth and poverty reduction. For the first time in over a generation, numbers living below theUS$1.25-a-day poverty line have fallen: some4 million people climbed out of poverty between2000 and 2007. With per capita income set toshrink in 2009, however, poverty levels could rise.

The impact of rising unemployment is alreadyregistering on education systems as householdbudgets come under pressure (World Bank andIMF, 2009). In Zambia, around a quarter of jobs inthe copper mining sector have been lost (te Veldeet al., 2009). Rising unemployment was alsoreported in copper mining in the DemocraticRepublic of the Congo after export prices collapsed.In both countries, there have been reports ofunemployed workers having to withdraw childrenfrom school (Hossain et al., 2009; Times of Zambia,2009). Women often bear the brunt of deterioratinglabour market conditions. One reason is that theyare often concentrated in the hardest-hit exportindustries, such as garments and electronics.Limited employments rights and social insurancefurther increase their vulnerability (Emmett, 2009;ILO, 2009b). Evidence from Cambodia’s garmentindustry points to women being required to worklonger hours for less pay, with adverseconsequences for education spending.

Household provision for education financing isdirectly affected by the loss of remittances, acrucial element of financial transfers from richerto poorer countries – the US$308 billion transferredin 2008 far exceeded international developmentassistance. Flows of remittances are projected todecline by 7% in 2009 (Ratha et al., 2009). Both the

decline and its overall effect will be uneven. Flowsto Latin America and the Caribbean are falling ina lagged response to the slowdown in the UnitedStates, with El Salvador and Mexico recordingdeclines in excess of 10% (Orozco, 2009). Ghanaand Kenya report reductions of a similar scale(IMF, 2009b). For several countries, the impactwill be very marked. In ten sub-Saharan Africancountries, including Ethiopia, Kenya, Liberia,Senegal, Sierra Leone and Uganda, remittancesare equivalent to over 5% of GDP, rising to 20% inLesotho (Committee of Ten, 2009). Also, with globalremittances falling and urban unemployment rising,transfers to rural areas are declining.

Much of the evidence on the impact of thesedevelopments on education is anecdotal. Even so,it points in a worrying direction. Remittances areoften vital to household spending on education.Evidence from Ghana and Uganda shows that asmuch as one-quarter of remittance income goesto education, pointing to potentially large lossesof household investment (te Velde et al., 2009). InEl Salvador and Haiti, where money sent from theUnited States contributes significantly to financingeducation, parents report growing difficulties inkeeping children in school as remittances decline(Grogg, 2009; Thomson, 2009). It is not just theloss of international remittances that is hurtingeducation. In China, unemployment has forcedan estimated 20 million migrants to return torural areas, and money that was previously beingremitted for education has dried up (Mitchell, 2009).

Evidence from previous recessions and otherexternal shocks shows how crucial it is for rich andpoor countries alike to address the human costs ofthe current downturn. The East Asia financial crisisof 1997 resulted in major reversals in child healthand education (Ferreira and Schady, 2008; Harperet al., 2009). In Indonesia, infant mortality increasedand the proportion of children not enrolled inschool doubled in 1998 (Frankenberg et al., 1999;Paxson and Schady, 2005a). The number of streetchildren also rose sharply (Harper et al., 2009).Drought and disrupted rainfall have deliveredsimilar setbacks to education in sub-Saharan Africa(Jensen, 2000; World Bank, 2007c). Not all theeffects on education are straightforward. In somemainly middle-income countries, school enrolmentincreases during crises, partly because risingunemployment and falling wages lower theeconomic returns to child labour (Ferreira andSchady, 2008). But the overall impact is universallyharmful to progress in education.

It is critical forrich and poorcountries alike toaddress the humancosts of thecurrent downturn

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

2 6

The challenge facing policy-makers today is to avoidrepeating the experience of past crises. As theeffects of the economic downturn are transmittedto more households, those lacking the resourcesto cope with the shock risk being pushed into adownward spiral. Short-term coping strategiessuch as cutting spending on health, nutritionand education can have damaging long-termconsequences for individuals and societies.Governments and the international communitycan contain the damage by investing in socialprotection. But a consistent lesson from previouscrises is that early, up-front investment in crisisprevention through social protection is moreeffective than treatment after the event.

Budget monitoring matters

Many governments in low-income countriesare reassessing public spending plans in theface of mounting fiscal pressure. Their room formanoeuvre depends on a range of factors, includingthe pre-crisis fiscal balance, recovery prospects,and domestic and international financing options.The impact of budget adjustments on publicspending plans for education will vary accordingto circumstance and policy choice. Options includecutting spending in real terms, scaling downplanned increases or maintaining current spendingplans through revenue raising and redistributionwithin the budget. Decisions made over the nextyear in these areas will have profoundconsequences for education financing. Publicspending cuts, or caps that are set below plannedlevels, will ultimately translate into fewerclassrooms built, fewer teachers recruitedand trained, and more children out of school.

Current monitoring exercises do not adequatelytrack budget decision-making processes.International data provide comprehensive cross-country coverage of public spending, butwith a significant delay. For example, this year’sReport documents expenditure for 2007. Whilevital for monitoring broad post-Dakar trends, suchinformation reveals nothing about the direction ofthe public spending plans that will define the future.

This information gap is difficult to defend. Data forthe current and previous budget years are availablein most countries, as are budget revision andreview documents. The problem is that the dataare not assembled and made publicly available byinternational or regional organizations. In the wordsof an analysis of education budgets in sub-Saharan

Africa carried out for this Report: ‘It is rathershocking in view of the strong emphasis givento monitoring progress … that there is not amore current database for analysing educationspending’ (Martin and Kyrili, 2009, p.14). Thecurrent crisis has added to the urgency of fillingthis information gap.

One central conclusion is that UNESCO should befar more effective in monitoring current-year publicspending on education and reviewing revisions tofuture spending plans.1 Through UNESCO regionaloffices and the UNESCO Institute for Statistics, aregional network of education ministries’ planningand budget directors could be established. Tofurther strengthen the monitoring process, nationalpoverty reduction strategy coordinators could beincluded, along with finance ministry officialsoverseeing medium-term expenditure strategies.

To assess the threat to Education for All financing,the Global Monitoring Report team commissioned Development Finance International to review the 2009 budgets of all thirty-seven low-income countries insub-Saharan Africa (Martin and Kyrili, 2009). Thisshould be viewed as both a partial and preliminaryexercise. It is partial because detailed andconsistent information on education expenditurefrom the 2009 budget was available for only twelvecountries, and it is preliminary because the budgetdocuments consulted reflect pre-crisis conditions.Broad budgetary patterns for the countries coveredcan be summarized under four headings:

Plans to increase expenditure on education inrelation to GDP and the overall budget. Five ofthe twelve countries are in this category: BurkinaFaso, Liberia, Mozambique, Sierra Leone andZambia. Liberia, Sierra Leone and Zambiaenvisaged a significant reallocation to educationwithin a growing budget. In Mozambique,education spending was projected to growsignificantly as a share of GDP but onlymarginally as a share of the budget, reflectinga planned rise in non-social sector spending.It should be stressed that governments in severalof these countries have raised concerns overtheir capacity to finance planned educationspending. A May 2009 report on Mozambique,for example, projects that revenue will be 1.3%below the level indicated in the approved budget,which could adversely affect spending plans.

Plans to maintain education spending at currentlevels in relation to GDP and total budget

1. The Education Sectorin UNESCO has startedto put in place some ofthe elements of such anapproach. In March 2009,it launched a ‘QuickSurvey’ to collect budgetinformation on publicexpenditure in education.UNESCO professionalstaff were invited to fill ina questionnaire capturingtheir assessment ofnational budget plans.Unfortunately, themethodology was notconducive to accuratereporting of budgetinformation, spendingplans or impacts of thefinancial crisis. UNESCOhas acknowledged theneed for more rigorousand timely collection ofcurrent-year budget data(UNESCO, 2009b).

Short-termcoping strategies

can havedamaging long-term

consequences for individualsand societies

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

D o u b l e j e o p a r d y : f o o d p r i c e s a n d f i n a n c i a l c r i s i s

2 7

spending. Two countries – Kenya and Uganda –fall into this category. Both have significantlyincreased financing for education in recent years.Here, too, crisis-related problems could hamperimplementation.

Plans to increase education spending as ashare of GDP but to maintain or cut the shareof education in the national budget. The countriesconcerned are Lesotho, Rwanda and the UnitedRepublic of Tanzania. Lesotho plans to raisethe ratio of education spending to GDP whilemaintaining the budget share. Rwanda’s budgetenvisages a rise in the share of educationspending in GDP but a slight fall in the budgetpercentage because of a shift towardsagriculture and infrastructure. The UnitedRepublic of Tanzania plans to maintain the GDPshare held by education spending but to reducethe budget share, as the country’s nationalpoverty reduction strategy entails dramaticincreases in expenditure on agriculture,infrastructure and water. All these plans arehighly susceptible to economic pressure, whichcould change patterns of budget allocation tothe detriment of basic education. For example,Lesotho’s response to the threat of risingunemployment was to shift spending prioritiesfrom pre-school and primary education totechnical and vocational training.

Plans to cut education spending as a shareof GDP and total expenditure. Two countries –Benin and Ghana – fall into this category. InBenin, the planned cut reflects reallocationof budget spending away from education andother social sectors. In Ghana, it is less a directresult of the economic crisis than an effect ofa domestic budget crisis inherited from theprevious government. In both cases, it is likelythat stagnant or declining economic growth willcompound the cuts, resulting in significantlyfewer resources available for educationspending. There is a danger that Benin’s strongprogress in recent years towards universalprimary education, documented in Chapter 2,will be reversed. In Ghana, efforts to addresseducation marginalization in the north couldbe undermined (see Chapter 3).

This overview contains good news and bad news.The good news is that current evidence indicatesthat few governments are cutting educationspending. The bad news is that the changing picturemay look worse than that captured in current

budget analyses. Most budgets of low-incomeAfrican countries reviewed by Development FinanceInternational were approved by parliaments at theend of 2008, before national economies registeredany significant impact of the crisis. Mid-termbudget reviews may result in marked adjustmentsin spending. Close monitoring of actual spendingon education, and of restrictions on spending, isvital. Formal revisions to 2009 budgets and publicspending plans drawn up amid changing fiscalconditions have to be carefully assessed, as dodiscrepancies between 2009 budget allocationsand actual spending. But the full impact of thedownturn is likely to be more fully revealed in 2010.There is already evidence of budget revision insome countries. For example, after copper pricescollapsed, Zambia’s government removed a windfalltax on mining companies that was to have financedan increase in education and other social spending(te Velde et al., 2009).

It is important to base budget monitoring exerciseson appropriate benchmarks. Much has been madeof the fact that, to date, relatively few low-incomecountries have cut public spending in general orpriority social sector spending in particular. Asfar as it goes, this is clearly a positive outcome.However, what ultimately matters for progresson the Education for All goals and wider humandevelopment measures is whether plannedincreases in public spending have beencompromised. Governments in many low-incomecountries have drawn up medium-term expenditureplans for education, often as part of wider povertyreduction strategies supported by donors. Theplans are linked to activities such as classroomconstruction, teacher recruitment, purchases ofteaching materials and special programmes formarginalized children. These activities are in turnaimed at specific targets for getting children intoschool and raising the quality of education. To theextent that budget pressures translate into levelsof expenditure that are lower than planned, theywill compromise any prospect of acceleratedprogress towards the Dakar goals.

What happens beyond the education sector isalso crucial. Progress in education is inevitablyinfluenced by developments in other key areas,including child and maternal health, and water andsanitation. The national and international responseto the economic crisis thus needs to reflect anintegrated strategy for protecting humandevelopment across a broad front.

Close monitoringof actual spendingon education, and of restrictionson spending, is vital

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

2 8

Expanding ‘fiscal space’:an Education for Allpriority

The term ‘fiscal space’ describes a factor that hasprofound consequences for governments’ capacityto finance vital social and economic programmes.Put most simply, it is about room for manoeuvrein national budgeting. Tax revenue is the primarysource of finance for public spending. Butgovernments can also resort to other revenue-raising measures, including domestic orinternational borrowing, printing money and, inthe case of the poorest countries, international aid.The options open to governments vary widely –but they are most limited in the poorest countries.

‘Fiscal space’ defines the budget parameterswithin which governments have to operate. TheInternational Monetary Fund (IMF) defines it as‘room in a government’s budget that allows it toprovide resources for a desired purpose withoutjeopardizing the sustainability of its financialposition or the stability of the economy’ (Heller,2005, p. 32). Less technocratic approaches wouldincorporate the financing of wider humandevelopment goals, including Education for All(Roy et al., 2007). In the context of the globalrecession, the issue facing governments is thatof using national budgets to strengthen demand,stabilize financial systems and maintain vital socialinvestments despite a shrinking revenue base.

Rich countries have responded to the financialcrisis by exploiting fiscal space on an epic scale.With their economies contracting, their financialsystems requiring support and demands on publicspending for social welfare rising, fiscal policy hasprovided a major stimulus. Overall fiscal deficits areprojected to increase by about six percentage pointsof GDP, with spending financed by a large increasein public debt.2 Much of this has been used to shoreup banking systems.3 While bank bail-outs are notstrictly comparable to aid flows in financial terms,the contrast between what has been mobilized inthe two cases is striking. The four largest assetinsurance programmes for commercial banks

obliged the governments of the United Kingdomand the United States to take on US$786 billion inpotential liabilities – over seven times the amount oftotal international development assistance flows.4

Fiscal policy has also played a wider role inadvanced economies. Public spending has gone to support demand and unlock credit markets,creating a countercyclical stimulus for recovery.Many governments have used that spending tostrengthen the social and education infrastructure.In the United States, the American Recovery andReinvestment Act (ARRA) passed by Congressin February 2009 delivered a prospectiveUS$789 billion stimulus to the economy. Thatstimulus also staved off a financing crisis ineducation that threatened to result in thousandsof teachers being laid off and many schoolsclosed (Box 1.1).

Unlike rich countries, most developing countriesoperate in a highly constrained fiscal environment.Some, including China and India, have been in aposition to counteract the impact of the downturnthrough increased public spending. But themajority of the poorest countries are walkinga fiscal tightrope. Overall tax revenue ratios areprojected to decline in well over half of all low-income countries and by more than 2% of GDPin one-quarter of them (IMF, 2009c). Meanwhile,pressures to increase spending arise from severalsources, including the need to finance socialprotection programmes. The combination oflimited fiscal space and revenue decline has thepotential to translate into painful public spendingadjustments, including in education.

The research for this Report by DevelopmentFinance International explored the dimensionsof the fiscal space available to thirty-seven low-income countries in sub-Saharan Africa that arefacing financing challenges in education (Martinand Kyrili, 2009).5 This ‘fiscal space assessment’starts by defining ‘sustainability thresholds’,based on comparative international evidence,in three key areas: domestic and internationalborrowing, revenue mobilization and aid.6

2. On average, public debt will climb from around 70% of advancedeconomy GDP in 2008 to a projected 100% by 2010 (IMF, 2009g).

3. Capital injections, debt guarantees and asset guarantees represented44% of GDP for the United Kingdom and 7% for the United States as ofJune 2009 (Martin and Kyrili, 2009).

Unlike richcountries,

most developingcountries lack room

for manoeuvre in nationalbudgeting

4. The programmes involved Citigroup, Royal Bank of Scotland, Lloyds and Bank of America (Panetta et al., 2009).

5. The countries are those classified by the World Bank as ‘IDA-only’:eligible only for concessional International Development Associationloans. Hence the list includes Cameroon, the Democratic Republic of the Congo and Djibouti, even though the latest World Bank data put themin the lower middle income category.

6. See Martin and Kyrili (2009) for a detailed explanation of the thresholds.

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

E x p a n d i n g ‘ f i s c a l s p a c e ’ : a n E d u c a t i o n f o r A l l p r i o r i t y

2 9

Sustainable borrowing. Given the region’s longhistory of unsustainable external debt, borrowingon international markets comes with high riskfor most low-income countries in sub-SaharanAfrica. The assessment sets a threshold forexternal debt based on the ‘Debt SustainabilityFramework’ developed by the IMF-World Bank.For domestic debt, it uses the IMF thresholdindicator of a nominal debt stock of 15% of GDPas an indicator for sustainability.

Sustainable domestic revenue levels. Raisingmore revenue is another way for governmentsto generate resources for public spending. Low-income African countries have made majorstrides in recent years by increasing taxes andexpanding the tax base, but it is widelyrecognized that there are limits to how much

they can increase tax collection. Governmentshave to avoid creating disincentives forinvestment and generating deflationary pressure,especially in the current context. In the absenceof a viable threshold indicator, the assessmentuses an ‘acceptable effort’ indicator for revenuecollection set at 17% of GDP (excluding grants).This is one of the convergence criteria for theCFA franc zone.

Sustainable aid levels. Another way for low-income African countries to expand fiscal spaceis to obtain more grants. While aid flows to Africahave increased substantially in recent years, theystill fall far short of overall pledges made in 2005and education-specific pledges made in 2000(see Chapter 4). Studies have indicated thatexcessive aid dependence can have damaging

In the UnitedStates, aroundUS$130 billion willbe injected intoeducation andrelated budgets

Governments across the developed world haveused national budgets to counteract the effectsof the economic downturn. In the United States, theUS$789 billion American Recovery and ReinvestmentAct (ARRA) of February 2009 is aimed at providinga platform for early recovery and protecting thesocial and economic infrastructure, with educationa high priority.

ARRA has attempted to turn the threat to educationposed by the recession into an opportunity. Withpublic finances damaged by slower economic growthand rising expenditure in other areas, educationspending was in jeopardy in many states. Thousandsof teachers faced the prospect of being maderedundant. Under ARRA, the federal governmentstepped into the breach left by collapsing statefinancing (which accounts for around 90% ofeducation spending). Around US$130 billion willbe injected into education and related budgetsto stabilize finances and extend opportunities forchildren from disadvantaged backgrounds. Thefollowing are among the most important measures:

States are to benefit from US$39.5 billiondesignated for public school districts and highereducation institutes under the ‘state fiscalstabilization’ fund.

School construction and upgrading projectswill receive US$22 billion.

Funding for targeted programmes aimed atspecial education and children from the mostdisadvantaged backgrounds will be increased byaround US$25.2 billion. ARRA will increase 2009

fiscal year spending on Title I — a set of specializedclassroom programmes supporting learning inschools with high concentrations of poor children —to US$20 billion from about US$14.5 billion.Spending on education for children with disabilitieswill rise to US$17 billion from US$11 billion.

Head Start and early Head Start pre-schoolprogrammes will receive an increase ofUS$2.1 billion.

About US$4.3 billion has been allocated to a ‘Race to the Top’ Fund aimed at recruiting andretaining effective teachers and raising standardsin low-performing schools.

ARRA has sparked a wide-ranging debate aboutthe respective roles of federal and state governmentsin education financing. With the Department ofEducation’s discretionary budget rising fromUS$60 billion in 2008 to a projected US$146 billionin 2010, the balance between state and federalfinancing has been dramatically changed. Butwithout the emergency financing, thousands ofteachers would have been made redundant, manyschools would have closed and education qualitywould have suffered. As one congressman put it,‘We cannot let education collapse; we have toprovide this level of support to schools.’

The world’s poorest countries cannot afford to leteducation collapse either. Yet, unlike rich countries,most lack the budget resources to provide the supporttheir education systems need to avert collapse.

Sources: US Department of Education (2009); National EducationAssociation (2009); Dillon (2009).

Box 1.1: The Obama rescue plan — protecting education during the economic downturn

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

3 0

consequences for economic growth andgovernance, but there are no clear parametersfor sustainable aid levels. Using evidence fromrecent studies, the Development FinanceInternational assessment assumes thatcountries receiving aid levels that exceed 25%of gross national income have no space toincrease their aid dependence.

These three pillars of sustainability cannot beviewed in isolation. Even if a country has scopeto gain access to more finance through borrowingor aid, it may decide not to exploit this fiscal spacebecause of the risk of macroeconomic instability.The fiscal space assessment therefore includeschecks for fiscal balance and inflation.

Overall fiscal space is assessed in two steps.First, each country’s ability to obtain and useresources through each of the three instruments –debt, domestic revenue and aid – is determinedby reference to the thresholds. Countries belowall three thresholds are described as having highfiscal space. Countries constrained on oneindicator are classed as having moderate space,on two indicators as having low space and on allthree as having no space. The second step is toadjust the outcomes to reflect the two checkindicators.7

Table 1.2 shows the results. After adjustmentsfor fiscal deficits and inflation risk, five countrieshave no fiscal space. At the other end of thespectrum, four have high space and the optionto resort to all three financing instruments.Seventeen countries have ‘low space’ and eleven‘moderate space’, indicating scope to resort toone or two instruments, respectively. Furtheranalysis points to a diverse set of policy optionsthat depend on national circumstances andhighlights choices facing governments and theinternational community over what type of

resources should be made available to protecthigh-priority spending in education and otherareas (Figure 1.4).

International aid. Twenty-five countries havethe fiscal space to use more developmentassistance, and eleven have no domesticalternatives. This implies that an increasein grant flows is the primary means open to low-income African countries seeking to avoidcuts and sustain spending plans in high-prioritysocial sectors.

Domestic revenue. Around fifteen countriescould raise more revenue on the basis of the17% of GDP norm, but seeking to raise revenuein the midst of a steep economic downturn islikely to damage recovery prospects.

Borrowing. Between eleven and fourteencountries could borrow more withoutcompromising their overall public debtsustainability. When external and domesticdebt are looked at together, however, thescope for expansion is limited.

A vital policy lesson can be drawn from thisassessment: increased aid has the mostimmediate potential for increasing fiscal space.Early action on a sufficient scale could provide thebudget resources needed to pre-empt potentiallydamaging public spending adjustments ineducation and other areas. It is critical to deliverthis aid before fiscal pressures convert thefinancial crisis into an irreversible long-termhuman development crisis, with attendantconsequences for progress in education.

7. Here the assumptionis that a country can useaid but cannot borrowmore if its fiscal deficitexceeds 3% of GDP, sincethis could have damaginginflationary effects.Similarly, if a country hasreached the ‘acceptableeffort’ threshold forrevenue mobilization,increasing aid grantsmight push up inflation,whereas if it has scopefor revenue-raising,the inflation effects couldbe neutralized.

Increased aid has the most

immediatepotential

for increasingfiscal space

Table 1.2: Fiscal space in sub-Saharan Africa, selected countries

Moderate(11 countries)

High(4 countries)

Low(17 countries)

None(5 countries)

Congo, Côte d’Ivoire, Djibouti,Eritrea, Ethiopia, Gambia, Guinea,Guinea-Bissau, Kenya, Malawi,Mauritania, Sao Tome andPrincipe, Senegal, Sierra Leone,Sudan, Togo, Zimbabwe

Benin, Burkina Faso, Cameroon,Central African Republic, Chad,Comoros, Lesotho, Madagascar,Mozambique, Niger, Nigeria

Burundi, Democratic Republic of the Congo, Ghana, Liberia,Zambia

Mali, Rwanda, Uganda, UnitedRepublic of Tanzania

Source: Martin and Kyrili (2009).

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

E x p a n d i n g ‘ f i s c a l s p a c e ’ : a n E d u c a t i o n f o r A l l p r i o r i t y

3 1

Figure 1.4: Many countries lack room for manoeuvre in budget management but could use more aidPolicy options available to increase resources and protect social sector spending, selected sub-Saharan African countries, 2009

Low fiscal space:Policy option=

Aid only(11 countries)

Congo, Côte d’Ivoire,Djibouti, Eritrea, Gambia,

Kenya, Malawi, Mauritania,Senegal, Sudan, Togo

Low fiscal space:Policy option=

Tax increase only(6 countries)

Ethiopia, Guinea,Guinea-Bissau,

Sao Tome and Principe,Sierra Leone, Zimbabwe

Low fiscal space:Policy option=

Borrowing only(1 country)Madagascar

Medium fiscal space:Policy options=

Aid and borrowing(6 countries)Benin, Chad,

Cameroon, Lesotho,Niger, Nigeria

Medium fiscal space:Policy options=

Aid and tax increase(4 countries)

Burkina Faso, C. A. R.,Comoros,

MozambiqueHigh fiscal space:

Policy options=Aid, borrowing and taxes

(4 countries only)Mali, Rwanda, Uganda,

U. R. Tanzania

Medium fiscal space:Policy options=

Borrowing andtax increase(0 countries)

Domesticand external

borrowing

Aid

Taxincrease

Source: Martin and Kyrili (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

3 2

The internationalresponse: missinga human dimension

The threat that the financial crisis poses tointernationally agreed human development goalsis widely recognized. The G20 communiqué ofApril 2009 acknowledged the ‘human dimensions’of the threat in particularly forthright terms(Group of Twenty, 2009, para. 25):

We recognise that the current crisis has adisproportionate impact on the vulnerable in the poorest countries and recognise our collectiveresponsibility to mitigate the social impact of the crisis to minimise long-lasting damage toglobal potential.

Subsequent gatherings have reaffirmed theconcern. At the G8 summit in July in L’Aquila,Italy, the governments of the world’s richestnations declared that they remained focused onthe human and social consequences of the crisis.‘We are determined’, their communiqué declared,‘to undertake measures to mitigate the impact ofthe crisis on developing countries, and to continueto support their efforts to achieve the MillenniumDevelopment Goals’ (Group of Eight, 2009c, para. 6). To what extent have political leaders inthe countries that caused the crisis acted on their‘collective responsibility’ to mitigate its effects?

Financial resources have been made availableon a large scale, both domestically andinternationally. Advanced economies have spentaround US$10 trillion shoring up their financialsystems by providing capital, loan guarantees,lending and asset protection. That figurerepresents around 30% of their combined GDP.Under the G20 recovery plan, the IMF has beenused to strengthen global liquidity and bolsterfragile financial systems. This national andinternational response has been vital to stavingoff a far deeper global crisis and creating theconditions for recovery. After a severe globalrecession, economic growth has turned positiveas wide-ranging public finance interventions havesupported demand and reduced financial risk.Yet the report card on support for the poorestcountries is deeply unimpressive.

Headline figures on global financing have maskedthree problems. First, the poorest countries havebeen largely bypassed (Woods, 2009b). As thepresident of the African Development Bank put it,‘only a small proportion of the resourcesannounced at the G20 summit in London willtrickle down to low-income countries’ (Kaberuka,2009). Second, much of the support that doestrickle down will arrive too late and on termsthat are inappropriate for the financing needsof the poorest countries.

The third concern is that much of what has beenpresented as ‘new and additional’ finance is in factrepackaged or reprogrammed aid. This ‘smokeand mirrors’ financial reporting has obscuredthe collective failure of developed countries todecisively deliver resources on the required scale.Some new resources have been made available,principally through the IMF. In the case of theWorld Bank, which G8 and G20 rhetoric placesat the centre of the crisis response for the poorestnations, very few additional resources have beenmobilized (Woods, 2009b). Instead, the institutionhas been left to reconfigure its resources tomount a response.

Consolidating current financing for low-incomecountries is problematic because of uncertaintiesover commitments. On an optimistic assessment,new concessional financing potentially available tolow-income countries amounts to between aroundUS$2 billion and US$3 billion annually for the nexttwo to three years.8 That figure has to be setagainst the annual revenue loss of US$80 billionfor sub-Saharan Africa alone in 2009.

It is easy to lose sight of what is at stake for theinternational development goals in education.The everyday concerns of parents struggling tokeep their children in school in a slum in Lusakaor a poor village in Senegal seem far removedfrom the international summits on the globalfinancial crisis. Yet the connections are real. Asrich countries take the first steps towards economicrecovery, the aftershock of the crisis is jeopardizingthe efforts of the world’s poorest households tosecure for their children an education that mightlift them out of poverty. Containing the aftershockwill require a strengthened focus on financingfor human development.

8. In the period to 2010, IMF concessional loans could rise by up toUS$8 billion. The estimate for this Report adds US$2 billion for variouscommitments undertaken through bilateral aid programmes andWorld Bank trust funds, though this is almost certainly an overestimate.

Much of what has been

presented as‘new and

additional’finance is in fact

repackaged orreprogrammed

aid

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

T h e i n t e r n a t i o n a l r e s p o n s e : m i s s i n g a h u m a n d i m e n s i o n

3 3

The crisis response

The framework for the international responseto the financial crisis was set at the G20 summitin April 2009, with the ensuing G8 summitsupplementing the agreement. The recoverystrategy gave the IMF wide-ranging responsibilityfor strengthening global liquidity by expandingcurrency reserves to prevent further financialcrises and by providing concessional finance forlow-income countries. The World Bank was givenresponsibility for financing measures aimed atstrengthening social protection and tackling foodsupply problems.

The checklist of global financing commitmentsand provisions for the poorest countries isexpansive and superficially impressive. Muchof the new financing has come through the IMF:

Boosting global liquidity and strengtheningfinancial stability. Under the G20 plan the IMFhas injected US$283 billion into the globaleconomy in Special Drawing Rights (SDRs),currency reserves that can be exchanged forhard currency. New SDR allocations effectivelysupplement IMF members’ existing currencyreserves, thereby providing liquidity to theinternational economic system.9 The IMF’scredit lines for emerging markets have alsobeen reinforced through the creation of anew facility and the strengthening of existingfacilities.10

Scaling up concessional financing. Measureshave been introduced to increase the IMFresources available to low-income countriesthrough the fund’s Poverty Reduction andGrowth Facility (PRGF). The measures couldincrease concessional lending by US$17 billionthrough to 2014, with up to US$8 billion by2010. Several new financial instruments havebeen created to provide more concessionalsupport to low-income countries.11 In addition,the IMF has modified its Exogenous ShocksFacility (ESF), a mechanism aimed at providingsupport to countries facing exceptionalproblems as a result of conflict, natural

disaster, falling commodity prices or rising foodprices (Bredenkamp, 2009a, 2009b; IMF, 2009a,2009d; Woods, 2009b).12

The G20 meeting signalled a broad agenda forthe World Bank. It included what was termed ‘a substantial increase in lending of US$100 billion’and increased bilateral contributions for a rangeof crisis-response facilities aimed at strengtheningsocial protection and wider poverty interventions(Group of Twenty, 2009). These include the newInfrastructure Crisis Facility, VulnerabilityFramework and Rapid Social Response Fund.The World Bank was also made institutional leadactor in the response to the global food crisis.At the G8 summit, governments pledged toprovide US$20 billion over three years to supportcountries struggling with higher food import bills(Group of Eight, 2009b).

The IMF and World Bank facilities have attracteda great deal of media attention. An impressionhas been created that rich countries have movedrapidly to extend to the world’s poorest countriesthe same principles applied in their domesticresponses to the crisis. That impression owesless to real financial transfers than to somequestionable reporting practices.

Consider first the IMF component of the globalrecovery package. The initial expansion of post-crisis lending bypassed the poorest countries,principally because it was directed towardsfinancial stabilization in Europe and someemerging markets. Of the eighteen new lendingagreements the IMF had approved by late July2009, 82% were directed to Europe and 1.6% toAfrica (Woods, 2009b). While low-income countrieswill have their currency reserves boosted by thenew SDR issue, the allocations are linked to thesize of national economies (the increasedallocation for France exceeds that for all of sub-Saharan Africa). Moreover, an expansion of the national currency reserve does notautomatically generate additional resourcesfor high-priority budgets.

What of the increase in concessional lendingthrough the IMF? As of October 2009, this was theonly source of new and additional financing linkeddirectly to the global financial crisis. The IMFclaims the new arrangements enable it to make up

12. Much of the additional IMF support to low-income countries in 2009came through the Exogenous Shocks Facility, whose financing termsare equivalent to those of the PRGF.

9. Low-income countries will receive an additional US$17 billion in SDRs (Gottselig, 2009).

10. In April 2009, the IMF announced the creation of a new flexiblecredit line and increased flexibility for its standard stand-byarrangements.

11. These are the Extended Credit Facility (medium-term support),the Standby Credit Facility (short-term and precautionary support)and the Rapid Credit Facility (emergency support).

The initialexpansion of post-crisislending bypassedthe poorestcountries

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

3 4

to US$8 billion available in 2009 and 2010, thoughone-quarter of that figure is accounted for byearly disbursement of existing loans. The G20framework makes about US$6 billion in newconcessional lending resources available to theIMF over 2009–2012 (around US$2 billion annually)for all low-income countries. The IMF itselfestimates that the increased lending capacity willcover only 2% of low-income countries’ externalfinancing needs (IMF, 2009d; Woods, 2009b). Actualtransfers of new financing will be contingent on therate of disbursement. Given that disbursementsthrough the PRGF are often disrupted becausecountries cannot comply with loan conditions,there are serious questions over the prospectsfor timely delivery.

The World Bank’s role in the internationalresponse to the crisis is characterized by alarge gap between words and money. Manycommitments in the G20 communiqué, notablythose directed to low-income countries, representnot new money but an imaginative ‘relaunch’ ofpast pledges.13 Others effectively exempt the G20countries from providing new and additionalfinancing, with bold language on scaling up socialprotection backed only by a vague pledge of‘voluntary bilateral contributions’.

The World Bank has been left to act on the G20agenda mainly by drawing upon its own resourcesand facilities. While strong pronouncements havebeen made declaring that World Bank supportto crisis-affected countries is at a ‘record high’,increased lending has been sustained not byhigher donor support, but by a combination ofearly disbursement of funds – front-loading –and reprogramming.

The Global Food Crisis Response Programme(GFRP) is a case in point. After eighteen monthsthe programme had disbursed US$795 million,or 68% of its original funds – far too slow a pacegiven the immediacy of the crisis (United NationsConference on Trade and Development, 2009).Interventions have ranged from support to schoolfeeding programmes in Burundi, Liberia andSenegal to safety-net programmes in Ethiopia,the United Republic of Tanzania and Yemen, andbudget support in Bangladesh, Cambodia andHonduras. These programmes provide vital socialprotection, but the bulk of GFRP finance comesnot from increased aid but from existing countryallocations, regional International DevelopmentAssociation (IDA) funds and resources transferred

from other facilities (Delgado, 2008). The only newsource of finance has been a multidonor trust fundthat channelled US$200 million to the GFRP. Mostof the US$20 billion pledged at the G8 summit forfood supplies also involves the diversion of existingaid commitments rather than new money.

Some World Bank programmes appear not to havetaken off on any scale. The Rapid Social ResponseFund was created to assist poor and vulnerablepopulations in developing countries, mainly fromthe World Bank’s own resources. As of September2009, only one programme appears to have beenapproved – a cash transfer and nutritionintervention for children under 5 in Senegal(World Bank, 2009i).

Other programmes have generated large headlinenumbers under the banner of ‘crisis response’ withlittle in the way of new financing. In 2009, the WorldBank significantly increased financing provisionsfor countries affected by the crisis. Commitmentsunder IDA reached US$14 billion in 2009 and a newUS$2 billion facility was created to provide earlysupport in key areas of social protection, health andeducation. Almost half the allocations available hadbeen disbursed by late 2009 (World Bank, 2009d).However, most of the new financing came fromfront-loading of IDA allocations for low-incomecountries (Figure 1.5). Burkina Faso, Liberia andSenegal, among others, received over 150% oftheir planned IDA allocations in 2009.

As a crisis response measure, front-loadingmakes sense. Faced with mounting budgetpressure and rising poverty, countries need earlyaid. For households confronting hunger, healthrisks and the challenge of keeping children inschool, delays in social protection carry a highprice. But front-loading does not increase theoverall resources available to governments overthe full cycle of programme support. Moreover,it comes with its own risks, including the riskof financing deficits in later years.14

The upshot is that the World Bank has beeninvolved in an elaborate financial reshuffle. Effortsby the institution itself to address the issue ofmaking new resources available have not been

13. One example is thepledge of US$100 billionin additional multilaterallending, originally madeseveral months before theG20 summit, with India,Indonesia and Ukraineidentified as being amongthe potential beneficiaries.

14. The World Bank is not alone in combating the human developmentemergency through creative accounting. Plans drawn up by the EU Commission in May 2009 announced an intention to mobilize8.8 billion euros (approximately US$12 billion at May 2009 exchangerates) in development financing as a crisis response, but almost all thecommitments and pledges behind this figure come from pre-existingcommitments (Woods, 2009b).

Bold language on scaling up

social protectionhas been backedonly by a vague

pledge of‘voluntary

bilateralcontributions’

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

T h e i n t e r n a t i o n a l r e s p o n s e : m i s s i n g a h u m a n d i m e n s i o n

3 5

wholly successful. Before the G20 summit,World Bank President Robert Zoellick called ondeveloped countries to put aside the equivalentof 0.7% of their stimulus package for a newVulnerability Fund (World Bank, 2009l). This wasan innovative attempt to create a financing basefor new and additional aid to countries lackingthe fiscal space to respond to the crisis, enablingthem to create the conditions for recovery andstrengthen social protection. Mr Zoellick notedthat the real issue at stake was a choice betweenan ‘age of responsibility or an age of reversal’(Zoellick, 2009). That formulation captures theoptions rich countries face with respect to theinternational development goals in educationand other areas. Evidence to date suggeststhat the ‘age of reversal’ is the default choice.

There are wider problems in the G20 response tothe crisis related to the respective roles of the IMFand World Bank. The latter would have been theobvious institution to lead the response to thespecial challenges facing low-income countries.It has a far stronger capacity than the IMF forrapid assessment of the budgetary implicationsof the economic downturn on financing for theMillennium Development Goals. It has also playeda leading role in supporting and developing socialprotection programmes. Moreover, theInternational Development Association, the World

Bank’s main source of financing for low-incomecountries, provides loans on more concessionalterms than the IMF’s Poverty Reduction and GrowthFacility.15 For all of these reasons, the World Bankand the IDA should have been the first line ofdefence in the response to the crisis.

The IMF’s track record in poverty reduction effortshas prompted further questions about its enhancedrole. In 2004, the IMF Independent Evaluation Officeconcluded: ‘Success in embedding the PRGF in theoverall strategy for growth and poverty reductionhas been limited in most cases’ (IMF, 2004).Several commentators have identified an inflexibleapproach to targets, enshrined in loan conditionsfor inflation, fiscal deficits and public spending,as a source of tension between the IMF approachto macroeconomic stabilization and the financingstrategies aimed at achieving the MillenniumDevelopment Goals. That tension has been evidentin debates over financing for education. Forexample, the Global Campaign for Educationconcludes a review of twenty-three IMFprogrammes by warning of a potential conflictbetween spending targets set in loan conditionsand financing requirements for teacher recruitment(Global Campaign for Education, 2009).

In the wake of the financial crisis, the IMF’s seniormanagement has pledged to adopt more flexibleapproaches to fiscal deficits and inflation (IMF,2009d; Sayeh, 2009). This is vital, because fiscalpolicy should counteract the crisis, not createdeflationary pressures. There is some evidenceof greater flexibility being applied at the countrylevel in sub-Saharan Africa. Even before the crisis,inflation targets had been loosened to reflect theimpact of higher food prices. In mid-2009, the IMFreported that fiscal targets had been relaxed ineighteen of the twenty-three countries with activeprogrammes (IMF, 2009d).16 However, questionsremain over the degree to which the recentdeclarations reflect a new approach tomacroeconomic management. Loan conditions inseveral countries examined in the United Nations’2009 Trade and Development Report – includingCôte d’Ivoire, Ethiopia, Pakistan and Senegal –

15. PRGF loans are provided at 0.5% interest and are repayable overten years with a five-year grace period. The PRGF has a grant elementof around 30%. IDA provides interest-free credits repayable over thirty-five to forty years with a ten-year grace period. The grantcomponent of IDA is roughly double that of the PRGF.

16. A preliminary review of programmes for thirty-three low-incomecountries indicates that the deficit is being allowed to widen in aroundtwenty cases (though in some instances just for 2009) and is stayingthe same or falling in the other countries (Martin and Kyrili, 2009).

The World Bankand the IDA shouldhave been the firstline of defence in the response to the crisis

Figure 1.5: The World Bank has front-loaded concessionalInternational Development Association loansEarly disbursement as a share of planned allocation under the International Development Association (IDA)

Source: World Bank (2009b).

LiberiaCôte d’IvoireBurkina Faso

SenegalEthiopia

BeninGhana

MaliC. A. R

TogoNigeria

Guinea-BissauBurundiLesothoGambia

KenyaNiger

% front-loaded

100 110 120 130 140 150 160 170 180

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 1

3 6

include the tightening of fiscal and monetary policy.The authors conclude: ‘Policy conditions attachedto these IMF loans are fairly similar to those ofthe past, including a requirement that recipientcountries reduce public spending and increaseinterest rates’ (United Nations Conference onTrade and Development, 2009). This would appearto be inconsistent both with the IMF’s policypronouncements and – more importantly – withthe need to avoid deflationary measures in theinterests of economic recovery and long-termpoverty reduction.

Looking ahead

The United Nations Secretary-General has warnedin stark terms that the financial crisis has thepotential to mutate into a long-term developmentemergency. ‘If we do not act together, if we do notact responsibly, if we do not act now,’ he said inMay 2009, ‘we risk slipping into a cycle of poverty,degradation and despair’ (United Nations, 2009b).The danger is that as the world economy pulls outof recession, the real victims of the crisis will beforgotten, including millions of children facing theprospect of losing their chance for an education.

The most immediate priority is for rich countriesto respond to the mounting budget pressure facinggovernments in low-income countries. That meansproviding more concessional financing beforeirreparable damage is inflicted on vital socialinfrastructure. While leaders of the G20 and the G8have adopted encouraging communiqués, deliveryhas been woefully inadequate. Behind the globalfinancial pledges, the world’s most vulnerablecitizens have been left to sink or swim with theirown resources. As social and economic pressuresmount, there is an imminent threat that progressin education will stall, damaging prospects foreconomic growth, poverty reduction and health.Political leaders in rich countries need to respondto the human crisis in poor countries with thesame level of resolve they have demonstratedin their domestic responses to the crisis.

Action is required at many levels. The followingare among the most urgent priorities:

Convene a high-level meeting on Educationfor All financing before the 2010 MillenniumDevelopment Goals summit. Financing gaps forachieving the 2015 Education for All goals havebeen systematically underestimated. Evidenceset out in this Report (see Chapter 2) suggeststhat the average annual shortfall in financingis around US$16 billion, rather than theUS$11 billion previously assumed. With slowereconomic growth in the poorest countries,prospects for closing this gap are deteriorating.Given the scale of the financing gap and thefailure of rich countries to support social andeconomic recovery in the poorest countries,the United Nations Secretary-General shouldconvene a high-level meeting to elaboratestrategies for making more resources availablebefore the Millennium Development Goalssummit in September 2010.

Scale up aid and provide early support.If developing countries are to protect andstrengthen public financing commitments inthe face of an economic downturn, they needa sustained and predictable increase in aid andup-front support to counteract revenue lossesfrom 2008 and 2009. The financial crisis hasadded to the urgency of rich countries actingon the aid commitments made in 2005 (seeChapter 4). Increased official developmentassistance should be backed by a temporarydebt moratorium for low-income countries for2009 and 2010, with the savings released forspending in key areas. Such a moratorium wouldcost around US$26 billion in total (United NationsConference on Trade and Development, 2009).

Make monitoring more effective. Waiting untilthe education crisis announces itself in officialdata is not a sensible course of action. Crisisprevention – which is eminently preferable toresponse after the event – requires far more

The UnitedNations

Secretary-General should

convene a high-levelmeeting on

EFA financing

E D U C AT I O N AT R I S K : T H E I M PA C T O F T H E F I N A N C I A L C R I S I S

T h e i n t e r n a t i o n a l r e s p o n s e : m i s s i n g a h u m a n d i m e n s i o n

3 7

effective and current monitoring of governmentbudgets, school attendance and dropout rates.UNESCO should take the lead in this area,working through national education and financeministries and coordinating a wider donorresponse. It is particularly important that theimplementation of 2009 budgets, real educationspending and 2010 budgets are subject to closescrutiny. Beyond outright cuts in public spending,monitoring should focus on disparities betweenplanned spending in education sector strategiesand actual spending.

Ensure that IMF support is provided on aflexible basis that is consistent with achievingthe Education for All goals. Statements by theIMF leadership pointing to greater flexibility inloan conditions on fiscal deficits, inflation andpublic spending are welcome, but concernsremain over whether this flexibility will bemaintained in 2010 and beyond. In drawing uploan conditions, IMF staff should be required toreport explicitly on consistency with the financingrequirements for achieving the core Educationfor All goals by 2015. Special priority should beput on the costs associated with teacherrecruitment, training and remuneration.

Increase support through the InternationalDevelopment Association. IDA is the mostappropriate multilateral financing vehicle formitigating the effects of the economic downturnin the poorest countries. While the World Bankhas demonstrated a capacity for innovation infront-loading IDA financing, transferringresources from other facilities and redirectingexisting country allocations, this approach isneither a sustainable nor a credible responseto a systemic crisis in financing for internationaldevelopment goals. Front-loading also raisesuncertainty over future financing for educationand other high-priority sectors. To guard againstthis uncertainty and place IDA financing on amore balanced footing, donors should undertakea binding commitment to increase the resources

available during the next replenishment. Donorsshould pledge to increase their support for WorldBank concessional loans from US$42 billion inthe fifteenth IDA replenishment to US$60 billionin IDA-16, which begins in 2010.

Make social protection a high priority.Protecting education budgets is just one of therequirements for sustained progress towards keyEducation for All goals. Rising household povertylinked to the economic crisis brings with it theprospect of increased child labour, deterioratingnutrition and reduced capacity for investmentin education. Social protection, through cashtransfers, nutrition programmes and targetedsupport in other areas, has been shown in manycountries to build the resilience of vulnerablehouseholds and strengthen their ability to copewith economic shocks without resorting todamaging measures such as withdrawingchildren from school. As Chapter 3 shows,government and donor support can makea huge difference in this area.

UNESCO shouldtake the lead in the monitoringof governmentbudgets, school attendanceand dropout rates

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

3 8

Chapter 2. Progress

Overcrowded and under-resourced:a classroom in Malawi

Keeping malnutritionin check: an Ethiopianinfant gets weighed

Continuing educationin Lebanon: learningempowers at all ages

© L

ouis

e Gu

bb/C

orbi

s©

Alfr

edo

Caliz

/PAN

OS

© R

EUTE

RS/

Zohr

a B

ense

mra

0102Education for All Global Monitoring Report

3 9

towards the EFA goals

Getting there in ruralChina: minority children

have further to go

In the Hindu Kush, Pakistan, a traditional Kalash storyteller

captivates his audience

© C

hret

ien

Eric

/Gam

ma/

Eyed

ea P

ress

e

© P

atri

ck L

e Fl

och/

Expl

orer

/Eye

dea

Illus

trat

ion

Monitoring national trends is the core task of the EFA GlobalMonitoring Report. This chapterexamines progress towards the goals adopted at Dakar.It documents differences acrosscountries and regions, as well asthe sometimes dramaticdifferences that coexist within asingle border. It highlights the roleof persistent inequalities in holdingback advances towards all of theEFA goals and considers the degreeto which governments and aiddonors are meeting – or fallingshort – of their promises to investin basic education. A costingexercise assesses the financing gapfor forty-six low-income countries,putting a price tag on the costof delivering a quality education to every child, youth and adult.

Introduction .................................................................. 41

Early childhood care and education ............................................................ 42

Universal primary education .............. 54

Youth and adult skills – expanding opportunities in the new global economy ................... 76

Youth and adult literacy ............................. 94

The quality of education ............................ 104

Estimating the cost of achieving Education for All ................... 119

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

4 0

P R O G R E S S T O WA R D S T H E E FA G O A L S

I n t r o d u c t i o n

4 1

Introduction

The Dakar Framework for Action, adopted by164 governments in 2000, is one of the mostcomprehensive, wide-ranging and ambitious ofall commitments undertaken by the internationalcommunity. It pledges to expand learningopportunities for every youth, adult and child –and to achieve specific targets in key areas by 2015.

With just five years to go to the target date, thischapter of the Report monitors progress towardsthe Education for All goals set under the DakarFramework. Effective monitoring is vital toachieving international development targets. Apartfrom keeping the goals themselves in the spotlight,it can highlight examples of success, provide earlywarning of failure, inform policy and supportadvocacy. At an international level, measuringprogress towards shared goals gives insight intothe strengths and weaknesses of nationalstrategies. It demonstrates what can be achievedin practice and shows that all countries, whatevertheir level of development, can make progressunder strong political leadership. Above all,monitoring provides a tool to hold governmentsaccountable for the degree to which they act onthe commitments they undertake at internationalsummit meetings.

This chapter starts by highlighting the importanceof early childhood care and education in creatingthe foundations for lifelong learning. It thenmonitors progress towards universal primaryeducation, an area which raises serious concerns.While the number of children not attending schoolcontinues to fall, the most recent school enrolmentdata suggest that the goal of universal primaryeducation by 2015 will be missed. Moreover,household survey evidence suggests that morechildren may be out of school than the officialdata indicate. Progress in school participationcontinues to outstrip progress in learningachievement, pointing to a widening gap betweenquantitative and qualitative indicators of progress.The chapter further shows that advances in adultliteracy fall far short of the goals. Meanwhile,technical and vocational education programmeshave – at best – a mixed record in respondingto the learning needs of youth and young adults.

Financing is critical to accelerating progresstowards the Education for All goals. Current globalestimates of the financing required to meet the2015 targets are outdated and methodologicallyflawed, primarily because they do not take intoaccount the cost of reaching disadvantaged groups.This Report provides an updated analysis of thefinancing needed to reach key targets, adjusted forthe incremental cost of extending opportunities todisadvantaged groups. The analysis shows thatfinancing gaps have been underestimated and thatdeveloping country governments and aid donors willhave to act with urgency to close these gaps.

Monitoring is about more than technicalmeasurement of progress. This chapter highlightsthe strong connections between progress towardsspecific goals and underlying problems ofinequality and social marginalization, whichChapter 3 examines further. While national datahelp illuminate broad trends, they can obscureunderlying disparities. Gender disparities arenarrowing in many areas, but young girls andwomen continue to face disadvantages at severallevels, from early childhood, through primary andsecondary school, and into adulthood. Widerinequalities linked to poverty, language, ethnicity,region and other factors also restrict opportunityon a global scale.

Global monitoring exercises inevitably revealcomplex and varied patterns. Two broad messagesemerge from the detail of national and regionalprogress reports on the Education for All goals.The first is good news: there is unequivocalevidence that the world is moving in the rightdirection, with many of the poorest countriesregistering impressive advances on many fronts.Their record demonstrates what is achievable – and shows that many of the 2015 goals are stillwithin reach. The second message is cautionary,with a ‘bad news’ element: on current trends,progress towards the Dakar goals is far too slowto meet the 2015 targets. An underlying problemis the failure of many governments to put higherpriority on policies that extend opportunitiesto the most marginalized sections of society.Failure to change this picture will result inthe international community falling far shortof the promise made at Dakar in 2000.

Progress towardsthe Dakar goals is far too slow to meet the 2015targets

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

4 2

Early childhood careand education

Goal 1: Expanding and improving comprehensiveearly childhood care and education, especially forthe most vulnerable and disadvantaged children.

‘Five- and six-year-old children are the inheritors of poverty’s curse and not its creators. Unless weact these children will pass it onto the nextgeneration like a family birthmark.’

These remarks by United States President LyndonB. Johnson (1965) retain a powerful resonance.Early childhood can create the foundations for a lifeof expanded opportunity – or it can lock children intoa future of deprivation and marginalization. There isstrong – and growing – evidence that high-qualitycare in the early years can act as a springboardfor success in school. In turn, education providesvulnerable and disadvantaged children with achance to escape poverty, build a more securefuture and realize their potential. The past decadehas witnessed rapid and sustained increases in thenumber of children entering primary school in theworld’s poorest countries. Yet every year millionsof children start school carrying the handicapthat comes with the experience of malnutrition,ill health and poverty in their early years.

For many that experience starts, quite literally,in the womb and continues through the early years.Maternal undernutrition and the failure of healthsystems to provide effective antenatal support,along with safe delivery and post-natal care,contribute to child mortality. They also help transmiteducational disadvantage across generations.Malnutrition before children enter school is anotherformidable barrier to education. Apart fromthreatening lives, it robs children of the opportunityto develop their potential for learning. That is whythe eradication of child malnutrition should beviewed not just as a development imperative inits own right, but as a key element in the Educationfor All agenda.

With some notable exceptions, governmentsacross the world have failed to accelerate progressin combating child hunger. And with the sharphike in global food prices during 2007 and 2008,and the economic downturn pushing more childreninto poverty, a picture that was already bleak hasbeen deteriorating.

Effective early childhood care and education cangive children a better chance of escaping whatPresident Johnson called ‘poverty’s curse’. Whilemuch has been achieved, the monitoring evidenceset out in this section suggests that far more has tobe done. To summarize some of the key messages:

Malnutrition needs to be recognized as both ahealth and an education emergency. Malnutritionis damaging the bodies and minds of around178 million young children each year,undermining their potential for learning,reinforcing inequality in education and beyond,and reducing the efficiency of investment inschool systems.

Improved access to maternal and child healthcare should be seen as a high priority foreducation, as well as for public health. Chargingfees for basic services is locking millions ofvulnerable women out of health systems andexposing their children to unnecessary risks.As in the education sector, the elimination ofuser fees should be treated as a high priority.

Governments need to tackle inequality in accessto early childhood care. Those in greatest need ofearly childhood care – and with the most to gainfrom it – have the least access. In both rich andpoor countries, parental income and educationheavily influence who attends pre-schoolprogrammes, pointing to the need for greaterequity in public provision and financing.

This section is divided into two parts. Part 1reviews progress and the current status of somekey indicators of child well-being and nutritionin developing countries, and highlights the stronglinks between maternal and child health. Part 2provides a snapshot of access to early childhoodprogrammes across the world, along withevidence that these programmes can play animportant role in equalizing opportunity andovercoming marginalization.

Malnutrition and ill health — a ‘silent emergency’ in education

Retarded growth in the womb, early-childhoodstunting and anaemia are not typically viewedas mainstream education issues. The evidencestrongly suggests that they should be. Each ofthese conditions can have profound andirreversible effects on a child’s ability to learn,undermining the potential benefits of education.

High-quality carein the early years

can act as a springboard forsuccess in school

P R O G R E S S T O WA R D S T H E E FA G O A L S

E a r l y c h i l d h o o d c a r e a n d e d u c a t i o n

4 3

Neurological science helps explain why educationprospects are shaped in utero and during theearly years of life. The period from about threemonths before birth is critical to the formationof neural pathways, while the first three yearsare marked by rapid development of languageand memory (Bennett, 2008). Normal braindevelopment during this period creates afoundation for future school achievement andlifelong learning (Harvard University Center onthe Developing Child, 2007).

Children who suffer nutritional deprivation in uteroor during their early years pay a high price laterin life. There is a powerful and growing body ofevidence that nutritional status during the firsttwo years of life strongly determines laterperformance in education (Alderman et al., 2001;Glewwe et al., 2001; Grantham-McGregor et al.,2007). Children who experience episodes of earlymalnutrition tend to score worse on tests ofcognitive function, psychomotor development, finemotor skills, activity levels and attention span(Alderman et al., 2006; Behrman, 1996; Maluccioet al., 2009). They also tend to start school laterand are at greater risk of dropping out beforecompleting a full primary school cycle. A recentstudy in Guatemala finds that the impact of beingstunted at age 6 is equivalent in its test scoreeffects to losing four grades of schooling(Behrman et al., 2008). The critical but widelyignored insight to emerge from the researchevidence is that what children are able to learnin school is heavily influenced by pre-schoolhealth and nutrition.

Research carried out for this year’s EFA GlobalMonitoring Report adds further weight to evidenceof the long-term impact of nutrition on cognitivedevelopment (Box 2.1). Drawing on the YoungLives Survey, a unique data set that trackschildren in Ethiopia, India, Peru and Viet Namthrough their early years, the analysis documentsmarked nutrition-related disadvantages revealedin test scores at age 4 to 5. By age 7 to 8, themalnutrition penalty is equivalent to the lossof a full term of schooling (Sanchez, 2009).

Child malnutrition — limited progressOpportunities for education are heavily influencedby the well-being of children before they enterschool. It is an unfortunate fact that, at the startof the twenty-first century, the twin scourges ofhunger and ill health continue to blight educationon a global scale.

One way to gauge how children are faring aroundthe world is to look at child mortality rates. Whiledeath rates are falling, the world remains far offtrack for the Millennium Development Goal of atwo-thirds reduction from 1990 levels by 2015.There were 9.3 million child deaths in 2008. Oncurrent trends the millennium goal target will bemissed by a figure equivalent to more than 4 millionadditional deaths in 2015. Set against this badnews is the fact that many of the world’s poorestcountries, including Ethiopia, Malawi, Mozambiqueand the United Republic of Tanzania, have cut childdeaths by 40% or more (UNICEF, 2008b).

Child mortality is intimately related to malnutrition.Progress towards the Millennium DevelopmentGoal target of halving malnutrition has beenpainfully slow, with most countries in South Asiaand sub-Saharan Africa off track. It is estimatedthat malnutrition is directly implicated in two of

Research carried out for this year’s Report strongly reinforces widerevidence on the contribution of malnutrition to educational disadvantage.

Using data from the Young Lives Survey, which tracked children in Ethiopia,India (Andhra Pradesh state), Peru and Viet Nam, the study examines therelationship between early nutrition and cognitive achievements at age 4 to5, measured on the Peabody Picture Vocabulary Test (PPVT), an internationallearning achievement scale. It also looks at the relationship between nutritionat age 7 to 8 and outcomes measured in terms of PPVT scores andaccumulated years of education at age 11 to 12 (for this cohort, Dercon[2008] presents similar evidence). In both cases, height for age is used as an indicator of nutritional history and status, standardized using the latestWorld Health Organization (WHO) growth curve references. Although thesamples are not nationally representative, they were designed to reflectcultural, ethnic and geographic differences within each country.

The results are striking. After controlling for an extensive set of child,parental and household characteristics, and taking into account the effect of community characteristics, the results point to a strong associationbetween nutritional status measured at 6 to 18 months and cognitiveachievement at age 4 to 5. An increase of one standard deviation in earlyheight for age is associated with an improvement of 4% to 12% of the PPVTstandard deviation in the Young Lives samples.

Similar findings emerge for the older cohort. In this case, an increase of one standard deviation in nutritional status measured at ages 7 and 8 is associated with a marked increase in school grade attainment thatrepresents 14% to 20% of the grade attainment standard deviation (about 0.2 to 0.4 additional years of schooling). Given the high levels of stunting for both cohorts in all the Young Lives samples, the resultsunderline the significant costs imposed by malnutrition on education.

Source: Sanchez (2009).

Box 2.1: Early malnutrition leads to long-term educational damage

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

4 4

every three deaths of children under age 5. Whilethere have been some advances towards improvedchild nutrition, and expanded access to Vitamin Asupplements and iodized salt, achievements fallfar short of the goals that have been set:

Childhood stunting.1 Around one in threechildren under age 5 – 178 million in total –suffers severe or moderate stunting. By the timethese children enter school, malnutrition willhave diminished their potential for learning – a disadvantage they will carry into adulthood.Apart from its damaging consequences forindividuals, malnutrition in early childhoodinevitably erodes the benefits of investment ineducation. The highest regional rates of stuntingare found in central and eastern Africa andSouth Asia. Of the forty-nine countries wherestunting prevalence rates are in excess of 30%,thirty are in sub-Saharan Africa (Figure 2.1).

Low birth weight. Recent international estimatessuggest that about 19 million infants – 14% of allnewborns – are delivered with low birth weight(UNICEF, 2008b). More than half of these birthstake place in South Asia: over one in four of theregion’s children are delivered with low birthweight (Figure 2.2). These children face aheightened risk of early mortality: low birthweight is an underlying factor in 60% to 80%of deaths in the first month. They also facelonger-term risks of disadvantage in healthand education. Low birth weight is stronglyassociated with loss of years in school andpoorer cognitive skills (Victoria et al., 2008),which undermine the potential benefits ofimproved access to secondary education. Manyof the 8.3 million Indian children born with lowbirth weight will carry a burden of disadvantagewith them into primary school. Moreover,almost half of all children under age 3 in Indiaare underweight for their age, pointing to fardeeper nutritional deficits.

Micronutrient deficiency. Early cognitivedevelopment can be severely impaired bymicronutrient deficiencies. It is estimated thatone-third of all pre-school children is affectedby iodine deficiency, a condition associated witha loss of ten to fifteen points on IQ tests evenin moderate forms. A similar proportion ofchildren is affected by Vitamin A deficiency,a major cause of blindness, ill health and poorconcentration (Victoria et al., 2008).

The factors behind malnutrition vary acrosscountries. Poverty, social inequalities and livelihoodinsecurity all play a role. National wealth is often a

1. Stunting, or low height for age, is caused by long-term insufficiency ofnutrient intake and frequent infections. It generally occurs before the ageof 2 and the effects are largely irreversible.

Sub-Saharan Africa

Arab States

East Asia and the Pacific

South and West Asia

Latin America and the Caribbean

CameroonKenya

UgandaCôte d’Ivoire

MaliBurkina Faso

GuineaBenin

C. A. R.D. R. Congo

EritreaLesothoNigeria

SomaliaU. R. TanzaniaEquat. Guinea

LiberiaZambia

Sierra LeoneChad

Guinea-BissauMozambique

ComorosAngola

RwandaMalawiEthiopia

MadagascarNiger

Burundi

DjiboutiMauritania

SudanYemen

PhilippinesMyanmarViet NamCambodia

DPR KoreaLao PDR

Timor-Leste

BangladeshPakistan

IndiaBhutanNepal

Afghanistan

PeruGuatemala

0 10 20 30 40 50 60

Severe and moderate stunting (%)

Figure 2.1: High levels of child stunting are holding back progress in educationSevere and moderate stunting among children under 5, selected countries, 2000–2007 1

Note: Countries included are those in which the proportion of stunted children is 30% or more.1. Data are for the most recent year available during the period specified.Source: Annex, Statistical Table 3A.

P R O G R E S S T O WA R D S T H E E FA G O A L S

E a r l y c h i l d h o o d c a r e a n d e d u c a t i o n

4 5

poor guide to deprivation. Guatemala is not oneof the world’s poorest countries, but it has oneof the highest levels of child stunting. Almost halfof the country’s children are malnourished – andin parts of rural Guatemala, where the populationis largely Mayan, the figure reaches 80%. Overthe past two years, drought and high food priceshave made things worse. But the underlyingproblem is extreme inequality in wealthdistribution, allied to the failure of governmentto mobilize resources for social protection.

Nutritional indicators have been deterioratingin many countries over the past two years.World agricultural prices rose sharply in thetwo years to 2008, affecting all major tradedfood staples. While prices have since fallen, theyhave stabilized at levels far higher than they werebefore 2007. Effects at the national level havevaried considerably, depending on the incidenceof poverty and dependence on food imports.However, higher food prices have almost certainlystalled global progress in cutting malnutrition.

Recent estimates from the Food and AgriculturalOrganization of the United Nations suggest thatthe number of malnourished people in the worldincreased from 848 million in 2005 to 963 millionin 2008, largely because of rising food prices(FAO, 2008). Another 44 million people mayhave been pushed into malnutrition during 2008(Commission on Growth and Development, 2008).

The damage inflicted by higher food prices hasbeen unevenly spread. Outcomes depend onwhether households are net sellers or buyersof food, on access to savings or credit and oncurrent nutritional status. For people living belowthe international poverty threshold of $1.25 a day,many of whom spend 50% to 70% of their incomeon food, higher food prices pose a stark choice:eat less or decrease spending in other areas(von Braun, 2008; World Bank, 2008a). Landlessrural households, low-income urban groups andfemale-headed households have been amongthe hardest hit. Many have cut already inadequatediets and switched from protein-rich foods tocheaper coarse cereals (Hauenstein Swan et al.,2009; von Braun, 2008). In Bangladesh, whererice and wheat prices almost doubled in 2007,it is estimated that a 50% increase in the priceof food staples increases the prevalence of irondeficiency among women and children by 25%(Bouis, 2008).

Short-term distress in the form of rising malnutritionwill have long-term consequences for education.As more children experience episodes ofmalnutrition in early childhood their prospects forlearning will be diminished. At the same time, risingpressure on household budgets will have widerconsequences as poor parents are forced to adjusthousehold budgets. There is evidence fromBangladesh, Jamaica and Kenya of householdscutting education spending to accommodate higherfood prices (Hossain et al., 2009; World Bank, 2008e).

High food prices have not been the only causeof rising malnutrition. In northern Sri Lanka,300,000 people were displaced by conflict in 2009.It is estimated that about 13% of the displacedwere children under 5. A survey covering six ofthe thirteen camps for displaced people foundthat one in four children was malnourished andone in three was moderately or severely stunted(Jayatissa, 2009). Failure to adequately protectthese children raises wider issues of humanitarianconcern. But the consequences for education willalso be severe.

Maternal health — critical, but neglectedThe health of newborn children – critical for latereducational chances – is intimately related tothe health of their mothers. Women who aremalnourished and suffering from micronutrientdeficiency face far higher risks during pregnancyand childbirth, and are more likely to give birth tounderweight babies. Restricted growth of the foetusduring pregnancy is a major risk factor for maternalhealth and child survival – and is likely to lead tofuture educational disadvantage.

Short-termdistress in theform of risingmalnutrition willhave long-termconsequences for education

Figure 2.2: Low birth weight sets the scene for lifelong disadvantageAverage % of infants with low birth weight, selected regions, 2000–2007 1

Notes: Regions presented are those used by UNICEF, which differ to some extent from the EFA regions. Low birth weight is defined as less than 2.5 kilograms.1. Data are for the most recent year available during the period specified.Source: UNICEF (2008b).

East Asia/Pacific

Latin America/Caribbean

Middle East/North Africa

Sub-SaharanAfrica

SouthAsia

0

5

10

15

20

25

30

% o

f inf

ants

with

low

birt

h w

eigh

t

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

4 6

Unsafe pregnancy and childbirth exact an immensehuman toll. An estimated half a million women losetheir lives each year from pregnancy and birth-related causes – and for every death another thirtywomen suffer severe long-term injuries. Almost allthese deaths and injuries could be averted throughaccess to antenatal care, skilled attendance duringpregnancy and emergency obstetric care. Poormaternal health, inadequate nutrition and limitedaccess to care are also implicated in the deaths ofthe 4 million newborns who do not survive their firstmonth (Lawn et al., 2006). Two conditions – birthasphyxia and sepsis with pneumonia – cause nearly60% of these deaths. The real cause, however, islimited access to skilled health professionals at birthand a failure to prioritize maternal and child healthin national policy (Thea and Qazi, 2008).

This ‘needless human tragedy’ (UNICEF, 2008b)goes beyond maternal and child mortality andimmediate health risks. Undernutrition in utero, lowbirth weight and heightened vulnerability to sicknessafter birth can cause direct structural damage to thebrain that impairs cognitive development and lockschildren into a future of underachievement. Widerhealth risks during pregnancy and childbirth alsohave consequences for education:

Maternal iodine deficiency in pregnancy causesan estimated 38 million children to be borneach year facing risks of mental impairmentand congenital abnormalities (UNICEF, 2007b).

Anaemia, which affects around half of allpregnant women, heightens the risks associatedwith pregnancy and reduces prospects for childsurvival (UNICEF, 2008b).

Around half of the stunting observed in infantsoccurs in the uterus and the remainder duringthe first two years of life (Victoria et al., 2008).

The absence of skilled health personnel duringdelivery costs lives and leaves children facinglifetime disadvantages. Asphyxia contributesto around one-quarter of newborn deathsand results in about 1 million children sufferinglearning difficulties and disabilities such ascerebral palsy (WHO, 2005).

Access to health provision is not the only barrier toimproved child and maternal care. Many underlyingproblems associated with pregnancy and childbirthreflect a failure to protect women’s rights. Lowstatus, heavy workloads, a lack of voice in matters

of sexual and reproductive health, early marriageand poor access to information all contribute.

Providing quality health careInadequate maternal and child health care isholding back advances in education. Progresstowards the Millennium Development Goal targetof a three-quarters reduction in maternal deathsby 2015 has been close to zero. Meanwhile, limitedimprovements in survival in the first month of lifeare preventing progress towards the target onchild mortality.

One of the most urgent priorities is providingquality health services. Intrauterine growthrestrictions and maternal micronutrient deficienciescan be readily detected through antenatal care andtreated at little cost. Access to facilities providingskilled attendance at birth, emergency obstetriccare and post-natal care could prevent over 80%of maternal and neonatal deaths, and set childrenon course for a healthy future (UNICEF, 2008b).Yet more than one in three births in developingcountries take place without a skilled birthattendant. Skilled attendance rates are lowestin South Asia (41%) and sub-Saharan Africa (45%)(UNICEF, 2008b). Not coincidentally, these are theregions with the highest maternal mortality rates.

Poverty undermines maternal health in severalways. It heightens exposure to threats such asmalnutrition and infectious disease. It can alsoreduce access to vital health care, either becausecare is lacking or because it is unaffordable tothe very poor. The poverty risk factor is graphicallycaptured in a UNICEF review of evidence fromfifty household surveys that found that neonatalmortality rates among the poorest 20% weretypically 20% to 50% higher than for the wealthiestquintile (UNICEF, 2008b). These health inequalitiesfuel education disparities later in life.

The poorest mothers and children are oftenunderserved along the whole continuum of care.In South Asia, being poor reduces by a factor offive the probability of having a skilled health personin attendance during delivery. Even controlling forpoverty, indigenous people and ethnic minorities areoften severely disadvantaged. In Guatemala, non-indigenous women are more than twice as likely astheir indigenous counterparts to give birth in apublic health facility with trained personnel. Thefactors excluding poor and vulnerable householdsfrom basic maternal and child health services varyby country but include cost, distance and the poor

More than one in three births in developing

countries takeplace without a skilled birth

attendant

P R O G R E S S T O WA R D S T H E E FA G O A L S

E a r l y c h i l d h o o d c a r e a n d e d u c a t i o n

4 7

quality of public care. Whatever the underlyingcauses of health disadvantage, the consequencesinclude educational disadvantage later in life.

The strength of the links between maternal healthand education is often overlooked. Some of thoselinks are very direct. Young women of middle tohigher secondary school age, 15 to 19, accountfor one in seven deaths related to pregnancy andchildbirth (WHO and UNICEF, 2003). The youngerthe age at pregnancy, the greater the health risksfor mother and child. Being born to a motherunder 18 increases the risk of infant mortality by60% and the children who survive are more likelyto suffer from low birth weight, undernutrition anddelayed cognitive development (Lawn et al., 2006;UNICEF, 2008b; WHO, 2005).

Empowerment through education is one of thestrongest antidotes to maternal risk. Women withhigher levels of education are more likely to delayand space out pregnancies, and to seek healthcare support. In South and West Asia, almost halfof women with no education give birth withouthaving received antenatal care, compared withnearly 10% for women with secondary education(Figure 2.3). The ‘education advantage’ is evenmore pronounced when it comes to having askilled birth attendant present during delivery.In Burkina Faso, mothers with primary educationare twice as likely to have a skilled attendantpresent as those with no education, and womenwith secondary education are almost four timesas likely. While the association between educationand improved maternal and child indicators isnot evidence of causation, the strength of theassociation points to the importance of the two-way link between investment in healthand investment in education.

Rapid progress is possibleSlow progress towards international goals inareas such as maternal health, child nutritionand survival is sometimes viewed as evidenceof the cost and complexity of effective measures.That assessment is flawed. Without understatingthe extent of the challenges, there is compellingevidence that rapid progress is possible.

Cost-effective measures that work includecomplementary feeding and vitaminsupplementation, a continuum of care duringpregnancy and childbirth, immunization and widerstrategies to tackle killer diseases such as malariaand pneumonia (Black et al., 2008). To make such

interventions available, countries need affordableand accessible health systems, allied to widermeasures for targeting vulnerable groups andcombating malnutrition. Bad news tends todominate the headlines, but there is positivenews too:

Scaling up maternal and child health services.Experience from Bangladesh and Nepal showsthat maternal and child survival can be improvedin low-income settings by increasing access toskilled attendants, antenatal care and familyplanning advice (DFID, 2008b). In the UnitedRepublic of Tanzania, health spending has beenincreased and focused on diseases that affectthe poorest districts. Coverage of key maternaland child health services has expanded, witha marked increase in the recruitment ofcommunity-based midwives and health workers.Child nutrition is improving, as reflected in a 40%decline in child mortality between 2000 and 2004(Masanja et al., 2008).

Achieving results through aid. The GAVI Alliance(formerly Global Alliance for Vaccines andImmunisation), formed in 2000, has supportedthe immunization of 213 million children, savingan estimated 3.4 million lives. From 2000 to 2006,deaths from measles in Africa fell by 90% (GAVIAlliance, 2009a). International partnerships onHIV and AIDS have increased the share of HIV-positive pregnant women receiving antiretroviraltherapy from 15% to 33%, helping preventtransmission to children (Global Fund, 2008a).

The links betweenmaternal healthand education areoften overlooked

Figure 2.3: Educated mothers have better access to antenatal careChildren under age 3 born without antenatal care, by maternal education, South and West Asia and sub-Saharan Africa, circa 2005

Notes: Figures presented are population weighted averages. The sample of countries used to estimatethe South and West Asia average represents more than 90% of the total population of the region andthe sample used to estimate the sub-Saharan Africa average more than 80%.Source: Macro International Inc. (2009).

0

10

20

30

40

50

South and West Asia Sub-Saharan Africa

% o

f chi

ldre

n un

der 3

bor

nw

ithou

t ant

enat

al ca

re

No education

Primary

Secondary or higher

Regional average

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

4 8

Removing cost barriers to vital maternal andchild health services. Inability to pay is a majorfactor limiting access to basic maternal and child health services (Gilson and McIntyre, 2005;Pearson, 2004). Recent experience fromcountries including Ghana, Nepal, Senegal,Uganda and Zambia provides evidence thateliminating charges for basic health services is often followed by a rapid rise in the uptake of services, especially by the poor (Deininger and Mpuga, 2005; Yates, 2009) (Box 2.2).

Putting nutrition at the centre of the povertyreduction agenda. Over the past two decadesViet Nam has achieved some of the world’s mostrapid reductions in child malnutrition. ItsNational Target Programme has focused onthe 2,374 communes with the highest rates ofpoverty and child malnutrition. Supplementaryfeeding programmes and maternal and childhealth care have figured prominently. NationalInstitute of Nutrition surveys indicate thatstunting rates fell by one-quarter from 1999to 2005 (Khan et al., 2008). In Brazil, the ZeroHunger programme, a concerted drive to combatmalnutrition, contributed to a fall in malnutritionrates in the north-east, the poorest region, from18% to 16% in the decade to 2005 (Ruel, 2008).

Implementing effective social protection.Programmes that provide parents with income,services and incentives can help combat earlychildhood deprivation. Large-scale programmessuch as Bolsa Familia in Brazil andOportunidades in Mexico directly link cashtransfers to participation in child nutritionprogrammes – and both have reduced stuntingand improved cognitive development (Fiszbeinet al., 2009) (Box 2.3).

The artificial separation of health and educationin public policy is particularly damaging for earlychildhood provision. Education planners oftenmeasure progress in primary education bynumbers entering classrooms, pupil/teacher ratiosand the quality of school infrastructure. There isa widespread view that children’s nutritional andhealth status before school age is a health policymatter. This silo mentality produces a distortedpicture of policy priorities. Millions of children enterschool having suffered irreparable damage to theirlearning potential as a result of malnutrition andmicronutrient deficiencies. Poor maternal healthand risks during pregnancy and childbirth areimportant contributory factors. The upshot of publicpolicy failure in the areas of nutrition and maternaland child health care is not just unnecessaryhuman suffering, but also the erosion of benefitsassociated with investment in education andprogress in getting children into school.

Early childhood educationprogrammes — a mixed record

Learning starts in the home, as children manipulateobjects and materials, explore the world aroundthem and develop language. During the crucial

The removal of cost barriers has played a critical role in opening upopportunities for education. Yet cost barriers to maternal and child healthcare remain largely intact, with damaging consequences for health andeducation. The inability of poor households to afford health costs oftenleads to fatal delays in treatment or to their wholesale exclusion from formal health care. Research in countries as diverse as Chad, India andSudan points to cost as a major factor restricting the use by poor women of maternal and child health services.

As in education, the scale of the barriers that fees create for the very pooris often revealed when fees are removed. When Uganda withdrew healthfees in 2001, the number of outpatients visiting hospitals went the same way as school enrolments after fees were withdrawn several years earlier:attendance rates doubled in less than a year, with the poorest groupsrecording the highest increases. After Burundi removed all health fees forpregnant women and children in 2006, average monthly births in hospitalspromptly rose by 61%. In Nepal, the removal of fees, allied to increasedinvestment in the recruitment and training of community health workers,has also increased access to care.

Many governments across Africa and beyond are reconsidering health fees.There is compelling evidence that charging for basic services is ineffective,inefficient — fees generate only 5% to 6% of health sector revenue — andinequitable. In the past two years, Burundi, Ghana, Kenya, Lesotho, Liberia,the Niger, Senegal and Zambia have abolished fees in key areas. Most majordevelopment agencies, including the WHO and the World Bank, have alsoadopted clear positions against fees. Meanwhile, some donors have providedadditional aid to countries that have removed fees, including France (for theNiger) and the United Kingdom (for Burundi, Ghana, Nepal and Zambia) .

Eliminating user fees on maternal and child health care should be seen as an urgent priority. However, it is not a stand-alone strategy. Rapid increasesin demand for already overstretched services can lead to deterioration inquality and long queues for treatment — outcomes that undermined thebenefits of free maternal health care in Ghana. As in the education sector,making access more affordable should be seen as one element in a broaderpackage of policy reforms. Increased investment to strengthen healthsystems, greater equity in public spending and improved governance are allimportant. And there is no substitute for recruiting and training more healthworkers. The shortage of trained health workers is estimated at over1 million in sub-Saharan Africa alone.

Sources: Yates (2009); Nabyonga et al. (2005); Batungwanayo and Reyntjens (2006); Cohen and Dupas (2007); Gilson and McIntyre (2005); Witter et al. (2009).

Box 2.2: Removing cost barriers to maternal and child health services

P R O G R E S S T O WA R D S T H E E FA G O A L S

E a r l y c h i l d h o o d c a r e a n d e d u c a t i o n

4 9

formative years, children develop the cognitive andwider skills that will prepare them for school. Pupilsfrom disadvantaged backgrounds often enter schoolcarrying a legacy of disadvantage in many areas,including lower levels of communication, languageand literacy skills. The effects of growing up in adisadvantaged home are seldom reversed later inlife – in fact, the gaps widen as children progressthrough their school years (UNESCO, 2005).

Narrowing the opportunity divideAn early start in education is particularly importantfor children from disadvantaged families. Poverty,low levels of parental education or speaking aminority language at home are among the mostpowerful transmitters of disadvantage acrossgenerations. Good-quality early childhood provisioncan cut the transmission lines.

By the time children enter school, disparities inlanguage skills linked to income and other factorsare often so marked that children can never catchup. Evidence from the United States demonstratesthat test scores at the age of 18 are predictable byage 5 (Heckman, 2008). Research in Ecuadorindicates that differences in vocabulary test scoresbetween children from different wealth groups arelimited at age 3 but that by age 5 the gap is far toowide to be closed in later school years (Paxson andSchady, 2005b) (Figure 2.4). In the United Kingdom,longitudinal studies show that test scores at22 months are a strong predictor for educationalqualifications at 22 years (Feinstein, 2003).Moreover, studies have shown that children fromlow socio-economic backgrounds but with highcognitive ability scores at 22 months are overtakenby children with lower scores from more affluentfamilies between the ages of 5 and 10 years.

Income differences are not the only source ofadvantage and disadvantage. Parental education,ethnicity and home language all exercise a stronginfluence on early childhood test scores andsubsequent educational achievement (Brooks-Gunnand Markman, 2005; EACEA, 2009; Leseman andvan Tuijil, 2005). The issue of language is especiallysalient. There is strong evidence from theOrganisation for Economic Co-operation andDevelopment (OECD) that having a home languagethat is different from the language used in schoolsignificantly decreases achievement for immigrantchildren in both primary and secondary school(Christensen and Stanat, 2007; Schnepf, 2004).Remedial action often meets with limited success.In Norway, 20% of migrant students placed in

special language training groups on enteringschool never leave them and in Switzerland mostmigrant children not deemed equipped to entermainstream classes are still in such groups aftertwo years (Field et al., 2007). Moreover, evidencefrom several countries shows that catching upthrough special classes often requires studentsto miss the normal curriculum (Karsten, 2006).

Early childhood education can play an important rolein offsetting social, economic and language-baseddisadvantage. Evidence from around the worldindicates that high-quality early care is goodfor all children, but particularly for those fromdisadvantaged backgrounds. The followingare among the findings to emerge from a rangeof rigorous evaluations:

The Head Start Impact Study in the United Statesrandomly evaluated about 5,000 3- and 4-year-olds. It found small to moderate statistically

High-quality earlycare is particularlyimportant for children fromdisadvantagedbackgrounds

In many developing countries, serious delays in children’s cognitivedevelopment damage their prospects in school and their productivity as adults. Understanding the causes of cognitive deficits and developingways to reduce them are critical policy priorities.

The Atención a Crisis programme in Nicaragua demonstrates the potentialbenefits of early intervention. Significant cash payments, representing onaverage about 15% of household income, were made every two months towomen in poor rural households. To be eligible, parents had to take childrenof pre-school age for regular visits to health centres, where they wereweighed and received vaccinations and food supplements.

This pilot programme, carried out during 2005 and 2006, included a careful evaluation. Results indicated that the programme improvedseveral dimensions of child development:

After only nine months in the programme, children aged 3 to 4 years had made up 1.5 months’ delayed personal-social and languagedevelopment on one set of test scores, rising to 2.4 months for children aged 5 to 6 years.

Participating households were found to have higher values for signs of parental stimulation, including the availability of books, paper and pencils, and the likelihood of parents reading to children.

Overall food expenditures increased among treated households,especially on nutrient-rich foods.

Wide-ranging preventive health benefits were identified. Participatingchildren were more likely to have had a growth check-up, receivedvitamin and iron supplements, and to have been treated with de-wormingdrugs. The reported health status of mothers had also improved.

Source: Macours et al. (2008).

Box 2.3: Cash transfer in Nicaragua — overcoming cognitive deficits

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

5 0

significant increases in four key cognitive scores,including pre-reading, pre-writing, literacy skillsand vocabulary. While Head Start children scoredbelow average for all children, reflecting racialand social background factors, the programmehalved the achievement gap that would have beenexpected in its absence (US Department of Healthand Human Services, 2005).2

Attending the French pre-primary educationsystem (école maternelle) increases classretention of low-income and immigrant childrenin primary school by 9% to 17%, with widerreported benefits for literacy and numeracy(Nusche, 2009).

Early childhood care can help overcomelanguage-based disadvantage and the problemsfaced by children of migrants (Cunha et al., 2005).In the Netherlands, children of Turkish andMoroccan immigrants who spent two years inkindergarten halved the average test score gapfrom the national average (Leseman, 2002).

In New Zealand, 12-year-olds who had partici-pated in high-quality early education performedbetter in reading and mathematics, aftercontrolling for household income (UNICEF, 2008b).

While these findings relate to developed countries,there is also evidence from developing countriesthat effective early childhood care and educationcan both raise learning achievements and narrowdisparities. That evidence was extensively reviewedin the EFA Global Monitoring Report 2007. Whilethe precise channels of influence are a subjectof debate, good-quality early childhood provisionclearly has the potential to weaken the influenceof parental factors on later education achievement.

Pre-primary education — slow and unequal expansion‘Pre-primary’ is an umbrella term covering awide range of providers and programmes, mostlyfor children aged 3 and above. Countries differenormously in the mix of public and privateprovision, and in financing arrangements andgovernance. As in other areas of education, data oncoverage say little about quality, but high-qualityprogrammes tend to start early, be based incentres, have a critical mass of trained teachersand involve parents (UNICEF, 2008b).

Participation in pre-primary education has beensteadily increasing. Some 140 million children wereenrolled in pre-school programmes worldwidein 2007, up from 113 million in 1999. The grossenrolment ratio (GER) climbed from 33% to 41%over the same period (Table 2.1). Increases havebeen most pronounced in sub-Saharan Africa, andSouth and West Asia, albeit from a low base. Onein seven children in sub-Saharan Africa is enrolledin an early childhood programme, compared wihone in three for all developing countries.

Looking beyond the regional data reveals a diversearray of country experiences. Among the countriesfor which data are available, seventeen states insub-Saharan Africa have coverage rates of lessthan 10%. In the Arab States, levels of pre-primarycoverage are far lower than average income mightseem to indicate: out of nineteen countries withdata for 2007, fourteen have GERs below 50%. Egyptand Saudi Arabia have lower levels of coverage thansome far poorer countries, including Nepal and theUnited Republic of Tanzania. Indeed, sub-SaharanAfrica has increased pre-primary enrolment atthree times the rate of the Arab States, with GERsrising by more than 20% since 1999 in severalcountries, including Burundi, Liberia and Senegal(Annex, Statistical Table 3A). The Arab States regionalso remains the only one with significant genderdisparity at early childhood level: just nine girls areenrolled for every ten boys.

2. In 2005-2006, 24% ofchildren from the poorest20% of United Stateshouseholds were incentre-based Head Startprogrammes, comparedwith 1% of children fromthe wealthiest 20%.Evaluations of earlier pilotchildcare programmes –such as the NorthCarolina AbecedarianProject and PerryPreschool Program –have also recorded wide-ranging benefitsassociated with pre-school, extending fromprimary education tocollege attendance,employment, wages andcrime reduction(Campbell et al., 2008;Karoly et al., 2005;UNESCO, 2008a;Schweinhart et al., 2005).Observed effects werestrongest for poorchildren and childrenwhose parents had littleeducation.

Figure 2.4: Wealth-based gaps in learning begin early and widen over timeTest scores across ages for the poorest and the fourth deciles in Ecuador, 2003–2004

Notes: The test scores used are from the Test de Vocabulario en Imágenes Peabody, the Spanish versionof the Peabody Picture Vocabulary Test. The figure presented here, a smoothed version of the original figure(which appears in the source document), has also been reproduced elsewhere (e.g. Fiszbein et al., 2009, and World Bank, 2006j ).Source: Paxson and Schady (2005b).

50

60

70

80

90

100

110

36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70

4th decile

1st (poorest) decile

3-year-olds 4-year-olds 5-year-olds

Age in months

Stan

dard

ized

test

scor

es

P R O G R E S S T O WA R D S T H E E FA G O A L S

E a r l y c h i l d h o o d c a r e a n d e d u c a t i o n

5 1

Developed countries vary considerably in theirblend of crèches, pre-primary schools, centre-based day care and home support. They also differin the balance between public and private financingand in the age groups that programmes reach.Some countries, notably in the Nordic area, havehigh rates of coverage for children under 3, thoughmost early childhood programmes in OECDcountries cover ages 4 to 6. The duration of pre-primary education varies from one to four years.In Sweden, full-time free early childhood educationis available to all children, from age 3, for elevenmonths of the year; in the United Kingdom, freeprovision is available part time for 3- and 4-year-olds (EACEA, 2009). Most European Union countriesprovide two years of free pre-school.3 By contrast,in the United States, there is no statutory right topre-school before age 5, though about 60% ofchildren in the pre-school age group were enrolledin 2007.

Differences within countries are often as markedas differences across borders. This is especiallytrue of countries that combine high levels ofdecentralization with subnational autonomy. TheUnited States provides a striking example. Virtuallyevery 4-year-old in Oklahoma can start school atage 4. In eight other states – including Florida,South Carolina and Texas – more than half of 4-year-olds attend a public pre-school programme.

At the other end of the range, twelve states haveno regular state pre-school education programmeand in eight states less than 20% of children areenrolled (Barnett et al., 2008). There are alsomarked differences in the quality of provision(Ackerman et al., 2009). Ten benchmarks havebeen established for assessing quality standards.4However, programmes in Florida are required tomeet only four benchmarks and Texas sets nolimits on class size or staff/child ratios. Spendinglevels per child also vary markedly: five statesspend more than US$8,000 per pupil while anotherfive spend less than US$3,000 (Barnett et al., 2008).

Reaching the vulnerable and disadvantagedGoal 1 of the Dakar Framework for Action commitsgovernments to expanding early childhood careand education ‘especially for the most vulnerableand disadvantaged’. This is for good reason.Children from disadvantaged households have themost to gain from early childhood care – and themost to lose from being excluded. Unfortunately,cross-country evidence strongly suggests thatthose who need it most receive it least.

Household poverty and low levels of parentaleducation are two of the most pronounced barriersto early childhood programmes. Evidence from asurvey of fifty-six developing countries shows thatbeing born into a poor household or having a

3. In the European Union, about 87% of 4-year-olds are in school (EACEA, 2009).

4. The standards include teacher and assistant teacher degrees and specialized training, in-service training provision, class size, staff/child ratios, support services, meals and monitoring. Just two states – Alabama and South Carolina – meet all ten benchmarks.

Children fromdisadvantagedhouseholds havethe most to gainfrom earlychildhood care

School year ending in School year ending in

113 139 24 33 41 26

80 106 32 27 36 3225 26 4 73 80 10

7 8 7 45 63 39

5 10 82 10 15 532 3 26 15 19 251 1 13 19 28 44

37 39 4 40 47 1837 38 4 40 47 19

0.4 0.5 12 61 67 1121 36 69 21 36 7116 20 22 56 65 17

0.7 0.8 16 65 74 1316 19 22 55 65 1719 20 6 75 82 9

9 10 5 50 64 30

World

Developing countriesDeveloped countriesCountries in transition

Sub-Saharan AfricaArab StatesCentral AsiaEast Asia and the Pacific

East AsiaPacific

South and West AsiaLatin America and the Caribbean

CaribbeanLatin America

North America and Western EuropeCentral and Eastern Europe

Table 2.1: Pre-primary enrolment and gross enrolment ratios by region, 1999 and 2007

1999

(%) (%)

2007

Changebetween 1999

and 2007 1999 2007

Changebetween 1999

and 2007

Source: Annex, Statistical Table 3B.

Total enrolment Gross enrolment ratios

(millions) (%)

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

5 2

mother with no education carries a large handicapwhen it comes to early childhood care, regardlessof age, gender or place of residence (Figures 2.5and 2.6). Living in one of Zambia’s pooresthouseholds cuts the chance of participating inearly childhood care by a factor of 12 comparedwith children in the wealthiest households, andthe factor rises to 25 in Uganda and 28 in Egypt(Nonoyama-Tarumi et al., 2008). Such figuresdemonstrate the degree to which early childhoodprovision is reinforcing inequalities associatedwith the home environment.

Why do children from disadvantaged householdsface the highest barriers to entry? In some cases,it is because facilities are too far from their homes.In others, facilities are accessible but unaffordable– a problem that has held back efforts to expandcoverage in Egypt (UNESCO, 2008a). However,several countries have succeeded in expandingaccess. In Chile, a programme aimed at achievingearly childhood care for all 4-year-olds hastargeted the poorest income groups (Box 2.4).

Rich countries have also struggled to meet equitygoals. There is extensive evidence from theEuropean Union that low-income families andimmigrants have less access to good-quality earlychildhood care (Arnold and Doctoroff, 2003;Nusche, 2009; Sylva et al., 2007).

Evidence from the United States also documentslarge disparities (Barnett et al., 2008). Familieswith incomes just above the poverty line facesome of the greatest difficulties in gaining access,demonstrating the importance of targetinghouseholds at this level. Maternal education alsohas a marked bearing on United States pre-schoolparticipation: attendance rates of 4-year-olds are55% for children of mothers who have droppedout of secondary school but 87% for children ofmothers with a college education (Barnett et al.,2008; Barnett and Yarosz, 2007).

On a more positive note, several governments arescaling up early childhood care as part of wideranti-poverty initiatives. In the United Kingdom,Sure Start, a flagship strategy introduced in 1997to tackle child poverty, social exclusion andeducational disadvantage, now reaches 2.4 millionfamilies (Every Child Matters, 2009).

In Chile, a programme

aimed at achieving

early childhoodcare for

all 4-year-oldshas targeted the pooresthouseholds

BoliviaNicaragua

HaitiCôte d’IvoireMadagascar

KenyaBotswanaViet Nam

LesothoPhilippines

ZambiaNigerTogo

RwandaMongolia

UgandaEgypt

0 5 10 15 20 25 30

Odds ratios for the likelihood of participationin early childhood programmes

Egyptian children fromthe richest families aremore than 25 timesmore likely than thosefrom the poorestto participate in earlychildhood programmes

Figure 2.5: Children from rich families are more likely to participate in early childhood programmes Likelihood of 3- and 4-year-olds participating in early learningprogrammes, children from the richest 20% compared with childrenfrom the poorest 20%

Notes: Using odds ratios (see glossary), this figure compares the likelihood of young children participating in early learning programmes, depending on the wealth status of their families. Specifically, odds ratios provide an estimate of the differences in probability of attending early childhood programmes betweenone reference group (the poorest 20%) and another (the richest 20%), and they areestimated from a logistic regression with five dependent variables: gender, age,place of residence, mother’s level of education and household wealth.Source: Nonoyama-Tarumi et al. (2008).

ZambiaEgypt

Dominican Rep.India

LesothoMadagascarCôte d’Ivoire

CameroonViet Nam

ChadNiger

Burundi

0 2 4 6 8 10 12

Odds ratios for the likelihood of participationin early childhood programmes

Children in Burundiwith educated mothersare ten times more likelyto participate in earlychildhood programmes

Figure 2.6: Children of educated mothers are more likelyto attend pre-school programmesLikelihood of 3- and 4-year-olds participating in early learningprogrammes, children of mothers with secondary education or higher compared with children of mothers with no education

Notes: This figure compares the likelihood of young children participating inearly learning programmes, based on the mother’s education level. The oddsratios are calculated in the same way as those in Figure 2.5 for differentreference groups: one consists of children with a mother with no education and another of children with a mother having secondary education or higher.Source: Nonoyama-Tarumi et al. (2008).

P R O G R E S S T O WA R D S T H E E FA G O A L S

E a r l y c h i l d h o o d c a r e a n d e d u c a t i o n

5 3

Another striking example comes from NewZealand. Since 2007, all 3- and 4-year-olds in thecountry have been entitled to twenty hours a weekof free early childhood education (Froese, 2008;May, 2008). Efforts are being made to improve thequality of early childhood education available to Ma-ori children. Curricula and teaching materialshave been modified through partnerships with Ma-ori groups. Scholarships and incentives havebeen expanded to attract Ma-ori-language speakersinto early childhood teaching. In the five years to2007, the number of Ma-ori-speaking educatorstripled and the share of Ma-ori primary schoolentrants having been to pre-school rose from 86%to 91% (New Zealand Ministry of Education, 2009).

Conclusion

The Dakar Framework for Action does not seta quantitative goal for early childhood careand education, so what targets – if any – shouldgovernments set? And what role shouldgovernments play in paying for and providing care?

There are no universal answers to these questions.As the EFA Global Monitoring Report 2007documented, many countries have set unrealistictargets. Countries struggling to get children intoand through basic education have to weigharguments for universal early childhood coverageagainst real resource constraints. At the same time,governments need to recognize the potentialefficiency and equity gains from investing in earlychildhood care. As one Nobel Prize-winningeconomist has written: ‘Early interventions targetedtowards disadvantaged children have much higherreturns than later interventions. … At current levelsof resources, society over-invests in remedial skillinvestments at later ages and under-invests in theearly years’ (Heckman, 2006, p.1902).

While that reflection is based on evidence from theUnited States, it is likely to have a wider application.The lesson to be drawn is that public investmentshould be geared towards narrowing disparities,targeting marginalized groups and providing good-quality services that are accessible to the poor.

Governments need to recognizethe potentialefficiency andequity gains frominvesting in earlychildhood care

Chile has some of the deepest and most persistenteducation inequalities in Latin America. Recentreforms are attempting to strengthen equity byexpanding and improving early childhood care.

After her election in 2006, President MichelleBachelet initiated a major overhaul of early childhoodcare, including raising public spending (Larrañaga,2009; OECD, 2009e). The most ambitious measureinvolves building 3,000 new childcare facilities andestablishing a national child development initiative,Chile Crece Contigo, for all children under 5, as partof the health care system.

Chile Crece Contigo, a result of collaboration bygovernment, child development experts and otherinterested parties (Frenz, 2007), aims to meet theneeds of vulnerable families and children duringthe critical phases of early childhood development.Families have access to a wide range of social andhealth services through primary care centres. Theirprogress is monitored via information technology.Implementation is managed by nine nationalministries and coordinated through regional,provincial and local governments.

