23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 1

REACHING THE POOR

The costs of sending children to school

a six country comparative study

Synthesis report
by Siobhan Boyle, Andy Brock, John Mace & Mo Sibbons
August 2002
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 2

Reaching the Poor - The costs of sending children to school

Educational Papers

Department for International Development: Educational Papers D

This is one of a series of Education Papers issued by the Policy Division of the Department N

For International Development. Each paper represents a study or piece of commissioned

research on some aspects of education and training in developing countries. Most of the
studies were undertaken in order to provide informed judgements from which policy
decisions could be drawn, but in each case it has become apparent that the material produced
would be of interest to a wider audience, particularly those whose work focuses on
developing countries. N

Each paper is numbered serially, and further copies can be obtained through DFID
Education Publication Despatch, PO Box 190, Sevenoaks, TN14 5EL, UK subject to
availability. A full list appears overleaf.
N

Although these papers are issued by DFID, the views expressed in them are entirely those of
the authors and do not necessarily represent DFIDs own policies or views. Any discussion
of their content should therefore be addressed to the authors and not to DFID.

Photography: Shafiqul Alam Drik. DFID

DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 3

Reaching the Poor - The costs of sending children to school

Educational Papers

DFID EDUCATION RESEARCH PAPERS

No.1 SCHOOL EFFECTIVENESS IN No. 7 PLANNING AND FINANCING

DEVELOPING COUNTRIES: SUSTAINABLE EDUCATION
A SUMMARY OF THE SYSTEMS IN SUB-SAHARAN
RESEARCH EVIDENCE. AFRICA.
D Pennycuick (1993) P Penrose (1993)
d
ISBN: 0 90250 061 9 ISBN: 0 90250 067 8

No. 2 EDUCATIONAL COST-BENEFIT No. 8 Not allocated

ANALYSIS.
J Hough (1993) No. 9 FACTORS AFFECTING
ISBN: 0 90250 062 7 FEMALE PARTICIPATION
IN EDUCATION IN SEVEN
No.3 REDUCING THE COST OF DEVELOPING COUNTRIES.
TECHNICAL AND C Brock, N Cammish (1991)
VOCATIONAL EDUCATION. (revised 1997).
L Gray, M Fletcher, P Foster, M King, ISBN: 1 86192 065 2
A M Warrender (1993)
ISBN: 0 90250 063 5 No.10 USING LITERACY: A NEW
APPROACH TO POST-
No. 4 REPORT ON READING LITERACY METHODS.
ENGLISH IN PRIMARY A Rogers (1994) Out of Print
SCHOOLS IN MALAWI. Available on CD-ROM and
E Williams (1993) Out of Print DFID website. Updated and
Available on CD-Rom and reissued on No 29.
DFID website
No.11 EDUCATION AND TRAINING
No. 5 REPORT ON READING FOR THE INFORMAL
ENGLISH IN PRIMARY SECTOR.
SCHOOLS IN ZAMBIA. K King, S McGrath, F Leach,
E Williams (1993) Out of Print R Carr-Hill (1994)
Available on CD-Rom and ISBN: 1 86192 090 3
DFID website
No.12 MULTI-GRADE TEACHING:
See also No. 24, which updates and A REVIEW OF RESEARCH
synthesises Nos 4 and 5. AND PRACTICE.
A Little (1995)
No. 6 EDUCATION AND ISBN: 0 90250 058 9
DEVELOPMENT: THE ISSUES
AND THE EVIDENCE.
K Lewin (1993)

ISBN: 0 90250 066 X

DFID
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 4

Reaching the Poor - The costs of sending children to school

Educational Papers

No.13 DISTANCE EDUCATION IN No.19 GENDER, EDUCATION AND N

ENGINEERING FOR DEVELOPMENT - A PARTIALLY
DEVELOPING COUNTRIES. ANNOTATED AND SELECTIVE
T Bilham, R Gilmour (1995) BIBLIOGRAPHY.
Out of Print Available on CD-ROM C Brock, N Cammish (1997)
and DFID website. Out of Print Available on CD-ROM
and DFID website. N
No.14 HEALTH & HIV/AIDS
EDUCATION IN PRIMARY & No.20 CONTEXTUALISING
SECONDARY SCHOOLS IN TEACHING AND LEARNING
AFRICA & ASIA. IN RURAL PRIMARY
E Barnett, K de Koning, SCHOOLS USING
V Francis (1995) AGRICULTURAL EXPERIENCE. N
ISBN: 0 90250 069 4 P Taylor, A Mulhall (Vols 1 & 2)
(1997) Vol 1 ISBN: 1 861920 45 8
No. 15 LABOUR MARKET SIGNALS Vol 2 ISBN: 1 86192 050 4
& INDICATORS. N
L Gray, AM Warrender, P Davies, No.21 GENDER AND SCHOOL
G Hurley, C Manton (1995) Out of ACHIEVEMENT IN THE
Print Available on CD-ROM CARIBBEAN.
and DFID website. P Kutnick, V Jules, A Layne (1997)
ISBN: 1 86192 080 6
No.16 IN-SERVICE SUPPORT FOR A
TECHNOLOGICAL APPROACH No.22 SCHOOL-BASED
TO SCIENCE EDUCATION. UNDERSTANDING OF N
F Lubben, R Campbell, HUMAN RIGHTS IN FOUR
B Dlamini (1995) COUNTRIES: A
ISBN: 0 90250 071 6 COMMONWEALTH STUDY.
R Bourne, J Gundara, A Dev,
No.17 ACTION RESEARCH REPORT N Ratsoma, M Rukanda, A Smith,
ON REFLECT METHOD OF U Birthistle (1997)
TEACHING LITERACY. ISBN: 1 86192 095 4 N
D Archer, S Cottingham (1996)
ISBN: 0 90250 072 4 No.23 GIRLS AND BASIC EDUCATION:
A CULTURAL ENQUIRY.
No.18 THE EDUCATION AND D Stephens (1998)
TRAINING OF ARTISANS FOR ISBN: 1 86192 036 9
THE INFORMAL SECTOR IN N
TANZANIA. No.24 INVESTIGATING BILINGUAL
D Kent, P Mushi (1996) LITERACY: EVIDENCE FROM
ISBN: 0 90250 074 0 MALAWI AND ZAMBIA.
E Williams (1998)
ISBN: 1 86192 041 5

DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 5

Reaching the Poor - The costs of sending children to school

Educational Papers

No.25 PROMOTING GIRLS No.32 SECTOR WIDE APPROACHES

Y EDUCATION IN AFRICA. TO EDUCATION.
E N Swainson, S Bendera, R Gordon, M Ratcliffe, M Macrae (1999)
E Kadzamira (1998) ISBN: 1 86192 131 4
ISBN: 1 86192 046 6
M No.33 DISTANCE EDUCATION
No.26 GETTING BOOKS TO SCHOOL PRACTICE: TRAINING &
PUPILS IN AFRICA. REWARDING AUTHORS.
D Rosenberg, W Amaral, C Odini, H Perraton, C Creed (1999)
T Radebe, A Sidib (1998) ISBN: 1 86192 136 5
ISBN: 1 86192 051 2
No.34 THE EFFECTIVENESS OF
E. No.27 COST SHARING IN EDUCATION. TEACHER RESOURCE
P Penrose (1998) CENTRE STRATEGY.
ISBN: 1 86192 056 3 Ed. G Knamiller, G Fairhurst (1999)
ISBN: 1 86192 141 1
No.28 VOCATIONAL EDUCATION
AND TRAINING IN TANZANIA No.35 EVALUATING IMPACTS (OF
AND ZIMBABWE IN THE EDUCATION PROJECTS &
CONTEXT OF ECONOMIC PROGRAMMES).
REFORM. Ed. V McKay, C Treffgarne (1999)
P Bennell (1999) ISBN: 1 86192 191 8
ISBN: 1 86192 061 X
No.36 AFRICAN JOURNALS
No.29 RE-DEFINING POST- LITERACY A SURVEY OF THEIR
IN A CHANGING WORLD. USAGE IN AFRICAN
A Rogers, B Maddox, J Millican, UNIVERSITY LIBRARIES.
K Newell Jones, U Papen, A Alemna, V Chifwepa,
A Robinson-Pant (1999) D Rosenberg (1999)
ISBN: 1 86192 069 5 ISBN: 1 86192 157 8

No.30 IN SERVICE FOR TEACHER No.37 MONITORING THE

DEVELOPMENT IN SUB PERFORMANCE OF
SAHARAN AFRICA. EDUCATIONAL
M Monk (1999) PROGRAMMES IN
ISBN: 1 86192 074 1 DEVELOPING COUNTRIES.
R Carr-Hill, M Hopkins, A Riddell,
No.31 PRODUCTION OF LOCALLY J Lintott (1999)
GENERATED TRAINING ISBN: 1 86192 224 8
MATERIALS.
I Carter (1999)

ISBN: 1 86192 079 2

DFID
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 6

Reaching the Poor - The costs of sending children to school

Educational Papers

No.38 TOWARDS RESPONSIVE No.43 COMPUTERS IN SECONDARY O

SCHOOLS SUPPORTING SCHOOLS IN DEVELOPING
BETTER SCHOOLING FOR COUNTRIES: COSTS AND R
DISADVANTAGED CHILDREN OTHER ISSUES. E
(case studies from Save the Children). A Cawthera (2001)
M Molteno, K Ogadhoh, E Cain, ISBN 1 86192 418 6
F
B Crumpton (2000)
S
ISBN: to be confirmed No.44 THE IMPACT OF HIV/AIDS ON
THE UNIVERSITY OF
E
No.39 PRELIMINARY BOTSWANA: DEVELOPING A
D
INVESTIGATION OF THE COMPREHENSIVE
ABUSE OF GIRLS IN STRATEGIC RESPONSE. E

ZIMBABWEAN JUNIOR B Chilisa, P Bennell, K Hyde (2001)

SECONDARY SCHOOLS. ISBN: 1 86192 467 4 A
F Leach, P Machankanja with P
J Mandoga (2000) No.45 THE IMPACT OF HIV/AIDS ON P
ISBN: 1 86192 279 5 PRIMARY AND SECONDARY
EDUCATION IN BOTSWANA: T
No.40 THE IMPACT OF TRAINING DEVELOPING A
D
ON WOMENS MICRO COMPREHENSIVE
ENTERPRISE DEVELOPMENT STRATEGIC RESPONSE.
F Leach, S Abdulla, H Appleton, P Bennell, B Chilisa, K Hyde,
J el-Bushra, N Cardenas, K Kebede, A Makgothi, E Molobe,
V Lewis, S Sitaram (2000) L Mpotokwane (2001)
ISBN: 1 86192 284 1 ISBN: 1 86192 468 2

No.41 THE QUALITY OF LEARNING No.46 EDUCATION FOR ALL:

AND TEACHING IN POLICY AND PLANNING -
DEVELOPING COUNTRIES: LESSONS FROM SRI LANKA.
ASSESSING LITERACY AND A Little (2003)
NUMERACY IN MALAWI AND ISBN: 1 86192 552 0
SRI LANKA.
D Johnson, J Hayter, NOW AVAILABLE CD -ROM
P Broadfoot (2000) containing full texts of Papers 1-42
ISBN: 1 86192 313 9