A concerted effort has been made to reach childrenfrom the poorest 40% of households. In that incomebracket, young children with mothers at work, inschool or seeking employment are eligible for freechild care in the sala cuna (under 2) or the jardíninfantil (ages 2 and 3).

Central to the strategy is commitment to quality.Efforts have been made from the outset to measureand assess the development of vocabulary, languageand wider skills through Un Buen Comienzo, aprogramme that runs in sixty schools in thirteencommunes of Santiago. Using rigorous evaluation,Un Buen Comienzo seeks to reduce the vocabularygap between children in low-income families andother children, improve pre-school attendance andreduce later reading difficulties. Teacher development,parental literacy and engagement, and child healthare emphasized.

Sources: Frenz (2007); OECD (2009e).

Box 2.4: Expansion of early childhood education in Chile

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

5 4

Universal primaryeducation

Goal 2. Ensuring that by 2015 all children,particularly girls, children in difficultcircumstances and those belonging to ethnicminorities, have access to and complete, free andcompulsory primary education of good quality.

The past decade has seen rapid progress towardsuniversal primary education. Some of the world’spoorest countries have dramatically increasedenrolment, narrowed gender gaps and extendedopportunities for disadvantaged groups. Schoolcompletion rates are also rising. Theseachievements provide a marked contrast to the‘lost decade’ of the 1990s. But there are limitsto the good news. In the midst of an increasinglyknowledge-based global economy, millions ofchildren are still out of school and countlessmillions more start school but drop out beforecompleting primary education. And there is nowa real danger that the global economic crisiswill stall, and perhaps even reverse, the gainsregistered over the past decade (see Chapter 1).

The post-Dakar record has to be assessedagainst the ambition set out in Goal 2 of the DakarFramework for Action: universal primary educationby 2015. Is the goal still attainable? The answerwill depend on decisions taken over the next twoyears by national governments and aid donors.The window of opportunity for ensuring that allprimary school age children currently out ofschool complete a full cycle of primary educationis rapidly closing. Getting all children into schoolby 2015 is still feasible, but the goal will not beachieved with a business-as-usual approach.

The World Education Forum (Dakar, 2000) gavenew impetus to education, both nationally andinternationally. Yet the hard fact remains that theworld will fall short of the targets set and that farmore could have been achieved. Many developingcountries could have done much more toaccelerate progress, notably through policies toovercome inequalities in education. Meanwhile,donors have a mixed record of delivery on theircollective commitment to back nationalprogrammes with increased financial support –an issue addressed in more detail in Chapter 4.

This section documents progress towardsuniversal primary education. Looking behind thenational data, it provides an in-depth look at someof the crucial challenges facing governments inthe countdown to 2015. The following are amongthe key messages to emerge:

Out-of-school numbers are dropping for primaryschool age children, but getting all children intoschool will require a far stronger focus on themarginalized. When the Dakar forum was held,over 100 million children of primary school agewere out of school. By 2007, the figure hadfallen to 72 million. This headline figure bearstestimony to national governments’ efforts.The bad news is that, on current trends, some56 million children could still be out of schoolin 2015. Changing this scenario will requirea far stronger commitment by governmentsto reach girls and other marginalized groups.It will also require a sharper focus on countriesaffected by conflict or engaged in post-conflictreconstruction.

Progress towards universal primary enrolmenthas been partial and mixed. Despiteencouraging progress, many of the poorestcountries are struggling to reach universalenrolment. Less attention has been paid tohigher-income countries with significant out-of-school populations, such as the Philippinesand Turkey. Such countries will have to targetmarginalization far more systematically todeliver on the Dakar commitments. Newresearch indicates that official enrolment datamay overstate the numbers of children in schoolat the appropriate age, suggesting that moreneeds to be done to address the problem oflate entry and dropout. Household survey datafor a number of countries indicate overestimatesof 10% or more in school attendance rates.

Gender barriers remain intact. There hasbeen progress towards greater gender parityin school enrolment. Even so, being borna girl carries with it a significant educationdisadvantage in many countries. Thatdisadvantage is reflected in the fact that girlsstill account for 54% of the out-of-schoolpopulation. Moreover, out-of-school girls arefar more likely than boys never to go to school.In twenty-eight countries, there are fewer thannine girls in primary school for every ten boys.Poverty further reinforces gender disparity.

Despiteencouraging

progress, many of the poorestcountries arestruggling to

reach universalprimary

education

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

5 5

Getting children into primary school is justa first step. Universal primary educationinvolves entering school at an appropriate age, progressing through the system andcompleting a full cycle. Unfortunately, millionsof children enter school late, drop out earlyand never complete a full cycle. Moreintegrated approaches to monitoring arerequired to measure the real state of progresstowards universal primary education.

Out-of-school adolescents are oftenoverlooked. Monitoring progress towardsinternational development goals in educationfocuses on the primary school age group.The situation of adolescents has been subjectto less scrutiny. There are some 71 millionchildren of lower secondary school agecurrently out of school. Many have notcompleted a full primary cycle and facethe prospect of social and economicmarginalization. Counting adolescentsdoubles the global headline figure for out-of-school children.

The section is divided into three parts. Part 1looks at progress towards one of the mostimportant requirements for achieving universalprimary education: getting all children into school.It looks beyond the headline numbers to explorethe characteristics of the out-of-schoolpopulation. Removing the barriers that keepchildren out of school is the first step towardsachieving universal primary enrolment – ensuringthat the entire primary school age group is inschool by 2015. Part 2 looks at enrolment trends.Part 3 examines the problem of retention andprogression through primary school, and thetransition to secondary education.

Numbers of out-of-school childrenare declining, but not fast enough

Malina is a 12-year-old living in a rural area ofRattanak Kiri, a remote hill district in Cambodia.She is a member of a minority ethnic group and has never been to school.

Lucy, 12, lives in the slum of Kibera in the Kenyancapital, Nairobi. When she was 8, she enrolled inprimary school, but in the second grade shedropped out. She wants to go back to school buthas to take care of her brother, and her mothercannot afford the fees, uniforms and books.

Victor is 14 years old. He lives on the streets ofManila and makes a living by selling newspapersat road junctions. He went to primary school forfour years, but left before completing it and has no prospect of returning.

Maria, 15, is in grade 4 of her local primaryschool in Panama, having started late, repeatedtwo grades and dropped out for a year whenshe was 12.

Compared with the 1990s, the first decade of thetwenty-first century has been one of rapid progresstowards universal primary education. Out-of-schoolnumbers are falling and more children arecompleting primary school. Yet the sheer size of theout-of-school population remains an indictment ofnational governments and the entire internationalcommunity. Denying children an opportunity to puteven a first step on the education ladder sets themon a course for a lifetime of disadvantage. It violatestheir basic human right to an education. It alsowastes a precious national resource and potentialdriver of economic growth and poverty reduction.

As the experiences of the children cited abovetestify, ‘out of school’ is a simple concept with manymeanings. Some children of late primary schoolage and even secondary school age have neverbeen to school. Others have started school butdropped out. Still others are in a state of flux,moving between in-school and out-of-school status.The out-of-school figures in this section refer onlyto children of primary school age who are not inschool. They represent the tip of the total out-of-school iceberg, since they do not cover adolescentsof secondary school age who have not completedprimary school. Even within the primary school agegroup, data for any one year provide only a staticsnapshot of a dynamic and complicated picture.

The snapshot for 2007, the latest year for whichdata are available, points to continued progressbut still large deficits. There are some remarkableachievements since 1999:

Out-of-school numbers are falling. Worldwide,the number of children of primary school agewho are out of school has declined by 33 millionsince the Dakar pledges were made, from105 million in 1999 to 72 million in 2007.Seven out of every ten out-of-school childrenlive in South and West Asia, and sub-SaharanAfrica (Figure 2.7).

There were 72 million childrenout of school in 2007

6. The key aim of the programme is to universalize elementaryeducation by 2010. Commitments include constructing and improvinginfrastructure in deprived areas, along with measures targeted towardsareas with large marginalized populations (scheduled castes, scheduledtribes, Muslims) or low female literacy (Ayyar, 2008; Govinda, 2009);see also Chapter 3.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

5 6

The gender gap is shrinking. The share of girlsin the out-of-school population declined from58% to 54%.

South and West Asia have achieved rapidprogress. The region more than halved its out-of-school population – a decline of 21 million.The region also cut the share of girls in the out-of-school total, from 63% to 58%.

Sub-Saharan Africa has registered strongprogress. During a period in which the size of itsschool age population increased by 20 million,sub-Saharan Africa reduced its out-of-schoolpopulation by almost 13 million, or 28%. Thestrength of the region’s progress can be gaugedby a comparison with the 1990s. Had the regionprogressed at the same pace as in the 1990s,18 million more children would be out of school.

The limits to progress also have to beacknowledged. Not only is the world off trackfor the Dakar commitments, but there is causefor concern over the pace of change:

The 2015 target will be missed. If the worldwere to continue the linear trend for 1999–2007,an estimated 56 million children would still beout of school in 2015.5 Slower economic growth,pressure on education budgets and rising poverty

associated with the global economic crisis couldsignificantly inflate this figure (Figure 2.8).

Progress has slowed. The post-1999 overviewprovides a positive gloss on some disturbingunderlying trends. Two-thirds of the total declinein out-of-school numbers took place during thetwo years to 2004, when the numbers droppedby 22 million. In the three years to 2007, the out-of-school population fell by just 8 million.The slowdown illustrates one of the centralchallenges now facing governments: the closercountries get to universal primary education,the harder it becomes to reach children still outof school. That is why sustained progress willrequire a stronger focus on marginalization.

South and West Asia dominated the reduction.Much of the decline took place in India, whichreported a fall of almost 15 million in out-of-school numbers in the two years after the 2001launch of the Sarva Shiksha Abhiyan (universalprimary education) programme.6

The deficit in sub-Saharan Africa remains large.Fully one-quarter of sub-Saharan Africa’sprimary school age children were out of school in2007 – and the region accounted for nearly 45%of the global out-of-school population. Half of thetwenty countries with more than 500,000 childrenout of school were in sub-Saharan Africa (seeFigure 2.12 below). Nigeria alone contributedover 10% of the global total. Progress in theregion has been uneven. Some countries withlarge out-of-school populations in 1999 havemade major advances; examples includeEthiopia, Kenya, Mozambique, the UnitedRepublic of Tanzania and Zambia. Ethiopia andthe United Republic of Tanzania each reducedout-of-school numbers by over 3 million between1999 and 2007. Countries making only limitedprogress include Liberia, Malawi and Nigeria.

Conflict remains a major barrier. Children livingin countries enduring or recovering from conflictare less likely to be in school. Many suchcountries lack publicly available data and soreceive less prominence in international debatesthan they merit. But a lack of reliable data shouldnot deflect attention from the scale of the

5. This figure should notbe compared with thepartial projection in theEFA Global MonitoringReport 2009, whichtreated a smaller group ofcountries using a differentmethodology, and did notinclude countries such asthe Democratic Republicof the Congo and theSudan.

An estimated 56 million

children couldstill be out of

school in 2015

0

20

40

60

80

100

120

Central Asia

North America and Western Europe

Central and Eastern Europe

Latin America and the Caribbean

Arab States

East Asia and the Pacific

South and West Asia

Sub-Saharan Africa

45

32

18

9

6339

6

84

1999 2007

Out-o

f-sch

ool c

hild

ren

(mill

ions

)

105 million

72 million

Figure 2.7: Numbers of out-of-school children are decliningOut-of-school children by region, 1999 and 2007

Source: Annex, Statistical Table 5.

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

5 7

problem. Best estimates suggest that morethan 25 million out-of-school children live inlow-income countries affected by conflict –around 35% of the global total.7 Finding waysto reach children in conflict-affected areasof countries such as Afghanistan, the CentralAfrican Republic, the Democratic Republicof the Congo and the Sudan is one of the mosturgent of all EFA challenges. Attaching moreweight to education in post-conflict recoveryis also vital. While Liberia now has peace andstability, 447,000 of its children were out ofschool in 2008 – an increase of almost 180,000over 1999.

Numbers are probably underestimated.Estimating the number of children from therelevant age group who are out of primaryschool is an inexact science. Administrativedata that schools report to ministries ofeducation are an important resource andnational reporting systems are becoming moreeffective. However, uncertainties overdemographic profiles (and hence the numberof children in each age group) can cloud theissue. Household surveys are another sourceof information, usually obtained throughparental reporting on whether their childrenattend school. There are often significant

inconsistencies between these two data sources.For twenty-nine countries in sub-Saharan Africa,and South and West Asia examined in thisReport, household surveys show around 50%fewer children in school – 22 million in total –than administrative data indicate. This isequivalent to an increase of 30% in the globalout-of-school estimate. Such findings illustratehow different measurement tools can generatedifferent results. They also demonstrate theimportance of national governments,international agencies and researchersworking together to build a more completepicture of the out-of-school population (Box 2.5).

Who and where are the out-of-school children — and what are their chances of entering school?Being out of school is not a fixed condition. Thecategory covers children who have dropped out ofschool temporarily or permanently, those who havenever been to school but might start late and thosewho will never go to school. Data constraints makeit difficult to unravel the precise characteristics ofthe out-of-school population. However, a modeldeveloped by the UNESCO Institute for Statistics(UIS) makes it possible to predict, on the basis ofpast evidence, what share of out-of-school childrenis likely to enrol in the future (UIS, 2009a).

7. The countries included areones that experienced armedconflicts resulting in at leasttwenty-five battle-relateddeaths per year over at leastthree years between 1999and 2007 or more than 1,000battle-related deaths in atleast one year during thesame period. Of these, onlycountries categorized asleast developed countries by the United Nations or low-income countries by the World Bank in 2007 wereincluded. The proportion of children out of schoolaccording to this definition is lower than the frequentlyquoted figure of more thanhalf. The higher figure isbased on calculations using a different methodology,which includes somecountries identified as‘fragile’ but not in conflict, aswell as some middle-incomeconflict-affected countries(Save the Children, 2009a).

Household surveydata suggest anincrease of 30% in the global out-of-schoolestimate

56 million

8 million

23 million

0

20

40

60

80

100

120

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

South and West Asia39 million

Rest of the world

105 million105 million

72 million

Sub-Saharan Africa45 million

Out-o

f-sch

ool c

hild

ren

(mill

ions

)

Figure 2.8: Missing the target — out-of-school trends projected to 2015Projected numbers of out-of-school children to 2015

Note: Projections based on regional compound growth rates for 1999–2007.Source: UIS database (data for 1999–2007).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

5 8

The model explores the importance of gender,income and location in determining whether ornot children are in school. These categories ofdisadvantage interact with each other and withwider factors – such as language, ethnicity anddisability – to create multiple barriers to schoolentry and survival. Chapter 3 explores the widerfactors and their interactions in detail.

Young girls. Disparities between boys and girlsare narrowing, but females still accounted for54% of the global out-of-school population in2007. Gender parity would cut the number of

girls out of school by over 6 million. Genderdisadvantage is most pronounced in the ArabStates, Central Asia, and South and West Asia.In Pakistan, girls accounted for 60% of out-of-school children in 2006.

Children from poor households. Parental wealthstrongly influences prospects of being out of school.Low average income in many of the countries withlarge out-of-school populations means that povertyextends far beyond the poorest 20%. However,as evidence from household surveys shows, thepoorest face distinctive problems. In India, children

Data on enrolment by age are often treated uncritically asan accurate record of how many primary school age childrenare actually in primary school. The information passes fromschools to education ministries and then to the UNESCOInstitute for Statistics (UIS), which compiles the internationaldata used to compare countries, monitor progress and informinternational meetings. But is the information accurate?

Research by the UIS compares enrolment figures generated bygovernments with data reported in Demographic and HealthSurveys of households in twenty-nine countries (Figure 2.9).The analysis found that, when compared with householdsurvey data, school registers tend to count more childrenwithin the official primary school age range. Correcting thatbias would have the effect of reducing the net enrolmentratio — by significant margins in some countries. In the UnitedRepublic of Tanzania, the discrepancy was equivalent totwenty-five percentage points. It was over ten percentagepoints in ten of the twenty-nine countries covered. For nineother countries the bias was in the other direction. The studyfound that the main factor behind the differences was notthe over-reporting of student numbers, but the misreportingof age. Household surveys actually report a greater numberof total students attending than administrative data whenall students, regardless of their age, are counted.

Age-specific reporting is the main source of discrepancy.Consider the case of Senegal — a country that illustratesthe broader pattern (Figure 2.10). For the 5 to 11 age group,administrative data reports an enrolment ratio consistentlyhigher than the attendance rate for the same age grouprecorded in household survey data. The discrepancy is verylarge. For age 8, the net enrolment ratio is reported to be 77%, compared with 58% for the net attendance rate. At about age 11, the reporting lines cross: at older ages,household survey data register higher levels of attendancethan net enrolment would imply.

The consistent pattern to emerge from the UIS study can bebriefly summarized. If the information in household surveys isaccurate, net attendance rates covering children in the official

primary school age range are lower than reported throughnet enrolment ratios based on administrative data. This wouldimply that there are more over-age children in school, andmore primary school age children out of school, than indicatedby official data.

Simple comparisons between administrative and householdsurvey data do not, however, provide a solid foundation forsuch conclusions to be drawn. There are many possible factorsbehind the discrepancies. One is demography. If nationalpopulation data over- or under-report the size of the primaryschool age population, net enrolment ratios drawn fromadministrative data will mirror the inaccuracy.

Uncertainty over the denominator (population size) can becompounded by education reporting systems. Registers maynot provide an accurate picture of the age of students, insteadtreating the class that children are in as a proxy for their age.In some cases institutional incentives may play a role: if thenumber of students in the appropriate grade for their agedetermines the allocation of grants or teachers, schools andlocal governments might have a tendency to inflate theschool register.

Household survey exercises have problems as well. Apart fromstandard sampling errors, the timing of surveys can have abearing on the results: for example, the data might be affectedby the agricultural calendar, drought or major external shocks.The reference period and phrasing of questions can also causecomplications. Household surveys examined in the analysispresented here ask whether children attended school at somepoint during the school year, not whether they are in school for the entire school year. Surveys can also systematically miss parts of the population that are difficult to reach.

All these considerations caution against drawing sweepingpolicy conclusions. The UIS emphasizes that its technical workcomparing household survey and administrative data does notprovide a basis for revising estimates of out-of-school children.What is clear is that important issues are at stake — and thatmore work is needed to clarify the real picture with respect

Box 2.5: Children count — but counting children in school is difficult

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

5 9

from the poorest 20% were over three timesmore likely to be out of school than children fromthe richest 20% in 2005 (Bruneforth, 2009b).

Rural children. Living in a rural area often putschildren at greater risk of being out of school. InBurkina Faso, Cameroon, Ethiopia, Malawi, theNiger, Senegal and Zambia, household survey datasuggest that rural children are more than twiceas likely not to be in school (Bruneforth, 2009b).

Many of those currently not in primary school willprobably never enrol. On the basis of past evidence

and the UIS model, an estimated 44% of out-of-school children are unlikely to make the transitioninto school (Figure 2.11). These 31 million childrenface the most acute disadvantages in education.The problem is most pronounced in sub-SaharanAfrica, where 59% of the out-of-school populationis unlikely to enrol. In South and West Asia, bycontrast, dropout is a more serious problem.More than 60% of the out-of-school populationhas dropped out, while one-third is unlikely everto enter. Almost half of the much smaller out-of-school population in the Arab States is unlikely toenter. In East Asia and the Pacific, the problem is

to the core issue in universal primary education: namely,how many children are really out of school. There is strongevidence that administrative data routinely overestimatenet enrolment by a considerable margin. In a separate reviewof the twenty-nine countries covered by the UIS study,this Report estimates that, if the household survey dataare accurate:

the out-of-school population in these countries wouldbe 66 million, rather than the 44 million reported inadministrative data;

the out-of-school population in India would be 16 millionhigher, more than twice the administrative data total;

in sub-Saharan Africa, Ethiopia and the United Republicof Tanzania would each have more than 1.8 millionadditional children out of school, Mozambique around600,000 and Uganda over 800,000.

Source: Bruneforth (2009a).

U. R. TanzaniaGuinea

MozambiqueBeninMali

EthiopiaUgandaMalawi

IndiaZambia

SenegalMadagascar

PeruRwanda

LiberiaBurkina Faso

NigerPakistan

PhilippinesZimbabweSwaziland

KenyaNigeriaGhana

CameroonD. R. Congo

NamibiaLesotho

Congo

-30 -20 -10 0 10 20 30 40

Rates are higherfor householdsurvey data

Rates are lowerfor householdsurvey data

% point difference between household surveyand administrative data

Figure 2.9: Different stories — administrative and householdmeasurement of children in school can differ greatlyDifferences between net enrolment ratio (administrative data) and net attendance rate (household surveys), selected countries,latest available year

Sources: Bruneforth (2009a).

0

10

20

30

40

50

60

70

80

90

Child

ren

in sc

hool

(%)

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Age in years

77%

58%

33%

51%

Household survey

Administrative data

Figure 2.10: In Senegal, estimates of children in school by age vary with data sourcesAge-specific enrolment and attendance rates, Senegal, 2006

Source: Bruneforth (2009a).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

6 0

overwhelmingly one of late entry, though close toone in five out-of-school children is unlikely evento enter school.

Country profiles mirror the regional differences(Figure 2.12). In four of the ten countries in sub-Saharan Africa with large out-of-school populations– Burkina Faso, Mali, the Niger and Senegal – morethan 70% of out-of-school children are expectednever to enrol.8 In Pakistan, almost half of the out-of-school population is unlikely to enrol. Thepattern is not restricted to low-income countries.One of the most striking results to emerge is theprofile of out-of-school children in Turkey, whereseven out of ten are unlikely to enter school.

For countries including Bangladesh, India, Indonesiaand Nepal, the big challenge is keeping childrenin school once they enrol. Identifying patterns ofexclusion are important for public policy design –the approaches needed to ensure that children notexpected to enrol have a chance to enter schoolare likely to differ from those addressing theconstraints facing children at risk of dropout.

How do the three markers for disadvantage –gender, wealth and location – shape prospectsthat out-of-school children will ever enrol?

Young girls face some of the highest barriers.Not only are they less likely than boys to be inschool, but those who are out of school are farmore likely than boys never to enter (Figure 2.13).In sub-Saharan Africa, almost 12 million girls areexpected never to enrol, compared with 7 millionboys. Countries in other regions face similarproblems. In Yemen, nearly 80% of out-of-schoolgirls are unlikely ever to enrol, compared with 36%of boys; in Pakistan the figures are 62% for girlsand 27% for boys. Gender disadvantages can cut inthe other direction: in Bangladesh, Brazil and SouthAfrica, it is more likely that boys will never enrol.However, it is clear that more rapid progress ingetting children into school will require measuresthat target the social, economic and culturalbarriers facing young girls.

Prospects for attending school are also heavilyconditioned by household location and wealth.Children from rural areas are at a particulardisadvantage (Figure 2.14). In Burkina Faso, ruralchildren are almost four times more likely than

8. Nigeria could be in a similar situation, but disaggregated data are not available.

In sub-SaharanAfrica, almost 12 million girls

are expectednever to enrol

World

Developing countriesDeveloped countries

Countries in transition

N. America/W. EuropeEast Asia/Pacific

Latin America/CaribbeanSouth and West Asia

Central AsiaArab States

Central/East. EuropeSub-Saharan Africa

0 20 40 60 80 100

% of total number of out-of-school children

Expected neverto enrol

Enrolled butdropped out

Expected toenrol late

Figure 2.11: Children in sub-Saharan Africa are the least likely to enter schoolDistribution of out-of-school children by school exposure, by region, 2007

Source: UIS (2009a).

NigeriaIndia

PakistanEthiopia

BangladeshNiger

PhilippinesBurkina FasoMozambique

GhanaYemen

BrazilKenya

MaliNepal

TurkeySouth Africa

IraqIndonesia

Senegal

0 2 000 4 000 6 000 8 000 10 000

Number of out-of-school children (000)

Expectednever to enrol

Enrolled butdropped out

Expectedto enrol late

Total

Figure 2.12: A child’s prospects of entering and staying in school vary by countryDistribution of out-of-school children by school exposure, selected countries, most recent year

Notes: Countries included had more than 500,000 children out of school in 2007 or the latest year available.For Iraq, Mozambique and Nigeria the breakdown is not available.Source: Bruneforth (2009b).

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

6 1

urban children to be out of school – and those notin classrooms are over four times less likely everto go to school. These disparities reflect some ofthe distinctive problems facing rural communities,including distance to school, poverty and genderdisadvantages.

Poverty strongly influences prospects for schoolentry. Children from the poorest 20% of householdsdominate the out-of-school populations in manycountries and are far less likely than higher-incomechildren ever to enrol. To take one example, aroundthree-quarters of children from the poorest 20% ofhouseholds in Ethiopia are not in school. Of these,over half are not expected to enter school(Figure 2.14). The heightened risk of never goingto school associated with low household wealthunderlines the importance of public policies toensure that poverty does not automatically leadto educational disadvantage.

Enrolment of school age childrenmoving too slowly

Getting children into school is just one of thestepping stones towards universal primaryeducation. As many children will drop out beforecompleting the primary cycle as are currently outof school. The critical challenge is not just gettingchildren into school but ensuring that, once there,they complete a good-quality education.

Universal primary education is easily identifiedafter the event. It exists when almost all primaryschool age children graduate at roughly the officialage. Measuring progress towards this goal ismore challenging. No single indicator providesthe complete picture, but a combination ofmeasures can help cast light on different partsof a complicated picture. Overall, there is clearevidence that school enrolment and completionare increasing across the world, but a narrowfocus on certain indicators may be leading to anunderestimation of the distance still to be travelledto achieve universal primary education.

Net enrolment ratios have been rising in the developing worldOne commonly used indicator, the net enrolmentratio, measures the proportion of students in theofficial primary school age group who are enrolledin school. In a system that has achieved universalprimary education, the vast majority of children inthe official age group will be in primary school.

Universal net enrolment, widely used as a measureof progress towards Goal 2, is a necessary but notsufficient condition for universal primarycompletion. Countries with a net enrolment ratioclose to 100% have most of their primary schoolage children in the school system, but the measuredoes not indicate where children are in the cycle.Some children may have dropped out and returnedto early grades, while others may be repeatinggrades having failed school tests.

The criticalchallenge is to ensure allchildren completea good-qualityeducation

MaliNiger

SenegalBurkina Faso

TurkeyYemenGhana

PakistanEthiopia

BangladeshIndia

NepalSouth Africa

Kenya

0 20 40 60 80 100

% of out-of-school children expected never to enrol

Boys Girls

Figure 2.13: Left behind: out-of-school girls are less likelyever to get into school% of out-of-school children who are expected never to enrol, by gender, selected countries, 2007

Notes: Countries included had more than 500,000 children out of school in 2007.Source: Bruneforth (2009b).

Expectedto enter late

Enrolled butdropped out

In school

Expectednever to enrol

Poorest 20% Richest 20%0

20

40

60

80

100

% o

f prim

ary s

choo

l age

child

ren

Ethiopia

0

20

40

60

80

100

Rural Urban

% o

f prim

ary s

choo

l age

child

ren

Burkina Faso

Figure 2.14: Poor and rural children have much less chance of going to schoolin Burkina Faso and EthiopiaSchool exposure of out-of-school children by location and wealth, Burkina Faso (2003) and Ethiopia (2005)

Source: Bruneforth (2009b).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

6 2

For all its limitations, the net enrolment ratio isuseful in providing an average picture of progressover time. That picture has been positive sincethe Dakar forum. Most developing countries thatstarted the current decade a long way fromuniversal primary enrolment have made significantstrides (Table 2.2). Since 1999, sub-Saharan Africaand South and West Asia have increased netenrolment ratios at five times and three timesthe rate of the 1990s, respectively, reaching 73%and 86% in 2007. However, regional aggregatesinevitably mask large intraregional differences.Sub-Saharan Africa has a particularly wide rangeof net enrolment ratios, from 31% in Liberia to 98%in Madagascar and the United Republic of Tanzania.In the Arab States, the spread extends from lessthan 45% in Djibouti to 75% in Yemen and over 95%in Bahrain and Egypt (Annex, Statistical Table 5).

Progress on enrolment has been unevenGlobal progress towards universal net enrolmentmasks a more complex picture. Countries aremoving forwards at different rates, some are notmoving – and others are moving backwards.

Figure 2.15 provides a summary progress report.Some countries have achieved extraordinaryadvances. The United Republic of Tanzania raisedits net enrolment ratio from around 50% in 1999

to 98% in 2006. Madagascar, Nicaragua and Zambiahave also broken through the 90% thresholdtowards universal primary enrolment. Benin startedout in 1999 with one of the world’s lowest netenrolment ratios and could now be on track foruniversal primary enrolment by 2015. As theeducation system expands, however, the challengeof extending opportunities to populations that arehard to reach will intensify (Box 2.6). Some of thecountries furthest from breaking through the 90%barrier towards universal net enrolment havenonetheless moved a long way, including BurkinaFaso, Ethiopia and the Niger.

Past net enrolment trends provide a limitedindicator of the potential for countries to achieveuniversal primary completion. As Figure 2.15demonstrates, very rapid progress on netenrolment is possible. However, countries withcurrent net enrolment ratios of less than 75% facevery steep challenges. Ensuring that all primaryschool age children progress through theeducation system is even more challenging,especially when schools are dealing with a largebacklog of over-age children.

Several countries in sub-Saharan Africa havetailored their ambitions to current circumstances.Burkina Faso’s original goal of attaining universal

Since 1999, sub-Saharan

Africa and Southand West Asia

have increasednet enrolment

ratios at fivetimes and three

times the rate of the 1990s,respectively

School year ending in School year ending in School year ending in

646 694 82 87 0.92 0.96

559 615 80 86 0.91 0.9570 66 97 96 1.00 1.0016 13 88 91 0.99 0.99

82 124 56 73 0.85 0.9035 41 78 84 0.87 0.90

7 6 88 92 0.99 0.98218 191 96 94 0.99 0.99214 188 96 94 0.99 0.99

3 3 90 84 0.97 0.97155 192 74 86 0.84 0.95

70 68 92 93 0.97 0.973 2 75 72 0.98 0.99

68 66 93 94 0.97 0.9653 51 97 95 1.01 1.0026 21 91 92 0.96 0.98

World

Developing countriesDeveloped countriesCountries in transition

Sub-Saharan AfricaArab States Central AsiaEast Asia and the Pacific

East AsiaPacific

South and West AsiaLatin America and the Caribbean

CaribbeanLatin America

North America and Western EuropeCentral and Eastern Europe

Table 2.2: Primary enrolment by region, 1999 and 2007

1999

(millions)

2007

1. Gender parity in primary education is measured by the gender parity index of gross enrolment ratios. See annex for details.Source: Annex, Statistical Table 5.

Total enrolment

1999 1999

(%) (F/M)

2007 2007

Net enrolment ratios Gender parity in primary1

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

6 3

primary enrolment by 2015 has been reviseddownwards in national plans to 70% (Bennell, 2009).Eritrea’s current education strategy neatlysummarizes the dilemma facing planners:‘achievement of universal primary education bythe global target date of 2015 would be extremelydifficult. Even if the financial resources were readilyavailable, it would be physically almost impossible

to provide the necessary infrastructure andassociated inputs (teachers, administrators, etc.)during the next eleven years to cater for allchildren of primary school age. … In view of this,it is projected that the net primary schoolenrolment ratio would reach 82% by 2015 andthat UPE would be achieved by 2019’ (EritreaMinistry of National Development, 2005, p.10).

20 30 40 50 60 70 80 90 10020

30

40

50

60

70

80

90

100

Net e

nrol

men

t rat

ios i

n pr

imar

y eud

catio

n in

200

7 (%

)

Net enrolment ratios in primary education in 1999 (%)

Net enrolment ratioshave progressedsince 1999

Net enrolment ratioshave deteriorated

since 1999Liberia

Eritrea

Niger Djibouti

Burkina Faso

MaliNigeria

Gambia Equat. GuineaCook Islands

Ethiopia

U. R. Tanzania Madagascar

Palestinian A. T.Oman

Togo

Guinea

Benin

Bhutan

MauritaniaNepal

Mozambique

Senegal

Yemen

Ghana Lesotho

Kenya

ZambiaNicaragua

Swaziland Lao PDR

Morocco

Cape VerdeSouth Africa

Malawi

Latvia

Dominican Rep.

Vanuatu

Botswana

S. Tome/Principe

CambodiaColombiaZimbabwe

EgyptGuatemala

Mongolia

Iran, Isl. Rep.Venezuela, B. R. Brazil

Hungary

Figure 2.15: Most countries improved their primary school enrolment between 1999 and 2007Change in net enrolment ratios in primary education, 1999–2007

Source: Annex, Statistical Table 5.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

6 4

Does the scaling down of ambitions markan unwarranted retreat from the politicalcommitments made at Dakar? Each country hasto assess what is achievable in the light of whereit currently stands, and the human and financialresources it has available. However, there isstrong evidence from several countries thatpolitical commitment allied to strong aidpartnerships can generate rapid progress.

Gender parity — some progress but a long way to goThe expansion of primary education has gonehand in hand with progress towards greatergender parity, but there are marked differencesacross and within regions, as witnessed by thegender parity index (GPI).

Twenty-eight countries had GPIs of less than 0.90 in2007; of these, eighteen are in sub-Saharan Africa.These countries have not yet achieved the goal ofgender parity in primary schooling, set for 2005.

There are also marked gender disparities in theArab States, though the largest gap is found in aSouth Asian country: Afghanistan, with just 63 girlsenrolled in school for every 100 boys. Large genderdisparities are inconsistent with sustained rapidprogress towards universal primary enrolment.

In countries at low levels of enrolment, such asBurkina Faso, Ethiopia and Yemen, moves towardsgender parity from a low starting point have helpedgenerate large increases in primary enrolment.The experience of Yemen demonstrates that rapidprogress towards gender parity from a low baseis possible and that sustained progress requiresa strong political commitment to equity (Box 2.7).Gender parity is usually inversely related toenrolment: the lower the enrolment, the greaterthe gender disparity (Figure 2.16). An exceptionis Senegal; while the country still has low netenrolment (72% in 2007), in the space of oneprimary school generation, the country has movedfrom a gender parity index of 86 girls per 100 boys

Twenty-eightcountries have

still not achievedthe 2005 goal

of gender parityin primaryschooling

Benin has been among the world’s fastest movingcountries on primary enrolment, with the netenrolment ratio rising from 50% in 1999 to 80%in 2007. The gender gap also narrowed, from just67 girls for every 100 boys in school in 1999 to83 girls in 2006. On current trends, Benin couldachieve universal primary enrolment by 2015.

Maintaining the trends will be difficult, however. As in other countries, rapid progress in scaling upenrolment has brought new policy challenges:

Raising completion rates. Achieving Benin’s goalof 100% primary school completion by 2015 willrequire far-reaching measures to ensure thatchildren enrol on time and complete a full primarycycle. Over-age entry remains a significant problem.The gross intake rate into the first grade is 115%,while in 2005 the net intake rate was less than50%. The disparity points to a concentration ofchildren over 6 years of age in the first grade.Getting children into school on time is importantfor increasing completion. Fewer than 20% of thosewho start school complete it at the correct age.

Addressing population pressures. With a populationgrowth rate of 3.2% and almost half of thepopulation under 15 years, Benin’s education systemwill need to expand just to stay in the sameposition.

Reducing regional disparities. There are markedinequalities across Benin. The gross intake ratefor the last grade of primary is only 36% in AliboriProvince (one of the poorest regions, withparticularly high levels of severe malnutrition forchildren under 5) compared with a national averageof 66%. Reaching vulnerable communities is vitalto sustained progress.

Tackling poverty. More than half of Benin’s ruralpopulation lives in extreme poverty. Children fromthe wealthiest quintile are at least twice as likely tocomplete the primary cycle as those in the poorestquintile. This has the effect of skewing educationfinancing towards children from the richest 20% ofhouseholds, who receive 57% of public expenditure on education compared with just 5% for the poorest.

The government has taken steps in its ten-yeareducation plan (2006–2015) to redress imbalances,including affirmative action for girls anddisadvantaged groups and regions — and strongbudget commitments. Education spending accountedfor 3.9% of GNP and 18% of budget spending in 2006.Just over half of the education budget is directed toprimary schooling. To ensure that Benin can go thefinal step towards universal primary education,international aid donors need to back up this nationalfinancing commitment.

Sources: World Bank (2009g); Benin Government (2008).

Box 2.6: Benin — on the right track, but tackling marginalization is a priority

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

6 5

in 1999 to an equal number of girls and boys in2007. However, not all progress towards genderparity has positive origins. In Equatorial Guinea,Liberia and Togo, greater parity has been driven notby expansion of the education system but by the factthat boys’ enrolment has declined (Figure 2.17).

With some of the world’s largest gender gaps,several countries in West Africa have adoptedpolicies aimed at strengthening parity as part ofthe wider strategy for achieving universal primaryeducation. Some of these policies focus onremoving one of the greatest obstacles to gender

equity: attitudes on girls’ and women’s place insociety. Working through village heads and religiousleaders, governments have mounted campaignsto communicate to parents the importance ofeducating daughters. Other strategies includepaying financial incentives, providing water andsanitation in schools (including separate latrinesfor boys and girls), recruiting female teachers andproviding incentives for their deployment to ruralareas, and giving teachers gender sensitizationtraining (UNESCO-IIEP, 2009). In remote ruralareas, distance to school is often a major securityconcern for parents of young girls. Governments

Senegal reachedgender parity in 2007 in thespace of oneprimary schoolgeneration

Figure 2.16: The relationship between enrolment and gender parity varies across countriesNet enrolment ratios and gender parity in primary education, 2007

Note: Gender parity in primary education is measured by the gender parity index of gross enrolment ratios. See annex for details.Source: Annex, Statistical Table 5.

0.60 0.70 0.80 0.90 1.00 1.1020

30

40

50

60

70

80

90

100

Net e

nrol

men

t rat

ios (

%)

Gender parity index

Genderparity

High

er en

rolm

ent

Liberia

Niger Djibouti

C. A. R.

Mali

Yemen

Pakistan

Eritrea

Burkina Faso

Nigeria

Iraq

Togo

Benin

Mozambique

Ethiopia

Guinea

Mauritania

Turks/Caicos Is

Gambia

Nepal

Nauru

Bangladesh

MalawiMorocco

Lao PDR

Congo

Timor-Leste

Equat. Guinea

Antigua/BarbudaGrenada

SenegalGhana

Oman

BurundiDominican Rep.

Cape Verde

SwazilandIndia

Cambodia

Girlsdisadvantaged

Boysdisadvantaged

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

6 6

have responded by attempting to bring classroomscloser to communities, often by building satelliteschools (see Chapter 3).

Aid donors can play an important role in supportingefforts to overcome gender disparity. In Chad, aUSAID-funded programme is addressing financialand cultural barriers to girls’ schooling by providingscholarships and backing community sensitizationcampaigns. Recognizing that attitudes cannot bechanged through top-down directives, theprogramme supports local agents for change,working through mothers’ associations, religiousfigures, local government and village leaders, andschool officials to promote girls’ education. The roleof imams in asserting the consistency of genderequality in education with the precepts of Islam hasbeen particularly important (Zekas et al., 2009).Initiatives such as these have helped make peoplemore aware that girls have a right to be educated.They also contributed to Chad’s progress between1999 and 2007 in narrowing the gender gap from58 girls per 100 boys to 70, with greater genderparity helping drive an overall increase inenrolment.

For many countries, sustained progress towardsgender parity will require advances on two fronts.Getting girls into school demands concerted actionto change attitudes and household labour practices.Keeping them in school once they reach pubertyposes another layer of challenges, especially incountries where early marriage is common andwhere girls’ disadvantage interacts with otheraspects of marginalization, such as poverty orethnicity. Countries including Bangladesh andCambodia have demonstrated that financialincentives can both increase the likelihood of girlsentering lower secondary school and raise demandfor primary schooling (Filmer and Schady, 2006;Fiszbein et al., 2009). However, public policyinterventions are required in many other areas ineducation and beyond.

In West Africa, some of the world’s poorestcountries with low enrolment ratios have shownthat political leadership and practical measurescan override gender discrimination in the household

Getting girls intoschool demands

concerted actionto change

attitudes andhousehold labour

practices

Figure 2.17: The gender gap is narrowing, but sometimes because enrolment is decliningChanges in net enrolment ratios and gender parity index of gross enrolment ratios in primary education, 1999–2007, selected countries

Note: Gender parity in primary education is measured by the gender parity index of gross enrolment ratios.See annex for details.Source: Annex, Statistical Table 5.

20

30

40

50

60

70

80

90

100

Equat. Guinea

Ethiopia

LesothoYemen

Togo

Benin

0.5 0.6 0.7 0.8 0.9 1.0 1.1

Gender parity index

Net e

nrol

men

t rat

ios (

%)

Genderparity

MaliLiberia

BurkinaFaso

Niger

Senegal

2007 The country has improved both participation and gender parity1999

2007 The country has improved gender parity but participation has deteriorated1999

In Yemen, one of the world’s poorest countries,enrolment increased from 2.3 million in 1999 to3.2 million in 2005 and gender disparities shrank.These achievements are all the more remarkablegiven Yemen’s deep poverty, rapid populationgrowth and dispersed rural population.

Girls have benefited from both the overall expansionof education and targeted interventions. Improvementin enrolment in recent years can be traced to policymeasures introduced in the late 1990s, including theuse of low-cost standardized school designs andconsultation with communities on school location.Basic education (grades 1 to 9) has been compulsoryand free in principle since the early 1990s, thoughlearners continued to pay for uniforms and textbooks.In 2006/2007 the Ministry of Education madeuniforms optional and eliminated textbook fees forgirls in grades 1 to 6 and for boys in grades 1 to 3.It has also taken measures to get more femaleteachers in rural schools.

Further progress will require policy measures thatweaken the interaction between gender inequalityand poverty. School attendance is lowest, and thegender gap widest, among the poor and in rural areas(Figure 2.18). Of the more than 900,000 primaryschool age children out of school in 2005, 70% weregirls and 88% lived in rural areas. Household surveydata show that only 28% of girls and 46% of boys in

Box 2.7: Yemen — making progress towards

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

6 7

and beyond. By the same token, failure to narrowgender gaps points to failure in these areas. Witha higher income and comparable net enrolmentratio, Pakistan lags far behind Senegal on genderparity. Pakistan’s primary net enrolment ratioin 2006 was 73% for boys but only 57% for girls.If Pakistan were to match Senegal’s performance,it would have 1.1 million more girls in school.Pakistan’s persistent gender disparities, whichmay be exacerbated by political movementshostile to girls’ education, are holding backoverall progress in enrolment (Box 2.8). Thethreat to gender equity is even more marked inneighbouring Afghanistan, where schools andteachers have been targeted with a view to drivinggirls out of school (see Chapter 3).

Gender disparity is not unidirectional. In a smallnumber of developing countries, girls’ enrolmentoutstrips that of boys. This may happen wheredemand for boys’ labour is higher. To take oneexample, poor rural families in Lesotho, particularly

those in highland areas, often rely on boys to herdcattle, with the result that dropout rates are highafter grade 3 (World Bank, 2005e). The positivenews is that the pace of increase in enrolment hasbeen faster for boys in recent years and genderparity has now been achieved.

Some countries are slipping or stagnatingPositive global trends on net primary enrolmentinevitably obscure negative national trends.Several countries with a long way to travel beforethey achieve universal primary enrolment arenot making progress – and some are registeringreversals.

Figure 2.15 shows that some countries with low net enrolment ratios and large out-of-school populations– notably Nigeria – are moving in the right direction,but at a snail’s pace. More disconcertingly, aroundtwenty-five developing countries with data for both1999 and 2007 experienced stagnating or decliningnet enrolment ratios.9

9. This trend is even moreapparent in 2006–2007, withforty countries that have yetto achieve universal primaryeducation not moving.

Between 1999 and 2007, aroundtwenty-fivedevelopingcountries haveexperiencedstagnating or declining netenrolment rates

the poorest quintile attended school. Such evidencepoints to parental attitudes and household labourpractices that attach less weight to girls’ educationthan that of boys.

Child labour patterns are also structured by genderdisparities. Poverty drives both boys and girls intoemployment, either because of household cash needsor because parents cannot afford education fees.Children of both sexes also spend time on householdchores. Around one-fifth of boys and one-quarter ofgirls are involved in child labour. However, while 70%of male child labourers attend school, only 52% offemales do. The disparity reflects longer work hoursamong girls, a division of labour that leaves girls withgreater responsibility for household labour and agreater weight attached to boys’ education.

The complex array of factors keeping children outof school in Yemen points to a need for a twin-trackresponse. Education policies can broaden schoolinfrastructure to reach more children and addressgender inequality through financial incentives,recruitment of female teachers and otherinterventions. At the same time, wider strategies areneeded to tackle rural poverty, curtail child labour andchallenge attitudes that devalue the education of girls.

Sources: Al-Mekhlafy (2008); Guarcello et al. (2006); Integrated Regional Information Networks (2006, 2007); Kefaya (2007); Ochse (2008).

universal primary education and gender parity

Figure 2.18: In Yemen, girls’ enrolment is lowest in the poorest and rural areasPrimary net attendance rates in Yemen, by gender, wealth and location, 2005

Source: UNESCO-DME (2009).

0

10

20

30

40

50

60

70

80

Net a

ttend

ance

rate

s in

prim

ary e

duca

tion

(%)

Poorest20%

Q2 Q3 Q4 Richest20%

Rural Urban

By wealth quintile By location

National averageBoys Girls

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

6 8

In Pakistan, only around

one-third of poorgirls are in school

Pakistan is off track for achieving universal primaryeducation by 2015. The country accounts for asignificant share of the global out-of-school problem.Failure to tackle gender disadvantages that intersectwith poverty and regional differences is at the heartof the problem.

Deep disparities based on location and wealth area feature of education in Pakistan. In the richesthouseholds, over 85% of children go to primaryschool, with little difference between boys and girls.Attendance rates for children from poor householdsare far lower, especially for females: only around one-third of poor girls are in school. Similarly, attendanceis higher and the gender gap smaller in urban areasthan in rural ones, and in the relatively wealthy Punjabprovince than in Balochistan and Sindh (Figure 2.19).

The North West Frontier Province stands out ashaving above average attendance for boys but wellbelow average attendance for girls. There is growingconcern that this gender gap could be wideningfurther. In the Taliban-occupied parts of the province,91 girls’ schools have been destroyed and25 damaged, with some boys’ schools also suffering.

The factors behind Pakistan’s deep gender disparitieshave been extensively researched. Distance to schoolmatters far more for girls than boys, reflectingsecurity concerns and household labour demands.

Girls’ enrolment drops off sharply with each 500-metre increase in distance from the closestschool admitting girls and this ‘distance penalty’accounts for 60% of the gender gap in enrolments.Cost factors can also disadvantage girls becausehouseholds tend to spend more on boys.

The presence of a government school in thecommunity has a significant positive effect on girls’enrolment. As there has been a marked trend towardssex-segregated primary education, the absence insome areas of all-girl government schools hasemerged as a major constraint on girls’ schooling.Insufficient recruitment of female teachers is anotherconstraint. Rural parents strongly prefer to have girlseducated by women, but the legacy of low investmentin girls’ education means few local women haveappropriate qualifications. It is also difficult to attractqualified female teachers to rural areas from otherparts of the country.

Education policy documents increasingly recognizethat more weight has to be attached to gender equity,but it is far from clear that the current policyframework provides concrete measures for translatingstatements into action. Policies indicate communityneeds as criteria for the location of new governmentprimary schools, for example; however, researchsuggests that community economic status and theextent of gender disparity have had little influence

over the placement of newgovernment schools.

Overall levels of publicfinancing remain low,education is weakly integratedinto national povertyreduction strategies and therehave been limited attempts tointroduce the type ofincentives for girls’ educationthat have been successful inBangladesh, which has movedfar ahead of Pakistan in termsof enrolment and genderparity.

Sources: Aly and National EducationPolicy Review Team (2007); Andrabiet al. (2008); Lloyd et al. (2007);O’Malley (2009); Pakistan Ministry of Education (2003).

Box 2.8: Pakistan — gender disparities hold back progress

Net a

ttend

ance

rate

s in

prim

ary e

duca

tion

(%)

By wealth quintile By location By region

30

40

50

60

70

80

90

PunjabNWFP*SindhBalochistanUrbanRuralRichest20%

Poorest20%

National averageBoys Girls

Figure 2.19: Pakistan’s primary school attendance is marked by gender, regional and wealth inequalitiesPrimary net attendance rates in Pakistan by gender, wealth, location and region, 2007

* NWFP: North West Frontier Province.Source: UNESCO-DME (2009).

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

6 9

Adding in countries that lack enrolment data forboth 1999 and 2007 would yield significantly morethan twenty-five lagging countries. For example,for the Central African Republic and Pakistan,which had net enrolment ratios below 70% in 2007and no data for 1999, there is strong evidence thatprogress has been limited. Other countries haveno data available on net enrolment for either 1999or 2007, including Afghanistan, the DemocraticRepublic of the Congo, Haiti, Sierra Leone, Somaliaand the Sudan. Here, too, there is strong evidenceto suggest that progress towards universal netenrolment, if any, has often been very slow, from alow base. Côte d’Ivoire, with a gross enrolment ratioof 72%, and the Sudan at 66% in 2007 are clearlyoff track.10 Although Afghanistan’s gross enrolmentratio has increased significantly (from 28% in 1999to 103% in 2007), in part due to the opening upof opportunities for girls’ education, there is stilla long way to go before all children enter andcomplete the cycle on time.11

Many countries experiencing slow progress orreversals are either in the midst of or recoveringfrom conflict. Developments in Liberia have beenparticularly disconcerting. After a brutal civil war,the country has now enjoyed several years ofpeace and has an elected president, Ellen Sirleaf-Johnson, with a strong commitment to education.However, its net enrolment ratio slipped from 42%in 1999 to 31% in 2008. The government plan foreducation acknowledges that ‘realistically, Liberiawill likely need more years beyond 2015 to achievethe UPE goals’ (Liberia Ministry of Education,2007b). How many more years will depend partlyon national efforts and partly on the degree towhich aid donors find innovative ways of supportingthose efforts (Box 2.9; see also Chapter 4). Thelarge recorded decline in enrolment in thePalestinian Autonomous Territories would alsoappear to be linked to the combined effects ofcivil conflict, military incursions, and restrictionson the movement of goods and people.

Eritrea is a further cause for concern. Aftersignificant progress increasing enrolment from1999 to 2006, the country experienced a reversalin 2007.12 Military tensions appear to be acontributory factor. Since the end of the 1990s,spending on education has more than halved asa share of GNP, from 5.3% in 1999 to 2.4% in 2006.Meanwhile, military spending has been extremelyhigh,13 crowding out urgently needed spendingon education infrastructure.

Going the final mile — some countries with high net enrolment face problemsMost of the countries facing difficulties inachieving universal net enrolment by 2015 havetwo characteristics in common. They started withlow initial enrolment ratios and they are very poor.There are exceptions to the rule. While enrolmentratios tend to rise with wealth, there are largevariations around the average – and some relativelywealthy countries perform worse than might beexpected. Moreover, some countries are in gravedanger of failing to achieve universal net enrolmentby 2015 despite having started at very high levelsof school participation.

Figure 2.20 demonstrates that wealth mattersfor education coverage. It charts the relationshipbetween average income and net enrolment ratios,

10. Because the grossenrolment ratio measuresthe enrolment of all childrenirrespective of their agerelative to the primary schoolage group, the net enrolmentratio would be far lower.

11. In 2007, the net intakerate into the primary systemin Afghanistan was just 55%.

12. The net enrolment ratiorose from 33% in 1999 to47% in 2006, but fell back to 41% in 2007.

13. Military expenditure isaround one-quarter of GDP,according to the 2007/2008budget (UNDP, 2007).

Autocratic rule, coups and fourteen years of civil war took a devastating toll on Liberia’s education system. Schools were destroyed, public servicescollapsed, investment fell and parental fears over security led to childrenbeing withdrawn from school. The election of President Ellen Sirleaf-Johnsonin 2006 created renewed hope, but recovery is proving arduous.

Liberia is one of the world’s poorest countries: three-quarters of thepopulation survives on less than US$1.25 a day. Education infrastructure isdilapidated and there are chronic shortages of trained teachers and teachingmaterials. As well as dealing with children who have enrolled since the end of the conflict, the education system must cope with population growth and the many displaced Liberian families returning from abroad.

Data limitations make it difficult to chart developments, but fragile gains inenrolment at the end of the 1990s are thought to have been reversed duringa renewal of violence from 2001 to 2003, with enrolment dropping by abouthalf for girls and one-third for boys because of insecurity and poverty.

In 2007, the Ministry of Education set out a strategy for moving from short-term emergency planning to long-term strategic planning. The strategy envisages strengthening quality and equity, in part by providing a regulatory umbrella that covers the diversity of education providers. In 2008, some 30% of primary enrolment was in private and mission schools,the rest in government and community-funded schools. The equity challengeis particularly daunting, given the large inequalities based on wealth, regionand gender.

Liberia’s experience raises wider concerns about the failure of aid systems.In countries recovering from conflict, the resources available to governmentare limited, so aid has a vital role to play. Aid donors were slow to supportreconstruction in Liberia, despite the endorsement of the country’s economicplan by the Fast Track Initiative. Chapter 4 explores the failure of current aidsystems to respond to the needs of countries such as Liberia.

Sources: Center for Global Development (2009); Liberia Ministry of Education (2007a); USAID (2007).

Box 2.9: Liberia — slipping back in a post-conflict country

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

7 0

while at the same time capturing the size of out-of-school populations. There are many reasons for theunderlying positive relationship between enrolmentand income. As countries grow wealthier, they andtheir citizens can spend more on education – and aseconomies grow they tend to generate demand forskilled labour. Of greater interest than the well-established average association is variation aroundthe mean. At the lower end of the enrolmentspectrum, countries such as Nigeria and Pakistanare outperformed by poorer countries such asBangladesh and Ethiopia. At the higher end,countries including the Philippines, South Africaand Turkey perform less strongly than expected.

The Philippines provides a particularly strikingexample of underperformance.14 With an averageincome four times that of the United Republic ofTanzania or Zambia, it has a lower net enrolmentratio. The unfavourable comparisons do not endthere. Whereas the United Republic of Tanzania

and Zambia have been steadily increasing netenrolment ratios, the Philippines has stagnated.Given the country’s starting point in 1999, achievinguniversal primary education by 2015 should havebeen a formality. There is now a real danger that,in the absence of decisive political leadership, thecountry will miss the goal. In 2007, out-of-schoolnumbers for children aged 6 to 11 broke throughthe 1 million mark and there were over 100,000more children out of school then than in 1999.Around one-quarter of those entering school dropout before grade 5. Other countries experiencingstagnation or slippage from high levels of netenrolment include Turkey, whose net enrolmentratio has remained unchanged since the beginningof the decade (UIS database).

Why have countries that were so close to universalnet enrolment at the end of the 1990s failed to gothe extra mile? One factor is the difficulty inextending opportunities to certain regions and parts

14. See Chapter 3for a fuller analysisof the reasons forthe challenges facingthe Philippines.

The Philippines and Turkey

perform lessstrongly thantheir national

wealth predicts

Figure 2.20: Most out-of-school children are in poorer countries, but some wealthier countries are underperformingPrimary net enrolment ratios, GNP per capita and out-of-school children headcount in low and middle income countries, 2007

Notes: Bubble size represents the number of out-of-school children. Countries in red have more than 500,000 children out of school.Source: Annex, Statistical Tables 1 and 5.

0 2 000 4 000 6 000 8 000 10 000 12 000 14 0000

20

40

60

80

100

C. A. R.

Eritrea

Liberia

Djibouti

Egypt

Colombia

Iran, Isl. Rep.Thailand

Net e

nrol

men

t rat

ios i

n pr

imar

y edu

catio

n (%

)

GNP per capita (PPP US$ 2007)

5 million out-of-school children

1 million out-of-school children

500,000 out-of-school children

TurkeyBrazil

Niger

Mali

Burkina Faso

Bangladesh

Ghana

Nigeria

Yemen

Pakistan

India

Indonesia

U. R. TanzaniaTajikistan

Dominican Rep.

Ethiopia

Philippines

South Africa

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

7 1

of society. Both the Philippines and Turkey faceproblems of deeply entrenched marginalization.In the Philippines, marginalization is stronglyassociated with poverty and location, with theAutonomous Region in Muslim Mindanao and someoutlying islands falling far behind. In Turkey,disadvantage is heavily concentrated among younggirls in eastern regions who do not have Turkish astheir mother tongue (Box 2.10). Chapter 3 exploresthe problem of reaching marginalized people in

more detail, but it is evident in the cases of thePhilippines and Turkey that current policies are not breaking down inherited disadvantage. Onecontributory factor is the low share of nationalincome invested in education. Turkey investedaround 4% of GNP in 2004, compared with 6%to 7% in Morocco and Tunisia. The figure wasjust 2.3% in the Philippines in 2005, comparedwith an East Asian regional average of 3.6%.15

Turkey’s advance towards universal primary educationhas stalled within touching distance of the goal. Muchhas been achieved over the past decade. But far more hasto be done to break down inequalities based on gender,region and wealth.

The country’s basic education law requires every childto undergo eight years of schooling and there is a singlecurriculum for all 6- to 14-year-olds. Primary schoolenrolment increased rapidly during the second half ofthe 1990s as a series of programmes expanded schoolconstruction, strengthened teacher training, increasedtextbook supplies and provided transport for childrenin remote villages.

Since 2000, however, progress has slowed. Enrolment ratioshave stagnated at around 90% since 2002 — far below thelevel predicted on the basis of Turkey’s average income.Some 640,000 children of primary school age were out ofschool in 2007. Around 60% were girls, pointing to deeplyentrenched gender inequalities. High levels of exclusion inthe early years are holding back progress at higher levels,with adverse consequences for future economic growth,employment and social development. Education quality isanother serious source of concern: Turkey figures among theworst performers on the learning achievement tests of the OECD Programme for International Student Assessment (PISA).

Turkey’s experience powerfully demonstrates the difficultiesgovernments face as they attempt to reach the mostmarginalized. One study using Turkey’s most recentDemographic and Health Survey highlights deep, overlappingand mutually reinforcing inequalities in opportunity foreducation, with gender disparities magnifying other gaps:

Gender. Between ages 8 and 12, 7% of girls never makeit to school, compared with 2% of boys. By age 15, femaleenrolment is almost twenty percentage points belowmale enrolment.

Region. The eastern region lags far behind the restof the country, mainly because of gender disparity.Enrolment ratios for girls in eastern Turkey, expressedas a share of the level for boys, peak at 85% at age 9and have dropped below 40% by age 15.

Rural location. Being born in a rural area isdisadvantageous for girls across the country. Outside ofthe eastern region, that disadvantage kicks in from age 13.In the eastern region it starts early: by age 15, fewer than20% of rural girls are enrolled.

Household wealth and other factors. Children in householdsthat are poor and whose parents have limited formaleducation are less likely to progress through the schoolsystem. Children in the wealthiest 20% of households arefive times more likely to reach higher education thantheir counterparts in the poorest 20%. The strength ofthe negative correlation between household circumstanceand education in Turkey is magnified by gender effects.For example, at age 16 boys of mothers with no educationare twice as likely as girls to be in school.

Such findings powerfully illustrate the distance Turkey stillhas to travel to make the right to education a reality forall of its citizens. As the authors of the research put it, theopportunity profiles that emerge from household surveysshow that ‘school enrolment in Turkey is evidently notindependent from circumstances at birth’.

Patterns of inequality in education raise concernsfor the future course of Turkey’s social and economicdevelopment. High levels of education inequality are holdingback efforts to strengthen economic growth, expandemployment and create a more equal society. Migrationfrom eastern to western regions, usually from rural to urbansettlements, spreads the legacy of education disadvantageacross the country. Large numbers of rural migrants toTurkish cities settle in squatter areas called gecekondulardistricts, which are centres of social marginalization andeducational disadvantage.

The scale of inequality also highlights the importanceof equity in public spending. It is critical to strengthenstrategies and incentives for reaching rural girls, especially — though not exclusively — in the eastern region. Addressingthe disadvantages faced by children of parents who do notspeak Turkish as a home language is another priority area.

Sources: Duman (2008); Ferreira and Gignoux (2008a); Otaran et al. (2003).

Box 2.10: Turkey — marginalization keeps universal primary education out of reach

15. The regional figure is the median for thosecountries with available data.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

7 2

From enrolment to completion and beyond — a difficult journey that is hard to measure

Universal primary education is an apparentlysimple goal that raises disarmingly complexquestions over measurement. Going back to firstprinciples, that goal is about all children enteringschool at an appropriate age, progressingsmoothly through the system and completing a full cycle.

For millions of children entering primary school,the journey through the system is often delayed,hazardous and short-lived. In half the countriesin South and West Asia, and sub-Saharan Africa,almost one child in three enrolling in school dropsout before completion. Even that figure, stark as itis, understates the problem: many children do notget past the first hurdle. In 2006, 13% of pupils inSouth and West Asia and 9% in sub-Saharan Africadropped out before completing the first grade(see annex, Statistical Table 7).16 Malawi andUganda have relatively high net enrolment ratios,yet between one-quarter and one-third of pupilsdrop out during the first grade, in some cases

never to return. Repetition of grades is alsocommon. In Burundi, nearly one-third of childrenin primary school in 2006 were repeating grades.

Charting progress towards universal primaryeducation in school systems marked by highlevels of late entry, dropout and grade repetitionis a challenging exercise. The tool kits used bygovernments and the international communitycomprise a range of instruments for measuringintake, grade progression and completion. Eachinstrument provides important information. Yetthey provide only a partial and in some casesinconsistent insight to where countries are onthe road to universal primary education.

Figure 2.21 illustrates the point. It looks at twoof the most widely used measures of progresstowards universal primary education. The first isthe gross intake rate into the last grade of primaryschool, which expresses the share of childrenentering the last grade as a proportion of theofficial age group for that grade. It includes over-age children who started school late or repeatedgrades. The second measure is the net enrolmentratio, discussed earlier. It provides information on

16. These figures areregional medians for thecountries that have therelevant data available.

In half thecountries

in South and West Asia,

and sub-SaharanAfrica, almost

one child in threedrops out

Liber

iaEr

itrea

Cong

oTim

or-L

este

Gam

bia

Ghan

aLe

soth

oPa

lest

inia

n A.

T.Do

min

ican

Rep.

Turk

eyBo

livia

Colo

mbi

aEg

ypt

Iran,

Isl.

Rep.

Tuni

siaTh

aila

ndIn

done

sia Peru

Ecua

dor

El S

alva

dor

Zam

bia

Cam

bodi

aIn

dia

Angu

illa

Mor

occo

U. R

. Tan

zani

aGu

atem

ala

Sout

h Af

rica

Nica

ragu

aNa

mib

iaLa

o PD

RSw

azila

ndM

adag

asca

rNe

pal

Beni

nBa

ngla

desh

Mal

awi

Mau

ritan

iaGu

inea

Togo

Pakis

tan

Sene

gal

Buru

ndi

Moz

ambi

que

Ethi

opia

Mal

iBu

rkin

a Fa

soC.

A. R

.Ni

ger0

20

40

60

80

100

120

(%)

Net enrolment ratio in primary educationGross intake rate into last grade of primary

Gross intake rate into last grade of primary is higherthan net enrolment ratio (indicates late entry, high levels

of repetition and/or re-entry following drop out)

Gross intake rate into last grade of primary is lower than net enrolment ratio(indicates high levels of drop out in the early grades)

Figure 2.21: Children’s precarious pathway from school entry to completionNet enrolment ratios and gross intake rates to last grade, selected countries, 2007

Source: Annex, Statistical Table 5 and UIS database.

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

7 3

the number of primary school age children in thesystem, but not on where they are in the system,on how many started school over the official age oron the level of grade repetition. Both measuresprovide important but partial information – and therelationship between the two is highly variable.Countries with gross intake rates that are higherthan net enrolment ratios (those on the left of thefigure) are characterized by high levels of over-age

entry to last grade. Those in the opposite position –Burundi is an example – are characterized by lowlevels of internal efficiency. But neither measureoffers more than a partial insight into how near acountry is to achieving universal primary educationor how far it may be from that goal.

Cohort tracking can provide a more integratedperspective. Figure 2.22 illustrates one possible

Cohort trackingprovides anintegratedperspective onprogress towardsuniversal primaryeducation

Primaryschool entryage pupils

100

Cohort thatenters at thecorrect age

67

Net cohortsurvival to

grade 5

32

Net cohortcompletion

rate

27

Net intake rateinto first grade

of primary (67%)

Survival to grade 5(47% of the cohort

entered)Primary completion

rate (40% of thecohort entered)

Figure 2.22: Children who start primary school have varying chances to complete the last gradeNet intake rates into first grade of primary through to net cohort completion rates, selected countries, 2006

* *

Eritr

ea

Burk

ina

Faso

Bots

wan

a

Mau

ritan

ia

Togo

Nige

r

Beni

n

Moz

ambi

que

Buru

ndi

Kuw

ait

Sene

gal

El S

alva

dor

Mal

awi

Lao

PDR

Nica

ragu

a

Brun

ei D

arus

s.

Cape

Ver

de

Guat

emal

a

Geor

gia

Ukra

ine

Angu

illa

Mon

golia

Arub

a

Mad

agas

car

Cam

bodi

a

Iraq

Mor

occo

Alge

ria

Pana

ma

Pakis

tan

Ecua

dor

Mau

ritiu

s

Tajik

istan

0

10

20

30

40

50

60

70

80

90

100

(%)

Net intake rates

Net cohort survival to grade 5

Net cohort completion rates

In Nicaragua, the net intake rate into first grade was 67% in 2006. The survival rate to grade 5 was47% and the primary completion rate 40%.These observed rates allow us to estimate the prospects

of a cohort of pupils aged 6 (the primary school starting age) completing the six-year cycle.If repetition and dropout rates remain unchanged, of 100 pupils aged 6, 67 will enter the

first grade of primary school at the correct age. Of these, 32 will survive until grade 5,and 27 will graduate from the final grade.

Notes: The lines for each country illustrate the prospects for a cohort of 100 children of primary school entry age completing the cycle if the education system remains in its current state (taking account of current rates of repetition and dropout). Ideally, all children should enter school at the official starting age. The net intake rate is therefore used as the entry point. The net cohort survival rate to grade 5 and the net cohort completion rate are obtained by multiplying the net intake rate by, respectively, the survival rate to grade 5 and the primary cohort completion rate. All countries with available data are included.* Countries whose primary education cycle is less than five years.Source: Global Monitoring Report team calculations based on Statistical Tables 4 and 7 (annex).

Net cohort completion rates, the example of Nicaragua

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

7 4

approach. Starting from the proportion ofchildren entering school at the official age, ituses administrative data to track their progressto grade 5 and, unlike the gross intake rate forthe last grade, subsequent completion. Forcountries seeking to make the transition fromschool systems characterized by late entry,grade repetition and low completion to a moreregular cycle consistent with progress towardsuniversal primary education, the net cohortcompletion rate is a potentially usefulmeasurement tool.

One advantage of cohort tracking is that it providesa credible measure of distance from universalprimary education. In the case of sub-SaharanAfrica, it underlines the daunting scale of thechallenge ahead. While intake rates are going up,delayed entry is endemic. Half of all countries inthe region had 50% or more children enteringschool later than the official starting age in 2007.Assuming a five-year school cycle, this impliesthat governments in the region would have todouble the net intake rate by 2010 to makeuniversal primary entry possible by 2015. Forsome countries, the challenge is to raise the netintake rate while building on a strong but limitedcompletion record. In Burkina Faso, most childrenentering school at the appropriate age progressthrough to completion – but the net intake ratein 2006 was just 27%. Conversely, Malawi and

Nicaragua have net intake rates over 60% in 2006where fewer than half the official age entrantsmake it through to completion.

Out-of-school adolescentsThe focus on out-of-school children of primaryschool age has deflected attention from a far widerproblem. Millions in the lower secondary school agegroup are also out of school, either because theyhave not completed primary school or could notmake the transition to lower secondary school.

Recent data analysis suggests that nearly 71 millionadolescents were out of school in 2007 – almostone in five of the total age group (Table 2.3).17

Viewed through this wider lens, the out-of-schoolproblem is twice as large as it is typically reportedto be. The problem is most widespread in sub-Saharan Africa, with 38% of adolescents out ofschool, and South and West Asia with 28%. As withprimary school age children, adolescent girls aremore likely than boys to be out of school. Globally,54% of out-of-school adolescents in 2007 weregirls. In the Arab States the figure was 59%(Bruneforth and Wallet, 2009).

Equally disconcerting is the fact that manyadolescents in school are still enrolled at theprimary level (Figure 2.23). This is the case for 39% of lower secondary school age adolescentsin sub-Saharan Africa, for example.

17. For the purposes ofthis analysis, adolescenceis defined in terms of theofficial lower secondaryschool age range.Although the range variesby country, it is typicallyshorter than that forprimary school. The lowersecondary cycle is usuallytwo to four years,compared with five toseven years of primaryschooling in mostcountries.

Nearly 71 millionadolescents were out of

school in 2007,54% were girls

71 791 11 70 921 18 142 712 14

68 638 12 68 197 21 136 835 152 334 4 1 538 4 3 872 4

819 6 1 187 6 2 006 6

32 226 26 21 731 38 53 957 305 752 14 4 009 18 9 761 15

271 5 302 4 573 49 039 5 10 319 10 19 358 7

18 031 10 29 905 28 47 937 172 989 5 1 885 5 4 873 51 931 4 1 319 4 3 250 41 552 7 1 452 7 3 004 7

World

Developing countriesDeveloped countriesCountries in transition

Sub-Saharan AfricaArab StatesCentral AsiaEast Asia and the PacificSouth and West AsiaLatin America and the CaribbeanNorth America and Western EuropeCentral and Eastern Europe

Table 2.3: Number and % of children and adolescents of primary, lower secondary or basic education age not enrolled in primary, secondary or higher education, 2007

Totalout-of-school

(000) (000) (000)

As % of theprimary age

groupTotal

out-of-schoolTotal

out-of-school

As % of the lower secondary

age group

As % of the basic education

age group

Source: Bruneforth and Wallet (2009).

Primary educationLower secondary

education

Basic education(primary and lower secondary

combined)

P R O G R E S S T O WA R D S T H E E FA G O A L S

U n i v e r s a l p r i m a r y e d u c a t i o n

7 5

The transition from primary school to lowersecondary school is hazardous for many children.Problems that may be evident at the primary levelare often magnified at the secondary level. Cost,distance to school, labour market demand and –especially for girls – deeply engrained social,cultural and economic barriers figure prominently(Otieno and K’Oliech, 2007). Because secondaryschools are often further from home, theimportance of distance as a barrier to entryincreases. This is especially true for poorhouseholds facing labour shortages and forchildren in rural areas (Mingat and Ndem, 2008).In Mauritania and Senegal, the average journeytime to the closest secondary school is eightyminutes in rural areas. The average distance tothe closest lower secondary school in Senegalis twenty-five times farther than to the nearestprimary school (Glick and Sahn, 2009). Distancecan compound the effects of poverty, with poorhouseholds often unable to cover the cost eitherof transport or of boarding school places. Girlsface a distinctive set of barriers: longer distancesmay reinforce security concerns and, in somecontexts, early marriage prevents them fromprogressing beyond primary school.

The transition to lower secondary school is nowat the centre of the Education for All agenda inmany countries. In sub-Saharan Africa, universalbasic education is an increasingly prominent policygoal. For example, Ghana has adopted a basiceducation cycle embracing six years of primaryand three years of lower secondary; in Zambiathe cycle is seven years of primary and two yearsof lower secondary.

There are good reasons for the shift in emphasistowards a longer basic education cycle. As morechildren get into and progress through primaryschool, demand for secondary school places isgrowing. There is also evidence of high social andprivate returns to education beyond the primarylevel. Yet governments also face tough choices.In countries that have been unable to deliveraffordable, good-quality basic education to largesections of the population, the shift in emphasisraises important questions for equity in publicfinance. Aid partnerships can help relieve thefinancing constraints. However, it is important forgovernments and donors to avoid a premature shiftin policy priorities. With millions of children stillexcluded from primary education and the worldoff track for the 2015 goals, there is a great dealof unfinished business awaiting urgent attention.

Conclusion

As in previous years, the progress report onuniversal primary education is a story of ‘glasshalf empty, glass half full’. Much has beenachieved – but the international community hasa long way to go if it is to deliver on the promisesmade in Dakar and in the Millennium DevelopmentGoals. The slow-down in getting children intoschool since 2004 is a particular concern. Anotheris the evidence of a large mismatch betweenadministrative data on school enrolment andhousehold survey data on school attendance.The out-of-school problem may be far biggerthan has previously been assumed, pointingto a need for an urgent policy response at boththe national and international levels.

The transition from primaryschool to lowersecondary schoolis hazardous formany children

Figure 2.23: Many adolescents are out of school, or still in primary schoolDistribution of lower secondary school age children by education level and % out of school

0

20

40

60

80

100

% o

f low

er se

cond

ary s

choo

l age

pop

ulat

ion

LatinAmerica/Caribbean

East Asia/Pacific

ArabStates

Developingcountries

Sub-SaharanAfrica

South/West Asia

Out of schoolIn primaryeducation

In secondary educationor higher

60

20

2123

72

5

23

39

3828

12

6075

15

10

64

18

18

Source: Bruneforth and Wallet (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

7 6

Youth and adult skills —expanding opportunitiesin the new global economy

Goal 3. Ensuring that the learning needs of all young people and adults are met throughequitable access to appropriate learning and life-skills programmes.

The Dakar Framework for Action does not providetargets for youth and adult skills. Rather thanagree to quantifiable benchmarks, governmentssigned up to a third EFA goal that amounts to avague aspiration. One consequence has been aprotracted and unresolved debate over what, ifanything, that aspiration means in terms of policycommitments (King and Palmer, 2008). Unlike otherparts of the Dakar Framework, goal 3 has been thesubject of quiet neglect. It has been conspicuous byits absence not just from the agendas of high-leveldevelopment summits but also from the campaignsof non-government organizations.

That situation is unfortunate. In the emergingknowledge-based global economy of the twenty-first century, learning and skills play an increasinglyimportant role in shaping prospects for economicgrowth, shared prosperity and poverty reduction(Sapir, 2005). A country’s most important resourceis not its raw materials or its geographical locationbut the skills of its people. Countries that fail tonurture these skills through effective learningface a bleak future, with human capital deficitshindering economic growth, employment creationand social progress (Commission on Growthand Development, 2008; Kok, 2004; OECD, 2004).Within countries, unequal access to opportunitiesto develop skills will be reflected in deepeningsocial and economic disparities. Youthunemployment, one of the most serious andpersistent challenges facing governments acrossthe world, is in part a reflection of a misalignmentbetween skills development and the economy.As one recent report put it: ‘Achieving worldclass skills is the key to achieving economicsuccess and social justice in the new globaleconomy’ (Leitch Review of Skills, 2006, p. 9).

Rich and poor countries alike increasinglyrecognize that they will pay a high price if they failto strengthen national skills (DFID, 2007, 2008a).The global economic crisis has raised the stakes,

pushing learning and skills up the politicalagenda. While all sections of society have beenaffected, the economic downturn has left itsdeepest imprint on vulnerable unskilled workers,especially the young (ILO, 2009a; OECD, 2009d).Governments across the world are grappling withthe twin challenges of providing immediate supportto the vulnerable during a period of turbulencewhile equipping people with the skills they needto re-enter labour markets with higher levelsof productivity.

This section looks at some of the lessons tobe drawn from current approaches to skillsdevelopment. Narrowing the wide-angle lensof goal 3, the focus is on skills and learningopportunities for young people provided throughtechnical and vocational education programmes.These programmes can play an important rolein strengthening the transition from school tothe world of work, in offering second chances andin combating marginalization. In many countries,however, technical and vocational education isin such bad shape that it merits its reputationas a form of second-class schooling. There areno quick fixes for this situation, but four broadmessages emerge from this section:

Give young people the training they need.Governments, trade unions and employersneed to cooperate to devise effective technicaland vocational education that equips youngpeople with the skills they need for successin employment. Too much vocational educationdelivers skills of limited relevance to economicand social needs, and at high cost, oftenbypassing the poor and the informal sector.It is often driven by inappropriate curriculaand qualifications, with providers insulatedfrom employers’ real demands.

Skills need a broad base. Successful transitionfrom school to work requires the developmentof broad skills, with an emphasis on problem-solving and ‘learning to learn’, alongsidemore specialized abilities. Early streaminginto specialized vocational educationthrough academic selection should be avoided.Governments also need to address thewidespread perception of technical andvocational education as a safety net for failingstudents or those from poor family backgrounds.Raising the quality and improving the relevanceof technical and vocational education is themost effective antidote to that perception.

The globaleconomic crisis

has pushedlearning andskills up the

political agenda

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

7 7

Strengthen basic education. Effective andequitable skills will not be developed incountries where a majority of the populationdoes not reach secondary school. Strengtheningbasic education is a key element in providingtechnical and vocational training.

Work towards greater equity. In many countries,technical and vocational education fails to reachlarge numbers of marginalized young people,notably young women. Far more could be doneto broaden vocational education opportunities,by offering ‘second chance’ programmes andby better integrating vocational training intonational poverty reduction strategies. Designingflexible programmes for young people who havenot completed secondary school or gone beyondprimary education can help combat youthunemployment.

This section is divided into five parts. Part 1 outlinesthe diverse ways in which countries approach thetask of supplying technical and vocational educationand provides a bird’s-eye view of global participationin vocational education at the secondary schoollevel. Part 2 looks at one of the most sensitivebarometers of the mismatch between trainingand the economy – youth unemployment. Whilethe global economic crisis is leaving its mark onpeople across the world, marginalized youngpeople are often bearing the brunt.

Part 3 examines what can happen to technicaland vocational programmes when good intentionsare undermined by lack of finance, poor designand weak linkage to labour markets. It highlightsthe particular challenges governments face in theArab States, India and sub-Saharan Africa. Part 4explores how vocational education can help youngadults avoid marginalization by offering them asecond chance to acquire the skills they need. Part 5considers what kinds of policies lead to effectivetechnical and vocational education programmes thatfacilitate the transition from school to employment.

Technical and vocational education

The fundamental purpose of technical andvocational education is to equip people withcapabilities that can broaden their opportunitiesin life, and to prepare youth and young adults forthe transition from school to work. Skilldevelopment in technical and vocational educationmatters at many levels. For individuals, the skillscarried into the labour market have a major

influence on job security and wages. For employers,skills and learning play a key role in raisingproductivity. For society as a whole, raising theoverall level of skills, ensuring that young peopleare not left behind and aligning the supply of skilledlabour with the demands of industry are criticalto social cohesion. This section focuses principallyon the role of vocational education, rather thanon training provided by companies, in thegeneration of skills and capabilities.

Vocational programmes vary across countriesTechnical and vocational education programmesemerged in developed countries during thenineteenth century to support industrialdevelopment. Their subsequent evolution and theiradoption in developing countries reflect complexinstitutional relationships between educationand economic systems.

There are many models of provision. While somecountries provide general education in schools,with companies or special training institutesoffering vocational options, other countries offerdistinctive vocational options in secondary school.Apprenticeship programmes are an importantpart of technical and vocational education provision,though here, too, arrangements vary. Several broadapproaches can be identified:

Dual systems. Some countries combine school-based and work-based training in dualsystems, integrating apprenticeships into theformal education structure. OECD countriesthat typically offer this option include Denmark,Germany, Switzerland and, more recently,Norway (OECD, 2007a). The well-known Germandual system, which has been widely copied indeveloping countries, creates opportunities forstudents to combine school-based classes within-company training (Barabasch et al., 2009).Four key stakeholders are involved: the federalgovernment, the state government,representatives of employer organizationsand trade unions (German Federal Ministryof Education and Research, 2006). Benefits ofthe German system include firm-based trainingthat equips students with skills suitable for thejob market, an assured pool of skilled workersand private sector contributions to financing.Vocational training has played an importantrole in combating youth unemployment andreducing wage inequalities. On a less positivenote, early tracking into vocational educationhas contributed to deep inequalities in

For individuals, the skills carriedinto the labourmarket have amajor influence on job security and wages

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

7 8

educational achievement, with the schoolsystem actively reinforcing social and economicdivisions.18 Germany has some of the largesteducation disparities between schools and socio-economic groups in the OECD countries,with the children of immigrants far more likelyto be tracked into vocational education.19

School-based systems. Several countrieshave traditionally maintained a division of rolesbetween school-based general education andcompany-based training. In Japan, full-timevocational schooling is followed by full-timeemployment in enterprises linked to the school(OECD, 2009c). As with the German dual system,vocational training in Japan has historicallyhelped facilitate quick settlement of schoolleavers into secure employment. However,unlike in the German system, with its focuson firm-based training, in Japan studentsin vocational tracks typically leave full-timeeducation to enter companies that providetraining linked to their schools.

Mixed models. Many countries operate hybridprogrammes, providing vocational educationstreams within the school system. This is acharacteristic of the French model, thoughFrance also operates a small parallel ‘dualsystem’ (Grubb, 2006). The United Kingdomoperates several ‘school and work’ programmesinvolving apprenticeships and general education(UK Learning and Skills Council National Office,2007). However, the links between employersand educators have traditionally been lessinstitutionalized than in the German orJapanese systems.

In most countries, governments hold primaryresponsibility for setting the overall directionof vocational education policy and for overseeingand regulating standards. A wide range of otherinterested parties is involved, however, includingemployers, trade unions, civil society and privateagencies. Many countries have created nationaltraining authorities to oversee and coordinateactivities, with remits that extend from the designof vocational curricula in schools to oversightof training in specialized institutions and incompanies. Occupational and standard-settingbodies, along with national qualificationframeworks, seek to establish uniform andpredictable standards, enabling employersto assess potential employees’ skills.

Beyond the school, there is a wide range of trainingproviders. In some cases, government agenciesplay the lead role in financing and providing trainingthrough specialized institutes. Other countries,such as Chile and Mauritius, have split financingfrom the provision of training and adopted acompetitive model for procuring training services.20

In some countries, the private sector occupies animportant position in both financing and providingtraining. The diversity of governance models isevident in Latin America (CINTERFOR/ILO, 2001;Gallart, 2008). In Colombia and Costa Rica, whichhave highly effective training models, the publicsector plays the dominant role in finance andprovision. By contrast, Brazil’s Serviço Nacionalde Aprendizagem Industrial (SENAI), one ofthe most successful vocational systems inthe developing world, is administered by theConfederação Nacional da Indústria (Box 2.11).

Vocational education is costly Evidence from developed and developing countriessuggests that technical and vocational educationis relatively costly to provide. In the fourteen OECDcountries for which data are available, expenditureper student is around 15% more than in generaleducation (OECD, 2008b). Evidence from sub-Saharan Africa suggests that vocational educationis up to fourteen times more expensive than generalsecondary education (Johanson and Adams, 2004).

Public financing plays the central role in payingfor vocational provision through the secondaryschool system. In dual systems, training costs aretypically shared by governments and employers.For example, in Germany, companies coverapprenticeship costs while regional governmentspay for the school-based component (Ryan, 2001).Many governments mobilize private finance fornational training programmes through payroll taxeslevied on companies. Egypt’s Training Finance Fundis supported through a 1% levy on payroll taxes(DFID and World Bank, 2005). Twelve countriesin sub-Saharan Africa impose a similar levy, albeiton a far narrower tax base (Adams, 2007b).

Companies play an important and expanding role There is strong evidence that investment in trainingfor young people in the workplace is good for thecompanies involved, for individuals and for nationaleconomies. However, governments have to addressthe fact that workplace training is not always sharedfairly. Levels of investment in training tend to risewith the size of the company and the level of

18. ‘Tracking’ refers to the practice of separatingstudents into differentschool types, typicallyacademic vs. vocational,at the scondary level.

19. Decisions over thetracking of students areoften taken as early asage 10 to 12. The top tierof schooling – gymnasium– paves the way touniversity. Only 18% ofimmigrant children makeit to this top track,compared with 47% ofGerman students.Meanwhile, 40% ofimmigrant children attendthe lowest branch – twicethe share for studentsfrom German families.

20. The Industrial andVocational Training Boardin Mauritius is oneexample. In Chile, theServicio Nacional deCapacitación y Empleo(SENCE) has no capacityfor delivering training butcontracts services from a range of public andprivate providers.

In mostcountries,

governments,employers,

trade unions, civil society and private

agencies areinvolved in TVET

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

7 9

education of the workforce. While over 80% ofcompanies in Kenya and Zambia with more than150 employers are active trainers, under 5% ofthose with fewer than 10 workers fall into thiscategory (Adams, 2007b; Tan, 2006). Severalcountries have adopted innovative approachesaimed at extending company-based opportunitiesfor skills development. Singapore’s SkillsDevelopment Fund and Malaysia’s HumanResources Development Fund are financed by a 1%levy on wages, with the revenue used to subsidizetraining for workers in smaller companies.21

Vocational training through secondary schoolsTechnical and vocational education is offeredthrough a bewildering array of institutionalarrangements, public and private providers, andfinancing systems, so cross-country comparisonshave to be treated with caution. The weaknessof many national reporting systems, combinedwith a lack of consistency, adds a further layerof complication (UNEVOC and UIS, 2006).

Mapping a diverse sector. Detailed mapping oftechnical and vocational education reveals somebroad patterns. The most common format is entryin middle school or upper secondary school, orthrough college courses combining general andvocational learning. Most courses at this levelorient students towards labour markets, thoughsome offer a route into tertiary or generaleducation. Some developed countries, includingFrance and Germany, introduce ‘pre-vocational’courses in lower secondary, often targeting themat what are deemed the less academic students.In many developing countries, early tracking isthe rule rather than the exception. In the UnitedRepublic of Tanzania, two out of three vocationalstudents are tracked after primary school, withthe remainder entering specialized technicalschools after completing general education(Kahyarara and Teal, 2006).

Participation in technical and vocational educationhas increased alongside the general expansion ofsecondary education, but the degree to whichsecondary education has been ‘vocationalized’ variesmarkedly (Lauglo and Maclean, 2005). In 2007, 16%of secondary school students in developed countrieswere in technical and vocational education,compared with 9% in developing countries.22

Technical and vocational shares were lowest insecondary enrolment in sub-Saharan Africa (6%),and South and West Asia (2%). (Table 2.4)

Behind these regional averages are very largedifferences between countries (Annex, StatisticalTable 7). In thirteen of the twenty-five countries insub-Saharan Africa with data, the share of technicaland vocational education in secondary enrolmentwas less than 5%. In Latin America, coverageranges from less than 5% in Brazil, the DominicanRepublic and Nicaragua to over 30% in Argentinaand Honduras. Developed countries, too,demonstrate wide variation. Reported enrolment intechnical and vocational education at the secondarylevel ranges from less than 20% in fourteencountries, including France, Spain and the UnitedKingdom, to over 45% in the Netherlands.

Secondary school enrolment – unequalconvergence. One way to assess participation intechnical and vocational education is to measurethe proportion of secondary school students whoare enrolled in such programmes. But to avoidgetting a distorted picture, the fact that countriesvary widely in levels of secondary schoolparticipation must be taken into account. Whiledeveloping countries have been increasingparticipation in secondary education and beyond,that process has been highly unequal.

Table 2.5 shows the limits of current progress.Developed countries have achieved near universalsecondary education and progression into tertiaryeducation has increased, with the gross enrolmentratio reaching 67% in 2007. Developing regions arecatching up at varying speeds and from different

21. Singapore’s fund reaches65% of enterprises withbetween ten and forty-nineworkers. Malaysia’s includesa facility for supporting smallenterprises in developingtraining plans and offersincentives for larger firmswith excess training capacityto offer places to workersfrom smaller firms.

22. This uses theconventional benchmark ofISCED 2 and 3 for lower andupper secondary educationlevels.

In the UnitedRepublicof Tanzania, two out of threevocationalstudentsare tracked afterprimary school

The best-known graduate of Brazil’s Serviço Nacional de AprendizagemIndustrial (SENAI) is President Luiz Inácio Lula da Silva, who trainedthere as a mechanic.

SENAI operates one of the world’s largest integrated vocational systems, administered by the Confederação Nacional da Indústria(National Confederation of Industry). Delivering courses through about 700 training centres in twenty-seven states, it trains 2.8 millionprofessionals a year. Working with government agencies, SENAI hasestablished rigorous, world-class standards for training and certification,enabling graduates to switch between employers and states.

Financed through a payroll tax on industry, the SENAI system is managedby entrepreneurs. Companies play an important role in identifyingpriority areas for training and in the design of courses. Administrationalso involves national and regional governments, and trade unions.

Source: SENAI (2009).

Box 2.11: Private vocational training in Brazil:widespread and successful

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

8 0

518 721 47 54 024 46 7 10

409 125 47 37 044 47 6 983 335 49 13 553 43 16 1626 261 48 3 428 40 12 13

35 580 44 2 221 39 2 627 453 47 3 157 43 7 1110 891 48 1 271 46 11 12

165 769 48 23 658 49 11 14162 324 48 22 550 49 11 14

3 445 48 1 109 44 34 32125 705 44 2 412 27 1 2

58 547 51 6 275 54 9 111 294 50 51 49 2 4

57 253 51 6 225 54 10 1162 401 49 8 645 43 14 1432 375 48 6 385 39 17 20

World

Developing countriesDeveloped countriesCountries in transition

Sub-Saharan AfricaArab States Central AsiaEast Asia and the Pacific

East AsiaPacific

South and West AsiaLatin America and the Caribbean

CaribbeanLatin America

North America and Western EuropeCentral and Eastern Europe

Table 2.4: Enrolment in technical and vocational education (TVE) by region, 2007

Total

(000) (%) (%)(000)

Female Total

% ofschool agepopulationFemale

% of totalenrolment insecondary

Source: Annex, Statistical Table 8.

Total enrolmentin secondary Total enrolment in TVE

School year ending in

60 0.92 66 0.95 18 0.96 26 1.08

52 0.89 61 0.94 11 0.78 18 0.96100 1.00 100 1.00 55 1.19 67 1.29

91 1.01 90 0.98 39 1.21 58 1.29

24 0.82 34 0.79 4 0.67 6 0.6660 0.89 65 0.92 19 0.74 22 1.0585 0.99 95 0.98 18 0.93 24 1.1065 0.96 78 1.01 14 0.75 26 1.0064 0.96 77 1.01 13 0.73 25 0.99

111 0.99 105 0.96 47 1.24 53 1.3145 0.75 52 0.85 7 0.64 11 0.7780 1.07 89 1.08 21 1.12 34 1.1953 1.03 58 1.03 6 1.30 7 1.3681 1.07 90 1.08 22 1.12 35 1.19

100 0.99 100 1.00 61 1.23 70 1.3387 0.98 88 0.96 38 1.18 62 1.25

World

Developing countriesDeveloped countriesCountries in transition

Sub-Saharan AfricaArab States Central AsiaEast Asia and the Pacific

East AsiaPacific

South and West AsiaLatin America and the Caribbean

CaribbeanLatin America

North America and Western EuropeCentral and Eastern Europe

Table 2.5: Gross enrolment ratios in secondary and tertiary education, 1999 and 2007

1999

Total GPI(F/M)

Total GPI(F/M)

2007School year ending in

1999

Total GPI(F/M)

Total GPI(F/M)

2007

Source: Annex, Statistical Tables 8 and 9A.

Gross enrolment ratiosin secondary

(%)

Gross enrolment ratiosin tertiary

(%)

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

8 1

starting points. Secondary gross enrolment levelsranged from 34% in sub-Saharan Africa to 65% inthe Arab States and 90% in Latin America in 2007.Tertiary enrolment was just 6% in sub-SaharanAfrica, compared with 22% in the Arab States and35% in Latin America. These regional averagesconceal large intra-regional disparities. While theaverage secondary participation level was 90% inLatin America and the Caribbean, it was less than70% for some countries in the region, includingEl Salvador, Guatemala, Honduras and Nicaragua.

Gender disparities in secondary school have animportant bearing on opportunities for technicaland vocational education. The two regions with thelargest gender disparities are South and West Asia,and sub-Saharan Africa. While the former hasachieved a marked improvement in gender paritysince 1999, the latter has moved in the oppositedirection: the secondary-level GPI for sub-SaharanAfrica has slipped from 0.82 to 0.79. This points tothe importance of public policy interventions tostrengthen opportunities for young girls to makethe transition from primary to secondary school.In Latin America and the Caribbean, where moregirls than boys attend secondary school, there hasbeen no progress in narrowing the gender gap.

Gender inequalities are often more pronouncedin technical and vocational education than ingeneral education. In South and West Asia, andsub-Saharan Africa, girls accounted for 44% ofstudents in secondary school in 2007, but just 27%and 39%, respectively, in technical and vocationaleducation. In nine of the eleven Arab states forwhich data are available, girls accounted for lessthan 40% of enrolment. The same is true for twelveof the twenty-five countries in sub-Saharan Africawith reported data. These disparities tell only asmall part of a far wider story of gender inequality.In many cases, young girls in technical andvocational streams are being trained for traditionalfemale occupations, often in areas characterizedby low pay. Moreover, returns to vocationaleducation are often lowered by genderdiscrimination in employment and wages.

Prospects for successful vocational educationprovision are inevitably shaped by the widerlearning environment. One of the lessons fromsuccessful countries in East Asia and elsewhereis that high levels of literacy, numeracy and broad-based general education are the real foundationfor acquiring flexible and transferrable vocationalskills. Many countries lack the foundation.

Consider the prospects for 15-year-olds in differentparts of the world (Figure 2.24). In OECD countries,85% of 15- to 19-year-olds are in full-timeeducation and at 15 a student can expect tocontinue for seven more years (Kuczera et al., 2008;OECD, 2008b). This compares with less than oneyear in South and West Asia. In sub-Saharan Africa,the average 15-year-old does not attend school.In countries including Burkina Faso, Ethiopia andMozambique, more than 75% of young people whodo not go to school report having no education(Garcia and Fares, 2008).

Failures in basic education have importantconsequences for technical and vocationaleducation. In sub-Saharan Africa, and South andWest Asia, technical and vocational educationreaches 1% to 2% of the total secondary school agegroup (Table 2.4). One reason for this is that, inmany countries in both regions, only a small shareof the secondary school age population reachesthe middle grades of secondary school.

One important policy conclusion to be drawn fromthe data in these regions is that no national policyfor developing skills is likely to succeed unlessgovernments dramatically increase the flow ofstudents into secondary school.

The foundations for learning are established inprimary school and nurtured in the early secondary

In sub-SaharanAfrica, the average15-year-old doesnot attend school

World

Developing countriesCountries in transition

Developed countries

Sub-Saharan AfricaSouth/West Asia

Arab StatesEast Asia/Pacific

Central AsiaCentr./East. Europe

Latin America/CaribbeanN. America/W. Europe

0 2 4 6 8 10 12 14 16

School life expectancy (years)

Age 15*

Figure 2.24: By age 15, many students in developingcountries are nearing the end of their schoolingSchool life expectancy from primary to tertiary education, by region, 2007

* This is a theoretical threshold that assumes an intake age of 6 in all regions.Source: Annex, Statistical Table 4.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

8 2

school years. People lacking these foundationsare not well placed to develop the type of flexibleproblem-solving capabilities needed to underpinmore specialized learning. For countries wheremuch of the youth population either does not reachsecondary school or lacks basic literacy andnumeracy, technical and vocational education insecondary school can only have limited successas a national skills development strategy. It maymake little sense to rapidly scale up investmentin technical and vocational education in countriesenrolling only a small proportion of the secondaryschool age group. Directing resources towardsimproving access and the quality of education incore subjects is likely to prove far more effectiveand equitable (Lauglo and Maclean, 2005).

Youth unemployment reveals the skills gap

The broad aim of technical and vocationaleducation is to equip young people and adults withthe skills and knowledge they need to cross thebridge from school to work. The economic crisishas made that crossing even more hazardous.Young people who fail to make the transition oftenface the prospect of long-term unemployment andsocial marginalization, and run a higher risk ofbeing drawn into illicit activities (Adams, 2008;Brewer, 2004).

While the picture varies by region, governments’records in tackling youth unemployment over thepast decade have been disappointing. With globalunemployment rising sharply in 2009, the recordcould deteriorate further as young people are hithardest by the job crisis.

Pre-crisis trends were not encouragingEducation and demographic trends, coupledwith rapid economic growth before the 2008economic downturn, might have been expectedto reduce youth unemployment, with the averagenumber of years spent in school increasing andthe youth share in the working age populationdeclining in all regions, with the notable exceptionof sub-Saharan Africa.