No.42 LEARNING TO COMPETE:

EDUCATION, TRAINING &
ENTERPRISE IN GHANA,
KENYA & SOUTH AFRICA.
D Afenyadu, K King, S McGrath,

H Oketch, C Rogerson, K Visser (1999)

ISBN: 1 86192 314 7

A

DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 7

Reaching the Poor - The costs of sending children to school

Educational Papers

Other DFID Educational Studies Also Available: A MODEL OF BEST PRACTICE AT LORETO
DAY SCHOOL, SEALDAH, CALCUTTA.
REDRESSING GENDER INEQUALITIES IN T Jessop (1998)
EDUCATION. N Swainson (1995)
LEARNING OPPORTUNITIES FOR ALL.
FACTORS AFFECTING GIRLS ACCESS TO DFID Policy Paper (1999)
SCHOOLING IN NIGER. S Wynd (1995)
N
THE CHALLENGE OF UNIVERSAL
EDUCATION FOR RECONSTRUCTION. PRIMARY EDUCATION.
D Phillips, N Arnhold, J Bekker, N Kersh, DFID Target Strategy Paper (2001)
E McLeish (1996)
CHILDREN OUT OF SCHOOL.
AFRICAN JOURNAL DISTRIBUTION DFID Issues Paper (2001)
PROGRAMME: EVALUATION OF 1994 PILOT
PROJECT. D Rosenberg (1996)

TEACHER JOB SATISFATION IN

DEVELOPING COUNTRIES. R Garrett (1999)

All publications are available free of charge from DFID Education Publications
Despatch, PO Box 190, Sevenoaks, TN14 5EL, or by email from [email protected]

DFID
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 8

Reaching the Poor - The costs of sending children to school

Acknowledgements

This study was conceived during discussion about practical issues identified by two of the T
research team (Andy Brock and John Mace) during implementation of education projects f
internationally. The positive response of DFID to the initial request for financial support q
for the research resulted in an agreement for the comparative study and the long road to E
the completion of the report. We thank DFID for their financial support, but more M
significantly for their moral, intellectual and administrative support in the UK and in the s
study countries. r
w
The research could not have been completed without the study teams cooperation in each r
of the countries. Many people were involved in the implementation of the research: we d
would like to thank them all for their contributions. The team leaders Professor B Baidya, o
Dr J Milimo, Mr M Mahiuddin, Dr M S Haq, Dr Amooti, Dr J Munene and Mr b
Weerackody respectively, are to be commended for their persistence and competence in
bring the country research activities to successful conclusions. In Zambia, changes were
made to the senior research team throughout the conduct of the study, including three
changes of the senior data analysts position. We thank Dr Milimo for working for the
duration of the study, despite the difficulties encountered due to the changes in data
analysts. In Uganda, we thank all of the team members who worked extremely hard to
complete the study in the final year without the guidance of their team leader, Dr Amooti.

We extend our condolences to the research group in Bangladesh whose director Dr S Huq
suffered an untimely death before the start of the work. His unfortunate demise had a
significant impact on the research, halting the process of team selection and training in
mid-flow. Although a new director Mr S.N. Anwar, to whom thanks are due, was
appointed, the team clearly missed the charisma, drive and technical competence of the
person who was to have led their research efforts. The Bangladesh research team suffered
further when the original Team Leader, Dr K S Ahmed, dropped out of the study. His
replacement, Dr T Anwar, unfortunately also died shortly after taking up the post. Special
thanks go to Mr Mahiuddin and Dr M S Haq for completing the study in Bangladesh
under such difficult circumstances.

We would like to note our thanks to the Government officials in each of the countries
without whose permission to conduct the research we could not have proceeded. Their
contributions during the seminars presenting the findings of each of the country studies
strengthened the validity of the research outcomes.

i DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 9

Reaching the Poor - The costs of sending children to school

Acknowledgements

This report has been produced by Siobhan Boyle, CEC (principal researcher responsible
for the majority of in-country support and training activities, the preparation of
quantitative and qualitative research tools, and data analyses), Dr John Mace, Institute of
Education (African research director), Andy Brock, CEC (Asian research director), and
Mo Sibbons, CEC (who ensured mainstreaming of equity issues and provided professional
support throughout). However, the report is based on the data and reports produced by
research teams in six countries. Inevitably, six teams working in very different contexts
will not produce consistently even results and the quality of the data, the analysis and the
reporting will vary. The CEC team have made every effort to ensure consistency in the
data analysis, although this was based on data of differing quality. In consequence, none
of the tables presented in this report should be used to extrapolate for other purposes, nor
be taken out of the context of this specific research remit.

D DFID ii
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 10

iii DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 11

Reaching the Poor - The costs of sending children to school

Contents

Acknowledgments i 4 Methodology and Sample 25

Abbreviations viii Characteristics

Abstract ix 4.1 Selection of Research Countries 25

4.2 Selection of Research Groups 25

1 Report Summary 1 4.3 Selection of Research Sites 26

1.1 Key Outcomes 1 4.4 Wealth Ranking 26

1.2 Summary of Synthesis Report 2 4.5 Selection of Research Households 26

1.2.1 Valuation of Education 2 4.6 Selection of Participatory 27

1.2.2 Quality of Education 3 Approaches Groups (PAGs)

1.2.3 Household Incomes and 3 4.7 Community Triangulation 27

Expenditures 4.8 Feedback to Communities 27

1.2.4 Education Costs 4 4.9 Country Seminars 27

1.2.5 Willingness and Ability to Pay 5 4.10 Methodology 27

1.2.6 Household Sacrifices 5 4.11 Bangladesh 28

1.2.7 Equity Issues 6 4.11.1 Selected Sample, Characteristics 28

1.2.8 Conclusion and Policy Responses 8 4.11.2 Primary Activities of 29

Household Heads

2 Introduction 10 4.11.3 School Attendance 29

2.1 The Research Question/Study 11 4.12 Nepal 30

Objectives 4.12.1 Nepal Selected Sample 31

2.2 Outline of the report 12 Characteristics

4.12.2 Primary Activities of 32

3 Cost Sharing in Education: 13

Household Heads

Concepts and Issues

4.12.3 School Attendance 32

3.1 Introduction 13 4.13 Uganda 33

3.2 The Meaning of Cost Sharing 13 4.13.1 Uganda Selected Sample 35

3.2.1 The Basic Theory 15 Characteristics

3.3 The Ability and Willingness to 16 4.13.2 Primary Activities of 35

Pay for Education Household Heads

3.3.1 Ability to Pay 16 4.13.3 School Attendance 35

3.3.2 Willingness to Pay 17 4.14 Zambia 37

3.4 Equity, Poverty and Gender 19 4.14.1 Zambia Selected Sample 38

3.4.1 Defining Equity 19 Characteristics

3.4.2 Equity and Equality 20 4.14.2 Primary Activities of 39

3.4.3 Poverty 21 Household Heads

3.4.4 Gender 23 4.14.3 School Attendance 39

3.4.5 Other Equity Concerns 24 4.15 Kenya 40

4.16 Sri Lanka 43

D DFID iv

23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 12

Reaching the Poor - The costs of sending children to school

Contents - Continued

5 Discussion of Findings 45 5.6.2 Cost as a Barrier to Education 76

5.1 Introduction 45 5.6.3 Household Direct and Indirect 79

5.2 Value and Benefits of Education 45 Costs of Schooling

5.2.1 Household Perceptions on the 5.6.4 Seasonality of Household 85

Value of Education 45 Incomes and Education Costs

(i) Value of Education 45 5.6.5 The Unexpected Costs 87

(ii) Gendered Perceptions 46 of Schooling

(iii) Childrens Perceptions 48 5.6.6 Monetary vs Non-monetary 88

(iv) Changing Attitudes 48 Payments

5.2.2 Benefits of Education 50 5.6.7 Summary 88
T
5.2.3 Value and Benefits of 54 5.7 The Ability and Willingness to 89

Education vs Participation Pay for Education

5.3 Constraints to Access and 55 5.8 Household Sacrifices 92

Barriers to Education 5.8.1 Introduction 92

5.3.1 The Effect of Poor Quality as a 55 5.8.2 Background 94

Barrier to Education 5.8.3 Household/Individual Benefits 95

5.3.2 Reasons for Never Attending 57 5.8.4 The Findings 95

School 5.8.5 Current Sacrifices Made 96

5.3.3 Other Reasons for Poor 58 for Education

Attendance and Leaving School 5.8.6 Responses to a Drop in the 97

5.3.4 Impact of Social Reproductive 60 Costs of Education

Activities on Education Access 5.8.7 Responses to a Rise in the 99

5.3.5 Summary 61 Costs of Education

5.4 Quality of Education 62 5.8.8 Sacrifices of Household Assets 100

5.4.1 Introduction 62 5.8.9 Other Issues 100

5.4.2 Perceptions of Quality 62 5.9 Equity Issues 102

5.4.3 Quality of Education as a 68 5.9.1 Poverty 102

Determinant of Demand 5.9.2 Local Perceptions of Poverty 103

5.5 Income and Expenditure 70 5.9.3 Poverty, Gender and 104

5.5.1 Expenditures on Education as 71

Government Legislation
a Proportion of Household
5.9.4 Ethnicity 105
Expenditures
5.9.5 Orphans and HIV/AIDS 106
5.5.2 Income and Expenditure by 72 5.9.6 Gender and Access to Education 107

Country 5.9.7 Enrolment Rates 108

5.5.3 Summary 75 5.9.8 Reasons for Fewer Girls 110

5.6 The Costs of Education 76 Attending: Gendered Perceptions 4

5.6.1 Introduction 76 5.9.9 Gender Differences in Outcomes 111

v DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 13

Reaching the Poor - The costs of sending children to school

Contents - Continued

6 6 Policy Issues 116 5.2: In your View Has Education 49

9 6.1 Quality as Well as Quantity 116 Always been Equally Important for
6.2 Diverse Contexts, 117 Boys and Girls (% of respondents) ?
5 Diverse Policies 5.3: Reasons for Changes in Attitudes 49
6.3 Intra Household Diversity 119 (all Respondents)
7 6.4 Local Decision Making, 120 5.4: Would You Change School 56
Accountability and Transparency and why?
8 6.5 Summary 121 5.5: Why Would You Not 56
Change School?
8 Tables 5.6: What are the Problems with 56
9 4.1: Characteristics of Households 28 the School
Sampled: Bangladesh 5.7: Out of School in the Last Month? 58
2 4.2: Main Activities of Respondent 29 5.8: Health and School Fees as 59
2 Household Heads by Gender % Reasons for Non-attendance in the
4 4.3: Primary School Attendance % 29 Last Month (as % of All Reasons
5 4.4: Secondary School Attendance % 30 for Non-attendance)
5 4.5: District Level Indicators 30 5.9: What is the Main Reason for 59
6 4.6: Characteristics of Households 31 Pupils to Leave School?
Sampled: Nepal (Absolute numbers) 5.10: The Effect of Water Collection 60
7 4.7: Main Activities of Respondent 32 on Childrens Education
Household Heads by Gender % 5.11: The Effect of Fuel Collection 61
9 4.8: Primary School Attendance % 32 on Childrens Education
4.9: Secondary School Attendance % 33 5.12: Existence of Good Quality School 63
0 4.10: District Level Indicators 33 5.13: Characteristics of a Good
0 4.11: Characteristics of Households 35 Quality School (%) 65
2 Sampled: Uganda 5.14: Sources of Income (%) 70
2 4.12: Main Activities of Respondent 35 5.15: Household Expenditure 71
3 Household Heads by Gender % Pattern (%)
4 4.13: Primary School Attendance % 36 5.16: Uganda: Income Index 73
4.14: Secondary School Attendance % 36 5.17: Main Barriers to Education 77
5 4.15: Characteristics of Households 38 5.18: Per Student Household 81
6 Sampled: Zambia Spending on Education
7 4.16: Main Activities of Respondent 39 5.19: Spending on Secondary 82
8 Household Heads by Gender % Education as a Factor of
0 4.17: Primary School Attendance % 39 Spending on Primary Education
4.18: Secondary School Attendance 40 5.20: Types of Non-monetary Payments 89
1 5.1: Importance of Primary and 46 made by Children and/or Parents
Secondary Education 5.21: Parental Willingness to 90
(% of Respondents) Work Harder