Instead, the International Labour Organization (ILO)reported a 13% rise in youth unemployment, from63 million in 1996 to 71 million in 2007. Labourmarket demand is one factor behind this trend.Economic growth has not generated employmenton the scale that might have been anticipated.At the same time, rising youth unemployment

during a period of sustained economic expansionpoints to a mismatch between skills acquired ineducation and labour market demand. Theupshot is that young people bear the brunt ofunemployment. Before the crisis, the global youthunemployment rate stood at 12%, or around threetimes the adult unemployment rate (ILO, 2008a).In every region, youth unemployment rates arehigher than those for older workers. Youth aged 15to 24 make up one-quarter of the world’spopulation but almost half of the unemployed.

Young people are now in the front line of the globaleconomic downturn. Recent estimates suggest thatworld unemployment could be 39 million higherby the end of 2009, compared with 2007, and thatyouth unemployment may rise by between 5 millionand 17.7 million. The youth unemployment rate isprojected to increase from around 12% in 2008 tobetween 14% and 15% in 2009 (CINTERFOR/ILO,2009). Employers are more prone to dismiss youngworkers – especially unskilled young women –because youth tend to have the least secureemployment conditions and are often not coveredby labour regulations (CINTERFOR/ILO, 2009).

Youth unemployment patterns vary across thedeveloping regions (Figure 2.25). The ILO reportsthat the Middle East and North Africa have thehighest unemployment rates, with about one-fifth of 15- to 24-year-olds unemployed. In Egypt, youthaccount for more than 60% of the unemployed.Gender discrimination, both in terms of jobsegmentation and wages, is deeply entrenched inArab States’ labour markets (Salehi-Isfahani andDhillon, 2008). In Egypt, fewer than one-quarterof women aged 15 to 29 are economically active –one-third the male rate. The transition from schoolto work is also more difficult for girls, with fewerthan 25% of young women finding work withinfive years (Assad and Barsoum, 2007). Employerdiscrimination, early marriage and claims on thelabour of women at home all reinforce genderdisadvantage in labour markets.

Demography and poverty combine to leave sub-Saharan Africa facing particularly stark challengesin youth employment. The region’s share of theworld’s youth population, currently about 17%, willbe some 25% by 2025. Almost two-thirds of thepopulation is under 25. The transition from schoolto work is enormously difficult for this growingpopulation. Every year between 7 million and10 million young Africans enter labour marketscharacterized by high unemployment, low

Youth aged 15 to 24 make up

one-quarter of the world’s

population but almost half of

the unemployed

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

8 3

productivity, chronic insecurity and poverty-levelincomes (Garcia and Fares, 2008).

Unemployment is just one of the problemsyoung people encounter as they seek to enterthe workforce. Many face protracted delays insecuring their first jobs. In much of the MiddleEast and North Africa, the average duration ofunemployment for first-time job seekers ismeasured in years rather than months. In sub-Saharan African countries including Ethiopia,Malawi, Mozambique and Zambia, young peopleface about five years of reported inactivity beforefinding work (Garcia and Fares, 2008).

Education is not an automatic panacea fordelayed employment. In many Arab states, youngpeople with secondary and tertiary education facelonger periods of unemployment than their peerswith only basic education. Similarly, in severalcountries of sub-Saharan Africa, includingBurundi, Cameroon, Kenya and Nigeria, youthwith secondary and tertiary education have higherrates of unemployment than those with lowerlevels of attainment (Fares et al., 2005; Garciaand Fares, 2008).

Comparisons across developing regions have tobe made with caution. Gender parity in reportedyouth unemployment in South Asia and sub-Saharan Africa does not imply gender equityin labour markets. In both regions, many youngwomen provide unpaid labour in the householdand do not participate in paid employment.23

Similarly, lower levels of youth unemploymentdo not necessarily correspond to higher levels ofdecent employment. Poverty forces millions ofpeople into insecure, low-wage jobs in the informalsector. The ILO estimates that 300 million youngpeople are ‘working poor’ who live on less thanUS$2 a day (CINTERFOR/ILO, 2009).

Developed countries also face acute problemsEconomic recession in OECD countries is pushingunemployment to record levels. In developedcountries as a group, unemployment is projectedto peak at 7.3% in 2010, compared with 5.5% in2007 (OECD, 2009d). The scenario could worsenif economic recovery is delayed.

As in developing regions, the economic downturnin rich countries comes against a discouragingbackdrop for youth employment (Figure 2.26).Despite strong economic growth from 1997 to 2007,the youth unemployment rate in OECD countries

23. In South Asia, the ILOreports just 22% of femaleyouth in employmentcompared with 58% of male youth.

In Burundi,Cameroon, Kenyaand Nigeria, youthwith secondaryand tertiaryeducation havehigher rates of unemploymentthan those withlower levels of attainment

Figure 2.25: Gender inequalities reinforce high levels of youth unemploymentYouth unemployment rates by region* and gender, 2007

* Regions presented are those used by the ILO, which differ to some extent from the EFA regions.Sources: ILO (2008b); OECD (2009f ).

0

5

10

15

20

25

30

35

East AsiaWorld LatinAmerica/Caribbean

MiddleEast

North AfricaSouth-EastAsia and

the Pacific

Sub-SaharanAfrica

South Asia

Yout

h un

empl

oym

ent r

ates

(%)

Female

Male

Average

Figure 2.26: In most OECD countries, youth face greaterrisk of unemploymentYouth and adult unemployment rates, selected OECD countries, 2008

Source: OECD (2009f ).

Norway

Denmark

Rep. of Korea

Austria

Australia

Japan

United Kingdom

Sweden

United States

Canada

OECD

Italy

France

Greece

Germany

Turkey

Spain

0 5 10 15 20 25

Unemployment rates (%)

Youth (15 to 24)

Adult (25 to 54)

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

8 4

fell only a little, from 15% to 13% (CINTERFOR/ILO,2008). While some countries, including Australia,Canada, France and Spain, achieved markedreductions, six other countries, including theUnited Kingdom, experienced increased youthunemployment.

Young people with low skills are especiallyvulnerable, as was evident even before the deeprecession took hold. In the OECD, for example,skilled jobs have been created at five times the rateof unskilled jobs since 2000 (OECD, 2008b). Theskills gap helps explain the apparent paradox of highgrowth and stagnant youth unemployment in manycountries. In the OECD countries as a group, peoplewith low skills are twice as likely to be unemployedas those with high skills, increasing to four times aslikely in the United States. The rising premium onskills has increased the penalties faced by those inthe OECD’s large pockets of educational deprivation.

Spiralling youth unemployment has added asense of urgency to national debates over technicaland vocational training. While the impact of theeconomic slowdown is being felt across society,it has fallen most heavily on the young and peoplewith low skills (OECD, 2009d). Young peopletypically find it hard to get established in the labourmarket because of their lack of experience, whichmakes them especially vulnerable in a downturn.The young in general and those with low levels ofqualification in particular are emerging as primevictims of the slump.

A side effect of the downturn is that it has pushedtechnical and vocational education and training tothe centre of the political agenda. In France, whereeven before the crisis almost one in five youngpeople was out of work, a quarter of them for morethan one year, the government has launched anemergency youth employment programme focusedon apprenticeships (CINTERFOR/ILO, 2009). InJapan, though youth unemployment rates are lowerthan in France, around one-third of workers aged 15to 24 are in temporary work with insecure contracts(OECD, 2009c). Here, too, measures have beenintroduced to facilitate school-to-work transitionthrough firm-based training. Comparable measuresinvolving incentives for young people to stay ineducation, training and apprenticeships are beingused across the OECD.

Looking beyond the immediate responses, itis important for governments to use the crisisas an opportunity to put in place the long-term

investments and policies – in education andbeyond – that are needed to combat themarginalization of young people.

Good intentions, poor results:problems in the developing world

Much can be achieved through good-qualityvocational education and training. But in manydeveloping countries, vocational programmeshave suffered from a combination of underfinancing,poor design and weak links to labour markets. Insome regions – notably sub-Saharan Africa andLatin America – deep cuts in spending during the1980s and 1990s further compromised quality invocational education (Johanson and Adams, 2004).Public investment has produced disappointingresults, calling into question the potential forvocational education to fuel economic growthand reduce poverty.

The poor track record is reflected in student andteacher preferences. In many countries, vocationaloptions are viewed either as a last resort or as apossible route back into general education, ratherthan as a stepping stone to employment. This isespecially true of sub-Saharan Africa, where thereluctance of parents to put their children intovocational streams is supported by evidenceconfirming that general education generatesfar higher returns than do vocational alternatives(Kahyarara and Teal, 2006). Thailand adoptedthe German dual system in 2005; successivegovernments have attempted to expand vocationaleducation to combat child labour and themarginalization of young people who drop out ofschool. However, while secondary school enrolmenthas doubled, vocational enrolment has failed to takeoff, reflecting concerns of parents and studentsabout the quality of provision and the weaknessof links to job markets (World Bank, 2008g).

The Middle East: fragmentation and weak links to employmentFaced with the world’s highest levels of youthunemployment, governments in the Middle Easthave identified vocational education as a priority.Two broad models have emerged. At one extreme,students in Egypt are tracked early, but vocationalgraduates suffer as much unemployment as theirsecondary school counterparts (Kamel, 2006;Salehi-Isfahani and Dhillon, 2008). In the IslamicRepublic of Iran, where tracking into vocationaleducation starts later, it is seen as a sign of failure,prompting many students to drop out. (Box 2.12).

Spiralling youthunemployment

has added a sense of

urgency to national

debates overtechnical

and vocationaltraining

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

8 5

The Islamic Republic of Iran’s experience demonstratesthe challenges facing policy-makers across the MiddleEast. Over the past twenty years, the country has maderapid strides in education. Participation at secondary levelhas increased, average years in education have nearlydoubled and gender inequalities have narrowed, especiallyin urban areas. Vocational education, however, reinforcesa mismatch between skills and jobs that perpetuates highyouth unemployment.

The education system in the country is heavily orientedtowards the university entrance exam, the concour, which parents and students see as a route to secureemployment, usually in the public sector. Compulsoryeducation ends at around age 15, when students areevaluated and directed on to three separate tracks: theacademic curriculum (Nazari), technical and vocationaleducation (Fanni-Herfei) and basic skills through on-the-job training (Kardanesh). The aim of the latter two isexplicitly to focus on job skills, but the system fails onseveral fronts.

Tracking brings high levels of attrition. Of the femalestudents who began their secondary education in2003/2004, nearly one-third dropped out after tracking(Figure 2.27). Most students pursue the Nazari track witha view to passing the concour, spurning the vocationaltracks because of their low perceived status and quality.But of the nearly 1.5 million who proceed each year to the concour, 1.2 million fail and leave school lackingqualifications and job skills.

Iranian policy-makers increasingly recognize the problemswith the current system. Of particular concern are themisalignment of education and labour markets, and thepoor quality of vocational education, which operatesthrough a network of highly centralized public trainingcentres. Many of these lack equipment and well-trainedinstructors, and they produce qualifications thatemployers see as having limited relevance.

The concour system creates further problems. Most ofthe exams are multiple choice, and teaching methodsemphasize rote learning. Students other than thoseentering elite engineering and medical schools oftenemerge ill prepared to enter productive enterprises.

The mismatch between education and employment isbecoming increasingly stark. Steady economic growth has reduced overall unemployment, but youthunemployment remains over 20%. Those who completedupper secondary education have the highest level ofunemployment (Figure 2.28). Measured in terms ofemployment, the benefits of education are dwindling,along with the skills base of the Iranian economy.

Education is only part of the story. Labour market rigidityand discrimination also play a role. Gender barriers to

employment appear to be rising, with unemployment rates among women aged 20 to 24 now twice the level for men of that age group.

Whatever the underlying causes, the skills mismatch is a pressing political concern. The rising proportion ofyoung people in the population means the labour force is expanding by almost 4% a year, or nearly 1.2 millionpeople. Many of them will face social exclusion if theIslamic Republic of Iran fails to create enough jobs andgive its people the skills they need to fill them.

Sources: Salehi-Isfahani and Egel (2007); Povey (2005).

Box 2.12: Training, skills and youth exclusion in the Islamic Republic of Iran

Grade 3(2005)

Grade 2(2004)

Grade 1(2003)

Grade 3(2005)

Grade 2(2004)

Grade 1(2003)

Female Male

0

20

40

60

80

100

0

20

40

60

80

100

Attrition Kardanesh(basic skills throughon-the-job training)

Fanni-Herfei (technical andvocational education)

Nazari(theoretical/academiccurriculum)

100

58 52

87

1312

2129

100

35 29

1314

16

12

3644

% o

f the

coho

rt

% o

f the

coho

rt

Figure 2.27: In the Islamic Republic of Iran, vocational tracking comes with high dropout ratesCohort tracking at lower secondary level, students entered in 2003

Source: Salehi-Isfahani and Egel (2007).

0

10

20

30

40

50

60

Illiterate Primary Lowersecondary

Some uppersecondary

Uppersecondary

Tertiary Total

Unem

ploy

men

t rat

es (%

)

Male Female

Figure 2.28: Unemployment increases with level of education, but Iranian women are especially penalizedUnemployment rates by educational attainment and gender, 2005

Source: Salehi-Isfahani and Egel (2007).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

8 6

High youth unemployment in the Middle Eastis about far more than a failure of vocationaleducation. Slow economic growth, rigid labourmarkets and gender discrimination have all stymiedjob creation. In many cases, education systems arepart of the employment problem. Courses aregeared towards rote learning for university entranceexams that are seen as a route to public sectoremployment. The upshot is that millions of youthleave school without employable skills and millionsmore emerge from university lacking the capabilitiesneeded to compete for entry into private sectoremployment (Salehi-Isfahani and Dhillon, 2008).

Most parents and students in the Middle Eastsee vocational education as unattractive becauseit receives meagre budget resources, is oftendelivered by badly trained teachers lacking inmotivation, bears little relation to the skillsemployers seek and produces certificates thatare not subject to uniform standards. Part of theproblem in many countries is that the privatesector has a limited voice in setting priorities andstandards (DFID and World Bank, 2005). As a result,the skills delivered through vocational programmesare often of little relevance. In addition, governancetypically falls to a range of ministries andgovernment agencies, so it is often fragmentedand poorly coordinated. There are some notableexceptions. In Egypt, innovative partnerships arebringing together governments, business anddonors.24 And Morocco has adopted far-reachinggovernance reforms aimed at improving quality,relevance and equity (Box 2.13).

In India, limited reach and duplicationTechnical and vocational education systems in manycountries suffer from inadequate reach as well aslimited benefits for participants. In India, only 3% ofrural youth and 6% of urban youth have had any kindof vocational training (India Ministry of Health andFamily Welfare, 2006; India Planning Commission,2008). The country’s Industrial Training Institutesand various craft centres are not accessible tothe vast majority of the poor. India also has someof the world’s largest reported gender disparitiesin technical and vocational education, with girlsaccounting for just 7% of enrolment at thesecondary level and their courses heavilyconcentrated in traditional areas such as nursingand sewing. In general, the benefits of vocationaltraining are not immediately apparent. Some 60% ofgraduates from Industrial Training Institutes are stillunemployed three years later (World Bank, 2006g).Industrial apprentices are more likely to get work,but generally not in the trade for which they trained.

Governance problems have hampered India’s effortsto strengthen vocational education. Responsibilitiesare split among the Ministry of Labour, the Ministryof Human Resource Development, other nationalbodies and state authorities. Duplication andfragmentation are widespread, there is little controlover quality and the certification system is poorlyunderstood by employers. Companies and employerorganizations are only marginally involved, thoughefforts are being made to strengthen theirengagement.

Sub-Saharan Africa: failing to reach the marginalizedGovernments in sub-Saharan Africa face someof the toughest challenges in reforming technicaland vocational education. Finance is part of theproblem – institutions across the region sufferfrom a familiar combination of underinvestmentin equipment, low pay for instructors and problemsrecruiting qualified staff. But not all the difficultiescan be traced to financial causes.

Many countries track students into vocationaleducation far too early – often in the face ofconcerted resistance from parents. Parentalconcerns are often well grounded. Evaluationspoint to low rates of absorption of graduatesinto employment – under half in some countries,including Madagascar, Mali and the United Republicof Tanzania (Johanson and Adams, 2004). Theresulting unemployment, even in countries whereemployers face shortages of skilled secondary

24. One prominentexample is the Mubarak-Kohl initiative, anarrangement involvingthe Ministry of Education,the German technicalcooperation agency GTZand businessassociations. Thegovernment providespremises, GTZ suppliestechnical experts andequipment, and businessassociations contributetraining opportunities andallowances. So far,around 16,000 traineeshave been trained in1,600 companies through45 technical secondaryschools.

In the MiddleEast, millions of

youth leaveschool without

employable skills

In Morocco, vocational education has beenoverhauled in the past decade. It has its ownministry and a national office for vocationaltraining and work promotion. Syllabuses areadapted to trainees’ general education level, withan emphasis on a combination of specific skills and broader capabilities. Vocational schools haveachieved good results, with more than half ofgraduates finding a job within nine months. Theproportion of female trainees is rising, reaching44% in 2006. The vocational system is expandingas the government seeks to foster the skillsneeded by new sectors such as vehiclemanufacture, aeronautics and agro-industry.

Source: African Development Bank/OECD (2008d).

Box 2.13: Morocco — strengtheningvocational governance

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

8 7

school graduates, points to a mismatch betweenlearning and labour market needs.

The high cost of vocational education is anotherfactor. Partly because class sizes are much smallerthan in general education and the cost of equipmentis higher, vocational education faces far higher percapita costs – about twelve times the average forprimary school and four times that for secondaryschool (Atchoarena and Delluc, 2001).

Problems in vocational education are a legacyof past policy failures and a difficult environment.The quality of provision suffered enormously withdeep cuts in spending under structural adjustmentprogrammes in the 1980s and 1990s. Widerproblems have also been evident. Vocationalsystems were designed to meet the needs of formalsector employers, notably in government (Adams,2008; Africa Commission, 2008). For at least threedecades, however, formal sector job creation hasstagnated while informal sector employment hasgrown in importance. In most countries, informalemployment and self-employment dominate inboth rural and urban areas, typically accounting forover 80% of total employment.25 Providing trainingto those employed in the informal sector involvesreaching people with lower levels of education.A survey covering Kenya, Senegal, the UnitedRepublic of Tanzania, Zambia and Zimbabwereported that half of informal sector workershad only primary education, if any (Haan, 2006;Liimatainen, 2002).

The need to reduce poverty makes it vital to reachthese people, yet most vocational systems fail todeliver. Traditional apprenticeships and on-the-jobtraining are by far the most important routes toskills development for the vast majority of Africanyouth (ILO, 2007; Wachira et al., 2008). On oneestimate they account for up to 70% of overalltraining (Liimatainen, 2002). The strength oftraditional apprenticeships is that they provide youthwho have low levels of education with practical,employable skills (Monk et al., 2008). On a morenegative note, apprenticeships tend to be biasedagainst young women and the very poor. They alsoperpetuate the use of traditional methods, offeringlittle theoretical knowledge (Adams, 2008).

Vocational education could help redress the equitybalance by targeting those who face the most acutedisadvantages. Unfortunately, evidence fromnational evaluations points in the opposite direction.Research in Ghana has highlighted a bias towards

regions and social groups that are already better off(Box 2.14). The broader failure to integrate technicaland vocational education into strategies for reachingmarginalized groups is clear in results from recentevaluations (based on Garcia and Fares, 2008):

In Burkina Faso, only one-third of interventionsinvolving technical and vocational education wereoriented towards disadvantaged groups, mainlythrough micro-credit programmes.

In the United Republic of Tanzania, out of twenty-eight programmes reviewed, only threetargeted the poorest youth, one targeted youthwith no education and three targeted rural areas(where the vast majority of the poor live).

In a region where 95 million young men andwomen have no education and are unemployed,have low-paying jobs or have withdrawn from thelabour force, second-chance programmes arevirtually non-existent. A review covering BurkinaFaso, Ethiopia, Uganda and the United Republicof Tanzania concluded that ‘most second-chanceinterventions are small in scale, underevaluatedand face severe challenges for sustainabilityand scalability (Garcia and Fares, 2008, p. xxx).’

The problems evident in vocational education in sub-Saharan Africa are widely recognized bygovernments, regional organizations and aid donors(Africa Commission, 2008; COMEDAF II+, 2007).Across the region, vocational education isundergoing major reform. Several countrieshave created or strengthened national trainingauthorities, reformed qualification systems andcreated structures giving the private sector astronger voice:

In Cameroon, the four ministries involved invocational education have developed a sector-wide plan linked to the national poverty reductionstrategy (African Development Bank, 2008a).

In Ethiopia, new curricula have been drawn upand qualification systems restructured to bolsterthe development of skills that labour marketsneed (African Development Bank, 2008b).

In Rwanda, a strategy adopted in 2007 setsout ambitious goals for changing the image ofvocational education. A Workforce Development Authority has been created to oversee coordinationand facilitate private sector involvement (AfricanDevelopment Bank/OECD, 2008f).

25. Reporting conventionsmake it difficult to compareacross countries (Adams,2008). The reported shareof informal employment intotal employment rangesfrom over 90% in Mali(where agriculture isincluded) to 22% in theUnited Republic of Tanzania(where agriculture isexcluded).

Vocationaleducation is failingto target thosewho face the most acutedisadvantages

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

8 8

There are also signs that vocational education is re-emerging as a priority in developmentassistance. Several countries, notably Germanyand Japan, have been giving precedence tosupport for the sector.

It is too early to evaluate the results of thelatest wave of reform. In some cases, oldmodels have proved highly resilient.

Mozambique’s government set out a bold strategy,the Integrated Professional Reform Programme,aimed at bringing vocational planning under asingle umbrella, with a unified qualification andaccreditation programme (African DevelopmentBank/OECD, 2008e). Two years before the end ofits first phase, however, there has been littleprogress in implementing it.

In Ghana,vocational

programmes havesuffered from

fragmentedadministration

and poor quality

Since independence half a century ago, politicalleaders in Ghana have seen technical and vocationaleducation as a means of generating jobs. Yetvocational programmes have suffered fromfragmented administration, a proliferation ofqualification standards and poor quality.

Public vocational education in Ghana operatesthrough two tracks. The first, extending from lowersecondary to post-secondary, is administered by theMinistry of Education and Social Service and operatesthrough Technical Training Institutes. The second track is run by National Vocational Training Institutesattached to the Ministry of Manpower, Youth andEmployment. Several other ministries, agencies and private institutions are involved, each offering its own programmes.

The pipeline into vocational education starts in juniorsecondary school, but parents and students tend toshun vocational streams, with just 5% of studentsentering public vocational institutions. The share ofadults aged 20 to 26 years with formal vocationaltraining stood at just 2% in 2005.

Reviews of Ghana’s vocational system haveconsistently highlighted problems of coherence andcoordination. Political oversight has been minimal.Despite what one report describes as a ‘dizzyingarray’ of examinations, programmes have failed to provide the skills employers seek. One reason is a multiplicity of certification and testing standardsdeveloped without employer advice.

The quality of instruction is far from satisfactory. Ill-trained instructors, low salaries and outdatedequipment all contribute. While some publicinstitutions do provide high-quality training, theyremain the exception.

There are few evaluations of the benefits of vocational education for Ghana’s youth. The availableevidence suggests that graduates of the publicsystem, including polytechnics, are prone to highunemployment. This is unsurprising given that

teaching is geared towards the demands of the smallformal sector, rather than an informal sector that onone estimate delivers 80% to 90% of skills training.

The cost side of the equation is better understood.Vocational programmes account for about 1% of theeducation budgets. However, recurrent per capitacosts in 2006 were five times higher than in primaryeducation and almost three times higher than insenior secondary.

Equity is another major concern. While policydocuments emphasize the importance of linkingvocational education to the national poverty reductionstrategy, marginalized groups are effectively excluded.Participation rises with income levels, with the richestquintile seven times more likely than the poorest to have received vocational education. Regionalinequality is marked: the northern region, Ghana’spoorest, has one of the lowest levels of vocationalenrolment. There is a bias towards males, especiallyin urban areas. And vocational graduates are twentytimes more likely to work in the formal sector thanbe self-employed as farmers, reflecting a bias against agriculture.

Rather than counteracting the disadvantagesassociated with limited access to education,apprenticeship programmes have the opposite effect —young people with an incomplete primary educationare half as likely to make it into apprenticeship asthose with a secondary education.

The government has adopted reforms aimed atestablishing a more efficient and equitable system.The Council for Technical and Vocational Educationand Training was created in 2006 as an autonomousoversight body, along with the Skills Training andEmployment Placement (STEP) programme whichtargets low-skilled unemployed youth seekingapprenticeships. It is too early to evaluate the latest reforms.

Sources: Adams et al. (2008); African Development Bank/OECD(2008c); Akyeampong (2007); Ghana Ministry of Education, Youth and Sports (2004a, 2004b); Palmer (2007).

Box 2.14: Vocational education in Ghana — limited access and poor quality

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

8 9

As in other regions, governments in sub-SaharanAfrica have to strike a delicate balance betweengeneral and technical and vocational education.The overwhelming priority for the region is toincrease enrolment, retention and progressionthrough basic education into secondary school.Vocational education has the potential to play afar greater role, however, not least in providingsecond-chance opportunities to marginalized youth.Public investment and international aid should bedirected towards creating opportunities for thepoor, in rural areas and in informal employment,with private spending and investment by companiesfinancing training for higher-income groups.

Offering young people a second chance

Technical and vocational education can extendopportunities for young people still in school.But what of the millions of young adults who havenever gone to school or have left education withlevels of achievement falling far short of whatthey need? Can vocational education offer aneffective ‘second chance’ for avoiding a future ofmarginalization? The economic crisis has given thatquestion renewed relevance because young peoplesuffer most when labour demand is reduced.

Comparing the effectiveness of ‘second-chance’programmes and wider targeted interventionsfor combating marginalization through vocationaleducation is inherently difficult because underlyingpatterns of marginalization vary. In developedcountries, the problem is concentrated at theupper secondary level. Data from the United Statesindicate that nearly 6.2 million of the country’s 16- to 24-year-olds – 16% of the age group –have left secondary school with no diploma.26 InFrance, about 18% of young people lack minimumsecondary school qualifications (OECD, 2009b).

The yardstick for measuring educationmarginalization in developing countries is different.Millions of young people in Latin America and theArab States, especially those from the pooresthouseholds, have just one or two years ofsecondary school, or less. In many low-incomecountries, only a minority of young people havebeen to secondary school at all – and anincomplete primary education is often the norm.27

Evaluations around the world show that ‘second-chance’ programmes can make adifference. Comprehensive approaches that

provide training as part of a wider package aremore likely to succeed. In the United States, theJob Corps programme offers 16- to 24-year-oldseducation and training alongside a wide rangeof support services (Schochet et al., 2003).28

Modelled partly on the Job Corps experience,the Jóvenes programmes in several Latin Americancountries, including Argentina, Chile, Peru andUruguay, have been particularly successful inreaching the marginalized (Box 2.15).

Skills for the twenty-first centuryThe Jóvenes programmes are effective becausethey provide an integrated framework for reachingthe marginalized and linking employment withskills training. That key principle also underpinsanother programme in Latin America andthe Caribbean, Entra 21, launched in 2001 bythe International Youth Foundation to equipunemployed youth with information technologyskills. An evaluation in six countries pointsto encouraging results for both employmentand earnings (Box 2.16).

Remedial education combined with flexiblecourses targeted at marginal populations providesanother way to offer the young a second chance.In Bangladesh, a large-scale programme operatedby a non-government organization targets youngpeople who have dropped out of formal education.Classes designed to facilitate early catch-up arefollowed by vocational programmes developedwith companies (World Bank, 2006j). In Chile,the Califica programme is aimed at youth andyoung adults who lack formal secondary education.It includes a secondary education equivalencycomponent that enables people over 18 to studyin a certified institution and to gain a certificatethat facilitates access to a wide range of vocationalcourses (Gallart, 2008).

Successful second-chance programmes haveto be accessible and affordable to people livingin poverty, be flexible enough to fit in with the livesof their target population and be seen to deliverresults (Jimenez et al., 2007). One of the mostsuccessful models has emerged in Mexico. TheOpen Secondary School system, aimed at youngadults who have dropped out of secondary school,offers second-chance opportunities in thirty-threesubjects covered in grades 10 to 12. There areno entrance requirements and no time limiton completion, and students can determinetheir own schedules. The average period forcompletion is three to five years, after which

26. It is estimated thatsecondary school dropoutsearn US$485,000 lesson average during theirlifetime than do secondaryschool graduates (Center forLabor Market Studies, 2007).

27. These issues are takenup in Chapter 3.

28. The evaluation found thatstudents graduating fromJob Corps programmesgained an average increasein income of around 12%(Schochet et al., 2003).

‘Second-chance’programmes canhelp combat youthmarginalization

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

9 0

Unemployed young people whose education hasbeen disrupted often struggle to break into skilledjobs. The Entra 21 programme is aimed atremoving barriers to entry through innovativeapproaches that give people the skills they needto overcome marginalization.

The programme began in 2001 throughcollaboration between the International YouthFoundation and the Inter-American DevelopmentBank, in six Latin American countries: Bolivia,the Dominican Republic, El Salvador, Panama,Paraguay and Peru. Courses combine technicaltraining, internship and job placement with lifeskills and job-seeking skills. Employers help withprogramme design and job placement. Anevaluation of the first phase, which covered20,000 people, found major benefits:

Among those who registered for Entra 21,69% were neither studying nor employed;after completion the figure was 24%.

The share of graduates from the programmein formal education was 42% — double theshare at the time of entry to the course.Another 21% were working and studying.

While most of the jobs documented werein the formal sector (between 75% and 90%,depending on the country), there were severalexamples of youth-led microenterprisedevelopment in the informal sector, especiallyin El Salvador and Peru.

The second phase of the programme includesmeasures aimed at strengthening the focuson marginalized youth. It targets 45,000 youngpeople from low-income households and5,000 facing increased risk as a result of internaldisplacement or physical disability.

Source: Lasida and Rodriguez (2006).

Box 2.16: Entra 21 — tackling marginalization

Experience from Jóvenes programmes in LatinAmerica provides some important insights intothe conditions for successful youth training.

Initiated in Chile in 1990, Jóvenes programmesare now well established across the region. Theyreach out to young people, combining technicaltraining and internship with basic life skillsand other support services. More than 60%of participants come from low-income families.The programmes tend to raise the probabilityof employment and higher wages. In Argentina,Proyecto Joven increased employment and wagesby about 10% compared with a control group.Although implementation and managementstructures vary, evaluations show that successfulprogrammes in Argentina, Chile, Peru andUruguay share some common elements:

Strong targeting. Programmes are aimedat youth from low-income families and thosewho have low educational attainment andlimited work experience. In some cases,preference is given to household heads withchildren, in order to combat child poverty.

Training is linked with work and wider skills.Most programmes provide training, workexperience, literacy and numeracy courses, anda wide range of auxiliary packages, including jobsearch assistance. The training component isaimed at helping participants attain semi-skilledstatus in trades for which there is demand.Work experience takes place under the auspicesof a company, which assumes a tutoring rolebut is not obliged to pay trainees or guaranteeemployment. Training and work experienceusually last about six months and includebroader life skills such as communication,teamwork and self-esteem.

Management and coordination. The stateassumes control of programme design,supervision and full or partial financing,but in most countries, training delivery isdecentralized. The private sector providesa link to the job market. In Chile, theprogramme operates through about1,000 training providers, ranging fromcompanies to non-government organizations.

Sources: Betcherman et al. (2004, 2007); Gallart (2008);Godfrey (2007).

Box 2.15: Linking skills and employment — Jóvenes programmes in Latin America

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

9 1

students can use their qualifications to re-enterthe education system (Flores-Moreno, 2007).

While such examples demonstrate what is possible,second-chance education remains a highlyneglected area. Effective government coordinationof the wide range of public, private and other non-government actors involved is rare, partlybecause planning for second-chance programmesis seldom integrated into mainstream education.

There are other ways for governments toenhance skills development and combat youthunemployment. One of the most obvious is toensure that more young people complete theireducation and achieve a qualification. In the UnitedStates, the American Recovery and ReinvestmentAct includes financing provisions: youth who havenot finished secondary school can re-entereducation through a community college, vocationaltraining or apprenticeship. Several states haveintroduced programmes led by experiencedprincipals and teachers aimed at facilitatingsecondary school completion, offeringcomprehensive after-school and vacationteaching (CNN.com/US, 2009).

Governments can also combine education andemployment measures. Providing incentives forcompanies to offer apprenticeship and vocationalprogrammes to unskilled young people is oneoption. For example, the OECD has argued thatFrance should gear public assistance andincentives for apprenticeships towards unskilledyoung people and set a benchmark that increasesthe share of unskilled youth starting training from40% to 50% (OECD, 2009b). In the United Kingdom,which has some of the deepest skill-basedinequalities in the OECD, post-crisis interventionshave been generating employment and trainingfor long-term unemployed youth (Box 2.17).

Programmes that deliver results

How successful are technical and vocationalsystems in providing young people with skills,meeting company demands and tackling theproblems of youth unemployment, low wagesand insecurity? There are no easy answers to thesequestions. Vocational programmes do not operatein isolation. Macroeconomic conditions, labourmarket regulations and investment patterns havea major bearing on their effectiveness. Vocationaleducation has the potential to make a differencein the lives of young people. Yet that potential is

weakened in countries relying on top-down, supply-driven models in which governmentsdetermine priorities. Moving towards a demand-driven approach that responds to the needs ofindividuals, companies and the economy is theoverriding priority for reform.

Most rigorous evaluations of technical andvocational education programmes come fromdeveloped countries. Reviews that control forselection bias broadly suggest that vocationaleducation improves employment prospects butdoes not necessarily lead to higher pay (Adams,2007a; Bishop and Mañe, 2005; Ryan, 2001).Evidence from Europe indicates that apprenticeshipsystems reduce youth unemployment and raiseentry into higher-wage occupations (Gangl, 2003;Quintini et al., 2007). Traditional apprenticeshipprogrammes are marked by strong gender bias,however. They achieve far less for women in termsof jobs, careers and wages (Adams, 2007b).

Traditionalapprenticeshipsoffer far less for women interms of jobs,careers and wages

Even before the global downturn, job prospects for young Britonswere deteriorating and school-leavers without qualifications facedsevere employment disadvantages. From 2002 to 2007, the youthunemployment rate increased from 11% to 14%. With the recession,it has jumped to 17% — the highest level since 1993. Relativelyunskilled youngsters leaving school with poor qualifications arebearing the brunt.

Many of the weaknesses in the United Kingdom’s vocational traininghave deep historical roots. Apprenticeship systems have been basedon voluntary provision by employers, with little governmentinvolvement. Moreover, vocational qualification systems havesuffered from high levels of fragmentation and overspecialization.

Reforms were introduced in 2007 aimed at closing the skills gap.Under new legislation, young people will be required to participatein education and training until they obtain a qualification or turn 18.The qualification system is being overhauled and consolidated aroundseventeen new diplomas, and is set to become operational in 2015.These will be composite qualifications combining theoreticaland practical learning, and including an apprenticeship element.In parallel, long-term job seekers aged 18 to 24 are being offereda range of support and training options.

Responses to the financial crisis have built on this framework. Underthe 2009 budget, every 18- to 24-year-old unemployed for a yearor more is guaranteed an offer of training or a job, with fundingmade available through local authorities and voluntary organizations.Questions remain about the degree to which the training offered willequip young people for employment.

Sources: Children England (2009); OECD (2008c); UK Learning and Skills Council (2008).

Box 2.17: Skills and employment in the United Kingdom

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

9 2

As policy-makers seek to address the twinchallenges posed by rising unemployment andan increasingly knowledge-based economy,some important lessons may be drawn fromthe better-performing programmes – alongwith some cautionary notes.

Reinforce the links between education and labourmarkets. A major strength of the dual system inGermany is the direct link it establishes betweenschool, work experience and practical educationin vocational courses. Companies train studentsto acquire skills relevant to the needs of theenterprise and, through the involvement ofgovernment agencies, the wider economy. Usinga very different approach, Japan’s system hasprovided students with a route into company-basedtraining and employment. Contrasts with countriesincluding France and the United Kingdom, wherelinks between education and companies have beenfar weaker, are striking. In the United States,Career Academies operate through less formalcontractual arrangements, but establish stronglinks between students, companies and educators,combining practical employment opportunities withteaching and job counselling. Rigorous evaluationthat controls for selection bias points to strongbenefits, including an average earnings increaseof about 11% (Kemple and Willner, 2008).

Recognize that past achievements are noguarantee of future success. Rapid economicchange is continually shifting the environmentfor vocational education. Germany’s dual systemhas been coming under pressure as employmentgrowth slows in metalworking, engineeringand the automobile sector. The number of newapprenticeship places available is in decline(German Federal Ministry of Education andResearch, 2006). In Japan, the ‘lost decade’ ofprotracted recession of the 1990s led companiesto lower their commitments to training and long-term employment. This is reflected in thelarge and growing share of young workers ininsecure or temporary work (OECD, 2009c).The experience of Germany and Japan serves tohighlight the important role of economic growthand employment creation in creating demandamong employers for technical and vocationaleducation and training. It also underlines theneed for state action to renew vocationalprogrammes in the light of changingcircumstances, a task heightened by thecurrent economic downturn.

Rethink the outmoded separation of technicaland vocational education from general education.Successful participation in knowledge-basedemployment markets characterized by rapidchange requires problem-solving and creativethinking as well as specific technical skills. Thereis a growing sense in which ‘what you know’ is lessimportant than ‘what you are able to learn’. Rigidtracking into vocational training, especially at anearly age, diminishes the prospect of developingflexible skills and restricts individuals’ choices.Vocational students need sufficient academiceducation to broaden their occupational choicesand general students need an opportunity todevelop practical skills. Innovative reformers arebreaking down barriers between vocational andgeneral education. In the Republic of Korea,academic and vocational students in secondaryschool share as much as 75% of a joint curriculum,creating opportunities for transition in bothdirections (Adams, 2007b). The share of studentsenrolled in vocational education at the secondarylevel has been declining as the emphasis shiftsto general education to equip students for post-secondary specialization.29 Several other countries,including Australia and Switzerland, have activelyrevised qualification systems to allow for greatermobility between general and vocational education(Hoeckel et al., 2008a; Hoeckel et al., 2008b).

Develop capability-based qualification systems,involving the private sector. In job markets shapedby rapid technological change, young people needexpertise that can be applied to acquiring a widerange of skills. Many countries are introducing orstrengthening national qualification frameworks,testing students on the basis of broad abilitiesand allowing training to be used for transferrablecredits into technical and general education(Adams, 2007b; Hoeckel et al., 2008b; Young, 2005).Involving companies in the development ofcapability-based training is important becausethey are well placed to pick up employmentmarket signals. In Australia, programmesdeveloped through industry associations andeducation authorities have been introduced in thefinal year of secondary school. At the same time,national skills bodies are bringing togetheremployers, teachers and education ministriesto develop and deliver curricula that are relevantto the needs of industry. One of the big challengesis to coordinate the diverse array of partnersinvolved into an administrative framework thatavoids fragmentation and duplication.

29. In the Republicof Korea, the shareof secondary schoolenrolment in technicaland vocational educationdropped steadily fromaround 45% in the mid-1990s to 29% in 2005.

There is agrowing sense in which ‘what

you know’ is lessimportant than

‘what you areable to learn’

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t s k i l l s — e x p a n d i n g o p p o r t u n i t i e s i n t h e n e w g l o b a l e c o n o m y

9 3

Integrate vocational programmes into nationalskills strategies. Some of the most successfulmodels demonstrate that long-term planningof skills development can play a critical role inraising productivity, generating economic growthand creating employment. The Republic of Koreaand Singapore both aligned vocational programmeswith the needs of high-growth sectors, identifyingskills bottlenecks and, as the economy developed,gradually shifting the focus of training fromsecondary schools to specialized technicalinstitutes and higher education (Law, 2008; Lee,2008). More recently, Viet Nam has invested heavilyin technical and vocational education to improveskills in light manufacturing. There is scope forother developing regions to learn from East Asia,but the conditions for success are difficult toreproduce (Fredriksen and Tan, 2008). They includethe integration of vocational education into an activepolicy for industrial development, rapid economicgrowth, strong state capacity and – critically –rapid progress in expanding good-quality primaryand secondary education. Two distinctive featuresof the vocational success story in East Asia havebeen missing from the policy environment of manyother developing countries. The first is rapideconomic growth, which has created demand forskilled labour and resources for training. Second,provision of technical and vocational educationin countries such as the Republic of Korea andSingapore has been integrated into broad-basednational strategies for industrial development,employment creation and raising living standardsthrough higher levels of skills and productivity(Lall, 2001) (Box 2.18).

Conclusion

In recent decades, the rapid rise of knowledge-based economies, along with persistent youthunemployment and the marginalization of youngpeople lacking skills, has prompted governmentsto review and revalue technical and vocationaleducation. The economic crisis is another driverof change. Emerging reform models arechallenging the image of vocational programmesas second-class education.

Governments face very different types ofchallenges. The problems with the dual systemin Germany are not those of vocational education

in Ethiopia. As in other areas of education policy,vocational education is not amenable to quickfixes through the import of successful modelsfrom other countries. Policies have to be tailoredto reflect governments’ abilities to manage them,the realities of labour markets and educationsystems, and institutional history. What is clearis that no government can afford to ignore theimportance of skills and learning in supportingeconomic growth, combating poverty andovercoming social marginalization. Goal 3 ofthe Dakar Framework for Action sets out avision for the learning and skills agenda. Nowgovernments and the international communityurgently need to develop meaningful benchmarksfor measuring progress and credible policiesfor achieving greater equity.

By helping drive economic growth, overcome shortages of skilled labourand reduce social inequalities, technical and vocational education hasplayed a central role in turning Singapore into a high-income countrywith one of the world’s best-performing education systems. Theeducation minister has described the Institute of Technical Educationas ‘a shining jewel in our system’.

The Institute of Technical Education was established in the early 1990sin response to growing concerns over the education system’s abilityto meet the demands of a more productive economy and the needs ofthe young. It is meant for students who register low scores in generalacademic education. Courses are designed by government and industry.Companies value its graduates highly: over 90% of students wereemployed within six months of graduating in 2007. As the economyhas evolved, the institute has responded with innovative programmes,including partnering with global industry to set up centres of technologyin niche areas such as industrial automation, offering joint certificateswith companies such as Microsoft and linking with institutes in Germanyto offer diplomas in machine technology.

Perhaps the institute’s greatest success has been in combating thestigma associated with vocational education. Successive governmentshave invested heavily in training teachers, involving the private sectoras well, so that the institute’s facilities are comparable to those of thecountry’s universities. Qualifications from the institute can be usedas a route into tertiary-level technical education through polytechnics,or back into academic education through universities. The emphasison giving confidence to students and tackling the perception of technicaland vocational education as a sign of failure helps explain whySingapore’s model has succeeded where others have failed.

Source: Goh and Gopinathan (2008).

Box 2.18: Singapore’s ‘jewel in the system’

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

9 4

Youth and adult literacy

Goal 4: Achieving a 50 per cent improvementin levels of adult literacy by 2015, especially for women, and equitable access to basic and continuing education for all adults.

Youth and adult illiteracy is the price peopleand countries are paying for the past failures ofeducation systems. When people emerge from theirschool years lacking basic reading, writing andnumeracy skills, they face a lifetime of disadvantageas illiteracy diminishes their social and economicprospects and damages self-esteem. But theconsequences of illiteracy extend beyond theindividual. When people lack literacy, society asa whole suffers from lost opportunities for higherproductivity, shared prosperity and politicalparticipation (Fasih, 2008; Kinsella and He, 2009;UIS, 2008a; UNESCO, 2005). Beyond the individualand social costs, illiteracy is a violation of humanrights and a global blight on the human condition(Maddox, 2008; Oxenham, 2008). Eradicating itis one of the most urgent development challengesof the twenty-first century.

The international community has failed to riseto the challenge. At the World Education Forumin Dakar in 2000, governments pledged to achieve a50% improvement in levels of adult literacy by 2015.The pledge was ambitious, but the target wasachievable. Unfortunately, the goal will be missedby a large margin. In a world with 759 millionilliterate young people and adults, there has beena conspicuous lack of urgency and commitmentto literacy on the part of political leaders. The manyexceptions to the rule serve to demonstrate thatfar more could have been achieved – and that farmore can be done to get closer to the 2015 target.Among the key messages of this section:

Literacy remains among the most neglectedof all education goals. Progress towards the2015 target of halving illiteracy30 has been fartoo slow and uneven. With half the period forachieving the target having elapsed, the regionsfarthest behind have travelled between a half andtwo-thirds of the distance required. On currenttrends, there will be 710 million illiterate adultsworldwide in 2015. The evidence from monitoringis clear: unless far more is done to accelerateprogress, the 2015 targets will not be reached.

More rapid progress remains possible. Severalcountries have demonstrated through successfulpolicies that more rapid advance towards adultliteracy is possible. The National Literacy Missionin India and the Literate Brazil Programme(Programa Brasil Alfabetizado) both reflect astronger commitment to literacy by politicalleaders. Several countries have developed highlyinnovative programmes through partnershipslinking communities to governments and non-government groups. Better financing anda renewed effort to reach older adults arecritical to accelerated progress.

Far more has to be done to overcome the legacyof disadvantage in literacy. While gender gapsare narrowing, they remain very large – womenstill account for nearly two-thirds of the world’sadult illiterates. Failure to tackle genderdisparities and wider inequalities based onwealth, region, ethnicity and language areholding back progress.

This section is divided into two parts. Part 1provides a global overview of literacy and a post-Dakar progress report. It also looks ahead to 2015,providing a projection of where current trends willleave the world in relation to the goal of achievinga 50% improvement in levels of adult literacy.Part 2 looks at some of the countries that aremaking progress and identifies approaches thatare making a difference.

Progress since the Dakar forum

The precise meaning of ‘literacy’ continues to besubject to intense academic debate (Benavot, 2008;Fransman, 2005). Unlike the simple dichotomiesused in other areas – such as being ‘in school’ or‘out of school’ – there are no clear-cut dividing linesbetween the literate and non-literate. In any society,there is a continuum of literacy – and people witha fragile hold on literacy in youth can lose that holdin adulthood. However, academic debates overthe precise meaning of the word should not detractfrom common-sense depictions of what theexperience of illiteracy means to those affected.

Half a century ago, UNESCO defined a literateperson as someone ‘who can with understandingboth read and write a short simple statement onhis or her everyday life’ (UNESCO, 1958, p. 3).More recently, the Global Campaign for Educationhas extended this basic idea: ‘Literacy is aboutthe acquisition and use of reading, writing and

30. The target of achievinga 50% improvement inlevels of adult literacyis measured by lookingat the illiteracy rate,reflecting the originalformulation of the goal asexpressed in Jomtien in1990. The adult illiteracyrate is computed bydeducting the adultliteracy rate from 100.

Literacy remainsamong the mostneglected of alleducation goals

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t l i t e ra c y

9 5

numeracy skills, and thereby the development ofactive citizenship, improved health and livelihoodsand gender equality’ (Global Campaign forEducation and ActionAid International, 2005, p.13).A combination of these two broad definitionscaptures the reality of illiteracy as a conditionthat denies people opportunity.

The condition affects much of the world’s youthand adult population, especially women indeveloping countries. While all regions areaffected, a relatively small group of countrieswith large populations dominates the globalilliteracy headcount.

The illiteracy scourge continuesAn estimated 759 million adults – around 16%of the world’s population aged 15 and over – lackthe basic reading, writing and numeracy skillsneeded in everyday life (Table 2.6). More than halflive in South and West Asia, and another one-fifthin sub-Saharan Africa. Reflecting the legacy ofgender disparity in education, almost two in everythree adult illiterates are female (see annex,Statistical Table 2).

Measured in aggregate terms, adult illiteratesare heavily concentrated in a small group of large-population countries (Figure 2.29). Just twenty countries account for around 80% of global illiterates, with Bangladesh, China, India and Pakistan makingup over half the total. The data in this sectionhighlight the concentration of illiteracy in developingcountries. This should not deflect attention fromthe serious problems in rich countries, where largepockets of illiteracy contribute to wider patternsof social and economic marginalization (Box 2.19).

Aggregate figures mask differences in theincidence of illiteracy. Both South and West Asia,and sub-Saharan Africa have high illiteracy rates,with more than one in three adults affected in bothregions (Table 2.6). In sub-Saharan Africa, twelvecountries have illiteracy rates in excess of 50%;among these, in Burkina Faso, Guinea, Mali andthe Niger, more than 70% of the adult populationis illiterate (Figure 2.30). In the Arab States, theproportion is nearly one-third. Gender disparities area major contributor to the high adult illiteracy ratesin all three regions (see annex, Statistical Table 2).For instance:

About 759 millionadults lack thebasic reading,writing andnumeracy skillsneeded ineveryday life

Table 2.6: Adult (15 and over) illiteracy rates and numbers, by region, 2000–20071

Figure 2.29: Adult illiteracy is heavily concentratedin a small group of large-population countriesAdult (15 and over) illiterates (millions), top ten countries

Notes: The population used to generate the number of illiterates is from the United Nations Population Division estimates (2006 revision). For countries with national observed literacy data, the population used corresponds to the year of the census or survey.1. Data are for the most recent year available during the period specified. See the web version of the introduction to the statistical tables for a broader explanation of national literacy definitions, assessment methods, and sources and years of data.2. The illiteracy rate is calculated as 100 minus the literacy rate.Source: Annex, Statistical Table 2.

World

Developing countriesDeveloped countriesCountries in transition

Sub-Saharan AfricaArab StatesCentral AsiaEast Asia and the Pacific

East AsiaPacific

South and West AsiaLatin America and the Caribbean

CaribbeanLatin America

North America and Western EuropeCentral and Eastern Europe

16 759

20 7520.7 50.6 1

38 15329 58

1 0.77 1087 1067 2

36 3919 36

25 39 330.6 42 8

Illiteracyrates(%)2

Illiterates(millions)

Rest of the world, 218

Morocco, 10

Indonesia, 13

Brazil, 14

Egypt, 17

Nigeria, 23

Ethiopia, 27

Pakistan, 47

Bangladesh, 49

China, 71

India, 270

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

9 6

In Afghanistan, 87% of adult women and 57%of men were illiterate in 2000.

In Chad, Ethiopia and Mali, women are around1.5 times as likely as men to be illiterate.

In Algeria and Yemen, the illiteracy rates forfemales are more than twice those for men.

Contrary to common understanding, therelationship between average income and literacyis highly variable. For example, Egypt’s averageincome is comparable to that of Ecuador, but itsliteracy rate is 66% while Ecuador’s is 84%.Similarly, Algeria has a far higher level of averageincome than Bolivia but a lower adult literacy rate.In both cases, gender disparities explain much ofthe discrepancy (see annex, Statistical Table 2).

The contrasting profiles for national literacy pointto distinctive policy challenges. As well as makingsure that young people emerge from educationsystems with basic literacy skills, many countriesin sub-Saharan Africa, South and West Asia andparts of the Arab world need to extend literacyopportunities to a large share of the adultpopulation. In Brazil and Indonesia, whereilliteracy affects 10% or less of the adult population,policy-makers still have to address the task ofreaching highly marginalized groups and people,many of them in remote areas.

Measuring literacy is not an exact science. Nationalestimates are typically derived from census andhousehold surveys in which people are asked toreport on their own literacy status (Box 2.20).Because the idea of literacy is specific to differentcultures and contexts, the word itself can havedifferent meanings to different people (Fransman,2005; UIS, 2008a). National surveys often fail togenerate representative data for populations thatare hard to reach or people living in informalsettlements (Aderinoye and Rogers, 2005). Giventhat literacy levels are likely to be lower amongthese groups this can also lead to underestimationof the numbers of illiterates.

The progress reportThe world is far off track for the 2015 target ofachieving a 50% improvement in levels of adultliteracy. In the absence of a concerted internationaldrive to prioritize literacy, there is little prospectof the target being brought within reach. Yet theexperience of some countries, and of someprogrammes within countries, demonstrates thata great deal can be achieved in relatively little time.

Adult literacy rates in the developing world havebeen rising with every school generation. As morechildren enter school and leave with basic literacyskills, literacy rates inevitably rise. Literacyprogrammes have also played a positive role insome countries. From 1985–1994 to 2000–2007,

The world is faroff track for the

2015 target ofachieving a 50%

improvement in levels of

adult literacy

This section focuses on illiteracy in poor countries,but rich countries also have significant pockets ofdeprivation. Many adults lack the functional literacyskills they need to apply for jobs, read newspapersor understand documents — on housing, healthand the education of children, for example — that affect their lives:

In France, an estimated 9% of people aged 18 to 65 lack the basic reading, writing, arithmetic and other fundamental skills required for simpleeveryday situations.

In the Netherlands, 1.5 million adults (including1 million native Dutch speakers) are classified asfunctionally illiterate, implying that they are notequipped to process basic information.

In the United States, 14% of the population lacksthe literacy skills to perform simple, everyday tasks like understanding newspaper articles andinstruction manuals. Around 12% lack the literacy

skills needed to fill out a job application orunderstand labels on food and drugs. More thanone in five — 22% of the population — has ‘below basic’ quantitative skills, finding it impossibleto balance a chequebook or deduce from anadvertisement the amount of interest on a loan.

In England (United Kingdom), 1.7 million people (5% of those aged 16 to 65) perform below thelevel expected of 7-year-olds on the nationalcurriculum test, and 5.1 million perform belowthe level expected of 11-year-olds.

Literacy problems in rich countries are oftenconcentrated in areas of acute social disadvantage,among migrant groups and the poor. Illiteracyis a factor in low pay, insecure employmentand social exclusion.

Sources: Burd-Sharps et al. (2008); National Agency to FightIlliteracy (2007); National Literacy Trust (2009); Reading andWriting Foundation (2009).

Box 2.19: Rich countries — poor literacy

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t l i t e ra c y

9 7

the number of adult illiterates in the world fell by13%. Given that population growth pushed the adultpopulation up by around 30% in the same period,the net effect is clearly positive. Adult literacy levelsincreased more rapidly than in the 1990s (Qiao,2007), growing by 10% to reach 84% in 2000-2007.

The broadly positive global canvas hides someless encouraging developments (Figure 2.31).Almost all the decline in the number of illiterateadults in the developing world took place in justone region, East Asia and the Pacific. In Southand West Asia, population growth cancelledout the decline in numbers of illiterate adults.In sub-Saharan Africa, the number of illiteratesincreased by 19.5 million. The Arab States alsoexperienced an increase. Some countrieswitnessed large absolute increases in thenumber of illiterate adults: over 1 million inBurkina Faso, the Philippines, Senegal, theUnited Republic of Tanzania and Viet Nam,and 4 million in Bangladesh and Ethiopia(see annex, Statistical Table 2).

The idea that countries are powerless tocombat adult illiteracy is refuted by theexperience of countries that have achieved rapidprogress. The following are examples of positivechange between 1985–1994 and 2000–2007:

Much of the illiteracy reduction in East Asiacan be traced to China. The number ofadult illiterates there fell by 114 millionor 62% between the two periods, with anaverage increase in the number of adultliterates of 4 million between 1990 and 2000(NCEDR, 2008).

Adult literacylevels grew by10% to reach 84%in 2000-2007

New approaches to literacy measurement are attemptingto address long-standing data problems. One prominentexample is the Literacy Assessment and MonitoringProgramme (LAMP).

Conventional approaches to literacy measurement are oftenfundamentally flawed. Asking people to report whether theyare literate is of limited use in assessing real capabilities.Similarly, testing literacy by reference to words, objects andexperiences that have no relevance in the lives of the peoplebeing surveyed can understate achievement levels.

The LAMP approach tests literacy in three domains:continuous texts (prose), non-continuous texts (documents)and numeracy. Results reflect a continuum of achievement,and the tests are designed to be meaningful to respondents.Data generated through the tests are intended for nationaland cross-national comparisons. Developed by the UISand administered through ministries of education, LAMPsurveys are in the pilot stage in several countries.

Sources: UIS (2009c, 2009d).

Box 2.20: A new generation of literacy statisticsFigure 2.30: In developing countries, illiteracy can affectfrom one to three out of four adultsAdult (15 and over) illiteracy rates in countries with rates of 25% or more in selected regions, 2000–2007 1

Notes: For countries indicated with *, national observed literacy data are used.For all others, UIS literacy estimates are used. The estimates were generated usingthe UIS Global Age-specific Literacy Projections model. Figures in parentheses afterregion names indicate the number of countries with publishable data in the region.1. Data are for the most recent year available during the period specified. See theweb version of the introduction to the statistical tables for a broader explanation of national literacy definitions, assessment methods, and sources and years of data. 2. Data do not include all geographic regions.Source: Annex, Statistical Table 2.

*,2

*

**

*

*

***

*

*

***

*

****

IraqEgypt

SudanYemen

MauritaniaMorocco

IndiaNepal

PakistanBangladesh

BhutanAfghanistan

KenyaUganda

U. R. TanzaniaNigeriaMalawi

MadagascarZambia

CameroonAngola

D. R. CongoGhana

RwandaEritrea

BurundiLiberia

TogoCôte d’Ivoire

C. A. R.Mozambique

SenegalBenin

Sierra LeoneEthiopia

ChadGuinea

Burkina FasoNigerMali

0 10 20 30 40 50 60 70 80

Arab States (19)

South and West Asia (9)

Sub-Saharan Africa (41)

Adult illiteracy rates (%)

Regionalaverage29%

Regionalaverage36%

Regionalaverage38%

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

9 8

With the world’s largest illiterate population,India has been making progress. In 1985–1994not quite half of adults were literate. The figureis now slightly above two-thirds. Since the adultpopulation increased by 45%, this marks a realadvance. It also suggests that the country’sTotal Literacy Campaign, under the Auspicesof the National Literacy Mission, may be havingan impact (India Ministry of Human ResourceDevelopment and National University ofEducational Planning and Administration, 2008).

Many countries in sub-Saharan Africa haveachieved steep rises in adult literacy rates.Burkina Faso and Chad, with some of theworld’s lowest literacy rates, have respectivelydoubled and almost tripled their rates.

Several Arab states have achieved majoradvances. Egypt’s adult literacy rate hasincreased from 44% to 66%. Yemen hasincreased the adult literacy rate from 37% to 59%.

Gender parity is improvingRising literacy has been accompanied bydeclining gender disparities. Many countriesthat started with very large gaps between maleand female literacy, and from low overall levels,have been on a pathway towards parity.

Gender parity improved in all but eight of theseventy-nine countries with data.31 In Bangladesh,Burkina Faso, Burundi, Malawi, Nepal andYemen, female adult literacy rates doubled or

31 The gender parity index(GPI) declined in Ecuador,Ethiopia and Zambia.In Botswana, the GPIimproved from 1.09 to1.00. Gender parity hadalready been achieved in1985-1994 in the Maldives,Panama, Seychellesand Uruguay, and wasmaintained in 2000–2007(Annex, StatisticalTable 2).

Burkina Faso and Chad have

respectivelydoubled

and almosttripled their

literacy rates

0 10-10 20-20 30-30 40-40 50-50 60-60 70-70

Per cent change, 1985-1994 and 2000-2007

Adult illiterates

Adult population

Literacy rates

World

Developing countries

Developed countries

Countries in transition

Arab States

Central and Eastern Europe

Central Asia

East Asia and the Pacific

Latin America and the Caribbean

North America and Western Europe

South and West Asia

Sub-Saharan Africa

Figure 2.31: The number of adult illiterates is falling despite population growthChanges in adult (15 and over) illiterates, literacy rates and population, by region, 1985–1994 and 2000–2007 1

Notes: The population used to generate the number of illiterates is from the United Nations Population Division estimates (2006 revision). For countries with nationalobserved literacy data, the population used corresponds to the year of the census or survey. For countries with UIS estimates, the populations used are for 1994 and 2007.1. Data are for the most recent year available during each period specified.Source: Annex, Statistical Table 2.

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t l i t e ra c y

9 9

tripled, and have increased twice as fast asmale rates. Because adult literacy gaps trackdevelopments in basic education, this catchingup process mirrors a narrowing of the gender gapin basic education. Between the two benchmarkperiods, the number of adult female literatesincreased by 14%, compared with 7% for adultmales (see annex, Statistical Table 2).

This positive trend has to be placed in context.Women may be catching up, but in many countriesthey are starting from a long way back. Genderdisparities remain very deep – and the share ofwomen in the total number of illiterates hasincreased slightly. The process of convergence isthus starting from very unequal points (Figure 2.32).In the three regions with the lowest levels of literacyand largest gender disparities – the Arab states,South and West Asia, and sub-Saharan Africa –female adult literacy rates for 2000–2007 were stillbelow the average for male literacy in 1985–1994.On the current trajectory, it will take women in South and West Asia about fifty-six years to catch up.

Gender convergence in adult literacy is proceedingat different rates in different countries. Compare thecontrasting experiences of Bangladesh and India.Women aged 25 to 34 in Bangladesh have illiteracyrates 32% higher than men in the same age group.The gap reflects gender disparities that prevailed inthe education system when that generation went toschool. For 15- to 24-year-olds in Bangladesh todaythe gender gap has been eliminated. While India hasbeen narrowing the gap, 15- to 24-year-old femalesare still about twice as likely to be illiterate asmales in that group. Among the Arab states,Morocco has been making rapid progress towardsimproved literacy with every school generation buthas been less successful in closing the gender gap,as comparison with China underlines (Figure 2.33).

Improvement in access to education acrossgenerations is one of the motors driving increasedliteracy levels. In almost all countries, literacy ratesamong younger adults (15 to 24) are higher thanthe average for all adults (15 and over). In the ArabStates, South and West Asia, and sub-SaharanAfrica, youth literacy rates in 2000–2007 were16% to 24% higher than the average for all adults(see annex, Statistical Table 2). Age-groupdisparities are particularly marked in somecountries, including Botswana, Eritrea, the IslamicRepublic of Iran, Nigeria and Sri Lanka, where theproportion of illiterates among all adults is doubleor more the proportion for younger adults.

Illiteracy mirrors wider disadvantagesNational data on literacy can provide insights intothe average picture for a country while obscuringdisparities within countries, where adult illiteracymay intersect with income, parental education,ethnicity, language and disability. While womenare systematically disadvantaged, genderdisparities are magnified by wider structuresof disadvantage and marginalization.

Low income. Adults from the pooresthouseholds are far more likely to be illiterate.In Guatemala, 60% of adults living in extremepoverty and 42% of those living in non-extremepoverty are illiterate, compared with 17% of richeradults (Porta Pallais and Laguna, 2007). Similarly,the literacy rate for the richest Bangladeshihouseholds is 76%, compared with 28% forthe poorest (Bangladesh Ministry of Planningand UNESCO Bangladesh, 2008).

Ethnicity, language and group-baseddisadvantage. Minority language groups andindigenous people often register far lower levelsof literacy. In Viet Nam, the literacy rate is 94%among the majority Kinh population, but only 72%for ethnic minorities (Daswani, 2005). In Peru,illiteracy is much more prevalent among

On the currenttrajectory, it willtake women inSouth and WestAsia about fifty-sixyears to catchup with men

30 40 50 60 70 80 90 100

World

Developing countries

Developed countries

Countries in transition

Sub-Saharan Africa

Arab States

Central Asia

East Asia and the Pacific

South and West Asia

Latin America/Caribbean

Central/Eastern Europe

N. America/W. Europe

Adult literacy rates (%)

Female 1985-1994 Female 2000-2007

Male 1985-1994 Male 2000-2007

Figure 2.32: Being so far behind, women have further to travel to reach male literacy ratesAdult (15 and over) literacy rates, by region and gender, 1985–1994 and 2000–2007 1

1. Data are for the most recent year available during each period specified.Source: Annex, Statistical Table 2.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 0 0

indigenous-language speakers at 21% of adultscompared with 4% for Spanish-speakers (Cuetoet al., 2009). In South Asia, literacy gaps betweenlower and higher castes are pronounced. In Nepal,caste disparities are even larger than wealthand gender disparities (Nepal Ministry of Educationand Sports and UNESCO Kathmandu, 2007).

Disparities linked to location. Illiteracy tends tobe higher in poorer regions, rural areas and slums.Regional disparities often mirror national povertymaps. For example, in Brazil some of the pooreststates in the north-east – Alagoas, Maranhão,Paraíba and Piauí – have illiteracy levels twice ashigh as in the south-east (The George WashingtonUniversity, 2006). In India, the regional spectrumextends from almost no illiteracy in the stateof Mizoram to 50% illiteracy in Rajasthan (IndiaMinistry of Human Resource Development andNational University of Educational Planning andAdministration, 2008). Rural areas often lag farbehind urban areas (Kinsella and He, 2009).

In Pakistan, urban literacy rates are twice as highas the rural average. Within urban areas illiteracytends to be concentrated in informal settlementscharacterized by high levels of poverty (PakistanMinistry of Education, 2008).

None of these disadvantages exists in isolation.Being female is a near-universal indicator for loweraverage literacy in many Arab states and most ofSouth and West Asia, and sub-Saharan Africa. Butgender disadvantage is compounded by poverty,location and ethnicity (Figure 2.34). The wealth gapin the Philippines is particularly marked: women inthe poorest households have literacy rates averaging65%, compared with over 96% for women in thewealthiest households. In South Africa, white youthand adult women have near-universal literacy levels,compared with just 70% literacy among blackwomen. In Mexico, women who only speak anindigenous language are about fifteen times lesslikely to be literate than women who only speakSpanish, and women lacking a knowledge ofSpanish have literacy levels of just 5%. Literacyrates among Cambodian women living in Ratanakiri,a province dominated by indigenous hill tribes, arejust over a third of those among women in thecapital, Phnom Penh.

Prospects for achieving the 2015 targetCurrent trends in adult literacy will leave the worldshort of the target set for 2015.32 Progress has beenso slow that the target is out of reach. Even in abest-case scenario, not enough children will enteradulthood literate over the next five years to halvethe level of illiteracy. Continuing on the current trendwill leave a very large gap with the Dakar promise.

Projections provided for the EFA Global MonitoringReport 2010 give a ballpark estimate of the scaleof shortfall: by 2015 the adult illiteracy rate will havefallen between 29% and 34% in the three regionswith the highest rates. In other words, between halfand two-thirds of the journey to the 50% illiteracyreduction target will have been completed.

There is a very real human cost associated withthe gap. On the current course, an estimated710 million adults – 13% of the world’s adults – willstill lack basic literacy skills in 2015. Regional gapsbetween target and projected outcome are largestfor South and West Asia, sub-Saharan Africa andthe Arab States (Figure 2.35). Failure to achieve theDakar adult literacy goal will translate into verylarge deficits for many countries. In India the target will be missed on current trends by around 81 million

32. See note forFigure 2.35 for howthe literacy target hasbeen measured.

On currenttrends,

710 million adultswill still lack

basic literacyskills in 2015.

0

20

40

60

80

100

75-84 65-74 55-64 45-54 35-44 25-34 15-24

Illite

racy

rate

s (%

)

India, male

Bangladesh, femaleBangladesh, male

India, female

Age groups

0

20

40

60

80

100

75-84 65-74 55-64 45-54 35-44 25-34 15-24

Illite

racy

rate

s (%

)

Morocco, female

Morocco, male

China, maleChina, female

Age groups

Figure 2.33: Contrasting experiences in reducing illiteracy and the associated gender gap in four countriesAge illiteracy profile in selected countries, by age group and gender, 2007

Source: UIS database.

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t l i t e ra c y

1 0 1

people. Bangladesh will have 16 million more illiteratesthan it would if the 2015 goals were achieved.33

Among the sub-Saharan African countries,Mozambique will face a deficit of 2.6 million people,based on a target set using 1997 data (UIS, 2009d).

Changing the trend — making the literacy decade count

The disappointing progress towards the literacy goalset at Dakar reflects a collective failure of politicalcommitment. While there are many exceptions,governments and aid donors collectively have failedto attach sufficient weight to the eradication ofilliteracy. There are encouraging signs, however,that this could be starting to change.

Some are apparent at the international level. In 2003,the United Nations launched a literacy decade, with

33. The Bangladesh and India targets are set using 2001 literacy datafrom the UIS database.

0

10

20

30

40

50

60

70

80

90

100

Fem

ale

liter

acy r

ates

(%)

Burkina Faso Ethiopia India Guatemala Cambodia Mexico South Africa Philippines Peru

Ucayali region

Castellano ethnicityCapital regionWhite ethnicity

Distrito FederalPhnom Penhregion

Ladino ethnicityMetropolitan/Central region

Spanish languageSpanishlanguage only

Indigenouslanguage only

Oaxaca region

Mondol Kiri/Rattanak Kiri region

Other language

North-OccidentalregionIndian ethnicity

W. Cape region

Black ethnicity

Limpopo region

Christian

ARMM*Muslim

Cusco region

Quechua ethnicity

Kerala region

Addis Abebaregion

Ouagadougouregion

East region Somali regionFulfulde/Peul ethnicity

Sénoufo ethnicity

Rajasthan region

Poorest

Average

Richest

Urban

Rural

Figure 2.34: Within countries, women’s literacy rates are influenced by socio-economic and geographic factorsFemale adult (18 and over) literacy rates, by area, ethnicity, income, language, religion or region, selected countries, latest available year

Notes: Data for Burkina Faso (2003), Cambodia (2005), Ethiopia (2005), Guatemala (1999), India (2005) and Peru (2004) are from Demographic and Health Surveys. Data for Mexico (2005), the Philippines (2000) and South Africa (2001) are from population censuses. Countries are sorted by average female literacy rate.* Autonomous Region in Muslim Mindanao.Source: UNESCO-DME (2009).

Figure 2.35: At the present rate, regions furthest behind will miss the literacy target for 2015Adult (15 and over) illiteracy rates, by region: 2007, projected by 2015 and required by 2015 to achieve the goal

Notes: The goal of achieving a 50% improvement in levels of adult literacy is measuredby looking at the 1999 illiteracy rate, reflecting the original formulation of the goal asexpressed in Jomtien in 1991. The adult illiteracy rate is computed by deducting theadult literacy rate from 100. Regions are sorted by illiteracy rate goal by 2015.Sources: Annex, Statistical Table 2; UIS (2009b).

East Asia/Pacific

Latin America/Caribbean

Arab States

Sub-Saharan Africa

South/West Asia

100 20 30 40

Adult illiteracy rates (%)

Projected 2015

2007

Goal by 2015

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 0 2

governments recognizing that ‘literacy is crucialto the acquisition, by every child, youth and adult,of essential life skills that enable them to addressthe challenges they can face in life, and representsan essential step in basic education’ (UnitedNations, 2002, p. 2). While development decadescome and go, usually without meaningful impact,the literacy decade has given rise to intensiveregional discussions and raised the profile of theilliteracy problem (Robinson, 2009; UNESCO,2008b). The International Conference on AdultEducation scheduled for late 2009 (CONFINTEA VI)provides an opportunity to move from internationaldialogue to international action.

Literacy continues to receive insufficient attentionat many levels. It is not treated as a politicalpriority, it receives insufficient financialcommitment and efforts to incorporate strategiesfor literacy into wider poverty reduction plansremain underdeveloped (Caillods and Hallak, 2004;Giffard-Lindsay, 2008; Lindt, 2008). Even so, somegovernments have demonstrated through practicalaction that national programmes deliver results.Others have increased financing commitmentsfor literacy. And a vast array of partnerships andapproaches are now promoting literacy at thecommunity level (Oxenham, 2008).

Some governments have openly acknowledgedthat neglect of literacy was a serious policy failure(Lindt, 2008). One of the most far-reaching effortsto correct that failure is the Literate BrazilProgramme (Box 2.21). In the Islamic Republicof Iran, community learning centres initiated bythe Literacy Movement Organization, a governmentagency, have enrolled 3.1 million illiterates from2000 to 2006 in preliminary basic educationcourses. Around three-quarters of those enrolledsuccessfully complete their courses (Richmondet al., 2008). In Burkina Faso, the government hasadopted the bold target of increasing adult literacyrates from 28% to 40% by 2010. That target hasbeen backed by an increase in the share of theeducation budget allocated to literacy from 1% to7% – a move that has facilitated the expansion ofpermanent literacy training centres and centresfor non-formal basic education. Graduation fromthese centres grew by 24% from 2003 to 2007(Richmond et al., 2008).

Another example comes from India, where theNational Literacy Mission, launched in 1988, hasbeen revitalized. The eleventh five-year plan, whichends in 2012, has quintupled the mission’s budget

to the equivalent of US$21 billion. Programmeshave been redesigned to provide an integratedapproach that combines initial literacy trainingwith ongoing post-literacy courses.Decentralization is transferring authority to statesand districts, and a much stronger commitmenthas been made to preparing literacy materials inlocal languages. In 2009, the Government of Indiaalso signaled a stronger focus on gender andequity, first by recasting the National LiteracyMission as the National Female Literacy Missionand then by announcing a strategy for targeting.Commitments have been made to ensure that85% of targeted beneficiaries will be women andthat 50% will come from scheduled castes, tribesand minorities, with a focus on Muslims (IndiaMinistry of Human Resource Development, 2009).

Slow progress in improving literacy is sometimescited as evidence that little can be done for oldergenerations. There is extensive evidence, however,of problems with past approaches to raisingliteracy. For two or three decades after 1960,many governments attempted to combat illiteracythrough top-down courses that were ill suitedto the lives of the intended beneficiaries, badlydesigned and offered no follow-up. Dropout rateswere high and literacy acquisition limited. Theneeds of indigenous people and minority languagegroups were often ignored. Literacy programmeshave mirrored schools in denying people anopportunity to learn in their local language,diminishing the perceived value of their culturein the process.

Shortcomings persist in national programmes,particularly in targeting. Literacy initiatives oftenfocus on youth and young adults, with insufficientattention paid to older people – especiallywomen – who represent the bulk of the illiteratepopulation. India’s and Brazil’s programmesprincipally target people under 30 (India Ministryof Human Resource Development and NationalUniversity of Educational Planning andAdministration, 2008). Reaching older illiteratescan be difficult, but far more could be done toextend opportunities through livelihood-basedliteracy programmes.

Financing is another area of concern. It isencouraging that more governments are adoptingbold targets, but those targets are seldom backedby adequate budget support. It is not uncommonfor literacy to account for as little as 1% to2% of total education spending (Lindt, 2008).

Large-scaleprogrammes

in Brazil, Indiaand the IslamicRepublic of Iran

are deliveringpositive results

P R O G R E S S T O WA R D S T H E E FA G O A L S

Yo u t h a n d a d u l t l i t e ra c y

1 0 3

Many literacy programmes also continue to sufferfrom low rates of uptake and completion.

On the other hand, programmes that provide foractive learning through a relevant curriculum andoffer follow-up have achieved results. Many suchprogrammes are built on partnerships, extendingfrom local communities to non-government actorsand government agencies. One prominentexample is Reflect. Developed and coordinatedby ActionAid, an international charity, it focuseson the learners’ own literacy objectives, motivationand skills (Riddell, 2001). It not only promotes theuse of real texts from the environment but alsoencourages participants to generate their owntexts. It further aims to transform the broaderliteracy environment, for example by campaigningfor newspapers to use local language or texts moreaccessible to adult literacy learners (Aderinoyeand Rogers, 2005). The programme is currentlyapplied in Bangladesh, Pakistan and many othercountries (Duffy et al., 2008).

Bilingual education is critical to the success ofliteracy programmes aimed at indigenous peopleand ethnic minorities. Here, too, many governmentshave openly acknowledged the mistakes of the past.

Several governments in Latin America – includingthose of Chile, Guatemala, Mexico, Paraguay andPeru – and the UN Economic Commission for LatinAmerica and the Caribbean have developed aregional poverty reduction strategy incorporatingbilingual literacy training for indigenous groups(Latin American and the Caribbean DemographicCenter, 2009; Stockholm Challenge, 2008).

Conclusion

Much has been achieved through the scaling upof literacy initiatives since 2000. However, themonitoring evidence is unambiguous: the 2015targets will not be reached on the currenttrajectory. Far more has to be done to accelerateprogress. This will require stronger politicalleadership. Governments across the world continueto attach too little weight to literacy in nationalplanning. This is short-sighted. Illiteracy imposeshuge costs on society and the economy – andinvestments in literacy have the potential togenerate large returns in both areas.

The Literate BrazilProgramme hasprovided literacytraining to about 8 million learners

Around 14 million Brazilian youth, adults and elderlypeople lack basic reading and writing skills. TheLiterate Brazil Programme (Programa BrasilAlfabetizado) initiated by President Luiz Inácio Lulada Silva in 2003 is the first concerted national effortto consign illiteracy to the history books.

The programme is coordinated by the Ministry ofEducation but operates through a highly decentralizedstructure. It functions in 3,699 municipalities, justover 1,000 of which have been accorded prioritystatus because they have illiteracy rates over 25%.The effort primarily targets disadvantaged groupssuch as indigenous people, small farmers and farmworkers, child labourers (as part of the Programmeto Eradicate Child Labour) and people covered underthe Bolsa Familia social protection programme.

Literate Brazil is open to anyone over 15 with less thana year of education. Thus far it has provided literacytraining to about 8 million learners. Literacy classestypically last six to eight months and are attendedby groups of eighteen to twenty-five learners. Federaltransfers cover the cost of training and providinggrants to literacy facilitators, many of whom are

teachers. Literacy textbooks are produced in locallanguages and reflect local circumstances and needs.Innovative pedagogical approaches have beendeveloped. Teaching is organized around the ideaof ‘making people literate’ through dynamic learningprocesses that lead to the acquisition of reading,writing and numeracy skills. Students’ abilities areassessed and recorded by government agencies,and qualifications are provided that can be usedto enter formal education.

Literate Brazil has been about more than providingservices. Political leaders have challenged the cultureof silence and indifference surrounding illiteracy.A Ministry of Education programme has put thedevelopment of textbooks for literacy, a previouslyneglected subject, on the same footing as books forprimary and secondary school. Prizes are awardedfor the development of literacy materials and the bestentries are integrated into national programmes, withspecific prizes for black Brazilians, an educationallydisadvantaged group.

Sources: Brazil Ministry of Education (2008); Henriques and Ireland (2007); Ireland (2007, 2008); UIL (2009); UNESCO Brasilia (2009).

Box 2.21: Brazil — ‘making people literate’

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 0 4

The quality of education

Goal 6: Improving all aspects of the quality ofeducation and ensuring excellence of all so thatrecognized and measurable learning outcomesare achieved by all, especially in literacy,numeracy and essential life skills.

The core task of any education system is toequip young people with the skills they need toparticipate in social, economic and political life.Getting children into primary school, through theirearly grades and into secondary school is not anend in itself but a means of delivering these skills.Success or failure in achieving education for allhinges critically not just on countries deliveringmore years in school; the ultimate measure liesin what children learn and the quality of theireducation experience.

Many countries are failing the quality test. Out-of-school children face obvious disadvantages,yet less attention has been paid to the fact thatmillions of children emerge from primary schooleach year without having acquired basic literacyand numeracy skills. Unable to formulate or reada simple sentence, these children are ill equippedto make the transition to secondary school – letalone enter employment markets. The problemsextend to secondary schools, where many children– sometimes a majority – do not reach even aminimal level of competence.

Policy-makers, educators and parents need tofocus far more on the core purpose of education:ensuring that children acquire the skills thatshape their future life chances. That goal isdifficult to achieve – far more difficult, arguably,than getting children into school. Governmentsneed to revise approaches to teaching, learningand curriculum development. With the globalfinancial crisis having tightened already severebudget constraints, cost is often a barrier, butlearning achievement can be greatly improvedat low cost, in some cases by making better useof resources already being invested in education.

Public concern over the quality of educationis evident in many of the world’s richest nations,as well as the poorest. This section focuses onthe situation in developing countries. There arethree key messages:

While global gaps in access to school may benarrowing, gaps in school quality remainenormous. Evidence from learning achievementtests suggests that, in many developing countries,average students are performing close to or belowminimum competency levels. Global disparitiesare reinforced by inequalities within countries. Theproblem is not just one of relative performance;absolute levels of learning achievement areexceptionally low in many countries.

Getting the basics right is important – and manycountries are failing to build strong foundations.Children in the early grades are not masteringthe reading skills necessary for further learning.Without these foundations, returns on the hugeinvestment that governments and householdsmake in education will be sub-optimal. Readingskills can be improved relatively easily. Educationministries and teachers need to renew theirefforts regarding these basic skills.

Children do not start their schooling on anequal footing: more must be done to equalizeopportunity. Circumstances beyond children’scontrol, such as the income and education oftheir parents, the language they speak and wherethey live, influence their achievement at school.If the quality goal is to be achieved, ensuring thatall learners, regardless of background, achievebasic levels of learning needs to becomea central objective. Programmes to improveachievement for the most disadvantagedlearners are necessary.

The section is divided into three parts. Part 1highlights the large disparities in learningachievement among and within countries. Part 2explores early grade reading – one of thefoundations for learning. Part 3 looks at the widerchallenge of improving learning in schools and atglobal trends in teacher recruitment.

The learning gap — from global to local

In an increasingly knowledge-based world,prosperity, employment and poverty reduction –for countries and individuals – depend increasinglyon skills and capabilities delivered in the classroom.For large parts of the world’s population, however,education systems fall far short of legitimateexpectations. Poor quality in education isjeopardizing the future of millions of young people,many of whom face the prospect of lifelong illiteracy.

The core purposeof education is to

ensure thatchildren acquire

the skills thatshape their

future lifechances

P R O G R E S S T O WA R D S T H E E FA G O A L S

T h e q u a l i t y o f e d u c a t i o n

1 0 5

Cross-country inequalities and achievement deficitsWhile significant gaps remain, more and morecountries are participating in global and regionalassessment exercises that make it possible tomeasure disparities between countries in termsof the skills students attain after a given periodof learning.

The fourth cycle of Trends in InternationalMathematics and Science Study (TIMSS),conducted in 2007 among eighth grade students,shows large gaps in learning achievement betweencountries (Figure 2.36). One way of looking at thesegaps is to consider the range of results. Averagetest scores for students in the Republic of Korea,the top-performing country, were almost twiceas high as for students in Ghana, at the bottomof the league. Viewed from a different vantagepoint, the average student in El Salvador, Ghana,Indonesia and Morocco figures alongside orbelow the poorest-performing 10% of studentsin higher-performing countries.

Few of the poorest developing countriesparticipated in TIMSS 2007. Researchers have

attempted to address this limitation byreconfiguring scores from wider test exercises.The results confirm that low-income countries lagfar behind others in learning achievement (Altinok,2008; Hanushek and Woessmann, 2009). Oneassessment in India, conducted in the states ofOrissa and Rajasthan during 2005, used questionsfrom TIMSS to see how students in these statescompared with those in countries participating inthe original TIMSS survey (Das and Zajonc, 2008).The results showed that ninth grade students inOrissa and Rajasthan ranked alongside studentsfrom the poorest-performing TIMSS countries.

Learning assessments allow for more thanrelative measurement. TIMSS establishes a seriesof performance thresholds aimed at measuringstudent capabilities. At the low end (scores of400 or less), students have only the most basicknowledge of whole numbers, decimals and basicgraphs. At the upper end (over 550), students canapply their understanding and knowledge in avariety of complex situations. In eighteen of thecountries covered, including Botswana, Egypt andSaudi Arabia, the average student performs belowthe low threshold (Figure 2.36). This points to

There are largegaps in learningachievementbetween countries

Rep.

of K

orea

Japa

nHu

ngar

yEn

glan

d (U

. K.)

Unite

d St

ates

Czec

h Re

p.Au

stra

liaSw

eden

Scot

land

(U. K

.)Ita

lyNo

rway

Sing

apor

eSl

oven

iaM

alta

Cypr

usIsr

ael

Bahr

ain

Oman

Kuw

ait

Saud

i Ara

bia

Qata

r Ru

ssia

n Fe

d.Lit

huan

iaAr

men

iaSe

rbia

Mal

aysia

Bulg

aria

Ukra

ine

Rom

ania

Bosn

ia/H

erze

g.Le

bano

nTh

aila

ndJo

rdan

Turk

eyTu

nisia

Geor

gia

Iran,

Isl.

Rep.

Indo

nesia

Syria

n A.

R.

Egyp

tAl

geria

Colo

mbi

aM

oroc

coPa

lest

inia

n A.

T.Bo

tsw

ana

El S

alva

dor

Ghan

a0

100

200

300

400

500

600

700

800 Above theTIMSS highinternationalbenchmark

Below theTIMSS low international benchmark

TIM

SS m

athe

mat

ics sc

ale

scor

e

95th90th

5th10th

50th percentile

OECD Other high income Middle and low income

Figure 2.36: There are large gaps in learning achievement across countriesDistribution of TIMSS mathematics scale score for eighth grade students in 2007

Notes: The markers show the scale score of the indicated percentile. The fiftieth percentile is the median score for the country. This represents the middle of the distribution, with 50% of students scoring above and below the median. Low benchmark – students have some knowledge of whole numbers and decimals, operations and basic graphs.High benchmark – students can organise and draw conclusions from information, make generalizations and solve non-routine problems.Source: Mullis et al. (2008).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 0 6

failings in education systems. While average testscores are higher in OECD countries, their educationsystems still fail a large minority. For example,approximately 10% of students in England (UnitedKingdom) and the United States, and an even highershare in Italy, score below the low threshold.

Learning achievement does rise with averageincome, but with large variations and some strikingexceptions. Households and governments inwealthier countries can invest more in educationand this often leads to higher achievement. Forexample, average government spending on asecondary school student in Norway was US$13,388in 2005, compared with US$348 in Ghana (seeannex, Statistical Table 11).34 But the links betweenincome and learning are far from automatic. Amonghigh-income countries, the best performers – theRepublic of Korea and Singapore – outperformwealthier countries such as the United States. Themost striking exceptions are among Arab states. InQatar and Saudi Arabia, both high-income countries,three-quarters of students register below the lowestscore threshold – a performance comparable withthat of Ghana. In the middle-income countries ofAlgeria, Egypt and Morocco, more than half ofstudents register below the lowest threshold.

These results point to serious underlying policyproblems and help explain the widely observedfailure of Arab states to translate investment ineducation into improved skills, employmentcreation and economic growth.

International comparison highlights the degreeof inequality in learning achievement worldwide,with students from low-income countries faringespecially poorly, as Figures 2.37 and 2.38powerfully demonstrate. At age 10 or 11, in thefourth grade of primary school, fewer than onein five children in Japan or the Netherlandsscored below the intermediate benchmark on therelevant TIMSS scale. In Japan, almost all studentshad at least intermediate levels of proficiency,while in Qatar and Yemen almost no childrenscored above that level. Meanwhile, fewer than20% in El Salvador, Morocco and Tunisia scoredat the low benchmark (Figure 2.37).

Evidence from international assessments ofreading skills is even more disturbing. PISAassesses students with about eight years ofeducation. Students with reading literacy belowlevel 1 are identified as being at risk during thetransition to work. They are also unlikely to have

achieved sufficient proficiency to be able tobenefit from further education and other learningopportunities throughout life (OECD, 2007b).In Kyrgyzstan, 70% of students tested in PISAfailed to achieve level 1 proficiency in reading(Figure 2.38). In Brazil, Indonesia, Mexico andThailand, more than 40% of students were at level1or below. After eight years of schooling, thesechildren were unable to demonstrate levels ofliteracy that would typically be achieved by themiddle of primary school in OECD countries.

Sub-Saharan Africa is covered poorly byinternational learning assessments, but there isno shortage of evidence pointing to acute problems.Regional assessments conducted by the Southernand Eastern Africa Consortium for MonitoringEducational Quality (SACMEQ) from 2000 to 2002for Malawi, Namibia and Zambia found that over70% of grade 6 students in each country had notachieved basic numeracy. Students in Lesothoand South Africa did not do much better: over halfof all students failed to achieve basic levels ofnumeracy. The evidence from sub-Saharan Africaalso demonstrates that income is not the onlyfactor shaping learning achievement. Lesothoand South Africa have much higher per capitaincomes and government resources than Kenya,but they registered lower levels of primary schoollearning achievement.

Poor quality of education in childhood has a majorbearing on adult illiteracy. Young adults with noeducation or just a few years of school inevitablyfigure prominently in the ranks of adult illiterates.But so do some who have spent several years inschool. An analysis of adult literacy in twenty-onecountries in sub-Saharan Africa using householdsurvey data found that 22- to 24-year-olds with fiveyears of education had a 40% probability of beingilliterate (UNESCO-BREDA, 2007). People withseven years of education had a 20% chance ofbeing illiterate. These figures point not just to anenormous waste of human potential and restrictedopportunity, but to a failure of investment ineducation to deliver results.

In Latin America and the Caribbean, readingproficiency of primary school students also varieswidely (Figure 2.39). According to the recentSegundo Estudio Regional Comparativo y Explicativo(SERCE) assessment, less than half of all grade 3students in the Dominican Republic, Ecuador andGuatemala had more than very basic reading skills(UNESCO-OREALC, 2008).35 In contrast, over 85%

34. Expressed atpurchasing power parity in constant 2006US dollars.

35. This is based on grade 3 studentsachieving at or below level 1 performance, as defined by theassessment.

Poor quality education in

childhood has amajor bearing on

adult illiteracy

P R O G R E S S T O WA R D S T H E E FA G O A L S

T h e q u a l i t y o f e d u c a t i o n

1 0 7

of students in Costa Rica and Cuba had movedbeyond the basics. Primary school students inCuba performed extremely well. Over 40% of Cubangrade 3 students achieved the highest reading skillsassessed, more than double the share in Chile andMexico. At the other end of the achievement scale,Cuba has the smallest share of studentsperforming below the lowest grade.

Countries in South and West Asia also suffer fromsignificant deficits in levels of learning. In ruralPakistan, a recent survey found that only two-thirdsof students in grade 3 could subtract single digitnumbers, and only a small proportion could tellthe time or carry out simple multiplication anddivision (Das et al., 2006). In rural India, levelsof learning are equally troubling. In 2008, just 28%of grade 3 students could subtract two-digitnumbers and only a third could tell the time(Pratham Resource Centre, 2008).

In one important respect international learningassessments understate the problem. This isbecause they cover only children in school, ratherthan the entire age group. Factoring in out-of-schoolchildren would significantly lower the scores of

countries that are far from universal primaryenrolment at the level being tested. In Ghana, TIMSS2007 found that 17% of 16-year-olds scored abovethe low international benchmark, but only aroundhalf of the children of this age were in school.This implies that only 9% of Ghanaian 16-year-oldshave mastered the most basic maths skills. In high-income countries such as the Czech Republic andEngland (United Kingdom), most children at thesame stage are still in school and have far higher-level skills in mathematics (Mullis et al., 2008).36

Disparities in learning within countriesAn equal opportunity to learn is no less a humanright than an equal entitlement to be in school,regardless of parental income, gender, language orethnicity. The Dakar goals in some cases explicitlytarget greater equity. For example, achieving genderparity in learning achievement is an important partof goal 6 and a key component of achieving genderequality in education by 2015 (goal 5) (Box 2.22).In many countries, however, large disparities inlearning achievement point to deep disparitiesin opportunity. What students achieve is heavilyinfluenced by both the type of school they attendand the characteristics of their family backgrounds.

36. For the TIMSSassessment children areassessed in grade 8 of theireducation careers. In theCzech Republic and England(United Kingdom) theaverage age of children ingrade 8 was 14 comparedwith 16 in Ghana.

Internationallearningassessmentsunderstate theproblem as theyonly cover childrenin school

Japa

nNe

ther

land

sEn

glan

d (U

. K.)

Germ

any

Unite

d St

ates

Denm

ark

Aust

ralia

Aust

riaSw

eden

Hung

ary

Italy

Scot

land

(U. K

.)Ne

w Z

eala

ndCz

ech

Repu

blic

Norw

ay

Hong

Kon

g, C

hina

Sing

apor

eSl

oven

iaKu

wai

tQa

tar

Latv

iaKa

zakh

stan

Russ

ian

Fed.

Lithu

ania

Slov

akia

Arm

enia

Ukra

ine

Geor

gia

Iran,

Isl.

Rep.

Alge

riaCo

lom

bia

Mor

occo

Tuni

siaEl

Sal

vado

rYe

men

0

20

40

60

80

100

100

80

60

40

20

High

est b

ench

mar

k ach

ieve

d (%

of s

tude

nts)

OECD Other high income Middle and low-income

Advanced benchmark

High benchmark

Intermediate benchmark

Low benchmark

Below low benchmark

Figure 2.37: There are wide disparities across countries in primary school mathematics performance% of fourth grade students reaching the TIMSS international benchmarks for mathematics achievement, 2007

Notes: Low benchmark – students have some basic mathematical knowledge, an understanding of adding and subtracting with whole numbers, and can read information from simple bar graphsand tables. High benchmark – students can apply their mathematical understanding and knowledge in relatively complex situations and explain their meaning.Source: Mullis et al. (2008).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 0 8

The case for reducing disparities in learningachievement goes beyond education. Moreequitable education, combined with sustainedimprovement in overall quality, is likely to be goodfor economic growth and social cohesion. Achievinggreater equity will require a stronger focus onschools that serve the disadvantaged – and onthe factors beyond education that diminish learningachievement. Many countries have shown thatit is possible to combine equity with high levelsof overall achievement.

Measuring equity in learning achievement isinherently difficult. One approach is simply tomeasure the gap between the best- and worst-performing students. Figure 2.40 applies thismeasure to grade 4 mathematics results using data from TIMSS. The sizes of the bars showdifferences in test scores between the bestperformers (ninety-fifth percentile) and the worst(fifth percentile), expressed as a percentage ofthe average country score. In OECD countries,disparities in learning are typically smaller thanin other countries covered in TIMSS. In Germany,the difference between the best and worstperformers is about 42% of the mean score of 525.

Figure 2.38: Reading ability in secondary school also varies greatly across countries% of 15-year-old students reaching the PISA standard levels of reading proficiency, 2006

Rep.

of K

orea

Finla

ndCa

nada

Irela

ndNe

w Z

eala

ndAu

stra

liaNe

ther

land

sSw

eden

Belg

ium

Switz

erla

ndGe

rman

yJa

pan

Denm

ark

Unite

d Ki

ngdo

mFr

ance

Aust

riaIce

land

Norw

ayHu

ngar

yCz

ech

Rep.

Luxe

mbo

urg

Portu

gal

Italy

Slov

akia

Gree

ceSp

ain

Hong

Kon

g, C

hina

Liech

tens

tein

Esto

nia

Slov

enia

Mac

ao, C

hina

Israe

lQa

tar

0

20

40

60

80

100

100

80

60

40

20

Level 5

Level 4

Level 3

Level 2

Level 1

Below level 1

High

est b

ench

mar

k ach

ieve

d (%

of s

tude

nts)

OECD Other high income

Notes: Below level 1 – students have serious difficulty using reading literacy as an effective tool to advance and extend their knowledge and skills in other areas. Level 5 – students can complete sophisticated reading tasks, such as evaluating critically, building hypotheses, and locating and accommodating concepts contrary to expectations.Source: OECD (2007b).

Cuba

Chile

Costa

Rica

Mex

icoUr

ugua

yAr

gent

inaCo

lombia

Braz

ilEl

Salva

dor

Nica

ragu

aPe

ruPa

nama

Para

guay

Ecua

dor

Guat

emala

Domi

nican

Rep

.

0

20

40

60

80

100

20

40

60

80

100

Level 4

Level 3

Level 2

Level 1

Below level 1

Perce

ntag

e of s

tude

nts r

each

ing pe

rform

ance

leve

l

Figure 2.39: Latin America’s reading league has large performance gapsLevels of reading performance of third grade primary school students in Latin America, 2007

Notes: Level 1 – Students can locate information with a single meaning which is repeated in a prominent part of a text. Level 4 – Students can integrate and generalize information given in a paragraph, read a textand identify new information.Source: UNESCO-OREALC (2008).

P R O G R E S S T O WA R D S T H E E FA G O A L S

T h e q u a l i t y o f e d u c a t i o n

1 0 9

Disparities in learning tend to be much widerin low-income countries, with some of thewidest disparities found in Arab states. In Yemen,the difference between the best and worstperformers is 368 scale points, or 165% of themean score of 224. A study of mathematicsachievement in the Indian states of Orissa andRajasthan found that the range in test scoresbetween the best and worst performers was widerthan for all TIMSS countries except South Africa(Das and Zajonc, 2008).37

Differences between schools play a critical rolein the level of equity within education systems,as evidence from the OECD countries shows.Measured on a global scale, these countries haverelatively low overall levels of inequality in learningachievement. Where they differ is in the shareof inequality that can be traced back to schools.In Nordic countries such as Finland, Iceland andNorway, less than 10% of the variation in sciencescores is explained by school differences. At theother end of the scale, such differences accountfor over half the variation in test scores in Germany(Figure 2.41).38 Such findings demonstrate thedegree to which school-based factors can widen –or narrow – learning achievement gaps.

0

20

40

60

80

100

100

80

60

40

20

High

est b

ench

mar

k ach

ieve

d (%

of s

tude

nts)

Middle and low-income

Pola

ndLa

tvia

Croa

tiaLit

huan

iaTu

rkey

Chile

Russ

ian

Fed.

Urug

uay

Bulg

aria

Mex

icoTh

aila

ndAr

gent

ina

Serb

iaJo

rdan

Braz

ilCo

lom

bia

Rom

ania

Mon

tene

gro

Tuni

siaIn

done

siaAz

erba

ijan

Kyrg

yzst

an

In many countries, girls are less likely than boys to get into school.Once in school, though, they tend to perform as well as, or better than,their male classmates. While there are important gender-baseddifferences in learning achievement by subject, learning achievementin general is not characterized by deep inequalities.

OECD countries. In PISA 2006, average reading scores for 15-year-old girls were 8% higher than those of boys throughoutthe OECD. In mathematics, boys held an advantage over girls.The widest gap was found in Austria, where males’ test scores wereon average 5% higher. Gender differences in science tended to bestatistically insignificant.

Arab States. TIMSS 2007 covered thirteen of the twenty Arab States.In most of them, grade 8 girls outperformed boys in mathematics.In Qatar, girls’ test scores were 13% higher than boys’. A similarnumber of countries recorded a female advantage in science, withlarger gender gaps in many cases. In Qatar, girls scored on average25% higher than boys.

Central and Eastern Europe. PISA 2006 covered fifteen of the twenty-one countries in this region. All registered a large femaleadvantage in reading performance. In most, gender gaps inmathematics were statistically insignificant; in the remainingcountries boys tended to do slightly better than girls.

Latin America and the Caribbean. Information from sixteen countriesin the 2006 SERCE assessment in mathematics shows that boys inthe sixth grade performed better than girls. When reading wasassessed, girls outperformed boys, but in both subjects the averagedifferences were small.

Sub-Saharan Africa. Among Francophone countries participatingin the PASEC assessment, there were no large gender differencesin second and fifth grade performance in French or mathematics.For the thirteen countries participating in the 2000–2002 SACMEQassessments, gender differences in sixth grade English were on thewhole either statistically insignificant or small.1 In mathematics, abouthalf the participating countries showed no statistically significantgender difference. In the rest, males’ average scores tended to behigher but the differences were not large.2

These findings confirm that gender gaps in overall achievement aremodest. Where differences do exist, the data show that, except in theArab States, girls do better in languages and boys in mathematics andscience. Eliminating remaining gaps will be necessary if the goal ofeducation for all is to be achieved. However, it has to be recognizedthat current data provide an incomplete picture, especially for countriesthat do not participate in international and regional assessments.

1. Seychelles was the exception, where girls’ performance in English compared with that of boys was 0.65 of a standard deviation higher.

2. Seychelles was again the exception, where girls’ performance in mathematicscompared with that of boys was 0.38 of a standard deviation higher.

Sources: Bonnet (2009); Ma (2007); Mullis et al. (2008); OECD (2007b).

Box 2.22: Gender parity and learning achievement

37. The study compares test scores of students at the fifth and ninety-fifth percentile of the test score distribution.

38. The overall dispersion of test scores in Germany is 110% of the OECD average, compared with 81% in Finland (OECD, 2007b).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 1 0

Does greater equity come at the price of reducedaverage performance? Not in this case: Finland isboth more equitable and higher performing thanGermany. The greater equity is partly the result ofcomprehensive education systems that providesimilar opportunities for all. In recent years, Polandhas achieved substantial reductions in inequalityamong schools by extending the duration ofcomprehensive education (Box 2.23). One factorcontributing to school-based inequality in manyOECD countries is the grouping of students intorigidly separate ability streams, or into academicand vocational tracks or schools (OECD, 2007b;Schutz et al., 2008). In Germany, vocational tracks

at the secondary level have helped reduce youthunemployment, but early tracking has contributedto high levels of inequality across schools.

In many developing countries, differences inperformance across schools are linked to theteaching environment. School systems are oftenmarked by large variations in class size, availabilityof books and teaching materials, teacher qualityand school building standards.39 In the 2000–2002SACMEQ assessment, differences among schoolsaccounted for 37% of the variation in studentreading performance (Dolata et al., 2004).Research in India and Pakistan has also found thatdifferences in school characteristics help explaininequality in test scores (Das et al., 2006; Das andZajonc, 2008). In Bolivia and Chile, a study showedthat over half the large disparities in learning

39. Performancedifferences amongschools can also arisebecause households withsimilar social andeconomic backgroundsoften make the sameschool choices (e.g. publicor private schooling), orbecause governmentpolicy on how studentsare selected into specificschools has this effect.

Hong Kong, ChinaNetherlands

GermanyLatvia

SwedenJapan

DenmarkAustria

SingaporeLithuania

United StatesSlovenia

Czech Rep.Kazakhstan

ItalyRussian Fed.

Scotland (U. K.)England (U. K.)

NorwayAustraliaSlovakia

New ZealandHungaryArmeniaUkraineGeorgia

Iran, Isl. Rep.Algeria

ColombiaEl Salvador

MoroccoQatar

KuwaitTunisiaYemen

0 20 40 60 80 100 120 140 160 180

Disparity in scale score expressed as a percentageof the mean score

Figure 2.40: Learning gaps are higher in poor countriesLearning gaps in TIMSS mathematics scale scores for fourth gradestudents, 2007

Note: The bars show the difference in test scores between students at the fifth andninety-fifth percentiles, expressed as a percentage of the country’s mean test score.Source: Mullis et al. (2008).

FinlandIceland

NorwaySweden

PolandSpain

DenmarkNew Zealand

IrelandAustralia

CanadaUnited Kingdom

United StatesLuxembourg

PortugalSwitzerland

MexicoRep. of Korea

SlovakiaJapan

GreeceItaly

BelgiumTurkey

AustriaCzech Republic

NetherlandsGermanyHungary

0 10 20 30 40 50 60 70 80

% of total variance explained by differences across schools

Figure 2.41: When schools make a difference — inequality in studentperformance across schools varies widely in rich countriesVariation in PISA science scale scores in OECD countries explained by differences across schools, 2006

Note: The length of the bars shows the percentage of the total variation in studentperformance that is accounted for by performance differences between schools.Source: OECD (2007b).

P R O G R E S S T O WA R D S T H E E FA G O A L S

T h e q u a l i t y o f e d u c a t i o n

1 1 1

between indigenous and non-indigenous studentswere explained by the poor quality of schoolsserving indigenous students (McEwan, 2004).

These findings demonstrate that improving schoolquality and narrowing differences among schoolswill reduce inequality in student performance. In themid-1990s, Brazil created the Fundo de Manutençãoe Desenvolvimento do Ensino Fundamental e deValorização do Magistério (FUNDEF), a fund tofinance subnational spending on primary and lowersecondary education to ensure a more equitabledistribution of per-student spending across thecountry. Preliminary evidence suggests that thisredistributive policy has narrowed learninginequalities, though only by a small amount (Gordonand Vegas, 2005). In other countries, per-studentfunding formulas have been introduced to ensurethat resources are more equitably distributedacross regions and population groups.

School-based disparities do not operate in isolation.In many cases they interact with and reinforcewider disadvantage. Parental income and education,home language and other factors are all stronglyassociated with learning achievement levels, asthe following cases demonstrate:

In Pakistan, children from families in the richestthird of the population scored on averagebetween 0.25 and 0.5 of a standard deviationhigher than children from the poorest households(Das et al., 2006).

In Peru, in national assessments of mathematicsconducted in 2004, sixth grade pupils whosemother tongue was Spanish scored more thanone standard deviation higher than childrenwhose mother tongue was an indigenouslanguage (Cueto et al., 2009).

Fifth grade students from Cameroon’s Bamilekelanguage group scored 48% on the PASECliteracy test, compared with 56% for studentsfrom the Ewondo language group (Fehrler andMichaelowa, 2009).

In the 2006 SERCE assessments in Latin Americaand the Caribbean, students who undertook asignificant amount of work, inside and outsidethe home, had lower levels of mathematicsachievement on average. For example, inEl Salvador, sixth graders who worked hadaverage scores 6% lower than those of childrenwho did not (Bonnet, 2009).

A longitudinal study in Ethiopia found that 42%of 12-year-olds who had lost their mothersbetween ages 8 and 12 were unable to read,while for children with both parents living thefigure was 23% (Himaz, 2009). The studyattributed the difference to lower schoolenrolment among orphans, as well as poorerperformance in school.

In countries with more equitable systems,children’s backgrounds are less important indetermining levels of achievement. In countrieswhere there is a strong relationship betweenstudent background and performance, or wherelarge differences in student background exist,reducing differences in school quality is unlikelyto be enough to improve equity significantly.Targeted programmes to improve learning amongchildren who are being left behind will also beneeded (see Chapter 3). In Mexico, the ConsejoNacional de Fomento Educativo (CONAFE) providessupplemental funds, learning materials – includingtextbooks in indigenous languages – and teachersupport to schools in areas marked by consistentunderperformance and disadvantage. Evaluationsindicate that these efforts have narrowed the gapin primary school mathematics scores, thoughthey have had little impact on Spanish scores(Vegas and Petrow, 2008).

School-baseddisparities interactwith studentbackground andsocio-economicstatus

In 1999, Poland started providing an additional year of generaleducation before students were split into upper secondary schooltracks. By using three rounds of PISA it is possible to assess thereform’s impact on equity:

From 2000 to 2003, average variation in student performance in science fell from 51% of the OECD average to 15%. By 2006,Poland had one of the lowest levels of variation in scienceperformance among participating countries (Figure 2.41).Improvement in equity came about at the same time as generalimprovement in performance. For example, average readingperformance of 15-year-olds increased by twenty-nine score points between 2000 and 2006.

Most of the improvement occurred among students with poorperformance. From 2000 to 2006, the proportion of studentsfailing to score above level 1 in reading competency fell from 23% to 16%.

Students in the vocational track appear to have benefited most from greater integration of the system.

Source: OECD (2007b).

Box 2.23: Improving equity in Poland

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 1 2

Teaching reading in the earlyprimary grades is crucial for learning

Children lacking reading and comprehensionskills are unable to use textbooks and otherwritten materials and to take advantage of learningopportunities in or out of school. Research showsthat children who have difficulties acquiring theseskills in their early lives are likely to strugglethroughout their school careers (Jukes et al.,2006). Schools in many developing countries arefailing to equip young learners with these basicskills. If the EFA quality goal is to be achieved,it is crucial for reading difficulties to be detectedearly and acted upon quickly.

Understanding is central to the achievement ofreading skills. Children need to be able to readwith sufficient fluency to obtain meaning fromwhat is being read. Reading accurately and quicklyhas been shown in many cases to be stronglycorrelated with comprehension (Abadzi, 2006;Fuchs et al., 2001). In a small-scale survey ofPeruvian first and second graders, children whocould answer a set of three comprehensionquestions correctly read on average at seventy-seven words per minute, compared with fifteenwords per minute for children who only answeredone comprehension question correctly (Abadziet al., 2005).40 Early grade reading assessmentsaim to test comprehension by measuring readingfluency (Box 2.24). Estimates vary, but readingfluency in excess of forty words per minuteis thought to be required for comprehension(Abadzi, 2006).41

Small-scale reading assessments conducted inseveral low-income countries paint a worryingpicture (Table 2.7).42 While these tests are notnationally representative, they often point tovery low levels of fluency in reading. Averageperformance in many test sites falls far short ofautomatic reading, implying that large numbersof children are failing to achieve the basic readingskills necessary to facilitate further learning:

In Gambia, children in grades 1 to 3 were able,on average, to read six words correctly in aminute.

In Liberia, grade 2 students could read eighteenwords per minute. Although fluency increasedin grade 3, it was still below the estimated fortywords per minute required for comprehension.

Looking beyond the averages reveals the pooroutcomes for some children:

In Ethiopia, a 2008 study of grade 3 studentsin Woliso district found that 36% could not reada single word in Afan Oromo, the local language(DeStefano and Elaheebocus, 2009).

In Guatemala, two-thirds of students could readmore than forty words correctly per minute butwide disparities across language groups existed.Students whose mother tongue was Mam hadaverage reading speeds below forty words perminute whereas students whose mother tonguewas Quiche or Spanish read more than sixtywords per minute (Dowd, 2009).

Assessing reading skills early in primary schoolprovides an opportunity to identify children with lowlearning achievement and take remedial measuresthat can help prevent dropout and grade repetition.It is far less time-consuming and costly to preventlow achievement at an early age than to act later.

Evidence from several countries demonstratesthat policy interventions can make a differencein improving reading skills. Involving schools andcommunities is a key to success. A programmeoperated by a non-government organization inUttar Pradesh, India, has used ‘remedial readingcamps’ run by volunteer trainers to achieveimpressive improvements in early reading(Box 2.25). In the Malindi district of Kenya, teacherswere trained for five days on a set of carefullydesigned lessons to teach effective reading skillsto grade 2 students (Crouch et al., 2009). Significantimprovements resulted: comparing grade 2 resultsbefore and after teachers were trained showed thatreading speeds had improved by 80%, on average.While it is difficult to attribute all the improvementto the training, the study showed that this relativelysmall intervention and the information it generatedon the poor state of reading skills contributedsignificantly.43 Pilot studies in Mali and the Nigerin 2007 also demonstrated promising approachesto improving reading skills at relatively low cost(Mitton, 2008).

Whether such pilot programmes can be scaled upto improve reading across national education

40. The positivecorrelation betweencomprehension and oralreading fluency has beenthe subject of muchresearch. For additionalexamples see Kudo andBazan (2009) and RTIInternational (2008).

41. In the United States,children are identified asbeing at risk of developinglearning difficulties if theirreading speed is belowseventy words per minutein grade 2 and beloweighty in grade 3.

42. It is not possible tocompare across countriesin the table owing to thedifferent languages usedand differences in theages of studentpopulations.

43. The study was set up as a randomized trial but improvement inreading skills was seen in schools where the teachers had received thetraining as well as in the schools where they did not. The study showedthat leakage of teaching techniques and the transfer of teachers betweencontrol and treatment schools accounted for similar improvements inboth types of schools.

It is crucial for reading

difficulties to bedetected early

and acted upon quickly

P R O G R E S S T O WA R D S T H E E FA G O A L S

T h e q u a l i t y o f e d u c a t i o n

1 1 3

systems is open to question. Many of theseprogrammes succeeded because of intensiveefforts to mobilize communities, resources andlocal professionals. Duplicating these efforts isoften difficult for subnational authorities, let alonenational ministries. Even so, the success of pilotprogrammes shows that substantial improvementsare possible. Central governments and educationauthorities can make a difference to early learningthrough action at many levels, including trainingteachers more effectively and providing appropriateteaching materials.

In India, ‘remedialreading camps’have achievedimpressiveimprovements in early reading

2007 1-3 6 1,2002007 2 10 40 schools2008 2 18 4292008 3 28 407

2006 3 18 2,2062008 3 57 475

2007 1-3 19 502

GambiaMalindi, KenyaLiberia Liberia

NicaraguaJunin, Peru

Senegal

Table 2.7: Results from early grade reading assessments(correct words per minute)

Notes: Unless otherwise stated, sample size refers to the number of studentstested. The studies recorded in the table also measured other aspects of readingoutlined in Box 2.24 and in some cases tested reading in local languages as well.Sources: Castro and Laguna (2008); Crouch and Korda (2008); Crouch et al (2009);Jammeh (2008); Kudo and Bazan (2009); Sambe and Sprenger-Charolles (2008).

Year GradeConnectedtext fluency Sample size

Tests of reading in English

Tests of reading in Spanish

Tests of reading in French

Failure to develop reading skills in the early schoolyears can severely compromise later learning,undermining progression through grades andcontributing to early dropout. Early grade readingassessments help teachers identify problems andcorrect them. The components assessed include:

phonemic awareness — children can focus on,manipulate and break apart the sounds in words;

ability to use phonics — they can understand and apply knowledge of how letters are linked to sounds to form letter-sound correspondencesand spelling patterns;

fluency — they can read orally with speed, accuracy and proper expression;

vocabulary — they know an increasing number of words, both orally and in print;

comprehension — they can actively engage with and derive meaning from texts.

Most of these components are tested by counting the letters and words that children can sound outaccurately in one minute. While there are obviousdangers with mechanistic application, theassessments can be used to capture both wordrecognition and understanding. Similar approachesare used in adult literacy work. Early grade readingassessments are not designed to serve as tests forgrading students or ranking schools. They are mostuseful when integrated into a wider framework forbuilding children’s confidence and equipping schoolsand teachers to respond to their needs.

Sources: Abadzi (2006); Kudo and Bazan (2009); RTI International (2008).

Box 2.24: What are early grade reading assessments and what can they be used for?

A baseline survey conducted in 2005 in Jaunpurdistrict in Uttar Pradesh revealed poor acquisitionof basic skills. Among 7- to 14-year-olds, 60% couldnot read and understand a simple story designedfor first grade students. In this context, arandomized evaluation examined the impact on basic education skills of combinations of three interventions:

encouraging community participation byproviding information and facilitating discussionon the status of local schools and outlining tovillage education committees their roles andresponsibilities;

training community members to assesschildren’s learning and presenting these findingsat village meetings;

training local volunteers in simple techniques for teaching children to read and introducingreading classes after school.

The evaluation found that information-sharing didnot improve reading skills but that extra classesafter school had a big impact. Overall, theevaluation showed that it was possible to getchildren who were not fluent readers to read fairlyfluently by combining instruction in school withadditional reading classes which students attendfor two hours a day over a period of three months.

Source: Banerjee et al. (2008).

Box 2.25: India — remedial reading classes in Uttar Pradesh

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 1 4

Improving learning in schools

One of the most important requirements forsustained progress towards better quality ineducation is an improved learning environment,encompassing the physical school infrastructure,the learning process and the interaction betweenchildren and teachers.

Low achievement levels are often associatedwith a poor school environment. Badly ventilatedclassrooms, leaking roofs, poor sanitation andlack of materials represent significant barriers toeffective learning in many schools. Over half of ruralprimary students in Peru, the Philippines and SriLanka attend schools viewed by the head teacher asneeding major rehabilitation (UIS, 2008b).44 A recentsurvey of primary schools in two Nigerian statesfound that over 80% of classrooms in Enugu and50% in Kaduna either did not have a blackboard, orhad one that was barely usable (World Bank, 2008c).Such conditions are common in many countries.

The fact that the most marginalized childrenoften attend the poorest-quality schools adds totheir learning disadvantages. Urban-rural dividesfigure prominently in school quality disparities. Inthe Philippines, over 70% of urban grade 4 studentsattended schools with basic facilities such asblackboards and toilets, but only about 50% ofrural students attended schools with thesefacilities (UIS, 2008b). Improving learning in suchenvironments requires redistributing resourcestowards poorer areas.

Many studies highlight the positive role ofappropriate textbooks (Boissiere, 2004; Scheerens,2004). A detailed evaluation of Ghana’s basiceducation system found that improvements inmathematics and English test scores from 1988to 2003 had been brought about in part throughincreased availability of textbooks (White, 2004).

The longer children spend in school over the courseof a year, the greater their opportunity to masterthe curriculum and achieve learning objectives(Boissiere, 2004). The official number of teachinghours varies considerably by country (Benavot,2004), but time spent on effective learning is whatmatters for achievement. In effective classrooms,about 80% of class time is spent on learning – a benchmark that many schools in developingcountries fail to meet (Abadzi, 2006). Student andteacher absenteeism further reduces learning time.In Nepal, a detailed study of a small number of

primary schools showed that, while schools wereofficially open for 192 days, the average studentexperienced only 97 days of effective learning(Dowd, 2009). In Ethiopia and Guatemala, childrenwere in class and learning for a third of the timeschools were officially open (DeStefano andElaheebocus, 2009; Dowd, 2009). Better monitoring,improved teacher incentives and targeted supportfor students struggling to attend regularly can allincrease learning time and performance.

Increasing the amount of time children spendlearning can be difficult. Chronic overcrowdingof classrooms has led many countries to operatedouble-shift systems in schools. These offerpotential efficiency gains in terms of the number ofchildren covered, but the gains sometimes come ata price. In francophone Africa, double-shift teachinghas sometimes reduced learning achievement,primarily because children spend less time inschool (Michaelowa, 2001). In the longer term,additional classrooms can be built to accommodatea single-shift system. However, policy-makersneed to assess whether building classrooms isas cost-effective in improving learning as otherinvestments, such as providing more teaching andlearning materials. Creating conditions that enablechildren to remain in school, ensuring that teachersactually attend and organizing the school day todevote more time to learning are all low-costoptions with potentially high returns.

The important role of teachersTeachers are the single most important educationresource in any country. From early childhoodthrough primary and secondary school, thepresence of a qualified, well-motivated teacheris vital for effective learning. What students achievein school is heavily influenced by classroompractices and teachers’ skills (Aslam and Kingdon,2007). In many countries, shortages of trainedteachers remain a major barrier to achieving theEducation for All goals, especially amongmarginalized groups.

Pre-primary education. Early childhood teachersand carers play a crucial role in preparing childrenfor school and supporting their social, emotionaland cognitive development. The quality of care andteaching depends critically on the pupil/teacherratio, teacher training and the creation of an activelearning environment (Schumacher and Hoffmann,2008; UNESCO, 2005). Many countries do not meetminimum standards of quality, however. In Bolivia,India, Liberia,45 Nepal, Uganda and the United

44. This is the percentageof grade 4 pupils inschools where the headstated that the ‘schoolneeds completerebuilding’ or ‘someclassrooms need majorrepairs’.

45. In Liberia, thepupil/teacher ratioincreased from 14 to 142between 2006 and 2008.This is due to asubstantial decrease involunteer and otheruntrained teacherspreviously recruited tomeet the teacher demandfollowing the civil crisis(UIS database).

In Ethiopia and Guatemala,

children were in class and

learning for onlya third of thetime schools

were officiallyopen

P R O G R E S S T O WA R D S T H E E FA G O A L S

T h e q u a l i t y o f e d u c a t i o n

1 1 5

Republic of Tanzania, for example, the pre-primarypupil/teacher ratio was 40:1 or higher in 2007(see annex, Statistical Table 10A).

Data from within countries highlight particulardisadvantages facing the marginalized in thisrespect. In Kenya, the national ratio of pupils totrained pre-primary teachers is 54:1. In the arid,largely pastoral district of Turkana, one of Kenya’spoorest, the ratio is 123:1 (Ruto et al., 2009). InIndonesia, the share of pre-primary teachers withat least a diploma ranges from 60% in Banten, arelatively prosperous area, to only 1% in Maluku,a region with high levels of poverty (IndonesiaMinistry of National Education, 2007).

Primary education. Higher enrolment since 1999has gone hand in hand with an increase in therecruitment of primary teachers. Many countriesin sub-Saharan Africa – including Burkina Faso,Burundi, the Niger and Senegal46 – have morethan doubled the teacher workforce in most cases,improving the pupil/teacher ratio (see annex,Statistical Table 10A). As countries seek toaccelerate progress towards universal primaryeducation, they will need to sustain a concerteddrive to recruit and train teachers.

Despite the progress of the past decade, teachershortages remain a serious concern. Countries settheir own targets for pupil/teacher ratios, makingcross-country comparisons difficult (Bennell, 2009).However, the most widely used international ceilingfor the pupil/teacher ratio is 40:1 (Takala, 2003;World Bank Independent Evaluation Group, 2006). In 2007, 26 countries out of 171 with data were above this ceiling, all but four of them in sub-SaharanAfrica (see annex, Statistical Table 10A). While datacoverage is patchy, there are also concerns overthe ratio of pupils to trained teachers. Countriesincluding Madagascar, Mozambique, Sierra Leoneand Togo have ratios of pupils to trained teacherin excess of 80:1 (Figure 2.42). In fifteen of the fortycountries with data, the share of trained teachersin the workforce has declined since 1999, in somecases dramatically (see annex, Statistical Table10A). In Togo, it has fallen from 31% to 15% asrecruitment has shifted towards contract teachers.

National average pupil/teacher ratios can conceallarge disparities. A recent review of teacherdeployment patterns examined differences acrossregions in ten countries in sub-Saharan Africa(Pôle de Dakar, 2009). In some countries the ratiosvary by a factor of three. While low ratios are often

found in rural areas with highly dispersedpopulations, high ratios tend to be concentrated inareas marked by poverty and acute disadvantage.In Uganda, northern regions affected by conflict weremarked by pupil/teacher ratios in excess of 90:1 –nearly double the national average (Figure 2.43).

Urban-rural differences create another layer ofinequality. The pattern of disadvantage is highlyvariable but overall ratios tend to be higher in urbanareas (Zhang et al., 2008). In other countries suchas Malawi, though, the average urban pupil/teacherratio is 46:1, compared with 81:1 in rural areas(Mulkeen, 2009). However, trained teachers areoften concentrated in urban areas. Whereas 60%of teachers in the Ugandan capital, Kampala, aretrained, the figure falls to 11% in the rural district ofYumbe. In Lesotho, nearly a quarter of teachers inlowland areas are unqualified, compared with abouthalf in the mountainous and less accessible areas(Mulkeen, 2009). These areas also tend to havehigher repetition rates and, like other rural areas,poorer test scores (Mulkeen and Chen, 2008).

To some degree, such deployment patterns reflectself-perpetuating processes of selection and inbuiltdisadvantage. Trained teachers are more likelyto choose to work in urban areas, especially insystems where their remuneration is linked to

46. In Senegal, this increasein teachers is due to thecreation of more schools,upgrading of schools withincomplete primaryeducation cycles and double-shift teaching (UIS, 2009b).

In northernUganda, therewere 90 pupils per teacher — nearly doublethe nationalaverage

MauritaniaCôte d’Ivoire

KenyaCambodia

RwandaU. R. Tanzania

UgandaBurundiEritreaCongo

Burkina FasoMozambique

BangladeshMadagascarSierra Leone

0 20 40 60 80 100 120

Ratios

Pupil/teacher Pupil/trained teacher

Figure 2.42: Trained teachers are sometimes in short supply Ratios of pupils to teachers and pupils to trained teachers in primary education, selected countries, 2007

Note: Among countries with available data, only those with pupil/teacher ratios ator above 40:1 are included. Countries sorted by the gap between the pupil/teacherand the pupil/trained teacher ratios.Sources: Annex, Statistical Table 10A; UIS database.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 1 6

parental contributions. Opportunities forprofessional development are also more likelyto be concentrated in urban areas, enablingurban teachers to gain qualifications more readilythan their rural counterparts (Bennell andAkyeampong, 2007). Cities may be seen aspreferable to rural areas for other reasons,ranging from the quality of housing, amenitiesand schools to the proximity of friends and family.Concerns over living in remote and unfamiliar ruralcommunities can also play a role. Such factorsplay a part in the preference of female teachersfor urban areas in many countries. In Ugandaand Zambia, the share of female teachers in urbanprimary schools is about 60%, compared with 15%to 35% in rural areas (Mulkeen, 2009).

Projected primary school teacherrequirements to 2015Future teacher recruitment needs vary enormously by region. They are determinedpartly by current deficits and partly by a complexmix of demographics, enrolment trends andnumbers of children still out of school. The UIS hasre-estimated the total number of primary educationteachers that will be required to achieve the goal ofuniversal primary education by 2015 (UIS, 2009e).47

The numbers underline the scale of the challengefacing many countries.

An additional 1.9 million teachers will have tobe recruited to reach universal primary educationby 2015.

Two-thirds of the additional teachers – around1.2 million – will be needed in sub-Saharan Africa.

The Arab States account for around 15%of the additional teachers required.

The effort needed to close these gaps varies bycountry (Figure 2.44). Many governments will haveto expand recruitment by 4% to 18% annually.For some countries, this means maintaining therate of increase registered since 1999. Others willneed to step up the pace of recruitment and budgetfor new posts, including Chad, Côte d’Ivoire,Djibouti, Eritrea, Kenya and Uganda.

In addition to increasing recruitment to achieveuniversal primary education, governments haveto replace teachers expected to retire or leavetheir posts before 2015.48 Taking into accountthe need to replace teachers drives up the regionaland global recruitment numbers (UIS, 2009e):

An additional 8.4 million primary teachers willhave to be recruited and trained worldwide toreplace existing teachers expected to retire orleave their posts before 2015.

Nearly a quarter of these teachers – around2.1 million – will be needed in East Asia andthe Pacific.

North America and Western Europe accountfor 17%, South and West Asia for 19% and sub-Saharan Africa for another 15% of theadditional recruitment needed to replaceteachers leaving their posts by 2015.

A total of 10.3 million additional teachers will beneeded worldwide by 2015, if the 1.9 million newteachers required to achieve universal primaryeducation are added to the 8.4 million needed toreplace departing teachers. The number of extrateaching posts that need to be created may seemsmall compared to the teacher needs resultingfrom attrition. However, creating new posts requiresan increase in the overall budget allocated forteacher salaries. In many countries this requiresgreater effort than that of filling vacant posts.

47. Estimates of teacherneeds are based onassumptions regardingenrolment, repetition and pupil/teacher ratios.For technical details see UIS (2009e).

48. These projections are based on a teacherattrition rate of 5% and include additionalteachers needed to fillvacancies resulting from increased attritioncaused by the expansionof universal primaryeducation and populationgrowth. For results using other attritionassumptions see UIS (2009e).

An additional 1.9 million

teachers willhave to be

recruited to reach UPE

by 2015

0 20 40 60 80 100 120 140

U. R. Tanzania

Eritrea

Uganda

Zambia

Benin

Malawi

C. A. R.

Pupil/teacher ratios

Lowest Average Highest

Figure 2.43: National averages can hide large differences in pupil/teacher ratiosProvincial disparities in primary education pupil/teacher ratios,selected sub-Saharan African countries, circa 2005/2006

Note: Except where indicated, data do not distinguish between civil servantteachers and community teachers, or between public and private sector teachers.Benin data exclude community teachers. C. A. R. data include community teachers.Countries sorted by average pupil/teacher ratio.Source: Pôle de Dakar (2009), Table 5.4.

P R O G R E S S T O WA R D S T H E E FA G O A L S

T h e q u a l i t y o f e d u c a t i o n

1 1 7

The capacity of countries to finance increasedrecruitment varies, but for many the prospect ofclosing the gap will hinge partly on aid donors.A sustained push to recruit teachers will increasethe future recurrent costs that governments haveto plan for. For low-income countries with a limitedrevenue base, multiyear aid commitments overfive to ten years, backed by predictable delivery,will be vital to the sustainable financing of teacherrecruitment. This will require both an increasein aid and a radical change in aid managementpractices – issues discussed in Chapter 4.

Recruitment is just one part of a far wider setof issues that governments have to address.In many of the world’s poorest countries, theproblem involves not just low teacher numbersbut also poor teacher morale. Attracting andretaining well-qualified candidates is increasinglydifficult, as many countries’ high attrition ratesshow. Low pay is endemic, a problem that inmany countries not only hinders recruitmentof able candidates but forces many teachers tosupplement their salaries, thus reducing the timethey devote to teaching. In Bangladesh, teachers instate-aided schools are paid less than US$1 a day,and two-thirds report undertaking additionalincome-earning activities (Financial ManagementReform Programme, 2006).

Teacher salaries are a contentious issue in manycountries. Governments face an obvious dilemma:how to increase teacher recruitment withoutcreating unsustainable budget pressures. Somecountries have attempted to address this dilemmaby reducing salary costs, notably by hiring fewerteachers under standard civil service pay termsand hiring more contract teachers at lower levelsof remuneration and benefits. Finance ministriesand several aid donors have actively encouragedmore contract employment. The risk is that thiswill lead to recruitment of less qualified candidatesand to even more pressure on teachers tosupplement their incomes through other formsof employment, with attendant implications formorale. There is evidence from West Africa thatincreased recourse to contract teachers hascompromised education quality (UNESCO, 2008a).While a balance has to be struck betweenaffordability and good teaching, the limits to cost-cutting also have to be recognized.Governments and donors need to ensure thatteacher pay and conditions reflect a commitmentto delivering good-quality education througha well-qualified and motivated workforce.

In many of theworld’s poorestcountries, theproblem involveslow teachernumbers or poorteacher morale

Figure 2.44: The rate at which new teaching posts arecreated will need to increase if universal primary educationis to be achieved by 2015Annual growth in teacher posts needed to reach universal primaryeducation by 2015, selected countries

Benin

Senegal

Guinea

Burundi

Liberia

Congo

Mali

Togo

Viet Nam

Gambia

Palestinian A. T.

U. R. Tanzania

Kenya

Comoros

Rwanda

Côte d’Ivoire

Uganda

D. R. Congo

Djibouti

Burkina Faso

Niger

Chad

Eritrea

Afghanistan

Bangladesh

Timor-Leste

Sudan

Malawi

Papua N. Guinea

C. A. R.

0 5 10 15 20

Observed 1999-2007

Needed 2007-2015

Annual growth rates (compound) in teaching staff (%)

Observed growth higher than needed

Observed growth lower than needed

No observed data to compare

1

Note: Only countries in which the current number of teachers must expand by at least 30% to reach the projected number needed to achieve universal primaryeducation on time are included.1. Kenya’s observed growth rate between 1999 and 2007 was -0.1% but forexpositional purposes this is not shown here.Source: UIS (2009e).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 1 8

Initial training and professional developmentare also crucial to morale and effective teaching.Teachers are the product of the education systemsthey teach in. Where these systems are of lowquality it is even more important for teachers toreceive effective training and support throughouttheir careers. Teachers need to understand thecontent of the curriculum and be able tocommunicate it to students of varying ability. Inmany countries, initial training is not good enoughto develop these skills. To make matters worse,many teachers do not even receive initial training.In Mozambique, one recent evaluation found that41% of primary school teachers were untrained(Mulkeen and Chen, 2008). In-service training,which is vital to build on initial skills, is alsopoorly developed in many low-income countries(Leu, 2004; Lewin and Stuart, 2003).

Conclusion

The ultimate aim of schools is to equip childrenwith the skills and knowledge they need to realizetheir potential, develop secure livelihoods andparticipate in society. Evidence presented in thissection suggests that many schools are failing tomeet even minimum standards for the quality ofeducation. Millions of children, especially thosefrom socially marginalized groups, are completingtheir primary education without having acquiredbasic literacy and numeracy skills. At the secondarylevel, too, many education systems in developingcountries are characterized by low levels oflearning and high levels of inequality. Equippingschools to provide good quality education willrequire governments to focus more strongly onrecruiting and training teachers, supplyingtextbooks and developing classroom practicesthat promote active learning. Support for literacyand reading in early grades has an especiallyimportant role to play, as these skills createthe foundation for future learning.

P R O G R E S S T O WA R D S T H E E FA G O A L S

E s t i m a t i n g t h e c o s t o f a c h i e v i n g E d u c a t i o n f o r A l l

1 1 9

Estimating the costof achieving Educationfor All

The Dakar Framework for Action includes strongcommitments on education financing. Developingcountries pledged to ‘enhance significantlyinvestment in basic education’ (para. 8[i]), while richnations promised an increase in aid, mainly in theform of grants and concessional finance, to ensurethat ‘no countries seriously committed to educationfor all will be thwarted in their achievement of thisgoal by a lack of resources’ (para. 10).

Ten years later, finance remains a major barrierto Education for All. Developing countries havestepped up their efforts to make domesticresources available, albeit on an uneven basis.Many could do much more to raise theirinvestment in education by strengthening revenuecollection and equity in public spending. Theinternational community has also increased aidfor basic education, but collectively donors havefallen far short of delivering on their commitmentsmade in Dakar.

This section assesses the scale of the Educationfor All financing gap. It sets out the results of acosting exercise covering forty-six low-incomecountries.49 The exercise looks at the financingrequirements for achieving a range of goals in basiceducation and beyond. Among the key findings:

The financing gap is far larger than previouslyassumed. Even with an increased domesticresource mobilization effort, low-incomecountries face a financing gap of aboutUS$16 billion for basic education (literacy, pre-primary and primary education),representing 1.5% of their collective GDP.50

This is a third higher than the previous estimate.Factoring in lower secondary education wouldincrease the gap to US$25 billion.

Low-income countries need to strengthenefficiency and equity in education financing.There is considerable scope for making moredomestic resources available by improvingrevenue collection, giving education a higherpriority and focusing more on basic education.Increased revenue collection and greater equitycould enhance domestic financing for basic

education by about 0.7% of GDP. This representsUS$7 billion, or two-thirds of current levels ofspending in the countries included in the study.Several countries have the potential to doublethe share of GDP allocated to basic education.Exploiting that potential should be part of theEducation for All contract between developingcountries and donors.

Aid donors need to undertake a ‘Gleneagles plus’aid commitment. Aid levels for basic educationin the forty-six countries covered need to risesixfold from their current level, fromUS$2.7 billion to US$16 billion.51 Even ifdonors act on the commitments made atthe 2005 Group of Eight summit in Gleneagles,Scotland, and substantially increase aid tothe poorest countries, the level will still fallUS$11 billion short. An emergency pledgingconference should be convened to mobilizethe additional financing required.

Reaching the marginalized requires additionalfinance. Failure to take into account the costsassociated with reaching marginalized groupshas contributed to systematic underestimationof the financing gap. It costs more to extendeducation opportunities to the mostdisadvantaged than it costs to reach better-offhouseholds. The new study estimates thatadditional measures to extend primary schoolopportunities to social groups facing extremeand persistent deprivation will cost US$3.7 billion annually.

The revised Education for All financing gappoints to challenges for both aid recipients anddonors. Developing countries need to increasethe level of ambition for public spending ineducation at a time when slower economic growthis putting budgets under pressure. Most majordonors, for their part, are gripped by recession andrising fiscal deficits. Some are cutting aid budgets.Others are reviewing future commitments. As thesepressures mount, it is important that governmentsrecognize the crucial role of education investmentsin creating the foundations for recovery and futurepoverty reduction efforts.

Part 1 of this section sets out the cost estimatesand the assumptions behind them. Part 2 presentsthe findings on the Education for All financing gap.

49. Full details of themethodology and resultsof the study are available in(EPDC and UNESCO, 2009).

50. These figures relate toaverage financing gaps andGDP levels in low-incomecountries between 2008and 2015.

51. Chapter 4 reports thatannual aid commitmentsto basic education averagedUS$4.9 billion in 2006and 2007 (see Figure 4.7).The low-income countriesincluded in this costingexercise received 55% ofthose commitments andUS$2.7 billion of total aid tobasic education. Chapter 4shows that low-incomecountries received 60% ofall aid to basic education,but the figure includes somesmall countries excludedfrom the costing exercise,as well as India, whichthe OECD-DAC defined asa low-income country atthe time of writing.

Ten years afterDakar, financeremains a majorbarrier toEducation for All

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 2 0

Costing the commitment to Education for All goals

Too often, governments and aid donors haveadopted bold goals at international developmentsummits but failed to put in place the financingmeasures needed to achieve them. How closelyare the Education for All goals aligned withcurrent financing?

Several studies have addressed this question.In 2003, the World Bank carried out a detailedanalysis of the financing required to achieveuniversal primary education in low-incomecountries (Bruns et al., 2003). Basing its estimateon assumptions about economic growth, revenuecollection, public spending and aid levels, the studyput the annual financing gap at US$3.6 billion(constant 2000 prices). The first EFA GlobalMonitoring Report 2002 adjusted this estimate formore moderate economic growth, the impact ofHIV and AIDS, and the inclusion of cash-transferprogrammes targeted at girls and poorhouseholds. These adjustments increased theestimated financing gap for universal primaryeducation to US$5.6 billion (constant 2000 prices).The EFA Global Monitoring Report 2007 updatedthis estimate to reflect the fact that aid levelshad been lower than expected. Rough estimateswere also made of additional financing for earlychildhood care and education, and literacy.52

These adjustments produced an annual financinggap estimate in low-income countries of US$11 billion (constant 2003 prices) – a figurethat has been widely used as a reference pointby the international community.

The costing exercise undertaken for thisReport provides a comprehensive review andreassessment of the financing gap (EPDC andUNESCO, 2009). Using the latest available nationaldata, the study updates the global estimate forlow-income countries.53 It covers a wider range ofeducation goals than in the 2003 study, recognizingthat Education for All is about more than universalprimary education. Another significant departureis the estimation of costs for reaching the mostmarginalized. Earlier studies assumed the costof extending education to out-of-school childrenwas the same as the average cost of providingeducation to those in school, but this assumptionis flawed. Many children in the most marginalizedgroups live in remote areas and suffer chronicpoverty and extreme gender disadvantage.Reaching these children requires higher levels

of spending, not just on providing schools andteachers but also on supporting demand foreducation.

There are strong grounds for factoring in theseadditional costs. Marginalized children have thesame right to education as others and that rightcarries with it a claim on financial resources.Equity in public spending means governmentsmust assess what it takes to deliver equivalentopportunities to children in very differentcircumstances. The fact that marginalized childrenhave benefited less than others from past publicspending reinforces their claim to fairer treatment.Moreover, failure to consider the financing neededto reach those who have been left behind willguarantee that many countries miss theEducation for All targets.

Financing gap estimates cannot be consideredin isolation. The same level of financing in twodifferent countries can produce widely divergentresults. Countries vary not just in their individualcost structures, but also in efficiency and equityin public spending on education. Some countriesachieve more for less because they have moreefficient procurement systems, school constructionprogrammes and textbook supply arrangements.National differences in teacher remuneration,the biggest single item in most education budgets,can have an enormous bearing on relative coststructures. The level of equity matters becauseit influences the degree to which increased publicspending translates into advances for the mostmarginalized. For all these reasons, averagecosts vary widely by country. The marginal costsassociated with reaching disadvantaged groupsare likely to depend on factors such as the depthof poverty and structures of inequality.

The limitations of global costing exercises have tobe recognized. Such exercises can help establishbroad orders of magnitude for the financingrequired to achieve specified goals. But theycannot substitute for detailed estimates drawn upat the national level. Bottom-up estimates providemuch clearer insights into the financing needed toachieve policy goals. It is a matter of concern that,almost a decade after the World Education Forumin Dakar, governments and donors continue toaddress this task in such a fragmented andhaphazard fashion – an issue taken up further inChapter 4. The estimates for the Education for Allfinancing gaps presented in this Report are basedon the most recent data available (Box 2.26).

52. The financing gapestimate for the literacytarget was half theUS$2 billion annualestimate of the cost ofliteracy programmes inall developing countries –a much larger group thanthe low-income groupused for the othercomponents of theestimate (Van Ravensand Aggio, 2005). It wasalso assumed that thefinancing gap for earlychildhood care andeducation was similarto that for literacy.

53. The study covers forty-six of the forty-ninecountries classified bythe World Bank as lowincome as of April 2009.The study excludesSolomon Islands andSao Tome and Principebecause their populationswere below 1 million andthe Democratic People’sRepublic of Koreabecause of lack of data.It includes the Sudanbecause southern Sudanhas a separate educationsystem and can beconsidered low income.

Countries need to consider the

financingrequired to reach

those who havebeen left behind

P R O G R E S S T O WA R D S T H E E FA G O A L S

E s t i m a t i n g t h e c o s t o f a c h i e v i n g E d u c a t i o n f o r A l l

1 2 1

However, their indicative and provisional natureis readily acknowledged.

Estimating the financing required to meet theEducation for All goals poses several problemsbecause the goals set in Dakar do not all includequantitative targets. In addition, quantifyingfinancing gaps means measuring the differencebetween estimated costs and domestic financingcapacity. Determining the latter involvesidentifying the degree to which low-incomecountries can make domestic resources available,taking into account economic growth prospectsand public spending levels. The followingsubsections set out the parameters for theseareas.

Identifying the targetsIn the Dakar Framework for Action, governmentsmade a commitment to achieve universal primaryeducation by 2015. This is a clear, quantifiable andmeasurable goal, though its precise meaning isopen to interpretation. There are also quantifiabletargets for adult literacy. Other goals of greatimportance lack clear targets. Examples includethe injunctions to improve education quality andensure access to appropriate learningprogrammes for young people and adults. Insome cases, goals relating to quality and equitydefine important principles but do not establishclear benchmarks. Targets chosen for thiscosting exercise cover four areas (Table 2.8):

Early childhood care and education. It is widelyrecognized that good early childhood educationis important not just in its own right but also asa way to improve participation and learningachievement in primary education. Buildingon previous work, this exercise adopts a targetof providing free pre-primary education to allchildren living below the poverty line (VanRavens and Aggio, 2007, 2008). This translatesinto an average gross enrolment ratio of 52%by 2015 for countries included in the exercise.

Universal primary education. For the purposesof this exercise, it is assumed that all primaryschool age children enter school on time andprogress through school with limited repetitionand no dropout, implying a net enrolment ratioof 100% by 2015.54

Lower secondary education. The DakarFramework does not include targets forsecondary education, but increasingparticipation at this level is important. ThisReport therefore includes a costing exercisethat assumes that all children completingprimary education by 2015 will make thetransition to lower secondary school, implyingan average gross enrolment ratio of about 88%by 2015 for countries included in the exercise.

Adult literacy. The Dakar target of halvingadult illiteracy will require wide-ranginginterventions. Part of the target will be achieved

54. This assumption has theeffect of understating costsassociated with reachingover-age children currentlyout of primary school.

The costingexercise providesa comprehensivereview andreassessment of the EFAfinancing gap

The study has drawn on the most recent cross-countrydata in preparing the global financing estimates.Variables covered include the size of the school agepopulation, school system structure and capacity,student progression rates (e.g. promotion, repetitionand dropout rates) and key costs such as those forteachers, classrooms and textbooks. The three mainsources of data are:

the United Nations World Population Prospectsdatabase, for information on school age populationsand projections of population growth;

the UNESCO Institute for Statistics, for informationon enrolment, student progression rates, teachers,classrooms and education financing;

the World Bank and International Monetary Fund,for information on overall government revenueand economic growth projections.

These were supplemented by over thirty othersources, including national education sector reportsand plans, and public expenditure reviews. Informationwas also collected directly from UNESCO offices andother studies commissioned for the costing exercise(Box 2.27). In the few cases where no national datawere available, regional aggregates were used. Therewere often large differences in reported data for agiven country, particularly with respect to educationcosts. Every effort was made to use the best availabledata. However, in some cases there are large marginsof error. Overall, the cost estimates should be treatedas indicative of the magnitude of financing gaps inlow-income countries.

A detailed outline of the methodology, data andresults is available in EPDC and UNESCO (2009).

Box 2.26: Information used for the global cost estimates

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 2 2

Reducing the cost burden on

households is a priority for

improving access

Table 2.8: Targets for the global costing exercise

Criteria for achievement by 2015Goal

Average for 46 low-income countries(circa 2007)

Targetfor 2015

Early childhood care andeducation

Universal primary education

Expansion of lower secondaryschooling

Adult literacy

Gender parity and equality

Education quality

Provision of pre-primary schooling for allchildren living below the poverty line

Provision of school places of good quality for all children of primary school age

Provision of places in lower secondary schoolfor all children completing primary school

Provision of sufficient literacy programmeplaces for illiterate adults to ensure thatilliteracy rates are halved from 1999 levels

Achievement of gender parity in primaryenrolment rates and lower secondary transition rates, and male and female literacyrates at or above target levels

Inclusion of a range of quality-enhancinginterventions at each education level

Pre-primary gross enrolment ratio = 16%

Primary gross enrolment ratio = 95%

Primary net enrolment ratio = 72%

Primary to secondary transition rate = 69%

Lower secondary gross enrolment ratio = 44%

Adult literacy rate = 59%

—

—

52%

108%1

100%

100%

88%

80%

Full parity

See Table 2.9

Notes: Targets for early childhood education and adult literacy are country-specific. The targets given in the table are unweighted averages for all countries covered.1. GER targets are country-specific but imply full enrolment of primary school age children with a maximum of 10% repetition.

Estimating education costs for countries affectedby conflict is problematic. In many such countries,access to the type of data required for ameaningful assessment of need is often lacking.

Innovative work for the global costing exercise set out in this chapter has made it possible to include several conflict-affected countries.UNESCO carried out detailed country-levelanalysis for the Democratic Republic of the Congoand for the Sudan, where conflict has seriouslycompromised education planning and datacollection. The analysis drew on recent surveys,including a 2006/2007 education census for theDemocratic Republic of the Congo (the first intwenty years), as well as detailed evidence oncosts from a range of donor, international agencyand national ministry sources.

This research draws attention to several importantconcerns. In both countries, the collapse of publicfinancing for education has shifted the burden to

households, which must cover half of overall costs in the Democratic Republic of the Congo and a third in the Sudan. Reducing the burden on households is a priority for improving access.

The case studies also highlight differences withineach country. In the Democratic Republic of theCongo, a legacy of weak governance and conflictstretching back over many years has resulted in a highly fragmented education system. Conflictand insecurity in some regions, notably the east,continue to hamper reconstruction prospects. In the Sudan, conflict has led to the developmentof separate political administrations and paralleleducation systems in the north and south.Financing for these systems varies. The bestestimates indicate that the north devotes 13% of government revenue to education, comparedwith 6% in southern Sudan, leading to largedifferences in spending per pupil. Primary schoolpupil/teacher ratios are 33:1 in the north and 51:1in the south.

Source: Chang et al. (2009).

Box 2.27: Basic education financing in the Democratic Republic of the Congo and the Sudan

P R O G R E S S T O WA R D S T H E E FA G O A L S

E s t i m a t i n g t h e c o s t o f a c h i e v i n g E d u c a t i o n f o r A l l

1 2 3

through increased participation and improvedquality of education at the primary level, withliterate school leavers driving down illiteracyrates. The residual element, representing about42% of the necessary decline, is assumed tooccur through adult literacy programmes.

An assumption underpinning the estimatespresented in this Report is that education isprovided at these levels without fees. This isconsistent with the Dakar Framework for Action.However, in many countries, particularly thoserecently involved in conflict, this would representa substantial shift in the burden of educationcosts from households to the state (see Box 2.27).

Setting targets for the cost parametersThe second step in the exercise is to developtargets for key parameters using country-levelinformation on costs. Recurrent costs per capitaand capital costs in education vary across andwithin regions, with significant implications forglobal cost estimates. Two factors account for

most of the variation. First, differences in averageefficiency associated with prices for importantinputs – such as teacher wages, building materialsand textbooks – inevitably influence cost structures.Second, countries have different norms and ruleson teacher remuneration, pupil/teacher ratios,school construction and other inputs. Table 2.9summarizes targets for the core cost parametersused in the estimates.

In setting the parameters, several difficultfinancing questions were considered. Teacherremuneration is one of the most significant andcontroversial areas in any costing exercise foreducation. This is typically the single biggestcomponent in the education budgets of low-incomecountries, often accounting for three-quarters oftotal spending. It follows that technical efficiencygains can dramatically reduce costs: adjustingsalaries in sub-Saharan Africa to levels found inSouth and West Asia would cut average costsby 40%. However, the issues at stake gobeyond considerations of technical efficiency.

Teacher salariesare typically thesingle biggestcomponent in theeducation budgetsof low-incomecountries

Table 2.9: 2015 targets for main cost parameters

Pre-primaryParameters Primary Lower secondary

Teacher salariesSub-Saharan Africa Other countries

Pupil teacher ratio

Percentage of non-salary costs in recurrent spending

School building and rehabilitation1

Share of private enrolment

School rehabilitation (% of classroom to be replaced)Low income countriesConflict affected countries

Cost drivers

4.5 times GDP per capita3 times GDP per capita

40

33%

$ 13,500 per classroom

10%

25%50%

4.5 times GDP per capita3 times GDP per capita

20

33%

$ 13,500 per classroom

Maintain current levels

25%50%

6 times GDP per capita3.5 times GDP per capita

35

40%

$ 17,000 per classroom

10%

25%50%

Targeted programmes for the marginalized

Demand side interventions (e.g. conditional cash transferprogrammes, school feeding programmes)

Supply side interventions (e.g. incentives for qualifiedteachers to work in remote areas, increased resources for schools serving marginalized groups)

—

—

5% of GDP per capita per marginalized student

Additional 33% of per pupil recurrent costs per

marginalized student

7.5% of GDP per capita per marginalized student

Additional 33% of per pupil recurrent costs per

marginalized student

Effect on per-pupil recurrent costs (constant 2007 US$)

Estimated current unit costs

Per-pupil costs in 2015 with additional policy measures

106

102

68

125

119

162

Notes: Per-pupil costs for 2015 do not include additional costs of demand-side and supply-side interventions for reaching the marginalized.1. Includes maintenance, estimated at 2% of construction cost. Classroom construction and rehabilitation include the cost of building school infrastructure (including latrines, offices, water supply, etc.) and providing access for children with disabilities.Source: EPDC and UNESCO (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 2 4

For example, lowering teacher salaries maycut costs but lead to low morale, making it moredifficult to recruit a workforce with sufficient skillsand forcing teachers to supplement their paywith other work.

Capital cost estimates raise another set ofdifficulties. By definition, achieving Educationfor All requires school infrastructure that isaccessible to all children and of sufficient qualityto ensure safety and provide an appropriatelearning environment. Costs of classroomconstruction vary enormously. Reasonable-practice standards point to a cost of aboutUS$13,500 per classroom, rising to US$17,000for lower secondary school.55 Classroomsobviously need to be built to accommodatechildren currently out of school. But thedilapidated state of the school infrastructure inmany countries means there is also a need forextensive investment in rehabilitation. Onerecent survey suggests that 30% of classrooms in low-income sub-Saharan Africa need replacing(Theunynck, 2009). Conflict-affected countriesface particularly pressing problems. To takeone example, half of Liberia’s classrooms weredestroyed or sustained major damage duringthe civil war (Liberia Ministry of Education, 2007b).

The parameters set for this Report’s costingexercise are derived from international evidenceon norms and current practice in key areas.They include the following:

Teacher salaries. Individual countries have toaddress issues of efficiency, norms and standardsfor teachers in the light of national circumstances.The costing exercise does not prejudge theappropriate teacher salary level. Instead, it takesthe current regional average for primary and lowersecondary salaries in sub-Saharan Africa as a long-term target that all countries in the region willconverge on.56 For countries outside sub-SaharanAfrica, the benchmark is lower (see Table 2.9).57

Rules and norms. While there is some debate aboutoptimal pupil/teacher ratios, here the bar is set at40:1 for primary school, reflecting the target used inprevious costing exercises. Effective teaching alsorequires access to stationery, textbooks and otherlearning equipment. Ensuring that one-third of therecurrent budget is directed towards non-salarycosts (rising to 40% for lower secondary education)should enable most low-income countries to meetbasic needs in this respect.58

Wider capital costs. As well as covering the costof future enrolment, budgets have to absorb thecost of replenishing infrastructure. A conservativeestimate is that about a quarter of the classroomsin low-income countries need replacing, risingto half in conflict-affected countries. As withthe other targets, it is assumed that all thisreplacement takes place by 2015.

Cost of adult literacy programmes. In line withprevious studies, the unit cost of adult literacyprogrammes is estimated at 8.9% of GDP percapita for countries in sub-Saharan Africa and5.3% for all other countries (Van Ravens andAggio, 2005, 2007).

Reaching the marginalizedPrevious global cost estimates for educationhave assumed that the average cost of reachingout-of-school children is roughly equivalent toa national average benchmark. That assumptionis misplaced. Specific programmes targeting highlymarginalized groups including child labourers,the extreme poor, ethnic minorities, girls, childrenwith disabilities, and locations such as remoterural areas and slums have to be financed.Moreover, extending education programmesto these groups and areas is likely to raise percapita spending requirements.

Top-down estimates are a particularly blunt toolfor assessing the financing required to reach themarginalized. Policy-makers need to consider theinterlocking constraints that keep marginalizedchildren out of school or that disrupt theirparticipation and limit their learning achievements(see Chapter 3). Detailed poverty assessmentsand planning processes that draw on the evidenceand perspectives provided by the marginalizedthemselves are critical to policy design.

With this caveat in mind, international evidenceyields some useful insights. Cash transferprogrammes that provide social protection canplay an important role in insulating vulnerablehouseholds from external shocks, enabling themto keep children in school. In some contexts, suchprogrammes have played a particularly crucialrole in allowing girls to enter and stay in school.Under the right circumstances, school feedingprogrammes can also provide strong incentivesfor children to attend school (as well as crucialhealth benefits). Effective programmes of thiskind typically cost about 5% of GDP per capita(Bundy et al., 2009a; Fiszbein et al., 2009a)59.

55. Classroomconstruction costs includeadditional infrastructurerequired for an effectivelearning environment,such as furniture, latrinesand water supply. Unitcosts are based on anaverage of low- and high-cost constructionscenarios in Theunynck(2009). Unit costs forlower secondary areassumed to be 25%higher.

56. The regional averagealso corresponds toaverage salary targets innational education plansfor sub-Saharan Africa(Bennell, 2009a).

57. For the low-incomecountries included in thecosting exercise that arenot in sub-Saharan Africa,average teacher salariesare 2.5 times GDP percapita in primary schooland 3.0 times insecondary.

58. The costs of reachingthe marginalized areexcluded when non-salaryspending as a proportionof total recurrent cost is calculated.

59. The wide range of policy interventionsneeded to addressmarginalization areexplored in detail inChapter 3.

Specificprogrammes

targeting highlymarginalized

groups are likely to have higheraverage costs

P R O G R E S S T O WA R D S T H E E FA G O A L S

E s t i m a t i n g t h e c o s t o f a c h i e v i n g E d u c a t i o n f o r A l l

1 2 5

International evidence on the incremental coststhat might be associated with creating a high-quality learning environment is more fragmentedand inconsistent (Chanamuto, 2009). Gettingteachers to schools in remote rural areas, slumsand other marginalized environments requiresincentives, but on what scale? Providingschooling to children whose lives have beenblighted by poverty, hunger, stigmatization andlow expectations is likely to requiresupplementary teaching and additional teachingmaterials, but there is no established benchmarkfor estimating the additional financing required.

For the purposes of the costing exercise,three criteria are used to introduce equity-basedfinance:

Assessing the size of the school age populationrequiring additional support. Drawing on a newstatistical source – the Deprivation andMarginalization in Education (DME) data setintroduced in Chapter 3 – the Report establishes,for each country, the share of the populationaged 17 to 22 with fewer than four years inschool. This is used as a proxy indicator for theproportion of the school age population that ismarginalized and in need of additional incentivesand school resources to participate in basiceducation. There are obvious limitations to thismeasure, including the fact that it captures pastoutcomes rather than the current situation.However, the four-year benchmark is a usefulmeasurement of the scale of marginalization.Chapter 3 explores the use of this measureof marginalization in greater detail.

Providing incentives for marginalized children.The costing exercise includes financial provisionfor incentives aimed at marginalized groups.It assumes a cost per child of 5% of GDP per capitafor primary school students and 7.5% for lowersecondary school students.

Creating an incremental financing coefficient.There are no ready-made standards that can beapplied on a cross-country basis. For the purposesof the costing exercise, the cost parameter forreaching the marginalized is set at an increment of33% above average recurrent costs. This is broadlyconsistent with the sparse evidence available onthe cost of financing teacher incentives and othermeasures to bring good-quality education tomarginalized children (Chen and Mulkeen, 2008;Mulkeen, 2009a).

The global costThe aggregate costs that emerge from the analysisare anchored in national data for the forty-six low-income countries covered.60 For each of the Education for All targets selected, the normsfor education inputs are applied to the size of thepopulation that has to be reached in each country.This makes it possible to identify the number ofteachers, additional classrooms and teachingmaterials required. The cost parameters forthese inputs are then applied, with adjustmentsfor reaching the marginalized. Table 2.10 showsthe resulting cost projections. To summarize:

Cumulative costs over 2008–2015 for the basic education goals run to US$286 billion, or US$36 billion annually (in constant 2007 US$).Current spending on basic education is about

60. The base year is 2007 and estimates are based onthe 2008–2015 period unlessotherwise indicated.

The financingrequired toachieve the basiceducation goals ismore than doublecurrent levels of spending

0.8 11.1 – 11.9 4.7 16.660.4 220.4 5.1 285.9 127.8 413.7

7.5 27.5 0.6 35.7 16.0 51.7

39 40 – 40 36 3941 27 – 30 30 30– 14 – 11 12 11

20 20 – 20 22 21

Current domestic resources (circa 2007)Cumulative cost (2008-2015)Average annual cost (2008-2015)

TeachersClassroom constructionProgrammes to reach the marginalizedOther

Table 2.10: Costs of achieving Education for All in low-income countries

Pre-primary

Notes: Breakdown of costs for basic education subtotal relates only to pre-primary and universal primary education. Subtotals are based on non-rounded figures.* The estimated adult literacy costs for the low-income countries covered are about three times the costs estimated in the original study (see Van Ravens and Aggio, 2005).Source: EPDC and UNESCO (2009).

PrimaryAdult

literacy*Basic education

sub-totalLower

secondary Total

US$ billions (constant 2007 prices)

Breakdown of costs between 2008-2015 (%)

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 2 6

US$12 billion a year. In other words, thefinancing required to achieve the basiceducation goals is three times the current level.

Some US$3.7 billion, or 14% of annual spendingon primary education, is required to financeprogrammes and interventions aimed atreaching the marginalized at the primary level.

Additional teacher costs account for 40%of the required spending for basic education,with classroom construction adding justunder one-third.

Factoring in lower secondary education raisesthe average annual cost by US$16 billion.

Human resources figure prominently in the costestimates (Table 2.10). Collectively, the countriescovered in the exercise have to recruit 3.2 millionmore primary and pre-primary teachers toachieve the basic education goals. At the primarylevel, the global cost of financing recruitmenton this scale is some US$9.1 billion annually.61

Translated into current national budget terms,this implies a significant increase in spending.Ten countries need to more than double spendingon primary teacher salaries and thirteen countriesmore than triple spending, from 2007 levels(Figure 2.45).

Overcoming school infrastructure deficitswill require large increases in investment.An estimated 6.2 million additional classroomswill be needed in primary and pre-primaryeducation to accommodate the increase inenrolment required to achieve the targets set.62

Current rates of construction fall far belowthe level required in most countries. In Burundi,Rwanda and Uganda, recent rates of classroomconstruction are less than 15% of the raterequired to achieve universal primary education(Figure 2.46). The estimated aggregate costsof expansion are highest in countries with thebiggest out-of-school populations. However,a large group of countries will have to increasespending on classroom construction andrehabilitation far above current levels to bringthe Education for All targets within reach.

61. This figure is reachedby multiplying theadditional teachersneeded from 2008 to 2015by the target teachersalaries shown inTable 2.9.

62. This includes newclassrooms to achievetargeted pupil/classroomratios and thereplacement of oldclassroom stock.

An estimated6.2 millionadditional

classrooms willbe needed

in primary andpre-primary

education

Figure 2.45: Spending on teachers has to rise% increase required from 2007 in spending on primary schoolteachers to achieve universal primary education by 2015

Sierra LeoneLao PDR

TogoMadagascar

KenyaTajikistan

SenegalD. R. CongoMauritania

Burkina FasoNiger

Côte d’IvoireGhana

EthiopiaZambia

MaliMozambique

BeninNorth Sudan

U. R. TanzaniaUganda

BangladeshRwandaBurundi

CambodiaNigeria

ChadPakistan

HaitiLiberia

South SudanGambiaGuinea

SomaliaEritrea

C. A. R.Malawi

AfghanistanGuinea-Bissau

Increase in spending required (%)

0 100 200 300 400 500 600 700 800

Note: Excludes countries that do not require increases in spending on primaryschool teachers to achieve universal primary education.Source: EPDC and UNESCO (2009).

Figure 2.46: Many countries need more classrooms Current classroom construction as % of required rate of constructionnecessary to achieve universal primary education by 2015

Notes: Estimated classroom needs for universal primary education include replacingold classrooms (the assumption is that classrooms have a forty-year life span) andrebuilding existing stock. The period covered by figures for actual annual growth inclassroom stock varies by country. See Table 1.1 in Theunynck (2009) for details.Sources: EPDC and UNESCO (2009); Theunynck (2009).

MauritaniaMadagascar

GuineaBurkina Faso

ChadMalawiRwandaUgandaBurundi

0 20 40 60 80 100

Current annual classroom construction as a percentageof annual construction requirement

Rwanda is currently constructing only 13% of therequired classrooms annuallyneeded to achieve UPE by 2015.

P R O G R E S S T O WA R D S T H E E FA G O A L S

E s t i m a t i n g t h e c o s t o f a c h i e v i n g E d u c a t i o n f o r A l l

1 2 7

Estimating the financing gap — and measures to close it

The cost of achieving the internationally agreedEducation for All goals has to be assessed againstthe financing available. National budgets are theprimary source of education financing. As theExpanded Commentary on the Dakar Frameworkfor Action recognized, developing countries willhave to do far more to make resources availableby ‘increasing the share of national income andbudgets allocated to education and, within that,to basic education’ (para. 46). Over and abovethese broad commitments, action is needed tostrengthen the efficiency and equity of educationspending, and to curb the diversion of resourcesassociated with corruption.

Most of the countries covered in the costinganalysis have the capacity to increase domesticspending on basic education. Increasedgovernment revenue, stronger budget commitmentand redistribution within the education budget allhave a role to play. But even with a strongerdomestic effort, many countries will be unable tofinance all the investment required. The analysisfor this Report estimates the Education for Allfinancing gap as the difference between the totalinvestment requirement indicated by the costingexercise and the domestic financing capacity ofgovernments making a ‘best effort’ to channelresources to education.

National governments can raise a substantialshare of the additional resources neededAlongside national income, the domestic resourceenvelope available for public financing of theEducation for All goals is ultimately determinedby three factors. The first is the share of nationalincome collected as government revenue. Thatshare rises on average with the level of per capitaincome, albeit with large variations by country thatreflect policies on taxation, the level of naturalresource exports and other national characteristics.The second factor is the proportion of revenuedirected into the overall education budget. The thirdis the share of the education budget allocated tobasic education. The proportion of national incomedirected towards basic education provides asummary overview of the level of public basiceducation financing.

Figure 2.47 presents the country-by-countrypicture. It shows the gap between current levelsof spending on basic education and the levels

required to achieve the goals set in this Report’scosting exercise. On average, the forty-six countriesneed to increase public spending on basiceducation by 2.5% of GDP to meet Education for Allgoals.63 However, there are very large variationsaround this average.

To what extent can low-income countriesincrease spending on basic education fromtheir own resources? Any attempt to addressthat question is highly sensitive to assumptions

63. This is an average figure weighted by the size of low-income countries in terms of their GDP.

Figure 2.47: Current national spending falls short of the levels needed to achievebasic education goalsCurrent and required spending on basic education as a share of GDP

Northern SudanPapua New Guinea

YemenKyrgyzstanTajikistan

MauritaniaLao PDR

MyanmarSouthern Sudan

PakistanCambodia

KenyaNepalGhana

BangladeshHaiti

Côte d’IvoireSenegal

Low income country averageBenin

NigeriaZambiaGuinea

TogoMadagascar

U. R. TanzaniaRwanda

AfghanistanGambia

Sierra LeoneMali

NigerMozambique

UgandaChad

EritreaBurkina Faso

EthiopiaC. A. R.

ZimbabweLiberia

SomaliaGuinea-Bissau

D. R. CongoMalawiBurundi

0 2 4 6 8 10 12 14 16 18 20

2007 estimated spendingon basic education

Estimated additionalspending required forbasic education in 2015

Basic education spending as % of GDP

Notes: Spending in 2007 is an estimate of domestic spending on education and excludes grants. Northern and southern Sudan are included separately in the costing study because of their separate education systems (see Box 2.27). Excludes Uzbekistan and Viet Nam which are projected not to require additional spending on basic education.Source: EPDC and UNESCO (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 2 8

about economic growth, revenue collectionand public spending patterns. For this analysis,a ‘best effort’ benchmark was established toassess national financing capacity. Factoring inprojected economic growth,64 a significant sourceof extra revenue, it is possible to estimate theadditional resources that can be generated by2015 if the countries covered (a) increase theaverage share of government revenue in GDP toat least 17%; (b) raise the share of revenue goingto education to at least 20%; and (c) ensure thatabout 70% of the education budget is devotedto pre-primary, primary and lower secondary.65

Where countries are already exceeding thesethresholds, it is assumed that current valuesare maintained up to 2015.

Applying these thresholds points to the scopefor a far stronger level of national effort. If everycountry covered in the study reached eachthreshold, it would expand the financial resourceenvelope for basic education on average byabout 0.9 percentage points of GDP by 2015.Put differently, it would provide slightly more

than a third of the additional resources requiredto achieve the basic education goals by 2015.

The aggregate picture inevitably obscuressignificant differences between countries(Figure 2.48). Some, such as Benin andMozambique, are close to the ‘best effort’thresholds in all target areas. By contrast,Chad combines high levels of revenue-raisingwith low levels of financial commitment toeducation. Nigeria raises 34% of national incomein government revenue, but has one of the lowestlevels of commitment to primary education amongthe forty-six countries covered. Pakistan performspoorly on all three counts: government revenuerepresents a small share of national income, theshare of revenue spent on basic education isamong the lowest for any of these low-incomecountries and the share spent on primaryeducation is the very lowest in the group. Thecountry has the potential to more than triplethe share of GDP currently allocated to basiceducation, suggesting that successivegovernments have failed to address the education

64. Resource projectionsuse the latest IMFeconomic growthforecasts (IMF, 2009f).

65. Targets for the exactshare of the budgetdevoted to Education forAll depend on the lengthof primary and lowersecondary cycles. SeeEPDC and UNESCO (2009)for details. Economicgrowth projections aretaken from the April 2009IMF forecast (IMF, 2009f).

Many countriescan mobilize

additionalresources for

basic education

Figure 2.48: Many countries can mobilize additional domestic resources for basic educationCurrent and additional resources countries devote to basic education as a share of GDP

Note: Excludes countries where the thresholds are already exceeded.Source: EPDC and UNESCO (2009).

North

Sud

anSo

uth

Suda

nGu

inea

Mal

iKe

nya

Beni

nM

ozam

biqu

eSe

nega

lBu

rund

iZa

mbi

aPa

pua

New

Gui

nea

Mal

awi

Som

alia

Lao

PDR

Rwan

daCô

te d

’Ivoi

reM

adag

asca

rU.

R. T

anza

nia

Tajik

istan

Haiti

Nepa

lLib

eria

Yem

enVi

et N

amLo

w in

com

e co

untry

ave

rage

Sier

ra Le

one

Mau

ritan

iaUg

anda

Ghan

aTo

goGu

inea

-Biss

auKy

rgyz

stan

C. A

. R.

Nige

riaBu

rkin

a Fa

soEt

hiop

iaD.

R. C

ongo

Zim

babw

eCa

mbo

dia

Chad

Mya

nmar

Bang

lade

shGa

mbi

aAf

ghan

istan

Pakis

tan

Eritr

ea

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Basic

edu

catio

n sp

endi

ng a

s % o

f GDP

Increasing scope for additional domestic resource mobilization

Additional resources mobilized in 2015 if thresholds met

2007 estimated resources for basic education

P R O G R E S S T O WA R D S T H E E FA G O A L S

E s t i m a t i n g t h e c o s t o f a c h i e v i n g E d u c a t i o n f o r A l l

1 2 9

financing challenge with sufficient urgency.In the cases of Chad and Nigeria, the problemis less one of revenue mobilization than the lowpriority attached to education in general andbasic education in particular.

It should be emphasized that the ‘best effort’thresholds used are an imperfect guide to publicpolicy. Revenue-raising capacity partly dependson export structures. Countries with largemineral assets may be better placed than othersto increase revenue collection. For countriesemerging from conflict, such as Nepal and SierraLeone, increasing the share of national incomecollected in revenue may be a slow processinvolving the restoration of credible publicinstitutions and confidence in government. Theestimates here should therefore be treated as anevaluation of what is possible under reasonableconditions, not as a full assessment of what eachcountry can achieve in practice. With the dataavailable, it is difficult to generate preciseEducation for All financing estimates forcountries such as Afghanistan, Liberia andSierra Leone, but there are strong grounds forrecognizing, as most aid donors have done, the

urgent need for a large up-front increase ineducation finance, given the limited capacity ofthese countries’ governments to raise that finance.

Donors need to increase aid to close the remaining gapSuccessive issues of the EFA Global MonitoringReport have drawn the attention of the donorcommunity to the gap between aid levels and thelevel of financing required to meet the Dakartargets. The revised global cost estimate suggeststhe gap is far larger than previously assumed. Anyprospect of accelerated progress towards the 2015targets hinges critically on a scaled-up donoreffort. The bottom-line message to emerge fromthe costing exercise is that two-thirds of theadditional resources required will have to beprovided through aid.

The residual aid component of the Education for Allfinancing requirement can be extrapolated fromthe costing exercise. Figure 2.49 summarizes thefinancing gap that remains once prospects foradditional domestic resources have beenexhausted. Table 2.11 provides an approximatebreakdown of this financing deficit by education

The financing gap for basiceducation is around US$16 billionannually

Figure 2.49: Financing gaps are large and unlikely to be eliminated by current donor pledgesBreakdown of annual resource needs to achieve basic education goals

Notes: Breakdown of annual resource needs does not add up to the total due to rounding. The percentage increase in aid between 2005 and 2010 associated with the Gleneagles targets (see Chapter 4) are usual to project 2005 basic education commitments to 2010 for each country covered.Sources: EPDC and UNESCO (2009); OECD-DAC (2009d).

Additional domestic resourcesfrom economic growth

Additional resourcesfrom domestic prioritization

0

5

10

15

20

25

30

35

40

Financing gapUS$16 billion

Currentestimatedresources

US$12 billion

US$4 billion

US$3 billionCons

tant

200

7 US

$ bi

llion

s

Additional aid to basic

Current aid to basic

Financing gap

Remaining shortfall US$11 billion

US$2 billion

US$3 billion

education if Gleneaglescommitments met

education (2006-2007)

Average annualresources needed

to finance EFA(2008-2015)

US$36 billion

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 2

1 3 0

sector and region. The deficit that will have tobe covered by increased development assistanceis projected to widen to 2015 before narrowing asthe domestic resource base expands and theneed for additional capital spending declines.Results of the analysis include the following:

Estimates of the financing gap for basiceducation are about 30% higher than theprevious global estimates.

Assuming that all low-income countries reachthe ‘best effort’ thresholds by 2015, theaggregate average annual financing gap in basiceducation for the low-income countries coveredis equivalent to about 1.5% of their collectiveGDP.66 The cumulative deficit for basic education,calculated on a country-by-country basis, isaround US$16 billion annually from 2008 to 2015.

Current aid levels cover only a small part ofthe Education for All financing deficit. For thelow-income countries included in this exercise,development assistance for basic educationamounts to US$2.7 billion (Figure 2.49).67

A sixfold increase in aid to basic educationwill therefore be required if the basic educationgoals are to be achieved.

Sub-Saharan Africa accounts for about 66%of the financing gap, or US$10.6 billion.

Low-income countries affected by conflictaccount for 41% of the gap, or US$6.7 billion.

Adding the costs of lower secondary educationincreases the gap to US$25 billion – a figure thatillustrates the enormous increase in resourcesrequired if countries are to universalize access.However, without addressing the financing gapsat the basic education level and building stronglearning foundations, increased investment inpost-primary education is unlikely to be equitableor to lead to the skills improvement thatgovernments and parents demand.

The global costing exercise raises importantquestions for the international community. Withjust five years remaining to the target date for theEducation for All goals and the wider MillenniumDevelopment Goals, the United Nations Secretary-General has called on donors to act on their 2005commitments, made at Gleneagles tosubstantially increase aid by 2010. Such a movewould clearly help narrow the education financinggap, but it would not fully close it. Holding constantthe distribution of aid between low-income and66. This is calculated

by dividing the averagefinancing gap by theaverage projected GDPof all countries includedin the costing exercisefrom 2008 to 2015.

67. Chapter 4 showsthat aid commitments tobasic education in 2006and 2007 averagedUS$4.9 billion (seeFigure 4.7). The low-income countriesincluded in the costingexercise received 55%of these commitments.

Sub-SaharanAfrica accounts

for about two-thirds of the

financing gap, or US$10.6 billion

5.8 66 23 29

9.8 68 28 48

0.6 42 37 51

16.2 66 27 41

8.8 60 35 42

25.0 64 30 42

Pre-primary

Universal primary education

Adult literacy

Basic education financing gap

Lower secondary

Total financing gap

Table 2.11: Average annual financing gaps in low-income countries, 2008–2015

(constant 2007US$ billions) (%) (%) (%)

Financing gap Sub-Saharan Africa South AsiaConflict-affected

countries

Education level

Note: The financing gap is the difference between the total investment requirement indicated by the costing exercise and levels of domestic financing associated with all countries reaching ‘best effort’ thresholds by 2015.Source: EPDC and UNESCO (2009).

P R O G R E S S T O WA R D S T H E E FA G O A L S

E s t i m a t i n g t h e c o s t o f a c h i e v i n g E d u c a t i o n f o r A l l

1 3 1

middle-income countries, and between differentlevels of education, full delivery of the 2005commitments would leave a deficit of US$11 billion(Figure 2.49). That scenario points to the case foran urgent reassessment of aid commitmentsand distribution patterns. With an internationalsummit on the Millennium Development Goalsplanned for 2010, donors should as a matter ofurgency convene a pledging conference to closethe Education for All financing gap.

Conclusion

The limitations and uncertainties associatedwith global financial costing models have to beacknowledged. Yet the results of the exerciseset out here provide a clear warning sign. In theabsence of an urgent, concerted effort to makenew and additional resources available foreducation, there is little prospect of the world’spoorest countries getting on track to meet the 2015targets. If the policy goal is to ensure that all theworld’s primary school age children are ineducation systems by 2015, the investment cannotbe delayed. The global costing exercise underlinesthe importance of low-income developing countriesand donors doing far more. However, the roleof donors is critical because governments inthe poorest countries lack the resources toclose the Education for All financing gap.

01

02

Education f

or

All

Glo

bal M

onitori

ng R

eport

Chapter 3

1 3 2

Compounded disadvantage:low-caste girls face thegreatest obstacles, India©

Am

i Vita

le/P

ANOS

0102Education for All Global Monitoring Report

Reaching the marginalized

1 3 3

On the move:pastoralist communities

require flexible solutions

© G

iaco

mo

Piro

zzi/P

ANOS

Support makes a difference:more girls in Yemen are now

going to school

© A

bbie

Tra

yler

-Sm

ith/P

ANOS

Who are the marginalized? What are the factors contributingto their exclusion and lack ofeducational opportunity? Thischapter looks at the mutuallyreinforcing interactions betweenpoverty, gender, ethnicity,geographic location, disability,race, language and other factorsthat create cycles of disadvantagein education. It also shows howintegrated anti-marginalizationstrategies can enable all children – regardless of circumstance – to enjoy their right to education.

Introduction .................................................................. 135

Measuring marginalizationin education .................................................................. 138

Getting left behind .............................................. 164

Levelling the playing field ....................... 186

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 3 4

R E A C H I N G T H E M A R G I N A L I Z E D

I n t r o d u c t i o n

1 3 5

Education is the great engine of personaldevelopment. It is through education that thedaughter of a peasant can become a doctor…thata child of farmworkers can become the presidentof a great nation. It is what we make out of whatwe have, not what we are given, that separatesone person from another.

– Long Walk to Freedom:The Autobiography of Nelson Mandela (1994, p.144)

Introduction

Education has the power to transform lives. Itbroadens people’s freedom of choice and action,empowering them to participate in the social andpolitical lives of their societies and equipping themwith the skills they need to develop their livelihoods.For the marginalized, education can be a routeto greater social mobility and a way out of poverty.Forged in a society that restricted educationon the basis of skin colour and discrimination,Nelson Mandela’s words powerfully capturethe role of inclusive education in broadeningopportunities and building inclusive societies.

This chapter focuses on marginalization ineducation. Marginalization is the subject ofmuch debate. There is a voluminous literatureon how to measure it and how to differentiatethe concept from broader ideas about inequality,poverty and social exclusion.1 Many importantissues have been raised. However, debate overdefinitions can sometimes obscure the politicaland ethical imperative to combat marginalization.Writing on the idea of justice, Amartya Sen arguesthat there are limits to the value of perfectingdefinitions. ‘What moves us,’ he writes, ‘is notthe realisation that the world falls short of beingcompletely just (…) but that there are clearlyremediable injustices around us which we wantto eliminate’ (Sen, 2009, p. vii).

The starting point in this Report is thatmarginalization in education is a form of acuteand persistent disadvantage rooted in underlyingsocial inequalities. It represents a stark exampleof ‘clearly remediable injustice’. Removing thatinjustice should be at the centre of the nationaland international Education for All agendas.

The focus of this chapter is on schools and basiceducation. While marginalization typically startslong before children enter school and continuesinto adult life, schools are in a pivotal position.They can play a vital role in counteracting earlychildhood disadvantage and help break thetransmission of illiteracy across generations.But schools can also reinforce disadvantageand perpetuate marginalization.

The experience of marginalization in educationtoday is seldom a consequence of formaldiscrimination. Legal restrictions on opportunity,such as those that characterized apartheid SouthAfrica, are rare. Yet informal discrimination iswidespread. It is embedded in social, economicand political processes that restrict life chancesfor some groups and individuals. Marginalizationis not random. It is the product of institutionalizeddisadvantage – and of policies and processesthat perpetuate such disadvantage.

Half a century ago, governments around the worldmade a clear statement of intent on education.In the 1960 Convention Against Discriminationin Education, they imposed what amounts toa comprehensive ban not just on discriminationby legal intent, but on processes that have theeffect of causing discrimination. As Article 1 ofthe Convention puts it,

the term ‘discrimination’ includes any distinction,exclusion, limitation or preference which, beingbased on race, colour, sex, language, religion,political or other opinion, national or social origin,economic condition or birth, has the purposeor effect of nullifying or impairing equalityof treatment in education and in particular:

(a) Of depriving any person or group of personsof access to education of any type or at any level;

(b) Of limiting any person or group of personsto education of an inferior standard[.] (UNESCO, 1960, Article 1, para. 1).

Underpinning this provision is the simple butcompelling idea of equal opportunity. That ideais at the heart of many international human rightsprovisions, starting with the 1948 UniversalDeclaration of Human Rights. The 1989 Conventionon the Rights of the Child establishes a bindingobligation on governments to work towards fulfillingthe right to education ‘progressively and on thebasis of equal opportunity’ (United Nations, 1989,

1. See, for example, Kabeer (2005), Sayed et al. (2007), Klasen (2001), Ferreira and Gignoux (2008),World Bank (2005f).

Marginalization in education is a form of acute and persistentdisadvantagerooted inunderlying socialinequalities

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 3 6

Article 28). The right to equal opportunity foreducation is also enshrined in most countries’national laws and constitutions. Indeed, fewhuman rights are more widely endorsed – andmore widely violated.

Millions of children are denied their human rightto education for the simple reason that theirparents cannot afford to keep them in school.Social and cultural barriers to education formanother formidable obstacle. In many countries,the education of girls is widely perceived as beingof less value than that of boys, with traditionalpractices such as early marriage adding anotherlayer of disadvantage. Members of ethnic minoritiesoften face deeply entrenched obstacles to equalopportunity. Denied an opportunity to learn in theirown language and faced with social stigmatization,they are set on an early pathway to disadvantage.Millions of children with disabilities across theworld also face far more restricted opportunitiesthan their peers, as do children living in regionsaffected by conflict.

None of these disadvantages operates in isolation.Poverty, gender, ethnicity and other characteristicsinteract to create overlapping and self-reinforcinglayers of disadvantage that limit opportunity andhamper social mobility.

The interaction between marginalization ineducation and wider patterns of marginalizationoperates in both directions. Being educated isa vital human capability that enables people tomake choices in areas that matter. The lack ofan education restricts choices. It limits the scopepeople have for influencing decisions that affecttheir lives. People lacking literacy and numeracyskills face a heightened risk of poverty, insecureemployment and ill health. Poverty and ill health,in turn, contribute to marginalization in education.So does the fact that the marginalized have onlya weak voice in shaping political decisionsaffecting their lives.

Reaching marginalized children requires politicalcommitment backed by practical policies. Whengovernments met in 1990 at the World Conferenceon Education for All in Jomtien, Thailand, theyrecognized the need to overcome extremeinequalities holding back progress in education.They declared that ‘consistent measures must betaken to reduce disparities’ and called for activecommitment to reach ‘underserved groups’,including the poor, remote rural populations,

ethnic, racial and linguistic minorities, refugeesand migrants, and those affected by conflict(UNESCO, 1990, Article 3). The Dakar Frameworkfor Action reaffirmed the commitment to ‘explicitlyidentify, target and respond flexibly to the needsand circumstances of the poorest and the mostmarginalized’ (UNESCO, 2000, IV, para. 52).

While some countries have made impressiveefforts to back up such words by extendingeducational opportunities to their mostmarginalized populations, action has generallyfallen far short of the commitments made atJomtien and Dakar. Marginalization has remaineda peripheral concern. The assumption has beenthat national progress in education wouldeventually trickle down to the most disadvantaged.After a decade of steady but uneven nationalprogress, it is time to abandon that assumption.In many countries, large swathes of societyare being left behind as a result of inheriteddisadvantages. Breaking down thesedisadvantages will require a far stronger focuson the hard to reach.

Tackling marginalization is a matter of urgencyon several counts. The targets for 2015 adoptedin the Dakar Framework for Action – includinguniversal primary education – will not beachieved unless governments step up theirefforts to reach the marginalized. Sustainingprogress in basic education and creating thefoundations for advances in secondary educationwill require a renewed drive to extendopportunity to individuals and groups facingthe most deeply entrenched disadvantages.Progress in combating marginalization ineducation would dramatically improve thediscouraging scenario that Chapter 2 describes.

The case for action on marginalization goesbeyond the 2015 targets. Extreme and persistentdeprivation in education carries a high pricefor societies as well as for individuals. In theincreasingly knowledge-based and competitiveglobal economy, depriving people of opportunitiesfor education is a prescription for wastage of skills,talent and opportunities for innovation andeconomic growth. It is also a recipe for socialdivision. Marginalization in education is animportant factor in the widening of social andeconomic inequalities. Working towards moreinclusive education is a condition for thedevelopment of more inclusive societies.

Extreme andpersistent

deprivation ineducation carries

a high pricefor societies

as well as forindividuals

R E A C H I N G T H E M A R G I N A L I Z E D

I n t r o d u c t i o n

1 3 7

The core message of this chapter is thatovercoming marginalization must be at the heartof the Education for All agenda. Education shouldbe a driver of equal opportunity and social mobility,not a transmission mechanism for social injustice.The familiar routine of governments endorsingequal opportunity principles, reaffirming humanrights commitments and signing up forinternational summit communiqués on educationis not enough. Overcoming marginalizationrequires practical policies that address thestructures of inequality perpetuatingmarginalization – and it requires political leadersto recognize that marginalization matters.This chapter has four main messages:

Governments across the world aresystematically violating the spirit and the letteron United Nations conventions obliging them towork towards equal opportunities for education.The failure of many governments to act decisivelyin tackling marginalization in education callsinto question their commitment to the humanright to education – and it is holding backprogress towards the Education for All goals.The scale of the marginalization crisis ineducation is not widely recognized, partlybecause the marginalized themselves lackan effective voice.

Disaggregated data can play an importantrole in identifying social groups and regionscharacterized by concentrated marginalization.All too often education policies are developedon the basis of inadequate information aboutwho is being left behind. Data have a vital roleto play in providing an evidence base fordeveloping targeted interventions and widerpolicies. This chapter sets out a new statisticaltool – the Deprivation and Marginalization inEducation (DME) data set – that looks beyondnational averages to provide insight intopatterns of marginalization.

Mutually reinforcing layers of disadvantagecreate extreme and persistent deprivationthat restrict opportunity. Poverty and genderinequalities powerfully magnify disadvantageslinked to ethnicity, language, living in ruralareas and disability, closing doors to educationalopportunities for millions of children. Moreover,stigmatization and social discrimination arepotent drivers of marginalization in education.

Good policies backed by a commitment to equitycan make a difference. Education systems canplay a central role in overcoming marginalizationby giving disadvantaged children access to agood-quality learning environment, includingproperly financed schools, motivated and well-trained teachers, and instruction in anappropriate language. But strategies in educationhave to be backed by wider interventions,including investment in social protection,legal provisions to counteract discriminationand wider empowerment measures. Thechallenge is to ensure that education policiesand broader anti-marginalization policiesoperate within a coherent framework.

The chapter is divided into three parts. Part 1provides a snapshot of the scale of extreme andpersistent deprivation in education. Drawing onthe DME data set, it measures marginalization bylooking at numbers of years spent in school. Part 1also explores problems in education quality ascaptured in measures of learning achievement.Part 2 looks at the social and economic processesbehind the data. It explores some key forces behindmarginalization, including poverty, gender, ethnicityand location. Part 3 provides an overview of policiesand approaches that can break down the structuresthat perpetuate marginalization in educationand beyond. While each country is different andthere are no ready-made ‘anti-marginalization’blueprints, there are models for good practice.These models can help inform policy choicesfor governments seeking to act on the obligationto ensure that all of their citizens enjoy a rightto education.

Overcomingmarginalizationmust be at the heart ofthe Education for All agenda

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 3 8

Measuring marginalizationin education

Introduction

In all countries, whatever their level ofdevelopment, some individuals and groupsexperience extreme and persistent disadvantagein education that sets them apart from the rest ofsociety. They are less likely to enter school, to startschool at the correct age or to complete a fullcycle of education, and they are more likely to leaveschool with lower levels of achievement. As well asbeing a sign of social deprivation in its own right,disadvantage in education is a cause and an effectof marginalization in other areas and a powerfultransmitter of deprivation across generations.

Defining who is marginalized is problematicbecause there is seldom an agreed definition ofthe term within any one country, let alone acrosscountries. Establishing what marginalization entailsin education presents another set of problems.Most people would accept that it encompassesquantitative deprivation, as measured by yearsin school or the level of education attained.But it also incorporates a qualitative dimension.The marginalized typically demonstrate lower levelsof educational achievement. The Convention on theRights of the Child calls on governments to providean education that leads to the ‘development of thechild’s personality, talents and mental and physicalabilities to their fullest potential’ (United Nations,1989, Article 29). For many children, though,the experience undermines learning potential,disempowers and stigmatizes them (Klasen, 2001).

This section identifies some of the characteristicsthat predispose individuals and groups to extremeand persistent disadvantage in education. Whileall countries endorse the principles of equalopportunity and universal rights, the evidenceshows that, when it comes to opportunities foreducation, some people are more equal thanothers – the marginalized being the least equalof all. Inequalities linked to parental income,gender, ethnicity, race and other factors continueto restrict life chances and fuel marginalization.

Understanding marginalization is one of theconditions for overcoming it. Too often,governments express commitment to equalopportunity in education but fail to monitor what

is happening to the individuals and groups beingleft behind. One of the central messages of thissection is that countries need to invest in morerobust and consistent data analysis to identifyareas of concentrated disadvantage. The newinternational data set prepared for this Reportprovides a tool that governments, non-governmentorganizations and researchers can use to makethe marginalized more visible.

Using a quantitative analysis of marginalizationin low-income developing countries, this sectiondraws on the DME data set to identify individualsand groups facing heightened risk ofmarginalization, with respect both to absolutedeprivation, defined in terms of years in school,and to disadvantage relative to the rest of society.The section looks also at individual and group-based disadvantage with respect to learningachievement. While the dimensions andcharacteristics of marginalization differ betweendeveloped and developing countries, rich countriesare also characterized by extreme and persistentpatterns of deprivation.

The Deprivation and Marginalizationin Education data set

Measuring marginalization in education is notstraightforward. Household surveys and otherdata provide insights into the relationship betweenpoverty, ethnicity, health, parental literacy and othercharacteristics on the one side and education onthe other. But while these are all characteristicsassociated with marginalization, they do not operatein isolation. The marginalized in education are oftenpoor and female, and from an ethnic minority livingin a remote rural area. Understanding how differentlayers of disadvantage interact is a first steptowards breaking the cycles of disadvantagethat push people into marginalization.

Invisibility adds to measurement problems.Concentrated in slums or remote rural regions,the marginalized are often hidden from viewand government agencies sometimes have limitedaccess to detailed data for monitoring theircondition. All too often the same agenciesdemonstrate a marked indifference to the socialcircumstances of the marginalized, reflectingthe indifference of political elites.

The new DME data set assembled for this Reportis a statistical tool that helps chart the dimensionsof marginalization and identifies patterns of

Countries need to invest in more

robust andconsistent data

analysis toidentify areas

of concentrateddisadvantage

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 3 9

individual and group disadvantage. The data aredrawn from Demographic and Health Surveysand Multiple Indicator Cluster Surveys coveringeighty developing countries, including thirty-eightlow-income countries.2 Data from these sourceshave been reconstituted to concentrate on keydimensions of education marginalization. Theanalysis presented in this chapter focuses onthree core areas:

The bottom line: education poverty. Themarginalized typically fall below a socialminimum threshold for years of education.To measure absolute deprivation, this analysistakes four years as the minimum required togain the most basic literacy and numeracy skills.People aged 17 to 22 who have fewer than fouryears of education can be thought of as beingin ‘education poverty’. People with fewer thantwo years can be thought of as living in‘extreme education poverty’.

The bottom 20%. Time spent in education isone indicator for the distribution of opportunity.Using the DME data set, relative marginalizationis measured by organizing individuals aged 17to 22 according to the number of years theyhave accumulated in education. The analysisthen uses the results to identify the individualand group characteristics of the bottom 20% – the 20% with the fewest years of education.

The quality of education. Acquiring thelearning skills that people need to escapemarginalization means more than justspending time at school. What children actuallylearn depends on a wide range of factors,including the quality of education and homecircumstances. The analysis looks atmarginalization in learning achievementusing national and international evidence.

Patterns of marginalization reflect underlyinginequalities in opportunity. One advantage ofthe DME data set is that it provides detailedinformation on individual and group characteristicsof the marginalized, including wealth, gender,location, ethnicity and language. That informationprovides insight into the weight of ‘inheritedcircumstances’. These represent conditionsover which people have little control but whichplay an important role in shaping theiropportunities for education and wider life chances(Bourguignon et al., 2007; Ferreira and Gignoux,2008; World Bank, 2005f).

Measuring marginalization is not a narrowlydefined technical matter. It is an integral partof the development of strategies for inclusiveeducation. The DME data set helps increasethe visibility of the marginalized and provides aresource that can help inform policy design andpublic debate. Summary tables are presentedat the end of this section and the full data setis available in electronic form.

The scale of marginalization

Falling below the minimum threshold —education poverty Time spent in education is one of the mostimportant determinants of life chances in allsocieties. There is no internationally agreedbenchmark for education deprivation analogousto the US$2.00 and US$1.25 a day internationalpoverty thresholds. However, people with fewerthan four years of schooling are unlikely to havemastered basic literacy or numeracy skills, letalone built a foundation for lifelong learning.Those with fewer than two years are likely to faceextreme disadvantages in many areas of theirlives. Of course, learning achievement ultimatelydepends as much on the quality of education ason time spent in school. But the four year and twoyear thresholds are bottom lines that this analysistreats as indicators for ‘education poverty’ and‘extreme education poverty’, respectively.

Figure 3.1 uses these thresholds to provide asnapshot of education deprivation for sixty-threemostly low-income countries. It covers areference group of young adults aged 17 to 22.Even taking into account over-age attendance, thisis far enough beyond the standard primary schoolcompletion age to provide a credible picture ofwho has completed four years of education.

Three broad themes emerge. The first is thescale of global deprivation and inequality. In richcountries, the vast majority of young adults in thisage range will have accumulated ten to fifteenyears of education. In twenty-two of the countriescovered by the DME data, 30% or more of 17-to 22-year-olds have fewer than four years ofeducation; in eleven of these countries, the figurerises to 50%. Nineteen of the twenty-two countriesare in sub-Saharan Africa, with Guatemala,Pakistan and Morocco making up the remainder.

The second theme concerns cross-countrydifferences. On average, as one would expect,

2. Demographic and HealthSurvey data are collected aspart of the MEASURE DHSproject implemented by ICFMacro. Seehttp://www.measuredhs.com/.Multiple Indicator ClusterSurveys are collected byUNICEF. Seehttp://www.unicef.org/statistics/index_24302.html.