D DFID vi
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 14

Reaching the Poor - The costs of sending children to school

Contents - Continued

5.22: Spending on Social Functions 94 Figures

A
as a Factor of HH Spending 3.1: Cost Sharing Theory 15

on Education 3.2: Effect of Subsidies 16

5.23: Sacrifices/Responses Made to Meet 94 3.3: Demand for Education 18

Costs of Social Expenditures (%) 4.1: Historical Time Line for 42

5.24: Responses Categorised as 98 Kapsinendet Village

Investment/Consumption 5.1: Household Benefits 52

5.25: % Respondents Identifying

98 5.2: Individual Benefits 53

Basic Needs (Food, Health,

5.3: Existence of Good Quality Schools 64

Clothing, Shelter)
5.4: Proportion of Household 72

5.26: Another Child Can be Sent 98 Expenditure Spent on Education

to School 5.5: Expenditure/Income Patterns 74

5.27: Bangladesh Primary Scholarship 103 5.6: Household Composition of 84

Examination, 1999 Primary and Secondary

5.28: Local Perceptions of Poverty 104 Education Costs

5.29: Gross Enrolment Rates for
108 5.7: Seasonal Calendar Kenya 86
the Six Study Countries,
Kapsinendet village
1990 and 1995
5.8: Seasonal Attendance, Bangladesh 87
5.30: Uganda Net Enrolment Rates 109 5.9: Demand for Education 91

5.31: Currently Attending School: 109 5.10: Drop-out Patterns, Sri Lanka 112

All Respondents to SSI:

% of School Age Children from References 123

Respondent Households

Attending School

5.32: Nepal School Leaving 114

Certificate Results 1995

vii DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 15

Reaching the Poor - The costs of sending children to school

Abbreviations

Abbreviation Meaning Abbreviation Meaning

5 AIDS Acquired Immuno-deficiency MOES Ministry of Education and Science
Syndrome
6 NEP North and East Province
8 ARC AIDS Related Condition NER Net Enrolment Rate
2 ARI AIDS Related Infection
NGO Non-Government Organisation
BBS Bangladesh Bureau of Statistics
NORAD Norwegian Government
CEC Cambridge Education Consultants Development Aid
2
CSO Civil Society Organisation
3 OECD Organisation of Economic
DAC Development Assistance Committee Cooperation and Development
4
2 DEO District Education Officer PA Participatory Appraisal
DFID Department for International PAGs Participatory Approaches Groups
Development
4 PO Poorest Off
EFA Education For All
4 PRS Poverty Reduction Strategy
FGD Focus Group Discussion
PRSP Poverty Reduction Strategy Paper
FH Female Headed
PTA Parent Teacher Association
6 GCE AL General Certificate of Education,
A level
Rs Rupees
SBO Slightly Better Off
GDP Gross Domestic Product
7 SCF Save the Children Fund
GER Gross Enrolment Rate
1 (UK and Northern Ireland)
GOSL Government of Sri Lanka
2 SHH Single Headed Household
GOU Government of Uganda
SSI Semi-Structured Interview
HDI Human Development Index
3 STD Sexually Transmitted Disease
HIV Human Immuno Deficiency Virus
TLM Teaching and Learning Material
HHs Households
ULIE University of London, Institute
HSH Household with Spouse and Head of Education
IDT International Development Targets UN United Nations
IMF International Monetary Fund UNICEF United Nations Childrens Fund
IPRSP Interim Poverty Reduction UPE Universal Primary Education
Strategy Paper
Ush Ugandan Shillings
K Zimbian Kwacha

VDC Village Development Committee

Ksh Kenyan Shillings

WB World Bank
LSMS Living Standard

Measurement Studies
WID Women in Development

MDGs Millennium Development Goals

MH Male Headed

D DFID viii
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 16

Reaching the Poor - The costs of sending children to school

Abstract

This comparative research study focuses on the main barriers to education for the poorest F
households in Bangladesh, Nepal, Sri Lanka, Kenya, Uganda and Zambia. Although the i
study set out primarily to look at the burden of education costs on the poorest households a
very rich data on other barriers to education (e.g. physical access, quality of education, c
vulnerability/poverty, and health,) have been gathered and are discussed. The study looks t
at what motivates parents to send their children to school (and keep them there) through
their perceptions of the quality and value of education. Illuminating views concerning the W
barriers, the quality and value of education from out of school children and children in v
school are also presented. c
s
The study shows that for all groups in our sample (the poorest and slightly better off) the
costs (monetary and non-monetary) of education are a great burden on the households G
and act as a significant barrier to education. There are a plethora of charges associated C
with schooling (direct and indirect). Even where education is nominally free, charges at v
schools are often levied under another name development funds, contributions etc. The i
extent of the burden of costs is obvious when spending on education as a proportion of p
discretionary household expenditure is looked at. In Uganda and Zambia the proportion d
is 33%, Bangladesh 32%, and Nepal 17%. Where optional policies operate e.g. school c
uniform not being mandatory peer pressure or unpublished school policies may add to i
the burdens of costs for poor families. In Uganda and Zambia a substantial part of u
reported costs were for uniforms these are not compulsory, but many schools still insist p
on them. For the poorest households the indirect costs are also considerable, with seasonal g
variations relating to the demand for labour. This seasonal cycle of opportunity costs a
impact on attendance patterns, which in turn influence permanent premature removal a
from school. a
s
One of the clearest threads running through the reports of all study countries is the strong d
sense that the poorest income groups are making very sophisticated choices about s
schooling their children. These choices are based on assessments of the quality of h
education available, value for money, and investment potential. While there is often a m
tendency to dismiss the poorest as either blind followers or recalcitrant laggards, (an often
quoted phrase being the poorest are unaware of the importance of education), our I
study indicates that neither stereotype is appropriate. There is a notable willingness T
amongst the poorest to pay (though ability is often limited or non-existent), and to make c
sacrifices for, what they perceive to be good quality education. t
t
However, the study brings no surprises when it reveals that teachers, parents and pupils all A
have different views on what constitutes quality in schooling. Quality is seen b
predominantly by parents in terms of the availability and competencies of teachers. From a
the perspective of the children, the issue of quality is intimately related to violence in
school for boys, largely corporal punishment, for girls, mainly sexual harassment.

ix DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 17

Reaching the Poor - The costs of sending children to school

Abstract

Frequent examples, especially in the African countries, are given of beatings and
intimidation which affect childrens motivation to attend school. Sexual abuse by teachers
s and fellow pupils sometimes resulting in unwanted pregnancies and forced drop out are
common problems for girls. Whilst parents and teachers played down the seriousness of
these issues, children emphasised them.

While these findings about perceptions of quality may not be new, they do make one point
very clear education policies which seek to address issues of access, attendance and
completion must address the quality of the service being offered if they are to ensure
sustainable demand, not just temporary surges in interest.

Gender inequality also acts as a considerable barrier to education in the countries studied.
Cultural and societal norms determine what a girl or a boy should or should not do at
various ages, and, for both, these norms affect their demand for schooling. In general, it
is the girl who is most affected, and from an early age. The expectation of household
productive engagement (water and fuel collection, younger sibling care, and general
domestic tasks) affect a girls attendance at school and her ability to undertake extra
curricular activities or out of school studies. Later, in those populations where the practice
is still a normal rite of passage, circumcision results in girls dropping from school;
unwanted pregnancies are an unfortunate reason for many girls to cease their schooling,
particularly in the African countries in the study. Early marriage persists as an influence on
l girls schooling, but this has a greater influence in the Subcontinent countries. For boys,
and again variably, productive activities might reduce their demand for schooling,
although this may be seasonal rather than permanent. This fluctuating demand affects
academic performance, which in turn influences whether a boy will continue with his
studies (this is also a factor for girls, but the causes of their academic failure are
g different). Interestingly, in most locations parents claimed not to distinguish between the
sexes of their children in whether they supported their schooling or not. Their children
had a different perception, though, both boys and girls noting that their parents were
more supportive of boys.

Ill-health acts as another barrier to education and causes many children to miss school.
The study shows that in three of the four main study countries health was the reason
children were out of school in the month prior to the fieldwork. The reduction in contact
time further exacerbates a poor record on childrens learning time in school, and reduces
the likelihood of their achieving academically. The effect of HIV/AIDS on schooling in the
African countries studied is also important. The pandemic affects demand, with orphans
being the most unlikely to be schooled, and associated ill-health affects money income
availability to meet the school fees and other school costs.

D DFID x
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 18

Reaching the Poor - The costs of sending children to school

Abstract - Continued

In conclusion, the most significant issues raised in the study are the need for the voices of
the poor and of the child to be heard and responded to, and for flexible systems to exist T
capable of being sensitive to widely differing local needs but, without compromising t
standards of provision. Some policy implications are listed which endorse DFID education f
strategies. The broadening of perspectives on what is quality is of importance. c
Of particular note is the focus on the child: improving quality should ensure that the o
actual learning experience of each child is at the heart of education, there is an a
environment where all children feel safe and valued, and that there is respect for the
fundamental rights of children(DFID Education Target Strategy Paper 2001). Education T
management (at all levels), teacher training (pre- or in-service), and academic supervision a
has to focus on what is the child learning? rather than what is the teacher teaching? b
h
c
b
r
u
r
s
r
c
a
I

xi DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 19

Reaching the Poor - The costs of sending children to school

1 Report Summary

1.1 Key Outcomes Although costs are the key factor

This comparative study of cost sharing in preventing full access, a wide range of
three Asian and three African countries has other socio-cultural factors play important
focused on the question: what impact does roles. Policies which are insensitive to the
cost sharing in school level education have complexity of such decision making are
on the expenditure choices of households likely to fail. Allowing flexibility in policy
among the poorest income groups? implementation at the level closest to the
household offers the best hope of success.
n The study set out to provide an in-depth
analysis of the burden of education costs The poorest are disproportionately
both direct and indirect on the poorest vulnerable to economic and natural
households in some of the poorest shocks (both macro and micro) their
communities. In doing so, the aim has not demand for education can be said to be a
been to generalise or to extrapolate, but vulnerable demand. The death of a
rather to obtain a more comprehensive breadwinner because of AIDS, or sudden
understanding to develop a colour picture changes in the price of staple foods can
rather than a black and white one. The completely extinguish a childs chances
study examines some of the effects of and of education. Special, targeted policy
responses to national and local policies on measures are needed for such groups.
cost sharing in education from those least Poor health too is widespread and
able to afford these services. emerges as a major cause of irregular
In doing so, some key features have emerged: attendance.