Measuringmarginalization isan integral part ofthe developmentof strategies forinclusive education

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 4 0

the share of the population with fewer thanfour years or fewer than two years of educationfalls as the national average for years ofeducation rises. Countries averaging more thaneight years of education typically have fewerthan 10% falling below the four-year threshold.This broad association conceals as much as itreveals, however. For example, Egypt averagesmore years of education than Kenya but has alarger share of 17- to 22-year-olds with fewerthan four years of education. Such comparisonspoint to deeply entrenched national inequalitiesthat are obscured by national average figures.

Comparisons of the depth of education povertypoint in the same direction. In countries with verylow average years of education, the majority ofpeople falling below the four-year threshold alsohave fewer than two years of education. However,Pakistan has a lower share of the population withfewer than four years than Rwanda, but a 50%

higher share with fewer than two years. Thesecomparisons illustrate the variation in the degreeto which all sections of society share in averageprogress in education.

The third theme to emerge from Figure 3.1 is thescale of national disparities based on income andgender. Wealth-based inequalities are a universalsource of disadvantage in education. Being borninto the poorest 20% significantly raises the riskof falling below the four-year threshold. In almosthalf of the countries including Cambodia, Ghana,Guatemala, India, Nicaragua, Nigeria and Yemen,the incidence of four-year education deprivationamong the poor is double the national average.In the Philippines, being poor increases thelikelihood of a 17- to 22-year-old having fewerthan four years in education by a factor of fourcompared with the national average.

Wealth-basedinequalities are auniversal sourceof disadvantage

in education

Tajik

istan

, 200

5

Indo

nesia

, 200

3

Phili

ppin

es, 2

003

Peru

, 200

0

Boliv

ia, 2

003

Colo

mbi

a, 2

005

Syria

n A.

R.,

2005

Lao

PDR,

200

5

Turk

ey, 2

003

Dom

inica

n Re

p., 2

007

Viet

Nam

, 200

2

Nam

ibia

, 200

6

Mon

golia

, 200

5

Swaz

iland

, 200

6

S. To

me/

Prin

cipe,

200

0

Egyp

t, 20

05

Vene

zuel

a, B

. R.,

2000

Gabo

n, 2

000

Surin

ame,

200

0

Keny

a, 2

003

Cong

o, 2

005

Leso

tho,

200

4

Ugan

da, 2

006

Hond

uras

, 200

5

Ghan

a, 2

003

Zam

bia,

200

1

Cam

eroo

n, 2

004

Togo

, 200

5

Haiti

, 200

5

Nica

ragu

a, 2

001

Indi

a, 2

005

Mal

awi,

2004

D. R

. Con

go, 2

007

Nige

ria, 2

003

Yem

en, 2

005

Average number of years of education: between 6 and 8 yearsAverage number of years of education: more than 8 years

0

10

20

30

40

50

60

70

80

90

100

Shar

e of

the

popu

latio

n w

ith fe

wer

than

4an

d fe

wer

than

2 ye

ars o

f edu

catio

n (%

)

Extreme education poverty:Population with fewer than 2 years of education

Education poverty:Population with fewer than 4 years

of education

Figure 3.1: Measuring education poverty across countries% of national population, the poorest households, and girls in poorest households aged 17 to 22 with fewer than four years and fewer than two years of education, selected countries, most recent year

Source: UNESCO-DME (2009).

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 4 1

Gender effects magnify poverty effects – and viceversa. Being poor and female carries a doubledisadvantage in many countries. Figure 3.1highlights the distance that separates girls in thepoorest households, not just from the nationalaverage but also from boys in poor households.Gender disparities play an important role inexplaining the relatively high level of educationpoverty in Egypt. Young women in the country aretwice as likely as young men to have fewer thanfour years of education – and four times as likely ifthey are poor women. The incidence of deprivationamong poor women in Egypt is higher than insome other countries, such as Honduras, Ugandaand Zambia, at far lower levels of average income.Young women from the poorest households inMorocco are more likely to have fewer than fouryears in education than their counterparts inSenegal. In Yemen, 90% of poor young womenaged 17 to 22 years have fewer than four yearsin education compared with 30% for poor males.

While data on those aged 17 to 22 provide insightinto the legacy of deprivation, current attendancepatterns reflect the degree to which disadvantageis transmitted across generations. Figure 3.2 showsincome and gender disparities in sub-SaharanAfrica, and South and West Asia are narrowing overtime but remain very large. The household surveyevidence in the DME data set indicates that 38% ofchildren aged 7 to 16 from the poorest householdsin sub-Saharan Africa and 26% in South and WestAsia have never been to school. It also providesworrying evidence of the limited progress achievedin reaching sub-Saharan Africa’s poorest 20% ofchildren, especially young girls. The share of youngadults aged 17 to 22 from the poorest householdswho never attended school was higher in Southand West Asia than in sub-Saharan Africa. Thatpicture is dramatically reversed for children aged 7-16 years, suggesting that social convergencein school attendance is moving more slowly in sub-Saharan Africa.

Being poor and female carriesa doubledisadvantage inmany countries

0

10

20

30

40

50

60

70

80

90

100

Iraq,

200

5

Mya

nmar

, 200

0

U. R

. Tan

zani

a, 2

004

Cam

bodi

a, 2

005

Bang

lade

sh, 2

004

Nepa

l, 20

06

Pakis

tan,

200

6

Guat

emal

a, 1

999

Liber

ia, 2

007

Buru

ndi,

2005

Mor

occo

, 200

3

Gam

bia,

200

5

Rwan

da, 2

005

Beni

n, 2

006

Mad

agas

car,

2004

Côte

d’Iv

oire

, 200

4

Guin

ea-B

issau

, 200

5

Sier

ra Le

one,

200

5

Guin

ea, 2

005

Moz

ambi

que,

200

3

Sene

gal,

2005

Ethi

opia

, 200

5

Som

alia

, 200

5

Chad

, 200

4

Mal

i, 20

01

Burk

ina

Faso

, 200

3

Nige

r, 20

06

C. A

. R.,

2000

Average number of years of education: fewer than 6 yearsSh

are

of th

e po

pula

tion

with

few

er th

an 4

and

few

er th

an 2

year

s of e

duca

tion

(%)

The gender effect:Girls from the poorest households

who are in education poverty

The wealth effect:People from the

poorest householdswho are in

education poverty

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 4 2

Inequalities stemming from income and genderhelp explain the inconsistent relationship betweennational wealth and acute deprivation in education.Average years of education tend to rise anddeprivation to diminish as income rises, butcountries vary enormously in the degree to whichthey convert rising income into declining educationdeprivation (Figure 3.3). Comparisons acrosscountries at different levels of income revealsome striking results for those aged 17 to 22:

While it has a per capita income comparable toViet Nam’s, Pakistan has more than three timesthe share of the age group with fewer than fouryears of education.

With double the average income level of Lesotho,Morocco has twice the population share withfewer than four years of education.

At the same average income level as Egypt,Jordan has an incidence of education povertyseven times lower.

Average income in Gabon and Turkey is morethan double the level in the Dominican Republic,but all three countries have comparablepopulation shares below the four-year threshold.

Such comparisons caution against assumingthat economic growth automatically dissolvesextreme deprivation in education. Wealthincreases the resources available to householdsand governments for investment in education.Yet the high levels of variation point to theimportance of other factors in expandingopportunity for the disadvantaged – notably,the effectiveness of public policies.

Income and gender disparities do not operate inisolation. Education inequalities in both dimensionsintersect with inequalities linked to location,ethnicity, language, disability and other factorsto limit opportunity and reinforce marginalization.

In many countries, rural households in generaland poor rural households in particular lag farbehind their urban counterparts. Rural locationcompounds wealth and gender disadvantages,reflecting the impact of cultural attitudes andthe unequal burden of household labour. It alsointersects with the wider patterns of group-baseddeprivation captured in Figure 3.4:

In Egypt, income differences overlap with rural-urban and gender divides. Rich urban boys and girls both average just over ten years in education. Poor rural males average fewer than eight years,declining to under five years for girls. The ruralpart of Upper Egypt is an area of particularlydeep disadvantage. Over 40% of the populationlives in poverty and rural females in the regionaverage just over four years of schooling – a levelsimilar to the national average in Côte d’Ivoire.

India’s wealth divides in education are among thelargest in the world – and they are reinforced byregional and gender disparities. While the richest20% average over eleven years in school, thepoorest have an average education expectancythat places them just above the four year‘education poverty line’. Poor rural females arewell below that line. Averaging three years ineducation, they are in a position comparable tothe national average for Chad. The average poorrural woman aged 17 to 22 in Bihar averagesfewer than two years in education.

In Nigeria, the average poor rural female isjust above the two-year threshold for extremeeducation deprivation, with less than 40% thenational average for years of school and aroundone-quarter the average for rich urban males.There is a three-year gap between poor rural

In Egypt, rich urban girls

averageten years in

education,declining to

under five yearsfor poor

rural girls

0

10

20

30

40

50

60

70

80

SWA richest 20%

SWA poorest 20%

SSA poorest 20% girls

SSA poorest 20%

SWA poorest 20% girls

SSA richest 20%

% o

f the

pop

ulat

ion

that

has

nev

er b

een

to sc

hool

23–27 17–22 7–16

Age group

Figure 3.2: Slow progress for Africa’s poorest children% of the population that has never attended school, by age group, sub-Saharan Africa, and South and West Asia, circa 2005

Notes: SSA stands for sub-Saharan Africa, SWA for South and West Asia. Estimates are population weighted averages.Source: UNESCO-DME (2009).

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 4 3

females and poor urban males. Poor ruralHausa women are identifiably at the bottomend of the national distribution for opportunitiesin education, averaging just a few months ofschooling. At the other end of the scale, richboys and girls average around 10 years ineducation. The Nigerian case powerfullyillustrates the mutually reinforcing effectsof poverty, rural location and cultural factorsin creating extreme disadvantage.

Inequalities associated with specific livelihoodsoften contribute to national disparities. Theexperience of pastoralists is a particularly starkexample. Living in remote areas, with childrenheavily involved in tending cattle and livelihoodsthat involve movement across large distances,pastoralists face major barriers to educationalopportunity. Those barriers of time and distanceare sometimes reinforced by problems ineducation policy, including failure to offer relevantcurricula, provide appropriate textbooks andrespond to the realities of pastoral livelihoods.

And they interact with labour practices, culturaltraditions and belief systems to perpetuate deepdisparities based on gender.

National household survey and census dataprovide insight into the scale of this disadvantage.In Ethiopia, Kenya and Uganda, for example,pastoralist groups are at the bottom end of thedistribution for educational opportunity (Figure 3.5).In Uganda, 85% of Karamojong pastoralists aged17 to 22 have fewer than two years in school,compared with a national average of over six years.In West Africa, the Peul group, also called theFula, Fulani and Poular, is among the mosteducationally disadvantaged in countries includingBenin, Chad, Mali and Senegal.

Current school attendance patterns point to acontinuation of extreme educational disadvantageacross generations, with pastoralist childrenparticularly unlikely to be attending school, asFigure 3.5 shows. In Benin, nearly 90% of Peulchildren of primary school age do not attend

In Ethiopia, Kenya and Uganda,pastoralist groupsare at the bottomend of thedistribution for educationalopportunity

0 10 20 30 40 50 60 70 80 90 1000

2 000

4 000

6 000

8 000

10 000

12 000

14 000

GNP

per c

apita

(PPP

US$

)

Education poverty: population with fewer than four years of education (%)

Burkina Faso

C. A. R.NigerMali

Senegal

Morocco

Gabon

Turkey

TFYR Macedonia

Colombia

Suriname

Dominican Rep.

EgyptJordan

Côte d’Ivoire

Gambia

Guinea-BissauLiberia

HondurasMongolia

D. R. Congo

Bolivia

Kyrgyzstan

TajikistanUganda

Viet Nam

Lesotho

Pakistan

Figure 3.3: Education poverty falls with rising income — but the association variesGNP per capita and % of the population aged 17 to 22 with fewer than four years of education

Sources: UNESCO-DME (2009); annex, Statistical Table 1.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 4 4

MaleFemale

0

2

4

6

8

10

12

14

Aver

age

year

s of e

duca

tion

Ukraine

Cuba

Bolivia

IndonesiaEgypt

Honduras

Cameroon

Bangladesh

Chad

C. A. R.

Richest 20%

Urban

Urban

Rural

Rural

Poorest 20%

Poor rural UpperEgypt male

Poor rural UpperEgypt female

Rural Upper Egypt

Education poverty

Extreme education poverty

Egypt, 2005

0

2

4

6

8

10

12

14

Aver

age

year

s of e

duca

tion

Ukraine

Cuba

Bolivia

Indonesia

IndiaHonduras

Cameroon

Bangladesh

Chad

C. A. R.

Richest 20% Urban

Urban

Rural

Rural

Poorest 20%

Poor Bihar male

Poor Bihar female

Bihar

MaleFemale

India, 2005

Figure 3.4: The education inequality treeAverage number of years of education of the population aged 17 to 22 by wealth, gender, location, and other selected drivers of marginalization, latest available year

Source: UNESCO-DME (2009).

Country average

Pastoralist group

9

7

17

96

92

51

53

26

60

39

88

85

57

7143

17 68

80

33 91

94

85

40

17

53

87

60 90

9

9

62

51

61 83

41

38

30

50

84

42

35

79

87

78

Ethiopia:Afar/related

m

f

m

f

m

f

m

f

m

f

Kenya:Somali

Uganda: Karamojong

Benin:Peul/related

Nigeria:Fulani

Senegal:Poular

Extreme education poverty: population aged17 to 22 with fewer than 2 years of education (%)

% out of primary school

0 20 40 60 80 100 0 20 40 60 80 100

Figure 3.5: Pastoralists face extreme education deprivation% of the population aged 17 to 22 with fewer than two years of education and % of primary school age children not attending primary school, by gender and membership of selected pastoralist groups, latest available year

Notes: Gender-disaggregated data are not available for Uganda. % out of primary school: proportion of children of primary school age not attending primary school.Sources: UNESCO-DME (2009);census, calculations by Harttgen and Klasen (2009).

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 4 5

school. Being a Somali in Kenya increases therisk of being out of school by a factor of five ormore, depending on gender. Over 60% of Somaligirls are not in primary school – some seventimes the national average.

Regional disparities figure prominently in theprofile of educational disadvantage. Inequalitiesbetween regions in the same country are oftenfar larger than inequalities between countries.Figure 3.6, which charts the share of the nationalpopulation aged 17 to 22 with fewer than fouryears of school across regions of selectedcountries, shows that regional differences havea strong influence on educational opportunities.In Nicaragua, the share of the population withfewer than four years of school ranges from lessthan 7% in Managua to almost 60% in Jinotega.

Marginalized regions are often characterized byhigh levels of poverty, concentrations of ethnic

minority populations and conflict. In Chad’seastern Barh Azoum district, fighting betweengovernment and rebel forces has led to large-scaleinternal displacement. The area is also home toa large population of refugees from the Sudandisplaced by Janjaweed militias (InternalDisplacement Monitoring Centre, 2009). Over 90%of the district’s population aged 17 to 22 has fewerthan four years of education and school attendancerates are among the country’s lowest. In Uganda,strong national progress towards universalprimary education has obscured large pocketsof regional marginalization. Education data starklyreveal the devastating impact of conflict andpoverty in the north of the country. In the north-eastern districts of Kotido, Moroto andNakapiripirit, where security concerns andviolence linked to cattle raiding have contributedto wider factors holding back progress ineducation, around 90% of those aged 17 to 22have fewer than two years of schooling (Box 3.1).

Geographic inequalities are often closely linkedto social and economic inequalities, rural-urbandifferences, ethnicity and language. In Cambodia’smost disadvantaged provinces, Mondol Kiri andRattanak Kiri, large concentrations of hill tribeslive in remote areas with high levels of poverty.Fewer than one in three residents aged 17 to22 have more than four years of education(Figure 3.6). Gender disparities in the area aremarked: young women average just 1.8 years ofschool, compared with 3.2 years for young men.These outcomes reflect the combined effectsof poverty, isolation, discrimination and culturalpractices, as well as policy failures in education.

In the Philippines, there is a close fit betweenthe regional incidence of poverty and the regionalincidence of young adults aged 17 to 22 withfewer than four years of education. One of themost educationally disadvantaged areas is theAutonomous Region in Muslim Mindanao, whereyears of conflict have exacerbated poverty anddisplaced 750,000 people (Box 3.2).

Another example comes from Mexico, where rapidprogress has been made over the past decade,with social protection programmes and targetedtransfers eroding regional and income-basedinequalities. While regional disparities have fallenover time, they nevertheless remain (Table 3.2):

The southern ‘poverty belt states’ of Chiapas,Guerrero, Michoacán, Oaxaca and Veracruz

In Cambodia’smostdisadvantagedprovinces, youngwomen averagejust 1.8 years ofschool, comparedwith 3.2 years for young men

0

2

4

6

8

10

12

14

Aver

age

year

s of e

duca

tion

Ukraine

Cuba

Bolivia

Indonesia

Nigeria

Honduras

Cameroon

Bangladesh

Chad

C. A. R.

Richest 20%Rural

Urban

Rural

Urban

Poorest 20%

Poor rural Hausa male

Poor rural Hausa femaleRural Hausa

MaleFemale

Nigeria, 2003

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 4 6

figure prominently at the bottom end of theeducational opportunity scale. Average yearsof education range from 5.7 for females inChiapas to over 10 in the Federal District.

Whereas 11% of those aged 17 to 22 havefewer than four years of education, for Guerrerothe figure rises to 19% and for Chiapas 26%.

Indigenous people and ethnic minorities faceparticularly severe disadvantages in education.Some disadvantages faced by indigenous groupsand ethnic minorities are poverty-related. VietNam’s more than fifty ethnic minority groupsaccount for 13% of the population but 40% ofpeople living below the poverty line (Truong Huyen,2009). In Bolivia and Guatemala, almost three-quarters of indigenous people are poor, comparedwith half of the non-indigenous population (Halland Patrinos, 2006). Higher levels of poverty are

associated in turn with discrimination andcultural stigmatization, creating obstacles toeducation. In Bolivia, Aymara speakers aged 17 to22 accumulate two years fewer in school than doSpanish speakers and for Quechua speakers thefigure is four years. In Guatemala, average yearsin school range from 6.7 for Spanish speakersto 1.8 for Q’eqchi’ speakers.

Poverty and gender discrimination exacerbateeducation deprivation among indigenousminorities. From Guatemala and Peru to Cambodiaand the Lao People’s Democratic Republic,indigenous young adults are far more likelythan the non-indigenous to experience extremeeducation deprivation, especially if they are poorand female. An indigenous person aged 17 to 22in Peru has two years less education than thenational average; poor indigenous girls are twoyears further still down the scale (Figure 3.10).

In Guatemala, the average

number of yearsin school ranges

from about 6.7for Spanish

speakers to 1.8for Q’eqchi’

speakers

0

20

40

60

80

100

Peru Viet Nam Nicaragua Uganda U. R. Tanzania Cambodia Pakistan Benin Chad Burkina Faso

Educ

atio

n po

verty

: pop

ulat

ion

with

few

er th

an 4

year

s of e

duca

tion

(%)

Lima Red River Delta

ManaguaKampala Zanzibar South

Phnom Penh

Punjab

Littoral

N’Djamena Ouagadougou

Jinotega

Moroto

Tabora

Mondol Kiriand Rattanak Kiri

Balochistan

Alibori

Barh Azoum

EstAdministrative region

Country average

Huanuco MekongRiver Delta

Figure 3.6: Many countries have large regional disparities in education poverty% of population aged 17 to 22 with fewer than four years of education, by region, selected countries, latest available year

Source: UNESCO-DME (2009).

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 4 7

Uganda has made rapid advances in primary educationover the past decade. Numbers of out-of-school childrenhave fallen sharply, completion rates are improving andgender disparities are shrinking. Sustaining progresstowards universal primary completion will require arenewed effort to reach some of the most marginalizedpopulations. Census and household survey data helpidentify these populations.

Poverty remains a major barrier. Over 20% of 17- to 22-year-olds in the poorest quintile of the populationhave fewer than two years of schooling — four timesthe level for the richest quintile. Increased investmentin education and the abolition of school fees haveimproved access for the poor. Even so, 16% of those aged7 to 16 from the poorest households are not attendingschool, pointing to a need for further measures.

Parts of Uganda have been left far behind. Conflictand the activities of the Lord’s Resistance Army in thenorthern districts of Acholi, Apac, Gulu, Kitgum andLira have had devastating consequences for education.School closures, parental fears over abduction and chronicteacher shortages have held back progress. Insecurityhas undermined livelihoods and reinforced poverty,making it difficult for parents to meet indirect educationcosts. Some 40% of Acholi parents cite cost as the reasonfor their children dropping out of school, although inabilityto meet costs and insecurity are mutually reinforcing.

Other northern districts with large pastoralist populationsare among the most educationally marginalized in thecountry. In Kotido, 83% of 17- to 22-year-olds havefewer than two years of education — and only one-fifthof children are currently in primary school (Figure 3.7and Table 3.1).

Gender disparities are another impediment to progressin the north. Traditional practices often lead to girls asyoung as 12 being married. Early pregnancy is anotherproblem. One survey found that almost 10% of school

dropout in the Acholi subregion was linked to pregnancyor early marriage. Fears over the safety of girls attendingschools in conflict-affected areas added to these concerns.And where poverty forces households to choose whogoes to school, cultural attitudes lead many to expressa preference for boys’ education.

Conflict has made it more difficult to attract teachersto the north. For example, in late 2006, 500 teachingpositions were advertised in Kitgum, but only 180 viableapplications were received. High rates of teacherabsenteeism reflect underlying problems. Many schoolslack teacher housing, so teachers have to commute longdistances, sometimes along insecure routes. Teacherincome also tends to be far lower than in more prosperousareas, partly because poverty reduces the supplementshouseholds pay.

The fragile peace in the north gives the government anddonors an opportunity to support an ‘education catch-up’.Seizing the opportunity may require a review of publicfinancing. Mapping of educational disadvantage highlightsthe special needs of the north, but on a per capita basisthe area receives roughly the same in governmenttransfers as the rest of the country. There is a strong casefor preferential financing for this disadvantaged area.

Sources: UNICEF (2007d); Higgins (2009); Women’s Commission for Refugee Women and Children (2005).

Box 3.1: Uganda — universal primary education is in sight, but large pockets of marginalization persist

0 20 40 60 80 100

Extreme education poverty:population with fewer than 2 years of education (%)

NationalCentral region

KotidoNakapiripirit

Moroto

Selected northern districts

Figure 3.7: Education poverty is high in some of Uganda’s northern districts% population aged 17 to 22 with fewer than two years of education

83 83

89 90

79 7679 7783 7982 7684 8222 1978 7587 8221 1817 1980 7282 7888 81

National

Central region

Northern regionAdjumaniApacAruaGuluKitgumKotidoLiraMoyoMorotoNakapiripiritNebbiPaderYumbe

Table 3.1: Primary net attendance rates in selected regions and districts ofUganda, by gender, 2002

Male Female

Primary net attendancerates (%)

Source: Census, calculations by Harttgen and Klasen (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 4 8

Education indicators for the Philippines are below what might beexpected for a country at its income level. There is a real dangerthat the country will fail to achieve universal primary educationby 2015. Household survey data help identify the large pockets ofextreme and persistent deprivation that are holding back progress.

The net enrolment ratio was 92% in 2007, which is comparablewith countries at far lower levels of average income, such asZambia, and below the levels attained by other countries in theregion, such as Indonesia. Around 1 million children are out ofschool — a slight increase over the level in 1999.

Extreme poverty and regional disparities are at the heart of the mismatch between national wealth and education outcome.The gap separating the poorest 20% from the rest of society is far wider than in most countries in the region (Figure 3.8).Those aged 17 to 22 in the poorest quintile average about sevenyears of education — more than four years fewer than in thewealthiest 20%. Data on school attendance provide evidencethat current policies are not reaching the poorest. Around 6% of 7- to 16-year-olds from the poorest households are reported as not attending school or to have ever attended. Extremeeconomic inequalities fuel education inequalities, notably bypushing many children out of school and into employment.

Regional data reveal deep fault lines in opportunity (Figure 3.9).Nationally, about 6% of those aged 17 to 22 have fewer than fouryears of education. In the best-performing regions — Ilocos andthe National Capital Region — the share falls to 1% to 2%. At theother extreme, in the Autonomous Region in Muslim Mindanaoand Zamboanga Peninsula over 10% fall below this threshold.

The disparities are driven by a wide array of factors. The impactof high levels of poverty is exacerbated by conflict in Mindanao,and by the remoteness and wider disadvantage experienced by indigenous people in the Eastern Visayas and Zamboanga.

National authorities face difficult policy choices if the Philippinesis to achieve universal primary education by 2015. Far moreweight has to be attached to reaching marginalized populations and providing them with good quality education. Social protection and conditional cash transfer programmes, such as those in Braziland Mexico, could play a vital role in combating child labour andextending educational opportunities to the poor. Another urgentpriority is local language teaching in indigenous areas.

Box 3.2: The Philippines — leaving the marginalized behind

0

2

4

6

8

10

12

Year

s of e

duca

tion

Poorest Secondpoorest

Middle Secondrichest

Richest

Wealth quintiles

Figure 3.8: The Philippines has large wealth gaps in educationAverage number of years of education of the population aged 17 to 22,Philippines, 2003

Source: UNESCO-DME (2009).

65 54 6.6 5.7 2668 64 7.1 7.0 1959 57 7.1 7.4 1469 61 7.1 6.7 1668 65 7.3 7.1 20

69 66 8.2 8.2 11

75 74 8.9 9.4 684 84 10.0 10.1 374 72 9.0 9.0 6

ChiapasGuerreroMichoacánOaxacaVeracruz

National average

Baja CaliforniaDistrito FederalMéxico

Table 3.2: Selected education indicators, by region, Mexico, 2005

Secondary net attendancerates (%)

Male Female (%)*

Years of education*

Fewer than 4 years ofeducation

* Data for population aged 17 to 22.Source: Census, calculations by Harttgen and Klasen (2009).

Male Female

Disadvantaged southern states

Selected northern and central states

Large pockets of extreme

and persistentdeprivation are holding

back progress

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 4 9

School attendance patterns revealed in householdsurveys point to the prospect of marked disparitiesbeing transmitted across generations.

Disadvantages associated with language arefound across all regions. Having the officiallanguage of instruction as a home languagesignificantly lowers the risk of having fewer thanfour years in education at age 17 to 22. HavingKurdish as a home language in Turkey carriesa 30% risk of having fewer than four years ofschooling compared with less than 5% forTurkish speakers. While these language effectsare strongly associated with regional povertydifferences, they are also important in theirown right (Figure 3.11).

In countries where the official language is not themost common language spoken at home thereare strong links from language to marginalizationin education. There are some thirty countries of

The diversity of the challenges sets limits to what the centralgovernment can do. Regional and subregional authorities needto develop and implement policies that respond to local needs.However, the central government could do more to createan enabling environment. The education system suffers

from chronic shortages of teachers and classrooms, rising class sizes and low levels of learning achievement.Addressing these problems will require an increase in the 2.1% share of national income directed towards education in 2005 — one of the lowest levels in the world.

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%0%

5%

10%

15%

Mimaropa

Income poverty

Educ

ation

pove

rty

Cordillera Admin. Region

BicolCaraga

Central Visayas

Northern MindanaoEastern Visayas

Soccsksargen

Western Visayas

Cagayan ValleyNational Capital Region

Central Luzon

CalabarzonIlocos

Zamboanga Peninsula

Autonomous Region inMuslim Mindanao

Davao

Figure 3.9: Children in poor, remote, or conflict-affected regions of the Philippines suffer higher levels of education poverty% of the population aged 17 to 22 with fewer than four years of education and prevalence of poor families by region, Philippines, 2003

Notes: Education poverty is measured as the proportion of 17- to 22-year-olds with fewer than four years of education. Income poverty rate is the proportion of families whose income puts them below the poverty line for each region.Sources: UNESCO-DME (2009); Philippines National Statistical Coordination Board (2006).

0

2

4

6

8

10

Aver

age

year

s of e

duca

tion

Bolivia Peru Colombia Guatemala

Country average

Indigenous female

Poor, indigenous female

Indigenous

Figure 3.10: Wealth and gender widen indigenous education disparities in Latin AmericaAverage number of years of education for indigenous people aged 17 to 22, selected countries,latest available year

Notes: The indigenous average is the weightedaverage for the indigenous groups for whichdata were available. These were: Bolivia(Aymara, Guarani and Quechua ethnicity);Guatemala (Chorti, Kanjobal, Kaqchiquel,K’iche’, Mam, Poqomchi’, Q’eqchi’ and Tzu’Utihillanguage); Peru (Aymara and Quechua ethnicity).For Colombia, the ‘indigenous ethnicity’ censuscategory was used.Sources: UNESCO-DME (2009); census,calculations by Harttgen and Klasen (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 5 0

sub-Saharan Africa where the official languageis not the most common home language (Alidouet al., 2006). This means that in many caseschildren are taught at primary school in alanguage other than their mother tongue,which contributes to extreme language-baseddisparities. In Mozambique, speakers of Jaua aged17 to 22 average one year in education comparedwith five years for speakers of Portuguese; over80% have fewer than four years in education. InNigeria, education poverty levels, defined by thefour-year threshold, range from less than 10% forYoruba speakers to over 60% for Hausa speakers.Across the region, home language has a strongbearing on prospects for getting more than fouryears of education.

The interaction between language, ethnicity andlocation is a potent source of marginalization ineducation. Household survey data can help identify

the regions and individuals most severely affected.One striking illustration comes from Turkey. Inmost regions, 2% to 7% of those aged 17 to 22 havefewer than four years of education, but in theeastern region the figure rises to 21%. Youngwomen speaking a non-Turkish home language –predominantly Kurdish – are among the mosteducationally marginalized. They average just threeyears of education – less than the national averagefor Senegal (Figure 3.12).

The ‘bottom 20%’: relative deprivationMarginalization is not just about deprivation inabsolute terms. It is also about falling behind therest of society. The individual and group-baseddisadvantages discussed above figure prominentlyin explaining the profile of those left behind ineducation. This section looks at the characteristicsof the ‘bottom 20%’ in education.

In Nigeria,education

poverty levelsrange from less

than 10% forYoruba speakersto over 60% forHausa speakers

0

20

40

60

80

100

Peul

Turkish

Jaua

Chigorogonza

Somali

Hausa

Portuguese

Shona

Urdu

NepaliArabic

Kurdish

Luganda

BembaKiswahili

Masai

Mijikenda Ateso-KaramojongNyanja

SaraikiBaluchi

Sindhi

Punjabi

MaithiliBhojpuri

YorubaIgbo

Luvale

Bambara

Educ

atio

n po

verty

: pop

ulat

ion

with

few

er th

an 4

year

s of e

duca

tion

(%)

Turkey Kenya Uganda* Zambia Nigeria Nepal Pakistan Mozambique Mali

Most spoken language

Language group

Country average

Official language

Figure 3.11: The language gap in educational opportunity% of population aged 17 to 22 with fewer than four years of education, by language spoken, selected countries, latest available year

Sources: UNESCO-DME (2009); *Uganda census, calculations by Harttgen and Klasen (2009).

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 5 1

Who are the bottom 20%? Household survey datamake it possible to group people aged 17 to 22on the basis of accumulated years of school. Dataanalysis can also be used to decompose groupmembership by identifying social characteristicssuch as household wealth, gender, ethnicity andlocation. Unlike the thresholds of deprivation usedin the previous section, the ‘bottom 20%’ providesa relative national scale. People at the lowest endof the distribution in, say, the Philippines or Turkeyhave more years of school than their counterpartsin Chad or Mali. What they share is the experiencein childhood of restricted opportunity relative toother members in their country.

Household surveys have been widely used tochart overall inequality in education. The newdata analysis prepared for this Report makes itpossible to look beyond overall inequality to thecharacteristics of the ‘bottom 20%’. The data canbe used to assess both the weight of discretevariables such as income, language and genderand – with limitations – the cumulative effectsof these variables.

Household wealth. Being born into the poorest 20%of households in a country is strongly associatedwith heightened risk of being at the bottom endof the distribution for educational opportunity(Figure 3.13). In Colombia, Mongolia, Nicaragua,the Philippines and Viet Nam, the poorest 20%account for twice their population share in thebottom 20% of the education distribution.

Ethnicity and language. In some countries, ethnicand language minority groups account for a largeshare of the bottom 20% (Figure 3.14). In Nigeria,over half the ‘education poor’ are Hausa speakers –a group that makes up one-fifth of the population.Reflecting the legacy of disadvantage experiencedby indigenous Q’eqchi’ speakers in Guatemala,membership of this language group more thandoubles the risk of being in the bottom 20% foryears in school.

Region and location. Regional differences inyears spent in education are often far larger thandifferences between countries (Figure 3.15).Areas such as northern Kenya, eastern Turkey,rural Upper Egypt and northernmost Cameroonare heavily overrepresented in the lowest 20%of the education distribution for their countries.Single region figures can understate the level ofdisadvantage. In Cameroon, three regions withjust one-quarter of the overall population account

Venezuela, B. R.

Madagascar

India

Mongolia

Viet Nam

Nicaragua

Bolivia

Colombia

Philippines

Nigeria

Pakistan

Ghana

Jordan

0% 20% 40% 60% 80% 100%

Poorest 2nd poorest Middle 2nd richest Richest

In countries such as India,Madagascar and theBolivarian Republic of Venezuela, the poorest fifthchildren make up morethan half of the bottom20% by years in school.

In many countries,the poorest two-fifths areheavily over-representedin the bottom 20%by years in school.

Com

posit

ion

of ‘b

otto

m 2

0%’

Figure 3.13: The poorest households are more likely to be left behind in educationDecomposition of the bottom 20% of the education distribution by wealth quintile, selected countries,latest available year

Note: The ‘bottom 20%’ is the 20% of 17- to 22-year-olds with the fewest years of education.Source: UNESCO-DME (2009).

0

2

4

6

8

10

12

14

Aver

age

year

s of e

duca

tion

Ukraine

Cuba

Bolivia

Indonesia

TurkeyHonduras

Cameroon

Bangladesh

Chad

C. A. R.

Richest 20% Urban

Urban

Rural

Rural

Poorest 20%Poor Kurdish male

Poor Kurdish female

Kurdish

MaleFemale

Education poverty

Extreme education poverty

Figure 3.12: Poverty, ethnicity and language fuel education marginalization in TurkeyAverage number of years of education of the population aged 17 to 22 by wealth, location,gender and Kurdish language, 2005

Source: UNESCO-DME (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 5 2

The disadvantages that drive people into thebottom 20% in education do not operate in isolation.They intersect and magnify the wider socialinequalities that restrict opportunities in education.This is illustrated in Figure 3.16, which uses DMEstatistics to look at the impact of two or threeoverlapping dimensions of deprivation. The impactof clustered disadvantage is evident from thecombined effects of poverty, gender and othermarkers for disadvantage. These effects can becaptured by reference to the ‘extreme educationpoverty’ benchmark of fewer than two years ineducation and the more recent disadvantagesreflected in the school attendance rates forprimary school age children:

Being a rural girl in the Cambodian hill provincesof Mondol Kiri and Rattanak Kiri increases therisk of not being in school by a factor of five.Three-quarters of the group have fewer thantwo years in school, compared with a nationalaverage of 12%.

In Guatemala, girls from poor households ofIndian ethnicity have primary net attendancerates of 60% compared with a national averageof 82% and they are over three times more likelyto have fewer than two years in school.

In Turkey, one of the most marginalized groupsis Kurdish-speaking girls from the pooresthouseholds. Around 43% at ages 17 to 22 havefewer than two years of education, while thenational average is 6%.

In Nigeria, poor Hausa girls face some of theworld’s most severe education deprivation.Some 97% of 17- to 22-year-olds have fewer thantwo years of education and just 12% of primaryschool age Hausa girls attend primary school.

Each of these examples involves a relatively largepopulation group. They represent a statisticallysignificant national policy challenge. But combatingmarginalization is also about identifying smallgroups facing intensive deprivation. Figure 3.17uses the DME data set to illustrate the high levelsof marginalization experienced by a number ofsmall population groups. To take one case in point,almost 90% of the Mushahar community in Nepal,a largely landless low-caste group, is in the bottom20%. The average time spent in school for thoseaged 17 to 22 in this group is less than threemonths, and only 29% of girls and 41% of boysattend primary school. Similarly, in Viet Nam

Cambodia: Mondol Kiri and Rattanak KiriKenya: North-eastern

Guatemala: north-westernUganda: north

Ghana: Upper eastAzerbaijan: Aran

Swaziland: LubomboSierra Leone: East

Congo: SouthZambia: Eastern

C. A. R.: Mambéré-KadeïMongolia: Khangai

Gambia: Lower riverGuinea-Bissau: East

Liberia: north-centralGuinea-Bissau: north

Burundi: NorthViet Nam: Mekong River Delta

Turkey: EastEgypt: rural Upper

Ghana: NorthernNigeria: north-west

Cameroon: Extreme North

0% 20% 40% 60%

Proportionin population

Proportion inbottom 20%

Figure 3.15: Some regions face acute education deprivation% of selected regions* in the bottom 20% of the education distribution, population aged 17 to 22, selected countries, latest available year

* Regions presented in the graph are the first level of administrative division, except those in italics which are geographical areas.Sources: UNESCO-DME (2009); census, calculations by Harttgen and Klasen (2009).

*

Gambia: Pulaar

Guinea-Bissau: Balante

Pakistan: Saraiki

Guatemala: Q’eqchi’

Mexico: indigenous

Nepal: Maithili

Turkey: Kurdish

Nigeria: Hausa

0% 20% 40% 60%

Proportionin population

Proportion inbottom 20%

Figure 3.14: Language often predicts risk of being in the bottom 20%% of selected language groups in the bottom 20% of the educationdistribution, selected countries, latest available year

Note: The ‘bottom 20%’ is the 20% of 17- to 22-year-olds with the fewest years of education.* The indigenous language category in Mexico consists of those who speak indigenouslanguages only and do not speak Spanish.Sources: UNESCO-DME (2009); Mexico census, calculations by Harttgen and Klasen (2009).

for three-quarters of the population in the lowesteducation quintile. In Nigeria, 86% of the lowesteducation quintile is in two regions – the north-westand north-east, which account for 43% of thepopulation.

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 5 3

In Nigeria, 97% of poorHausa girls have fewer thantwo years ofeducation

0

1

2

3

4

5

6

7

8

9

Mushaharethnicity

Nepal

1 5 7 4 1 3 1 2 3

89 43 32 43 25 45 88 42 77

Barh Azoumregion

Chad

Gourmantchéethnicity

BurkinaFaso

Peul/relatedethnicity

Benin

Gourmantchéethnicity

Niger

Somaliethnicity

Ethiopia

Hmongethnicity

Viet Nam

Indigenouslanguage only

Mexico

North-easternregion

Kenya

Aver

age

year

s of e

duca

tion

Share of population (%)

Proportion who are inthe ‘bottom 20%’ (%)

Country

Group

Figure 3.17: Small groups, big disadvantagesAverage number of years of education for selected marginalized groups, population aged 17 to 22 selected countries, latest available year

Sources: UNESCO-DME (2009); census, calculations by Harttgen and Klasen (2009).

Nigeria, poor, Hausa, girls

Kenya, rural, Somali, girls

Ghana, northern region, rural, girls

Pakistan, rural, Sindhi, girls

Turkey, poor, Kurdish language, girls

Guatemala, poor, Indian ethnicity, girls

India, poor, Uttar Pradesh, girls

Cambodia, rural, Mondol Kiriand Rattanak Kiri, girls

Group

Country

0 20 40 60 80 100

6 43

7331

17

8 96

9725

12 74

84

6619

20 57

Extreme education poverty: population withfewer than 2 years of education (%)

0 20 40 60 80 100

58 88

6841

32

30 91

6612

15 77

61

8260

67 76

Primary net attendance rates (%)

Figure 3.16: Overlapping disadvantages erode education opportunitiesPrimary net attendance rates and % of the population aged 17 to 22 with fewer than two years of education, selected countries, latest available year

Source: UNESCO-DME (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 5 4

nine out of ten Hmong, members of an ethnicminority group living in northern highland regions,are in the bottom 20% of the national distributionfor years in school.

The quality deficitMarginalized individuals and groups do not justaccumulate fewer years of education. When they arein school they often receive a poor-quality education,leading in turn to low levels of learning achievement.

Many of the world’s poorest countries have beenmore successful in expanding access than raisingquality. As Chapter 2 shows, average learningachievement is often shockingly low even forchildren who complete a full primary educationcycle. The achievement deficit is widely spreadacross the population, but is typically concentratedamong individuals and groups facing widerdisadvantages in access to education.

Factors such as household wealth, parentaleducation and home language exercise a pervasiveinfluence on learning achievement. That influencehas been extensively documented in developedcountries but less widely explored in the world’spoorest countries. Research carried out for thisReport examined data on learning achievementcollected for sub-Saharan Africa, through thePASEC and SACMEQ regional assessmentprogrammes, to identify characteristics associatedwith students performing at the top, middle andbottom of the test score range. The results arestriking. As early as grades 5 and 6, there is astrong association in many countries betweenwealth and test scores. In Kenya and Zambia, theaverage household of children scoring in the top10% has twice as many consumer durables as theaverage household for children in the lowest 10%.Parental literacy is also strongly associated withtest scores (Fehrler and Michaelowa, 2009).

In Latin America, too, assessments reveal the lowachievement of students belonging to marginalizedpopulations. The PISA assessment programmeuses a composite set of indicators to construct asocio-economic background index for parents of 15-year-olds tested. The results point to a strongassociation between parental socio-economicstatus and learning outcomes. In Brazil, Mexicoand Uruguay, children of parents in the top quartileachieved a mathematics score 25% to 30% higherthan those in the poorest quartile (Vegas andPetrow, 2008). In a national assessment in Uruguay,only 36% of sixth-graders from ‘very unfavourable’

backgrounds passed the mathematics test and 55%the language test, as opposed to 72% and 87%,respectively, of those from ‘favourable’ backgrounds(Vegas and Petrow, 2008).

Education outcomes are often substantially worsefor indigenous people and ethnic minorities. In LatinAmerica, there is extensive evidence of test scoregaps between indigenous and non-indigenouschildren. In Guatemala, indigenous children in bothrural and urban areas scored between 0.8 and 1standard deviation below non-indigenous childrenin grades 3 and 6 Spanish tests – a gap of around17% (McEwan and Trowbridge, 2007). Differences inmathematics tests were smaller but still significant.Recent research from Peru recorded exceptionallylarge gaps in indigenous and non-indigenouslearning achievement (Cueto et al., 2009). At theend of primary school, the gap in mathematics andlanguage scores was above a full standard deviation(1.22 and 1.07, respectively).

Home language often has an important influenceon test scores. Research using data from the 2007TIMSS assessment identifies a strong associationbetween students performing below the lowestinternational benchmark and the frequency withwhich the language of the test is spoken at home.In Turkey, grade 8 students who report ‘always oralmost always’ speaking the test language at homeare 30% less likely to score below the internationalmathematics benchmark than those who reportspeaking it ‘sometimes or never’ (Altinok, 2009).Evidence from PASEC and SACMEQ also pointsto a strong link between home language and thelanguage of instruction in influencing test scores(Fehrler and Michaelowa, 2009).

Language, ethnicity and regional factors can combineto produce complex patterns of disadvantage. InViet Nam, a large-scale survey of grade 5 studentsin 2001 found strong disparities in achievementamong provinces, with school location and students’ socio-economic background and ethnicity also having a strong influence (World Bank, 2004). Ethnic minority students who spoke no Vietnamese at home weremuch less likely to read ‘independently’ thanstudents whose home language was Vietnamese.

Marginalization in rich countries

Education is an increasingly important engineof social and economic success in rich countries.While education can break the transmission ofcycles of disadvantage across generations, it can

Householdwealth, parental

education andhome language

exercise a pervasive

influence on learning

achievement

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 5 5

also reinforce them. Many of those with the lowesteducation levels come from families characterizedby social disadvantage.

Getting a good education can create a virtuouscircle of life chances. There is extensive evidencethat education improves prospects not just forearnings and employment but also for health, civicengagement and social mobility (Lochner, 2004;Machin et al., 2006). Conversely, low levels ofeducation are associated with entrenchedemployment disadvantage, restricted socialmobility and a wide range of social problems.When individuals and groups emerge fromeducation systems with low levels of achievement,they and their children face a heightened risk ofmarginalization in many aspects of their lives.Education systems provide a mechanism foroffsetting social disadvantage, but whenopportunities and outcomes are skewed theycan reinforce social divisions.

There are obvious differences in the experienceof education marginalization in rich and poorcountries. One is in the degree of absolutedeprivation. Almost nobody in the rich world entersadulthood with fewer than four years of education,let alone fewer than two years. Relative deprivationis another matter. Many education systems inrich countries have entrenched patterns ofmarginalization linked to poverty, the socialand economic status of parents, ethnicity, raceand other factors.

Marginalization in education in France, Germany,the United Kingdom or the United States is clearlynot the same as in Cambodia or Mali. Yet there aretwo parallels. First, the playing field for opportunityis highly uneven: some groups and individuals entereducation systems facing a heightened risk offailure. Second, education systems themselvesoften reinforce and perpetuate wider socialdisadvantages.

Dropping out of schoolLeaving school too early is strongly linked withmarginalization. Young people with only a lowersecondary education have limited opportunities torealize their potential and develop their learningskills. They face disadvantages in employment andare at greater risk of poverty and social exclusion.

School dropouts represent a significant educationunderclass in many countries. In the EuropeanUnion, 15% of people aged 18 to 24 in 2006 left

school with only lower secondary education andwere not in further education or training. Theshare affected ranged from just over 10% insome countries, including France and the UnitedKingdom, to 20% in Italy and 30% in Spain. Cross-country research has identified parentalwealth, child poverty, ethnicity and gender asmajor factors influencing dropout rates(European Commission, 2008).

Evidence from the United States illustrates thepattern of risk factors associated with being outof school. In 2006, about 8% of people aged 16 to19 were neither enrolled in school nor working.Family poverty contributed strongly to being outof school. Some 17% of youth from poorhouseholds were out of school, compared with5% from non-poor households. Race and ethnicitywere also important, with 11% of African-Americanand Hispanic youth reported as out of school –double the share for white and Asian youth(US Department of Education, 2007).

These data reflect underlying social disadvantageslinked to school dropout. One high-profile nationalreport documented a secondary school dropoutepidemic in the United States (Bridgeland et al.,2006), with around 1 million school leavers eachyear lacking a diploma. The epidemic is unequallyspread. African-American and Hispanic youth arehighly disadvantaged. Whereas the graduation ratefor white students is 84%, it falls to 72% for Hispanicand 65% for African-American students (Heckmanand LaFontaine, 2007). Parental poverty and lowlevels of education are other major risk factors.Among student characteristics, low test scoresand pregnancy contribute strongly to dropout rates.While the factors behind dropout are varied andcomplex, the consequences are uniformly severe.Students who drop out typically earn 30% to 35%less than students with a secondary school diploma(Tyler and Lofstrom, 2009).

Learning achievementIn a country with equal opportunities for learning,it would be impossible to predict educationoutcomes on the basis of individual or groupcharacteristics. No country has achieved this state,but countries differ markedly in the degree towhich social circumstances shape educationopportunity and in the degree to which educationsystems counteract marginalization.

Students from more advantaged socio-economicbackgrounds generally perform better in tests of

In the UnitedStates, thegraduation rate for white studentsis 84%, but fallsto 65% forAfrican-Americanstudents

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 5 6

learning achievement. Analysis of national datafrom the 2006 PISA science tests given to 15-year-olds shows that, on average, socio-economicbackground explains 14% of the variation inperformance. There is marked variation aroundthe average. Socio-economic characteristics weighfar more heavily in some countries, such asFrance, Germany, the United Kingdom and theUnited States, than in others, including Finland,Japan and the Republic of Korea, all of whichachieve higher average scores (OECD, 2007b).

Figure 3.18 illustrates the weight of inheritedcircumstance in shaping learning achievement.It suggests that high levels of inequality areparticularly damaging for children fromhouseholds at the lower end of the socio-economicdistribution. Consider the following comparisons.The share of the national variation in PISAmathematics scores explained by socio-economicstatus is far greater in Germany than in Finland,with German children in the lowest socio-economic group twice as likely to score at thelowest level in mathematics tests. The contrast

between (less equal) France and (more equal)the Republic of Korea is equally striking. Doesthe higher level of equity achieved in Finland andthe Republic of Korea come at the price of loweraverage performance? On the contrary, bothcountries have higher mean test scores in PISAthan France or Germany.

Household poverty, a core element in socio-economic disadvantage, is strongly associated withlow levels of education achievement. In England,students receiving a free school meal – a sign ofhousehold deprivation – have far lower averagetest scores than other students. The score gap inEnglish is 16% and the gap in mathematics is 29%.The share of this group leaving school with highscores on national tests is one-third the nationalaverage (Vignoles, 2009; UK Department forChildren, Schools and Families, 2008).

Wealth-based performance differences in Franceare equally marked. Almost half the children fromthe poorest households are significantly behindtheir peers by sixth grade. By age 15, around 15%of the poorest students are at least two yearsbehind the ninth grade performance level – threetimes the national average. By age 17, almostone in five poor youth have given up their studies(France Council for Employment, 2008).

Poverty effects combine with other factors thatcontribute to marginalization. In the United States,schools with high concentrations of poverty (withover 75% of students eligible for free or subsidizedlunch) had the lowest percentage of whitestudents, the highest percentage of African-American and Hispanic students, and the highestpercentage of students who reported alwaysspeaking a language other than English at home.They also had the highest percentage of fourth-graders being taught by a teacher with fewerthan five years of experience (US Departmentof Education, 2007). Test score gaps reflect thecumulative disadvantage. On the internationalTIMSS scale for mathematics in grade 8, theUnited States ranks ninth out of forty-eightcountries. Hispanic students, however, score justabove the level of Malaysia. On an internationalscale, schools with high concentrations of povertyand African-American students score betweenthe average levels of Malaysia and Thailand(Figure 3.19). These very large test score effectspoint to limited success by the education systemin counteracting wide social disadvantages.

Socio-economicdisadvantageweighs more

heavily on testscores in somecountries than

others

0 5 10 15 20 250

1

2

3

4

5

6

Socio

-eco

nom

ic di

sadv

anta

ge in

dex

Variance in mathematics scores explained by socio-economic status (%)

Turkey

France

Hungary

Slovakia

Germany

Iceland

Japan

Rep of Korea

Finland

Denmark

Sweden

United StatesSwitzerland

Stude

nts w

ith lo

w so

cio-ec

onom

ic sta

tus a

re m

ore l

ikely

to be

amon

g the

wor

se pe

rform

ers

Socio-economic status explains a larger shareof the differences in mathematics scores

Belgium

Greece

Figure 3.18: Socio-economic disadvantage in education weighs more heavily in some countries than othersOdds ratio for likelihood of lowest socio-economic status students aged 15 being among the bottomperformers and % of mathematics score variance explained by socio-economic status, OECD countries

Note: The socio-economic disadvantage index is the relative likelihood of students with the lowest socio-economicstatus (SES) scoring below or at proficiency level 1 when compared to student with the highest SES.Sources: OECD (2006a, 2007a).

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 5 7

Migrant students in many countries face a farhigher risk of education marginalization thannative students do. Their participation in schoolis more likely to be disrupted by leaving early – and migrant students often lag in learningachievement. Research based on evidence fromPISA surveys shows that, in most OECD countries,first-generation immigrants typically lag anaverage of about 1.5 years behind their nativecounterparts (OECD, 2007b). In several countries,including Germany, the Netherlands andSwitzerland, the proportion of immigrant studentsfailing to reach level 2 in the 2006 PISAassessment was at least three times as highas the proportion of native students (Figure 3.20).

Countries vary also in the degree to which theyare narrowing learning achievement gaps. Whilethe gap is narrowing in Sweden and Switzerland,it is widening in Germany and the Netherlands(OECD, 2007a). Education policy is just part of theexplanation for these trends. Patterns of migrantdisadvantage are closely associated in manycountries with home language, country of origin,neighbourhood effects and other kinds of socialdeprivation. But education systems can helpnarrow or widen the gap.

Early tracking of students into different abilitystreams and types of school has been found inseveral cross-country studies to be associatedwith greater inequality in achievement withoutany discernible benefits for average performance(Hanushek and Wößmann, 2006). Being labelledas ‘low ability’ at an early age may lead studentsto internalize low expectations and lose motivation.Differences in tracking policies may help explainwhy students of Turkish origin tend to performbetter in Switzerland (where tracking is delayed)than in Germany (which tracks students early),two countries where many migrants arechannelled into vocational streams (Nusche, 2009;OECD, 2006b). They also go some way towardsexplaining the very large variation in performancebetween schools in Germany linked to socio-economic status. In Finland, less than 5% ofoverall performance variation of students can betraced to inequalities between schools, comparedwith over 70% in Germany – twice the OECDaverage (OECD, 2006b).

Racial and ethnic minority groups experiencesome of the most severe education disadvantage,which can be traced to deeply engrained and oftencenturies-old patterns of cultural discrimination

and stigmatization. Low educational achievementreflects the durability of these patterns, interactingwith social and economic inequalities to perpetuatesocial exclusion.

0

10

20

30

40

Stud

ents

bel

ow p

rofic

ienc

y lev

el 2

(%)

Switzerland Netherlands Germany

Second-generationimmigrant students

Native students

Figure 3.20: Second-generation immigrants in rich countries perform far below native students in science% of students aged 15 scoring below proficiency level 2 on PISA mathematics scale,* secondgeneration immigrants and native students in Germany, the Netherlands and Switzerland, 2006

* ’Level 2 on the PISA proficiency scale represents the baseline level of mathematics proficiency at whichstudents begin to demonstrate the kind of skills that enable them to actively use mathematics: for example, they are able to use basic algorithms, formulae and procedures, to make literal interpretations and to applydirect reasoning’ (OECD, 2007a, p. 107).Source: OECD (2007a).

350

400

450

500

550

600

650

Iran, Isl. Rep.Egypt

TunisiaTurkey

Thailand

Malaysia

United States

SingaporeRep. of Korea

Aver

age

mat

hem

atics

scor

e

Top 10% performers

Asian

White

United States national average

HispanicHigh poverty schoolsAfrican-American

Bottom 10% performers

Figure 3.19: Same country, different worlds of learning achievementTIMSS average mathematics scores in the United States and selected countries, grade 8, 2007

Source: Gonzales et al. (2008).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 5 8

One particularly stark example of marginalizationis the experience of the Roma community.3Assessing the full extent of the deprivation faced byRoma children in education is difficult, as data areoften partial and unreliable (Box 3.3). The data thatare available tell their own story. In most centraland eastern European countries no more than 20%to 25% of Roma children attend secondary schooland the vast majority of those are enrolled invocational education. Many drop out of primaryschool. It is estimated that 15% to 20% of Romachildren in Bulgaria and 30% in Romania do notcontinue beyond fourth grade. The problem is notrestricted to central and eastern Europe. It isestimated that half of Italy’s Roma children arein primary school but fewer than 2% progress toupper secondary education. While data are scarce,education outcomes for Roma fall well below thelevels for the majority population (Open SocietyInstitute, 2007).

Roma education experiences underline thedamage that can be inflicted by bad policies.In many countries, education policies andpractices have the effect of creating segregation.Geographic concentration is one factor. InBulgaria, an estimated 70% of Roma childrenstudy in schools where the share of the majoritypopulation is less than 50%. Moreover, Roma

children are often more likely than their peersto be diagnosed as ‘special needs’ students andplaced in separate schools (Open Society Institute,2007). In Hungary, one report found that ‘aboutevery fifth Roma child is declared to be mildlymentally disabled’ (Roma Education Fund, 2007,p. 32). Such practices reflect cultural attitudes andnegative stereotyping. One Council of Europe reporton Slovakia found that up to half of Roma childrenin special elementary schools were there asa result of erroneous assessment (EuropeanCommission against Racism and Intolerance, 2009).

The legacy of marginalization facing indigenouspeople in rich countries has received insufficientattention in international education debates. ForNative Americans, the Aboriginals of Australiaand the Ma-ori of New Zealand, the imprint ofdiscrimination, stigmatization and socialbreakdown is clearly visible in education data.Only 34% of indigenous Australians aged 15 to24 are in education, compared with 55% of theirnon-indigenous peers. Indigenous people alsoscore lower on reading and numeracy tests,especially if they live in remote areas (Figure 3.21).In very remote areas, the share of indigenousAustralians falling below the national minimumbenchmark for reading is more than double thelevel for all students. In New Zealand, there is

3. Roma – often knownas Gypsies – live primarilyin central and easternEurope and are the mostpopulous subgroup of the Romani.

In most centraland eastern

Europeancountries no

more than 25%of Roma childrenattend secondary

school

With an estimated population between 8 millionand 12 million, Roma are one of Europe’s largestminorities. They are also among the mostmarginalized. Throughout Europe, Roma faceinstitutionalized discrimination, limited opportunitiesfor participation in many aspects of society and pooraccess to good-quality education.

Lack of data makes it difficult to measure the scaleof Roma marginalization. It also limits public debateand the development of effective policy responses.Census data often undercount Roma because thesocial stigma attached to Roma identity leads manyto misreport or refuse to report their identity.Administrative data are also frequently lacking.In some countries, such as Romania and Slovakia,this is because of privacy legislation that restrictsreporting on ethnicity.

While data on Roma are scarce overall, the absenceof reliable statistics on education is a particular

weakness. Problems noted by the European RomaRights Centre range from under-reported births tounreliable and inconsistent data on school enrolment,dropout and other indicators collected by schoolauthorities. A qualitative study in Bulgaria found that administrative data failed to report a significantnumber of out-of-school Roma children becausehouseholds were not registered or school databaseswere incomplete.

Pressure to improve the scope and reliability ofmonitoring data on Roma has been building. TheDecade of Roma Inclusion 2005–2015, an initiativesupported by the World Bank, the United NationsDevelopment Programme, the European Commission,the Council of Europe and the Open Society Institute,has led to a range of initiatives aimed at challengingthe use of human rights laws to prohibit datacollection and at improving ethnic data disaggregationand clarifying ‘Roma identity’.

Sources: Open Society Institute (2007); European Roma RightsCentre (2007); European Commission (2009a).

Box 3.3: Monitoring gaps and marginalization — Roma in Europe

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 5 9

encouraging evidence that Ma-ori children –especially girls – are catching up with non-Ma-ori.Even so, the achievement gaps remain large.Whereas 65% of all students leave school withthe National Certificate of EducationalAchievement (NCEA) level 2 qualification, thefigure drops to 44% for Ma-ori children. Ma-orilearners are three times as likely as non-Ma-orito leave school with no qualification (New ZealandMinistry of Education, 2009).

Speaking a minority language is also oftenassociated with low levels of educationachievement. In many countries, large numbersof children are taught and take tests in languagesthat they do not speak at home, hindering the earlyacquisition of reading and writing skills. Theirparents may lack literacy skills or familiarity withofficial languages used in school, so that the homeenvironment reinforces learning opportunity gapsbetween minority and majority language groups.

International and national learning assessmentsconfirm the importance of home language as afactor in test scores. The TIMSS 2007 assessmentfound that fourth- and eighth-grade students whoreported ‘always speaking’ at home the languagein which the test was conducted score significantlyhigher. For fourth-grade science students whoreported only ‘sometimes speaking’ the testlanguage at home, the test score was 10% lower.For students who reported ‘never speaking’ thetest language at home, the score was 20% lower(Martin et al., 2008).

Conclusion

Making sure that everyone has a chance to developtheir potential through education is an importantchallenge for all countries. Equal opportunity ineducation is a basic human right. Moreover, fairand inclusive education is one of the most powerfullevers available for making societies moreequitable, innovative and democratic. Overcomingthe extreme and persistent disadvantages thatmarginalized groups experience is a vital elementin the wider agenda for inclusive education.Extending opportunity to these groups requiresmore than the general expansion of education andthe improvement of average learning achievementlevels. It requires policies that target the underlyingcauses of disadvantage in education and beyond.

Data have an important role to play in theformulation of such policies. Disaggregatedhousehold survey data such as those availablein the DME data set (Table 3.3) can provide policy-makers with the means to identify socialgroups and areas characterized by high levelsof deprivation. They can also provide insightinto the interaction between different patternsof disadvantage, informing approaches totargeting it and the development of strategiesaimed at equalizing opportunity. That is whyinvestment in data collection and analysis shouldbe an integral element of any national povertyreduction strategy.

Investment indisaggregateddata collectionshould be anintegral element of any nationalpoverty reductionstrategy

Metropolitan Provincial Remote Very remote0

20

40

60

80

100

% o

f pop

ulat

ion

who

mee

ts re

adin

g be

nchm

ark

Indigenous

All students

Figure 3.21: Indigenous Australians perform consistently below the student average in reading% of population meeting reading benchmarks at grade 7, indigenous Australians and all students, by location, 2006

Sources: Australia Department of Education (2008), Table A3.4, p. 190; Biddle and Mackay (2009).

Table 3.3: Deprivation and Marginalization in Education, selected data, latest year available1

Total Male FemaleRichest

20%Poorest

20%

Absolutewealth

gap

(Share of the poorestwealth quintile in the

bottom 20% of theeducation distribution,

by years in school) Total Male Female

Rural girls fromthe poorest

quintileYear ofsurvey

Average number of years of education ‘Bottom 20%’

Education poverty(Share of the population with fewer

than four years of education)

Albania

Armenia

Azerbaijan

Bangladesh

Belize

Benin

Bolivia

Bosnia/Herzeg.

Burkina Faso

Burundi

Cambodia

Cameroon

C. A. R.

Chad

Colombia

Congo

D.R. Congo

Côte d’Ivoire

Cuba

Dominican Rep.

Egypt

Ethiopia

Gabon

Gambia

Georgia

Ghana

Guatemala

Guinea

Guinea-Bissau

Haiti

Honduras

India

Indonesia

Iraq

Jamaica

Jordan

Kazakhstan

Kenya

Kyrgyzstan

Lao PDR

Lesotho

Liberia

2005 9.9 9.9 9.9 11.5 8.7 2.7 32.5

2005 9.2 9.3 9.2 9.8 8.2 1.6 21.5

2006 10.6 10.8 10.5 11.4 9.7 1.7 34.9

2004 5.8 5.9 5.7 8.0 3.6 4.4 40.4

2005 8.5 8.4 8.6 8.7 7.4 1.3 28.1

2006 4.6 6.0 3.4 7.7 2.2 5.6 31.6

2003 9.6 9.8 9.3 11.2 6.4 4.8 56.6

2005 11.2 11.1 11.2 11.7 10.4 1.4 29.4

2003 2.3 2.7 2.0 5.6 0.6 4.9 28.2

2005 4.6 4.8 4.5 6.6 2.6 4.0 32.9

2005 6.0 6.5 5.5 8.2 3.4 4.8 36.3

2004 6.4 6.9 6.0 8.8 3.5 5.3 50.8

2000 0.9 1.2 0.7 1.7 0.5 1.2 22.0

2004 2.7 3.9 1.8 6.7 1.0 5.7 32.9

2005 9.1 8.7 9.5 10.9 6.4 4.5 47.2

2005 7.2 7.3 7.1 8.9 5.4 3.5 37.8

2007 6.2 7.0 5.4 9.1 4.4 4.7 29.9

2004 4.3 5.5 3.3 6.3 2.1 4.2 23.8

2005 11.3 11.1 11.6 11.7 10.6 1.1 25.1

2007 9.2 8.7 9.8 10.8 6.2 4.6 43.0

2005 8.9 9.4 8.5 9.8 6.4 3.4 39.5

2005 3.1 3.8 2.5 7.5 1.6 5.9 32.9

2000 7.3 7.6 7.0 8.4 5.0 3.4 41.1

2005 5.4 6.2 4.7 8.1 2.4 5.7 30.0

2005 11.6 11.5 11.8 12.6 10.3 2.2 38.0

2003 7.1 7.5 6.8 9.2 3.3 5.8 52.5

1999 5.8 6.0 5.7 8.8 1.8 7.0 47.9

2005 3.5 4.9 2.3 5.8 1.0 4.8 31.7

2005 3.8 4.7 3.0 6.2 1.7 4.5 36.0

2005 6.4 6.4 6.3 8.5 3.3 5.2 45.4

2005 7.6 7.1 8.1 10.5 4.0 6.5 42.4

2005 7.2 8.1 6.5 11.1 4.2 6.9 50.2

2003 9.0 9.0 9.1 11.1 6.6 4.5 41.1

2005 5.8 6.2 5.3 7.2 3.9 3.3 39.6

2005 10.3 10.3 10.4 10.5 10.0 0.5 27.5

2007 11.5 11.3 11.8 12.1 11.4 0.6 21.6

2005 12.3 12.1 12.5 13.2 11.4 1.8 27.0

2003 7.5 7.5 7.5 9.5 5.4 4.1 35.2

2005 10.6 10.5 10.7 10.9 10.4 0.5 17.4

2000 8.2 8.8 7.7 8.1 5.9 2.3 34.6

2004 6.8 5.9 7.7 8.8 4.6 4.2 40.4

2007 5.0 5.7 4.5 7.2 2.6 4.6 46.1

1.0

0.9

2.3

27.4

3.5

47.8

6.1

0.6

70.9

36.6

26.8

21.4

89.2

67.3

6.7

12.4

25.4

49.2

1.1

7.9

15.0

61.1

9.8

40.4

0.7

20.3

35.6

56.4

49.7

23.2

18.1

23.6

4.4

20.6

0.5

2.0

0.4

12.0

0.5

13.6

15.4

35.7

1.1 0.9

1.4 0.6

1.3 3.1

26.4 28.1

4.7 2.1

33.4 60.5

4.1 8.1

0.7 0.4

66.5 75.1

34.5 38.5

22.5 30.9

16.1 26.2

86.0 92.3

53.0 79.0

8.4 5.0

11.1 13.4

15.6 34.5

36.5 59.5

1.4 0.8

10.3 5.5

9.7 20.2

51.8 70.2

8.2 11.2

31.6 48.0

1.0 0.5

15.9 23.9

33.1 37.9

39.6 71.1

37.7 59.9

23.4 23.1

21.4 14.9

15.6 30.7

4.3 4.5

15.3 26.2

0.3 0.7

1.9 2.1

0.4 0.3

12.3 11.7

0.7 0.3

9.6 17.2

25.3 5.5

27.6 42.4

0.9

2.3

6.4

50.1

2.8

85.0

30.1

2.0

93.1

64.6

62.8

63.9

99.7

92.2

18.9

27.2

57.0

84.8

0.7

22.0

47.9

87.4

24.2

80.2

0.7

67.2

86.2

94.5

90.3

59.0

42.8

60.1

12.6

56.9

0.0

2.5

0.2

27.6

1.6

42.5

11.9

79.3

(Years)Country (%) (%)

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 6 0

Most deprived region2Education

poverty

Mostdeprivedregion tocountryaverage Total Male Female Total Male Female

Rural girlsfrom thepoorestquintile

Education poverty(Share of the population with fewer

than four years of education)

Extreme education poverty(Share of the population with fewer

than two years of education)Share of the population

aged 7–16 with no education

Albania

Armenia

Azerbaijan

Bangladesh

Belize

Benin

Bolivia

Bosnia/Herzeg.

Burkina Faso

Burundi

Cambodia

Cameroon

C. A. R.

Chad

Colombia

Congo

D.R. Congo

Côte d’Ivoire

Cuba

Dominican Rep.

Egypt

Ethiopia

Gabon

Gambia

Georgia

Ghana

Guatemala

Guinea

Guinea-Bissau

Haiti

Honduras

India

Indonesia

Iraq

Jamaica

Jordan

Kazakhstan

Kenya

Kyrgyzstan

Lao PDR

Lesotho

Liberia

South

Aragatsotn

Dakhlik Shirvan

Sylhet

Stann Creek

Alibori

Potosi…

Est

Nord

Mondol Kiri and Rattanak Kiri

Extrême Nord

Vakaga

Barh Azoum

Atlantica

Nord

Nord-Kivu

Nord

Occidente

Elías piña

Rural Upper Egypt

Somali

Ngouni and Nyanga

Basse

Mtskheta-Mtianeti

Northern

North-west

Kankan

Bafatá and Gabu

Centre

Copán

Bihar

West Kalimantan

North…

Central

Akmola oblys

North eastern

Naryn

South

Thaba-Tseka

North Western

2.0 2.1

3.2 3.4

5.8 2.5

34.6 1.3

5.5 1.6

91.2 1.9

15.2 2.5… …

92.3 1.3

52.1 1.4

70.4 2.6

65.4 3.1

98.1 1.1

96.7 1.4

8.5 1.3

18.9 1.5

44.5 1.8

77.0 1.6

2.0 1.8

24.7 3.1

28.1 1.9

86.3 1.4

13.3 1.4

79.4 2.0

2.4 3.3

61.6 3.0

61.3 1.7

73.7 1.3

77.1 1.6

50.8 2.2

41.5 2.3

42.8 1.8

12.9 2.9

24.3 1.2… …

2.5 1.2

1.3 3.3

74.7 6.2

2.3 5.0

17.6 1.3

30.6 2.0

54.0 1.5

0.9

0.7

1.1

19.5

2.3

40.7

1.7

0.3

66.9

24.4

12.4

15.0

76.5

57.6

2.8

5.4

15.4

41.9

0.6

4.0

12.5

49.8

5.2

38.6

0.6

16.6

19.1

52.0

37.3

10.3

7.8

20.3

1.5

13.1

0.4

1.6

0.3

7.9

0.3

7.8

8.2

21.9

1.1 0.7 0.6 0.7 0.5 0.9

1.0 0.4 3.7 4.0 3.4 5.6

0.7 1.4 2.2 1.8 2.7 4.7

17.7 20.8 10.6 11.9 9.2 13.5

3.5 1.1 0.7 0.7 0.8 0.6

26.1 53.4 28.3 23.3 33.8 54.1

1.0 2.2 1.6 1.5 1.7 4.0

0.4 0.3 0.8 0.9 0.7 0.5

61.9 71.6 63.2 59.5 67.0 83.5

22.0 26.6 19.2 17.4 21.0 38.5

8.7 15.9 9.1 9.6 8.7 17.3

9.4 20.1 13.0 10.8 15.2 33.2

70.8 82.0 48.2 42.9 53.6 78.3

43.6 69.1 54.2 47.6 60.8 80.3

3.6 2.1 1.7 2.0 1.3 4.2

4.6 6.0 4.5 4.0 4.9 8.8

7.5 22.8 18.9 16.2 21.8 35.7

29.7 51.8 33.0 30.0 35.9 52.5

0.7 0.6 0.3 0.3 0.3 0.1

5.2 2.7 1.9 2.4 1.5 3.7

6.9 18.2 10.8 8.3 13.5 27.3

39.2 60.4 48.1 46.5 49.9 62.4

4.5 5.9 2.6 2.6 2.6 3.1

30.3 45.8 29.0 27.8 30.2 48.7

0.8 0.4 0.5 0.4 0.7 1.1

12.1 20.4 21.0 20.5 21.5 52.5

15.5 22.5 13.0 10.9 15.2 34.2

35.4 66.6 44.9 40.8 49.2 68.1

23.8 48.8 28.0 27.3 28.6 44.2

9.4 11.2 11.6 12.5 10.7 26.7

9.4 6.3 5.1 6.1 4.1 9.4

12.1 27.7 13.2 10.7 15.9 29.1

1.4 1.7 2.0 2.2 1.9 5.4

8.4 18.1 9.9 6.1 13.8 35.1

0.2 0.7 0.3 0.2 0.4 1.3

1.5 1.7 0.8 0.9 0.7 1.2

0.4 0.3 0.6 0.5 0.6 0.3

7.0 8.8 6.7 6.3 7.0 17.1

0.5 0.2 2.1 2.7 1.6 2.3

5.7 9.7 5.4 4.7 6.1 19.9

14.7 1.7 5.9 8.5 3.1 5.3

14.4 28.2 43.9 42.3 45.5 69.2

Country(%) (%) (%)Ratio

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 6 1

Table 3.3 (continued)

Total Male FemaleRichest

20%Poorest

20%

Absolutewealth

gap

(Share of the poorestwealth quintile in the

bottom 20% of theeducation distribution,

by years in school) Total Male Female

Rural girls fromthe poorest

quintileYear ofsurvey

Average number of years of education ‘Bottom 20%’

Education poverty(Share of the population with fewer

than four years of education)

TFYR Macedonia

Madagascar

Malawi

Mali

Mongolia

Montenegro

Morocco

Mozambique

Myanmar

Namibia

Nepal

Nicaragua

Niger

Nigeria

Pakistan

Peru

Philippines

Rwanda

S. Tome/Principe

Senegal

Serbia

Sierra Leone

Somalia

Suriname

Swaziland

Syrian A. R.

Tajikistan

U. R. Tanzania

Togo

Trinidad/Tobago

Turkey

Uganda

Ukraine

Venezuela, B. R.

Viet Nam

Yemen

Zambia

Zimbabwe

2005 10.5 10.1 10.9 11.9 7.3 4.7 22.5

2004 4.5 4.5 4.5 9.8 1.7 8.1 53.5

2004 6.2 6.5 5.9 8.6 4.8 3.8 29.1

2001 2.6 3.3 2.0 5.8 0.7 5.1 26.1

2005 8.8 8.5 9.1 10.7 6.0 4.7 52.5

2005 11.2 11.2 11.2 12.2 9.3 2.9 39.6

2004 5.7 6.6 5.0 8.4 2.0 6.4 41.3

2003 3.2 3.7 2.8 5.0 1.9 3.2 27.8

2000 5.6 5.7 5.5 7.7 3.6 4.1 32.7

2007 8.4 8.0 8.8 10.5 6.8 3.6 32.2

2006 5.9 7.1 5.0 8.7 3.3 5.4 40.9

2001 6.6 6.2 7.0 9.5 2.5 6.9 51.8

2006 1.7 2.4 1.2 5.5 0.6 4.9 26.3

2003 6.7 7.4 6.0 9.7 3.5 6.2 48.5

2007 5.7 6.5 5.0 9.0 2.4 6.5 34.0

2004 10.1 10.2 10.0 11.5 7.4 4.1 53.3

2003 9.4 8.9 10.0 11.0 6.3 4.7 55.4

2005 3.8 3.8 3.8 5.5 2.9 2.6 28.5

2000 8.4 8.9 8.0 11.3 5.9 5.4 30.8

2005 3.2 3.9 2.7 5.1 1.2 4.0 28.6

2005 11.1 10.9 11.3 11.9 8.3 3.6 17.1

2005 3.8 4.8 2.6 6.9 1.4 5.6 32.3

2005 3.1 4.8 2.0 6.6 0.4 6.2 31.5

2000 7.6 7.4 7.9 9.5 6.1 3.4 34.0

2006 8.0 7.8 8.2 9.4 6.4 3.0 34.7

2005 8.1 8.1 8.0 9.8 6.2 3.6 37.4

2005 9.6 10.3 9.1 10.8 9.2 1.6 24.9

2004 5.5 5.6 5.4 7.9 3.9 4.0 37.6

2005 6.5 7.2 5.5 8.2 4.1 4.1 34.9

2000 10.4 10.2 10.7 11.9 9.4 2.5 36.9

2003 8.1 8.8 7.4 9.2 6.3 2.9 28.7

2006 6.5 6.9 6.2 8.5 4.3 4.1 40.6

2007 13.0 12.8 13.1 13.3 12.4 0.9 21.5

2000 9.1 8.5 9.6 11.8 5.5 6.2 52.6

2002 8.3 8.4 8.3 10.4 5.3 5.2 51.0

2005 6.7 8.6 4.9 … 3.2 … 51.6

2001 6.4 6.7 6.1 8.9 4.0 4.9 39.8

2006 8.6 8.7 8.6 9.9 7.0 2.9 45.8

3.2

48.3

23.6

68.8

9.0

2.4

37.1

56.5

25.1

8.9

29.9

23.3

76.9

27.9

34.5

3.5

5.0

45.3

10.8

57.4

1.1

53.7

63.5

11.6

9.1

6.8

3.7

26.7

22.3

0.6

7.9

16.7

0.3

15.0

8.7

28.7

20.3

2.8

3.9 2.6

46.9 49.8

20.6 26.2

60.3 76.0

10.6 7.5

1.8 3.0

27.0 45.9

49.8 62.7

22.0 27.9

12.0 6.2

16.8 39.0

27.5 19.2

68.4 82.8

19.6 35.4

25.9 42.5

2.3 5.0

6.7 3.2

45.1 45.4

8.7 12.5

50.3 63.3

1.3 1.0

42.0 67.3

44.6 75.7

11.7 11.6

10.6 7.7

5.1 8.9

2.2 5.1

23.9 29.0

15.7 30.5

0.6 0.7

3.8 12.0

12.7 20.1

0.5 0.2

20.5 11.0

8.7 8.7

10.4 46.2

16.0 24.2

3.0 2.7

7.0

83.1

37.4

94.0

22.4

3.8

87.6

87.1

56.1

12.1

67.4

60.7

94.7

66.7

84.9

16.9

16.3

59.1

17.8

87.6

3.8

88.9

98.3

13.0

18.8

27.0

7.7

48.8

64.4…

21.4

45.3

0.0

19.4

29.7

89.9

46.5

8.0

(Years)Country (%) (%)

Notes:1. Data are for the population aged 17 to 22 unless otherwise stated.2. Regions presented are the first official administrative division level except those in italics which are geographic areas rather than official administrative divisions.Source: UNESCO-DME (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 6 2

Most deprived region2Education

poverty

Mostdeprivedregion tocountryaverage Total Male Female Total Male Female

Rural girlsfrom thepoorestquintile

Education poverty(Share of the population with fewer

than four years of education)

Extreme education poverty(Share of the population with fewer

than two years of education)Share of the population

aged 7–16 with no education

TFYR Macedonia

Madagascar

Malawi

Mali

Mongolia

Montenegro

Morocco

Mozambique

Myanmar

Namibia

Nepal

Nicaragua

Niger

Nigeria

Pakistan

Peru

Philippines

Rwanda

S. Tome/Principe

Senegal

Serbia

Sierra Leone

Somalia

Suriname

Swaziland

Syrian A. R.

Tajikistan

U. R. Tanzania

Togo

Trinidad/Tobago

Turkey

Uganda

Ukraine

Venezuela, B. R.

Viet Nam

Yemen

Zambia

Zimbabwe

Vardar

Toliary

Central

Mopti

Eastern

Central

Gharb-Chrarda-Beni Hssen

Cabo Delgado

Eastern Shan

Kunene

Central (Madhyamanchal)

Jinotega

Tillabéri

North West

Balochistan

Cajamarca

Aut. Region in Muslim Mindanao

Kibungo

South

Matam

Southeast

Northern

Central south

Sipaliwini

Lubombo

Ar-Raqqah

Khatlon

Tabora

Savanes

Tobago

East

Northern Region

North…

Mekong river delta

Hajjah

Eastern…

7.5 2.3

68.2 1.4

26.5 1.1

88.6 1.3

19.0 2.1

4.0 1.7

51.9 1.4

75.1 1.3

74.5 3.0

28.0 3.1

35.0 1.2

57.1 2.4

86.6 1.1

58.0 2.1

47.6 1.4

16.8 4.8

14.1 2.8

55.4 1.2

24.9 2.3

78.7 1.4

2.1 1.8

66.7 1.2

72.7 1.1

62.1 5.3

16.2 1.8

22.4 3.3

4.9 1.3

50.3 1.9

54.5 2.4

2.1 3.3

21.4 2.7

29.0 1.7

0.7 2.2… …

18.2 2.1

54.4 1.9

39.5 1.9… …

2.3

27.5

11.0

65.1

4.8

2.2

29.9

40.2

9.2

5.6

22.8

13.5

73.7

25.3

30.7

1.2

1.8

20.8

5.8

50.4

1.0

48.5

57.1

6.0

5.2

4.9

2.1

20.5

15.1

0.6

6.0

8.3

0.3

4.0

4.3

23.6

11.5

1.3

2.5 2.1 2.2 1.9 2.5 1.5

26.3 28.8 17.3 18.0 16.6 30.0

8.3 13.4 9.7 10.5 8.9 12.3

56.7 72.1 52.7 48.4 57.0 75.3

5.9 3.8 2.6 3.3 1.9 5.2

1.6 2.7 1.4 1.5 1.2 0.7

18.8 39.4 10.5 7.4 13.6 34.9

33.4 46.5 26.7 23.3 30.2 44.0

7.4 10.8 9.0 9.0 9.0 20.5

7.3 4.1 4.4 4.8 4.0 6.5

10.0 31.8 9.3 4.9 13.6 26.2

16.1 10.9 11.9 13.6 10.2 32.4

64.4 80.3 60.4 53.5 67.5 79.1

17.0 32.7 25.4 20.9 30.0 54.7

21.6 39.2 23.6 17.1 30.2 63.0

0.6 2.0 0.7 0.7 0.7 1.8

2.1 1.3 2.4 2.8 1.9 5.9

20.1 21.4 10.9 11.6 10.3 13.7

4.6 6.9 6.4 6.5 6.4 15.0

42.8 56.6 39.0 37.9 40.0 59.1

1.2 0.8 0.6 0.4 0.8 2.7

37.1 61.8 24.8 22.9 26.7 46.5

39.3 68.6 57.6 48.7 65.2 94.6

5.2 6.8 3.8 3.8 3.7 2.5

5.4 5.1 4.3 4.9 3.6 6.8

3.4 6.7 2.4 2.1 2.8 7.8

1.7 2.4 5.5 4.6 6.5 8.0

16.8 23.8 19.3 20.0 18.5 28.5

9.8 21.8 15.2 11.9 18.8 34.3

0.6 0.6 0.2 0.2 0.3 …

2.1 9.9 4.3 2.3 6.3 12.3

5.7 10.5 6.3 6.1 6.5 17.1

0.5 0.1 1.0 1.2 0.7 0.4

5.7 2.7 4.6 5.5 3.6 12.0

4.1 4.4 1.7 1.4 2.0 5.7

6.7 39.8 21.1 14.3 27.9 58.2

8.9 14.0 23.1 22.9 23.4 39.9

1.3 1.3 1.9 2.1 1.7 2.9

Country(%) (%) (%)Ratio

R E A C H I N G T H E M A R G I N A L I Z E D

M e a s u r i n g m a r g i n a l i z a t i o n i n e d u c a t i o n

1 6 3

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 6 4

Getting left behind

Introduction

Children at risk of marginalization in educationare found in all societies. At first glance, the livesof these children may appear poles apart. The dailyexperiences of slum dwellers in Kenya, ethnicminority children in Viet Nam, a blind girl froma low-income home in Pakistan and a Roma childin Hungary are very different. What they have incommon are restricted opportunities to developtheir potential, realize their hopes and build abetter future through participation in education.

The first part of this chapter identified some of themost marginalized social groups. It documentedmutually reinforcing disadvantages linked topoverty, gender and ethnicity. This part looksbeyond the data to the processes and powerrelationships that diminish opportunity. Itconcentrates mainly on primary school agechildren in developing countries, while recognizingthat early experience tips the balance againstmany children before they enter school and thateducational marginalization continues intoadulthood (see Chapter 2).

Unravelling the threads behind marginalizationin education can be difficult. Many factors areinvolved. Poverty often makes educationunaffordable and pushes children out ofclassrooms and into employment. Genderintersects with low income to create forces ofmarginalization that are less tangible and lesseasily measurable than poverty but no lessdamaging. The low value placed on girls’ educationcan make them the last into school and the first outwhen poverty strikes. Cultural attitudes and beliefs,stigmatization and discrimination also fuelmarginalization, locking children into cycles oflow expectation and underachievement. Moreover,many of the processes leading to marginalizationin education can be traced to deeply entrenchedpower relationships that perpetuate poverty andgender disadvantages and group-based inequalities.

The interaction is two-way. Marginalizationin education is in part a consequence ofmarginalization in other areas. But it is also acause of marginalization. Education systems havethe potential to mitigate social disadvantage, yetoften they either fail to utilize that potential or theyactually magnify underlying problems. As this part

of the chapter documents, acts of commission andomission in education policy can place good-qualityschooling far beyond the reach of the marginalized,reinforcing wider social divisions in the process.

The interaction between marginalization ineducation and wider forms of social exclusiondoes not follow general rules. The national andsubnational context matters, as does the specificform of disadvantage that marginalized childrenexperience. Even so, recurrent themes cut acrossdifferent environments and experiences. This partof the chapter looks at these themes, identifyingthe global drivers that fuel the local patternsof marginalization explored in the previous part.The first section looks at poverty as a barrierthat perpetuates disadvantages in education,partly by pushing children into work. Thesecond examines issues behind group-basedmarginalization, tracing the routes throughwhich ethnicity, language, stigmatization andpoverty often interact to create vicious circles oflow expectation and low achievement. The thirdsection considers location-specific factors thatintersect with livelihoods, highlighting problemsfaced by slum dwellers, remote rural communitiesand conflict-affected regions. The fourth examinesdisability and the fifth HIV and AIDS – issues thathave a marked impact on education.

Poverty and child labour

Household poverty is one of the strongestand most persistent factors contributing tomarginalization in education. The transmissionmechanisms are well known. Poor householdshave fewer resources to invest in their children’sschooling, health and other assets. Poverty isalso a source of vulnerability. When poor peopleare hit by economic shocks, droughts or healthproblems, they often lack the resources to copewithout cutting spending in key areas, includingchildren’s schooling. Education can act as apowerful catalyst in breaking cycles of poverty.But poverty itself is a strong constraint onopportunities for education, fuelling thetransmission of disadvantage across generations.

Global poverty trends: a mixed recordThe sheer scale of global poverty makes it aformidable barrier to Education for All. In 2005,nearly 1.4 billion people were living on less thanUS$1.25 a day. More than half the population ofsub-Saharan Africa and 40% of people in SouthAsia fell below this absolute poverty threshold.

Stigmatizationand

discriminationalso fuel

marginalization ineducation,

locking childreninto cycles of low

expectation andunderachievement

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 6 5

Many millions more were living just above thethreshold, surviving on less than US$2 a day,rendering them highly vulnerable to acutepoverty (Figure 3.22).

Global aggregate figures can obscure the depth ofpoverty. The average daily consumption level of apoor person in sub-Saharan Africa is just US$0.73 –a figure unchanged in twenty-five years (Chen andRavallion, 2008). The incidence and depth of povertyare more marked in sub-Saharan Africa than anyother region. Depth of poverty matters because ithas a bearing on capacity for coping with shocks.For people surviving on US$0.73 a day, even smalllosses can have catastrophic consequences fornutrition, health and schooling.

The good news is that the number of peopleworldwide living in extreme poverty has beenfalling. The decline is driven by strong performancein East Asia; progress in most other regions –notably sub-Saharan Africa – has been far lessencouraging. The bad news is that a combinationof rising food prices and the global financial crisishas slowed the pace of poverty reduction. For 2009,there may be 55 million to 100 million more peopleliving below the international poverty line than wasexpected before the crisis. On current economicgrowth projections, the number of people living inextreme poverty could rise in more than half of alldeveloping countries in 2009 (World Bank, 2009k).

Poverty trends in developed countries are alsoa source of concern. Evidence from the OECDsuggests that children are disproportionatelydisadvantaged by household poverty, with anaverage of 12% affected.4 Just as disturbing asthis number is the underlying trend. During adecade of sustained economic growth up to themid-2000s, child poverty rates grew as incomeinequality rose. Rising unemployment causedby the financial crisis is likely to lead to sharpincreases in child poverty during 2009 and 2010(see Chapter 1). The danger is that rising childpoverty will in turn fuel inequalities in education.

Poverty’s effects are transmittedto educationEducation can help lift people out of poverty byboosting productivity and opening doors to jobs andcredit. Conversely, lower educational attainmentis strongly associated with higher poverty levels.The evidence thus points to a negative cycle inwhich poverty begets education disadvantage, whichin turn perpetuates poverty. What drives this cycle?

The inability of poor households to supportinvestment in education is one significant factor.In many countries, parents have to pay a highproportion of their income to put their childrenin school. The costs include official school fees,informal and unofficial charges levied to supportteachers’ pay and other expenses, and paymentsfor uniforms and textbooks.

For the poorest households, schooling competeswith other basic needs, such as health care andfood. A study covering four slums in Bangladeshillustrates the extent of the financial burden. Forthe average household in these slums, expenditureon education amounted to 10% of their incomeper child in school, rising to 20% for the poorestone-fifth of households. Monthly expenditure perchild by the poorest households averaged aroundUS$2 out of an income of less than US$12 permonth. Overall, the largest single cost wasfor supplementary tuition, which many familiesdeem necessary for progress through school(Cameron, 2008).

Eliminating official school fees can help lowerfinancial barriers for the poorest households. From1999 to 2007, fourteen countries, most of them in sub-Saharan Africa, reported abolishing tuition fees.

4. The OECD defines povertyas living in a household withan equivalized householddisposable income of lessthan half the median for the whole population (OECD, 2009g).

For poor people in sub-SaharanAfrica, even smalllosses can havecatastrophicconsequences fornutrition, healthand schooling

0

200

400

600

800

1 000

1 200

1 400

1 600

1 800

2 000

East Asia/Pacific

Latin America/Caribbean

Sub-SaharanAfrica

SouthAsia

Mill

ion

peop

le

51 40 17 8

0.73 0.93 0.95 0.77

Incidence of poverty (%)

Average consumption(2005 US$)

Above poverty thresholds

US$2 - US$1.25

Less than US$1.25

Population living on less than US$1.25 a day

Figure 3.22: Mapping global povertySelected poverty indicators by region, 2005

Notes: ‘Incidence of poverty’ is the proportion of the population whose consumption expenditure is less than US$1.25 a day, at 2005 exchange rates. ‘Average consumption’ is the average daily consumptionexpenditure of those below this poverty line.Source: Chen and Ravallion (2008).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 6 6

This was followed by sharp increases in enrolmentin many countries, including Kenya, Uganda andthe United Republic of Tanzania (UNESCO, 2007).A study in Burundi in 2006, just after fees wereabolished, showed that over 40% of the pooresthouseholds reported some of their children wouldnot be in school had fees not been removed. Thisis consistent with research indicating that, beforefees were abolished, a third of children from thepoorest households were not in school becausetheir parents could not meet costs (World Bankand Burundi Government, 2008).

Poor people often report inability to affordeducation for their children, even in countries withnominally free primary schooling. In Cambodia,cost is among the most commonly cited reasonsfor children being out of school, even though thereare no official charges (World Bank, 2006a). InMalawi and Uganda, where fees were abolishedover a decade ago, many more children from poorhouseholds have entered school. Yet in bothcountries, half the households with children whohave dropped out cite lack of money as the mainproblem (World Bank, 2006h, 2007c). In a surveycovering fifty slums in Delhi, financial constraintswere given as the main reason for school agechildren being out of school or dropping out, eventhough education is nominally free (Tsujita, 2009).

Why has fee abolition failed to eliminate costbarriers? In some cases because legislationeliminating fees has been only partly implemented.In the Lao People’s Democratic Republic, tuitionfees are officially proscribed but about half theschools still levy them (World Bank, 2006c).Indonesia’s free basic education policy, introducedin 2005, provides incentives for schools to eliminatefees but allows them to opt out (World Bank,2006b). Another problem is that formal fees are justone part of the cost of education. In many cases,parents must also buy uniforms and textbooks. InSierra Leone, uniforms double the cost associatedwith school fees (World Bank, 2007b). Poor parentsin Nigeria no longer face tuition charges, but booksand uniforms cost more than fees once did(Lincove, 2009).5

Lowering costs is not a stand-alone strategy. Poorparents – like all parents – also consider the qualityof the education available. In some countries, theelimination of official fees has led to deterioration inquality, with surges in enrolment increasing classsizes and straining the school infrastructure. Toavoid such problems, governments need to assume

responsibility for maintaining education resourcesby raising public spending and sequencing reformsto increase the supply of teachers, classrooms andlearning materials (World Bank and UNICEF, 2009).

Social attitudes strongly condition the effectsof poverty. The degree to which parents valuechildren’s education inevitably influencesprospects of participation in school. For Hausagirls in northern Nigeria, the low value manyadults ascribe to their education is a powerfulsource of exclusion (Box 3.4).

Economic shocks can undermine educationWhile poverty is widely recognized as a barrierto educational opportunity, less attention has beenpaid to vulnerability. One characteristic of beingpoor is that precarious livelihoods carry aheightened risk of insecurity. The pooresthouseholds often find it impossible to shield theirchildren’s schooling from external shocks such asdroughts, floods or economic downturns. They oftenlive in hostile environments and have little accessto assets such as land, livestock, credit or savingsto see them through difficult times. In urban areas,the very poor often work in informal sectors withlow wages and limited security.

Cross-country research on past economic crisesand climate events shows that the effects of shockson schooling tend to be more pronounced in low-income countries than in middle-income countries(Ferreira and Schady, 2008). The children of thepoorest households are most likely to sufferadverse consequences as regards education, healthand nutrition. This risk adds to the threat of povertypersisting across generations.

External shocks can have direct and long-lastingconsequences for education. Droughts in sub-Saharan Africa have had significant effects onenrolment and years in school (Alderman et al.,2006; Ferreira and Schady, 2008). In Zambia,over one-third of those aged 7 to 14 belongedto households that experienced some form ofeconomic shock during 2005. Shocks involvingloss and destruction of property were particularlydamaging for education, raising the probability ofchildren being involved in full-time work by 14% inlow-income households (Understanding Children’sWork, 2009). In Indonesia, the 1997 financial crisisled to significant declines in enrolment amongprimary school age children, especially in thepoorest households (Thomas et al., 2004).

5. Costs of uniforms andtextbooks climb sharplyupon the transition fromprimary to secondaryschool. One review oftextbook provision in sub-Saharan Africa foundthat in eleven countriestextbook costs wereentirely financed byparents and that, next totuition fees, this was thelargest item in householdspending. (Read et al.,2008).

‘When my childis in school…

I have to pay forhis uniforms,

so money is givenout while it is not coming in.[…] School isvery costly.’

Parent, Nigeria

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 6 7

Other effects of such shocks may be lessimmediate. Malnutrition in young children of poorfamilies, for example, may not just lower schoolattendance but also impair cognitive development,learning ability and earnings potential. Whenchildren are born in a drought year or experiencemalnutrition early in their lives, the effects can beseen a decade later in their health and nutritionalstatus, and their educational attainment (Aldermanet al., 2006; Alderman et al., 2009). In Ethiopia,children born in a year in which drought affectedtheir district are 41% more likely to be stunted atage 5 than children born in a non-drought year(UNDP, 2007). Economic shocks can pushhouseholds into long-term poverty. One study inIndonesia found that about half of poverty in 2002could be traced back to the 1997 economicdownturn, even though recovery was well underway (Ravallion, 2008). Underlying gender disparitiesoften lead to girls bearing the brunt of economicshocks. In rural Pakistan, for example,unanticipated economic losses reduced thelikelihood of girls being in school, but not boys(Lewis and Lockheed, 2007). Similarly, in rural

Uganda, crop losses led to sharp declines in girls’enrolment and performance in examinations, whilethe impact on boys was much smaller (Björkman,2005).

The current economic downturn, along withincreases in food and fuel prices, has increased thevulnerabilities that come with poverty. It is too earlyto establish the impact on education with anyaccuracy. One survey in Bangladesh found that thesharp rise in food prices in 2007 and 2008 hadforced half the poor households covered to removechildren from school as a cost-saving measure(Raihan, 2009). As Chapter 1 indicates, there havealso been reports of declines in enrolment andincreased absenteeism in other countries. In Kenyaand Zambia, for example, crisis-related poverty hasleft some children hungry and too weak to walk toschool. Dropout has increased due to inability tocover the costs of schooling and the need for childlabour (Hossain et al., 2009). More broadly, thereare grounds for concern that a combination ofsluggish economic growth, rising unemployment,falling remittances and slower poverty reduction

In 2007 and 2008the sharp rise in food prices forced half of poor householdssurveyed inBangladesh toremove childrenfrom school

Any international ranking of opportunity in educationwould place Hausa girls in northern Nigeria near thebottom of the scale. In 2003, half of primary schoolage girls in Kano state were out of school and inJigawa state the figure was 89%. Being poor andliving in a rural area compounds the disadvantage —in this category, over 90% of Hausa women aged 17to 22 have fewer than two years of education.

Northern states such as Jigawa, Kaduna and Kanoare among the poorest in Nigeria. There is evidencethat household deprivation hurts girls’ educationin particular, as poverty intersects with social andcultural practices, beliefs and attitudes.

Some parents attach limited value to girls’ education.As one research report put it, ‘from birth’ a girl ‘maybe considered as a costly guest in her own home.Her schooling is likely to be considered a waste oftime and money, and she is diligently trained to behome as a bearer of many children and a free sourceof labour’ (Rufa’i, 2006, p. 86).

Hausa girls who go to school tend to start late.Around one-quarter of girls aged 6 to 14 in schoolin Kaduna and Kano were over the usual age fortheir grade. To compound the problem, marriageat 14 or even younger is common and typically signalsthe end of education.

Northern Nigeria is predominantly Muslim. Manyparents send their daughters to Islamic schools outof distrust for formal public education, concern overthe quality of government schools or the distanceto them, or fear of sexual harassment in school oron the way there. Yet the quality of Islamic schoolingis highly variable — and the education many younggirls receive there is both limited and short-lived.

The experience of Hausa girls illustrates some ofthe wider challenges involved in reaching those onthe margins of education. There are public policymeasures that can make a difference, such as buildingclassrooms closer to communities, eliminatinginformal school fees, integrating Islamic schools thatmeet quality standards into the government systemand improving quality through better teacher training.But in northern Nigeria the most tenacious barriersto girls’ education are often embedded in parentaland community attitudes and gender practices.Removing those obstacles requires more equitableeducation policies, including wide-ranging incentivesfor girls’ education, backed by social and politicaldialogue to change attitudes.

Sources: Rufa’i (2006); Akyeampong et al. (2009); UNESCO-DME (2009).

Box 3.4: Hausa girls in northern Nigeria — losing out in education

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 6 8

will add to the pressures on the poorest households,with potentially damaging consequences foreducation (see Chapter 1).

Child labour remains a barrier to educationChild labour is a deeply entrenched obstacle toEducation for All. Household poverty forces millionsof children out of school and into paying jobs or –especially for young girls – domestic chores.

The International Labour Organization putthe number of child labourers aged 5 to 14at 166 million in 2004 (Hagemann et al., 2006).Not all child labourers are kept out of school.Most combine school and work, though often withdamaging effects on their education. Some workbecause their parents cannot afford to send themto school. Others work to help their families makeends meet or to provide labour in the home.Understanding the interplay between educationaldisadvantage and child labour is critical not onlyfor education, but also for child welfare and widernational poverty reduction efforts.

Child labour ranges in scope from young girlscollecting water and firewood with their mothersto young boys tending cattle and engaging in paidwork, and to more extreme and dangerous formsof work. The worst forms of child labour are adirect source of marginalization in education.Over half the children engaged in labour in 2004were in hazardous work, involving dangerousconditions, long hours or hazardous machinery(Blanco Allais and Quinn, 2009).6 Such children canbe seen every day scavenging for rubbish in Manila,working on building sites in New Delhi or sellingnewspapers at traffic junctions in Haiti. They arealso forced into more invisible forms of labour,such as involvement in sex work.

The degree to which children combine work andschool varies by country. There are no upper limitbenchmarks, but children working about thirtyhours a week or more are unlikely to attend school(Edmonds, 2007); (Box 3.5). Moreover, it cannot beassumed that ability to combine work and school isconducive to learning. Evidence from eleven LatinAmerican countries indicates that this is detrimentalto educational achievement (Gunnarsson et al.,2006). In each country, child labourers achievedsignificantly lower scores in language andmathematics tests in third and fourth grades,controlling for school and household characteristics.Even modest levels of child labour at early ageshad adverse consequences for cognitive abilities,

with regular work being most detrimental(Gunnarsson et al., 2006; Sánchez et al., 2009).

Poverty has a very direct bearing on patterns ofchild labour. Poorer children are more likely thanwealthier children to work outside the home andless likely to combine work with school (BlancoAllais and Quinn, 2009). In Zambia, children fromhouseholds in the lowest income quintile aremore likely not only to work, but also to facehazardous work conditions (UnderstandingChildren’s Work, 2009).

In urban areas, many child labourers live on streets,either with destitute parents or with other children.These children experience particularly stark formsof marginalization in education. One study coveringseven cities in Pakistan found that fewer than 5%of children living on streets had completed primaryeducation (Tufail, 2005). A survey in Bangladeshfound that only 8% of street children were inschool at the time of the survey and only 14%had completed third grade of primary school(Foundation for Research on Educational Planningand Development, 2003).

Child labour in rural areas is often less visible, butno less widespread or damaging. A 2007 survey ofchildren on cocoa plantations in Côte d’Ivoire andGhana documented striking examples of childrenapplying toxic pesticides, working in extreme heatand using dangerous implements. In Côte d’Ivoire,many children in cocoa production had beentrafficked from Burkina Faso and Mali as bondedlabourers (Payson Center for InternationalDevelopment, 2008). Côte d’Ivoire and Ghana haveintroduced laws aimed at curtailing the practice(World Cocoa Foundation, 2009), but theeffectiveness of national action and regionalcooperation remain of concern. More broadly,governments are often more adept at adoptingstatements against child labour than at addressingthe underlying causes of the problem.

Child labour often magnifies poverty-related genderdisadvantage. A common thread across manycountries with large gender disparities in educationis the disproportionately large share of thehousehold labour burden that young girls carry.In the Lao People’s Democratic Republic, for bothurban and rural populations, the average time spentin school falls with poverty and young girls in poorhouseholds spend less time in classrooms thanyoung boys. Poor rural girls spend just over twohours a day studying and five hours working, on

6. Hazardous child labouris defined by theInternational LabourOrganization as work indangerous or unhealthyconditions, or under poorsafety and healthstandards and workingarrangements, that couldresult in a child’s death,injury or illness.

Evidence fromLatin Americaindicates that

combining workwith schooling is

detrimental toeducational

achievement

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 6 9

average. Young boys spend slightly more time thangirls in remunerated employment, while young girlsspend more than twice as much time as boys onhousehold activities (King and van de Walle, 2007).The upshot is that young girls from the pooresthouseholds are less likely than boys to combineschool and work, and more likely to be out of school(Hallman et al., 2007).

Economic shocks can increase the impetus towardschild labour. Crop losses, sudden increases inhousehold health costs or parental unemploymentcan pull children out of school and push them intopaying jobs. In the Kagera region of the UnitedRepublic of Tanzania, transitory income shockscaused by crop losses were associated with a 30%increase in hours worked by children aged 7 to 15and a 20% fall in school attendance (Beegle et al.,2006). This example illustrates the interactionbetween vulnerability and disadvantage in education.Households with a limited coping capacity can beforced to compromise the long-term welfare ofchildren to secure short-term survival.

Child labour confronts policy-makers with wide-ranging challenges. Preventing educationalmarginalization by saving children from having towork requires not only more effective legislationbut also economic incentives aimed at keepingchildren in school.

Group-based disadvantages

Education for All is a principle rooted in the ideasof human rights and equal citizenship. It does notallow for distinctions based on ethnicity, race,language or culture. Yet these group-basedidentities are among the deepest fault lines ineducation. In many countries, children born toparents who are members of an ethnic or linguisticminority, a particular racial group or a low casteenter school with poor prospects of success andemerge with less education and lower achievementthan do children without these disadvantages.

The processes that lead to group-basedmarginalization do not lend themselves togeneralization but they include formal and informaldiscrimination, stigmatization and social exclusionlinked to social, economic and political powerrelationships. Many of these processes have deephistorical roots in slavery, dispossession orsubjugation. The experiences of the K’iche’ inGuatemala, Aboriginals in Australia, low-castepeople in India and Kurds in Turkey have evolved

from complex histories and are perpetuatedthrough disparate structures. Yet there are somesignificant common threads, with marginalizedgroups facing high levels of social discrimination,fewer employment opportunities, more limitedrights, and limited prospects for social andeconomic mobility. All too often their experiencein school reinforces and perpetuates theirmarginalization.

Child labour is the rule rather than the exception in Mali and Zambia. Manychildren in both countries work longer than the average adult in rich countries,with damaging implications for education. However, the consequences varyin scale and severity.

About half of 7- to 14-year-olds in Mali and Zambia were working in 2005,predominantly in rural areas. An alarmingly large proportion of these children— about 80% in both countries — were reported as involved in hazardous work.

Behind these comparable headline figures there were complex variationsbetween school and work. Whereas most working children in Zambiacombined the two activities, in Mali about a third of children were reported tobe just working and only around 20% combined school and work (Figure 3.23).The average time spent working helps explain the difference. Child labourersin Mali logged an average of thirty-seven hours working each week,compared with twenty-four hours in Zambia.

These working children have lower levels of school attendance at every age,especially in Mali. School attendance gaps are relatively small in Zambia upto age 13 or 14, again underlining the more marked trade-off between schooland work in Mali.

Why does child labour in Zambia seem more compatible with education?Some children in Mali — notably those with inflexible employment conditionssuch as those working as domestic labourers and in manufacturing that limitthe scope for combining school and work — appear to face particularlysevere disadvantages. Mali has more and deeper poverty, and greatergender disparities in education. School-related factors, including distanceto school, the duration of the school day and flexibility of the school calendar,could also be significant.

Source: Understanding Children’s Work (2009).

Box 3.5: Mali and Zambia — combining child labour and schooling

School, 41% Work and school, 35% Work,12%

Neither,12%Zambia

School, 29% Work, 32% Neither, 18%Work and school,21%Mali

Figure 3.23: Patterns of school and work vary.Children aged 7 to 14 by involvement in economic activity and schooling, Mali and Zambia, 2005

Note: Work does not include household chores.Sources: Understanding Children’s Work (2009), based on Mali National Child Labour Survey, 2005 and Zambia Labour Force Survey, 2005.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 7 0

Social deprivation and educationalmarginalizationGroup-based marginalization has multiplesources. Some, such as race, ethnicity andlanguage, are intimately tied up with the culturalidentity of the group in question and with theexperience of social discrimination. Other factorsare related to poverty, health status and widersocial circumstances. The borders between theseunderlying sources of disadvantage are blurred.For example, ethnicity and language are oftentwo sides of the same coin and ethnic or linguisticminorities may face higher levels of poverty.What is clear from the evidence set out in the firstpart of this chapter is that group identity is often anaspect of ‘multiple exclusion’ that has a significantbearing on participation and achievement ineducation (Lewis and Lockheed, 2007).

The situation of indigenous groups in LatinAmerica powerfully illustrates the multipledimensions of deprivation. Indigenous people,especially women and children, have less accessto basic health services. They are also more likelyto suffer from nutritional problems. In Ecuadorand Guatemala, about 60% of indigenous childrenunder 5 are malnourished – roughly twice thenational averages (Larrea and Montenegro Torres,2006; Shapiro, 2006). In Ecuador, non-indigenouswomen are three times as likely to receiveantenatal care and have a skilled attendant presentat birth (Larrea and Montenegro Torres, 2006).Being indigenous raises the probability of beingin poverty by between 11% and 30%, dependingon the country (Hall and Patrinos, 2006).

Poverty magnifies the barriers facing indigenouschildren, especially girls. In Guatemala, indigenousgirls from extremely poor households enrol inschool 1.2 years later than indigenous girls fromnon-poor households, on average, and are farmore likely to drop out. Among 7- to 12-year-olds,Mayan boys and girls are twice as likely as non-indigenous children to combine school and work.For non-enrolled indigenous females, lack ofmoney and housework are cited by parents asthe main reason for children being out of school(Hallman et al., 2007).

The experience of indigenous people in LatinAmerica also draws attention to the interactionbetween marginalization in education andemployment. Over the past decade, someindigenous people in Latin America have narrowedthe gap with the majority population in terms of

years in school. But gains in education haveenhanced their prospects for employment andhigher wages far less than for non-indigenouspeople, pointing to discrimination in labour markets(Hall and Patrinos, 2006). This helps explain whyprogress in reducing poverty among indigenouspeople has been slow despite expanded accessto education. The persistence of high levels ofhousehold poverty helps explain in turn why childlabour, a major cause of school dropout, hastended to fall more slowly among indigenouspeople than among non-indigenous people.

Australia provides a striking example of extrememarginalization amid high levels of overalldevelopment. The country consistently figures inthe top five on the United Nations DevelopmentProgramme’s Human Development Index. Yet in2001, it was estimated that Aboriginals and TorresStrait Islanders in Australia would rank around 103– below the Philippines and around the level ofViet Nam (Biddle and Mackay, 2009; Cooke et al.,2007). Social disadvantage on this scale inevitablyaffects what Aboriginal children achieve in school.

The marked racial divisions evident in the UnitedStates’s education system are also wrapped up insocial disparities. Gaps in learning achievementare evident early on. On average, African-Americanchildren register lower cognitive developmentlevels by the age of two (Fryer and Levitt, 2006);(Table 3.4). Part of the difference can be traceddirectly to poverty and to parental education. Othersignificant factors include the number of books inthe home and time spent reading (Ferguson, 2007).These disparities point to the importance ofconcerted pre-school strategies for overcominggroup disadvantage, as discussed in Chapter 2.

Similarly, the restricted opportunities experiencedby Roma children in school are intimately linkedto poverty, unemployment, poor housing and poorhealth. A survey has found that one-quarter of theRoma population in southern and eastern Europelives in dilapidated housing. The poverty rate forRoma in Romania is almost three times thenational average (UNICEF, 2007a). The invisibilityof Roma in national education programmesreinforces their exclusion: in Hungary, mosteducation policies do not mention Roma, thecountry’s most educationally disadvantagedcommunity (Open Society Institute, 2007).

High economic growth and rapid poverty reductiondo not automatically dissolve deeply entrenched

‘The ridicule we face prevents

us from comingto school and

sitting withhigher-caste

children.’Musahar girl,

India

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 7 1

group-based disadvantages. Since the early 1990s,poverty in Viet Nam has been cut by two-thirds,far surpassing the Millennium Development Goaltarget. Despite the gains, however, the averagepoverty rate among the country’s 10 million ethnicminority people is 52%, compared with 10% forthe majority Kinh (World Bank, 2009d). Minoritiesalso have worse health, nutrition and educationindicators, and less access to basic services.Partly because of these inequalities, the benefitsof rapid economic growth have trickled down moreslowly to ethnic minority groups. And the povertygap has widened over time. At the end of the1990s, the poverty rate among the non-Kinhpopulation was two and a half times higher thanthe average for Kinh. By 2006, it was five timeshigher (Baulch et al., 2009).

The wider social and economic inequalities drivinggroup-based marginalization in Viet Nam haveimportant consequences for education. Whileeducation figures for ethnic minority groups areimproving, they still lag far behind those of theKinh population. One-quarter of minority childrenenter school late, compared with 5% for Kinhchildren. Around 30% of minority householdsreport at least one child dropping out of primaryschool, double the Kinh share (World Bank, 2009d).Two of the four top reasons for dropping out –inability to afford school fees and need for childlabour at home – are directly related to poverty.

Low status and social identityLow status is intrinsic to marginalization. Inparts of South Asia, social practices relatingto group status are often based on complex ideasabout caste. While caste-based discriminationis frequently outlawed through legislation,underlying practices and attitudes are oftendifficult to change.

In India, the 1950 Constitution banned‘untouchability’ and provided measures tocompensate for the extreme social, education andeconomic disadvantage arising out of that status.Yet, despite progress in many areas includingeducation, deep caste-based disparities remain(Box 3.6). Belonging to a scheduled caste or tribelowers prospects of school attendance.7 Being agirl and living in a rural area brings a further layerof disadvantage. In 2004/2005, just 57% of ruralgirls aged 12 to 14 from scheduled tribes and 66%from scheduled castes were in school, comparedwith a national average of 80% (Figure 3.24).

11 25 13

42 30 37

71 56 64

37 28 33

24 21 23

Poverty rate (%)

Cognitive development2-year-olds: Per cent demonstratingproficiency in listening comprehension

2-year-olds: Per cent demonstratingproficiency in expressive vocabulary

4-year-olds: Per cent proficient at letterrecognition

4-year-olds: Average overallmathematics score

Table 3.4: Poverty and early cognitive development by race, United States

White

Sources:Poverty rate: Annual Social and Economic Supplement (ASEC) to the 2008 Current Population Survey (CPS),in DeNavas-Walt et al. (2008).Cognitive development: National Center for Education Statistics (2009). Data for 2-year-olds collected in 2003–04; for 4-year-olds in 2005–06.

AfricanAmerican

Nationalaverage

‘The higher-caste students tell us that we smell bad,’ one girl said. Anotheradded, ‘The ridicule we face prevents us from coming to school and sittingwith higher-caste children.’ These girls from the hamlet of Khalispur, nearthe city of Varanasi, belong to the Musahar or ‘rat catcher’ communityof eastern Uttar Pradesh, India.

Khalispur has a government primary school. Despite an entitlement toreceive a stipend, midday meals and uniforms, few Musahar girls attend.The testimony of some of them powerfully demonstrates the force of social attitudes in creating disadvantage: for these girls, school is a placewhere they experience social exclusion, as stigmatization undermines the self-esteem vital to effective learning. Subtle forms of discriminationreinforce caste hierarchies in the classroom. ‘We are forced to sit on the floor,’ one girl said. ‘The desks and benches in the classroom are meant for the children from the higher castes.’

The Musahar community, which spans eastern Uttar Pradesh and Bihar,has high levels of poverty and low levels of literacy among adults. Apartfrom catching rats in rice fields, the livelihoods of the Musahar typicallyrevolve around crushing and carrying stones, supplying brick kilns, makingleaf plates and performing casual day labour. In contrast to some other low-caste groups, the Musahar have a weak political voice.

According to Musahar elders, government policies have improved butsocial attitudes have not: ‘They do admit our children to school and wenow have legal rights, but the behaviour of children from other castesand the teachers is a problem. Our children do not dare attend the school.’

Interviews courtesy of Sudhanshu Joshi, Global March Against Child Labour

Box 3.6: Living with stigma — the ‘rat catchers’ of Uttar Pradesh and Bihar

7. Scheduled castes are the former untouchables and scheduled tribes are India’s indigenous populations. Both are listed in schedules appended to India’s constitution as groups deserving affirmative action measures.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 7 2

To what extent do these differences stem fromdistinctive caste and tribe disadvantages ratherthan wider social and economic factors? That is akey question for policy-makers seeking to equalizeopportunity. Research for this Report helps providea partial answer (Bhalotra, 2009). Using householdsurvey data, and controlling for household andindividual characteristics, the study found thatabout 60% of the attendance gap for scheduled-caste children aged 6 to 14 could be attributed tohousehold characteristics, mainly poverty and lowerparental education. For scheduled-tribe childrenin the same age group, household characteristicsweighed less heavily, accounting for about 40% ofthe attendance gap. One conclusion to be drawn formembers of both scheduled groups is that povertymatters a great deal in perpetuating educationaldisadvantage. However, the non-poverty componentis larger for scheduled tribes partly because of theweight of social and cultural discrimination.

Public attitudes have consequences that go beyondschool attendance. Institutionalized stigmatizationcan erode self-confidence and levels of expectation,undermining children’s potential for learning.

One particularly striking illustration comes from anexperimental investigation into the impact of casteperceptions on test scores (Hoff and Pandey, 2004).Children aged 11 and 12 were chosen at randomfrom a low caste and three high castes, and givena series of puzzles to solve. When caste was notannounced to the participants, it had no bearingon the initial score or on the improvement in scoreregistered in subsequent test rounds. But whencaste was announced before the test, the scores forlow-caste children fell dramatically (Figure 3.25).8These findings underline the degree to whichsocial identities that are a product of history,culture and personal experience can createpronounced education disadvantages throughtheir effects on individual expectations.

The critical role of languageLanguage and ethnicity are deeply intertwined.Having a distinctive language is often a crucialelement of personal identity and group attachment.Just as a local language may be a point ofassociation for members of an ethnic group, itcan also be an element in their marginalization.People who cannot speak a country’s dominantlanguage may have less access to written andspoken sources, restricting their opportunitiesfor employment and social mobility (Smits andGündüz-Hosgör, 2003; Smits et al., 2008). Parentswho do not speak the official language in whichtheir children are being educated may have lessopportunity to engage with teachers, educationauthorities and homework. And their childrenmay not grasp what is being taught if teachersdo not speak their home language. The resultinginequalities in opportunity are a major factor

8. In three test rounds,scores for low-castechildren fell by 14%, 25% and 39%.

National average

Scheduled caste girls, rural

Scheduled tribe

Scheduled tribe girls, rural

Scheduled caste

20

40

60

80

100

Scho

ol a

ttend

ance

rate

(%)

6–11 12–14 15–18

Age group

Figure 3.24: In India, scheduled castes and tribes remaindisadvantaged at all levels in educationAttendance rates by age group in India, by community, rural/urban residenceand gender, 2004/2005

Notes: The attendance rate for an age range is the proportion of children of that age range who report attending school at the time of the survey. The age ranges correspondapproximately to primary education, upper primary (or ‘middle’) education and secondaryeducation, respectively, in the Indian school system.Source: Bhalotra (2009) based on National Sample Survey data (61st round).

Caste not announced Caste announced0

1

2

3

4

High caste

Low caste

Aver

age

num

ber o

f puz

zles s

olve

d

Figure 3.25: Social stigma can undermine test performanceExperimental impact of the announcement of caste on solvingpuzzles in India

Note: Children aged 11 and 12 were given a packet of fifteen maze puzzles and asked to solve as many as they could in fifteen minutes.Source: Hoff and Pandey (2004).

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 7 3

in marginalization in countries where ethnicityand language are strongly associated with socialdeprivation.

The sheer scale of linguistic diversity in the worldtoday and its consequences for achievement ineducation are not sufficiently recognized. Thereare nearly 7,000 spoken languages. Every worldregion is multilingual. Sub-Saharan Africa has1,200 to 2,000 languages (Alidou et al., 2006).Cameroon alone has more than 200 languages,of which thirty-eight are written. In East Asia,Thailand has over seventy and Indonesia morethan 737. Latin America’s indigenous peoplesspeak an estimated 551 languages (Dutcher, 2004).

Education systems seldom reflect linguisticdiversity. Many countries stress the importanceof children learning in their mother tongue orhome language. Nevertheless, about 221 millionschool age children speak languages that areused at home but not recognized in schoolsor official settings (Dutcher, 2004).

The degree of alignment between home andschool language has a critical bearing on learningopportunities. Children who study in their mothertongue usually learn better and faster than childrenstudying in second languages (UNESCO Bangkok,2008; Woldemikael, 2003). Pupils who start learningin their home language also perform better in teststaken in the official language of instruction later intheir school careers (UNESCO Bangkok, 2008). Thebenefits extend beyond cognitive skills to enhancedself-confidence, self-esteem and classroomparticipation (Alidou et al., 2006).

Decades of cognitive research have establishedthe language conditions most conducive tolearning (Alidou et al., 2006; Dutcher, 2004;UNESCO Bangkok, 2008). Translating thosefindings into policies that create an enablingenvironment for ethnic and linguistic minoritiesis not straightforward. Linguistic diversity createschallenges within the education system, notablyin areas such as teacher recruitment, curriculumdevelopment and the provision of teachingmaterials. Moreover, language policy in educationis not just about learning but is intimately wrappedup in power relationships and history.

In many countries, the dominant languages usedin education are connected with social, political andcultural subjugation. Colonization has left a deepimprint. For most pupils entering primary school

in francophone Africa, French is still their firstlanguage of instruction (Alidou et al., 2006). Duringthe 1880s, authorities in New Zealand banned theteaching or use of the Ma-ori language in nativeschools, arguing that it was an impediment to‘national progress’. One hundred years later, thelanguage was spoken by less than one-quarter ofthe Ma-ori population and drifting towards extinction(Wurm, 1991). Across much of Latin America,language was key to the exclusion and exploitationof indigenous people by Spanish-mestizo elites(Klein, 2003). Indigenous organizations in the regionhave seen ‘decolonization of the school’ as a vitalpart of wider political emancipation.

Governments have often seen the forging of a common linguistic identity as crucial to thedevelopment of a national identity (Daftary and Grin,2003). The Turkish Constitution of 1923 includes aprovision that ‘no language other than Turkish shallbe taught as mother tongue to Turkish citizens atany institutions of teaching or education’ (Kaya,2009, p. 8). While legislation adopted in 2002 allowsgreater flexibility, access to minority languageprimary education remains limited.

Language policy in education raises complexissues and potential tensions between groupidentity on the one hand, and social and economicaspirations on the other. Parents in many countriesexpress a strong preference for their children tolearn in the official language, principally becausethis is seen as a route to enhanced prospects forsocial mobility (Alidou et al., 2006; Cueto et al.,2009; Linehan, 2004). Labour market factors oftenfigure prominently. In response to changing jobopportunities and the earnings premium associatedwith use of English, lower-caste girls and youngwomen in Mumbai are switching from primaryand secondary schools teaching in Marathi to thoseteaching in English (Munshi and Rosenzweig, 2003).

Education systems have to perform a delicatebalancing act. First and foremost, they need tocreate the enabling conditions for effective learning.Ideally, this implies learning the official languageas a subject in primary school while receivinginstruction in the home language. It also impliesa school curriculum that teaches the majoritypopulation respect for ethnic minority languageand culture. But education systems also haveto ensure that children from disadvantagedminority backgrounds learn the skills they needto participate successfully in social and economiclife, including language skills.

About 221 millionschool agechildren speaklanguages at homethat are notrecognized inschools or officialsettings

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 7 4

Breaking long-established institutional patternsis difficult. This is true even in countries withgovernments that acknowledge the disadvantagesethnic minorities face, as the experience ofViet Nam shows (Box 3.7). In Latin America, mostcountries have intercultural bilingual educationpolicies, some of them dating from the 1920s.Today, such programmes aim at incorporatingindigenous languages into national educationsystems by giving children a chance to learn intheir home language before moving on to Spanish.Despite some significant achievements, however,these programmes face major challenges inseveral countries:

Limited coverage. Many indigenous childrendo not have access to intercultural bilingualeducation. In Guatemala and Paraguay,legislation provides for bilingual education injust the first three grades of primary schooland, in reality, children are often taught onlyin Spanish. In Guatemala, 74% of children aged 7 to 12 years were reported as receivingclasses only in the Spanish language in 2006(López, 2009). In Peru, only around 10% ofindigenous children attend interculturalbilingual schools. Coverage is far lower inurban than in rural areas (Cueto et al., 2009).

Poor quality. Where indigenous languageteaching is available, it is often of poor quality,with schooling compounding disadvantageslinked to social and economic deprivation.Of about 900 teachers working in indigenous

communities in Paraguay, a third have completedonly basic education and fewer than two-thirdsreport speaking the local language (López, 2009).In Peru, which has been implementing theintercultural and bilingual model since 1972,one study in the south of the country found thathalf of teachers in intercultural and bilingualeducation schools could not even speak thelocal indigenous language. Moreover, bilingualmaterials provided by the Ministry of Educationwere not being used (Cueto et al., 2009).

Limited scope. Intercultural bilingual educationfocuses on more effective integration ofindigenous children into mainstream education.For many indigenous groups this objective istoo limited. In several countries, indigenouspolitical movements have mobilized behinddemands for education reforms and forcurriculum content that focuses on widerpolitical concerns. In Bolivia, indigenouseducation councils have been pressing for a neweducation law that emphasizes multiculturalism,ethnic diversity and the values of indigenousculture. In Guatemala, where indigenouspeople’s rights were brutally suppressed duringthe civil war, the period since the Peace Accordsin 1995 has been marked by the development ofa vigorous Mayan political movement focusing onlanguage as one element in a broader campaignagainst discrimination. In both countries, manyindigenous political leaders are looking tostrengthen intercultural education to addressdeeper problems of discrimination and inequality,

In parts of Peru,many teachers inbilingual schools

can not speak thelocal indigenous

language

The government of Viet Nam recognizes thatproblems facing ethnic minorities are a major barrierto universal primary education. It has established anextensive system of financial transfers targeted athouseholds and communes with large minoritypopulations. A 1999 law allowing minority languagesto be used in education recognizes the importanceof home language.

Implementing that law has proved difficult, however.Part of the problem is a serious shortage of ethnicminority teachers. While ethnic minority childrenaccount for 18% of the primary school age population,ethnic minority teachers make up just 8% of theteaching force. Moreover, few of these teachers areposted to ethnic minority areas. And not all have the

training or experience to teach bilingual education.As a result, Kinh remains the dominant language of instruction for most ethnic minority children.

Demographic factors also appear to have animportant bearing on the language of instruction.Analysis undertaken for this Report compared homelanguage education in Lao Cai, a mountainousnorthern province with a large ethnic minoritypopulation, with that in Phu Yen, a south-centralcoastal province in which ethnic minorities account for just 5% of the population. Minority groups in Phu Yen have far less access to home languagecourses, partly because their children attendoverwhelmingly Kinh-dominated schools.

Sources: Truong Huyen (2009); World Bank (2009d); UNESCO-DME (2009).

Box 3.7: Tackling the ethnic divide in Viet Nam

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 7 5

and to help change power relationships insociety (López, 2009; Luykx and López, 2007).

Location and livelihoods

Disadvantages linked to poverty, ethnicity andlanguage are often reflected in human geography.Children living in slums, remote rural areas orconflict-affected zones are typically among thepoorest and most vulnerable in any society.Potentially, they have the most to gain fromeducation. Yet they live in areas with the mostlimited access to basic services, includingeducation. Restricted education and livelihoodopportunities reinforce the poverty trap. Thissection looks at institutionalized disadvantageslinked to location that perpetuate marginalizationin education.

Right to education denied to slum dwellersKibera is one of the largest slums in sub-SaharanAfrica. Located next to the Royal Nairobi GolfCourse and a short distance from leafy suburbsthat are home to some of Kenya’s wealthiestpeople, it has an estimated population of 1 million.Most lack access to clean water, sanitation andother public services. It is a short walk fromKibera to some of Kenya’s finest primary schools,yet the vast majority of the slum’s children arelocked out of even the most basic opportunitiesfor education.

Kibera is a microcosm of a wider problem.Half the world’s population now lives in citiesand urban growth is highest in the developingworld (UN-HABITAT, 2008). In the midst ofurban prosperity and opportunity, almost everymajor city has large islands of slums that arecentres of social deprivation. On one estimate,one in three urban dwellers in the developingworld – 900 million in total – resides in a slum (UN-HABITAT, 2006). In an increasinglyurbanized world, slum populations are growingby over 20 million a year as rural poverty andthe lure of opportunity create a steady streamof new arrivals.

Not all slum environments are equivalent in thescale of deprivation. One study comparing slumsin Nairobi and Dakar, Senegal, found that whilethe inhabitants of the latter were poorer, theywere four times more likely to have access towater and electricity. Just under a third ofNairobi’s population lives in slums. Children inthese settlements face disadvantages at many

levels. Less than 6% of households have pipedwater in their homes and even fewer have accessto sanitation facilities. Poor sanitation andinadequate garbage collection cause majorhealth problems. Children in Nairobi’s slumsface higher mortality rates than those in ruralareas (World Bank, 2009f).

Many governments have little idea how manychildren live in informal settlements and arefailing to respond to the major new educationpolicy challenges created by the rapid growthof slums. Because many settlements are ‘illegal’,they are not recognized in government plansor provided with public water, sanitation, healthor education services (UNESCO-IIEP, 2009).

What schooling is available is often suppliedby non-government organizations, churches orprivate entrepreneurs, with little governmentsupport or regulation. As evidence from slumsin Dhaka, Bangladesh, shows, the poor generallyhave little if any choice of education provider(Box 3.8). The financing of education in slumssuch as those in Nairobi is largely private: parentshave to pay for poor-quality private schooling,while non-slum children have access to freegovernment education (Oketch et al., 2008).Household poverty, poor child health and nutrition,and extensive child labour combine to createa formidable barrier to education. Even whereschools are not far away, security concernspresent an additional hurdle to access: 60%of girls interviewed in Kibera expressed fearof being raped. It was not uncommon for boysand girls to have witnessed physical violence.A common response to fear of violence andharassment in slums is to stop going to school(Erulkar and Matheka, 2007; Mudege et al., 2008).

Restricted entitlements are among the mostpotent elements of educational marginalizationin slums. Parents often cannot secure theirchildren’s human right to education becausethey lack official residency status. For purposesof school registration, the authorities do notrecognize that these children even exist. Onestudy of 400 slum-dwelling households in Delhifound that only half of primary school agechildren were in school, compared with a citywideenrolment rate in excess of 90% (Tsujita, 2009).Although government schools were withinwalking distance, only a third of children inthe sample had a birth certificate, which ismandatory for admission to government schools.

In Delhi’s slums,only a third ofchildren surveyedhad a birthcertificate, which is mandatory for admission to governmentschools

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 7 6

Residency requirements were another majorbarrier, as migrants from other states make upa large share of the slum-dwelling population.Rural migrants to urban areas in China facesimilar problems, with the hukou (householdregistration) system restricting access to basiceducation (Box 3.9).

Many governments lack credible public policiesfor providing education and other basic servicesin fast-growing informal settlements. Authoritiesoften claim that legal entitlement to educationand other services in all slums would act asincentives for accelerated rural-urban migration.While this concern is not without foundation,depriving children of their right to an educationthrough government inaction is not anappropriate response.

Remote rural areas are underservedRural children face heightened risks ofmarginalization in education, especially if they arepoor and female. Rural-urban divides in educationoften overlap with wider inequalities. In manycountries, rural areas tend to have higherconcentrations of poverty and less access tohealth care. Marginalization in education bothmirrors and magnifies these disparities.

Low population density in rural areas oftenmeans children have to travel greater distancesto school, sometimes across difficult terrain. Inaddition, rural parents tend to be less educated.These concerns are compounded by governmentfailure to provide schools or attract good teachersto the countryside. Traditional cultural practicesand attitudes also play a role.

‘Do we buy food or enrol

our child in school?’

Parent, Bangladesh

Dhaka, the capital of Bangladesh, is one of the world’s fastest-growing cities. An estimated 300,000to 400,000 new migrants arrive each year and thevast majority head for informal settlements. Aroundone-third of the city’s population — 4 million people —live in slums, many of them in flood zones.

Education figures for Dhaka’s slums are among theworst in Bangladesh. One study of four slums foundthat just 70% of children were enrolled at the primarylevel, many of them in schools run by non-governmentorganizations.

The study also found high inequality within the slums.The children of better-off families, such as those withsmall businesses, were far more likely to be not just in school, but in a government or private school.Children from the poorest households were less likelyto be in school and almost half those enrolled reliedon non-government organizations (Figure 3.26). The parents of these children were predominantly day labourers and rickshaw drivers.

Schools run by non-government organizations play an important function in Dhaka’s slums. Unlike privateschools, they are usually free and offer flexible hoursand classes. Their quality varies, however, and manyoffer only three or four years of basic education, withlimited scope for transition into the formal educationsystem. In some respects non-government educationis a symptom of the vacuum created by limited publiceducation. Only a quarter of Dhaka’s slums have a government school. Most of these schools are in well-established slum areas, while newer, less formal settlements are left to fend for themselves.

As in other slums worldwide, insecure tenurecontributes to marginalization in education. Lackingtenancy rights, slum dwellers are in a weak position to demand education and public finance. Moreover, as many city authorities periodically bulldoze informalsettlements, some non-government providers areloath to invest in school buildings.

Sources: Cameron (2009); World Bank (2007a); Centre for UrbanStudies et al. (2006).

Figure 3.26: Poor slum dwellers in Bangladesh depend on non-government education provision% of children aged 6 to 11 enrolled, by type of school and wealth,selected slums of Dhaka, 2008

0%

20%

40%

60%

80%

100%

Not in school

Madrasa/other

Private

NGO

Government

Poorest Middle Richest

Wealth quintiles

Box 3.8: Slums in Dhaka — marginalization with rapid urban growth

Note: Includes only ages 6 to 11 and enrolment in grades 1 to 5.Source: Cameron (2009).

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 7 7

Remoteness is one of the strongest factors inmarginalization. The poorest households in manyrural areas are the furthest from roads, markets,health services and schools. In Nicaragua, theincidence of extreme poverty is 20% higher in thecentral rural region, where people have to traveltwice as far as the national average to reacha school or health clinic (Ahmed et al., 2007).

Distance to school is often a major determinant ofparticipation by ethnic minorities. In India, childrenfrom scheduled tribes, many of them living indispersed communities in remote areas, facesome of the longest treks to school in the country(Wu et al., 2007). In the Lao People’s DemocraticRepublic, schools in rural and predominantly non-Lao Tai areas are less likely to offer a fullprimary education cycle, and the availability oflower secondary schools is far more restrictedcompared with Lao Tai areas. Only 80% of ruralnon-Lao Tai children have a primary school in theirvillage and only 4% have a lower secondary school.The shares for the majority Lao Tai children aresignificantly higher (88% and 17%, respectively).Such differences help explain why only 46% ofpoor non-Lao Tai girls aged 6 to 12 attend schoolin rural areas, compared with 70% of poor ruralLao Tai girls (King and van de Walle, 2007).

Lack of nearby facilities has implications for boththe time and the energy needed to get to school.Country surveys in West Africa from the 1990srevealed high average walking distances in severalcountries, including 7.5 km in Chad, 6.6 km in Mali,5 km in Senegal and 4 km in the Central AfricanRepublic. Distances are likely to be higher thanthese averages in remote areas (Filmer, 2004).

Even relatively short distances to schoolcan significantly reduce demand for education.A 2002–2003 survey of 179 villages in thewestern Sahelian region of Chad found that fordistances over a kilometre, enrolment declinedsteeply, with fewer than 10% of children typicallygoing to school. Physical barriers such as riversand forests could considerably increase the timerequired to reach school (Lehman et al., 2007).

Girls’ attendance is particularly sensitiveto journey times. Household surveys in manycountries identify distance as a major factorin parents’ decisions to keep daughters outof school (Kane, 2004, and World Bank, 2005d,cited in Theunynck, 2009; Glick, 2008; Huismanand Smits, 2009). Explanations vary, butconcerns over security and domestic labourneeds figure prominently.

In the westernSahelian region of Chad, enrolmentdeclines steeplywhen children liveover a kilometrefrom school

In China, children’s right to education can run upagainst residency requirements that limit access to schooling.

The full extent of rural-urban migration in China is unknown. One estimate is that 98 million ruralmigrants live in China’s cities, including 14 millionchildren. Attracted by employment and an escapefrom rural poverty, many migrants live in informaltemporary housing in areas with limited publicservices. Migrant children are among the mosteducationally marginalized in China, largely because of the registration system called hukou.

Under the hukou system, city schools can only admit students registered as official inhabitants with a permanent home in the school district. School budgets are based on the number of officialstudents registered by authorities. Individual schoolscan admit unregistered children, but typically require parents to pay a fee to compensate for thelack of government funds. This arrangement makeseducation unaffordable for many migrant families.

Education figures for major cities reflect theconsequences of the hukou system. Only two-thirds of Beijing’s 370,000 migrant children were enrolled in public schools. Another quarter were reported as attending unauthorized migrant schools. Theseschools, a response to exclusion from the publiceducation system, are of questionable quality and some have been forced to close.

Chinese authorities, acknowledging the problemsfacing rural migrants, have introduced reforms. City authorities have been required to accommodateholders of rural hukou with temporary residence and employment permits, reducing the pressure on schools to charge fees. Even so, the children of many migrants, including those working in the informal sector, continue to face restrictedopportunities for education.

Sources: Han (2009); Liu, He and Wu (2008); Liang et al. (2008).

Box 3.9: China’s hukou system has restricted education opportunities for migrant children

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 7 8

Education systems themselves are oftenunresponsive to pastoralist demands. Pastoralistlivelihoods are inherently mobile, geared towardsseasonal calendars for grazing and wateravailability. Formal education planning, in contrast,is commonly organized around a fixed schoolinfrastructure and a fixed national schedule forthe school term and school day. Such planningfails to take into account the realities and demandsof pastoral livelihoods. The misalignment betweenthe education supply model and livelihood realitiesmeans demands for schooling are often unmet.

School infrastructure is not the only problem.Pastoralists often see curricula as having littlerelevance to their lives. They are typically absentfrom the images and stories in primary schooltextbooks, reinforcing the cultural distancebetween home and school. If pastoralism ismentioned at all, it may reflect the view of manynon-pastoralists that the practice is outdated andignorant (Krätli, 2006), rather than a specializedand sustainable livelihood.

Early marriage for girls is another barrier toeducation in some pastoralist communities. Sois a deeply engrained belief that female educationmay be of less value. A proverb of the Gabracommunity in northern Kenya says: ‘God first,then man, then camel, and lastly girl.’ This

‘The educationsystem that fits

us will be the onethat follows us,that follows our

animals.’Village elder,

northen Kenya

Nasra Hassan, 7, has had a taste of education. She was enrolled in standard one at Basaa Primary Schoolin the Merti Division of Isiolo, a remote district ofNorthern Kenya. But then the drought hit. The currentdrought has left an estimated 4 million Kenyans inneed of emergency food aid. Pastoralist areas havebeen among the worst affected. Child malnutrition is rising and households have seen their livestockherds decimated.

The harm to education has been less visible — but noless damaging for long-term efforts to reduce poverty.Nasra’s parents no longer have the money they needto pay for her education. And as herders have totravel farther and farther in search of water for theiranimals, there are fewer people at home to help withhousehold chores, so Nasra is expected to spend moretime looking after the smaller animals and collectingwater for home use. Instead of studying, she is nowbusy washing, cooking, and fetching water andfirewood. The drought has forced her out of school.

The drought is not the only barrier to educationamong pastoralist children. Many parents and villageelders have ambivalent attitudes to schooling, partlybecause they are acutely aware of the trade-offs theyface. As one parent eloquently put it, ‘We have tochoose between wealth and knowledge — betweenhaving a prosperous herd and having educatedchildren. We need our children to tend the cattle, even though we know they need an education.’

The tension between securing livelihoods and gainingeducation is a recurrent theme in pastoral areas.Formal education happens in a fixed context — theclassroom. By contrast, pastoralist survival oftendepends on children following herds over large areas.

Resolving the dilemma will require more flexible andmore mobile ways to provide education. As one villageelder in Isiolo said, ‘The education system that fits uswill be the one that follows us, that follows our animals.’

Interview courtesy of SOS Sahel

Box 3.10: Kenya’s pastoralists — ‘we need schools that follow our herds’

Pastoralist lifestyles demand better education responsesPastoralists in sub-Saharan Africa and South Asiaface extreme educational disadvantage (see Figure 3.5above; Dyer, 2006). By one rough estimate, as manyas 8.5 million children from nomadic householdsdo not attend school globally (Carr-Hill, 2009).

Why do pastoralist children face such restrictedopportunities for schooling? Livelihood pressuresare an important factor. Pastoralists are not alwaysthe poorest rural people, especially if their livestockassets are taken into account. But they often relyheavily on boys for tending cattle and girls fordomestic chores, restricting children’s timeavailable for formal schooling (Ruto et al., 2009)(Box 3.10). Education loses out because labourdemands take priority.

On the other hand, pastoralists often seeeducation as a route to more diverse and lessinsecure livelihoods. This finding emerges fromresearch in the Somali region of Ethiopia andamong the Turkana of Kenya and the Karamojongof Uganda (Devereux, 2006; Krätli, 2006; Ruto et al.,2009). Paradoxically, environmental degradation,drought and cattle raids may be stimulatinginterest in the role formal education can play inproviding skills needed to cope with contemporarylivelihood challenges (Dyer, 2006).

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 7 9

explains a reluctance to sell camels to finance girls’education, unlike for boys (Ruto et al., 2009, p.11).The social attitudes behind such sentiments aredeeply damaging for girls’ education.

The diversity of pastoralist experience cautionsagainst generalization. Yet even in countries makingstrong progress in primary education, pastoralistchildren are often being left far behind. Kenya isnow looking beyond primary schooling to universalsecondary education, but that vision contrastsstrongly with reality in the country’s ten most ariddistricts. Inhabited predominantly by pastoralistcommunities, these districts have some of thecountry’s lowest enrolment ratios and largestgender disparities, with net enrolments lessthan 30% for boys and 20% for girls in the threeworst-performing districts located in the NorthEastern Province (Figure 3.27).

Armed conflict fuels educationalmarginalizationArmed conflict contributes to marginalization ineducation in many ways. Most obviously, it exposeschildren to the risk of violence and trauma. Inaddition to driving people from their homes andcreating large refugee populations, conflict candestroy schools and create risks for pupils andteachers. Moreover, conflict can leave a legacyof distrust, instability and weak governance foundin many of the world’s most fragile states, withgovernments often unable or unwilling to providebasic services.

While firm evidence of the impact of armed conflictis limited, international data clearly reflect a closeassociation between conflict and marginalization.Over one-third of primary school age children whodo not attend school – 25 million in total – live inconflict-affected poor countries (see Chapter 2).Many of these countries have among the world’sworst child health and education figures. InSomalia, one in seven children does not surviveto age 5 and just 22% of those who do reachprimary school age are in school – one of theworld’s lowest enrolment levels (UNDP, 2009).

Mass displacement caused by conflict locksmillions of children into a future of extremedisadvantage in education. Forced to flee theirhomes, parents often have to resettle in areas illequipped to provide good basic education. At theend of 2008, there were an estimated 42 millionforcibly displaced people worldwide: around26 million were displaced within their own

countries and 16 million had to flee across borders(Internal Displacement Monitoring Centre, 2009;UNHCR, 2009). Children aged 5 to 17 comprisearound one-third of the global population of forciblydisplaced people (Women’s Commission forRefugee Women and Children, 2004). Indigenouspeoples and ethnic minorities make up adisproportionate share of displaced populations.

International debates on refugees often focuson issues affecting rich countries. Yet developingcountries bear the brunt of cross-borderdisplacement. Countries including Chad, Kenya,Uganda and the United Republic of Tanzania haveabsorbed millions of people displaced by conflicts inthe Democratic Republic of the Congo, Somalia andthe Sudan. Pakistan is host to the world’s largestrefugee population, having absorbed over 2 millionpeople uprooted by violence in Afghanistan(UNHCR, 2009; Winthrop, 2009a). Struggling toachieve universal primary education for their ownchildren, these countries are ill equipped to provideeducation to large, vulnerable, extremely poorrefugee populations that often speak differentlanguages. The international aid system offers onlylimited support. Children end up either studyinga curriculum that is alien to them or with noschooling at all. In Pakistan, a refugee censusin 2005 estimated that 1 million Afghan refugeechildren were out of school (Winthrop, 2009a).

Internal displacement can also create wide-rangingproblems for education, overloading the system inareas of resettlement. Pakistan’s recent experience

In Pakistan, a refugee censusin 2005 estimatedthat 1 millionAfghan refugeechildren were out of school

Figure 3.27: Many of Kenya’s arid districts are left behindNet enrolment ratios in public primary schools for northern arid districts of Kenya, 2007

National average

Moyale

Isiolo

Marsabit

Tana River

Samburu

Turkana

Mandera

Ijara

Garissa

Wajir

0 20 40 60 80 100

Public primary school net enrolment ratios (%)

North Eastern Province

MaleFemale

Source: Ruto et al. (2009), based on 2007 data from Ministry of Education Statistics Unit (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 8 0

again illustrates the scale of the problem.With 2.5 million people displaced from the NorthWest Frontier Province in 2008 by fighting betweenthe government and Taliban militants, schools inother parts of the country came under pressure(Winthrop, 2009b). In the Philippines, hostilitiesin 2008 and 2009 between government forcesand armed groups led to the displacement of750,000 people, severely disrupting children’sschooling (Amnesty International, 2009) (Box 3.11).

Violent conflict can touch the lives of childrenin many ways, including enforced recruitmentas soldiers. At the end of 2007, child soldierswere directly involved in armed confrontations

in seventeen countries, including Afghanistan,Chad, Somalia and the Sudan (Coalition to Stopthe Use of Child Soldiers, 2008). In Somalia, theTransitional Federal Government has reportedlyrecruited over 1,000 children into its armed forces,most of them directly from schools (UN GeneralAssembly Security Council, 2009).

Apart from missing out on education, childsoldiers often suffer psychological trauma,hampering prospects for a return to education.During the civil war in Sierra Leone that startedin 1991, over 15,000 children are estimated tohave been forced to serve in military groups.After the end of the conflict in 2002, schoolingwas seen as a way for the former soldiers torecover some of their lost childhood. However,schools were ill-equipped to provide thepsychosocial support necessary to enable themto readjust to normal life (Betancourt et al., 2008).

Other children experience trauma as a resultof being part of a civilian population caught inviolent conflict. The process of reconstructingeducation in Gaza will require not only repairingphysical infrastructure but also measures tosupport traumatized children (Box 3.12).

In some cases, education is targeted as asymbol of government authority, with schoolssubject to armed attack, and pupils and teachersthreatened with murder, injury, abduction andrape. In Afghanistan, 670 schools were closed inearly 2009 because of security threats, depriving170,000 children of education. In the threesouthernmost provinces of Thailand, separatistgroups hostile to Buddhist values and Thai-language teaching have attacked schools. Inthe past five years, 99 teachers have beenreported killed and 296 schools have beenfirebombed (O’Malley, 2009).

Groups within the Taliban in Afghanistan andPakistan have targeted girls’ schools, both tochallenge government authority and to assertvalues hostile to equal opportunity in education.In the Swat district of Pakistan, the Talibandestroyed 108 girls’ schools and damaged 64 otherschools between 2007 and May 2009. During 2008,local Taliban leaders ordered a ban on womenteachers and girls’ education. In response,900 schools closed or stopped admitting girls andfear created by the decree led to the withdrawalof 120,000 girls from school (O’Malley, 2009).

‘Students areoften absent

because theyspend hourslining up forrations and

water.’Abdul,

Philippines

Muhammed’s new home is a tent on the grounds of a school, yet he has little time to attend class. For him and many otherchildren in an evacuation camp, helping his parents supplementmeagre food rations is now his priority. ‘I can only go to classesin the morning because I have to look for vegetables andfirewood outside the camp and return before dark,’ he said.

Muhammed, 13, is the eldest of five children who are takingrefuge with their parents and grandparents in a camp set up in the Datu Gumbay Piang Elementary School in Maguindanao.

Heavy clashes between the military and separatist rebels in the Mindanao region of the Philippines have left hundreds ofthousands of civilians stranded in evacuation camps, often set up in schools such as this one. The Datu Gumbay Piang centrehas reportedly become home to the highest number of internallydisplaced persons since the outbreak of the fighting.

For the moment Muhammed and his family consider themselveslucky to have a tent to live in. ‘Some of the refugees have nochoice but to make their homes inside the classrooms or takeshelter under the school buildings when it rains,’ said BernieAbdul, an evacuee working in the school.

Most of the children come to class to escape the dismal livingconditions in their tents. But there is no immediate escape fromthe destruction and violence they have witnessed. ‘When thechildren are in class, they are either lethargic or very nervousbecause we often hear howitzers being fired not far from us.’

Muhammed is not the only child in the camp who is unable toattend school regularly. Abdul explained: ‘Students are oftenabsent because they spend hours lining up for rations and waterat the pump or because they’re sick. Living in an unhealthyenvironment without running water and sanitary facilities hasaffected the children physically and emotionally as well.’

Interviews conducted by Ross Harper Alonso for this Report

Box 3.11: The human face of conflict in the Philippines

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 8 1

Disability

Disability is one of the least visible but most potentfactors in educational marginalization. Beyondthe immediate health-related effects, physicaland mental impairment carries a stigma that isoften a basis for exclusion from society and school.The impact is often worse for poorer households.

Attitudes towards disability have changed overtime. Until relatively recently, the ‘medical model’was dominant: those with disabilities were seenas having a condition that set them apart fromthe rest of society. That attitude gave rise todiscrimination, isolation and stigmatization. It isnow increasingly accepted that, while disabilitiesinvolve varying levels and types of impairment,it is social, institutional and attitudinal barriersthat limit the full inclusion of people withdisabilities. Understanding disability in thisway highlights the importance of identifyingand removing the barriers. Education has a keyrole to play in changing attitudes.

Poverty is both a potential cause and a consequenceof disability. In several countries, the probabilityof being in poverty rises in households headedby people with disabilities (McClain-Nhlapo, 2007).In Uganda, evidence from the 1990s found that theprobability was as much as 60% higher (Hoogeveen,2005). Those with disabilities are much less likelyto be working. Other family members may also beout of work (or school) to care for them. Inadequatetreatment, along with poor families’ inability toinvest sufficiently in health and nutrition, reinforcesthe problems people with disabilities face (Bird andPratt, 2004). These links to poverty, combined withstigma and discrimination, are a significant factorin their educational marginalization.

While globally comparable, reliable data arenotoriously difficult to obtain, one widely citedsource estimates that 150 million childrenworldwide live with disabilities (WHO and UNICEF,2008).9 Around four in five children with disabilitiesare in developing countries. In addition, manymillions of children live in households withparents or relatives who have disabilities. At allages, levels of both moderate and severe disabilityare higher in low- and middle-income countriesthan in rich countries. They are highest in sub-Saharan Africa (WHO and UNICEF, 2008). Thescale of disability and its concentration in theworld’s poorest countries contributes significantlyto marginalization in education.

Systematic under-reporting of disability is a seriousproblem. To take one example, a 2004 census inSierra Leone reported only 3,300 cases of mentalimpairment, while a detailed national survey theyear before had estimated the real figure to be tentimes higher (World Bank, 2009c). One reason forunder-reporting is that stigmatization often makesparents and children reluctant to report disability.

Many impairments can be traced back to poverty,poor nutrition and restricted access to basicservices (Yeo and Moore, 2003). Asphyxia duringbirth, often resulting from the absence of a skilledattendant, leaves an estimated 1 million childrenwith impairments such as cerebral palsy and learning difficulties (UNICEF, 2008b). Maternal iodine deficiency leads to 18 million babies being born withmental impairments and deficiency in vitamin Aleaves about 350,000 children in developingcountries blind (Micronutrient Initiative et al., 2009).

Conflict contributes to disability directly throughphysical threats and indirectly through effects onpoverty, nutrition and health care. For every childkilled in warfare, it is estimated that three are left

9. In the 1970s, the WorldHealth Organizationestimated that 10% of theglobal population lived with a disability. This roughestimate is still in use today,suggesting that there areabout 650 million people withdisabilities. It is the basis forthe estimate of 150 million of children with disabilities.

Disability is one of the least visiblebut most potentfactors ineducationalmarginalization

Conflict in 2008 and 2009 gravely affected the education system in Gaza.The circumstances surrounding the violence are subject to claim andcounter-claim. In a report presented to the United Nations GeneralAssembly, Justice Richard Goldstone documented evidence of both sidestargeting civilian populations. What is not in question is the scale of thehuman and physical damage inflicted by Israeli military actions.

Part of the damage can be counted in terms of lives lost and people injured.It is estimated that 164 students and 12 teachers were killed. Many moresuffered long-term injuries. Infrastructure was severely affected. Whileestimates vary, Justice Goldstone reported that some 280 schools andkindergartens were identified as destroyed or badly damaged. Restrictionson transport of building materials have delayed reconstruction.

Less easy to document are the effects of childhood trauma. Violent conflicthas left deep scars in Gaza society. Research in Gaza has identified post-traumatic stress disorder as a major problem for young people, with 69%of adolescents affected and 40% reporting moderate or severe depression.Such conditions create severe educational disadvantage.

The scale of violence experienced by civilian populations in 2008 and 2009has compounded the disadvantage. Many children have returned to schoolsuffering from anxiety, the emotional shock of losing parents or siblingsand the memory of acts of extreme violence. The consequences foreducation are likely to be far reaching and long lasting.

Sources: O’Malley (2009); United Nations (2009a); Elbedour et al. (2007).

Box 3.12: Education destruction and reconstruction in Gaza

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 8 2

with an impairment (UN Enable, 2009 cited inPeters, 2009). Road accidents, a less widelyrecognized cause of impairment in childhood, areendemic in many of the world’s poorest countries.It is conservatively estimated that 10 millionchildren are injured each year on the world’s roadsand many are left with permanent impairments.Over 80% of road-related injury and death occursin developing countries (WHO and UNICEF, 2008).The consequences for education of these deathsand injuries have been subject to insufficientscrutiny, notably by national agencies and donorsinvolved in road construction.

The link between disability and marginalization ineducation is evident in countries at different endsof the spectrum for primary school enrolment andcompletion. In Malawi and the United Republic ofTanzania, having disabilities doubles the probabilityof children never having attended school, and inBurkina Faso it increases the risk of children beingout of school by two and a half times (Kobiané andBougma, 2009; Loeb and Eide, 2004; United Republicof Tanzania Government, 2009). In these countries,inadequate policy attention to disability is clearlyholding back national progress towards universalprimary education. In some countries that arecloser to achieving that goal, people with disabilitiesrepresent the majority of those left behind. InBulgaria and Romania, net enrolment ratios forchildren aged 7 to 15 were over 90% in 2002 butonly 58% for children with disabilities (Mete, 2008).

‘Disability’ is a generic term covering a multitudeof circumstances. Children with, say, severe autismare likely to face very different education-relatedchallenges than children who are partially sighted,or who have lost a limb. Impairments that affectthe capacity to communicate and interact in wayscommon in mainstream schools can imposeparticularly high practical and social obstaclesto participation in education.

A closer look at national data often reveals markedlydifferent consequences for various impairments.In Burkina Faso, children reported as deaf or mute,living with a mental impairment or blind were farless likely to be enrolled in school than those witha physical impairment. In 2006, just 10% of deaf ormute 7- to12-year-olds were in school (Kobiané andBougma, 2009); (Figure 3.28). The attendance ratefor children with a physical impairment was 40%,only slightly below those with no impairment. InUganda, recent evidence suggests dropout ratesare lower among children with visual and physical

impairments than among those with mentalimpairments (Lang and Murangira, 2009).

Children with disabilities face many challengesin education. Three of the most serious involveinstitutionalized discrimination, stigmatizationand neglect, from the classroom to the localcommunity and in the home. Children withdisabilities are often isolated within their societiesand communities because of a mixture of shame,fear and ignorance about the causes andconsequences of their impairment.

One qualitative study of attitudes towards childrenwith autism in Ghana revealed they were widelydescribed as ‘useless and not capable of learning,(…) stubborn, lazy, or wilfully disobedient’ (Anthony,2009, pp.12–13). In a statement with widerapplication, the Ghanaian Ministry of Education,Sports and Science has powerfully captured thesocial prejudices that shape the educationdisadvantages associated with disability: ‘Theeducation of children with disabilities is undervaluedby families, there is a lack of awareness about thepotential of children with disabilities, children withdisabilities in mainstream schools receive lessattention from teachers and there is an over-emphasis on academic achievement and examinationas opposed to all round development of children’(Ghana Ministry of Education, Science and Sports,2008, pp. 60–61).

Education systems and classroom experiencecan help counteract the marginalization thatchildren with disabilities face. However, they oftenhave the opposite effect. Insufficient physical access,shortages of trained teachers and limited provisionof teaching aids can diminish opportunities. Manyschools, particularly in remote rural areas or inslums, are physically inaccessible to some childrenwith disabilities. Children with sensory or mentalimpairments can find schools noisy, confusingand threatening. The grossly inadequate level ofprovision for children with disabilities in generalschools often drives parents and groupsrepresenting people with disabilities to demandseparate provision (Lang and Murangira, 2009).

That demand is both understandable and is asymptom of wider problems. Putting children withdisabilities in special-needs schools or institutionscan reinforce stigmatization. It can also deny thema chance to participate in mainstream education,build relationships and develop in an inclusiveenvironment. Moreover, special schools are often

In Burkina Faso,just 10% of

children with a hearing or speech

impairment were in school

in 2006

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 8 3

chronically underfunded and lack either skilledteaching staff or the equipment needed to delivera good education.

Education planners need to recognize that givingchildren with disabilities a level of access and qualityof education equivalent to that enjoyed by otherchildren often entails increased financing. Additionalresources are needed to provide teachers withspecialized training and children with speciallydesigned learning materials to realize theirpotential. Families may also require additionalfinancial support. One study in Bangladesh foundthat the parents of children with disabilities facedcosts for aids, appliances and health care thatwere three times the average household budgetfor raising children (Chowdhury, 2005, in Marriottand Gooding, 2007). Overcoming a legacy ofinstitutionalized disadvantage can be difficult evenin countries with a strong commitment to moreinclusive education, such as India (Box 3.13).

Giving childrenwith disabilities a level of accessand quality of educationequivalent to thatenjoyed by otherchildren oftenentails increasedfinancing

Figure 3.28: Burkina Faso’s children with disabilities face deep but varied levels of disadvantage% of children aged 7 to 12 and 13 to 16 attending school, by nature of impairment,Burkina Faso, 2006

Source: Kobiané and Bougma (2009), based on data from the 2006 Burkina Faso census: Recensement Général de la Population et de l’Habitation.

0

10

20

30

40

50

Scho

ol a

ttend

ance

rate

s (%

)

7- to 12-year-olds 13- to 16-year-olds

No impairment

Physical impairment

Visual impairment

Mental impairment

Hearing/speech impairment

Education planning documents in India enshrine astrong commitment to inclusive education. The aimis to provide all children with disabilities, irrespectiveof the type or degree of impairment, with educationin an ‘appropriate environment’, which can includemainstream and special schools as well as alternativeschools and home-based learning. Delivering on thiscommitment requires a concerted political effortbacked by reforms in service provision.

Yet disability remains a major brake on progresstowards universal primary education in India. Whilethere are inconsistencies in national data, estimatessuggest that school participation among children with disabilities never rises above 70%, far below the national average of around 90%. According to a World Bank analysis of India’s 2002 National SampleSurvey, children with disabilities are five and a halftimes more likely to be out of school.

Disaggregation of the data highlights importantvariations. Almost three-quarters of children withsevere impairments are out of school, compared with about 35% to 40% among children with mild or moderate impairments. The most likely to beexcluded are children with mental illness (two-thirdsof whom never enrol in school) or blindness (over half never enrol).

Public attitudes are among the greatest barriers to equal education for people with disabilities in India.Children with mental impairments face some of the

most deeply entrenched prejudices. In a public attitude survey covering the states of Uttar Pradeshand Tamil Nadu, almost half of respondents said suchchildren could not attend either regular or specialschool. Another commonly held view was that thosewith mental impairments would not find decentemployment. People from households with a disabledmember shared the general view, reflectingstigmatization in the home.

Institutional constraints reinforce public attitudes.In 2005, just 18% of India’s schools were accessibleto children with disabilities in terms of facilities such as ramps, appropriately designed classrooms andtoilets, and transport.

National education policies reflect growing awareness of the problems associated with disability.Measures introduced so far range from providing aidsand appliances in schools to stipends for children with disabilities. Public awareness problems havehampered implementation, however. In a survey inTamil Nadu and Uttar Pradesh, almost three-quartersof households that included a member with a disabilityreported being unaware of their eligibility for aids and appliances, and only 2% had directly benefited in 2005. Less than half of these households wereaware that stipends were available and only 4% hadreceived them.

Sources: National Sample Survey Organization (2003); Singal (2009); O’Keefe (2007); District Information System for Education (2009).

Box 3.13: Prejudice limits educational opportunities for children with disabilities in India

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 8 4

Having disabilities is not the only situation affectingchildren’s educational opportunities. Children whoseparents have disabilities often face tensions betweenschooling and care demands at home. Theseindirect consequences of adults with disabilities,known as the ‘cascade effect’, are often very severe.

National census data provide an insight into thescale of the problem. Having a poor parent with adisability increases the likelihood of 7- to 16-year-olds never having been to school by twenty-fivepercentage points in the Philippines and thirteenpoints in Uganda – a reminder of how poverty,disability and education interact (Table 3.5).

People affected by HIV and AIDS

HIV and AIDS are principally a global healthcrisis, but one with profound and wide-rangingconsequences for education. As well asthreatening lives, keeping children out of schooland compromising learning, HIV and AIDSreinforce wider problems arising from povertyand social discrimination, such as economicpressure, orphanhood and stigmatization.

An estimated 33 million people were living withHIV in 2007, two-thirds of them in sub-SaharanAfrica. The region is home to 90% of the 2 millionchildren below age 15 living with HIV. Mostcontracted the virus during pregnancy, birth orbreastfeeding – easily preventable forms of HIVtransmission (UNAIDS et al., 2008). Withoutantiretroviral therapy, about 90% of these childrendie before reaching school age (Pridmore, 2008).Those who live may suffer associated problems,such as respiratory infections, malnutrition anddiarrhoeal disease, more often and more severelythan do healthy children, affecting their capacityto attend school and learn.10

Some of the most devastating effects of HIV andAIDS on education are not reflected in school data,

for an obvious reason: many victims do not reachschool age. Around 270,000 children under 14 diedof AIDS-related illnesses in 2007 (UNAIDS et al.,2008). In many countries HIV and AIDS arereinforcing deep gender disparities in education.In high-prevalence southern African countries,such as Malawi, South Africa and Swaziland, HIVinfection rates for girls and young women aged 15to 24 are 1.8 times to 5.5 times the rates for men(Stirling et al., 2008). These disparities can harmgirls’ prospects of completing primary school andmaking the transition to secondary school.

With limited savings and assets, and dependent onphysical labour for income, the poorest householdsare the least equipped to cope with the health costsof HIV and AIDS (UNAIDS et al., 2008). Many mustsacrifice spending in other priority areas, includingeducation. Research in Cambodia found that, topay for health care, two-thirds of families affectedby HIV and AIDS reported spending less onchildren’s needs, including nutrition – potentiallycompromising children’s capacity for learning(Alkenbrack et al., 2004). Household members’ill health can also compromise education byincreasing demand for child labour (Pridmore, 2008).

Becoming an orphan due to AIDS can inflict severedamage on education prospects. Some 15 millionchildren under 18 have lost one or both parents toAIDS. Evidence from fifty-six countries with recenthousehold survey data indicated that orphans whohad lost both parents were 12% less likely to attendschool than non-orphans, on average (UNAIDS et al.,2008). Behind this figure are marked variations,some influenced strongly by the gender of thedeceased parent. In some sub-Saharan Africancountries, including Ethiopia, Kenya, Malawi andthe United Republic of Tanzania, children whosemothers died were more likely to move to anotherhousehold and less likely to stay in school (Beegleet al., 2009; Evans and Miguel, 2007; Himaz, 2009;World Bank, 2007c). While the death of a father

10. While access toantiretroviral therapy hasrisen extremely rapidlyover the past few years,increasing the numberof HIV-positive childrenin school and survivalrates among theircaregivers, in mostcountries the scale-uprate is insufficientto reach universalaccess goals by 2010(UNAIDS, 2009).

Becoming anorphan due to

AIDS can inflictsevere damage

on educationprospects

2000 3 21 28 6 30 44

2002 10 19 23 26 39 49

Philippines

Uganda

Table 3.5: Education indicators by disability status of head of household and wealth, Philippines and Uganda

AverageYear Disabled*Disabled* from

poorest 20% Average Disabled*Disabled* from

poorest 20%

* ‘Disabled’ refers to self-reported disability status of the household head.Source: UNESCO-DME (2009).

7- to 16-year-olds who have never been to school (%)

17- to 22-year-olds with fewer than 4 years of education (%)

R E A C H I N G T H E M A R G I N A L I Z E D

G e t t i n g l e f t b e h i n d

1 8 5

in Ethiopia did not significantly affect schoolenrolment, the death of a mother reducedenrolment among both boys and girls by around20% and disrupted attendance by enrolled children(Himaz, 2009).

Stigmatization and institutionalized discriminationoften reinforce education disadvantages associatedwith HIV and AIDS. In Thailand, a qualitative studyfound that those with HIV were denied admissionto school, in violation of national laws. Educatorsexpressed concern that other parents would reactnegatively to the enrolment of HIV-positive students(Save the Children UK, 2006). To some degree,discriminatory school practices hold up a mirrorto society. One large household survey in Indiaindicated that 58% of women and 43% of men fromhouseholds not affected by HIV and AIDS would notsend their children to a school with an HIV-positivechild (Loudon et al., 2007). The same survey foundthat stigma was a major reason for dropout. Youngchildren reported losing interest in their studies,becoming depressed and dropping out becauseof taunts by peers, while adult caregivers reportedthat stigma and discrimination by teachers werethe major educational barrier.

One effect of stigmatization is to force HIV andAIDS underground. In a study examining theeducational needs of HIV-positive learners inNamibia and the United Republic of Tanzania,every HIV-positive child interviewed cited experienceof the negative consequences of disclosure andemphasized greater safety in silence (UNESCOand EduSector AIDS Response Trust, 2008). Suchfears can be well founded. In Brazil and Haiti,teens infected with HIV reported experiencingviolence and fighting among their peers in schoolas a response to their HIV-positive status (Abada-Barrerío and Castro, 2006; Loudon, 2006).

Governments’ failure to respond with sufficienturgency to the threat posed by HIV and AIDS ineducation is often part of the problem. While therehas been an increase in the number of orphanedchildren able to access school thanks to publicpolicy interventions, much needs to be done.A survey of eighteen national education plans in sub-Saharan Africa that have been developed since2005 found that just ten had specific strategies forchildren affected by HIV and AIDS, and that onlyEthiopia, Kenya, Namibia and Rwanda includeddetailed integrated strategies (UNESCO-IIEP, 2009).To failures in policy planning can be added a morewidespread failure by political leaders to lead public

awareness campaigns aimed at challengingmisperceptions and overcoming stigmatization.One policy response to stigmatization has been toprotect learners by not identifying their HIV status,but this can have unfortunate results. In Namibia,it has led to an absence of information on how manylearners are HIV-positive, and hence a lack of specialarrangements or allowances for them (UNESCOand EduSector AIDS Response Trust, 2008).

Conclusion

Identifying the underlying causes of marginalizationin education is a step towards the development ofpolicies aimed at equalizing opportunity. Childrendo not choose the circumstances into which they areborn. Yet the wealth of their parents, and their owngender, ethnicity or language can greatly influencetheir achievement in education and beyond.

This chapter has highlighted the interactionof poverty and social attitudes in creatingdisadvantages that limit opportunities for education,restrict mobility and perpetuate marginalization.What happens in the education systems is criticalbecause schooling can act either as a great levelleror as a driver of disadvantage. But overcomingmarginalization in education requires policiesthat target wider problems rooted in poverty,stigmatization and unequal power relationships.

There are no policy blueprints. Marginalizedpeople across the world share many experiencesin common. By the same token, the circumstancesthat shape these experiences are highly varied.This is true even within countries. For example, thefactors that drive the marginalization in educationamong pastoralists in northern Kenya are verydifferent than those driving marginalization inNairobi’s slums. Poverty is a near universal sourceof extreme disadvantage in education, thoughpoverty does not operate in isolation. The poverty-related disadvantages experienced by young girlsor ethnic minorities are reinforced by socialattitudes that undermine self-confidence and lowerthe perceived value of education. These differencesmatter because successful interventions againstmarginalization have to tackle specific underlyingcauses that may be missed by blanket interventions.

The ultimate goal for education policy is to create anenvironment in which effort and talent, rather thanpre-determined circumstances, determine learningachievements and life-chances. The next part of thischapter explores routes for attaining this goal.

Teens in Brazil and Haiti reportedexperiencingviolence at schooldue to their HIV-positive status

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 8 6

Levelling the playing field

Marginalized people are often conspicuous bytheir absence from national debates on educationreform. The implicit assumption of many policy-makers is that, as national education systemsbecome more effective, the benefits will eventuallytrickle down to the most disadvantaged sectionsof society. That assumption is flawed. Increasingpublic spending on education, raising averagelearning standards and strengthening overallaccountability are necessary conditions forovercoming marginalization. But they will not besufficient to break the cycles of marginalizationdocumented in this Report. Reaching themarginalized will take a concerted effort to tacklethe interlocking structures of disadvantage thatlimit opportunity. The diversity of the processesperpetuating marginalization means there are nosimple panaceas or blueprints for reform. To theextent that any general conclusion can be drawn,it is that all governments can, and should, do moreto put marginalization at the centre of educationreform debates.

How can governments break the cycles ofeducational disadvantage that trap so manychildren, restricting their opportunities and fuellingmarginalization in other areas? This part of thechapter identifies broad clusters of policies:

Make education affordable. Governments inmany countries have withdrawn formal schoolfees, but this is not enough. Indirect costs andinformal charges continue to keep school outof reach for millions of children. Eliminating allschool fees is a first step towards improvingaffordability. Incentives covering other costslinked to school attendance can also play a vitalrole in enabling marginalized children toparticipate in school.

Ensure that schools are accessible. Distanceto school remains a major barrier to educationfor all. This is especially true for girls because ofthe security risks associated with long distancefrom home. Classroom construction can reducedistance and improve physical accessibility tobring schools closer to marginalized people,provided governments target investment withequity in mind. Ensuring that school constructionprogrammes prioritize remote rural areas andurban slums is key. Some marginalized groups –

notably pastoralists – have been bypassed asa result of inflexible models of school provision.More flexible models, including multigrade andmobile schools, can open the doors to education.

Develop an inclusive learning environment. Allchildren deserve a good-quality education buttypically those who enter school carrying theweight of disadvantage receive the worst. Theyare often taught by poorly trained teachers,sometimes in a language they do not understand.They often lack textbooks – and when booksare available, they frequently include materialthat depicts negative stereotypes. Governmentscan address these problems by creating anenvironment of non-discrimination and equalopportunity. Providing incentives for skilledteachers to work in areas characterized by highlevels of marginalization is a starting point.Supporting intercultural and bilingual educationcan strengthen achievement amongdisadvantaged ethnic minorities. Ensuring thatteachers and schools are equipped to supportchildren with disabilities is also important forinclusive education. Channelling extra resourcesand pedagogical support to ‘failing’ schoolscan benefit areas of greatest need.

Rights and redistribution matter. Translatingthe human right to education into concreteentitlements requires action at many levels.National laws can prohibit formal discriminationand create an environment enabling greaterequity. Laws are most effective when linkedto political mobilization and the developmentof broad-based alliances to advance Educationfor All. In addition, governments and donorsneed to strengthen social protection measures,using cash transfers and risk-managementinterventions such as employment programmesto build the resilience of vulnerable households.National budgets can play a vital role inequalizing educational opportunities betweenricher and poorer people and regions.Redistributive public spending can help tonarrow gaps. Conversely, failure to prioritizeequity in national budgets can reinforceexisting disparities.

‘Joined-up’ national strategies. Marginalization ineducation is the result of interlocking deprivation.Breaking down disadvantage requiressimultaneous public action across a broad front,with education interventions integrated into widerpolicies for social inclusion, including strategies

Reaching themarginalized willtake a concerted

effort to tacklethe interlocking

structures of disadvantage

that limitopportunity

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

1 8 7

for tackling social and cultural discrimination,and poor nutrition. In many countries progresstowards more inclusive education is being heldback by piecemeal, under-resourced andfragmented policy planning.

This part of the chapter starts by setting outthe framework for understanding the levels ofintervention required to combat marginalization.It identifies three broad layers explored in thesubsequent sections: policies for improving accessand affordability; the learning environment andfactors influencing education quality; and thebroader enabling environment for tacklingmarginalization in education, including povertyreduction measures and legal entitlements. Theconclusion highlights the importance of joiningup all aspects of these policy approaches into anintegrated framework for tackling marginalization.

The analytical framework

Consider the experience of five primary school agechildren who are all out of school. One is a Hmonggirl living in a remote hill region of the Lao People’sDemocratic Republic. The nearest school is a two-hour walk away and classes are taught in Lao, alanguage she does not understand. The secondchild lives a few metres from a public school undera sackcloth tent in Manila. He spends his daycollecting and selling rubbish to buy food forhimself and his siblings. The third is a young girlin northern Nigeria who has a brother in schoolbut has dropped out herself because she is aboutto be married. The fourth, a Masai boy from Wajirin northern Kenya, tends cattle during a long trekto grazing land. In a small Brazilian town, the fifthchild, who has a severe hearing impairment, doesnot go to school even though there are severalnearby. Local teachers lack training to teach a deafchild and her parents cannot afford a hearing aid.

Each of these children experiences marginalizationin education. Yet the underlying causes vary.Distance to school, the language of instruction,child labour and the affordability of education,discrimination and low expectations, and traditionalcultural practices and beliefs all play a role.Disentangling the forces behind marginalizationis vital, for obvious reasons. Raising teachingstandards in schools in Manila will not help childrenexcluded from those schools by poverty and childlabour. Increasing the overall education budgetin northern Nigeria’s Kano state may not deliverthe intended results if half the state’s children –

the female half – face restricted opportunitiesbecause of the lower value attached to theirschooling by parents or practices such as earlymarriage. Building a new school in Wajir willnot necessarily help educate the children ofMasai communities whose livelihoods dependon being mobile.

One way of thinking about marginalization is toidentify some of the key ingredients for overcomingit. Figure 3.29 presents these ingredients in aschematic outline.

Accessibility and affordability. Proximity ofschools to communities is an obvious conditionfor participation in education, especially foryoung girls, as gender disparities in manycountries widen with distance. Schools also needto be affordable. Just as poverty can leave peoplehungry amid plentiful food, so it can lock poorchildren out of education even when schools areavailable. Public policy can ensure that childrenare not disadvantaged by the location or physicalaccessibility of classrooms or by cost barriersto education.

The learning environment. Most teachersattempt conscientiously to do a good job, oftenin difficult circumstances. Yet millions of childrenface restricted opportunities to learn in anappropriate language and millions more aretaught by overstretched, undermotivated,

The inclusiveeducation triangleindentifies threebroad strategiesfor tacklingmarginalization

Learning environment! Allocating teachers equitably! Recruiting and training teachers from marginalized groups! Providing additional support to disadvantaged schools! Developing a relevant curriculum! Facilitating intercultural and bilingual education

Entitlementsand opportunities

! Developing poverty reduction strategies! Tackling early childhood deprivation! Enforcing anti-discrimination legislation! Providing social protection! Allocating public spending more equitably

! Cutting direct and indirect costs! Providing targeted financial incentives! Investing in school infrastructure! Bringing classrooms closer to children! Supporting flexible provision! Coordinating and monitoring non-state provision

Accessibilityand affordability

Figure 3.29: The Inclusive Education Triangle

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 8 8

untrained teachers in overcrowdedclassrooms lacking basic teaching materials.While the problems are often system-wide,it is marginalized children who experiencethem the most acutely. Strategies to combatmarginalization need to ensure that schoolsserving the poor attract skilled teachers whocan teach in an appropriate language withcultural sensitivity, and that sufficient andrelevant teaching materials are available.

Entitlements and opportunities. Schools can playan important role in combating marginalizationin education and beyond, but there is a limit towhat they can do. Mitigating the impact of povertyon education requires measures that increaseand stabilize the incomes and food securityof poor households. Legal provisions can setstandards and equip people with rights thatunlock opportunities for education, providedthey are enforceable. And public spending canhelp counteract the disadvantages associatedwith poverty. In each of these areas, actionsby governments can create an enablingenvironment for greater equity. At the sametime, political mobilization by the marginalized,or by civil society more widely, is often apowerful catalyst for change.

Each point of the triangle needs to be viewed inrelation to the others. Making primary educationaccessible and affordable without tacklingproblems in education policy is clearly nota prescription for combating marginalization.Conversely, raising the average level of learningfor the majority while leaving behind a substantialminority is a route to more marginalization.The wider pattern of entitlements and enablingconditions is vital because it shapes theenvironment in which the abstract ‘human right toeducation’ is translated into meaningful claims andsubstantive rights. What ultimately matters is thedevelopment of an integrated policy response thataddresses the multiple and overlapping structuresof disadvantage that restrict opportunities formarginalized learners. One powerful exampleof such a response at a community level comesfrom Harlem in New York (Box 3.14).

The lesson that emerges from this section isthat schools have the potential to make a greatdeal of difference to the lives of the marginalized.But the processes that drive marginalization startearly in life – long before children enter school.As Chapter 2 makes clear, evidence from

developing countries shows that malnutritionbefore age 2 undermines cognitive developmentand weakens learning achievement. Evidence fromrich countries shows that much of the attainmentgap at the end of secondary school is predictablebefore age 5, and that learning achievement isstrongly associated with household wealth andparental education (Blanden and Machin, 2008;Feinstein, 2003). Schools can at best mitigatedisadvantages accumulated in early childhood.That is why nutrition, maternal and child health, andearly childhood care and education are central to anintegrated approach for overcoming marginalization.

Expanding access and improvingaffordability for excluded groups

Around 72 million children of primary school ageare out of school, either because they have neverentered the education system or because they havedropped out. Many millions of adolescents enteradulthood without the basic learning skills they needto realize their potential. Changing this picture andaccelerating progress towards the goal of universal

Evidence fromrich countries

shows that muchof the attainmentgap at the end ofsecondary school

is predictablebefore age 5

Numerous initiatives have attempted to close the racial and social divide in American education, but few have achieved a breakthrough in equalopportunity. The Harlem Children’s Zone Project isdifferent. Begun in 1997, it traces its roots to 1970scommunity activism. The failure of social programmesto improve education, tackle unemployment andrespond to the breakdown in family and communitylife that came with crack cocaine use and streettrading prompted community leaders to explore new avenues.

In contrast to narrowly based ‘school reform’ models,the Harlem Children’s Zone Project recognizes thatpoverty, gun crime and drugs are part of a widerculture of low expectations and underachievement.The intent of the project is to create a ‘tipping point’by covering at least 65% of children and their parentsliving in the blocks where the project operates. It seesthis as ‘a threshold beyond which a shift occurs awayfrom destructive patterns and towards constructivegoals’ (Harlem Children’s Zone, n.d., p. 3).

An ambitious, integrated ‘pipeline’ model startsbefore birth with support for maternal health andparenting skills, continues through pre-school tosecondary school and college, and encompasseshousing, social services and nutrition. The emphasis

Box 3.14: ‘Tipping points’ in Harlem

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

1 8 9

primary education by 2015 requires action atmany levels. In most countries in danger of missingthat goal, improving opportunities in educationmeans lowering cost barriers and bringing schoolscloser to marginalized children.

Cutting the costs of entry to schoolMany countries have laws or constitutionsenshrining the right to free primary education. Yetchildren often are excluded from education becausetheir parents cannot afford informal school fees.A 2005 survey by the World Bank covering ninety-three countries found that only sixteen chargedno fees at all, even though the vast majority madefree education nominally available (World Bank andUNICEF, 2009). In reality, free primary schoolingremains the exception rather than the rule.

Recent experience powerfully demonstrates thedamaging effects of charges on primary educationfor equity. Countries eliminating user fees forprimary education have typically seen largeincreases in enrolment, especially amongdisadvantaged groups (Plank, 2007). Even in

countries that have moved to eliminate formalcharges, however, cost may remain a barrier,with many poor parents continuing to cite inabilityto afford education as the reason their childrendo not attend. Why has the move to ‘free’ educationfailed to remove this cost barrier for the parentsof many marginalized children?

Local school-financing practices have sometimescounteracted national policies. When Ghanaintroduced a policy eliminating fees in 1996, therewas initially only a limited increase in enrolment.The reason: schools faced with a loss of revenueintroduced informal fees of their own. In response,the government introduced school grants to makeup for the lost fee income – a policy interventionthat led to rising enrolment levels (Maikish andGershberg, 2008). Several other countries, includingEthiopia, Kenya, Mozambique and the UnitedRepublic of Tanzania, replaced schools’ formeruser-fee revenue with grants. In addition toreducing pressure on household budgets, schoolgrants give governments a vehicle for targetingdisadvantaged groups and regions, for example

‘I have spentdays withouthaving a full mealbut never letFaruk think aboutleaving school.’Faruk’s mother, Bangladesh

is on quality: kindergartens have one teacher for every four children. But scale is also expanding rapidly.From twenty-four blocks in 1997, by 2007 the HarlemChildren’s Zone Project had expanded to ninety-sevenblocks with 7,400 children.

Education is one of the core elements. In 2004, threeschools dubbed ‘Promise Academies’ were opened withfunding from government, philanthropists and charities.Many of the children come from highly marginalizedbackgrounds: 10% live in homeless shelters or fostercare. Management of the schools is geared towards the pupils’ need for intensive support. The learningenvironment includes an extended school day, after-school teaching and remedial classes atweekends. Efforts have been made to recruit and retain high-quality teachers. The schools provide meals and medical care (many students come fromhouseholds without health insurance).

Early results have been very promising. Researchersfrom Harvard University found that students whoenrolled in the sixth grade gained more than a fullstandard deviation in math, and between one-third and one-half of a standard deviation in EnglishLanguage Arts (ELA), by eighth grade: ‘Taken at facevalue, these effects are enough to reverse the black-white achievement gap in mathematics (HCZ students

outperform the typical white student in New York City and the difference is statistically significant) and reduce it in ELA. Students in the HCZ elementaryschool gain approximately one and three-quarters of a standard deviation in both math and ELA, closing the racial achievement gap in both subjects’(Dobbie and Fryer, 2009, p. 3).

Can the project’s achievements be replicated on a national scale? The Obama administration hasoutlined plans to reproduce it in twenty cities under a programme of ‘Promise Neighborhoods.’ Rolling out such an initiative will require more than copying a ready-made blueprint. The high level of communitymobilization and the innovation demonstrated bycommunity leaders over many years cannot be readily duplicated. Moreover, expansion to poorneighbourhoods across America will require large-scalepublic investment during a period of acute budgetaryconstraints. But the prize of building on theaccomplishment of the Harlem Children’s Zone Projectis potentially enormous. The costs of narrowing the deep divides in American education have to be assessed against the wider social, political andeconomic costs of allowing marginalization to diminishthe potential of the country’s children.

Sources: Dobbie and Fryer (2009); Harlem Children’s Zone(2007, n.d.); Shulman (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 9 0

by attaching more weight to rural areas withhigh concentrations of out-of-school populations(World Bank, 2009j).

Like all parents, those of marginalized childrencare about the quality of education. If fee abolitionleads directly to heavily overcrowded classrooms,shortages of teaching materials and unmotivatedteachers, parents may question the real valueof ‘low-cost’ education. Evidence from a rangeof countries that have withdrawn fees showsthat sequencing reform is vital (World Bank andUNICEF, 2009). Increasing investment in teacherrecruitment and textbook provision in anticipationof rising enrolment is likely to prove more effectivethan action after the event. Similarly, bringing moremarginalized children into school increases theimportance of complementary action in otherareas, including school-based nutritionprogrammes (World Bank, 2009j).

Fee abolition is only a partial response to widerpoverty constraints affecting demand for education.Making schools affordable to parents of the mostmarginalized children is likely to involve removingor cutting costs for uniforms, textbooks and othermaterials. In western Kenya, one study based ona randomized experiment found that studentsreceiving a free uniform who did not previously ownone were 13 percentage points more likely to attendschool. For those who already owned a uniform, theestimated impact was small and insignificant (Hollaand Kremer, 2009). Such evidence illustrates theneed to look at the overall cost barriers confrontingpoor households, rather than at user fees inisolation. Experience from a broad group ofcountries points to the positive effects of measuressupplementing the abolition of fees:

In Nepal, the 2004–2009 education strategyincluded scaling up a stipend programmetargeted at low-caste Dalit children. In 2003,about 384,000 out of 527,000 eligible Dalitchildren received stipends (World Bank, 2006d).Scholarships and other incentives have also beenmade available for girls. Another targeted grantprovides a cash transfer to children fromhouseholds in which no member has completeda primary education. Despite some problemsin targeting, the programme appears to havehelped girls and children from disadvantagedbackgrounds into education (Acharya and Luitel,2006; Research Centre for Educational Innovationand Development, 2003).

Viet Nam has introduced a range of financialsupport mechanisms targeting ethnic minoritystudents. However, school costs are still cited asa cause of children dropping out of school. UnderProgramme 135, a poverty reduction strategytargeting 2,100 communes with very low humandevelopment scores, the government provideschildren attending semi-boarding schools witha monthly stipend. Those who do not live incommunes covered by Programme 135 butare poor or live in a ‘commune with extremedifficulties’ receive lower stipends. Everywhere,ethnic minority students receive free textbooksand notebooks (Truong Huyen, 2009).

Several countries have targeted orphansand other vulnerable children. A programmein Mozambique provides around 3,400 orphansand other vulnerable children with vouchersto buy shoes, clothing and stationery. Onestudy points to positive results for enrolment(Ellis et al., 2009).

Stipends at the secondary school level can beeffective in counteracting marginalization in primaryeducation. In some countries, there is evidence thatparents unable to meet secondary school costs willwithdraw their children from primary school beforecompletion. An innovative programme in Cambodiaattempted to forestall that decision. In a pilotscholarship programme supported by the JapanFund for Poverty Reduction, girls who reachedthe final grade of primary school were eligiblefor grants of around $45. The cash was providedto families, conditional on their children attendingsecondary school. It was estimated that theprogramme increased enrolment amongparticipants by around 30%. An evaluation foundthat enrolment effects rose with household poverty.For girls from the poorest 20% of households,enrolment increased by 50%, compared with 15%for girls in the wealthiest two quintiles (Filmer andSchady, 2008; Fiszbein et al., 2009).11 TheBangladesh Female Secondary School StipendProgramme has also introduced wider conditionsfor transfers. It covers school fees and additionalpayments for girls who stay in school, remainunmarried to age 18 and pass exams. The stipendsare credited not just with increasing secondaryschool enrolment by around twelve percentagepoints, but also with creating incentives forhouseholds to ensure that girls complete primaryeducation (Khandker et al., 2003). Girls’ primaryschool enrolment now exceeds that of boys.12

11. The pilot project, from2002 to 2005, targetedonly girls. A follow-upprogramme, CambodianEducation Sector SupportProject – Scholarships forthe Poor, targets bothboys and girls withdifferent levels of support.It has also had markedeffects on enrolment andattendance (see Annex,p. 294).

12. Another programmein Bangladesh targetingprimary school childrenfrom poor ruralhouseholds has been lesssuccessful, partly becauseeligibility criteria haveexcluded some of themost marginalizedchildren, including manyliving in slums andinformal settlements aswell as those attendingmadrasas and schoolsrun by non-governmentorganizations(Al Samarrai, 2008).

For poor girls in Cambodia,

recieving a secondary

schoolscholarship

increasedenrolment

by 50%

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

1 9 1

Bringing classrooms closer to marginalized childrenPhysical access to classrooms remains a majorbarrier to Education for All. There is no universalbenchmark for the appropriate distance to school.One estimate suggests that 2 km, or a thirty-minutewalk, should be viewed as an upper limit(Theunynck, 2009). However, much dependson context and circumstance. Where mountains,forests or rivers limit accessibility, even shortdistances can entail long journey times andhigh levels of risk.

Increased and more efficient public spending onclassroom construction is one way to expandaccess. Classroom shortages inevitably increasedistance to school – and many countries have acuteshortages. Low-income countries in sub-SaharanAfrica are currently running a deficit of around1.7 million classrooms. To close that deficit by 2015,the number of classrooms needs to be doubled(EPDC and UNESCO, 2009). Recent estimatesfor ten sub-Saharan African countries that areoff track for the 2015 goals suggest that thenumber of classrooms is growing at less thanhalf the required rate (Theunynck, 2009).

The location of new schools and classroomsis critical for underserved groups. Too often,classroom construction programmes fail toprioritize areas and groups with greatest need.This is despite the proven benefits of greater equity.In Ethiopia, classroom construction has been acentral part of the national strategy to accelerateprogress towards universal primary education.Of the 6,000 schools built since 1997, over 85%are in rural areas, significantly reducing averagedistances to school. The out-of-school populationhas declined by 3 million and gender disparitieshave narrowed, underlining the effect of distanceon demand for girls’ education (UNESCO, 2008a).

Combining technology and community participationcan help education planners identify underservedgroups and areas. Some countries, includingEthiopia, have used geographic information systemsto generate information on the spatial distributionof schools, their proximity to pupils’ homes andgeographic features such as roads, rivers andmountains (Attfield et al., 2002). Communities cansupplement this information with local knowledgeon the ‘cultural distance’ that gender, social andethnic factors can create between schools andmarginalized people. Such social mapping is oftenimportant. Assessments in India’s Rajasthan state

in the 1990s found that over 90% of children livedwithin 1.5 km of a primary school, yet enrolmentrates were below 50% because social divisions,including caste, made many parents unwillingto send children to school (Govinda, 1999). Thisillustrates how social distance can reinforce spatialdistance in marginalizing disadvantaged groups.

Children with disabilities – particularly those withvisual, physical and severe mental impairments –face obvious disadvantages in negotiating thejourney to school and, in many cases, in access tothe classroom and other facilities, such as toilets.These disadvantages are reflected in the limitedimpact of school fee abolition on their enrolment.On one estimate, only one in six Kenyan childrenwith disabilities was attending school after thefee abolition (Mulama, 2004). Difficulties withaccessibility cannot readily be separated fromwider factors that exclude children with disabilitiesfrom school. In many cases, parental concerns overchildren getting to and into school are compoundedby concerns over their experiences in classrooms.

Improving access for children with disabilitiesrequires policy interventions at many levels.Regulations on school design can play an importantrole in making participation in school possible.Many children with disabilities are effectivelyexcluded from school by the absence of low-costramps and appropriate toilet facilities. Getting toschool raises wider problems. Public transportsystems in many countries are inaccessible topeople with disabilities. Sparsely-populated ruralareas, where distance to school is the greatest,often have no public transport at all. In urban areas,where the condition of streets often hampersmobility for people with disabilities, the absenceof transport effectively prevents many childrenwith disabilities from reaching school. Parentalresponses to surveys underline the importanceof transport. One survey in Bangladesh found thatparents of children with disabilities saw theabsence of a specialized transport system fromhome to school in rural areas and the lack ofsubsidized support for rickshaw transport asmajor constraints (Ackerman et al., 2005).Education authorities can play a role in addressingaccess problems through regulations on schooldesign, providing subsidized transport and bringingschools closer to homes.

Some of the most severe classroom shortages arefound in areas where conflict has destroyed schoolinfrastructure. After conflict ends, rapid

In sub-SaharanAfrica the numberof classroomsneeds to bedoubled by 2015

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 9 2

reconstruction and concerted efforts to get childreninto school are vital. Rwanda’s government backeda school rehabilitation programme with a strenuousre-enrolment campaign aimed at overcomingparental security fears and rebuilding trust.Although it took four years for enrolment to returnto the levels recorded before the 1994 genocide,by 2005 access was above the level that a simpleextrapolation of the trend from 1985 to 1992would have predicted (Obura and Bird, 2009).

Adapting schools to local contextsUnderstanding local context is critical to developingpolicies for inclusive education. Many marginalizedchildren live in scattered communities in remoteareas where low population density can significantlyraise the average cost of providing schools andteachers. Household poverty and livelihood systemscan also keep children out of school when familiesrely on children to tend cattle or help with farmwork and domestic chores. Other marginalizedchildren live in slums that are not legallyrecognized and may face problems linked tohousehold vulnerability. Making schools accessiblerequires innovative policy responses gearedtowards specific circumstances.

In many countries, low-population density ruralareas are marked by highly concentrated patternsof marginalization in education. Individual villagesor groups of villages in regions such as the Andeanhighlands of Peru and Bolivia may have far fewer,and more widely dispersed, primary school agechildren than other areas. These children arelikely to face longer journeys, with harsh terraincompounding the problem of distance. Attendinga school in a ‘neighbouring’ village might involvefording streams and negotiating steep slopes.During the monsoon season in Bangladesh,children living on chars (sand islands in rivers)may have to swim or use banana-leaf rafts toget to school.

Several countries have developed ‘satellite school’models aimed at addressing such problems.Schools are organized into clusters, usuallyconsisting of a central, relatively well-resourcedschool and several smaller satellites. The lattermay be one-room schools with one personteaching more than one grade in the same class.

In Bolivia, clusters of schools, known as núcleos,have been created to expand the reach of theeducation system into underserved highland andjungle areas. Each cluster comprises a central

school, offering the full cycle of grades up tosecondary school, and several satellite schoolsoffering the first three primary grades in multigradeclasses. Students and teachers can be redirectedto different schools within the cluster to makecoverage more even. This system has played a vitalrole in expanding access to education amongindigenous children in highland areas. By providinginstruction in Bolivia’s three main indigenouslanguages, as well as Spanish, núcleos alsopromote bilingual and intercultural education(Giordano, 2008). The reform helped increase thepublic education system’s coverage. For instance,in 1992, 82% of urban but only 41% of ruralstudents completed grade 6; by 2001 it was 85%in urban areas and 74% in rural areas (Contrerasand Talavera Simoni, 2003).

Satellite systems have to address difficult problemsin managing progression through grades. Thenúcleo system in Bolivia aims to ensure thatchildren complete their basic education at theconsolidador, or central school. Another approachis to create satellite schools that provide a fullprimary cycle, such as those developed for remoterural communities in Burkina Faso (Theunynck,2009). The advantage of such a system is that itallows for continuity. But does the provision ofmultigrade teaching across more grades potentiallycompromise the quality of provision?