Despite their poverty, the poorest Gendered decision-making is a dominant

households are acutely concerned about feature of many household decisions
the quality and relevance of education involving trade-offs between schooling
services. Both the economic and non for different children girls usually suffer.
economic judgements they make about The gender implications of policies to
schooling their children are strongly encourage enrolment and attendance
affected by their perceptions of the must be kept constantly to the fore.
quality of services offered. There is a
notable willingness amongst the poorest In some cases schools are identified as
to pay, or make sacrifices for, what they centres of violence, where children are
perceive to be good quality education. afraid to attend either because they will
Quality is seen predominantly in terms of be beaten or sexually harassed within
the availability and competencies of school or to and from school. It is a
teachers. source of constant wonder that children
and their parents continue to support
schooling with such singularly
unattractive characteristics.

D DFID 1
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 20

Reaching the Poor - The costs of sending children to school

1 Report Summary

1.2 Summary of Synthesis Report time in the valuation of education by parents

This report is a synthesis of six other reports including ever declining valuations (Kenya). O
which themselves covered a total of 20 study Such changing valuations are often associated t
sites. Of necessity therefore, considerable with macroeconomic changes in the demand t
amounts of information have been omitted for graduates and diploma inflation. They g
in an attempt to elicit common and special are also associated with cultural changes in v
characteristics which shed light on the research the independence of family members and the c
question. The report has been organised perceived benefits of education to either the o
under seven subsections, relating to the household or the individual. In such i
detailed study objectives key characteristics circumstances the value of education as a t
of those sections are highlighted below. consumption or investment good may be b
changing even within the individual household, s
1.2.1 Valuation of Education complicating the analysis of motivations. a
In many of the study sites, separate interviews
with parents and teachers revealed high levels The study indicated very clear perceptions W
of misunderstanding and suspicion between by poor parents of education as a vehicle for d
parents and teachers. In all sites amongst all escape from the intergenerational cycle of A
income groups, the study shows very strong poverty providing conditions were right. o
valuation of education by parents. However, For those at the margin even modest a
teachers frequently cited poor parents as investment is not possible. Most powerfully s
uncommitted, ignorant of the benefits of however, was the strong sense which came m
education and an obstacle to getting children through in many interviews and focus groups s
into school rather than an ally. Though there that the poor have very clear views on the S
was sympathy amongst many for the hardships product schooling they are purchasing. i
poor families had to suffer to send children Even if they themselves have no experience p
to school, there was also criticism for what of schooling they judge the quality of the a
was perceived to be lack of fulfilment of service by certain key characteristics and t
parental duties and a willingness to hide their consumption/investment decisions are
behind repetitive excuses. deeply affected by their assessment of the F
quality of the education service they believe t
Parents and children for their part indicated is being offered. This is most keenly v
a sophisticated understanding of the benefits observed in the relative value accorded p
of education but showed no blind attachment primary and secondary education. The price F
no sense of education above all else. of secondary education in most countries is c
Education amongst these groups is very considerably higher than primary and many w
much a dispensable commodity to be reckoned of the poorest households cannot see the s
against more important survival needs such justification for the sacrifices needed to p
as food and health. school children at this level in terms of the p
benefits brought to the household. p
This is not a static situation and a number of a
community discussions indicated changes in t

2 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 21

Reaching the Poor - The costs of sending children to school

1 Report Summary

s 1.2.2 Quality of Education While these findings about perceptions of

. One of the clearest threads running through quality may not be shatteringly new, they do
d the reports of all study countries and sites is make one point very clear education
d the strong sense that the poorest income policies which seek to address issues of access,
groups, as much as the richest, are making attendance and completion must address the
very reasoned judgements about schooling quality of the service being offered if they
e children based on assessments of the quality are to ensure sustainable demand, not just
of education available, value for money, and temporary surges in interest. Moreover, these
investment potential. While there is often a issues need to be addressed at the poorest
tendency to dismiss the poorest as either most remote schools as much as the better
blind followers or recalcitrant laggards, our off schools and at the same time as better off
d, study indicates that neither stereotype is schools. Often there is an implicit or explicit
appropriate. trickle down approach which simply leaves
the poorest schools even further behind.
When it comes to quality of education, the
definitions given by respondents in Participatory 1.2.3 Household Incomes and Expenditures
Approaches Groups (PAGs) and interviews The data on incomes and expenditures is very
overwhelmingly related to the attendance variable in quality and it is difficult to draw
and competence of teachers. In urban areas any but the most qualified implications from
student performance was more frequently it. The data on Household (HH) income is
mentioned perhaps a reflection of greater probably the least reliable because of the
s supply in urban areas (in one community in unwillingness of interviewees to reveal their
Sri Lanka teachers were described as arriving true incomes. A further problem concerns
in envelopes i.e. they were administratively the question of how much income is paid in
posted, but never arrived) in any case, kind, particularly in rural economies where
achievement is also intimately related to the exchange system is not as sophisticated,
teacher competence. monetarised, as in the urban settings. What
is very clear is that in most sites there was a
From the perspective of the children interviewed, difference between the Poorest Off (PO) and
the issue of quality is also intimately related to Slightly Better Off (SBO) in their incomes
violence in school for boys, largely corporal and ability to pay the cost of their childrens
punishment, for girls, mainly sexual harassment. education. For all groups in our sample the
Frequent examples, especially in the African costs of education were a significant burden
countries, are given of beatings and intimidation on the HHs and paying for education would
which affect childrens motivation to attend involve significant sacrifices.
school. Sexual abuse by teachers and fellow
pupils sometimes resulting in unwanted Most households suggested that education
pregnancies and forced drop out are common has the second largest expenditure devoted to
problems for girls. It is notable that parents it after food costs. This emphasis on education
and teachers played down the seriousness of was confirmed in both Semi-Structured
these issues children emphasised them. Interviews (SSIs) and PAGs.

D DFID 3
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 22

Reaching the Poor - The costs of sending children to school

1 Report Summary

When removing non-discretionary items Household size is a significant determinant c

from household expenditure of the small of whether children will have the opportunity m
amount used for discretionary expenditure a to be schooled. In Kenya where the average n
considerable proportion (about one third) of household size in each of the study sites was
this is devoted to education1. 10-14 it is not surprising to find that children F
attend and drop out of school as a result of b
1.2.4 Education Costs other children arriving at school age, or n
For both the poorest and the slightly better according to the season. These choices on f
off groups, the costs of education are the which children to school are also often Z
predominant reason given for children in the gendered (though not always to the e
household never having attended school. disadvantage of girls). In such households w
This finding is even more true of those the opportunity costs of schooling children T
children who have dropped out of schooling. take on a stark meaning with many mouths m
However, for children who had been out of to feed and high costs of schooling. n
school at some time in the month prior to
the interview, health is the main reason. In each of the study sites it was notable that
There are rural and urban differences the there were a plethora of charges associated T
costs are more important in the urban areas with schooling both direct and indirect. As w
where school fees are higher and demanded observed elsewhere this may account for the t
with less flexibility. tendency to under-report the costs of l
education when focusing only on education r
As would be expected SBO households spend as a proportion of household expenditures. p
more on childrens school costs than PO In many places where education is free, d
households. Although we cannot be certain, charges at schools are levied under another W
it is possible that this is due to two reasons name development costs, contributions, t
either SBO households are able to school etc. Where optional policies operate e.g. b
more of their children or they are able to school uniform not being mandatory peer i
afford higher priced services and pay for pressure or unpublished school policies may
supporting materials (including tuition for add to the burdens of costs for poor families. R
example). Education expenditure on girls is In Uganda and Zambia a substantial part of p
higher than on boys at the primary level reported costs were for uniforms these are c
probably reflecting higher costs rather than not compulsory, but many schools still insist a
higher valuation (which tended to be said to on them. T
be equal or lower than for boys). As a factor e
of expenditure at primary level, expenditures Unexpected costs contributions to s
at secondary level are higher for boys in teachers leaving or joining, funerals, repair e
three of the four main study countries. of buildings affected by natural calamity etc. w
also featured as a major difficulty. While the m
poor are willing to make non-monetary c
s
1
See Section 5.8.1 for a definition of discretionary / non-discretionary expenditures.

4 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 23

Reaching the Poor - The costs of sending children to school

1 Report Summary

contributions, in Uganda and Zambia the It would also appear that different groups
y majority of respondents indicated these were exhibit different elasticities of demand, the
not voluntary. shape and slope of the curve differing between
geographical regions or ethnic groups for
n Finally, the times when school costs have to example. A kinked demand curve may be in
be paid and the fact that, generally, they are operation in which a fall in the price of
needed in cash, creates additional hardships schooling creates a disproportionately higher
for many poor families. For example, in demand for education, perhaps driven by the
Zambia peak expenditures on both food and poorest and most marginal groups.
education are experienced in January to March
when incomes are at their lowest in rural areas. Although there is a difficulty in gauging how
These months are also the months when realistic the responses from respondents are
malaria is most prevalent and money is since the questions put were hypothetical
needed for medicines. (especially concerning willingness to pay for
improved services), the case of Uganda does
1.2.5 Willingness and Ability to Pay offer an example of what may happen when
The difficulty of distinguishing between costs are dropped (an example of the kinked
s willingness and ability to pay is well known, demand curve). After the initial surge of
though the two often tend to be confused in enrolment, there has been a falling back
literature, government studies and project perhaps as parents have realised that some
reports. This study seeks to investigate these costs have been shifted rather than abolished
phenomena further and to try to make and also as they make judgements about the
distinctions within the group being researched. benefits of schooling in overcrowded
Willingness to pay was investigated both for classrooms.
the education services currently consumed
by respondents children and for notionally 1.2.6 Household Sacrifices
improved services. Not only does the study establish considerable
willingness to pay for education services,
. Responses showed that there was apparent but, it is also possible to infer that many
price elasticity of demand even for the households are already making considerable
current level of services in most countries sacrifices of discretionary, non-discretionary
and also for notionally improved services. expenditures and capital assets.
There were clear differences, as would be
expected, between responses for primary and Households in all study sites were found to
secondary schooling with the latter more be making regular trade-offs between
elastic than the former. Interestingly, evidence consumption of education and consumption
was also found that this willingness is also of other discretionary expenditures such as
matched by the ability to pay, though with a entertainment or non-urgent medical
caveat: the concept of ability here also covers treatment. These included decisions on the
sacrifice. number of children to school. Such decisions
are often gendered ones, or related to the