That question is an important one. About one-thirdof all primary school age children in developingcountries are now taught in multigrade settings(Little, 2006b). Evidence from some countriessuggests multigrade teaching can enhance accesswithout compromising quality. Reviews of the well-established Escuela Nueva, a multigradesystem in Colombia, have found higher achievementin Spanish and mathematics than in other primaryschools, controlling for other characteristics(Forero-Pineda et al., 2006). Evidence from BurkinaFaso, Pakistan and Togo similarly suggests thatmultigrade classes can perform at least as well assingle-grade schools (Little, 2006b). Still, not allmultigrade schools are successful and muchdepends upon the effectiveness of institutionalsupport mechanisms (Little, 2006a).

The Escuela Nueva system and, to a lesser extent,comparable programmes in Chile and Guatemalahave been successful partly because they are linkedto wider reforms. Research has highlighted theimportance of investment in adequately trainedteachers to work in a multigrade setting, the

Evidence fromsome countries

suggestsmultigrade

teaching canenhance access

withoutcompromising

quality

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

1 9 3

development of curricula and teaching materialsthat are responsive to student needs and parentalconcerns, and teaching approaches that encouragestudents to participate actively in the learningprocess and to work independently and creatively(McEwan, 2008). Strategies to overcomemarginalization need to combine innovativemultigrade teaching with support in these key areas.

In pastoral areas, problems posed by low populationdensity are compounded by mobile lifestyles.Improving access to education for pastoralistchildren requires a break with traditional thinking –and an evidence-based assessment of what works.One such response has been the development of‘mobile schools’ that follow the community, withteachers delivering instruction at times whenchildren are not herding. Initiatives in both Ethiopiaand Kenya experimented with mobile schoolprogrammes, supplemented by boarding schools.

While these approaches have created newopportunities, they have often lacked a coherentpolicy framework or sufficient investment ofresources (Rose, 2003; Ruto et al., 2009). Somecountries are now starting to take a moreintegrated approach. In northern Kenya, improvedpolitical representation of arid areas has gonehand in hand with the development of broad-basedstrategies to overcome education marginalization.Much will depend upon the level of support,financial and political, that these strategies attractfrom the central government and upon the successof wider poverty reduction strategies (Box 3.15).

Enforced mobility often comes with vulnerabilitiesthat lead to educational marginalization. Refugees,internally displaced people and children migratingto find work in urban areas are all examples. Mostchildren in slums wage a daily battle for survivalthat involves long hours working for little income.

Improving accessto education forpastoralistchildren requires a break withtraditional thinking

Marked by unpredictable rainfall and unreliablefood supply, along with cattle rustling and banditry,life for pastoralists in the arid lands of northernKenya is precarious. The region’s underdevelopmentreinforces the daily challenges: only one district townis connected to the national electricity grid. Againstthis harsh backdrop, the arid lands were hardesthit by a devastating drought and famine that sweptthe country in 2009, killing entire herds and sendingmalnutrition soaring. Turkana children had to hike30 km for water and some Turkana men abandonedtheir families, unable to face the shame of beingunable to feed their children. Ethnic conflict roseover the last remaining pieces of fertile grazing land.

Education reforms have had a limited impacton the lives of pastoralists. In most of the restof Kenya, fee abolition led to a surge in enrolment,but it made little difference in pastoral areas. Inthe North Eastern Province, fewer than 40% ofchildren were enrolled in school in 2007, four yearsafter fees were abolished. Pastoralists’ mobilelifestyle and extreme vulnerability mean thatreducing the cost of schooling alone was insufficientto enable their children to gain access to education.

To make a difference, an integrated approach todevelopment in the region is needed, along withstrategies directly aimed at providing an educationrelevant to the lives of pastoralists. Such an approachhas not been apparent until very recently. Until thelate 1990s, the north in general and pastoralists inparticular were largely ignored. In education policy,

the focus was on persuading pastoralists to abandontheir livelihoods and settle in one place where theycould more easily be provided with services. Thispicture has been changing with the emergence ofpastoralist civil society organizations and a significantpastoralist group in Parliament — a development thathas increased the voice of one of the country’s mostmarginalized groups. The creation of a Ministry ofState for the Development of Northern Kenya andother Arid Lands in April 2008 is one of the boldeststatements of the government’s intention to addresschallenges in the north more proactively.

As part of its strategy to address the developmentneeds of the region, the new ministry was influentialin developing a Nomadic Education Policy, draftedin 2008. Innovations include incorporating traditionalknowledge in the curriculum, providing grants tomobile schools, establishing feeder schools withinlocal communities, modifying the formal systemto suit the nomadic calendar, recruiting teachers(particularly females) from nomadic areas throughaffirmative action, and using radio and mobilephones for outreach.

The problem is that the new ministry has a broadmandate with an insufficient budget. For 2009/2010,the ministry was allocated a mere 0.5% of thegovernment budget. Without more serious financialbacking, there is a real danger that the ministry’sinitiatives will fail.

Sources: Gettleman (2009); Ruto et al. (2009); World Bank (2009f).

Box 3.15: Reaching pastoralists in northern Kenya

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 9 4

Improving access to education for these childrenis often difficult, but it is not impossible. The keyis to identify the children and ensure thateducation is provided on a flexible timetablein an accessible environment.

Targeting excluded regions and groups ofteninvolves more than the physical presence of aschool. Some governments and non-governmentorganizations have used technology in an effortto shrink distances in education. Such technologycan complement teacher-student contact by beingavailable at times when children cannot make it toschool (whether in the evening or during seasonswhen they are needed to work) (CambridgeDistance Education Consultancy, 2009). In China,education authorities have developed a range ofdistance-learning models, using DVDs and satellitebroadcasts to provide teaching to schools in remoterural areas. While the benefits of distance learningin primary school can be compromised by theabsence of a teacher, in this case the policy wasaccompanied by investment in training localteachers. Large-scale evaluations in Gansu andHubei – among the most deprived provinces inwestern China, with particularly low literacy rates –found improvements linked to distance learning,with most teachers reporting evidence of studentstimulation (McQuaide, 2009).

Providing a second chance to out-of-school children and adolescentsMany marginalized children and youth lack a wayback into education. Adolescents who have neverattended school or who dropped out early havelow levels of literacy and numeracy. Many of theover 71 million adolescents estimated to be outof school are denied a second chance, oftenbecause of inflexibility in national educationsystems. Facilitating re-entry into education isa key strategy for empowering youth and youngadults to escape poverty.

Non-government organizations often provideeducation that is complementary to formalschooling, and can put children and youth on aroute back into the formal system. The scale ofthis provision is not widely recognized. One surveyin sub-Saharan Africa recorded 154 programmesin 39 countries reaching 3.5 million children(DeStefano et al., 2006). While the quality of sucheducation is highly variable, the scale of demanddemonstrates that complementary educationprogrammes fill an important gap. The moresuccessful programmes combine flexible timing

of classes with strong support for learners aswell as courses and curricula geared towardsrelevant skills.

Re-opening the doors to education is a majorchallenge for education policy. Some programmesfocus on building bridges between skills trainingand employment for marginalized youth and adults.The Jóvenes programmes in Latin America areone example (see Chapter 2). Over-age childrenand adolescents who have missed out on primaryeducation have different needs. Acceleratedlearning programmes have been developed inseveral countries to provide them with opportunitiesto cover the primary education curriculum over ashorter period. An important requirement for bothtypes of intervention is that they lead to recognizedqualifications, allowing graduates to re-enterthe formal school system or to gain meaningfulemployment. This means programmes run by non-government organizations must beacknowledged by governments and integratedinto their national plans.

Such programmes have been beneficial in reachingvarious marginalized groups, from Bangladeshinomads and street children (Box 3.16) to peoplein the most educationally disadvantaged regionof Ghana (Box 3.17). They also play a vital rolein post-conflict settings, where a generationof children may have missed out on education.Sierra Leone’s post-conflict reconstructionstrategy targeted children aged 10 to 16 througha programme called Complementary RapidEducation for Primary Schools. Although under-resourced, the schools in the programme broughteducation to thousands of children. These childrenperformed as well as other primary schools innational tests. As a result, many participantstransferred to regular primary and secondaryschools, and are reported to have continued to dowell (Baxter and Bethke, 2008; Johannesen, 2005).

Responding to non-state initiativesWhen governments fail to provide marginalizedchildren with an appropriate education, localcommunities often develop their own schools.How governments respond to such local initiativescan have an important bearing on educationopportunities for marginalized groups.

In Zambia, some of the poorest communities set uptheir own schools after a breakdown in the nationaleducation system in the 1990s. In 2006, about one insix basic-level students were attending one of these

Innovativeprogrammes run by non-government

organizationsmust be

integrated intonational plans

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

1 9 5

Bangladesh has made rapid but uneven progresstowards universal primary education. Previously deepgender inequalities have been eliminated in primaryeducation and rural areas have been catching up withurban areas. Enrolment among children living in extremepoverty has been less impressive, however, and themarginalization of this group remains a barrier touniversal primary education. Initiatives developed by non-government organizations, which reach over 1 millionof the country’s most marginalized children, providepowerful evidence that this barrier can be removed.

One example comes from the country’s riverbanks. The800,000 strong Bede, or River Gypsy, community lives onboats in groups of ten to fifteen families. The Bede, amongthe poorest people in the country, live off trinket selling,fishing, pearl-diving, snake-catching and traditional healing.These activities involve travel over long distances. Becausethey are not settled, the Bede have traditionally lacked theresidency rights necessary to claim school places. Evenwhen they do have formal rights, their mobility makes itdifficult for their children to attend school regularly, soteachers are reluctant to enrol them or provide books.

Since 2006, a national non-government organization, theGram Bangla Unnayan Committee, has provided educationthrough twenty-one ‘school boats’ that follow the Bede

community. Teachers are recruited from the communityand given basic training. The boats provide educationfor two to three years, after which children living withsedentary relatives can gain admission to governmentprimary schools.

Street children are another highly marginalized group.Recognizing the limited success of government effortsto reach these children through formal schooling, non-government organizations opened learning centres as partof the Basic Education for Hard to Reach Urban WorkingChildren programme. In its first phase, the programmetrained 346,000 urban working children aged 8 to 14 inbasic literacy, numeracy and life skills. They took two-yearcourses that were equivalent to three years of governmentprimary schooling. Participants were among the mostdeprived children in the country. One survey revealed thatthree-quarters of them had never been to school and that83% of participants’ families earned less than US$2 per day.

Accessibility problems were addressed by locating learningcentres near children’s places of work and shortening theschool day to two-and-a-half hours. Few children droppedout of the programme. A remaining challenge is to finda way to enable them to enter the formal system.

Sources: Bangladesh Government (2008); Khan and Chakraborty (2008); Maksud and Rasul (2006); Nath (2009); UNICEF (2008a); World Bank (2008d).

Box 3.16: Reaching the most marginalized in Bangladesh through floating schools and programmes for child labourers

Northern Ghana faces some of the country’s mostacute educational deprivation. School attendancerates in the region are among the lowest in thecountry and many children reach adulthood withno more than a few years of education. Parentscite distance to school, cost, seasonal labour demandand, for girls, early marriage as major barriers.

An innovative programme run by non-governmentorganizations is attempting to provide out-of-schoolchildren in northern Ghana with a second chance.School for Life offers an intensive nine-month literacycourse for children aged 8 to 14, with the aim ofpreparing them to re-enter primary school. Teachingschedules are designed to accommodate seasonaldemands on children’s time. Students are given freebooks and uniforms are not required, reducing thecost of attendance.

The School for Life curriculum is designed tomake education meaningful to rural families whofeel that formal schools fail to respect the dignityand strengthen the self-esteem of their children.Students are taught in local languages by locallyrecruited facilitators, many of them volunteers,who receive in-service training.

School for Life has achieved impressive results.Between 1996 and 2007, it reached around85,000 children in eight districts, with no discerniblegender gap. An evaluation in 2007 found that over90% of students completed the course, 81% metthird-grade literacy and numeracy standards and 65%entered the formal education system. Governmentdata indicate that School for Life graduates enteringformal school perform above the average inmathematics and English.

Sources: Casely-Hayford et al. (2007); Hartwell (2006); Mfum-Mensah (2009).

Box 3.17: Addressing educational deprivation in northern Ghanathrough complementary education provision

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 9 6

8,000 community schools. These schools play a vitalrole in providing access to education for childrenin slums and poor rural areas. Government supportis erratic: many community schools are staffed byvolunteer teachers and lack teaching materials.Yet scaling up government support could be a cost-effective strategy to combat marginalizationin education (de Kemp et al., 2008; DeStefano et al., 2006). To be effective, partnerships betweengovernments and non-state providers servingmarginalized groups need to be well-defined, withgovernments taking responsibility for long-termfinancing, the provision of teaching materials andthe monitoring of quality (Akyeampong, 2009).

In some countries, religious schools fill gaps ingovernment education. Some of these schoolsreach highly marginalized groups and regions.In Kano state, Nigeria, which has some of theworst education indicators in sub-Saharan Africa,around 2.9 million children and youths aged 6 to21 attend some kind of Islamic school – roughlytwice the combined attendance in government andprivate schools. About half of these schools arecommunity-owned schools, some of which teachthe national curriculum and receive state support.Aid donors sometimes express concern overwhether Islamic schools foster ‘anti-Western’values. Yet these schools reach some of Nigeria’smost deprived children and they are often in parta response to poor quality in the state system(Bano, 2008). Here, too, there is potential forthe government to work with non-state actorsto extend education opportunities in marginalizedareas. Integrating these schools into thegovernment system, and providing supportby training teachers and supplying textbooks,would help ensure that their students achievedbasic literacy and numeracy skills.

Private schools may also fill gaps in education.There may, however, be adverse consequencesfor equity (UNESCO, 2008a). In some cases, it canmean that the poorest slum households pay foreducation while free government schooling isavailable to those in less poor urban areas. InKenya, the government has responded by providingcapitation grants from the Free Primary Educationbudget to private schools willing to comply withministry guidelines. Many schools do not comply.The government could take more responsibilityfor regulating these schools, but this is a difficulttask, given that they often operate under thegovernment radar. A longer-term solution wouldbe for the government to fulfil its commitment

to free primary education for all by extendingits provision to slum dwellers (Oketch et al., 2008).

The learning environment

Governments across the world have signedup in large numbers to the principle ofinclusive education. At the core of this ideais a compelling vision, set out in the SalamancaStatement and Framework for Action on SpecialNeeds Education, of ‘the need to work towards“schools for all” – institutions which includeeverybody, celebrate differences, support learning,and respond to individual needs’ (UNESCO andSpain, Ministry of Education and Science, 1994,p. iii). Translating the vision into practice requirescreating learning environments that include allchildren, giving priority to those who aremarginalized and excluded.

The learning environment in which childrenparticipate is shaped by a vast array of factors.Parental influence, home background, studentcharacteristics, the school and the educationsystem as a whole all play a role. The interactionbetween these layers and the factors thatmarginalize children is quite complex. Poverty,gender, ethnicity, minority language and disabilitydo not automatically consign children to amarginalized future, in education or beyond.

Classroom experience, the focus of this section,can help counteract disadvantage but may reinforceit. Schools that give marginalized children accessto well-trained and motivated teachers, instructionin a language they are familiar with, a relevantcurriculum and adequate teaching materialsare powerful vehicles for combating socialdisadvantage. Many schools lack some or allof these ingredients. All too often, the mostmarginalized children are taught by the least skilledteachers in the most poorly resourced schools.Tackling this problem requires education systemsand political leaders to recognize and respond tothe special needs and constraints facing childrenwho have been denied opportunities for education.

Allocating teachers to marginalized areas and schoolsWell-trained teachers can help mitigate thedisadvantages of marginalized children. Suchchildren stand to gain the most from high-qualityteaching, but are the least likely to receive it. Theproblem is not restricted to developing countries.In France, teachers in lower secondary schools

Translating thevision of inclusive

education intopractice requirescreating learning

environmentsthat give priority

to those who are marginalized

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

1 9 7

belonging to Zones d’Éducation Prioritaire (ZEPs,or Priority Education Areas) are likely to have lessexperience than teachers in other schools andteacher turnover is much higher than the nationalaverage (Duru-Bellat, 2009). Problems are mostacute, however, in poor countries with deprivedareas facing acute shortages of skilled teachers.

Recruitment and deployment practices are atthe heart of the problem. Many teachers, youngwomen in particular, are understandably reluctantto move to remote areas, especially when theyare characterized by high levels of poverty andlack transport, health services and other facilities.Teachers may be similarly reluctant, for careerreasons, to serve in what are seen as failingschools. Experienced teachers may use theirseniority to get assigned to the smallest classes(often in higher grades), leaving the largest classes,where the marginalized are at particular risk ofdropping out, to the least experienced or leastqualified teachers.

Changing patterns of recruitment anddeployment can help overcome the problemsthat marginalized children face. As the followingexamples demonstrate, it is important toencourage people from marginalized communitiesto become teachers as well as to ensure thatthe most experienced teachers are allocatedto underperforming areas and schools:

Recruit teachers from marginalized groups.Recruiting from marginalized groupscan promote positive identities, combatdiscrimination and ensure that children learnin their own language. But expanding suchrecruitment is not straightforward. Somecountries give ethnic minorities preferentialaccess to teacher training. This approach hasachieved some success in Cambodia, whichwaives the grade 12 entry requirement forcandidates from areas where upper secondaryeducation is unavailable. Increasing the pool ofteachers from ethnic minorities has been foundto have benefits in terms of their understandingof the local culture and motivation to stay inremote areas, as well as ensuring they are ableto teach effectively in the vernacular language(Benveniste et al., 2007).

Ensure that teachers are deployed to the schoolswhere they are most needed. Uneven distributionof teachers can result in shortages, particularlyof qualified teachers, in the most disadvantaged

regions and schools. Even in countries thatallocate teachers on the basis of studentnumbers, teachers can find ways to avoid difficultpostings. In Indonesia, which uses a nationalformula for teacher deployment, there aremarked inequities across schools and districts.For instance, 68% of urban primary schools havetoo many teachers, while 66% of remote primaryschools have shortages (World Bank, 2008f).Some governments have adopted strategiesand rules aimed at achieving more equitabledistribution:

– Better access to and use of data on pupil/teacher ratios in the Philippines has helpedreduce disparities in teacher deployment.Using a ‘rainbow spectrum’ to make disparitiesvisible, districts are colour-coded accordingto pupil/teacher ratios. Making the informationreadily available and easily understandablehas led to better channelling of new teachingpositions to shortage areas and systematictransfer of vacant teaching positions fromsurplus to shortage areas. As a result, all7,237 new teaching posts created in 2006 wereallocated to red or black zone schools, namelythose most in need (World Bank, 2006e;UNESCO, 2007).

– In Eritrea, many teachers start their careers aspart of their national service, which facilitatesenforcement of deployment rules. Teachersare allocated at the national level to one ofthe country’s six regions, then to schools withinthe region. They have no choice of location.This has resulted in a more even distributionof teachers. Average pupil/teacher ratios rangefrom 30:1 to 53:1, with the most rural regionshaving the lowest ratios. However, the leastexperienced teachers are allocated to themost challenging schools (Mulkeen, 2009).

Provide financial incentives. More equitablerules for teacher deployment may not be enough.Financial and other incentives – such as hardshipor travel allowances, subsidized housing, studyleave and training opportunities – are oftenrequired to encourage teachers to go todemanding schools or to areas with difficultliving conditions. Incentives need to be highenough to attract good teachers. Evidence fromseveral countries shows that the incentivesoffered for teaching in marginalized areas areoften too limited to have much effect (Kelleher,2008; Mulkeen, 2009; Mulkeen and Chen, 2008;UNESCO, 2008a). In Bolivia, teachers receive

Some countries,such as Cambodia,give ethnicminoritiespreferential accessto teacher training

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

1 9 8

extra pay for teaching bilingual studentsand working in rural areas, but on averagethe bilingual bonus is 0.3% of annual salaryand the rural bonus 1.1% (Vegas and Umansky,2005). Such low incentives are unlikely todeliver results. Ultimately, inducements forrelocation have to be seen by well-trained andexperienced teachers as adequate compensationfor transfer. The more successful examplesinclude the following:

– In the Gambia, a special allowance wasintroduced in 2006 to attract and retainteachers in schools more than 3 km froma main road. The allowance represents 30%to 40% of average salary. By 2007, 24% ofteachers in several regions had requestedtransfers to hardship posts, with negligiblenumbers requesting transfers in the oppositedirection (Mulkeen, 2009).

– In Mozambique, bonuses are aimed atattracting the most experienced teachersto remote areas. Schools are placed into fourcategories, from urban to the most isolated,and teachers are paid a bonus depending onschool location and their qualifications.Bonuses effectively double the salary of themost qualified teachers; the least qualifiedreceive no bonus (Mulkeen and Chen, 2008).

– In Uganda, a recent study on teacher attritionfound housing to be a key factor in assuringretention, especially in rural areas. Thegovernment responded by allocating a grantfor the construction of teacher housing in2005 (Mulkeen and Chen, 2008).

– Several Latin American countries haveintroduced incentive packages including careerdevelopment to encourage teachers to workin remote areas. For example, teachers livingin isolated areas of Ecuador get not only abonus but also priority in being granted tenure.The incentives have helped reduce disparitiesin pupil/teacher ratios, but have also tendedto attract the least experienced teachers toremote areas (Mpokosa and Ndaruhutse, 2008).

Train teachers to address marginalization.Beyond recruitment and deployment, teachersneed the skills to address marginalization inthe classroom. Brazil’s FUNDEF programmedevoted 60% of its resources to recruiting andtraining more teachers in poorer states.Qualified teachers helped students to avoidgrade repetition and dropout, and possibly alsoto enter the first grade on time (Vegas, 2007).

Even experienced teachers need training tochallenge attitudes to the marginalized andto equip them to teach effectively in classroomswith children from a diversity of backgrounds.This rarely happens, however; when it does,the initiative often comes from non-stategroups, reflecting inability or lack of intereston the part of governments. In some cases,partnerships between state and non-stateactors have emerged. In the Amazonian regionof Peru, the Programa de Formación deMaestros Bilingües de la Amazonía Peruana,a teacher-training programme co-directedby the Ministry of Education and an indigenousorganization, led to non-indigenous andindigenous experts cooperating to train bilingualteachers and familiarize them with indigenousculture (López, 2009).

Ability grouping seldom helps the marginalizedClassroom practices often reinforcemarginalization. An example is the separationof children into ‘ability’ groups at an early age.Children from disadvantaged backgrounds maybe more likely to be assigned to low ability groups,sometimes because of language problems. Oncein a low ability group, disadvantaged learners oftenfall further behind. Evidence from rich countriesstrongly suggests that grouping children by abilityearly in the education cycle reduces equity and canlead to weaker overall results (Duru-Bellat, 2009;Lleras and Rangel, 2009). Research using datafrom the Early Childhood Longitudinal Study in theUnited States shows that, among African-Americanand Hispanic students, reading achievement gainsmade in the first grade are lower for students whoare assigned to low-ability groups than for studentswith similar characteristics who are taught in non-grouped classes (Lleras and Rangel, 2009).Similarly, research in France shows that studyingin a mixed-ability class helps weaker students andthat removing streaming has a strong equalizingimpact on achievement (Duru-Bellat, 2009).

Tracking, or separating children into differenttypes of school (such as vocational versus generaleducation) according to academic ability at thesecondary level, also has adverse consequences.A study based on data from the TIMSS, PIRLS andPISA assessments, covering forty-five mostly OECDcountries, finds that the effect of early trackingaccounts for one-quarter of the ‘equality gap’between the most inequitable and most equitablecountry, and is also associated with lower mean

Teachers needtraining to

challenge theirattitudes to the

marginalized andto equip them toeffectively teach

children from a diversity ofbackgrounds

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

1 9 9

performance (Hanushek and Wößmann, 2006).In Germany, early tracking seems to be a factorbehind the country’s large education inequalitiesand particularly the marginalization of Turkishyouth (Crul, 2007). Recognizing the equityimplications of tracking, many European countriesadopted a unified secondary school system in the1960s and 1970s. There is evidence that the moveweakened the link between family backgroundand educational attainment, with associatedbenefits for those who would have been sentto the lower tracks (Brunello and Checchi, 2007).

The effects of academic segregation and trackingare widely debated. Evidence from developingcountries is both fragmented and limited. However,there are strong equity grounds for planners inrich and poor countries alike to avoid early trackingand to treat academic selection within schoolswith caution. Both can reinforce exclusion.

Targeting financial and pedagogical supportto disadvantaged schoolsOne way of targeting marginalized children is totarget their schools. Targeting criteria can includelocation, ethnolinguistic composition or the shareof poorly performing students, with governmentsusing a range of regulatory instruments andfinancial mechanisms to raise standards. Moreintensive support to teachers and school heads,more specialized pedagogical support to studentsand more per student financing are among theoptions. One targeted programme in Uruguayis credited with improving learning outcomesin some secondary school grades by up to 30%by combining financial and pedagogical support(Cerdan-Infantes and Vermeersch, 2007; Crouchand Winkler, 2008).13 In Chile, the 900 SchoolsProgramme provided intensive support to theworst-performing 10% of elementary schoolsby training teachers, gearing courses to studentslagging behind or with behavioural problemsand providing textbooks. Evaluations have shownthat grade 4 test scores improved significantlyfor students in the programme, mainly as a resultof the introduction of more appropriate pedagogicalpractices in the classroom and facilitationof a cooperative environment within schools(García-Huidobro, 2006).

Not all school-based targeting has produced suchpositive results. For almost three decades Frenchgovernments have given additional support to Zonesd’Éducation Prioritaire serving disadvantagedstudents. In 2008, around 16% of secondary school

students were in schools with ZEP status. Theseschools have more teachers, so class size is lowerand students receive additional support. In addition,ZEP teachers receive higher pay. Yet severalstudies have found only a limited impact onstudent achievement (Duru-Bellat, 2009). Whyhave ZEP schools not achieved better results?One reason is that the additional resources arespread too thinly over a large number of schools,so class size is reduced by only two students onaverage. Schools have also had trouble attractingexperienced teachers (Moisan, 2001). High teacherturnover makes it difficult to organize strategiesthat could improve achievement (Duru-Bellat,2009). A comparable programme in England(United Kingdom), Excellence in Cities, producedmore positive results, yet it too fell short ofexpectations (Box 3.18).

Experience from programmes targetingdisadvantaged schools shows that they can make adifference provided the level of additional financingis sufficient and they are accompanied by incentivesto attract and retain qualified teachers.

Learning in an appropriate language and through a relevant curriculumInclusive education for ethnic and linguisticminorities requires schools that offer a relevantcurriculum in an appropriate language. Sitting ina primary school classroom listening to a teacherproviding instruction in a language they do notunderstand is a short route to marginalization.Bilingual education facilitates learning in a familiarlanguage and equips students with the nationallanguage skills they need to make the transitionto secondary school and, eventually, to employmentand full participation in social and political life(Alidou et al., 2006; Dutcher, 2004; UNESCOBangkok, 2008).

Evidence from several countries in sub-SaharanAfrica demonstrates that bilingual education canimprove learning achievement. One example comesfrom the Écoles Bilingues created in Burkina Fasoin the mid-1990s. After five years of instruction inlocal language and French, 85% of pupils in theseschools successfully passed the primary schoolexamination in 2002, compared with a nationalaverage of 62% (Alidou et al., 2006). In Zambia,the successful introduction on a pilot basis of locallanguage teaching in the late 1990s was followedin 2002 by reforms that introduced seven locallanguages into primary school education (Alidouet al., 2006; Linehan, 2004). Ethiopia has gone

13. Learning assessmentswere used to identify weakerschools. Teachers in selectedschools received intensivetraining together with on-going support throughoutthe year and were paid anincentive. The schooltimetable was lengthenedfrom half a day to a full day.

Evidence fromseveral countriesin sub-SaharanAfricademonstrates that bilingualeducation canimprove learningachievement

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

2 0 0

further than many countries, seeking to combinemother tongue instruction with Amharic andEnglish in grades 1 to 8. One recent review oflearning assessment data concluded that ‘thoseregions with stronger mother tongue schoolinghave higher student achievement levels at Grade 8in all subjects, including English’ (Heugh et al.,2006, p. 6). In Mali, bilingual schools have beenassociated with large declines in dropout andrepetition (World Bank, 2005c).

Overcoming underlying causes of marginalizationassociated with language requires more thanbilingual provision. Language is wrapped up withcultural identity and schools have a vital role to playin addressing the social attitudes that devalue somecultures. That is why education reform in someLatin American countries has sought to combineintercultural and bilingual education. In Bolivia,reforms that started in the mid-1990s introducedintercultural and bilingual education on a nationalscale for the three most widely used indigenouslanguages. Bilingual teaching expanded rapidly,from 75,896 pupils in 1997 to 192,238 in 2002, or11% of all primary school pupils (Sichra Regalsky,n.d.). Alongside this change, curriculum reforms ledto the development of courses and textbooks thatattach more weight to the country’s multicultural

history and the role of indigenous peoples.In other countries, intercultural and bilingualeducation has suffered from poor design andweak implementation, with intercultural educationreceiving particularly limited attention. In Peru,which pioneered the approach in the region,it is largely limited to indigenous communitiesin remote rural areas, and many nominallyintercultural and bilingual schools offer no teachingin indigenous languages (Cueto et al., 2009).

Education systems can be instrumental inovercoming marginalization arising from languagedifficulties. The starting point is to align the rulesgoverning education with broader principles ofinclusion. Many countries have not yet done this.In the Lao People’s Democratic Republic, theconstitution forbids discrimination betweenethnic groups and emphasizes the importanceof expanding education in ethnic areas. Yet it alsoestablishes Lao as the official language, includingof instruction in school – an arrangement thatarguably discriminates against children fromthe 27% of the population that does not have Laoas the mother tongue (Benveniste et al., 2007).Legal recognition of the entitlement to be taughtin a familiar language is an important principlestill lacking in many countries.

Schools have a vital role

to play inaddressing the

social attitudesthat devalue

some cultures

England’s Excellence in Cities programme was aimedat improving pupil achievement in deprived urbanschools. Introduced on a pilot basis in 1999, it wasextended nationally until 2006. The programmereached in particular children from non-whitebackgrounds, those with English as an additionallanguage, those entitled to free school meals andchildren identified as having special education needs.

Eligible schools received higher than averagesupport per student. In 2005, this amountedto £120 per pupil per year, only 4.4% above theaverage allocation. Institutional support includedfour core elements, although specific interventionsvaried by setting. Local partnerships encouragedschools to work together in developing needsassessments and strategies. Learning Support Unitsassisted students failing to achieve academically andexperiencing behavioural problems. Mentors wereprovided to children making slow progress inlearning. A separate part of the programme soughtto identify and support ‘gifted and talented’ children.

Evaluations revealed some positive outcomes. Thegreatest impact was on mathematics achievement atage 14. Within the most deprived schools, however, theimpact was greatest for children previously achievingmedium and higher scores. No impact was found forstudents using support units and students with amentor at age 14 made less progress than thosewithout. Pupils designated as ‘gifted and talented’registered higher levels of achievement, but therewas no evidence of an Excellence in Cities effect.

One possible explanation why this programme failedto achieve stronger outcomes is that insufficientadditional finance was provided. Another factor isthat schools in deprived urban areas, includingthose covered by Excellence in Cities, were finding itincreasingly difficult to recruit and retain experiencedteachers. More fundamentally, it appears that theinitiative failed to override the wider structures ofdisadvantage in the home and beyond that pushchildren towards educational marginalization.

Sources: Vignoles (2009); Kendall et al. (2005).

Box 3.18: Achieving ‘Excellence in Cities’? A targeted intervention to support deprived urban schools in England (United Kingdom)

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

2 0 1

Delivering effective bilingual education requiresthe development of institutional capacity to trainbilingual teachers. This is an area in whichnational targets are often delinked from publicspending allocations and a longer-term strategyfor change. One reason Ecuador has been ableto deliver strong bilingual teaching is that it hasestablished five specialized teacher-trainingcolleges. Similarly, Bolivia has created threeindigenous language universities to supportbilingual training (López, 2009).

Children often enter classrooms weighed down bylow self-esteem and facing low expectations fromteachers. Schools can play an important role inchanging this situation. Having teachers from amarginalized community can help widen children’shorizons and raise their ambitions. And teachersthemselves can be trained to understand theproblems faced by ethnic minorities. The AustralianGovernment has set ambitious targets forovercoming disparities between Aboriginal childrenand the rest of the population. One is to halvethe gap in reading, writing and numeracy withina decade. Local initiatives provide pedagogicaland curriculum support to address marginalization

within the classroom. A pilot programme,Deadly Ways to Learn, has sought to buildrespect for Aboriginal languages (Box 3.19).

Curriculum reform and intercultural educationare not just about reaching the marginalized.They are also about combating marginalizationby challenging the stereotypes and the invisibilitythat sustain it. Textbooks can reinforce gender,racial and ethnic stereotypes that narrow thehorizons of many children. Intercultural educationhas a key role to play in building respect fordifferent cultures, combating prejudice, raisingawareness about social inequalities and fosteringdebate (Luciak, 2006).

Reaching children with disabilitiesRules, attitudes and systems that are unresponsiveto the needs of children with disabilities oftendeny these children an opportunity for education.Excluding children with disabilities restricts theirchoices, making it more likely that they will livetheir adult lives in poverty, and has wider costs forsociety. No country can afford an education systemthat limits the potential of millions of children tocontribute to social, cultural and economic life.

Ecuador has beenable to deliverstrong bilingualteaching byestablishing fivespecializedteacher-trainingcolleges

Aboriginal children in Australia face languageproblems at school that had escaped officialrecognition until recently. The 2006 census indicatedthat about 11% of the indigenous population aged 5to 19 speaks an indigenous language at home. Therate rises to 17% in remote Australia and 58% in veryremote Australia. The shares are likely to be greaterstill for Aboriginal English, which many consider adialect separate from the Standard Australian Englishtaught in primary schools, with a distinctive grammarand vocabulary. While most Aboriginal children enterschool speaking English, they often have no ideathat their language is different until teachers tellthem that it is wrong or inappropriate.

Language problems go beyond the classroom.Aboriginal languages have often been seen asinferior and subjected to ridicule, reflecting widerprejudices about culture, lifestyles and ability to learn.Language problems have often made it difficult forAboriginal children to understand lessons, absorbinformation and realize their potential in tests. Theresult has been a vicious circle of underachievement,with teachers often mistaking a language problemfor a learning difficulty.

The Deadly Ways to Learn programme is an attempt to change the ways teachers view Aboriginallanguages. It began as a pilot project in fourteengovernment, private and Catholic schools acrossrural and urban Western Australia. The name is aplay on ‘deadly’, which Aboriginals use in the sameway Standard Australian uses ‘great’. The projectincluded the preparation of books such as DeadlyWays to Teach and Talking Deadly to introduceteachers to the culture, identity and history thatinform Aboriginal language. Aboriginal educationofficers provide support and guidance to teachersin the selected schools. Curriculum and textbookreforms are also involved.

The programme highlights the importance of allstudents in Australia receiving an education thatis sensitive to the history, culture and languageof indigenous Australians, and that also takes intoaccount the backgrounds of people from otherminority groups. Schools have to become moreeffective in promoting respect, tolerance andmulticulturalism, and in combating the prejudiceschildren bring to school.

Source: Biddle and Mackay (2009).

Box 3.19: Promoting respect for Aboriginal languages in Australia

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

2 0 2

Governments across the world have recognizedthat inclusive education for people with disabilitiesis a human rights imperative. The Conventionon the Rights of Persons with Disabilities, whichcame into force in 2008, has strengthened theentitlements and rights of those with disabilities.It requires governments to ensure that peoplewith disabilities have access to ‘an inclusive,quality and free primary education and secondaryeducation on an equal basis with others in thecommunities in which they live’ (United Nations,2008. Article 24, para. 2b). As of September 2009,seventy countries had ratified the convention.

Putting the principles of inclusive education intopractice requires action at many levels, startingwith information. Most developing countrieshave poor data on the number of children withdisabilities or the incidence of specific impairments.Government estimates are often inconsistent,reflecting not only problems in monitoring andrecording but also, in many cases, the invisibilityof people with disabilities and the indifference ofpolitical leaders (USAID, 2005). Some countriesare working actively to strengthen the monitoringof disability. One example comes from the UnitedRepublic of Tanzania, where a 2008 survey provideda detailed profile of the prevalence, distributionand pattern of impairments across the country.It found marked regional disparities and a higherincidence of disability in rural areas (UnitedRepublic of Tanzania Government, 2009).

Approaches to reaching people with disabilitiesvary. Many governments, parents and groupsrepresenting them continue to view special schoolsas the most viable option (Lang and Murangira,2009). One survey in Uganda found that disabilitygroups and parents favoured this approachpartly out of concern about overcrowding andpoor resourcing in standard schools (Lang andMurangira, 2009). In some cases, children withsevere impairments do need education inspecialized institutions. However, special schoolscan reinforce social exclusion, denying childrenwith disabilities the opportunity to interact withtheir peers who do not have disabilities, reinforcingstereotypes and segmentation in the process.

Integrating children with disabilities into thestandard education system is a preferred policyoption because it can break down the segregationthat reinforces stereotypes. But integration is nota panacea. Children with severe disabilities mayrequire highly specialized support. Moreover,

integrating children with disabilities into poorlyresourced, overcrowded schools with restrictedaccess to toilets and other facilities is not aprescription for inclusive education, especially whenteachers are not equipped to meet their needs.Placing deaf children in schools where none of theteachers can communicate in sign language will dolittle to alleviate their disadvantages. And very fewschools in the poorest countries, or even in middle-income countries, have access to Braille textbooksor teachers able to teach Braille. It is thereforecritical that moves towards integration are part ofa broader strategy encompassing teacher training,school financing and other measures.

Several countries are developing educationsystems that are more responsive to the needsof children with disabilities. The Lao People’sDemocratic Republic has a network of 539 schools– three for each district in every municipality andprovince – that teach children with disabilitiesalongside their peers and provide specializedsupport. The schools give children with specialneeds opportunities to learn in an inclusiveenvironment, partly through investment inspecialized teacher training. The experienceaccumulated through the programme is informingwider school reforms (Grimes, 2009). In SouthAfrica, the focus has shifted from special schoolsto inclusive education in mainstream schools.Authorities have to identify the level of supportrequired by individual learners with disabilities(South Africa Department of Education, 2005;Stofile, 2008). Research in Eastern Cape, oneof the poorest provinces, found that inclusiveeducation produced significant gains, ranging fromimproved physical access to support for specializedteaching practices and increased admission oflearners with disabilities (Stofile, 2008).

Non-government organizations have played animportant part and in many poor countries arethe primary source of education for childrenwith disabilities. Through active engagementwith children with disabilities, their parents andeducation authorities, such groups are producingresults that demonstrate what is possible. In2003, a Bangladeshi non-government organization,BRAC, established a pre-school and primaryeducation programme aimed at increasingparticipation by children with mild special needs.Training teachers, providing equipment, adaptingthe curriculum and improving physical access,it had reached about 25,000 children by 2006(Ryan et al., 2007).

‘We welcomechildren with

disabilities nowbecause

we know thatthey have the

same right to education

as the others.’Teacher,

Nicaragua

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

2 0 3

Some non-government organizations andgovernments, including those of Uganda andthe United Republic of Tanzania, have supported‘itinerant teaching’ approaches, which enablespecialized teachers in central primary schoolsto reach a larger group of pupils in satelliteschools, and support and train teachers(Lynch and McCall, 2007).

Several countries are also attempting to buildlinks between existing special institutions andmainstream schools, with the specialized schoolsproviding learning materials and aids, in-serviceteacher training and support personnel. InEthiopia, with the support of the non-governmentorganization Handicap International, a school fordeaf students operates as both a special schooland a resource centre, supporting educationfor deaf learners in other schools and thedevelopment of sign language (Lewis, 2009).

These experiences demonstrate the potential forscaling up local initiatives, but governments need todevelop national plans to extend inclusive educationfor children with disabilities, including detailedtargets, strategies for improving access andlearning achievement, and comprehensive plansfor providing financing and training teachers. Thestarting point for such a plan is a credible needsassessment based on a national survey of theprevalence of disability.

Entitlements and opportunities

Education systems can do a great deal to addressthe inequalities that restrict opportunity forchildren from disadvantaged groups. They canmake schools more affordable and accessible,create conditions for effective learning, and actas a vehicle for changing attitudes and beliefs thatstigmatize children and corrode self-confidence.But prospects for greater equity in educationultimately depend on what happens to childrenbeyond school, through the social and economicstructures that perpetuate marginalization.

This section looks at the interaction betweeneducation systems and policies in other sectors.It concentrates on two thematic areas. The firstconcerns the role of laws, norms and rules inempowering marginalized people. Legalinstruments, international as well as national,can enhance equity not just by setting standardsfor public policy, but also by enabling marginalizedpeople to claim entitlements. Political mobilization

by the marginalized and other civil society groupsis another way of broadening rights-based claims.

The second area is redistributive finance. Manychildren are marginalized in education becausetheir families are poor and particularly vulnerableto external shocks, such as drought or economiccrisis. The geographic and historical factorsunderlying regional disparities also limitopportunity. In many cases, the poverty andeconomic differences that lead to marginalization ineducation are linked to unequal power relationshipsand to disparities in financing. Redistributivefinance can help redress disadvantages associatedwith poverty and regional inequality. In particular,social protection can be instrumental in makingeducation more affordable and less susceptibleto the economic shocks that pull many poorchildren out of school.

Enforcing rights and lawsConcerns with equity and fairness inform ethicaldebates worldwide, crossing political, religiousand moral divides. The United Nations Charterencapsulates those concerns in its commitmentto universal human rights. Legal institutions andcodes enshrine equity in common law traditions(Kritzer, 2002). And political movements for socialjustice mobilize around agendas emphasizing equalopportunity, non-discrimination and fair distributionof resources. The combined weight of internationalhuman rights agreements, laws and politicalmobilization can act as a powerful catalyst forovercoming marginalization in education.

The Universal Declaration of Human Rights remainsthe foundation for international human rightsentitlements. The contemporary human rightsregime operating under United Nations auspicescomprises a broad array of instruments, many ofwhich set standards for rights in education. Theseinstruments collectively form a comprehensiveframework for extending opportunities to childrenfacing exclusion or discrimination in educationon the basis of gender, race, ethnicity, languageor poverty (see Annex, p. 292).

International conventions and wider human rightsinstruments set norms, define shared principlesand establish an institutional framework foradvancing broad-based civil, political, social andeconomic rights. Principles of international laware often embedded in national legal codes andconstitutions. Yet more could be done to useinternational human rights agreements to empower

Politicalmobilization by the marginalizedand other civilsociety groupsis another way of broadeningrights-basedclaims toeducation

one example, the European Court of Human Rightshas ruled that the Czech Republic’s treatmentof Roma children is not legal because the policiesamount to de facto segregation (Box 3.20). Inthe United States, education campaigners havemounted legal challenges aimed at securinggreater equity in the distribution of public finance,along with wider institutional reforms (Box 3.21).

Both instances illustrate the importance of legalentitlements that can be used to hold governmentsaccountable. Many countries’ constitutions includethe right to free, non-discriminatory education forall, but constitutional principles are not alwaysenforceable. Article 45 of India’s constitutionmentions ‘free and compulsory education for allchildren’ up to 14 years but this ‘directive principle’could not be enforced in court. The Right ofChildren to Free and Compulsory Education Actadopted in 2009, however, now legally requiresstates to provide free education to children aged 6to 14 and reserves 25% of private primary schoolplaces for disadvantaged children (Economic andPolitical Weekly, 2009; India Ministry of Law andJustice, 2009).

The entitlement to a formal identity is a criticalasset for achieving greater equity in education.The Convention on the Rights of the Child requiresall signatories to guarantee the formal identity ofchildren through birth registration. Yet UNICEF

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

2 0 4

the marginalized. Ratification of United Nationsconventions often fails to lead to action that helpsthe marginalized. Part of the problem is that thecommittees overseeing the conventions have forthe most part failed either to hold governmentsto account or to provide transparent and publicassessments of national policies. The Committeeon the Rights of Persons with Disabilities, theindependent body of experts overseeing thenew convention, needs to provide a more robustdefence of human rights entitlements.

National legal systems have played a crucialrole in addressing equity and marginalization ineducation. A landmark ruling in the developmentof civil rights in the United States was the 1954decision in Brown v. Board of Education. TheSupreme Court determined that laws separatingchildren of different races into different schoolsviolated the equal protection clause of theAmerican constitution. The principles applied inthis case were subsequently extended to challengesegregation in other areas. Brown thus servedas a milestone in the struggle of African-Americansto gain equal civil and political rights.

Recourse to law offers marginalized groups anopportunity to contest discriminatory andinequitable practices. As was the case with Brown,legal rulings can have wider importance becauseof the general principles they establish. To take

India now legallyrequires states to provide free

education

The European Court of Human Rights has ruled on severalcases in which governments have been accused of violatingthe education rights of Roma children. Echoing themesraised in Brown v. Board of Education, the court has appliedthe principle of non-discrimination to cases of segregation.

Roma children across Europe are often assigned to ‘specialschools’ with little attention to their education needs.Cultural bias and discrimination by teachers and educationauthorities is widespread. In D. H. and others v. the CzechRepublic the court was asked to pass judgment on a casebrought by eighteen Czech nationals of Roma origin living in the Ostrava region of the Czech Republic who had beenassigned to schools for children with learning difficulties.Represented by the European Roma Rights Centre, theplaintiffs argued that the assignment was discriminatoryand therefore contravened the European Convention onHuman Rights. Evidence was presented that 56% of thechildren enrolled in special schools in Ostrava were Roma,and that half of all Roma children attended such schoolscompared with less than 2% of non-Roma children.

In 2007 the court ruled that such statistics, although not completely reliable, established a presumption ofindirect discrimination. This shifted the burden of proof to the defendant, who failed to show that the difference in treatment had an objective and reasonable justificationunrelated to ethnic origin. The court ruled that theassessments through which Roma children were selectedfor special schools were flawed, notably in failing toconsider linguistic and socio-economic conditions.

How successful was the case in addressing Romamarginalization? The trial provided a focal point for Roma and wider human rights groups and the judgmentestablished an important principle, but the European Roma Rights Centre has claimed that the Czech authoritieshave done little since to address segregation.

Sources: de Beco and Right to Education Project (2009); European Roma Rights Centre (2008).

Box 3.20: Roma children’s right to education — using the law to challenge the state

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

2 0 5

estimates that 51 million births per year gounregistered (UNICEF, 2007c). The lack ofregistration means parents and children may nothave the documentation they need to claim a placein school, establish an entitlement to stipends orvotes, or seek legal redress. Failure to registerbirths can also mean the most marginalizedchildren are bypassed in national statistics,rendering them invisible to policy-makers.

Several governments have demonstrated thatregistration gaps can be closed. In 2009, BurkinaFaso initiated a one-year programme aimed atregistering 5 million people, most of them womenand children, by providing free birth certificates(Integrated Regional Information Networks, 2009).Furnishing documentation does not have to beexpensive. Senegal’s drive to supply modern identitycards to all citizens over 15 is estimated to havecost just US$0.61 per recipient (Levine et al., 2008).

Legal instruments can also make a differencefor the millions of young girls every year who facehaving their education disrupted or terminated byearly marriage. By one 2005 estimate, almost halfof South Asian females aged 15 to 24 were marriedbefore age 18. Poverty, tradition and unequal powerrelationships between men and women all playa part in early marriage (Levine et al., 2008).These issues have to be addressed on many fronts,

but legal prohibition of early marriage, coupled withincentives to keep girls in school and campaigns tochange attitudes, can establish norms and a basisfor legal recourse.

Wider political mobilization is importantLegal provisions cannot be considered in isolation.Brown v. Board of Education was the culminationof a decade-long struggle by African-Americansand sympathetic whites against segregation andother discriminatory laws. The legal principles thatthe Supreme Court laid down were a landmark.But it was the civil rights movement that made theruling such a powerful force for change. Politicalmobilization, involving the marginalized and widersocial movements, has been essential in reforminglaws and rules on education.

Political mobilization against marginalization canbecome part of a wider movement. One strikingexample comes from Bolivia, whose educationsystem systematically reinforced subordination ofindigenous people. The 1994 Education Reform Lawhelped establish indigenous people’s right to learnin their own language and brought multiculturalisminto the curriculum. Education reform in turnplayed a role in political processes that broughtan indigenous political leader to power in 2005.Reforms have seen the strengthening of IndigenousEducation Councils, which held their own congress

Legal instrumentscan make adifference for the millions ofyoung girls everyyear who havetheir educationdisrupted by earlymarriage

Education groups in the United States have taken to the courts to address a wide range of concerns. The results have been mixed.

In Campaign for Fiscal Equity, Inc. v. State of New York,plaintiffs claimed the state’s school finance systemunderfunded the New York City public schools, therebydenying students their constitutional right to theopportunity for a sound basic education. Evidence waspresented that areas with high poverty, learners withdisabilities and large numbers of students learningEnglish faced special problems. After ten years ofproceedings, the courts finally found in favour of theplaintiffs. In 2007, the New York State Legislature enacted the Education Budget and Reform Act, increasingeducation funding by an unprecedented amount andestablishing transparency and accountability measuresfor the distribution of funds and school finance reform.

Antoine et al. v. Winner School District. This case involveda class action lawsuit brought by the American CivilLiberties Union on behalf of Native American students in South Dakota. Among other issues, the suit charged

that the school district disproportionately targeted NativeAmerican students for disciplinary action and maintainedan educational environment hostile to Native Americanfamilies. In 2007, a federal court approved a settlementrequiring the district to undertake institutional reforms,including hiring a full-time ombudsperson, nominated bythe Native American community, to serve as liaison withthe community and work with school officials, especiallyon disciplinary issues. Authorities also agreed to providetraining for teachers on ‘unconscious racial bias andeducational equity’, and to include Native Americanthemes in the curriculum.

Other cases with less positive outcomes includeHorne v. Flores, in which the Supreme Court in June 2009reversed a federal court decision upholding minimumstandards and necessary resources for the educationof English-language learners in Arizona primary schools,which have a very large population of Latino students.

Sources: Campaign for Fiscal Equity (2009); Child Rights InformationNetwork (2009); Orfield and Gándara (2009).

Box 3.21: Recent legal challenges to educational marginalization in the United States

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

2 0 6

in 2004 and have submitted proposals aimed atbroadening and strengthening multiculturalismin Bolivia’s schools (Gamboa Rocabado, 2009;Howard, 2009; López, 2009; Luykx and López, 2007).

The Bolivian experience draws attention to abroader feature of the interaction between politicsand law in combating marginalization in education.Political mobilization is important because it givesa voice to social groups facing discrimination andstigmatization. In New Zealand, the ko-hanga reolanguage movement provided a social, politicaland cultural focal point for empowerment of Ma-oripeople. Political mobilization has contributed todevelopment of a more multicultural educationsystem, which in turn has extended opportunitiesfor Ma-ori children (Box 3.22). In Bangladesh, anational non-government organization called NijeraKori (‘We do it ourselves’) has helped landlesslabourers, primarily women, strengthen their abilityto claim rights and entitlements (Chronic PovertyResearch Centre, 2008).

Political mobilization can also pose risks. Themarginalized are not a homogenous group, andpolitical parties, social movements and non-government organizations take up their problemsunevenly. In India, the rise of political parties

representing low-caste groups in northern stateshas been described as a ‘silent revolution’(Jaffrelot, 2003, p.10). Yet that revolution has donelittle to address poor schooling for low-castechildren, suggesting that political priorities havebeen in other areas (Mehrotra, 2006). Some highlymarginalized groups have a weak voice even withinbroad-based civil society lobbies seeking improvedaccess to education. The rural poor, ethnic minoritywomen, children with disabilities, slum dwellersand children in conflict zones are groups whosecauses have not been widely or effectively taken up.

Social protection: conditional cash transfers and beyondHousehold poverty is one of the most potentfactors in education marginalization. If a poor familyis hit by a disaster such as a drought, a flood,unemployment or a serious illness, it may have nochoice but to take children out of school. By helpingpoor people manage risk without compromisinglong-term welfare, social protection programmescan also broaden opportunities in education.

Such programmes take many forms. They rangefrom cash transfers to employment-based safetynets and interventions to support nutrition. Inaddition to reducing destitution, such programmes

By helping poor people

manage risk, social protection

programmes can broaden

opportunities in education

New Zealand’s ko-hanga reo movement has demonstratedwhat a powerful force indigenous language revitalization can be, not only for education but also for social cohesion.

In the 1970s, the Ma-ori language was on the edge ofextinction. A grassroots movement arose to save the languageby educating a new generation in total-immersion ‘languagenests’ (from which the movement takes its name). Today it is a national institution widely credited with sparking the language’s revival and fuelling a powerful assertion of Ma-ori identity in almost all walks of national life.

The concept is simple. Ma-ori under age 6 get their pre-schooleducation in a community- and family-based environmentwhere only Ma-ori is spoken. They spend their early yearssurrounded by the culture and values of their people. Ko-hanga reo are typically found in church halls, schools and marae, traditional Ma-ori community centres. Like manysocial movements, this one started small. It was begun in 1981by the government’s Department of Ma-ori Affairs but grewquickly as a grassroots, mostly volunteer-run movement.Thirteen years later there were 800 ko-hanga reo catering for 14,000 children.

With their ethos of self-help and commitment to continuityacross generations, ko-hanga reo became a source ofinspiration for young Ma-ori parents, many of whom could not speak their ancestral language. The movement nurtured a generation of bilingual Ma-ori speakers, with alumni numbers estimated today at 60,000. In 2008, one-quarter of all Ma-ori children enrolled in early childhood programmeswere in ko-hanga reo.

As graduating Ma-ori speakers turned 5 and started school,they generated demand for Ma-ori immersion schools (kura kaupapa). Today, there are sixty-eight kura kaupapawith 6,000 students. Year 11 Ma-ori students in immersionschools have recorded significantly better achievement ratesthan their Ma-ori peers in English-medium schools.

Ko-hanga reo have not solved the marginalization in educationthat many Ma-ori children experience. Ma-ori youth are stilltwice as likely as their non-Ma-ori counterparts to leave schoolwith no qualification. But the movement has played a crucialrole in challenging discrimination and forging a moremulticultural national identity.

Sources: Te Ko-hanga Reo National Trust (2009); New Zealand Ministry of Education (2008a, 2008b).

Box 3.22: New Zealand’s Ma–ori Renaissance

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

2 0 7

can create incentives supporting children’seducation, health and nutrition. They can betargeted not just at the very poor, but also atthe most marginalized groups or regions.

Cash transfer programmes have grown enormouslyover the past decade. Many of these programmesare conditional on specific behaviour, such askeeping children in school and attending healthclinics. In some countries, including Brazil andMexico, nationwide social assistance programmestransfer between 1% and 2% of national income totargeted households. In other countries, conditionalcash transfer programmes are more localized andoften project-based. The degree to which educationfigures in transfer conditionality and support varies.Some social protection programmes provide directsupport for education, including stipends, bursaries,fee waivers and funding for transport and books(Grosh et al., 2008); (see previous section). Inother cases, the education benefits associatedwith social protection are incidental, resultingfrom employment creation, nutrition programmesor other measures that enable households to getthrough difficult periods.

Comparisons have to be made with some cautionbecause of data constraints, and differences inevaluation methodology and in the programmesthemselves. Even so, evaluations of socialprotection programmes point to wide-rangingpositive effects (see Annex, p. 294).

Evaluations of social protection programmes havedocumented a range of positive effects, albeit withmarked variation across countries and groups. InMexico, Oportunidades has had a significant impacton children making the transition from primary tosecondary school, especially in rural areas (Fiszbeinet al., 2009). Nicaragua’s Red de Protección Socialwas targeted at children aged 7 to 13 who had notyet completed grade 4 of primary school. Evaluationresults indicated a thirteen percentage pointincrease in school enrolment, with the extremepoor registering the most marked gains (Villanger,2008). Employment guarantee programmes havealso delivered results, often in contexts markedby deep poverty and acute vulnerability. Ethiopia’sProductive Safety Net Programme is an example.Evaluations suggest that around 15% of cashpayments have gone to education, while half ofbeneficiary households report being able to keepchildren in school longer as a result of thetransfers (Slater et al., 2006) (Box 3.23).

Social protection is not a simple antidote tomarginalization. Levels of poverty, financingcapacity and institutional factors have a bearingon the type of social protection intervention likelyto deliver results in various contexts. The costand effectiveness of any programme will beshaped by factors such as:

the scale of transfer;terms of the transfer; andtargeting of beneficiaries.

The scale of transfer. Transfer levels varyconsiderably. One survey found that transfersranged from around 8% to 23% of the nationalpoverty line in Latin America and from 5% to 30%in sub-Saharan Africa (Yablonski and O’Donnell,2009). Large-scale conditional cash transferprogrammes in Brazil and Mexico have had amarked effect on poverty partly because the moneythey provide represents a significant increment inthe income of the very poor. Ethiopia’s ProductiveSafety Net Programme boosts child educationand reduces child labour when the transfersto households are sufficiently large.

When it comes to supporting poor and vulnerablechildren, more is clearly better. But policy-makersalso have to consider the marginal benefit ofincreasing transfers and the potential trade-offbetween reaching more people and providing larger transfers. In the Cambodia Education Sector Support Project scholarship programme, the 25% of students deemed most at risk of dropping out receivedUS$60 and the group next most at risk US$45.Comparing beneficiaries with non-beneficiaries,an evaluation found that while the US$45 transfersignificantly increased the probability of a girl beingin school, the additional US$15 had a modestadditional effect (Filmer and Schady, 2009). In otherwords, in this case there were diminishing marginalreturns to the investment.

Terms of the transfer. Many social protectionprogrammes provide cash transfers to createincentives for behavioural change. To put it crudely,parents get paid for keeping children in school,taking them to health clinics and presenting themfor weighing at nutrition centres. The size oftransfer influences the strength of the incentivecreated by this conditionality. Giving transfers towomen can result in a higher share of the moneybeing directed towards children – especially girls –than may be the case when men receive thetransfers (Kabeer, 2005).

Half of beneficiaryhouseholds inEthiopia keepchildren in schoollonger as a resultof the ProductiveSafety NetProgramme

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

2 0 8

Unconditional transfers can also generate strongbenefits. In Zambia, a pilot unconditional cashtransfer programme supported by German aidinvolved two districts, Kalomo and Kazungula,marked by large out-of-school populations andhigh levels of poverty. It resulted in significantdeclines in absenteeism among children from poorhouseholds in Kalomo and an increase in spendingon education in both districts (UnderstandingChildren’s Work, 2009). Thus, social protectioncan have an effect even in countries unable toimplement and monitor conditional transfers.

School feeding programmes provide another formof social protection. The World Food Programmeestimates that 59 million primary students attendschool in a state of malnutrition, with 23 millionof them in sub-Saharan Africa alone (World FoodProgramme, 2009). Well-designed school feedingprogrammes that include micronutrient fortification

and deworming provide significant nutritionalbenefits. They can increase school attendanceand educational achievement (Bundy et al., 2009b;Kristjansson et al., 2007; Miguel and Kremer, 2004).Many programmes incorporate a strong genderdimension by making special provision for girls’nutrition. One survey in sub-Saharan Africacovering 32 countries and 4,000 primary schoolsreceiving World Food Programme support foundthat school feeding had marked benefits on schoolparticipation (World Food Programme, 2007).

What is less clear is the scale of the benefits andthe most effective delivery mechanism. Schoolfeeding programmes raise many of the same issuesfor policy-makers as social protection in otherareas. The key to success is equitably and cost-effectively delivering an adequate incentive in termsof the amount of rations provided. There is someevidence that programmes combining take-home

Well-designedschool feeding

programmes can providesignificant

nutritional andeducational

benefits

Ethiopia’s Productive Safety Net Programme is thelargest social protection programme in sub-SaharanAfrica outside South Africa. Launched in January2005, the Productive Safety Net Programme nowprovides regular cash or food transfers to more than7 million people whose food sources are unreliable.It has produced significant benefits for education.

The programme aims to protect highly vulnerablepeople against shocks and to build their assets. During periods of stress, one adult per household isguaranteed the option of working in an employmentprogramme that provides payment in cash or in kindas food. In effect, the programme is a social insurancemechanism. It offers people a chance to manage riskwithout having to sell productive assets, cut spendingon nutrition or take children out of school. It hasbenefits affecting education at various levels:

Children’s participation in education. Data for2006 indicate that about 15% of cash from theprogramme was used for education purposes.By 2008, spending on education was the mostcommon type of investment of programmeresources. Financial support has enabled manyfamilies to deal with shocks without taking childrenout of school. Half of the households interviewed in 2006 reported keeping their children in schoollonger rather than withdrawing them when cashor food was short; and one-third enrolled moreof their children in school. The benefits werestrongest in districts where transfers were incash rather than food.

Classroom construction. The public workscomponent of the programme has includedclassroom construction and upgrading of schools.In some villages, construction of classrooms hasallowed schools to add a grade, enabling pupils tostay on for another year and reducing the attritionassociated with transition to more distant schools.

Health and nutrition. Almost a third of recipientsspend cash from the programme on health servicesand the public works component has helped buildlocal clinics. The programme bolsters health andnutrition — receiving a relatively high transfer from the programme reduces the likelihood of low calorific intake by over ten percentage points.

Set against these positive outcomes are someimplementation problems. Employment-based supportcan create incentives for child labour. One study foundthat about 8% of workers in the programme wereunder 18. In families facing tight labour constraints,low transfers only partially alleviated resourceconstraints and in some cases pushed parents intocompensating for the transfer of their labour to theprogramme by increasing demands on young girls.An independent evaluation has concluded that theprogramme ‘could improve child schooling and reducechild labour provided that the transfers are largeenough’ (Hoddinott et al., 2009, p. 21).

Sources: Devereux et al. (2006); Hoddinott (2008); Hoddinott et al. (2009); Sharp et al. (2006); Slater et al. (2006);Woldehanna (2009).

Box 3.23: Ethiopia — Productive Safety Net Programme boosts children’s education

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

2 0 9

rations with on-site meals have the strongesteffect on enrolment, though there are large gapsin the evidence available (Bundy et al., 2009a).

Contrasting evidence from programmes inBurkina Faso underlines the importance of policydesign. In 2005/2006, the World Food Programmeassumed responsibility for all school feeding in thecountry’s Sahel region. In some schools it providedlunches to all pupils every school day; in others,girls with 90% attendance received monthly take-home rations of 10 kg of flour. The two models produced different results. While bothimproved enrolment, take-home rations extendedpositive nutritional benefits to younger siblings.An evaluation carried out after one year of theprogramme also found that both approachesincreased new enrolment among girls by fiveto six percentage points, but school lunches didnot appear to significantly affect boys’ enrolment.Absenteeism declined on average, but increasedamong girls in households facing severe labourconstraints. The reason: siblings took over the off-farm labour of girls eligible for school feeding, whoin turn took on more domestic labour. This resultedin higher enrolment but periodic absenteeism asgirls were occasionally pulled out of school forchores in the home (Kazianga et al., 2009).

Incorporating school feeding into wider anti-poverty programmes is also important. In Brazil,a school feeding programme covering 37 millionchildren has been a central part of the ZeroHunger strategy. It appears to have deliveredstrong results, in part because governmentagencies work through decentralized procurementstructures that are well resourced and regulated(Bundy et al., 2009a). The Mid-Day Meal Schemein India, which procures food centrally anddistributes it through a network of stores, hasachieved wide coverage. But while there is someevidence of nutritional benefits during droughtsand improved cognitive skills, the impact onenrolment is less clear cut. Moreover,implementation has been uneven, with widevariations in quality of food provided (Bundyet al., 2009a; Singh, 2008).

School feeding programmes have potential to playa greater role in combating marginalization, butproblems and limitations have to be recognized.By definition, such programmes do not reach out-of-school children. By targeting schools ratherthan individuals, they risk providing large transfersto children from high-income homes. In countries

lacking cost-effective procurement systems,this can result in a significant diversion ofresources away from those in greatest need.More fundamentally, some critics suggestschool feeding misses the target, since theprimary window of opportunity for addressingmalnutrition is during pregnancy and up to age 3 (World Bank, 2006f).

Targeting of beneficiaries. Social protectionconfronts policy-makers with difficult policychoices. Should social transfers be directed toindividual households or to districts and regionswith high levels of deprivation? Should theyhave narrow objectives, such as getting childreninto school, or target specific groups, such aschildren affected by HIV and AIDS, or havebroader objectives and target groups?

There are no simple answers. Much dependson governments’ capabilities and the scale anddepth of deprivation. In Mexico, Oportunidadeshas targeted districts and villages with poorhuman development indicators, as well asindividual households. Results include stronggains in education and decreases in child labour forindigenous children in southern Mexico (Lunde et al.,2009). For countries lacking the information orcapacity needed to implement finely tuned targetingstrategies, self-selection is an option. Ethiopia’sProductive Safety Net Programme targetsvulnerable regions on the basis of rural poverty anddrought indicators but participants choose whetheror not to work for the income on offer throughemployment programmes (Sharp et al., 2006).One potential problem with narrow targeting, inthe view of some commentators, is that it can leadto stigmatization. For example, there are concernsthat this could happen to people receiving transferslinked to HIV or AIDS status. The Kenyan socialtransfer programme for orphans and vulnerablechildren has attempted to address this problemby using wider eligibility indicators linked to poverty,orphanhood and other factors (Lunde et al., 2009).

Child labour is often neglected in poverty reductionstrategies (World Bank, 2005a). In a survey of forty-four recent national education plans, only eightidentified child labourers as a marginalized groupand of these just four mentioned specific strategiesto reach them (UNESCO-IIEP, 2009). Mali’s actionplan for accelerating progress towards universalprimary education mentions child labourers as avulnerable group, but contains no specific policies(Understanding Children’s Work, 2009).

Oportunidadesshows strong gainsin education and decreases in child labour for indigenouschildren insouthern Mexico

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

2 1 0

Social protection provides a mechanism forintegrating child labour into wider national povertyreduction efforts. Evidence from Latin America andbeyond highlights the potential. Reductions in childwork by beneficiaries of conditional cash transfershave been found in Brazil, Cambodia, Ecuador,Mexico and Nicaragua. In Cambodia, the averagechild receiving a transfer was ten percentagepoints less likely to work for pay. Reduction of childlabour as a result of these programmes is oftena by-product of school attendance conditions,or, as in Cambodia, a result of direct transfersfor education (Fiszbein et al., 2009).

Programmes could go further to target householdswhose poverty forces them to rely on child labour –but transfers need to be big enough to compensatefor the lost income. Targeting the Ultra Poor, aprogramme launched in 2002 by the Bangladeshinon-government organization BRAC, includeschild labour as one indicator of eligibility. Inthe programme, carefully targeted ‘ultra poor’households in rural Bangladesh receiveunconditional cash and asset transfers, credit,training and equipment. Income poverty has fallen,nutrition and health have improved, andbeneficiaries have increased their access toproductive assets. However, the effects on childlabour and enrolment have been more muted. Asone response, BRAC now includes school enrolmentas a monitoring benchmark for graduation fromultra poverty (Sulaiman, 2009). Conditions mayalso be needed to ensure that children are not keptout of school to take care of livestock assets thatthe household has been given. At the same time,benefits from the programme need to be sufficientto compensate for lost income from child labour.

Budgeting against marginalizationGovernment budgets are a major policy toolfor combating marginalization in education.Reaching the most marginalized often requireshigher spending than for wealthier areas, with aredistribution of public finance helping overcomeinherited disadvantage. Yet the marginalizedoften live in regions with little capacity to mobilizefinance. Without redistributive fiscal transfers,whole regions and historically disadvantagedgroups can be left behind.

Financial decentralization has often widenedopportunity gaps. Devolving responsibility forrevenue-raising can bring decision-making onfinancing closer to the communities affected,but it can also widen financing gaps between richer

and poorer regions, and between schoolswithin regions (UNESCO, 2008a). In China’shighly decentralized financing system, per studentexpenditure on junior middle schools is eighteentimes higher in Beijing and Shanghai than inthe poorest provinces (Dollar and Hofman, 2006).

Governments can seek to direct public spendingtowards marginalized regions and groups throughvarious mechanisms.

Mobilizing resources. Ensuring that excludedgroups get a stake in new sources of national wealthis one way to combat marginalization. In practice,this is often a politically fraught exercise becauseredistribution between subnational bodies involvescomplex bargaining by central government. TheBolivian Government has introduced several newfiscal transfer mechanisms financed by a DirectHydrocarbon Tax. Two of these are directlyredistributive. The tax finances a cash transferof around US$50 million to the Juancito Pintoprogramme. Covering close to 2 million children,it targets districts with high dropout and lowattendance. Another social transfer programmeprovides minimum income support. Together thetwo programmes represent around 2% of GDP.By far the largest part of the Direct HydrocarbonTax revenue takes the form of a block grant tosubnational governments. This transfer, estimatedin 2009 at US$902 million, or 9% of GDP, is not pro-poor and tends to favour gas-producingdepartments with relatively low poverty. Thus,the Direct Hydrocarbon Tax has increased overallfinancing for marginalized children in education,but has done little to narrow financing inequalities.Scaling up the Juancito Pinto programme wouldstrengthen equity by making the tax systemmore progressive (Gray Molina and Yañez, 2009).Other countries with significant mineral wealth,such as Angola, Nigeria and Peru, could alsosystematically target transfers to regions ofhigh deprivation in education.

Prioritizing equity. Many countries have adoptedrules for the transfer of public finance that attachweight to poverty-related factors, including deficitsin education (UNESCO, 2008a). One recent examplecomes from India. Before 2007, equity played onlya limited role in determining resource allocation.District population size was the main criterion usedin estimating need. A new formula attaches moreweight to social indicators, including a district-levelEducation Development Index. In 2005/2006, thedifferences in per child allocation between high

Withoutredistributive

fiscal transfers,whole regions

and historicallydisadvantaged

groups can be left behind

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

2 1 1

and low Education Development Index districtswere negligible, but in 2008/2009, districts in thelowest quartile on the index received twice asmuch per child as those in the highest quartile(Jhingran and Sankar, 2009; Figure 3.30). Brazilprovides another illustration of equity-basedfinancing, with the education budget weightedto provide additional support to the poorest statesand districts (Box 3.24).

Targeting regional development. Educationfinancing can be integrated into financingstrategies for regions with high levels of poverty,large ethnic minority populations and geographicdisadvantages. The effectiveness of suchprogrammes in narrowing regional disparitiesdepends on the level of redistribution andthe overall effect on public spending.

While almost all governments have someredistributive financing mechanisms in place,their effectiveness varies. The United Republicof Tanzania has adopted a needs-based financingformula for education, but it appears to havedone little to narrow financing gaps between localgovernment authorities. In fact, recent evidencesuggests the gaps may be widening, with damagingconsequences for equity in education. For each childaged 7 to 13, the richest thirty local governmentauthorities are allocated twice as much as thepoorest thirty. The pupil/teacher ratio is nearly 70:1in the poorest 20% of authorities and 44:1 in therichest. Such outcomes suggest that underlyinginequalities heavily outweigh redistribution. Thereis a strong relationship between spending per childin each authority and the pass rate at Standard 7(United Republic of Tanzania Government, 2008;World Bank, 2006i).

Budget systems vary in their level of commitmentto poverty reduction and the targeting ofmarginalized areas. Within Kenya’s unitary budgetsystem, a broad range of mechanisms is used tosupport decentralized spending. The ConstituencyDevelopment Fund allocates 3.5% of governmentrevenue for national poverty reduction efforts butattaches surprisingly little weight (around 25% inthe current formula) to poverty levels, as distinctfrom the overall population in the district. Thenational budget also identifies ‘core povertyprogrammes’ representing around 7% of totalplanned expenditure. They have played a key rolein financing free primary education but havesuffered from low levels of disbursement, limitedtransparency and the inclusion of programmes

with weak links to poverty reduction (WorldBank, 2009f). One result is that areas andgroups identified in this Report as centres ofmarginalization in education – notably the aridand semi-arid north-eastern areas inhabitedmainly by pastoralists – receive insufficientsupport (World Bank, 2009f). The Kenyan budgetframework thus suffers from both a weakcommitment to redistribution and poor delivery.

Countries with highly devolved financial systemsand deep geographical inequalities face distinctiveproblems. Poor states and regions have the leastcapacity to raise the revenue they need to delivergood-quality education. Yet they may be home tolarge populations facing restricted opportunitiesfor education. Overcoming marginalization is likelyto require higher levels of per capita spendingon the most disadvantaged, while the publicfinancing system is pulling in the other direction.The result is a vicious circle, with poverty andlow average income limiting access to education,and deprivation in education reinforcing povertyand regional inequalities.

Breaking the circle requires a strong commitmentto redistribution through public finance. Thatcommitment has often been lacking, as witnessed

Overcomingmarginalization is likely to requirehigher levels of per capita spendingon the mostdisadvantaged

400

800

1 200

1 600

Rupe

es p

er ch

ild (c

urre

nt p

rices

)

Districts with worst education indicators

Districts with besteducation indicators

2005/2006 2006/2007 2007/2008 2008/2009

Figure 3.30: Redistribution of public finance benefits the lowest performingdistricts in IndiaPer child allocations to worst and best performance quartiles on Education DevelopmentIndex, India

Notes: The allocations shown are those provided under the Sarva Shiksha Abhiyan programme. The district-level Education Development Index takes into account access (primary school coverage, ratio of upper primary to primary schools); infrastructure (availability of classrooms, toilets and drinking water);pupil/teacher ratio; enrolment of 6- to 14-year-olds; primary and upper primary school completion rates; and equity (girls’ enrolment, female literacy).Source: Jhingran and Sankar (2009).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 3

2 1 2

by the deep and persistent regional inequalitiesdocumented at the beginning of this chapter. Thereare exceptions; in recent years, for instance, Brazilhas used transfers from the national budget in aneffort to redress financing inequalities in education.It has succeeded in narrowing the gap, thoughlarge financing disparities remain (Box 3.24).

Conclusion

Most governments claim to have in place a policyframework for combating marginalization ineducation. Pledges to expand opportunities foreducation, improve school quality and enhancelearning standards for all are a staple part ofelection campaigns across the world.Unfortunately, the practical policies associatedwith such pledges are often fragmented andinsufficiently coordinated, and they fail to tackle

head-on some of the most powerful forces behindmarginalization. But accelerated progress towardsgreater equity is possible.

The building blocks for a concerted drive tocombat marginalization are well known. Sincethe World Education Forum in Dakar in 2000, manydeveloping countries have removed school fees.Primary school enrolment has often increasedsteeply as a result. Benefits for the marginalizedhave been most positively pronounced when thewithdrawal of fees has been combined withincentives for school attendance by disadvantagedgroups – such as young girls and street children –and social protection measures that reducevulnerability. Some countries have also addressedthe problems marginalized learners face in theclassroom, deploying qualified teachers tounderserved areas, providing additional resources

Brazil’seducation budget

is weighted to provideadditional

support to thepoorest states

and districts

In Brazil, greater equity in national budgeting has been a central pillar of wider national strategies aimedat breaking the links between poverty, inequality andmarginalization in education.

Bolsa Família, one of the developing world’s largestsocial protection programmes, transfers 1% to 2% ofBrazil’s gross national income to 11 million of its pooresthouseholds. The average transfer is around US$35.Most of this is spent on health, education and clothing.The programme has helped improve basic educationsignificantly.

Education budget reforms have attempted to addressdisparities associated with large inter-state wealthinequalities. Under Brazil’s devolved public financingsystem, the bulk of the revenue directed towardseducation finance comes from eight taxes. The federalgovernment uses a national formula to determine theshare of each tax going to education. Because taxrevenue is highly sensitive to wealth, it mirrors inter-state economic inequalities. State and municipalrevenues are complemented by transfers from thenational budget.

The federal government uses two levers to influencepublic spending outcomes. The first involves settingregulatory standards to establish national norms for per capita financing. The norms set a minimumthreshold for spending at each of twenty-one levelsof education, from pre-school to elementary school,secondary school and adult literacy. The norms areweighted for equity. The weighting favours rural overurban schools. It also provides indigenous people and

quilombolas, a highly marginalized group of blackBrazilians, with a level of support 20% above thebenchmark.

The second redistributive lever is transfer from thecentral government. States whose tax revenue leavesthem below the stipulated threshold are eligible forcomplementary federal financing. In 2008, nine stateswere in this position. These states, located in the poorernorth and north-east, are characterized by low averageincomes, high levels of poverty and some of the worsteducation indicators in Brazil.

Box 3.24: Redistributive public financing in Brazil

398 27 33377 21 33314 26 36348 21 42

618 13 48

744 8 52826 4 67

ParáBahiaMaranhãoCeará

National average

Rio Grande do SulSão Paulo

Table 3.6: Low and high performing Brazilian states on education and poverty indicators, 2007

1. Share of children aged 10 to 14 who are more than two years behind the gradethey should be in.2. Fundeb is the federal complement to state revenue.Source: Henriques (2009), based on the Brazilian National Household Sample Survey.

Familyincome percapita (R$)

Lateenrolment1

(%)

Secondarynet enrolment

(%)

Selected Fundeb recipient states 2

Best performing states

R E A C H I N G T H E M A R G I N A L I Z E D

L e v e l l i n g t h e p l a y i n g f i e l d

2 1 3

to ‘failing’ schools, and implementing interculturaland bilingual education programmes. Manygovernments have also recognized the needto prioritize disadvantaged areas in schoolconstruction. While public spending patternscontinue to favour wealthier groups and regionsin most countries, several countries haveacknowledged that levelling the playing field ineducation requires a commitment to redistributivefinancing in favour of the marginalized.

Non-government organizations have alsodemonstrated that progress is possible. They havebeen instrumental in developing and implementinginnovative strategies that reach some of themost marginalized, including street children andpastoralists. These strategies are increasingly beingintegrated into government systems. One examplehas been the development of second chance

programmes allowing children and youth deniedthe chance to develop literacy and numeracy skillsduring their primary school years the opportunityto develop skills for employment, gain qualificationsand re-enter the formal education system.

The evidence presented in this chapterdemonstrates that ‘reaching the marginalized’does not have to be an empty rhetorical pledge.There are strategies that work – but they have tocut across the borders of traditional policy-making.More important, they have to be integrated intoa coherent policy framework that simultaneouslytackles the multiple underlying causes ofmarginalization. Setting equity-based targetscan help to focus policy and ensure that themarginalized figure more prominently in nationalplanning frameworks and poverty reductionstrategies.

‘Reaching themarginalized’ does not have to be an emptyrhetorical pledge

Targeted regional support has significantly raisededucation spending in some of the poorest states.Federal transfers have increased per capita spending inCeará by 21%, rising to 55% in Maranhão. Very largefinancing gaps remain, however. Per capita spending inbetter-off states such as Espírito Santo, Acre and RioGrande do Sul, and in the city of São Paulo, greatlyexceeded spending in the eight states receivingcomplementary support in 2008 (Table 3.6; Figure 3.31).The upshot is that the states lagging furthest behind ineducation have the most limited resources for catchingup with better-performing states.

The problems do not end with inter-state disparities.Some states, including Rio Grande do Sul and MatoGrosso do Sul, may have high average income and percapita education spending but also very large pocketsof education marginalization among children of landlessagricultural labourers and small farmers. Similarly,children living in the slums of São Paulo and Rio deJaneiro have some of the most restricted opportunitiesfor education in Brazil. Current approaches to publicfinance do not systematically address these problems.

The experience of Brazil has wider internationalrelevance. Achieving equity is hampered by the sheerscale of inequality, highlighting the limits to the scopeof redistribution through the budget and pointing to aneed for structural reforms in other areas.

Source: Henriques (2009).

Pará

Bahi

a

Piau

í

Mar

anhã

o

Alag

oas

Cear

á

Pern

ambu

co

Para

íba

Amaz

onas

Natio

nal a

vera

ge

Sant

a Ca

tarin

a

Mat

o Gr

osso

do S

ul

Rio

Gran

de d

o Su

l

Espí

rito

Sant

o

Acre

São

Paul

o0

500

1 000

1 500

2 000

2 500

Real

per

pup

il

Government transfers

State spending

Figure 3.31: Federal government redistribution leaves large gaps in Brazil State spending per pupil, including the Fundeb transfer from central government, Brazil, 2008

Source: Henriques (2009), based on data from Fundeb. See http://www.fnde.gov.br/home/index.jsp?arquivo=fundeb.html.

01

02

Education f

or

All

Glo

bal M

onitori

ng R

eport

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

Outdoor learning in Uganda:adapting educationto community needs

© C

risp

in H

ughe

s/PA

NOS

2 1 4

An inclusive schoolin Nicaragua makesroom for everyone

Chapter 4

© H

andi

cap

Inte

rnat

iona

l

0102Education for All Global Monitoring Report

Conflict and displacement poseparticular challenges to education,

Afghanistan

2 1 5

© R

EUTE

RS/

Ahm

ad M

asoo

d

Window of opportunity: slum childrenin and out of school, Bangladesh

© U

NES

CO/S

amer

Al-

Sam

arra

i

The aid compact:falling short of commitments

Education budgets in developingcountries are under increasingpressure. Concessional aid couldhelp alleviate this pressure. Thischapter looks at the donor recordon aid delivery and finds a collectivefailure to act on the pledges madeat Dakar. It also assesses ongoingefforts to strengthen aideffectiveness and meet the needsof countries affected by conflict.Finally, it critically reviews themultilateral architecture for aid toeducation, concluding that far morecould be done to scale up financing,give developing countries a greatervoice in governance and engagewith the private sector. Reform ofthe Fast Track Initiative is identifiedas a priority – and lessons aredrawn from global initiatives in public health.

Introduction .................................................................. 217

Aid for education ................................................... 218

Reforming theFast Track Initiative .......................................... 248

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 1 6

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

I n t r o d u c t i o n

2 1 7

Introduction

International aid is a vital part of the Education for All compact. When rich countries signed theFramework for Action in Dakar in 2000, theypledged that no country committed to achievingthe EFA goals would be allowed to fail for wantof finance. The global economic downturn hasreinforced the importance of that commitment.Weaker economic growth and mounting pressureon government budgets threaten not only to slowprogress in education, but also to reverse the hard-won gains of the past decade. Counteringthat threat will require not just increasing aidflows but also improving the quality of aid.

International aid debates often focus on technicalissues surrounding aid delivery. The human faceof what is at stake is sometimes overlooked.Development assistance can help bring learningopportunities to the marginalized children andyoung people discussed in previous chapters. Thelimits to aid for education have to be recognized,but so does the potential for it to help remove thebarriers to school created by poverty, gender andother sources of marginalization.

This chapter examines some of the most pressingconcerns surrounding aid. The first section looksat how overall trends in development assistancecompare with collective commitments made bydonors in 2005. While overall aid is rising, severalmajor donors are falling far short of their pledges.In effect, the underperformers are ‘free-riding’on the efforts of others. The record on aid foreducation is disappointing and inconsistent withthe Dakar promise. Large financing gaps remainand commitments to basic education arestagnating. The narrow donor base for supportto basic education and the skewing of aid towardspost-secondary education contribute to theproblem. Poor countries affected by conflict, whichaccount for a large share of the world’s out-of-school children, are not receiving enough attention,with the result that opportunities to rebuildeducation systems and societies are being lost.

Improving aid quality is as important as increasingquantity. Development assistance is a scarceresource, and it is vital for donors and aid recipientsto work together to maximize the benefits itgenerates. As the global economic downturn hasraised pressure on donor budgets, the need tomake aid more efficient has taken on greatersignificance. The case for scaling up developmentassistance ultimately rests on demonstrating thatmore aid can improve access to schools as well asequity and quality in education. Donors and partnercountries have agreed to a wide-ranging agendafor strengthening aid effectiveness. The third High-Level Forum on Aid Effectiveness, held inAccra, Ghana, in 2008, gave renewed momentumto that agenda. The record on implementationhas been mixed, however, and many donors needto speed up reform.

One of the most important financing initiatives toemerge from the Dakar forum was the EFA FastTrack Initiative (FTI). The second section of thischapter looks at the record of that initiative. Theaim of the FTI is to galvanize political and financialsupport for accelerated progress towards universalprimary education and wider goals. Its coreprinciples are as valid today as they were in 2002,when the FTI was created. Unfortunately, very littlehas been achieved. The initiative has not facilitatedmobilization of new financing and its own limitedfinancial contribution has entailed high transactioncosts. Protracted delays between aid allocationdecisions and disbursement have underminededucation planning in many developing countries.Conflict-affected countries have also faceddifficulties in receiving support from the FTI.Nevertheless, the world needs a multilateral aidmechanism for education. As the FTI is notcurrently functional for this purpose, fundamentalreforms are needed to fix the financing andgovernance problems that undermine its abilityto deliver aid.

As the globaleconomicdownturn hasraised pressure on donor budgets,it is vital fordonors and aidrecipients to worktogether tomaximize thebenefits

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 1 8

Aid for education

Since the Dakar World Education Forum in 2000,the global aid environment has undergone aprofound shift. After a steep decline in the 1990s,development assistance budgets have been rising.An important catalyst for change was theMillennium Development Goals (MDGs). Donorsand developing country governments see increasedaid as vital support for policies aimed at reducingpoverty, getting children into school and achievingthe wider goals set out in the MDGs.

Donor commitments – and efforts by campaignersto hold donors to those pledges – reflect a positiveview of international aid. Some commentatorsargue, however, that aid undermines economicgrowth, distorts national priorities, fuels corruptionand delivers little for the poor (Easterly, 2003).As one prominent critic puts it: ‘Aid has been, andcontinues to be, an unmitigated political, economicand humanitarian disaster for most parts of thedeveloping world’ (Moyo, 2009, p. xix). Controversiesabout aid effectiveness go back several decadesbut recently have taken on a new lease of life,with some commentators calling for developmentassistance to be curtailed or even eliminated.Yet the evidence does not support this intensepessimism on aid effectiveness.

Consider first the argument that more aid meansless economic growth. If it were true, this wouldclearly be bad not just for poverty reduction, butalso for the financing of basic services such ashealth and education. But there is no robustevidence to support the claim that aid weakensgrowth prospects. From 2000 to 2008, as aid to sub-Saharan Africa almost doubled, economicgrowth averaged 5% to 6% a year – double theaverage of the 1990s. Meanwhile, the incidence of poverty fell from 58% to 51%, with absolute numbers below the poverty line declining for the first timein a generation (Chen and Ravallion, 2008).

Cross-country analysis looking further backsuggests aid has a broadly positive impact ongrowth, though high levels of aid dependenceover long periods can have adverse consequences(Clemens et al., 2004). Part of the problem withthe argument of aid pessimists is that it fails todifferentiate between types of aid. No one wouldexpect aid to basic education or child health todeliver early results for economic productivity.

But aid to productive infrastructure has supportedgrowth. One study finds that each US$1 in aid yieldsUS$1.64 in increased income in the recipientcountry (Radelet et al., 2005).

The association between aid and governanceis even more complex. Aid pessimists claim thatan assured and abundant supply of developmentassistance can reduce the incentives forgovernments to raise domestic revenue, creatinga cycle of dependence and weakening accountabilityto citizens. Another claim is that large inflows ofaid can help fuel corruption, especially in countrieswith weak public financial management systems(Brautigam, 2000). Yet, while there is no shortageof corruption among many governments receivingaid, cross-country studies have generally failed toestablish significant, clear or consistent causallinks between aid dependence and standards ofgovernance (Coviello and Islam, 2006; Moss et al.,2006). Moreover, aid has played an important rolein supporting the development of more accountableinstitutions in countries including Mozambique,Nepal and the United Republic of Tanzania.

Nevertheless, aid pessimists raise some importantissues. Economic growth in many aid-dependentcountries has been disappointing. That does notmean aid is the underlying reason, but there arestrong grounds for concluding that aid could haveachieved far more. Similarly, aid optimists tendto turn a blind eye to corruption. Too much aidthat could have been used to build classrooms,train teachers or stock health clinics has beenwasted or stolen – sometimes with the collusionof major donors – or otherwise ill-used becauseof poor governance (Wrong, 2008). There is nodoubt that aid is likely to work better in countriesthat are serious about tackling corruption andstrengthening governance.

Developments in education underscore thepotential for aid to make a difference. To citesome achievements in countries where aidfinancing is important:

Since the overthrow of the Taliban in 2001,Afghanistan has received sizable amountsof aid to restore its education system. Withsupport from many non-governmentorganizations, donors and United Nationsagencies, the government has respondedto the high demand for education from theAfghan people. Fewer than a million children,most of them boys, were enrolled in primary

The evidencedoes not

support intensepessimism on aid

effectiveness

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 1 9

education at the beginning of the decade. In 2007,over 4.7 million children went to school, morethan one-third of them girls.

Cambodia’s Education Sector Support Project,funded by several donors, provides scholarshipsthat help children from poor families make thetransition from primary to secondary school.The scholarships have had a marked effect:schools benefiting from the programme havesecondary enrolment rates 21% higher than non-participating schools (Fiszbein et al., 2009).

Over the past decade, Mali has embarked onan ambitious programme to accelerateprogress towards universal primary education.Twenty-two donors provide financial andtechnical assistance. External aid accounted fornearly three-quarters of the programme cost in2007 – excluding teacher salaries. The primarynet enrolment ratio increased from 46% at theend of the 1990s to 63% in 2007. While markedgender disparities remain, the ratio of girls toboys in primary school rose from 70% to 80%.A decade ago, children entered primary schoolswith very few books, but in 2008 every first gradepupil had two books (Ky, 2009).

In Mozambique, donors have pooled theirsupport for the national education strategy.Aid has played a key role in financing schoolconstruction in rural areas, recruiting andtraining teachers, and providing textbooks.From 1999 to 2007, the net enrolment ratio inprimary education increased from 52% to 76%.The number of children out of school fell byhalf a million.

When the Dakar forum was held in 2000,about 3 million children in the United Republicof Tanzania were out of school. The figure isnow less than 150,000. The country’s educationstrategy has combined measures aimed atimproving access, including the removal of usercharges, with increased investment in classroomconstruction, teacher training and textbooks.

These examples do not represent aid successstories in a narrow sense. They are the result ofnational policies and national political leadershipsupported by development assistance. No amountof aid can counteract poor policies and politicalindifference. But when increased aid is harnessedto strong policies, it is possible to rapidly expandopportunities for basic education. As the case of

Afghanistan shows, development assistance canalso help rebuild education systems in countriesaffected by conflict.

This section is divided into four parts. The level ofaid to education is a function of two things: overallflows of official development assistance (ODA) andthe share of those flows directed into education.Part 1 looks at the first part of that equation,assessing the record of donors in the light ofpledges to increase aid by 2010. Part 2 examinesthe level of aid to education, with a focus on basiceducation. Part 3 looks at progress towards moreeffective aid, focusing on aid predictability anddonor use of country reporting systems. Part 4considers the position of countries affected byconflict. The following are among the keymessages:

Development assistance works. Aid pessimistsargue that development assistance is failing theworld’s poor. The evidence on education doesnot support that claim. While much can bedone to strengthen aid’s effectiveness, it isdelivering results.

Overall aid levels are rising – but there is areal danger that donors will fall short of theirpledges. Taking into account current spendinglevels and forward spending plans, projected aidin 2010 may be US$20 billion less than targetlevels. Budget pressures and political decisionsin donor countries may exacerbate the gap.Delivering on commitments made to developingcountries in 2005 will require an emergencyresponse on the part of the donor community.

Free-riding has emerged as a serious problem.Donors have adopted bold collective targets,but national targets reveal highly variable levelsof ambition, and some countries – includingG8 members – are undermining collectivecommitments by failing to meet their fair shareof the burden.

There are large financing gaps for basiceducation and aid commitments are stagnating.With the 2015 deadline for achieving the Dakartargets approaching and many countries offtrack, it is urgent for donors to close the basiceducation financing gap. The stagnation incommitments for basic education remains aconcern, with several major donors orientingaid towards higher levels of education.

No amount of aidcan counteractpoor policies and politicalindifference

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 2 0

Governance problems continue to undermine aideffectiveness. Aid works best when its provisionis predictable and when it operates throughviable national reporting and public financialmanagement systems. Under the ParisDeclaration on Aid Effectiveness, donors andrecipients adopted ambitious targets in theseareas. However, progress has been limited.Unpredictable aid and failure to use nationalsystems weaken the ability of developing countrygovernments to undertake long-term financialplanning and add to transaction costs.

Conflict-affected poor countries receiveinsufficient support. Countries enduring oremerging from conflict often have large out-of-school populations, severely damaged educationinfrastructure, weak governance, and limitedfinancial, technical and human capacity. Whilethere are problems in building aid partnershipsin these countries, far more could be done.Opportunities to consolidate peace through thereconstruction of education systems are beinglost. Over one-third of out-of-school childrenare in conflict-affected poor countries. Yetdonors commit less than one-fifth of aid toeducation to these countries.

Overall aid pledges: the record on delivery

International support for education depends on thesize of the global aid envelope and the allocation ofresources within that envelope. Here we examineoverall aid levels and donors’ progress towardsthe benchmarks set by their own pledges.

Aid flows rose sharply in 2008 after two yearsof decline, but there is a real danger thatcommitments made in 2005 to increase overallaid by US$50 billion by 2010 – and to double aidto Africa – will not be honoured. Even beforethe global economic downturn, spending plansindicated that these targets would be missed bya wide margin. As budgets come under mountingpressure, the deficit could widen, with graveconsequences for international developmentgoals in education and beyond.

Aid levels are rising — but too slowly‘Despite the severe impact of the crisis on oureconomies, we reiterate the importance of fulfillingour commitments to increase aid,’ the leaders ofthe Group of Eight industrialized countries statedat their July 2009 summit in L’Aquila, Italy (Group

of Eight, 2009c, p. 35). Their joint communiquémarked the fourth such reaffirmation of a pledgemade at the Gleneagles summit and other high-level meetings in 2005. Commitments under thatpledge include an increase in overall aid fromthe US$80 billion spent in 2004 to US$130 billionby 2010, with around half the increase, orUS$25 billion, directed towards Africa.1

Measuring progress towards these benchmarksis complicated by several factors. High levels ofdebt relief in 2005 led to a sharp spike in reportedaid, followed by a comparative decline in 2006and 2007. Another difficulty relates to the way aidis measured. The OECD Development AssistanceCommittee (OECD-DAC) converted donors’initial pledges to targets that expressed aid asa proportion of donor countries’ gross nationalincome (GNI). With economic growth projectionshaving fallen, the same aid-to-GNI ratios translateinto less real aid. The question is whether theGleneagles pledge should be adjusted to reflectthe new growth projections.

Leaving the aid-to-GNI targets unchanged wouldcontradict the spirit of donors’ commitments.For aid recipients, what counts is real financingfor schools, teachers, clinics and roads, not thebookkeeping arrangements of the OECD-DAC.This Report, therefore, uses the original pledgeof increasing aid by US$50 billion by 2010 as thebenchmark for measuring progress.

Overall development assistance rose sharply in2008 as debt relief reverted to more normal levels.Spending on aid increased by around US$10 billionto US$101 billion in 2008 – a rise of more than 10%from the previous year (Figure 4.1). The share of aid in the GNI of rich countries also increased, to 0.30%.2

The positive news on the recovery in aid flows iscounterbalanced by the prospect of large shortfallsagainst the targets set. Two years before the 2010deadline, donors still have to mobilize an additionalUS$29 billion. In other words, they are less thanhalfway to meeting their pledges. Their currentlyplanned increases fall far short of the level requiredto close the impending 2010 deficit. As Figure 4.1indicates, the estimated increases leave a globalgap between target spending and actual spendingof around US$20 billion.

Africa accounts for a large share of the 2010financing gap. Donors are a long way from theaid spending targets they set for the region at

1. As the 2010 targetof increasing aid byUS$50 billion is expressedin constant 2004 prices,so are the figuresthroughout this part.

2. The 2008 aid data werestill preliminary at thetime of writing.

Donors still haveto mobilize

an additionalUS$29 billion — in other words,

they are lessthan halfway

to meeting their pledges

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 2 1

Gleneagles. The region has accounted for lessthan one-third of the global increase in aid from2004 to 2008, whereas the 2010 target share is 50%.The latest OECD-DAC survey of spending plansprovides an overview of the regional distributionof future aid flows. Preliminary findings show anincrease of only US$2 billion in programmed aidto Africa between 2008 and 2010. This representsone-third of the planned global increase in countryprogrammed aid and a marked slowdown in therate of increase in planned aid spending for Africa.3To achieve the 2010 target, donors need toincrease aid spending for Africa by US$18 billion(OECD-DAC, 2009a and 2009d).

Uncertainty about whether donors will meet theircommitments for 2010 is holding back educationplanning in some of the world’s poorest countries.Promises made at summit meetings cannot buildschools, pay teachers, buy textbooks or financeincentives for marginalized groups. These activitiesrequire real funds. Budget planners need to beconfident that donors will deliver on theircommitments – and donors’ collective performanceto date does not breed confidence. As the OECDputs it, ‘only a special crisis-related effort canensure that the 2010 targets for aid are met’(OECD-DAC, 2009b, p. 2). Failure to make that effortwill undermine education financing in recipientcountries and prospects for accelerated progresstowards the goals set at Dakar.