D DFID 5
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 24

Reaching the Poor - The costs of sending children to school

1 Report Summary

birth order of children. Considerable numbers depth. The country studies clearly demonstrate d
of households also reported making sacrifices income poverty as a determinant of demand. (
of non-discretionary expenditures such as Nonetheless, other dimensions of poverty r
cooking oil, salt or skipping meals entirely. such as lack of social capital, unheard or a
Yet other households were found to be non-articulated voice, limited empowerment t
selling household capital assets such as and poor health and nutrition make equal a
property and land in order to put children contributions. The considerable intra- and c
through school. inter-country differences in demand for a
education displayed by the communities in i
Yet, despite this, some apparently surprising the study are related to relative poverty c
expenditures account for substantial (broadly defined as above), their ability to s
proportions of household budgets e.g. command support from government or non f
spending on social functions (weddings, government organisations (especially church- t
funerals, circumcision festivals). This is based institutions) during particularly f
actually a useful barometer of the place of difficult periods (such as drought, famine, D
education at this time in the lives of these flood, cyclone, price changes and livestock s
people in these communities. Expenditures loss), culture and influences of historical t
whose economic return may appear limited events or processes. t
in comparison with education will continue h
to be made by even the poorest groups It also is apparent that vulnerability of c
because of the psychic and social returns households on the margins of poverty is an p
achieved. For education to compete it must important feature of many, particularly in p
also match these psychic and social returns. rural areas. Lives controlled by the vagaries a
That cannot be achieved when quality and of the climate and the international global o
relevance are low. economy are in a situation where negative
events can push them towards or into G
The withdrawal of children from school as a poverty which may or may not be permanent. s
response to increased costs or reduced At such times expenditures on education are C
household income remains a common perhaps the most dispensable, and childrens a
strategy response. The study shows that in education is interrupted until better fortune v
the hypothetical case of costs rising by 50% returns, or ceases when other opportunities a
half of all respondents said they would arise or time is a disincentive to return. i
withdraw one or more children from school. Demand for education can perhaps be seen a
In the case of severe shocks death of a as a vulnerable demand, commensurate with r
breadwinner, failure of several harvests the dynamics associated with poverty and c
special strategies are needed to help the most the vulnerable household. d
vulnerable families. s
However, supply-side factors were also c
1.2.7 Equity Issues demonstrated to have a significant effect on L
In the report issues related to the distribution demand. Not only do costs, notably g
of education demand are explored in some opportunity costs, contribute to decrease in p

6 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 25

Reaching the Poor - The costs of sending children to school

1 Report Summary

e demand, but the location of the school many girls to cease their schooling, particularly
. (distance and travel time), various factors in the African countries in the study. Early
related to teaching (class size, the curriculum marriage persists as an influence on girls
and Teaching and Learning Material (TLMs)) and schooling, but this has a greater influence in
t teachers (ability to teach, presence in the school, the Subcontinent countries. For boys, and
attitudes towards the pupils), and the physical again variably, productive activities might
condition of the school are also highly rated reduce their demand for schooling, although
as factors affecting whether a child will enrol this may be seasonal rather than permanent.
in school or not, will attend regularly and This fluctuating demand affects academic
continuously, or will remove themselves from performance, which in turn influences
school permanently or temporarily. Demand whether a boy will continue with his studies
- for secondary schooling was considerably less (this is also a factor for girls, but the causes
- than for primary schooling, and each of the of their academic failure are different).
factors mentioned above had greater relevance. Interestingly, in most locations parents
Direct money costs were higher for secondary claimed not to distinguish between the sexes
schools, and the older the child the higher of their children in whether they supported
the opportunity cost. Secondary schools tend their schooling or not. Their children had a
to be fewer in number than primary and different perception, though, both boys and
hence at a greater distance from most girls noting that their parents were more
children. For girls, as they progress through supportive of boys.
puberty into womanhood, the social
pressures to marry and unplanned pregnancies Ill-health causes many children to miss school.
also increase in significance as a depressor The effect of HIV/AIDS on schooling in
of demand. the African countries is important, although
this came through far more from the
Gender issues are complex, but in all places qualitative assessment methods than from
. significant as an influence on demand. the SSIs, and from informal discussions with
Cultural and societal norms determine what the researchers and the respondents. Despite
a girl or a boy should or should not do at the severity of the pandemic in Uganda,
various ages, and, for both, these norms Zambia and Kenya, it is not a topic for
affect their demand for schooling. In general, common discussion or overt reference. The
it is the girl who is most affected, and from pandemic affects demand, with orphans
an early age. The expectation of household being the most unlikely to be schooled, and
reproductive engagement (water and fuel ill-health affecting available money income
collection, younger sibling care, and general to meet the school fees and other school
domestic tasks) affects a girls attendance at costs. It is also the case that older children
school and her ability to undertake extra may well become heads of household in the
curricular activities or out of school studies. case of the death of parents which affects
Later, but variably, circumcision results in their and their siblings chances of completing
girls dropping from school, and unwanted education. On the supply-side, deaths of
pregnancies are an unfortunate reason for professionals, particularly in the younger

D DFID 7
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 26

Reaching the Poor - The costs of sending children to school

1 Report Summary

age groups, reduces the number of trained the quality of education being delivered, and G
personnel (including teachers) available, and ensuring the safety of children when they c
also adds a further time constraint on teachers attend school, is a pre-requisite if parents i
and officers who necessarily take time off to and children are to demand schooling on a w
attend funerals of their colleagues and families. regular and sustained basis. Raising revenue, b
or seeking non-monetary inputs for schools E
Other than HIV/AIDS, the health of children is possible, even from the poorest households e
in the poorest households is an important but only if what is being provided meets r
factor affecting their attendance in school. their and their childrens needs. Even when s
In any one month (although varying with they acknowledge the poor quality of the t
the seasons) it was noted that many children local school, parents indicate that they are a
are absent for days due to illnesses. In three making sacrifices to school their children. p
of the four study countries health was the Increased commitment to the local school p
reason children were out of school in the clearly seems to be possible if the school (or R
month prior to the study (if they were out of the education system) reciprocates. a
school). This reduction in contact time g
further exacerbates a poor record on childrens Some general policy implications emerge w
learning time in school, and reduces the none of which have easy policy solutions. t
likelihood of their achieving academically. Since the start of this research study the
international development agenda has shifted I
1.2.8 Conclusion and Policy Responses considerably, with a particular emphasis on d
In conclusion, the most significant issues debt relief for the highly indebted countries l
raised are the need for the voices of the poor and greater concentration on poverty W
and of children to be heard and responded reduction strategies. This shift permits the e
to, and for flexible systems to exist capable ready entry of our own conclusions into the r
of being sensitive to widely differing local current development discourse, and equally, t
needs but, without compromising standards our conclusions confirm the necessity for a w
of provision. This implies that an acceptance refocusing of development initiatives. p
of devolution of school management to the r
school and community level with substantial The complexity and dynamic nature of poverty o
autonomy of decision making is desirable, cannot be addressed through education alone.
with clearly described roles and responsibilities While there are some education policy H
for each level of the government education approaches that target resources to the most f
system. Cost-sharing mechanisms are more needy children which might have positive t
likely to be acceptable where the benefits of impacts, perhaps the more significant policy (
making contributions (of whatever kind) are implication is the need for inter-sectoral d
evident, and where transparency is evident. coordination of efforts. The role of Non- s
Government Organisations (NGOs) or Civil e
More effective supervision of schools capable Society Organisations (CSOs) could valuably F
of supporting a continuing improvement to be explored in this context. h
s

8 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 27

Reaching the Poor - The costs of sending children to school

1 Report Summary

d Government responsiveness to exceptional Learning from international experiences,

circumstances needs to be strengthened, systems such as the establishment of School
including the need to identify communities Management Committees/Boards should be
who are particularly vulnerable and likely to introduced or strengthened, and linked into
, be affected during years of especial adversity. a whole school development approach. The
Even though community involvement in responsibilities of such committees/boards
s education to identify needs and manage should extend to management of school
resource distributions may be desirable, this financing, through which appropriate local
should not result in Governments absolving resource raising and distribution may be
themselves of responsibility to provide encouraged. Appropriate training,
additional supports which are designed to supervision and support is needed if any
prevent vulnerable households moving into such approach is to be successful.
poverty. Within the framework of a Poverty
Reduction Strategy Paper (PRSP) or equivalent, This study demonstrates that, even where
and in line with millenium development primary education is ostensibly free, costs are
goals, national policy should be developed a significant factor inhibiting childrens
which responds to the education needs of education, and that demand for education is
the poorest and most disadvantaged. influenced by a complexity of service related
issues as well as direct and indirect costs.
d In order to respond to local needs relevant The main quality issues which dominate
decision-making should be devolved to the responses from parents as to why children
level closest to the household: the school. are not encouraged to maintain school
Whole school development planning can be attendance are related to inadequacies in
encouraged with commensurate training and teacher quality (subject knowledge,
resources. School development plans have pedagogy and attitudes to students), and
the potential to enable schools, in partnership availability of teachers (attendance in school,
with communities (primary stakeholders), to numbers employed and their distribution).
prioritise their development programmes in This leads us to conclude that greater
response to local needs, and allow heterogeneity attention to distribution, training, academic
y of needs to be recognised and responded to. supervision and management of the teaching
e. staff should be a primary concern, if children
Hand in hand with such developments, are to be encouraged to enrol and attend
financial allocations systems need to consider school on a sustained and regular basis, at a
the introduction of weighting mechanisms minimum through to the completion of the
(formula funding) which respond to differential primary cycle.
development needs. Greater resources
should be distributed to those with greater
l educational and socio-economic disadvantage.
Financial accountability both vertically and
horizontally can be assisted through the
school development planning process.

D DFID 9
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 28

Reaching the Poor - The costs of sending children to school

2 Introduction

This comparative research study focuses on The range of issues discussed in the following a
the main barriers to education for the poorest chapters shows that there are great challenges a
households2 in Bangladesh, Nepal, Sri Lanka, ahead if the Millennium Development Goals i
Kenya, Uganda and Zambia. Although the (MDGs) are to be made reality. The MDGs 4
p
study set out primarily to look at the burden pertinent to this study are: o
of education costs on the poorest households b
very rich data on other barriers to education Goal 1 Eradicate extreme poverty a
(e.g. access, poverty, quality of education) and hunger: a
have been gathered and are discussed. The Halve, between 1990 and 2015, the i
genesis of the study was the work being done proportion of people whose income
by various members of Cambridge Education is less than US$1 a day. 2
Consultants (CEC) and the University of
London, Institute of Education (ULIE), on Goal 2 Achieve universal primary T
practical applications of cost sharing initiatives education (UPE):
in education programmes overseas (many Ensure that, by 2015, children
funded by DFID) coupled with an increasing everywhere, boys and girls alike, will
awareness of the paucity of research at that be able to complete a full course of
time3 on household support for cost sharing primary schooling.
in education and the impact of cost sharing A
programmes on poor households. There was Goal 3 Promote gender equality and f
already considerable research on cost sharing empower women: l
which took the school as the unit of analysis Eliminate gender disparity in T
(Penrose 1993, Cumming et al 1995, Igwe primary and secondary education
1988) but, relatively little which focused on preferably by 2005 and in all levels
the household (exceptions include Bray of education no later than 2015.
1996a, b, 1999).
Previous targets set at the 1990 World
This study provides a clear understanding of Conference on Education for All, held in
the problems the poorest households face in Jomtien, Thailand, and reaffirmed at the
sending their children to school the barriers. World Education Forum in Dakar, Senegal,
Although education costs (direct and indirect) in 2000 have not been fully met. Although
were found to be a major barrier for the primary school enrolment has increased and
poorest households, there are other factors the gender gap has narrowed (primary school
5

which come into play (e.g. education quality, enrolment has increased by 82 million pupils C
gender disparity, health problems, access issues). since 1990, with 44 million more girls in 6

school in 1998 than in 1990), these gains A

g
s
2
In each of the countries studied the research teams agreed upon a locally acceptable definition of the household before s
the fieldwork began. w
3
Around the period 1997/98 e
e
4
For the full set of MDGs see the Millennium Report, We the Peoples: The Role of the United Nations in the 21st b
Century, Annan K. 2000 d
c

10 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 29

Reaching the Poor - The costs of sending children to school

2 Introduction

g are modest.5 Over 60 % of all children without cross sectional data on trade-offs made
s access to primary education are girls and within the household; the impact of
s increased access has come at a price. The choices made between spending on health
4
price is quality. Take, for example, the case and/or education and consumption/savings
of Uganda, where the UPE campaign has
been introduced. Although enrolment findings from studies within and across
almost doubled initially, this was found to be each sub-set of the relationship between
at the expense of quality. Uganda is not willingness and ability to pay the private
e isolated in this regard. costs of schooling and: a) household
income and expenditure priorities
2.1 The Research Question/Study (including income foregone); b) physical
Objectives access (distance from school);
The research question, as originally framed, was: c) household decisions on the number
What impact does cost sharing in school and gender of children to be schooled;
level education have on the expenditure and d) community attitudes or peer
l choices of households among the poorest pressure etc.
income groups?6
student and parent perceptions on the
As already mentioned the study has gone quality of education
further since its conception than simply
looking at the household costs of education. the value attached to education by
The issues covered include: students and parents.

some broad measurement of sample Overall, the study has set out to give a textured
household willingness and ability to pay and fine-grained picture of the effects of de
for education services facto and de jure cost sharing policies at the
school level. The study does not attempt to
an in-depth understanding of the schooling answer the question of who should pay for
choices made by sample households when education, nor how it should be paid for:
faced with expenditure alternatives for related questions, but conceptually distinct.7
different education services

l
5
World Education Forum, 2000 The Dakar Framework for Action. Education for All: Meeting Our Collective
s Commitments, UNESCO, France, p.13
6
School level here includes basic and secondary education.
7
Fiscal policy.
Although this study is not concerned with fiscal policy as such, since fiscal policy will almost certainly affect the
governments education cost sharing policies the following brief points should be noted. (For a fuller discussion of these
see Penrose (1998)). One major consideration is whether the resources raised outside the tax system are used as a
substitute for tax finance or to augment it. If it is the former no new resources are available and the question arises as to
whether the system has become more efficient as a result of the new method of raising resources. If the latter, the
efficiency issue still remains and in addition there is the possibility of education quality being enhanced. This itself may
encourage greater contributions from HHs and local communities. Tilak (1991) found a positive relationship in India
between government and HH expenditures on education. This finding must be treated with caution since correlation
does not prove a causal relationship and, as Penrose has found, the relationship may be sui generis and not exist in other
countries.

D DFID 11
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 30

Reaching the Poor - The costs of sending children to school

2 Introduction

The study was conducted in four countries 2.2 Outline of the report 3
(Bangladesh, Nepal, Uganda and Zambia) The report is structured thematically. I
where a SSI questionnaire was administered a
to an average of 250-300 respondents in a Section 3 reviews the meaning of cost sharing a
total of four geographical areas. The and the definitions of poverty and equity s
intention from the outset was to explore in used in the report. e
depth the responses of households, not to t
generalise, and qualitative data was collected Section 4 introduces the methodology of the t
through group and individual interviews to research in the six countries selected. o
triangulate data from the SSIs and other i
sources, and add depth and texture to the Section 5 synthesises the findings from the six i
information gathered. Supplementary data separate country studies drawing them h
was included from a further two countries together under the following themes: f
(Kenya and Sri Lanka). Only qualitative data h
through Participatory Appraisal (PA) Valuation of education
methods was collected in these locations. Constraints and barriers to education 3
(Selection of countries and sites is described Perceptions of quality A
in section 4). In a comparative report of this Household income and expenditure c
nature it is not possible to do justice to the Education costs/opportunity costs d
wealth of data gathered from each of the Willingness and ability to pay
countries involved in the study, nevertheless Household sacrifices
an attempt has been made to present the key Equity issues.
issues from each of the study countries.
Section 6 presents some key policy
implications.

T
f
p
b

12 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 31

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

3.1 Introduction The full opportunity cost should include the

In this section some of the basic concepts totality of what is foregone by undertaking
and issues related to cost sharing in education an activity. Thus, in addition to the costs
g are explained. Firstly, the meaning of cost borne by governments, usually salaries, they
sharing is outlined, followed by a brief should include payment for books, pencils,
explanation of the underlying economic paper etc., i.e. the private costs, and the
theory related to ability to pay and willingness domestic and farming support that children
to pay for education. The main equity issues cannot provide if they are at school, i.e. the
of concern when assessing cost sharing indirect costs. Whether these costs should be
initiative outcomes for access to, involvement included if children are under compulsory
x in and outcomes from education are outlined schooling age is debatable. The whole
here. (In a later section (5.9) the empirical question of how income foregone should be
findings of poverty, ethnicity, gender, and treated is explored in Blaug (1970, p50).
health (including HIV/AIDS) are described.) However, in developing countries, and for
many HHs in developed countries too,
3.2 The Meaning of Cost Sharing implementation of compulsory schooling
At the June 1998 Addis Ababa international legislation is not very effective and it becomes
conference on cost sharing the following appropriate to view attendance at school as a
definition of cost sharing was adopted: choice for HHs. Thus, if there is interest in
Cost sharing includes all officially analysing the effects of cost sharing on HHs
sanctioned contributions made by users to it is relevant to include foregone production
the financing and management of social below school leaving age, for only then will
services. Contribution can be made either the analysis include what are effectively all
by individuals, households, employers or the costs incurred by HHs. It will then also
the community. They can vary from cash be possible to understand how equity and
to contribution in kind, or can be in the efficiency of education provision is being
form of labour and/or participation in affected. (Of course, there may well be
management decisions. Cost sharing, difficulties in quantifying these costs precisely,
however, excludes private out-of-pocket costs but the definition should at least be
that individuals incur in time, travel or other conceptually correct.)
expenses when seeking access to these services.
These costs are nevertheless important to The second problem with the definition is
consider in assessing the impact of cost that in focusing only on the user it fails to
sharing on the poor. (our italics) (P3) take into account other parties who to a
lesser or greater degree share in the costs of
This definition is flawed on two counts. The education. Apart from the users the costs of
first concerns the exclusion of private out of education are borne by the general taxpayer,
pocket costs and the indirect costs incurred firms, NGOs, philanthropists, the local
by households whose children attend school. community and international agencies.

D DFID 13
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 32

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

The extent of the involvement of these other for education resources as they introduce i
parties in cost sharing will have significant universal basic education and attempt to meet t
impact on households.8 the increased demand for secondary education t
that results. At the same time governments w
Cost sharing is an umbrella term embracing discretionary spending is eroded by the m
such terms as cost-recovery, tuition fees, school demands of external and domestic debt, s
levies, user charges and matched funds. The coupled with the difficulties associated with t
conditions under which these various cost the stabilisation and restructuring i
sharing schemes are applied vary enormously programmes which so many are introducing. p
and in consequence so do the effects on Alternative sources of education revenues are w
households. required, and households/users are perceived e
as a possible source of these revenues. The m
Cost sharing schemes have arisen in a variety same sort of problems of budgetary constraints h
of ways and at a number of levels within usually lie behind cost sharing schemes f
national settings. In some cases they have been introduced at the school level, although the e
the result of school or community level problems are on a very much smaller scale, a
initiatives; at other times at the regional and but not necessarily for the HHs who have to m
national level, and at others at the instigation meet them. m
of international agencies or foreign governments. S
Whatever the level at which they have arisen, Efficiency Efficiency is concerned with the r
the reasons for their introduction are usually relationship between inputs and outputs. If p
one or more of the following: fiscal stress, more outputs can be obtained from a given c
efficiency or equity. set of inputs then efficiency is said to have
improved. If costs are introduced into the E
Fiscal stress Education budgets in most picture, efficiency is said to have improved s
countries, particularly developing countries, when the cost per unit of output has been
are being stretched by the increasing demand reduced. Of course, what the outputs are is p
b
r
8
Who should pay for schooling? An important principle of public finance is that he/she who benefits should pay. With
respect to education, a quasi-public good, who the beneficiaries are is a complex issue. One beneficiary is certainly the t
person receiving the education and therefore they should bear a part of the cost. But how large a part? Education is often
thought to provide spillover or external benefits i.e. benefits that accrue to society at large, rather than to the individuals l
who actually receive the education. Some of these benefits may be pecuniary and others not. If society benefits the
benefits principle is clear: society should contribute towards the costs of education, though precisely how much is difficult b
to estimate. A further two difficulties are that some benefits will accrue over time and when should these be paid for and,
a related concern, is that there may also be uncertainty about the extent of benefits. In addition there may also be e
negative benefits, either to society or the individual. One such negative benefit may be that the education received does a
not provide the individual with the skills needed by society and unemployment results.
Paying for education may also discourage certain groups from buying into education, even though they would have s
enjoyed benefits from doing so. This problem of adverse selection could arise through ignorance or an individuals strong
positive time preference they want the benefits now, rather than the future. In deciding how much an individual should
t
pay towards their education all these factors need to be considered by policy makers. o
How should education be paid for and when? One issue is whether payment should be in cash or in kind. For the cash starved
poor HHs paying in cash may be much more difficult than providing the school with labour, materials or farm produce. v
In rural and some urban communities income and expenditure are seasonal and out of kilter with payments required by u
schools. The solution seems simple: adjust school payments to the rhythm of HH incomes. However, if school spending
on materials and other supplies takes place at the beginning of the academic year and this spending is financed by fees the d
school would need to find some alternative source of income if the quality of schooling was not to suffer.
p

14 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 33

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

itself an issue and can raise problems since Gambia and Jamaica, appear to recognise the
t the outputs produced should, ideally, be importance of these arguments and have
n those desired by society. How it is decided introduced into their cost sharing schemes
what societys desired outputs are raises yet special arrangements for those unable to pay.
more questions. However, to return to cost
sharing, one of the interests that there is is 3.2.1 The Basic Theory
the extent to which cost sharing is Mingat and Tan (1986) provide the basic
improving the efficiency of education theoretical framework for cost sharing. Work
. provision. This raises a number of issues in the area since then has essentially been
e which are explored in the case studies. These concerned with elaborating the basic theory
d efficiency issues relate to both the or with empirical testing. The Mingat and
macro/societal level and to the micro/ Tan framework is presented below.
s household level effects of cost sharing. A
Figure 3.1
further efficiency issue concerns the
Cost Sharing Theory
efficiency with which the scheme is
S1
administered: could the fees be collected
more efficiently if different agencies and
methods were used to collect the fees? C2