Donor performance variesGlobal monitoring provides an aggregate pictureof how well rich countries are meeting theircollective commitment to developing countries.But it conceals significant differences betweendonors, some of whom perform much morestrongly than others.

While almost all donors have signed on to collectivecommitments, there is little uniformity in how theytranslate these into national targets. EuropeanUnion members have a shared commitment toreach a collective aid-to-GNI target of 0.56% by2010 and 0.70% by 2015. Some members havealready met the first target and others have setthe bar even higher. Japan’s national commitmententails an increase of US$10 billion between 2005and 2009 – nearly double what would have beenrequired to increase aid in line with the Gleneaglesgoal.4 Canada’s commitment is directed to doublingaid by 2010, but only from the nominal level of aidprovided in 2001. The United States has committedto doubling aid to sub-Saharan Africa between 2004

and 2010 but has no global aid target. Furthercomplicating the picture, OECD-DAC membersstart from very different baselines in terms oftheir 2004 aid levels.

There are no simple mechanisms for comparingnational aid targets, yet it is vital to submit donors’comparative performance to critical scrutiny. Oneway to place commitments and performance on acommon scale is to look at the ratio of aid to grossnational income. How do individual donors shapeup against each other?

Figure 4.2 provides a partial answer to thatquestion. It captures the wide variation behind theincrease in the aggregated OECD-DAC aid-to-GNIratio, along with the divergence in starting points.By converting the 2010 targets of individual donorsinto a common unit, it also provides a snapshot oftheir different levels of ambition.

Five countries surpass the United Nations targetof 0.7% and Sweden invests almost 1% of GNI inaid. Three of the four least generous donors – Italy,Japan and the United States – are all G8 countries.

3. Country programmable aid for Africa rose by 7% a year from 2004 to 2007. The planned annual increasefor 2008-2010 is 3%.

4. To meet the Gleneaglestarget, total aid would needto increase by 62.5% from2004 to 2010. Japan’scommitment translates into a 112% increase for the same period.

Uncertainty aboutwhether donorswill meet theircommitments for 2010 is holdingback educationplanning in someof the world’spoorest countries

0

20

40

60

80

100

120

140

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Tota

l ODA

, net

disb

urse

men

ts in

cons

tant

200

4 US

$ bi

llion

s

Total ODA

2010 target for total ODA

Estimated increase of total ODA

ODA to Africa

2010 target for Africa

Estimated increase of ODA to Africa

Total ODA

Total ODA to Africa

2010 globaltarget:US$20 billionshortfall

2010 Africatarget:US$18 billionshortfall

Figure 4.1: Africa faces the greatest projected shortfall in total aidTotal ODA, net disbursements for 1999–2008 and simulations for 2009 and 2010

Notes: The simulations are based on the targets of raising total aid to US$130 billion and aid to Africa to US$55 billion by 2010 (at constant 2004 prices). The estimated increases of total aid and of aid to Africa are based on the OECD-DACSecretariat’s 2009 survey of donors’ spending plans. It is assumed that aid not classified as country programmable willincrease at the same rate as country programmable aid.Sources: OECD-DAC (2009a, 2009d).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 2 2

While some countries, such as Germany and Spain,have significantly increased aid-to-GNI from themiddle of the range, others have registered amarginal increase (Italy), no increase (the UnitedStates) or have fallen back (Japan) from a low level.

Some donors have set the bar far higher thanothers. That is one reason it is problematic tocompare donors on the basis of progress towardsnational targets. Starting from a high level, Norwayand Sweden aim to reach an aid-to-GNI level of 1%,while EU members have set a collective target of0.56%.5 The financial target for Canadian aid wouldtranslate into a 0.34% aid-to-GNI ratio and meetingthe Japanese target would produce a ratio of 0.28%.

Another way to measure donors’ comparativeperformance is to look at ‘fair shares’, allocatingeach donor responsibility for delivering on a shareof the global pledge based on the size of their GNI.Figure 4.3 illustrates this approach with reference

to the G8 for the aid target of a US$50 billionincrease by 2010. As G8 countries account for two-thirds of global development assistance,the group’s collective and individual performanceis clearly of great importance in terms of reachingaid targets.

The data show that G8 performance has beenvariable. Preliminary 2008 data for France, Japanand the United States indicate they have made littleprogress towards their fair share target whileGermany and the United Kingdom are progressingat a rate that would see them surpass that target.Figure 4.4 looks at the broader group of OECD-DACdonors, providing a league table in terms of fairshares. It highlights the strong performance ofsome countries, including the Netherlands, Spainand Sweden, which have exceeded their share ofthe overall commitment.

The large prospective 2010 deficit facing sub-Saharan Africa is a source of growing concern.Aid represents a large share of revenue for theregion and is a vital source of finance for education.The 2005 Gleneagles commitment was promptedin part by donor recognition that the region was faroff track for many of the Millennium DevelopmentGoals and that a stronger aid effort could helpchange this picture. In the event, though, manydonors have failed to give higher priority to aidfor sub-Saharan Africa (ONE, 2009).

Four years after the 2005 pledges were made,confusion over targets and monitoring criteriacontinues to hamper effective scrutiny of donorperformance. At the 2009 G8 summit in L’Aquila,leaders agreed to explore a ‘whole of country’approach to development that takes into account‘a wide range of factors such as government aidand non-aid policies, private sector and civil societyefforts’ (Group of Eight, 2009c, p. 37). It is not clearwhat this means in practice. An obvious danger isthat the conspicuous failure of some donors todeliver on measurable aid pledges will be obscuredby a poorly defined reporting system designedto report on indicators that are not comparable – and in some cases not readily measurable.

The financial crisis threatensfuture aid flowsProspects for achieving the 2010 aid targets havediminished with the global economic downturn.Donor country governments are grappling withswelling fiscal deficits as they seek to balancea shrinking revenue base with rising expenditure

5. Individual EU countrytargets differ butrepresent 0.56% of EU GNI.

Some countries,such as Germany

and Spain, have significantly

increased aid-to-GNI

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2

Italy

Greece

United States

Japan

Austria

New Zealand

Spain

Australia

Canada

Germany

United Kingdom

Finland

Ireland

Belgium

Switzerland

France

Portugal

Netherlands

Sweden

Luxembourg

Denmark

Norway

Total DAC

DAC-EU countries

ODA as % of GNI

2004

2008 (increase)

2008 (decrease)

2010 target

Figure 4.2: Nearly all donors are falling short of their aid pledges for 2010OECD-DAC donors’ total aid as a share of GNI, 2004–2008 (net disbursements), and targets for 2010

Source: (OECD-DAC, 2009d).

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 2 3

for economic recovery and social protection.The impact of the financial crisis will inevitablydepend on the severity and duration of theeconomic slowdown. But there are growingconcerns that as aid budgets come under pressure,the pledges made to the poorest countries in 2005are even less likely to be honoured.

Experience points in a worrying direction.Aid contracted sharply and recovered slowlyafter the early 1990s’ financial crisis. Over thequarter century to 2004, aid tended to declineat times of rising public debt and deterioratingfiscal indicators in rich countries (Roodman, 2008;World Bank and IMF, 2009).

It is encouraging that political leaders havepublicly reaffirmed aid targets, notably at theDoha Conference on Financing for Developmentin late 2008, and the London summit of the G20and L’Aquila summit of the G8 in 2009. But actingon these reaffirmations in national budgetnegotiations will require strong political leadership,with governments and development advocatessetting out a compelling case in defence of aid.

The record to date has been mixed (Box 4.1).The United Kingdom has committed to maintainingthe real financial value of its aid budget, implyinga rising share for development assistance in GNI.Public spending reviews in Ireland and Swedenhave led to announcements of aid cuts for 2009and 2010, respectively, albeit in the context ofmedium-term financing plans that, if implemented,will restore aid levels. There are strong indicationsthat Italy’s aid budget may be cut, with no clearframework thus far for recovery and future growth.

Many donor countries have yet to set out clear post-crisis aid spending plans. This has addedto the uncertainty over prospects for achievingthe 2010 targets. The European Commission hasprepared one of the most detailed projectionsso far, using information from EU members.It indicates that overall EU aid spending in 2010will represent 0.50% of GNI, against the 0.56%target level (European Commission, 2009b).

The varying responses of donors to the economiccrisis are conditioned by many factors, includingfiscal pressures, the depth of the recession andprospects for recovery. It would be naïve to supposethat aid budgets can be entirely insulated fromwider economic developments, but politicalleadership can make a significant difference.

Consider the very different positions of Italy andSpain within the European Union. Not only is Italyone of the least generous EU donors, but it has alsomoved towards shared aid targets more slowly thanalmost any other member state. Meanwhile,Spanish aid has almost doubled as a share of GNI

0

2

4

6

8

10

12

14

16

18

CanadaItalyFranceUnitedKingdom

GermanyJapanUnitedStates

Cons

tant

200

4 US

$ bi

llion

s

Fair share of US$50 billion increase by 2010

Increase achieved from 2004 to 2008

Figure 4.3: Most G8 countries are falling short of their ‘fair share’ in aidG8 donor contributions needed to meet the 2010 aid targets according to the ‘fair share’ principle and increase in aid achieved from 2004 to 2008

Note: The fair share is based on each DAC donor country’s share of total DAC GNI. A donor’s fair share of the targeted US$50 billion increase in total net ODA from 2004 to 2010 is measured as its share of the total DAC GNI multiplied by the targeted amount. By convention, the United States share of total DAC GNI is capped at 33%.Source: OECD-DAC (2009d).

-250 -200 -150 -100 -50 0 50 100 150 200 250

Portugal

Switzerland

Denmark

New Zealand

Greece

Belgium

Finland

Australia

Norway

Austria

Netherlands

Spain

Luxembourg

Ireland

Sweden

Achievement of ‘fair share’ of 2010 aid target (%)

Increase from2004 to 2008

Decrease from2004 to 2008

Fair sharetarget achieved

2004

Figure 4.4: Most non-G8 donors also have a long way to go% of 2010 target DAC donors achieved by 2008 according to the ‘fair share’ principle

Note: The fair share is based on each DAC donor country’s share of total DAC GNI. A donor’s fair share of thetargeted US$50 billion increase in total net ODA from 2004 to 2010 is measured as its share of the total DAC GNImultiplied by the targeted amount. By convention, the United States share of total DAC GNI is capped at 33%.Source: OECD-DAC (2009d).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 2 4

since 2004 and Spain has overtaken Italy inabsolute financial terms. As Figure 4.5 shows,Spain is broadly on course to achieve its nationalgoal of a 0.56% aid-to-GNI level by 2010, while Italyis hopelessly off track in terms of the goals it hasendorsed at successive G8 summits, as well asthe 2005 EU goal. From a development financingperspective, this shortfall matters. Had Italy’s aidlevel moved towards the 2010 target at the samerate as Spain’s, its aid budget would have beenaround US$3.8 billion larger in 2008 (at constant2007 prices). In effect, Italy has been free-ridingon the strong performance of Spain and the othercountries that have pushed the European Uniontowards its aid targets.

The impact of the economic downturn on aid levelswill depend partly on how donors interpret theircommitments. With economic growth projections

declining, defining commitments in terms of aid-to-GNI ratios has potentially damagingimplications for overall aid: a fixed share of decliningnational income translates into less aid. This is not a technicality: adjusting the 0.56% EU aid-to-GNIcommitment to reflect lower projections couldresult in a loss of nearly US$9 billion from the pre-crisis growth forecast for 2010. If education’sshare of overall aid remains the same as in 2007,the adjustment could mean a loss of US$890 millionin aid to education from European donors in 2010.6

Currency movements linked to the financial crisiscould also affect development assistance flows.Appreciation of the US dollar against the currenciesof other major donors has deflated the value of aid.Preliminary analysis by the World Bank suggeststhat the losses could be in the order of US$3 billionto US$5 billion annually (World Bank and IMF, 2009).

6. These projections arebased on OECD-DACSecretariat simulations ofnet ODA disbursementsby EU members in 2010.The simulations show theODA levels resulting fromeach donor reaching its2010 aid-to-GNI target.The calculation requiresgrowth projections foreach donor country. Forthe pre-crisis calculation,the same growthprojections (fromJune 2008) were used asin OECD-DAC Secretariatsimulations (OECD-DAC,2008b). For the crisis-adjusted amounts, growthprojections from theOECD Economic Outlook(24 June 2009) were used.

Appreciation of the US dollar

against thecurrencies

of other majordonors has

deflated thevalue of aid

People in the world’s poorest countries played no part in creating the financial market implosion thatcaused the global economic downturn, but they stand to be among the biggest losers. International aid has a vital role to play in preventing the short-term globaldownturn from causing long-term damage to humandevelopment.

The impact of the economic crisis on poor countriesmakes it urgent for donors not only to deliver on pastaid commitments but also to respond to the additionalneeds arising from the crisis. The record to date hasbeen mixed:

The United Kingdom: maintaining the real value ofcommitments. Given the recession and a loweredforecast for GNI in 2010, the United Kingdom couldhave made deep cuts in its aid budget and stillremained on track for the 0.56% aid-to-GNI ratio.Instead, a public spending review in 2009 announcedplans to meet the real spending commitmentsprojected for 2009 and 2010. As a result, the countryis expected to reach an aid-to-GNI ratio higher thanthe 0.56% European target in 2010/2011.

Ireland: initial cuts but a commitment to recovery.The downturn has affected Ireland more than almost any other OECD country. With the nationalbudget absorbing the cost of a large-scale financialrescue package, deep budget cuts have beenprogrammed. Plans announced in 2009 will see the aid budget cut by 22%, reversing a rapidexpansion. However, medium-term budget plansreflect a continued commitment to achieving a 0.70% aid-to-GNI ratio by 2012.

The United States: announcing real increases. Thenew administration has announced an ambitious planto double national development assistance by 2015,from US$25 billion to US$50 billion, albeit from avery low aid-to-GNI level. The 2010 budget proposalshows a small decline in levels of aid compared with2009, but it still puts the United States on a path toachieve this goal, though the target date may changedepending on the speed of economic recovery.

Sweden: making cuts but maintaining a high level ofcommitment. The Swedish Government has signalleda major reduction in its 2010 aid budget. The decisionis linked to the country’s economic recession in 2009and to its practice of basing development assistancelevels on an aid-to-GNI formula. The cut will be about12%, but the government is committed to continuingto spend 1% of GNI on external aid.

Spain: continuing the upward trajectory. Despite theeconomic downturn that started in 2008, in early2009 the Spanish Government adopted a highlyambitious new policy framework setting out multiyearcommitments. Public spending targets are on courseto achieve the 0.70% aid-to-GNI ratio by 2012, threeyears ahead of the collective EU schedule.

Italy: deep cuts from a low base with littlepredictability and no recovery plan. It is difficult tosquare the actions of the Italian Government with the letter and spirit of the international communiquésit has signed since 2005. The European Commissioncalls Italy ‘the only Member State apparentlyabandoning its commitments’ (European Commission,2009b, p. 27). Grounds for that judgement include

Box 4.1: Aid and the financial crisis

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 2 5

ConclusionThis is a critical moment for leadership in thedonor community. Whatever the immediate budgetpressures, there are very good reasons for donorsto avoid or minimize aid cuts. Unlike rich countries,many of the world’s poorest countries are illequipped to protect their inhabitants against theeconomic downturn through fiscal expansion.Cutting aid at a time when poverty levels are rising,budgets are under pressure and financing gaps ineducation and health are widening would deal a fatalblow to hopes for accelerated progress towards theinternational development goals. More than that,it would erode the benefits of past aid investments.

The additional financing needed to achieve the 2010targets needs to be placed in perspective. Expressedin absolute financial terms, the US$20 billionfinancing gap appears large. Yet it is equivalent

to a tiny fraction of the estimated cost of bailingout financial systems in advanced economies –and to 0.05% of DAC donors’ collective GNI. Withoutdiscounting the very real budget pressures facingdonor governments, redeeming the aid pledge isaffordable. It might also be viewed as a small priceto pay given the expected gains in education,health and poverty reduction.

It would also boost the credibility of internationalcommitments to the Millennium DevelopmentGoals. Donors need to go beyond reaffirming theircommitments and adopt more credible approachesto monitoring their aid efforts. The long-awaited G8Preliminary Accountability Report adopted at theL’Aquila summit fundamentally failed the credibilitytest in important areas (Box 4.2). Looking to thefuture, the three top priority areas for action are:

Clarifying the commitment to real financialtargets. Developing countries widely interpreteddonors’ 2010 aid targets as constituting a realfinancial commitment to a US$50 billion increasein overall aid, with US$25 billion directed to Africa.All donors should clarify their commitment tothese targets, if necessary by adjusting aid-to-GNItargets for 2010. The United Kingdom decisionto maintain real aid commitments for 2010should serve as a model.

Cutting aid at atime when povertylevels are rising,budgets are underpressure andfinancing gaps in education and health arewidening woulddeal a fatal blow to hopes for acceleratedprogress

plans for significant cuts over 2009–2011 that willlower the aid-to-GNI ratio from 0.20% in 2008 to0.09% in 2011. Moreover, current public spendingplans are uncertain: national authorities havesignalled to the OECD that the constraints on Italy’s public finance will influence the aid trend(OECD-DAC, 2009c).

Canada: modest commitment to continued growth.Canada has pledged to double developmentassistance by 2010, but in nominal rather than real terms and from a baseline of 2001 rather than 2004. OECD estimates indicate that meetingthis commitment will translate into a 3% increasein aid from 2008 to 2010, lower than for any other G8 country.

France and Germany: continued uncertainty.Both countries have increased aid since 2004, but the European Commission does not considereither one on track for achieving the 2010 targets.As of mid–2009, neither country had preparedmulti-annual timetables setting out plans forachieving the targets in the light of weakereconomic forecasts.

Japan: planned increases from a low base. In 2008,Japan’s aid level rose for the first time since 2005,though the country still has one of the world’slowest aid-to-GNI ratios. While details remainunclear, Japan increased its 2009 developmentassistance budget by 13%.

Sources: European Commission (2009b); Sweden Ministry ofFinance (2009); DFID (2009a); Ingram (2009); Irish Aid (2009);OECD-DAC (2009b); Yoshida (2009); World Bank and IMF (2009).

0.0

0.1

0.2

0.3

0.4

0.5

0.6

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

ODA

as %

of G

NI

0.23%

0.56%

0.15%0.20%

0.51%

Trajectory for Spain to reach its 2010 target

Spain

Trajectory for Italy to reach its 2010 target

Italy

Spain

Italy

0.43%

US$3.8 billionin shortfallin 2008

Figure 4.5: Spain is on track to achieve its national aid target while Italy is off trackTotal aid as % of GNI, Spain and Italy, 1999–2008 (net disbursements), and targets for 2010

Source: (OECD-DAC, 2009d).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 2 6

Translating commitments into public spendingplans. Long-term development requirespredictable aid flows. Setting clear aid budgetplans is therefore a priority. Donors forced tomake deep budget adjustments as a result ofthe crisis should adopt the Irish and Swedishapproach of setting a course for recovery.

Monitoring donor delivery more closely. Theaid pledge is a collective commitment. Somevariation in individual donor performance isinevitable. However, in the case of the 2005 aidpledges it is difficult to escape the conclusionthat free-riding has become a problem, withsome donors having to compensate for theweak commitment of others. More rigorousmonitoring and public reporting is required.

Recent trends in aid to education

As governments look to the 2015 target date forachieving universal primary education and widergoals, prospects for accelerated progress willdepend in part on future aid flows. Sustained andpredictable increases in those flows can helpsupport more ambitious education strategies,supplementing the resources available to recruitteachers, construct classrooms and reach themarginalized. Aid delivered to education continuesto rise, but there is no evidence of a concerted driveto mobilize the additional resources needed toachieve universal primary education and othereducation goals. Looking ahead, there is a realdanger that reduced commitments to basiceducation will lead to lower levels of disbursementsover the next few years.

The share of education in overall aid has not changedAid priorities have shifted a great deal in recentyears, with the shares of overall aid devoted tovarious sectors rising, falling or staying the same.Education falls in the third category. The increasein support to education recorded since the Dakarforum in 2000 has been driven principally by theoverall increase in aid rather than redistributionfrom other sectors. In 2006–2007, educationaccounted for about 12% of all aid commitmentsto sectors, the same level as in 1999–2000.7In contrast, health has been a big winner in aidallocations, with an increased share of sectoraid from 11% in 1999–2000 to 17% in 2006–2007.This reflects a surge of bilateral, multilateral andphilanthropic aid directed through global funds andnational programmes.8 United Nations agencies,campaigners, governments and the private sectorhave succeeded in putting health at the centre ofthe international development agenda.

Education financing has not suffered directly as aresult of the rising share of health in aid spending.With overall aid flows increasing, a fixed share stillimplies an increase in real resources. Moreover,investment in health generates important benefitsfor education. What matters in the end is whetheroverall aid flows and aid targets are commensuratewith the commitments donors made in 2000 at the

7. Where a two-year period is indicated, figures have been calculated on the basis of two-year averages, in order to smooth out volatility of aidcommitments.

8. If education had risen at the same rate as health, direct aidcommitments to education would have been US$15.9 billion in 2006-2007. The actual figure was US$10.7 billion.

Commitmentlevels are

stagnating andthe trend

is highly erratic

The annual G8 summits have produced a steady stream of communiquésmaking impressive commitments on education. Recognition in recent yearsof the importance of tracking delivery on these commitments culminatedin the G8 Preliminary Accountability Report adopted at the 2009 summitin L’Aquila, Italy. Its contents fell far short of the required reportingstandards.

The report claims to account for ‘the progress made towards the Educationfor All goals and the Fast Track Initiative’ (p. 16). In fact, it treats G8commitments to the FTI as the sole measure of performance. In contrastto health, where the G8 has adopted a global financing target aimed atachieving international development goals, there is no global educationtarget. The US$1.2 billion FTI replenishment estimate represents a smallfraction of the global basic education financing gap.

To make matters worse, the accounting system for FTI support leavesmuch to be desired. The United Kingdom is the only G8 member to havebeen a major source of Fast Track finance. Successive summits havepledged to close the financing gap, with no effect on delivery. Theaccountability report obscures this failure by including aid for educationin countries receiving Fast Track support as aid to the initiative itself.

Further such reports should take a new approach to benchmarkingin three areas:

The G8 should adopt a credible figure for the global financing gap figurefor Education for All. This Report estimates that gap at US$16 billion.

The summit should agree a ‘fair share’ framework stipulatingcommitments of individual G8 members to investment in basiceducation, based on global financing gaps.

The accountability report should measure real FTI financialcommitments and G8 leaders should provide leadership in reformingand revitalizing the initiative.

Source: Group of Eight (2009a).

Box 4.2: The G8’s disappointing Accountability Report

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 2 7

Dakar forum. Unfortunately, they are not. Ifdonors increased aid in line with their Gleneaglescommittment and the share directed towardsbasic education remained constant, there wouldstill be a financing gap of some US$11 billionagainst the requirements identified in this Report.Donors need to urgently review both the overalllevel of planned aid and its distribution by sector.

Disbursements are still rising — but are commitments waning?Effective national planning also requires a clearindication of how much aid can be expected infuture years. Recruiting teachers in 2010 hasbudget implications for salaries in 2012. Similarly,bringing more children into primary school andensuring that they complete a basic educationrequires planning classroom construction andpurchases of books and other teaching materials.That is why aid commitments, which act as asignal for future disbursements, are important.

Disbursements and commitments are not directlycomparable: aid committed by donors this year maybe allocated to national programmes over one, two,three or more years. Another complicating factoris that several donors, notably some multilateralinstitutions, do not report disbursements to theOECD-DAC and so are not included in the analysisof disbursements in this Report.9

Disbursed aid has been on a steadily risingtrend both for education in general and forbasic education (Figure 4.6). Overall aid flows toeducation reached US$10.8 billion in 2007, morethan double the level in 2002.10 Aid disbursementsto basic education grew more slowly – fromUS$2.1 billion in 2002 to US$4.1 billion in 2007 –indicating a slight distribution shift towardssecondary and post-secondary provision: the shareof basic education in total education disbursementsfell from 41% to 38% over the period.

The picture for aid commitments contrastsstrongly with that for disbursements. Overallcommitment levels are stagnating and the trendis highly erratic (Figure 4.7). In 2007, reportedcommitments stood at US$12.1 billion, aroundthe same level as in 2004.

Basic education remains an area of particularconcern. While aid commitments rose in the yearsafter Dakar, with an increase of 58% between 1999–2000 and 2003–2004, the period since then has been marked by stagnation punctuated by episodes

of steep decline. In real terms, the US$4.3 billionreported in 2007 represented a cut of 22% from2006 – or about US$1.2 billion in real finance –so that commitments were below the 2003 level.The decline in commitments to basic education wasfar greater than that for education as a whole.11

Several factors contributed to the steep declinein aid commitments for basic education. In 2006,the Netherlands and the United Kingdom madelarge pledges to the Catalytic Fund of the FastTrack Initiative. The Netherlands also committedresources to UNICEF for education in countriesaffected by conflict and humanitarianemergencies.12 As a result, commitments fromthe Netherlands and the United Kingdom spikedin 2006 and declined the following year. Thedecline was only partly offset by an increasein commitments from other bilateral donors,

Total aidto education

5.2

7.57.9

8.3

9.7

10.8

2.12.6

3.74.1

Cons

tant

200

7 US

$ bi

llion

s

Total aid tobasic education2.9 3.2

2002 2003 2004 2005 2006 2007

Figure 4.6: Aid disbursements to education have been on a steadily rising trendTotal aid disbursements to education and basic education, 2002–2007

Note: Box 4.3 explains the calculation of total aid to education.Source: OECD-DAC (2009d).

Aid to education comes not only as direct allocations to theeducation sector but also through general budget support. As in previous EFA Global Monitoring Reports, this Report includes part of general budget support as aid to education. It also assumesthat half of all aid to education classified as ‘level unspecified’ is designated for basic education. Thus:

Total aid to education = direct aid to education + 20% of generalbudget support.

Total aid to basic education = direct aid to basic education + 10%of general budget support + 50% of ‘level unspecified’ aid toeducation.

Box 4.3: Assessing the total aid contribution to the education sector

9. The African DevelopmentFund, the Asian DevelopmentFund and the InternationalDevelopment Association(IDA) do not reportdisbursements to the OECD-DAC. Information on IDAdisbursements for educationfor this Report was obtaineddirectly from the World Bankand hence included in theanalysis of disbursements.

10. All figures in thissubsection are expressedin constant 2007 prices.Data on disbursement arenot available before 2002.

11. Commitments foreducation as a whole fellby 2% from 2006 to 2007.

12. These commitmentsamounted to US$553 millionfor the Catalytic Fund andUS$231 million for UNICEF.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 2 8

principally the United States, and some multilateralagencies. The net effect was a 31% decline inbilateral aid commitments to basic educationfrom 2006 to 2007, to below US$3 billion.Another important factor behind the decline wasthe timing of commitments to major aid recipients.Commitments to the twenty largest recipients ofaid to basic education, including Ethiopia, Maliand the United Republic of Tanzania, dropped fromUS$2.8 billion in 2006 to US$2.0 billion in 2007.

While fluctuations in commitments are an inevitablepart of aid programming, recent trends highlightserious systemic problems. One is that a smallgroup of donors dominates aid to education. In 2006and 2007, the five largest donors to education – France, Germany, the Netherlands, the UnitedKingdom and the World Bank’s InternationalDevelopment Association (IDA) – accounted for59% of total commitments to the sector. With acombined commitment of US$3.5 billion, Franceand Germany account for over one-quarter ofoverall aid to education. Similarly, the five largestdonors in basic education – the EuropeanCommission, the IDA, the Netherlands, the UnitedKingdom and the United States – accounted for61% of commitments (Figure 4.8). An importantconsequence of this concentration is that relativelysmall movements by one or two key donors canhave large global consequences, as the combinedeffect of the Netherlands’ and United Kingdom’said programmes showed in 2006 and 2007.

The record on aid commitments to basic educationis a matter of growing concern. Fluctuations on thescale recorded since 2003 raise questions over thepredictability of future disbursements. While thedata in this section are global, volatile commitment

levels have consequences for national budgetsand education planning in many aid-dependentcountries. Developing a broader base of donorsupport for education is one key to a less volatilepattern of commitments.

The distribution of aid to low- and middle-incomecountries has changed little since Dakar(Figure 4.9). In 2006 and 2007, low-incomecountries received just under half of all aid toeducation, on average, and almost 60% of aidto basic education. Middle-income countriesaccounted for nearly two-fifths of overall aid toeducation. Much of that goes to the post-secondarylevel, though these countries account for a quarterof aid to basic education.

Primary education needs to be given higher priorityCountries do not expand the choices open to peoplethrough primary education alone. Progress towardsuniversal primary education brings increaseddemand for secondary education – and secondaryschools have a vital role to play in training teachers.Investment in post-primary education is alsoimportant in developing skills that strengthenprospects for economic growth.

For all these reasons, aid to post-primary educationis justified in terms of the Dakar commitments.The challenge for donors – and for aid recipients –is to achieve the right balance of support for thedifferent levels of education. How successfully arethey meeting that challenge?

In signing the Dakar Framework for Action, donors pledged to increase the share they devoted to primary education and other forms of basic education.13

13. Aid to basic educationcovers pre-primary,primary, literacy and basiclife skills. Comprehensivedata on aid to forms ofbasic education otherthan primary are notreadily available, butprevious editions of theEFA Global MonitoringReport have shown thatthe amounts of aid forthese purposes are verylimited.

A small group of donors

dominates aid to education

Total aidto education

Total aid tobasic education

8.2 7.6 7.9

9.510.4

12.0

9.9

12.3 12.1

1999 2000 2001 2002 2003 2004 2005 2006 2007

7.9

10.010.9

12.2

Cons

tant

200

7 US

$ bi

llion

s

Cons

tant

200

7 US

$ bi

llion

s

1999-2001 2002-2003 2004-2005 2006-2007

3.3 4.0 4.8 4.93.2 3.2 3.4 3.4 4.5 5.6

4.05.5

4.3

Figure 4.7: After rising in the early part of the decade, aid commitments to basic education are stagnatingTotal aid commitments to education and basic education, 1999–2007

Note: The figure on the right takes two or three year averages in order to smooth out volatility and make the overall trend clearer.Source: OECD-DAC (2009d).

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 2 9

Implicitly, this pledge acknowledged that too littleaid was being directed to the primary level,especially in countries that were far from achievinguniversal primary education. Patterns since 2000do not indicate that any major correction has takenplace, however. Around one-quarter of aid isdirectly committed to basic education, which isslightly below the share reported at the time of theDakar forum.14 With the deep cuts in 2007, theshare of basic education in all direct education aidcommitments fell sharply. Beyond basic education,the post-secondary level dominates, accountingfor 38% of total commitments from 1999 to 2007.Sandwiched in between is secondary education,object of around 12% of education aid over theperiod, though the overall level of support for it isrising – seemingly at the expense of commitmentsto basic rather than post-secondary education.

This global picture is the result of highly disparatenational aid profiles. As Figure 4.10 shows,individual donors vary considerably in theircommitments to the different levels of education.Two of the six largest bilateral donors – theNetherlands and the United States – direct over60% of aid to basic education. Three others –France, Germany and Japan – commit over 55%to post-basic education, underpinning the globaldistribution of aid beyond the basic level. A closerlook at the data reveals a strong bias towards post-secondary, with over 70% of French and Germanaid directed towards this level. The figure alsoshows that France and Japan have significantly

14. This refers only to direct aid commitments to basic education, whichexcludes general budget support and ‘level unspecified’ aid (see Box 4.3).

0 500 1 000 1 500 2 000 2 500

IDA

EC

AsDF

AfDF

UNICEF

France

Germany

Netherlands

United Kingdom

Japan

United States

Spain

Canada

Norway

Australia

Belgium

Sweden

Austria

Ireland

Portugal

New Zealand

Switzerland

Denmark

Italy

Finland

Greece

Luxembourg

Constant 2007 US$ millions

1 563

858

185

157

76

2 034

1 510

1 097

843

744

317

288

275

269

193

151

132

89

73

73

65

59

50

48

44

30

Total aidto education

Total aid tobasic education

971

Figure 4.8: The lion’s share of aid is committed by a small group of donorsTotal aid commitments to education and basic education, by donor, 2006–2007 average

Notes: AfDF = African Development Fund, AsDF = Asian Development Fund, EC = European Commission, IDA = International Development Association.Source: OECD-DAC (2009d).

1999-2001 2002-2003 2004-2005 2006-2007

9% 9% 9% 14%

39% 42%36%

38%

52% 49% 55%49%Sh

are

in to

tal a

id to

edu

catio

n

Education

1999-2001 2002-2003 2004-2005 2006-2007

7%5%5%15%

21%28%29%

25%

72%67%65% 60%

Shar

e in

tota

l aid

to b

asic

educ

atio

n

Basic education

Low-income countries

Middle-income countries

Unallocated by country

Figure 4.9: The priority given to low-income countries has not changed since DakarTotal aid commitments to education and basic education, by country income group, 1999–2007

Note: Commitments to the FTI areincluded in the category ‘unallocated by country’. Commitments to the FTICatalytic Fund have increased in recentyears, which explains part of theincrease in the share of ‘unallocated bycountry’ over 2006–2007. Low-incomerecipient countries account for a largemajority of FTI commitments.Source: OECD-DAC (2009d).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 3 0

increased the share of aid commitments directedto post-secondary education. Meanwhile, someother donors have moved in the opposite direction.Spain is one example (Box 4.4).

Promotion of higher education often entails highlevels of aid spending in the donor country. The EFAGlobal Monitoring Report 2009 critically examinedaccounting practices associated with the reportingof post-secondary aid levels. In the case of Franceand Germany, more than four in every five dollarsof the aid reported to the OECD-DAC takes the formof ‘imputed student costs’. This essentially meansthat the estimated costs of teaching students fromdeveloping countries in French and German tertiaryinstitutions are counted as aid to the students’countries. In Germany, ¤701 million of the¤714 million allocated to higher education in theaid programme is spent in this way, representing

around 68% of German aid to education (GermanFederal Ministry for Economic Cooperation andDevelopment, 2009). While counting domesticspending on higher education as aid is consistentwith OECD reporting rules, civil society groupsin both countries and some French senatorshave regularly contested its legitimacy.

Such criticism does not imply that support forhigher levels of education is unimportant. Ashighlighted in Chapter 2, many donors haveneglected technical and vocational education andtraining. An exception is Germany: building on itsextensive experience at home, Germany spent¤77 million in 2007 supporting vocational educationand is one of the largest donors to the subsectorin the world. It finances the reform and expansionof vocational education in countries including Egypt,Ethiopia, Mozambique and Uganda. As part of a

Promotion ofhigher education

often entails highlevels of aid

spending in thedonor country

IDAEC

AsDFAfDF

UNICEF

FranceGermany

NetherlandsJapan

United KingdomUnited States

SpainCanada

AustraliaNorwayBelgiumAustria

SwedenIreland

PortugalNew Zealand

SwitzerlandItaly

GreeceDenmark

FinlandLuxembourg

IDAECAsDFAfDFUNICEF

FranceGermanyNetherlandsJapanUnited KingdomUnited StatesSpainCanadaAustraliaNorwayBelgiumAustriaSwedenIrelandPortugalNew ZealandSwitzerlandItalyGreeceDenmarkFinlandLuxembourg

0 500 1 000 1 500 2 000 -50 -40 -30 -20 -10 0 10 20 30 40

Constant 2007 US$ millions Change between 1999-2000 and 2006-2007in share of post-secondary education

in direct aid to education (percentage points)Level unspecifiedBasic Secondary Post-secondary

Figure 4.10: Only a few donors give priority to basic educationDirect aid commitments to education by level, 2006–2007 average, and change in the share of post-secondary education between 1999–2000 and 2006–2007

Notes: Direct aid to education falls into four subcategories: basic, secondary, post-secondary and ‘level unspecified’. Aid to education not allocated to a particular level of education is recorded as ‘level unspecified’. AfDF = African Development Fund, AsDF = Asian Development Fund, EC = European Commission, IDA = International Development Association.Source: OECD-DAC (2009d).

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 3 1

wider strategy for achieving the MillenniumDevelopment Goals, Germany has responded torequests from national governments to scale upsupport for skills development as a means ofraising wages and tackling youth unemployment (German Federal Ministry for Economic Cooperation and Development, 2009). German aid plays animportant role in supporting the reform of technicaland vocational education, and there is scope forother donors to follow the lead provided. The widerchallenge for all donors is to find the right balance.In countries where poverty remains a huge obstacleto achieving universal primary education, the casefor investing the bulk of scarce aid resources inhigher education that overwhelmingly benefithigher-income students is not credible.

Emerging donors, private giving and innovative financeOECD-DAC members continue to dominateinternational development assistance. Butimportant new sources of aid are emerging, someof which could give a significant boost to education.

Overall aid from countries that are not DACmembers is on a strong upward trend. While oftenreferred to as ‘emerging donors’, many membersof this diverse group have a long history of providing

aid to developing countries. In 2007, aid from non-DAC donors reporting to DAC15 amountedto US$5.6 billion – four times the level in 1999.The largest emerging donor is Saudi Arabia, whichspent US$2.1 billion on aid in 2007. Aid from Brazilhas been estimated at US$437 million and thatfrom India at US$1 billion. Official data are notavailable for China, but estimates point to a totalChinese aid budget of US$1-1.5 billion in 2006(OECD-DAC, 2009c).

Strong economic growth, the size of the externalbalances available to major economies such asBrazil, China and India, and growing cooperationin areas such as trade and energy could drivea sustained expansion in aid from non-OECDcountries. This makes it all the more important toimprove the flow of information and coordinationbetween all donors (Manning, 2006). Achievingthat outcome will require a broadening of aidgovernance structures, which need to be reformedto ensure that the views of emerging donors aretaken into account when developing policies andidentifying priorities.

Too little is known about the composition of non-DAC donors’ aid portfolios to assess their aid toeducation. China has supported school construction

15. Including the CzechRepublic, Hungary, Iceland,Israel, Kuwait, Poland, theRepublic of Korea, SaudiArabia, the Slovak Republic,Turkey and the United ArabEmirates.

Leadership by the SpanishGovernment couldplay an importantrole in renewingand revitalizingthe education aidagenda

The rapid emergence of Spain as a major donorhas given important impetus to the widerinternational aid effort. It has backed its increasedaid flows with a stronger commitment to equity,especially in education.

Spanish aid has undergone a remarkabletransformation since 2000. The aid-to-GNI ratiohas doubled to reach 0.43%. Spain’s rapideconomic growth means this translates into alarge increase in real financial transfers. Moreover,an initially narrow focus (on Latin America) hasbroadened, and the country has curtailed its tyingof aid to Spanish exports and commercial interests.Since the adoption of the Second Master Plan(2005–2008) there has been a much strongerfocus on poverty reduction. The plan, which tookthe Millennium Development Goals as its keyreference point, set out a detailed strategy foraligning a wide range of policies and ministriesbehind Spain’s development cooperation goals.

The plan placed Education for All squarely at thecentre of the aid agenda. At the time of the WorldEducation Forum, Spain directed most of its aid toeducation towards the tertiary level. Since Dakar,it has not only increased the level of aid but hasshifted it in favour of basic education. Total aidcommitments to education grew from an annualaverage of US$268 million in 1999–2000 toUS$316 million in 2006–2007, an 18% expansion.Over the same period, total aid to basic educationincreased by 79% to US$144 million, or just underhalf of Spain’s total aid to education.

Leadership by the Spanish Government could playan important role in renewing and revitalizing theeducation aid agenda. In the first half of 2010,Spain assumes the presidency of the European Union.Spain is well placed to call on other EU membersto demonstrate a stronger commitment to aid foreducation, including a commitment to greater equitybetween sectors.

Sources: Manzanedo and Vélaz de Medrano (2009); OECD-DAC (2009d).

Box 4.4: Spain: political will behind increased aid to basic education

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 3 2

programmes in sub-Saharan Africa. It has alsoincreased its support for training in its external aid.By 2007, more than 80,000 people from developingcountries had participated in short- and medium-term training courses, in such fields as agriculture,health, management and education, supported byChinese aid (Brautigam, 2008). The Republic ofKorea, which aims to become a member of the DACin 2010, has a strong focus on infrastructure forsocial services. Education is one of seven prioritysectors in the country’s Mid-Term ODA Strategyand accounted for 14% (US$70 million) of itsbilateral aid in 2007 (OECD-DAC, 2008c). InSeptember 2008, Saudi Arabia joined the launchof Education for All: Class of 2015, a new globalinitiative. It pledged US$500 million in concessionalloan financing for basic education – its first suchundertaking and one pointing to a greater sharefor basic education in its overall lending (Educationfor All: Class of 2015, 2008).

Data on private aid are not comprehensive, but theavailable evidence points to strong growth in recentyears. In 2007, private aid for international purposesreported to the OECD reached US$18.6 billion,which almost certainly understates the real flow(World Bank and IMF, 2009). Internationalfoundations – such as the Bill and Melinda GatesFoundation – and corporations dominate theseaid flows, largely directed towards public health.A recent survey showed that 43% of contributionsfrom United States-based foundations were aimedat health in developing countries, with only 6%directed to education (World Bank and IMF, 2009).

That picture could be starting to change. Severalnew education initiatives have emerged recently,many involving innovative private-publicpartnerships. In 2008, the Open Society Institutecontributed US$5 million to the Liberia PrimaryEducation Recovery Program. This is one of the firstcases of a private foundation, and multilateral andbilateral donors pooling resources in support ofa national education programme – an approachthat is well established in the health sector throughthe Global Fund to Fight AIDS, Tuberculosis andMalaria and the GAVI Alliance (see final section). In2007, the United Nations refugee agency (UNHCR)launched its ‘ninemillion’ campaign in cooperationwith Nike and Microsoft. It aims to raiseUS$220 million by 2010 to give 9 million refugeeand vulnerable children access to education, sportsand technology (UNHCR, 2007). Projects havealso been initiated in conjunction with the WorldEconomic Forum, in several cases with a focus

on the use of information and communicationtechnology in education. The Jordan EducationInitiative, supported by the Jordanian Government,private corporations and non-governmentorganizations, works with teachers and pupilsto promote interactive learning in 100 ‘DiscoverySchools’ (Light et al., 2008).

There is no shortage of innovative financingmodels to inform approaches in education. Manylessons can be drawn from experiences in thehealth sector. The International Finance Facilityfor Immunisation (IFFIm) has mobilized aroundUS$1.2 billion through government bond issues.The first Advance Market Commitment, amechanism aimed at creating incentives for thedevelopment of new drugs to treat poverty-relateddiseases, has generated US$1.5 billion (GAVIAlliance, 2009a). Climate change is another areaincreasingly characterized by creative thinking.Education aid agencies and campaigners, however,have been slow to respond to innovative financingmodels. It is vital to ensure that the interestsof the world’s 72 million out-of-school childrenare not crowded out of innovative financing bycompeting claims in other areas.

Avoiding that outcome will require more effectivecampaigning and advocacy, backed by more incisivepolitical leadership in the United Nations system.Opportunities for action have to be exploited. Oneexample is the 2010 football World Cup, which isbecoming an important focal point for internationalaction and campaigning on Education for All. In2009, France and the United Kingdom reaffirmed ajoint pledge to get an additional 8 million children inschool by the start of the World Cup, though details– especially with respect to the French aid budget –remain unclear. The Global Campaign for Educationis working with the Fédération Internationale deFootball Association (FIFA) and several majorEuropean football leagues in the lead-up to theWorld Cup to raise awareness of the educationproblems facing sub-Saharan Africa, along withsome additional financing (1 Goal, 2009). However,awareness raising and limited voluntarycontributions are not enough.

Innovative financing could go on benefitingeducation well after the 2010 World Cup events areover. An agreement by the major European leaguesto place a small (0.4%) EFA levy on futuresponsorship and media marketing revenue couldgenerate some US$48 million annually. Channelledthrough a reformed FTI or another multilateral

In 2007, private aid for

internationalpurposes

reported to theOECD reached

US$18.6 billion,which almost

certainlyunderstates the

real flow

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 3 3

mechanism, these resources would enable oneof the world’s most popular sports to make areal difference in the lives of some of the world’spoorest children (Box 4.5).

ConclusionIncreases in aid to education since 2000 have beenthe result of improvements in overall levels of aid,rather than a shift in donor priorities. Aid for basiceducation has also been rising, but there is a large

gap between current levels of provision andthe estimated US$16 billion required to achievethe EFA goals. This gap will widen if recent fallsin commitments to basic education translate intolower future disbursements. At Dakar, donorspledged to increase the share of aid to educationdevoted to basic education, but this shift has nottaken place. If progress to Education for All is tobe accelerated, donors need to make a concertedeffort to mobilize the additional resources required.

There is a largegap betweencurrent levels of provision and the estimatedUS$16 billionrequired toachieve the EFAgoals

In 2010, Africa will host the World Cup for the firsttime. The event will set a benchmark for globalsporting competitions. Apart from being the firstsuch event to be staged in Africa, it will be watchedby more people and generate more media andsponsorship revenue than any World Cup in history.With leadership from FIFA, its national members,clubs, footballers and supporters across the world,the World Cup could also set a benchmark for fightingdeprivation in education.

Directing to education just a small proportion of therevenue flowing into the industry could make a bigdifference in the lives of out-of-school children.Consider what might be achieved through a modestlevy on media and marketing revenue (Table 4.1).

The 2010 World Cup is setting new records. As of May 2009, it had generated US$3.4 billion incommercial revenue — a 48% increase over the 2006World Cup. The sale of media rights is the singlebiggest contributor. Revenue flowing to nationalmembers of FIFA in the rich world reflects the growth

of the global market for football. The five major rich-country leagues — England, France, Germany, Italy and Spain — account for commercial revenue of US$11.1 billion annually. Their broadcast andsponsorship revenue amounts to US$7.8 billion. To put that figure in context: it is more than double all international aid for basic education in low-incomecountries.

Placing a modest ‘Better Future’ levy on footballrevenue would generate potentially significantamounts. For example, a levy of 0.4% would mobilizearound US$48 million annually — less than someEuropean clubs spend on a single footballer yetsufficient to finance a basic education of decentquality for approximately half a million of the world’sout-of-school children each year to 2015.

A model that could provide guidance is that of Futbol Club Barcelona, which has created afoundation that receives 0.7% of the club’s ordinaryincome and directs it towards global povertyreduction efforts. To follow this good example, theproposal set out in this box would enable all majorfootball clubs to unite in a global philanthropiceffort. Directing the revenue towards a reformedFTI (see ‘Reforming the Fast Track Initiative’ below)would help maximize the benefits, minimizetransaction costs and revitalize multilateral aidfor education. Football could do for basiceducation what the Gates Foundation and otherphilanthropic interventions have done for theGlobal Fund in health.

The World Cup is an event that will be rememberedfor many important goals. But its most lasting legacycould be helping to bring basic education into thelives of some of the world’s poorest children anddemonstrating to governments that, with goodleadership, the goal of universal primary educationis still attainable.

Sources: Sportcal (2009); Deloitte LLP (2009).

Box 4.5: Education for All and the football World Cup

3 511 14 140 4302 068 8 82 7272 068 8 82 7272 044 8 81 7491 422 6 56 897

850 3 34 000

11 963 48 478 530

EnglandGermanySpainItalyFrance

World Cup

Total

Table 4.1: Football revenue and school levy

Notes: Based on a recurrent unit cost of US$100 per child in primary school. No account is taken of the capital costs (e.g. classrooms) required to provideprimary schooling. The commercial revenue for the World Cup is averaged over four years to provide an annual revenue figure.Sources: Sportcal (2009); Deloitte LLP (2009).

Annualcommercial

revenue(US$ million)

Revenuefrom 0.4%

school levy(US$ million)

Estimatednumber of

primary school places provided

Major football leagues

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 3 4

Ways to make aid more effective

The quality of aid, by its very nature, is more difficultto measure than quantity – but no less important.The 2005 Paris Declaration on Aid Effectivenessmarked an attempt by donors and aid recipientsto identify institutional arrangements that canstrengthen the impact of development assistance.Recognizing the national planning problems causedby uncertainty over the timing of aid flows, donorsacknowledged the need for greater predictabilityin delivery. They also recognized the importanceof supporting and working through national publicfinancial management systems. Greater controlby recipient countries and improved donorcoordination in support of national plans were seenas antidotes to donor-driven aid programmes thatbypassed national structures and reinforced aiddependence, often without delivering sustainableresults (Deutscher and Fyson, 2008). Specifictargets were adopted to change this picture by 2010.

The Paris agenda has a very direct bearing on aidfor education. Perhaps more than in any othersector, planning for education requires predictablemedium-term finance. The cost of paying teachers,meeting per pupil costs and financing textbookprovision stretches over many years. Forgovernments lacking a sustainable and predictablerevenue base, ambitious public investment ineducation is a high-risk enterprise. Donors’ use ofpublic financial management systems in educationis also critical. Reporting through national systems,rather than parallel donor systems, candramatically reduce transaction costs. Similarly,given the significant presence of many donorsin aid for education, aid agency coordination insupporting national plans can reduce the burdenon already overstretched education planners.

Progress towards the targets set in the Parisagenda has been mixed. Given the 2005 baseline,any overall assessment would be premature.There has been progress in most areas, albeit froma low base. However, its pace will have to pick upover the next three years if the goals are to beattained (Table 4.2). For instance, less than half ofdevelopment assistance is currently reported in aidrecipients’ budgets, against a 2010 target of 85%.

The 2008 High-Level Forum on Aid Effectivenessin Accra gave renewed momentum to dialoguebetween donors and developing countries on aidgovernance. Donors recognize the need to put inpractice the principles underpinning the Paris

agenda, but this will require a fundamental shift inthe way many donors manage their aid – which willin turn require an even more fundamental shift inhow they think about aid partnerships. This sectionlooks at four areas of the broad Paris agenda thathave important implications for education:

the predictability of aid;the use of national public financial managementsystems;donor coordination; andperformance-based aid.

Aid predictabilityPredictability is a hallmark of effective aid. Ifrecipients cannot rely on donor commitments,they cannot develop and implement medium-termfinancing plans for achieving education goals.

There is a great deal of room to improvepredictability. In 2007, less than half of aid arrivedon schedule. For some countries, the figure was farbelow that. In Yemen, just one-third of scheduledaid was disbursed. Benin was to receiveUS$477 million but just US$151 million actuallyarrived (OECD, 2008a). Such shortfalls can havehighly damaging effects in education, disruptingschool building programmes and limiting theresources available to hire teachers and providechildren with textbooks.

Not all the problems associated with unpredictableaid can be traced to donors. If recipientgovernments cannot account for previouslydisbursed funds or meet basic reporting conditions,there may be strong grounds for delaying aid.Donor aid management systems are often partof the problem, however, imposing unrealisticconditions or onerous reporting requirements.In most cases, there are problems on both sidesof the aid partnership. In the United Republic ofTanzania, disbursements for the Primary EducationDevelopment Programme (2001–2006) wereconsistently below commitments. Delays inapproval of work plans, poor quality audit reportsand demanding donor reporting requirements allcontributed. Using national systems can helpstrengthen predictability by removing a layerof transactions in reporting. Aid predictability isparticularly weak in conflict-affected countries.In 2007, less than half the aid scheduled fordisbursement was delivered in the DemocraticRepublic of the Congo, Nepal and Sierra Leone.In Chad and Liberia, none of the scheduled aidwas disbursed that year (OECD, 2008a).

Reportingthrough nationalsystems, rather

than paralleldonor systems,

can dramaticallyreduce

transaction costs

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 3 5

One underlying cause of poor predictability is weakdonor planning. Recipient countries are encouragedto develop three to five year expenditureframeworks in areas such as health and education,but donors have made little progress providingreliable multiyear aid estimates. While some donorshave legislated multiyear aid commitments, mosthave no binding commitments – and thisinformation is not always shared with recipients(OECD-DAC, 2009c). At the 2008 Accra forum on aideffectiveness, donors reaffirmed their commitmentto improving medium-term aid predictability andto providing regular, timely information for a threeto five year period on the levels of aid developingcountries can expect to receive. It is crucial thatthey act on these pledges.

Use of country public financial management systemsThe efficiency, integrity and transparency thatgovernments demonstrate in mobilizing, managingand spending public resources and in reportingto citizens are at the heart of good governance.Aid recipients have made progress in keepingcommitments they made in the Paris agenda tostrengthen public financial management systems,but donors are not keeping their promise to usethose systems as much as possible, thusweakening incentives for reform.

In many developing countries, progress instrengthening public financial managementsystems has been slow, not least because theinstitutional arrangements are complex (de Renzio,

While some donorshave legislatedmultiyear aidcommitments,most have nobindingcommitments

Operational development strategies

Reliable public financial managementsystems

Aid recorded in country budgets

Technical assistance coordinated

Donors use country public financialmanagement systems

Donors use country procurementsystems

Donors avoid parallel ProjectImplementation Units

Aid is disbursed on schedule

Aid is untied

Donors use programme-basedapproaches

Donors coordinate their missions

Donors coordinate their country studies

Transparent and monitorableperformance frameworks

Mechanisms for mutual accountability

17% 24% 12%

– 36% –

42% 48% 14%

48% 60% Achieved

40% 45% 13%

39% 43% 10%

1 817 1 601 18%

41% 46% 17%

75% 88% –

43% 47% 17%

18% 21% 14%

42% 44% 9%

7% 9% 7%

22% 26% 5%

75% of recipient countries havethese strategies

50% of countries improve quality

85% of all aid on budget

50% coordinated with country programmes

80% of aid to government using nationalsystems

80% of aid to government using nationalsystems

611 parallel implementation units

71% of funds disbursed in the yearscheduled

Progress over time in percentageof aid untied

66% of aid using these approaches

40% of donor missions coordinated

66% of country studies undertakenjointly

35% of recipient countries with these frameworks

100% of countries have reviews of mutual accountability

Table 4.2: Progress on Paris Declaration targets, 2007

Notes: The percentage of target achieved is calculated by dividing the change between 2007 and 2005 by the difference between the target and the baseline figure. The assessment is based on the thirty-three recipient countries included in the first monitoring survey. Targets for the use of country public financial management and procurement systems represent maximums, as targets vary by country depending on system quality in 2005.Source: OECD-DAC (2008a).

IndicatorParis DeclarationPrinciple

2005baseline

% of targetachieved 2010 targets

2007results

Ownershipand alignment

Aid is predictableand untied

Harmonizationwith partners

Managing for resultsand accountability

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 3 6

2009; de Renzio and Dorotinsky, 2007). However, thelatest OECD survey, from 2008, found that one-thirdof the forty-two low-income countries covered hadimproved their financial management systems byat least one measure in the Country Policy andInstitutional Assessment (CPIA), a World Bankdiagnostic tool that ranks performance on anascending scale from one to six (OECD, 2008a).

Donors set an ambitious target of channelling80% of aid through national systems by 2010,but between 2005 and 2007 the actual amountincreased from 40% to just 45%. Moreover, thequality of a country’s public financial managementsystem is a weak guide as to whether donors use it,as Figure 4.11 illustrates. Bangladesh scores loweron the CPIA scale than Mozambique, Rwanda orZambia, yet has a far higher share of aid usingnational reporting systems.

In any one country, donor perceptions of corruption,organizational incentives, legislation governing aid,the direction of reform and headquarters policiescan play a far more important role in shaping policythan a CPIA score. The extent to which individualdonors use national financial management systems

varies widely (Figure 4.12). More than 60% of aidfrom France, Japan, the Netherlands, Spain andthe United Kingdom goes through national financialmanagement and procurement systems, comparedwith 35% of aid from the European Commission andonly 5% of United States aid. Some countries, suchas France and Spain, have been willing to channelaid through weak national systems while supportingefforts to strengthen them.

Channelling aid through national systems gives aid-dependent countries far greater control overbudget planning and public spending, and reducesthe costly need to create parallel managementsystems. It makes little sense for the EuropeanCommission to require Zambia to meet separatereporting requirements when individual EUmembers are willing to work through the country’snational system. There is much greater scope fordonors to work creatively together in supportingand using effective national systems.

Furthermore, scaling up aid to the required levelthrough current financial arrangements is nota viable option. It would entail a proliferation ofseparate and parallel management structures

The quality of acountry’s public

financialmanagement

system is a weakguide as

to whetherdonors use it

0

10

20

30

40

50

60

70

80

90

% o

f aid

to g

over

nmen

t sec

tor u

sing

natio

nal p

ublic

fina

ncia

l man

agem

ent s

yste

ms

Central African Republic

SudanTogo

Mauritania

Bangladesh

Burundi

CambodiaKyrgyzstan

Yemen

Mongolia

Cape VerdeNiger

Nepal

KenyaCameroon

BeninMozambique

Mali

Quality of public financial management system (CPIA score)

Côte d’Ivoire Congo D. R. Nigeria

Lao PDR

MalawiAfghanistanHaiti

Zambia

Bolivia

Low High

U. R. Tanzania

Viet Nam

UgandaHonduras

Burkina FasoEthiopia

Rep. MoldovaRwanda

GhanaNicaragua

Papua N. GuineaSenegalMadagascarSierra Leone

1.5 2.0 2.5 3.0 3.5 4.0 4.5

Figure 4.11: The extent to which donors use recipients’ financial systems is not related to their qualityDonor use and quality of public financial management systems, 2007

Source: OECD (2009a).

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 3 7

that would overload the capacity of developingcountries, divert scarce human resources fromnational planning, weaken budget systems andultimately diminish the effectiveness of aid.

Aligning aid and coordinating activityAll donors are committed in their policystatements to aligning their activities with theplans of recipient governments. Better alignmentalso means improved coordination, with donorsworking collectively to support the goals set outin national plans.

One indicator of progress in this regard is theshare of programme-based aid. In 2005–2006,it accounted for some 54% of all aid to basiceducation, compared with 31% in 1999–2000. InBangladesh, donors have formed a consortium thatworks with the government on a unified programmeof support for primary education. Mozambique andZambia have also seen a strong shift towards pooledfunding for education, with donors working togetherthrough national systems and shared reportingstructures. In some cases, donors have cooperatedin supporting reforms in planning, reporting andauditing to facilitate a pooled financing arrangementand the scaling up of aid in support of the nationaleducation strategy (Box 4.6).

While improved donor coordination is deliveringresults, it can give rise to new tensions.Negotiations between aid-dependent countriesand groups of like-minded donors can reinforceunequal power relationships (Abou Serie et al.,2009). In the United Republic of Tanzania,Education Ministry officials saw dialogue withdonors as a source of intrusion, while donorsreported concerns over a perceived exclusionfrom discussions over programme implementation(Box 4.7). Such tensions highlight the complexityof aid partnerships and the importance of settingclear parameters for donor influence.

Managing for resultsUnder the Paris Declaration on Aid Effectiveness,developing country governments committed tostrengthening monitoring of the progress that aid isintended to facilitate and donors pledged to supportthese efforts and to use national data. ‘Managing for results’ is the shorthand description of this approach.

There is some evidence that the stronger focuson results is influencing national educationprogrammes supported by aid. In Bangladesh, forexample, the national primary education programme

IrelandSpainJapan

NetherlandsFrance

NorwayDenmark

FinlandUnited Kingdom

World Bank Sweden

Italy Germany

CanadaAsian Development Bank

African Development Bank Austria

SwitzerlandIDB

European CommissionBelgium

AustraliaNew ZealandUnited States

Portugal Luxembourg

0 10 20 30 40 50 60 70 80 90

Average percentage of aidto the government sector using national public

financial management and procurement systems

Figure 4.12: The use of recipient financial management systems varies by donorSelected donors’ use of national public financial management systems, 2007

Source: OECD (2009a).

Recent experience from Nicaragua highlights the importance of trustand good communication between government officials and donors in strengthening management systems to increase aid alignment.

With the adoption of the National Education Plan in 2001, Nicaraguaset out to harmonize external aid to education. Several instrumentswere introduced for managing aid, with an emphasis on using nationalprocedures for financial planning, reporting, auditing and procurement.A pooled fund financed by Canada, Denmark and the Netherlands hasbeen a particularly important resource. It provides predictable finance,which can be used flexibly to pay for non-salary activities agreed inthe Education Ministry’s annual plan.

The introduction of the pooled fund required a strengthening ofnational management and planning capacities. Close dialogue andfrank discussions between senior officials and donors on keymanagement elements were critical to the successful managementof the pooled funds. Donors have also agreed to accept a single financialaudit for the entire annual budget, replacing multiple donor audits.

Source: Jané (2008).

Box 4.6: Nicaragua — strengthening managementsystems through aid alignment

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 3 8

has made result-based management a prioritythrough improved information sharing since 2004.In the past, government departments rarely sharedinformation for sector planning. New systems arebreaking down this fragmentation step by step,with an annual report providing an overview ofperformance measured against key educationsector objectives (Bangladesh Government, 2009).

At its best, managing for results is aboutstrengthening the capacity of developing countrygovernments to determine what works on the basisof the best available evidence. The more extensiveuse of outcome indicators to measure policyeffectiveness is a wholly positive development.The drive towards results in aid managementis not without problems, however.

Some donors see performance-based funding as anobvious corollary of a commitment to managementby results. While such funding takes many forms,the broad approach is to create incentives forgovernments to strengthen policies that are notachieving targets and to reward those that areperforming. The United States Millennium ChallengeAccount, created in 2004, provides funding on thebasis of policy reforms and development results.The Global Fund model also uses incentives toimprove poorly performing programmes.In Senegal, for example, a grant to combat malaria

was stopped due to underperformance andrestored only once the national programmehad been strengthened. Negative reviews ofprogrammes in the Laos People’s DemocraticRepublic, Lesotho and Nigeria have also ledto policy reform (Global Fund, 2009d).

Does performance-based funding conflict with theprinciple of country ‘ownership’? All aid is to somedegree conditional on recipient governments beingseen as viable partners – and on results. Recipientsare likely to see performance-based funding aslegitimate if they have a role in setting goals anddeciding how best to achieve them (Abou Serie et al.,2009). In the case of the Global Fund, the centralrole of developing country governments and civilsociety in setting national targets, submittingfinancing and implementation plans, and jointlyreviewing progress creates a basis for countryownership. While there have sometimes beensevere tensions over the release of funds, in manycases governments already committed to reformappear to have accepted financing incentives.

Under different conditions it is a small stepfrom performance-based support to old-styleconditionality, or worse. Recent proposals infavour of ‘cash-on-delivery’ aid for educationillustrate the problem (Box 4.8).

The moreextensive use

of outcomeindicators to

measure policyeffectiveness is a

wholly positivedevelopment

From 2001 to 2006, nine donors pooled funds to supportthe Primary Education Development Programme in theUnited Republic of Tanzania, with the World Bankproviding additional support. Measures in the programmeincluded the abolition of school fees in 2001, theintroduction of capitation grants for primary schools and a major classroom construction programme. Publiceducation spending rose, backed by increased aidcommitments. Donors were closely involved with theMinistry of Education in designing and implementingpolicies, and in financing. While the programme hasbrought about remarkable improvements in basiceducation, there have been strains in the aid partnership:

Pooled fund disbursement was often delayed becauseof what donors saw as unclear quarterlyimplementation plans and late, inadequate progressreports from the government.

Different reporting requirements for the World Bank andfor donors working under the pooled fund overstretchedgovernment employees responsible for reporting.

Donors reported concerns over exclusion fromdiscussions at key stages of programmeimplementation.

Education Ministry officials felt that policy dialogue with donors was often intrusive, sometimes leading to additional aid conditions, and that donors did notrespect the principles of country ownership.

Civil society representatives reported that they wereoften crowded out by the large number of donors and a lack of access to information.

In 2007, all donors previously contributing to the pooledfund turned to general budget support. This was partly in response to difficulties managing the pooled fund andpartly because the Government of the United Republic of Tanzania said it preferred budget support.

Sources: Williamson et al. (2008); World Bank (2005b).

Box 4.7: Harmonization and alignment in the United Republic of Tanzania education programme

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 3 9

ConclusionTranslating the Paris Declaration principles intopractical strategies requires donors and recipientsto reconsider the distribution of political power inaid partnerships. Effective aid requires a nationalpolicy environment in recipient countries that isconducive to planning. It also requires donors to act

on their commitments to deliver more predictableaid. Donors also need to resist the temptationto micromanage aid, either formally (throughconditionality) or informally (through control overfinance). Delivering development assistance in waysthat strengthen national capabilities is not justmore effective – it is a route out of aid dependence.

By linkingpayments toverified results,cash-on-deliveryaid has thepotential to create perverseincentives

Linking aid to results has an intuitive appeal. If thegoal is decent quality education, why not rewardgovernments with a cash payment for every additionalchild who completes primary education or achievesabove a set score on a standardized test? This is thecentral idea behind cash-on-delivery aid, which aims to provide incentives for recipients to address theinstitutional and governance problems that canprevent aid from producing results.

The appeal of cash-on-delivery aid is its focus onresults. Payments to aid recipients would be made on the basis of verified improvement in outcomes (say, children completing primary education andreaching a specified learning standard) from anestablished baseline. Recipient governments would beleft free to decide on policies and on how to spend theaid they receive. While superficially offering a route togreater ownership, this model poses several problems:

Penalizing governments for outcomes they do notcontrol. School attendance figures and completionrates can be strongly affected by factors such asdroughts, floods, unemployment and economicgrowth. In theory, an external auditor could adjust achieved outcomes (and aid payments) by controlling for exogenous factors, and donorscould renegotiate their contract with aid recipients.In practice, unravelling the effects of variousinfluences requires data that are either unavailableor not likely to become available until much later.

Shifting the risk. Development is a risky business.Neither national governments nor aid donors knowin advance with any certainty which policy inputs(public investment, targeted incentives, governancereforms and so on) will work. By conditioning aid onbroadly shared policy inputs, donors share the riskof failure with the recipient. Basing aid on outputtransfers risk to the recipient. If a particular input,designed and implemented with a genuine intent to achieve a positive outcome, does not work, the would-be aid recipient loses out while the donoris unaffected. Governments might adopt policiesaimed at removing a set of barriers to education of the marginalized, only to find that the policiesproduce weaker results than expected, incurringcash-on-delivery aid penalties. In effect, this is

old-style conditionality on a no-risk basis for donors.Far from encouraging innovation in aid recipientcountries, cash-on-delivery could have the oppositeeffect, creating incentives to avoid risk-taking.

Diverting attention from the strengthening ofsystems. Cash-on-delivery aid places a premium onachieving short-term targets, such as getting morechildren through primary school, rather than long-term goals such as strengthening the educationsystem, improving child nutrition and training moreteachers. For governments that choose cash-on-delivery aid for quantitative targets, there are alsopotential tensions with qualitative goals, as hasbeen widely documented in the health sector.

Creating incentives for misreporting. By linkingpayments to verified results, cash-on-delivery aidhas the potential to create perverse incentives, withgovernments being rewarded for over-reporting —another phenomenon documented in the healthsector. Programmes under the auspices of the GAVIAlliance include a payment for every vaccinatedchild above a baseline. Research indicates that insome countries, including Bangladesh, Indonesiaand Mali, official data systematically understate the baseline and overstate subsequent coverage.

Bypassing ‘underperformers’. Cash-on-delivery aid effectively penalizes countries that miss theirtargets. This raises the immediate question of whatto do with such countries, many of which are likelyto be in the greatest need of support. Should theybe disregarded? Or should it be assumed that theprospect of increased aid will create an incentive for policy change?

Accelerated progress towards education for allrequires far-reaching changes in monetary and non-monetary incentives, backed by changes in rules foraccountability and reporting, aimed at changinginstitutional behaviour. Under some limited conditions,cash-on-delivery approaches might complementbroader performance-based incentives, but theyshould be developed in the context of national policy,not unequal negotiations between donors andrecipients.

Sources: Birdsall et al. (2008); de Renzio and Woods (2007);Lockheed (2008); Lim et al. (2008).

Box 4.8: Cash-on-delivery aid raises as many problems as it solves

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 4 0

Aid to conflict-affected countries

Low-income countries affected by conflict pose someof the greatest challenges for aid partnerships.People living in these countries need help to rebuildtheir livelihoods, health and education systems. Yetfor donors, working with conflict-affected countriesis difficult and often dangerous.

Analysis of the role of aid to education in conflict-affected states is not straightforward. There isno agreed definition or list of such states. Even ifa list could be agreed, the status of the countrieson it would vary enormously. The situation in theDarfur region of the Sudan is not the same as thatof Helmand Province in Afghanistan. Some preferthe broader term ‘fragile states’ to encompasscountries affected by conflict and those facingwider governance challenges, but this does littleto add clarity: almost all low-income countriesare fragile in some way.

There is broad agreement, however, that conflicthas had devastating consequences for education inmany poor countries, affecting millions of children.Whether they are injured or traumatized bybombing in Gaza, living in camps for displacedpeople in Sri Lanka or recruited as child soldiers innorthern Uganda, children are never immune to theimpact of conflict. Neither are education systems.Warring factions often destroy schools and targetteachers, and education suffers badly when conflictleads to a collapse of governance.

Childhood disrupted as a result of conflict is difficultto mend. Yet education can provide children andyouth with protection, a safe space and hope for thefuture. Similarly, the reconstruction of educationsystems in countries emerging from conflict canplay a vital role in underpinning peace, rebuildinglives and laying the foundations for stability (Aguilarand Retamal, 2009). The experience of Sierra Leonedemonstrates what is possible, while the failureto rebuild education in the Democratic Republicof the Congo demonstrates the corrosive effectof slow social reconstruction on peace processes.

Most donors recognize the importance ofsupporting education in conflict-affected countries.Yet they face difficult policy dilemmas. Donors wantaid to be effective, so they focus on conditions suchas country ownership, macroeconomic stability andgood governance. Few countries emerging fromconflict are in a position to meet these conditions.In addition, maintaining access to education during

humanitarian emergencies is enormously difficult.Such considerations help explain the highlyunequal, volatile and poorly coordinated patternof aid delivery to conflict-affected countries.Yet adequate education provision in these countrieswill not be achieved without scaling up aid.

Monitoring aid to conflict-affected countriesHow do countries affected by conflict fare inattracting aid, in comparison with other countries?This Report addresses the question by focusingon twenty poor countries meeting establishedcriteria for classification as conflict-affected(Harbom and Wallensteen, 2009; Uppsala ConflictData Program, 2009).16

The diversity of the group underlines theproblems in defining conflict-affected countries.Those covered include countries such as Liberiaand Rwanda that have embarked on successfulpost-conflict recovery strategies, countries thathave faced localized conflict (Senegal and Uganda)or far broader conflict (Côte d’Ivoire), and thosesuch as Afghanistan where reconstruction istaking place amid continued instability.

The impact of conflict on educational access isclear. Taken collectively, these twenty countriesaccount for about one in three children who areout of school.17 In many cases, national data makeit difficult to establish the full consequences ofconflict. For example, there are no reliableestimates of the out-of-school population in Darfur.In other cases, national data can obscure the extentof conflict-related damage to education. WhileUganda has made strong national progress towardsuniversal primary education, several northerndistricts affected by conflict have been left behind.

While aid to conflict-affected poor countries isrising from a low base, it still falls far short ofwhat is needed. For 2006–2007, just under one-fifthof overall aid to education and one-quarter of aidto basic education went to conflict-affected poorcountries (Figure 4.13). Data limitations makeit difficult to provide an accurate assessmentof the levels of aid required for education inthese countries. Indicative estimates for thisReport put the basic education financing gap inconflict-affected poor countries at approximatelyUS$7 billion or 41% of the total gap for low-incomecountries (Education Policy and Data Center andUNESCO, 2009). This is substantially more than theUS$1.2 billion of aid for basic education committedto these countries in 2006-2007.

16. The countries includedare ones that experiencedarmed conflicts resultingin at least twenty-fivebattle-related deaths peryear over at least threeyears between 1999 and2007 or more than1,000 battle-relateddeaths in at least one yearduring the same period.Of these, only countriescategorized as leastdeveloped countries bythe United Nations or low-income countries bythe World Bank in 2007were included.

17. These twentycountries account for 56%of those out of school inlow-income countries.

Thereconstruction

of educationsystems incountries

emerging fromconflict can play

a vital role inunderpinning

peace, rebuildinglives and layingthe foundations

for stability

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 4 1

Aid distribution within the group of conflict-affectedcountries is highly concentrated in Afghanistan,Ethiopia and Pakistan. They accounted for morethan half of total aid to basic education in conflict-affected countries in 2006–2007 (Figure 4.14).18

Comparisons across the group reveal strikingdisparities in levels of support. Afghanistanreceived US$19 per primary school age child –eight times as much as the Democratic Republicof the Congo at US$2 (Figure 4.15). Rwandareceived US$20 per child and Burundi US$13(Box 4.9). At US$4 per child, Liberia received lessthan half the group average in 2006–2007.

Patterns of aid allocation do not correspond towhat might be expected on the basis of a globalassessment of need. One reason may be that aidpriorities have emerged as a key element of aglobal security agenda. An obvious case in pointis Afghanistan, which receives a large amountof aid overall and for education in particular.This is partly because reconstruction of educationsystems is recognized both as a requirement forhuman development and greater gender equity,and as a vital element in state-building. However,it is also because of the perceived threat of thecountry to global security.

Comparisons between Afghanistan and theDemocratic Republic of the Congo illustrate theimportance of donor priorities. In the DemocraticRepublic of the Congo, long-running civil conflicts,fuelled in some regions by neighbouring states,have had devastating consequences for education.Household survey data indicate that more than4 million children are not in school (DemocraticRepublic of the Congo Ministry of Planning et al.,2008).19 While the country may be a sourceof regional instability, donors do not perceiveit as a global security threat, unlike Afghanistan.That may explain why it figures among the top tenrecipients for only one donor – Belgium, theformer colonial power. By contrast, Afghanistanwas among the top ten recipients of basiceducation aid for eight donors in 2007.20

18. The twenty countries received 16% of total ODA in 2007, similar to their share of total aid to education. In this group, Afghanistan,Ethiopia and Pakistan are also the biggest recipients of ODA overall (see OECD-DAC, 2009e), suggesting that education is following moregeneral patterns of donor priorities.

19. GMR calculations based on net attendance rate from DRC 2007 DHSand population data from UIS database.

20. Australia, Canada, the European Commission, Germany, the IDA,Japan, Sweden and the United States.

Total aidto education

Total aid tobasic education

1999-2001 2002-2003 2004-2005 2006-2007

15%

19% 19%

24%

Shar

e of

aid

toco

nflic

t-affe

cted

poo

r cou

ntrie

s

12%14%

17%15%

Figure 4.13: Conflict-affected poor countries receive a low share of aid to educationShare of total aid to education and basic education (commitments) allocated to conflict-affected poor countries, 1999–2007

Source: OECD-DAC (2009d).

Pakistan

Ethiopia

Afghanistan

Senegal

Uganda

Nepal

Rwanda

Sudan

D. R. Congo

Côte d’Ivoire

Burundi

Angola

Sierra Leone

Eritrea

Chad

Myanmar

Somalia

C. A. R.

Guinea-Bissau

Liberia

0 50 100 150 200 250 300 350 400

Constant 2007 US$ millions

Total aidto education

Total aid tobasic education

371

280

138

136

127

68

66

61

55

43

36

35

22

18

17

13

11

9

5

11

Figure 4.14: Distribution of aid to education among conflict-affected poor countries is unevenTotal aid disbursements to education and basic education in conflict-affected poor countries, 2006–2007 average

Source: OECD-DAC (2009d).

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 4 2

In Afghanistan, aid has played a critical rolein expanding education opportunities. Overall,however, the aid allocation patterns raisequestions about donor priorities regarding thedifferent recipient countries. In some cases,there are marked disparities in aid levelsbetween conflict-affected countries in the sameregion, or even neighbouring countries – suchas Burundi and Rwanda (Box 4.9).

From humanitarian to development aid — the missing link‘Support for the re-establishment and continuity ofeducation must be a priority strategy for donors andNGOs in conflict and post-conflict situations,’ wroteGraça Machel in 1996 (Machel, 1996, p. 47). Morethan a decade later, most aid for conflict-affectedcountries continues to be delivered through short-term, uncoordinated projects that fail to lay thefoundations on which to rebuild education systems.While donor policy statements increasinglyrecognize the importance of integrating short-termhumanitarian assistance with long-term social andeconomic reconstruction, progress towards a more‘joined-up’ policy framework has been limited.

Humanitarian assistance covers a broad spectrumof activities, but countries affected by violentconflict figure prominently among recipients.Such aid has increased since 1999–2000, thoughits share in total aid commitments declined from9% in 1999–2000 to 7% in 2006–2007 (OECD-DAC,2009d). Estimates suggest that educationaccounted for just 2% of total humanitarian aid – a meagre US$237 million in 2008 (Office forthe Coordination of Humanitarian Affairs, 2009).

In many conflict-affected countries, expenditureon security operations and emergency assistanceoverwhelmingly dominates donor support, withlong-term development in general – and educationin particular – taking a back seat. In Liberia, forexample, the cost of United Nations peacekeepingoperations has consistently been more than doubletotal aid flows since 2004. Only 2% of the total aidwas allocated to education in 2004–2007. Duringthis post-conflict phase, humanitarian aidcontinued to play a significant role (OECD-DAC,2009d). But it did not make up the shortfall foreducation: in humanitarian as in development aid,education accounted for just 2%. This suggeststhat longer-term, more sustainable approachesto supporting basic service delivery are not yetbeing addressed (Figure 4.16).

Another example comes from the DemocraticRepublic of the Congo. In the five years after thesigning of the 2003 peace accord, developmentaid was nominally higher than spending on UnitedNations peacekeeping. However, this was largelybecause, under an agreement signed just afterthe peace accord, creditors wrote off a large shareof the country’s debt stock, which is counted asaid even though it entails no real financial flows.Humanitarian aid has been a significant proportionof actual assistance from donors, reflecting thedifficult environment in which they operate.The 2003 peace accords swiftly broke down,as did subsequent accords. Today, humanitarianaid is dominated by food and emergency nutritionprogrammes, with long-term aid to agriculture,education and health figuring only marginally.In 2007, US$5 million, or only 1% of humanitarianaid, supported education interventions in theDemocratic Republic of the Congo, far shortof the US$27 million identified as a minimumrequirement for education in the 2007humanitarian action plan (United Nations, 2007a).

The experience of the Democratic Republic of theCongo illustrates both the relative neglect of social

Today,humanitarian aid

is dominated by food andemergency

nutritionprogrammes,

with long-termaid to agriculture,

education andhealth figuring

only marginally

2620

1917

141414

1310

99

66

55

44

322

129

Constant 2007 US$

EritreaRwanda

AfghanistanSenegal

NepalEthiopia

Sierra LeoneBurundi

Guinea-BissauAngola

UgandaPakistan

SudanSomalia

ChadLiberia

C. A. R.Côte d’IvoireD. R. Congo

Myanmar

Average conflict-affectedAverage other low-income

Figure 4.15: Spending per primary school child is low in conflict-affected poor countriesTotal aid disbursements to basic education per primary school agechild, 2006–2007 average

Note: Only low-income countries with a primary school age population above150,000 are included.Source: OECD-DAC (2009d).

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 4 3

The scale of theRwandan genocideand the failure ofthe internationalcommunity toprevent it played a role in eliciting a strong aidresponse

Burundi and Rwanda are known in French as lesfaux jumeaux — the non-identical twins. Both haveexperienced devastating episodes of violent conflictwhich have left a deep imprint on their educationsystems. One area where they differ is in the levelof support they have received from aid donors inrebuilding those systems.

The three months of genocide in Rwanda in 1994left 800,000 people dead and 3 million displaced,many of them in neighbouring countries. Some 80% of the country’s children experienced death in their immediate family and 90% saw dead bodies. In Burundi, the conflict was more protracted.From 1993 to 2005, out of a population of around6 million, 300,000 people were killed and 1.6 millionfled their homes.

Both countries emerged from conflict with shattered education systems. When a newgovernment assumed office in Rwanda in July 1994,the Education Ministry had no financial resources,no equipment or supplies, and limited manpower. In Burundi, the near-decade of conflict severelyweakened education planning and financing. By2000, just 40% of the school age population wereattending primary school, according to householdsurvey data. With large numbers of traumatized

children, a bitter legacy of mistrust, shortages of teachers and large financing gaps, both countries urgently needed strong donor support and increased aid.

The donor response has been unequal. Over2006–2007, Rwanda received US$20 per primaryschool age child. Burundi received just US$13, eventhough it is lagging behind Rwanda in progresstowards universal primary education, with threetimes as many children out of school.

Financing disparities of this magnitude are difficultto square with an independent assessment of need,governance or capacity. Other factors have drivenaid allocations. The scale of the Rwandan genocideand the failure of the international community to prevent it played a role in eliciting a strong aidresponse — and rightly so. Public pressure on donorsto act was reinforced by graphic media coverage of the genocide. Beyond the humanitarian impulse,many aid donors see Rwanda as a more significantstrategic actor than Burundi in the Great Lakesregion. The point of the comparison is not toquestion the level of aid to Rwanda, which hasachieved extraordinary progress, but to ask whydonors have not supported reconstruction inBurundi more strongly.

Source: Obura and Bird (2009).

Box 4.9: Non-identical donor responses to education systems in Burundi and Rwanda

reconstruction in humanitarian aid and thecomplexity of the problems facing donors. Food and nutrition are obvious priorities for emergency support. Yet the failure to put in place a viable strategy andadequate finance for education reconstruction maywell have contributed to wider factors that havedestabilized successive peace accords.

At one level the aid financing profiles for theDemocratic Republic of the Congo and Liberiareflect a compelling set of recovery imperatives. The problem is not that the international communityinvests too much in security and alleviating hunger.It is that too little is invested in other areas that areno less important to post-conflict reconstruction.

Peace, political stability, access to basic servicesand economic recovery cannot be viewed inisolation. In a post-conflict environment, failure inany one area can lead to collapse in others. Whenpeace settlements bring an end to violence but failto restore education systems, the thwarted hopesand ambitions of parents can fuel social tensions

and mistrust of government. Distributing foodto combat hunger without restoring the economicinfrastructure and productive systems that peopleneed for more secure livelihoods can erodeprospects for sustainable recovery. The bottom lineis that security in the broader sense is about morethan the absence of violence and hunger. It is aboutexpanding the real choices open to people andbuilding confidence in the future.

Working effectively in conflict-affected statesThere is no ready-made model for working inconflict-affected states. In some cases, peaceprocesses create an opportunity to work withgovernments committed to reconstruction. Inothers, donors work amid ongoing conflict, withthe risk of being seen as a supporter of one side – a risk that has resulted in a growing numberof attacks on aid workers. In still other cases,government unwillingness to participate in peaceprocesses or reconstruction may leave non-government groups as the only potential partnersfor aid agencies. While the problems are oftendaunting, there are always opportunities to engage.

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 4 4

The potential for engagement depends partlyupon donor governance practices. Stringent ruleson reporting may be beyond the capacity of manyconflict-affected states. Similarly, the scope forsupporting country ownership and using publicfinancial management systems – key elements ofthe 2005 Paris agenda for aid effectiveness – maybe limited. Donors have adopted principles to guidetheir work in fragile and conflict-affected states,21

but there is a lack of clarity over the alignment ofthese principles with those of the Paris agenda –

and over how to translate them into action(Oxford Policy Management and IDL Group, 2008).

One principle for international engagement infragile states emphasizes the importance ofacting fast, but also staying involved for longenough to give success a chance (OECD-DAC,2007). This often does not happen. One reasonis that countries emerging from conflict are thoughtto lack the governance systems to absorb largequantities of aid. That may be true immediately

21. These include‘Principles and GoodPractice of HumanitarianDonorship’ (2003) and‘Principles for InternationalEngagement in FragileStates and Situations’(2007). The latter include aprinciple emphasizing theimportance of aligning withlocal priorities in differentcontexts.

Multi- and nonspecified sector

0

200

400

600

800

2004 2005 2006 2007

UN peacekeeping Total aid

Humanitarian aid

Total aid to education

Cons

tant

200

7 US

$ m

illio

ns, d

isbur

sem

ents

Liberia

Humanitarian aid in 2007

Multisector

Food

Protection and human rights

Agriculture2%, Education

1%, Health

0

400

800

1 200

1 600

2 000

2004 2005 2006 2007

UN peacekeeping Total aid

Humanitarian aid

Total aid to education

Cons

tant

200

7 US

$ m

illio

ns, d

isbur

sem

ents

Democratic Republic of the Congo

Humanitarian aid in 2007

Water and sanitation, 1%

48%

25%

17%

6%

Food

Support and coordination

Health

Water and sanitation

Protection and human rightsAgriculture

1%, Education2%, Shelter Infrastructure, 1%

39%

19%

14%

9%

6%

5%4%

Figure 4.16: Peacekeeping and reconstruction in the Democratic Republic of the Congo and Liberia* United Nations peacekeeping expenditure, humanitarian aid and development aid disbursements, Democratic Republic of the Congo and Liberia, 2004–2007

Notes: Data from the United Nations Department of Peacekeeping Operations (UNDPKO) are used as a proxy for peace- and security-related expenditure. The data indicate the peacekeeping missions under UNDPKO command and the levels of financial support being provided to them by United Nations member states (see OECD-DAC, 2009e).* The sector breakdown in humanitarian aid is taken from the Financial Tracking Service, United Nations Office for the Coordination of Humanitarian Affairs (OCHA). Unlike for the OECD-DAC,reporting to OCHA is voluntary and so likely to be an underestimate. For some countries (notably Liberia in Figure 4.16) multisector aid can be a sizable proportion of humanitarian aid. In Liberia it was mainly aimed at supporting the return and reintegration of refugees once the political and security situation improved (United Nations, 2007b).Sources: OECD-DAC (2009d); Office for the Coordination of Humanitarian Affairs (2009).

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 4 5

after a conflict, when security and the restorationof basic governance are an immediate priority. Butonce peace has taken root, there is often a potentialfor increasing aid (Collier and Hoeffler, 2002).

All too often in the past, donors tended to scaleback aid in countries that had emerged fromconflict but remained politically unstable two orthree years after a peace settlement. The upshotwas that aid declined at a time when publicconcerns were shifting from security to basicservices – and when post-conflict governmentswere building their capacity to use aid moreeffectively (Weinstein et al., 2004). There arealternatives. Recognizing that uncertainty overfuture aid flows could compromise efforts to buildon the Sierra Leone peace settlement, in 2002the United Kingdom Department for InternationalDevelopment made a ten-year commitment tosupport the government. Similar arrangementswere later put in place for Afghanistan, Ethiopiaand Rwanda (DFID, 2005).

Seizing opportunities for reconstruction requiresflexible policies and a strong commitment toworking in conflict-affected countries. Some donorsare integrating into their policies approaches toproviding education in conflict and emergencies(Brannelly et al., 2009). Even so, only ten of thetwenty-three OECD-DAC members have policycommitments to providing education in countriesaffected by conflict and fragility,22 and only fiveinclude education in their emergency policies(Save the Children, 2009b).23

The risks associated with working in conflict andpost-conflict environments can entail hightransaction costs for measures such as securityassessment, engagement with government andnon-government actors, and the design of practicalreporting and evaluation systems. Many donorshave developed innovative strategies for loweringrisks and transaction costs, adapted by context.24

Investing in pooled funds managed by anotherdonor with a strong track record in the recipientcountry is one approach (Box 4.10). TheNetherlands has committed US$231 million,25

around 15% of its direct aid to education in 2006,for the period up to 2010 to a joint programme withUNICEF aimed at supporting education in countriesin conflict and emergencies. Norway has reducedits bilateral aid for education in Afghanistan andincreased support provided through the AfghanistanReconstruction Trust Fund, a multidonor trust fundmanaged by the World Bank (Brannelly et al., 2009;

Save the Children, 2009b). Other innovativeapproaches to aid delivery in conflict-affectedcountries include the following:

In Guatemala, the 1996 peace accords includeda commitment to support the development ofindigenous education. With the help of fundingfrom Norway directed through Save theChildren’s Rewrite the Future Campaign,60,000 children are reported to have benefitedfrom improved education quality, with therecruitment and training of bilingual teachersand curriculum development playing importantroles (Save the Children, 2009b).

The development of a basic educationprogramme (focusing on pre-primary, primaryand adult education) in Nepal in 2004 shows itis possible for donors and government to worktogether even amid serious armed conflict. Inthis case, Denmark, Finland, Norway, the UnitedKingdom and the World Bank contributed to apooled fund. There are indications of some gainsin educational attainment as a result, despitethe conflict (Berry, 2007).

In Somalia, since the early 1990s the EuropeanCommission has supported education throughinternational non-government organizations,focusing on basic education, teacher training andvocational or life-skills training for disadvantagedyouth (Brannelly et al., 2009). Support hascontinued through periods of intense conflictand a non-functioning government.

Education in northern Uganda has beenseriously hindered by violent conflict, withschools and teachers targeted by the Lord’sResistance Army. Aid from the Netherlands hashelped finance a bursary programme for formerLord’s Resistance Army combatants in the north,along with other programmes helping childrenand youth catch up on missed schooling(Save the Children, 2009b).

Another example comes from Canada, which hasdramatically increased its overall aid budget foreducation and its support for conflict-affectedcountries. In seeking to align aid financing witha national policy commitment to the reconstructionof education systems, the Canadian InternationalDevelopment Agency (CIDA) has demonstrateda high level of flexibility. In Afghanistan and theSudan, it has allocated resources to multidonortrust funds and non-government organizations.

22. Australia, Canada,Denmark, the EuropeanCommission, Ireland, theNetherlands, Norway, Spain,the United Kingdom and theUnited States. The WorldBank, which participatesin the OECD-DAC as anobserver, also has a policyon providing education inconflict-affected countries.

23. Canada, Denmark, Japan,Norway and Sweden.

24. The first principle forInternational Engagementin Fragile States is to takecontext as the starting point,recognizing that capacity,political will and legitimacydiffer according to whethera country is in a prolongedconflict or recently emergingfrom conflict, for example.

25. At 2007 prices.

Only ten of the twenty-threeOECD-DACmembers have policycommitments to providingeducation incountries affectedby conflict and fragility

01

02

Educatio

n f

or A

ll G

lobal

Monit

orin

g R

eport

C H A P T E R 4

2 4 6

In the Palestinian Autonomous Territories, CIDAhas a strong focus on providing safe spaces forlearning and institutional support for UNRWA,the United Nations agency for Palestinianrefugees, rather than providing funding directlyto the Palestinian authorities. In Sierra Leone,a country in which Canada has no significantaid programme, support for reintegrating childsoldiers has been delivered via non-governmentorganizations (Mundy, 2009).

Aid for conflict-affected countries is not justabout development in a narrow sense. In manysuch countries, donor governments are engagedin wide-ranging military and security operations,diplomatic activity and the rebuilding of basicgovernance systems. These overlapping rolesentail threats and opportunities for effective aid.The threats derive from the risk that aid will beused, or be perceived by the people of recipientcountries, as one element in a wider militarystrategy. Yet the integration of aid into a widerpolicy framework can create opportunities formore effective delivery.

Some donors are attempting to organize theiraid programmes in conflict-affected countriesthrough a ‘whole of government’ frameworklinking development, defence and diplomacy(OECD, 2006c).26 For example, in the DemocraticRepublic of the Congo, the United Kingdom hassupported bridge-building programmes aimed atrestoring economic infrastructure under a pooledfinancing arrangement involving the defenceministry’s conflict prevention fund (DFID, 2009b).Another example, which illustrates the difficultrelationship between development and security,comes from Afghanistan, where Canada isdeveloping an education programme in a provincemarked by severe insecurity (Box 4.11).

Effective multilateral approaches to aid can playa vital role in supporting conflict-affected countries.Such mechanisms enable bilateral donors to poolresources and risk, and to avoid having to createtheir own delivery systems. One problem with theglobal aid architecture is the lack of a single unifiedmultilateral framework for education through whichdonors can channel resources to conflict-affectedcountries. Hopes that the Fast Track Initiativewould fill the gap have not been realized – anissue discussed further in the final section.

26. The ‘whole ofgovernment’ approachis related to one of theprinciples for internationalengagement in fragilestates, recognizing thelink between politicalsecurity and developmentobjectives.

One problem with the global

aid architectureis the lack of

a single unifiedmultilateral

framework foreducation

through whichdonors can

channelresources

Multidonor trust funds, which pool the fundsof several donors, are usually managed by theWorld Bank or a United Nations agency. A recentevaluation of such funds identifies them as thebest-practice post-crisis funding mechanism.They are among the most important coordination,harmonization and alignment mechanisms opento donors.

In difficult post-conflict environments, multidonortrust funds offer donors several advantages.They spread fiduciary risk and reduce the costof initiating programmes and providing support.For aid recipients, they can reduce transactioncosts and provide early delivery of urgentlyneeded support. Experience in Liberia has shownthat funds can be disbursed quickly to support reconstruction of education activities (see Box 4.15). Similarly, the Afghanistan Reconstruction TrustFund has helped pay the salaries of the expandedteaching force needed for the increased numbersof children entering school since 2002.

Not all multidonor trust funds have beenas effective. A World Bank-managed fund inSouthern Sudan covers about 12% of donorfunding to education. Disbursement has beenslow. This is partly because of weak capacityon the part of the new Education Ministry, but ithas also been reported that stringent applicationof World Bank procurement rules has made rapiddisbursement difficult.

Multidonor trust funds offer considerablepotential for scaling up support to conflict-affected countries. The money currentlyallocated to such funds is usually a small shareof total aid to conflict-affected countries, assome donors continue to provide assistancethrough separate projects or directly to non-government groups. This dilutes the potentialbenefits of pooling resources, placing additionaltransaction costs on recipient governments.

Sources: Scanteam (2007); Brannelly et al. (2009); Greeley (2007); Echessa (2009).

Box 4.10: Multidonor trust funds — a promising approach with mixed results

T H E A I D C O M PA C T: FA L L I N G S H O R T O F C O M M I T M E N T S

A i d f o r e d u c a t i o n

2 4 7

ConclusionThe international development goals for educationwill not be achieved without scaled up aid effortsin conflict-affected states. These states accountfor a large share of the out-of-school population.The recovery of their education sectors is hamperedby inadequate finance, weak technical capacityand chronic shortages of teachers. The difficultiesin providing support to the people of thesecountries are well known. Yet opportunitiesto rebuild education are being lost as a resultof overly rigid aid management practices andthe failure to develop an effective multilateralvehicle to support conflict-affected countries.

The internationaldevelopment goalsfor education willnot be achievedwithout scaled upaid efforts inconflict-affectedstates

The ‘whole of government’ model that Canadahas adopted in Afghanistan is an attempt tounify diplomacy, defence and developmentwithin a single policy framework. What doesthis mean in practice?

Multiple Canadian ministries have joined tocreate pooled funds, such as the Global Peaceand Security Fund, supporting ‘whole ofgovernment’ programmes not only inAfghanistan but also in Haiti, Iraq, the PalestinianAutonomous Territories and the Sudan. Thefunds are an integrated source of finance coveringeverything from police training to emergencyfood aid and education. One aim is to bridgethe divide between short-term humanitarianaid and long-term development aid.

Experience in Afghanistan has played animportant role in shaping the development ofthis approach. Education has been a focal point.In 2006 and 2007, Canada provided an averageof US$168 million in bilateral aid to Afghanistan.Nearly 13% was allocated to basic education,making Canada one of the country’s largestdonors to the sector.

Approaches to education have been shapedthrough a complex interaction between Canadian

security commanders and development experts.In 2007, Canadian forces were redeployed toKandahar Province. Leaders within the forcesidentified education as a priority concern andcalled for a strengthened focus on schools,teachers and textbooks. As the CanadianInternational Development Agency became moreinvolved, clearer guidelines on civil-militaryinteraction in the education sector werenegotiated to ensure that Afghanistan’s Ministryof Education, rather than Canada, was seen asdelivering education services. The Canadian forcessupport CIDA activities by providing securityescorts, assessing and planning infrastructuresuch as school perimeter walls and providingintelligence about local security.

Critics argue that linking military and humansecurity clashes with poverty-reduction priorities.They raise questions about Canada concentratingresources in Kandahar instead of offering broadernational support and about a blurring of civil andmilitary responsibilities. Nevertheless, Canada’sexperience provides lessons for efforts to supporteducation in conflict-affected contexts wheredonors might otherwise avoid working for fearof the high risks involved.

Sources: CCIC (2009a, 2009b); Simpson and Tomlinson (2006);Mundy (2009).

Box 4.11: Canada’s ‘whole of government’ approach in Afghanistan