Should the fees be collected at the school, S2

C1
regional or national level? There is no global
panacea; the answer depends on local C0
Private Demand
circumstances.
No. of Students
E1 E2 E3
Equity It has also been argued that cost
sharing will enhance equity (Mingat and Tan Mingat and Tan show how the supply of
1986, Monk 1990). In the first place pupils student places can be expanded, through the
previously denied adequate education may imposition of user charges. If the initial level
be able to receive it if the additional of user charges is C0 the government can only
revenues raised through cost sharing increase supply E1 places although demand at this
the number of school places available in the price is for E3 places. Thus the supply of
locations in which they are needed - this may places is E1, while the excess demand is the
t be seen as an example of horizontal equity, the difference between E1 and E3. If the
equal treatment of equals. The generation of government imposes a user charge of C1 the
additional revenue may also enable the number of student places can be increased to
support of the poorest members of society E2. At this point demand and supply are in
g
d
through the provision of special discounting equilibrium. This new position is a more
or even fee free schooling - an example of socially efficient position. This is usually the
d
vertical equity, the unequal treatment of case because a better-educated population is
unequals. These arguments are more fully perceived as bringing increased economic
developed in Mingat and Tan (1986). Many and social benefits to both society and the
policy makers, including those in Kenya, the individual. The assumption made here is that

D DFID 15
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 34

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

the government uses the increased revenues 3.3 The Ability and Willingness to t
generated by fees to increase the number of Pay for Education a
school places. However, there may be families The ability to pay for education is often p
who cannot afford to pay the user fees and if conceived to be about the availability of b
fees are a requirement of attendance at school financial resources to the purchasers of t
these students will effectively be debarred from educational services. If the purchaser, usually, m
education. This may be seen as inequitable, but by no means always, the household b
since attendance at school now becomes a (HH), has access to funds that exceed the e
function of the ability to pay the fee rather costs of education then they are able to pay. h
than education being provided as a right. If they do not have access to such funds they t
It may also be inefficient since able students are unable to pay. Willingness to pay is a
who could potentially make a valuable concerned with the attitude of the purchaser i
contribution to society and the economy are towards education and whether or not they h
now denied education. are prepared to pay for the educational a
service. Both terms require some elaboration. a
Figure 3.2
s
Effect of Subsidies
3.3.1 Ability to Pay e
As defined above, to discover whether a HH d
had the ability to pay would merely require t
S3
knowing the costs of schooling and relating r
S2 them to HH income; if income exceeded t
C1
costs they would be deemed able to pay. c
C0 However, a number of related issues have to e
Demand
No. of Students be considered about this view of ability to pay. o
E1 E4 E2 E3 One of these concerns the way in which HH
income available to pay for education is 3
Figure 3.2 shows how providing support for measured. The total income of the family is P
poorer students, for example through bursaries, quite different from income that may be d
can affect the total supply of student places. available for education spending. Perhaps the b
If the government imposes a cost C1 on most important distinction is between total c
students the number of places that would be income and discretionary income. Although G
demanded would be E2. However, at this price in principle all HH income could be available t
some of the poorer students would not be able for spending on education, in practice only f
to attend. In order to subsidise these students that income that is available after essential (
revenue generated from those students who purchases of such items as food, clothing t
could afford to pay is redistributed to support and housing have been met is/will money o
the poor students. This results in the total be available for education. r
number of places falling to E4, though now W
subsidised poor students fill some of these Of course, to make this point is to raise the s
places. question of what is essential. For example, i

16 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 35

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

there are minimum standards of food, clothing the educational infrastructure, hiring better
and housing necessary to preserve life, but teachers and building more local schools,
perceptions of this minimum will vary user fees might encourage school attendance.
between households, within households and If, as happened in one of the schools visited
through time. A further issue concerns the in Zambia, a gift of money to the headteacher
manner of payment. Does payment have to from the consultants for the purchase of
be made in money or can it be paid in kind, teaching materials was instead used to purchase
either wholly or in part? Does the payment alcohol. Subsequently, the headteacher got
have to be made at particular times, such as drunk. In such circumstances it is hardly
y the beginning of the school year, or is it surprising that parents (and visiting
allowed to vary? For rural communities this consultants, for that matter) are disinclined
r is particularly important as the time of to donate money to the school.
harvests, when there may be a greater
availability of discretionary money, is seasonal As is made clear in the discussion of pricing
. and usually has no relationship to the policy, the attitude to fees depends on families
schooling cycle and the incidence of perceptions of their marginal welfare and its
education costs. (The nature of these costs, relationship to marginal cost. If marginal
H direct and indirect, is discussed elsewhere in welfare is greater than the increment to costs,
the report.) Much of this discussion becomes mu>mc, they will be prepared to pay the
redundant if the costs of schooling are user fee. If mu<mc they will not. Although
themselves discretionary. This would be the demand curves are usually depicted as fixed
case if, for example, poor parents were lines, mu is fixed, this is certainly not the
exempted from school fees and the wearing case in the real world and as attitudes to
. of uniforms was optional. education change so the curves will shift.
H (Although the economic terminology may
3.3.2 Willingness to Pay be a little unfamiliar, there is overwhelming
Parents may be able to pay, but unwilling to evidence that this analytical framework is a
do so. At the extreme they may avoid payment useful one for understanding individual and
e by the simple expedient of not sending their HH behaviour (Monk, 1990, Lipsey 1963).)
child or children to school. As Gertler and
h Glewwe, 1989, make clear, and it remains A major factor in determining the shape and
e the case still, any debate about user fees position of the demand curve, the marginal
focuses on the price elasticity of demand utility function, is families perceptions of the
(though policy makers do not always use this quality of schooling and how schooling may
term) but the ultimate effect of school fees affect their future opportunities. For poor
on enrolment depends on how the funds families in developing countries education is
raised from these fees are used (p3). the major route to economic advancement
Without totally accepting this claim it seems for their children. As our evidence shows
sensible to recognise that if fees are used to human capital theory seems to provide a
improve the quality of teaching by improving very plausible explanation of the behaviour

D DFID 17
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 36

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

of families: the education of children, with reason, but the advantage of gaining new t
all the sacrifices involved, is a worthwhile knowledge is also that it enables students to d
choice since it is perceived as giving children get better jobs, an investment reason. s
the opportunity to enter highly paid jobs in i
the exchange sector of the economy and, in The difficulties of untangling the consumption s
addition to enhancing the childrens economic and investment reasons have been the subject i
welfare, enhance the economic wellbeing of of acrimonious debate since Schultz first a
their families through the repatriation of funds developed the idea of human capital in the i
to their parents and siblings. sixties, but our evidence does indicate that t
poor families consider education as the route s
Some insight into the demand function for that their children should take to enjoy greater a
education is provided through the quotes economic opportunities and escape from the p
from the case studies. For example, a female deprived existence that they experience c
respondent from Chinsali stated themselves they see education, at least in i
part, as an investment. Whether the demand r
You know our parents (ancestors) long time for education is for investment or consumption m
ago used to marry polygamously in order to reasons it can be depicted diagrammatically i
bear many children who would look after as in Figure 3.3 below. m
them in their old age. The function of c
Figure 3.3
polygamy has been taken over by education s
Demand for Education
d
Clearly, this parent saw education as an t
investment and this view seemed to be shared
by others as shown in the PAGs. In Chinsali P1 G
female parents saw education as providing s
P*
the individual with the opportunity to live t
P2 D2
in a big house, to be paid monthly, to a
Private demand
look after parents, all investment type for education
perspectives. But there were also other benefits, Quantity f
q1 q* q2
such as being independent, securing h
family happiness, read and write for us, In the diagram education is treated as a t
assist in health care which indicate that the homogeneous product to facilitate exposition e
investment was also for the benefit of the of the concept of elasticity of demand. In the c
family and community. This qualitative real world there would be a range of differently t
evidence is supported by the quantitative shaped demand curves reflecting households t
analysis which shows that the reasons for different perceptions of the value of the t
attending school were often economic, as different schooling opportunities available. g
above. The main reason given for attendance The diagram shows how much education o
at school was to gain new knowledge families are prepared to buy at different prices, b
(32%). This suggests that education is seen with the responsiveness of demand to changes c
as a good thing in itself, a consumption in price being determined by the slope of c

18 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 37

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

the demand curve, the price elasticity of if they did have to walk considerable distances,
demand. Although in the diagram the curve they would arrive at school tired with adverse
slopes down, i.e. as price rises less schooling effects for their ability to concentrate on
is demanded, this might be an over lessons at school.
n simplification. The effect on enrolments will
t in part depend on how the revenues raised Section 5.7 describes the evidence produced
are used. For example, if the fees raised by the study on the sample populations ability
improve the quality of schooling or reduce and willingness to pay for education.
the distance that children have to travel to
e school they may actually increase school 3.4 Equity, Poverty and Gender
r attendance. If the amount families are
e prepared to pay is greater than the marginal 3.4.1 Defining Equity
cost of the improvement user fees can be Gender equity as a concept is now well
increased, for in such cases social welfare is known and a commonly cited focus of
d raised. If willingness to pay were below international development aid projects and
n marginal cost attendance at school would fall programmes. Several web sites are devoted
if fees were increased. In this situation there to this theme, for example the Organisation
may be a case for government subsidy to of Economic Cooperation and Development
cover the difference and thus to increase (OECD) Women in Development (WID)
social welfare. The willingness to pay will collection of papers to be found through
differ according to the attitudes of families www.oecd.org. Because of this common
towards education and their ability to pay. association of the term equity with gender it
is frequently interpreted with this narrow
Gertler and Glewwe (1989) demonstrate focus. However, inequity associated with
some of the points made above. They divide gender is only one type of inequity. Others
the population in rural Peru into quartiles include inequities consequent on poverty; on
and show that secondary school enrolment is location; and resultant from disability or
15% higher for the wealthiest quartile than belonging to minority ethnic groups.
for the poorest. In the same paper they show Although the principles may remain the
how dramatically enrolments are affected by same, the policies, programmes, strategies
travel time to the nearest school, with and monitoring tools needed to promote
n enrolment rates being more than double for gender equity will differ from those needed
e children within one hours travel time from to rebalance disparities associated with other
y the school compared to those who have to forms of inequity.
travel for more than four hours. Presumably
the distance from school would have an even Distribution of resources within households
greater effect for primary enrolments for not as well as between households is of concern.
only would there be greater danger of In the case of education this concern manifests
, bullying or attack from animals for smaller itself most particularly as gendered access to
s children, they would also not have the physical schooling. Age is a significant factor in
capacity to walk considerable distances. And, household resource decision-making. In this

D DFID 19
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 38

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

study and elsewhere, when fertility rates are create learning opportunities for all children I
high it is often the younger siblings that lose especially those with particular needs associated m
out to their older brothers (usually) and with context, socio-economic circumstances, i
sisters (sometimes). The elderly, especially and those with special physical or cognitive t
widows, are frequently marginalised in needs. Some of these skills can be learned o
respect to access to household resources, given the right inputs and training opportunities, d
although elders opinions may be decisive. but there are additional aptitudes related to g
the personality of the individual which are y
This section starts by elaborating the term over and above learnt ability. s
equity: why equity and not equality? The o
following discussion on equity and equality More practically, the term equity allows us n
explores these concepts in relation to to think in terms of working towards the t
education. condition of being equal in the education a
system. Equity issues refer to matters of t
3.4.2 Equity and Equality disparity matters related to more or less; I
The terms equity and equality are often advantaged and disadvantaged; accessible and p
confused and although equity and equality inaccessible; majority and minority; abled and b
are similar in meaning, there are also disabled. Childrens access to schools and b
important differences. It is our opinion that their achievements in the classroom are likely s
the most appropriate term for strategy to be affected by poverty, disability, ethnic d
development purposes in a situation of minority and gender. The capacity of the m
considerable inequality is in fact equity. Let school to provide access to and quality of e
us examine why. education will be affected by the schools i
location and condition, and by the availability, d
If Education For All (EFA) is to be more attitudes and education of the teachers. r
than an international slogan, all children s
should have equal access to school and be Perhaps it could be summarised by saying r
provided with equal opportunities to learn. that all children should be equally able to go D
Similarly, to achieve the millenium to school and they should all have an equal G
development goal of UPE an education quality of schooling. However, this does not
system should be equally available and mean that they should have an equal and T
suitable for the learning needs of all children. identical delivery and content in their n
However in life, as in nature, nothing is schooling. Nor is there an expectation that m
equal. Children as users of schools all live in all students are able to have an equal a
different circumstances; they have different outcome. There will be individual or b
needs; they have different learning capacities. community preferences in the choices they s
Schools as providers of education cannot make in their education; there will be
provide all children with equal access and individual differences in academic ability. An 9

d
teachers are not all gifted with an equal equitable system needs to be able to respond r
a
capacity to teach and the skills needed to to these preferences and to individual needs. w
r
c

20 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 39

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

It is important to recognise that in order to materials or uniforms to girls; or lowering

d move towards equality of outcomes, or entry qualifications to the teaching profession
, indeed equality of opportunities and process, for women in order to increase the numbers
there are often circumstances where inequality of women teachers. Similar examples would
of inputs is necessary. For example, in many also be relevant for those from the poorest
s, developing country contexts the number of households; or for children with disabilities
girls who attend school, survive through five who require additional support in order to
years or more of primary education and get equal opportunities. Marginalised
succeed in being promoted to junior secondary communities outside the mainstream of
or secondary schools is far fewer than the society may need additional incentives to
number of boys. There are many reasons for encourage their entry in mainstream schools,
this gender inequality, some to do with or the mainstream schools may need to
aspects of the schooling provided and some consider ways of adjusting their service delivery
to do with the cultural and social context. in such a way that they remove barriers to
In many of these circumstances simply access of these minority children. For these
d providing an equal amount of schooling for numerous reasons, the term equity is preferred
d both boys and girls will not ensure that they to equality, implying differences in inputs
both equally have access to schools: the and sometimes processes in order that equality
y status quo will remain. Because of current of opportunities to achieve can occur.
disparities, equal treatment of women and
men is insufficient as a strategy for gender This section looks at issues related to poverty,
equality. Equal treatment in the context of gender and equity in relation to cost sharing.
inequalities can mean the perpetuation of
y, disparities. Achieving gender equality will 3.4.3 Poverty
require changes in institutional practices and Universal Primary Education, Education for
social relations through which disparities are All, and the International Development
reinforced and sustained (OECD Targets (IDTs) emphasise the need to
o Development Assistance Committee (DAC) include all children, irrespective of gender,
Gender Guidelines, 1999). ethnicity, religion or disability, in effective
t schooling and at a minimum in completing a
This explains why unequal inputs are often primary cycle of schooling9 . Schooling, or
needed to redress current imbalances and education, is both important in its own right
move towards a more equal opportunity to (and is a fundamental human right), and as a
achieve. Examples might be removing cost determinant of poverty: Inadequate
barriers to girls entry in school: providing education is one of the most powerful
scholarships for girls, giving education determinants of poverty and unequal access

9
It is usual to find the word education used rather than schooling in references to the IDTs. This may be misleading and
denies other forms of education. For example, the importance of children learning traditional skills from their elders, the
d relevance of learning local cultural practices, the significance of understanding social norms are as important for a childs
ability to get on in their own environment as formal learning of the school curriculum. Despite feeling uncomfortable
with the imprecise use of the term education, in this paper the terms schooling and education are used interchangeably, in
recognition of the generally accepted interpretation and use of the latter, and in spite of the fact that the study is specifically
concerned with formal schooling.

D DFID 21
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 40

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

to educational opportunity is a strong correlate of deprivation (using income or consumption o

income inequality (World Bank, 2001). measures) but also low achievement in M
A countrys efforts to reduce poverty might education and health. These latter are of f
be seriously compromised if current rates of concern in their own right, but when they p
education are not expanded. A large body of accompany material deprivation they are of b
research points to the catalytic role of basic particular concern. Vulnerability and exposure t
education for those individuals in society who are to risk, voicelessness and powerlessness are H
most likely to be poor girls, ethnic minorities, features of poverty accepted by the World p
orphans, people with disabilities and people living Bank and provide a far broader definition of i
in rural areas (op cit). poverty than that espoused in the earlier p
1990 World Development Report. The free t
Education is seen as a catalyst for poverty market is not seen as a perfectly functioning (
reduction, then. However, what is also entity, but one whose effects are distorted or s
significant, and a key motivator for this constrained by social, cultural and s
research, is that poverty is a determinant of institutional factors. The multi-dimensional D
access to schooling. A circular argument aspects of poverty require a broad range of s
emerges when it is recognised that in order policies. The interactions of different poverty p
for poverty to be addressed, amongst other related factors reinforce one another, creating o
things, education needs to be available and cycles of deprivation which require multiple t
effective, but that for education to be available interventions. Uni-directional reliance on d
certain features associated with poverty increased Gross Domestic Product (GDP) to t
which prevent access need to be addressed. improve the well-being of the poorest has s
The concern of this study is to determine at long since been abandoned in favour of t
what level, if any and for whom, cost sharing integrated social policies targeted towards t
would be acceptable. Who are the children the most disadvantaged or deprived sections u
who do not get to school, or who do not of the population. a
remain once they do enter the primary cycle? i
Will cost sharing if used to improve the This rather simplistic statement of a truism
quality of schools actually encourage them, long recognised by social policy or socio T
and their parents to send them to school, or economic analysts10 should not lead to d
will it/does it provide a cost-barrier? assumptions that once recognised simple d
solutions may be introduced. r
The World Bank defines poverty as a V
pronounced deprivation in well-being The current emphasis on the production of a
(World Bank 2001). It goes on to state that PRSPs by the poorest (heavily indebted) e
poverty encompasses not only material countries d
a
10
See for example the many works of Amartya Sen, writing in this way about poverty as early as 1976. His entitlements r
theory conceptualised out of his analysis of periods of famine, particularly the 1958 Bengal famine, and later a capabilities
approach, demonstrate the multidimensionality of deprivation. The significance of institutional, legal, political and social c
factors in the functioning of markets effectively distorts market transactions in favour of those who have entitlements and
capabilities, and further disadvantage those who do not enjoy those same rights. The widening of the gaps in income c
distribution which have accompanied the introduction of free market ideology to the transitional economies provides a
distressing example of the consequences of removing entitlements, and reducing capabilities. n

22 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 41

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

of the world as the passport to International gaps in educational access were between the
Monetary Fund (IMF) and World Bank highest socio-economic strata and the
funds indicates the urgency and validity lowest. It is considered that there are
placed by these multilateral institutions on sufficient research studies demonstrating the
broadening government policies to include gaps in educational attainment between the
e the voice of prioritised needs of the poor. rich and the poor in all of the countries
However, the difficulties faced in translating studied to make this unnecessary. This study
policy statements into practical and puts more emphasis on understanding the
f implemented actions which favour the mechanisms that prevent the poorest from
poorest, and thus removing the barriers to accessing education using qualitative
their full participation in economic success research methods (the hypothesis being that
(and increased access to improved social many of these constraints would be income
r service provision, including schooling), are or price related mediated through localised
still to be solved. A recent analysis for social, cultural and historical factors), and
DFID of sub-Saharan Africas PRSPs from a analysing the cost (price) constraints and
social policy and sustainable livelihoods peoples ability and willingness to pay for
y perspective (Thin, 2001) found that analysis education through quantitative analysis.
g of poverty was disaggregated
to identify who and where the most 3.4.4 Gender
disadvantaged were. The pro-poor strategies The economic benefits of education are by
o themselves, however, were not developed in now well established. Education improves
such a disaggregated way. They recommend the quality of life. It promotes health, expands
that Future PRS processes should be encouraged access to paid employment, increases
to show how the interventions proposed are productivity in market and non-market work,
underpinned by information on poverty and by and facilitates social and political participation
analysis of opportunities for specific kinds of (Bellew and King, 1993, p285, cited in Leo-
improvements among specific categories of people. Rhynie and the Institute of Development
and Labour Law, 1999).
The significance of this concern with
disaggregation of analysis and the means to Over and above the economic benefits that
disaggregate policy implementation is are evident from widespread education, the
reinforced by the outcomes of this research. education of women provides considerable
Very clearly, each of the countries demonstrates additional benefits to society. However, this
a significant heterogeneity of social, seems to be a stronger effect when schooling
economic and cultural structures and is widely spread amongst the population,
distributions, both between the countries and when women have had more than three
and within them. The locations for the years of schooling. The social externalities
research in each of the countries were created when overall levels of education are

chosen to represent the more deprived above a certain threshold add to the
communities. Given the scope of the study economic case for ensuring universal primary
no attempt was made to assess how wide the education (Colclough, 1993, p31).

D DFID 23
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 42

Reaching the Poor - The costs of sending children to school

3 Cost Sharing in Education: Concepts and Issues

Educated women have smaller families, fewer 3.4.5 Other Equity Concerns 4
of their children die in infancy, and the Two other sets of equity issues are included I
children who survive are healthier and better in the description of empirical findings in p
educated. Moreover, educated women are section 5.9. Ethnicity (5.9.4) and the f
better prepared to enter the paid labour problems associated with being an orphan t
force, which is critical to the welfare of the (5.9.5) are described. In some countries i
many female-headed households in belonging to minority ethnic groups has a
developing countries (Bellew and King, significant bearing on entitlements. This
1993, p285, cited in Leo-Rhynie, 1999). impacts directly and indirectly on a childs
ability to access schooling. Higher levels of
This view of education carries significant poverty are often experienced by
weight in international funding agency marginalised minority groups, or people not 2
environments, and the importance attached belonging to the politically dominant ethnic
to the economic and social benefits of group, reducing potential participation in
education can be seen in the UPE and IDTs. school. The social exclusion of some 3
Beyond this instrumentalist perspective, minority groups persists, despite legislation,
there is also a strong case to be made for in some parts of the world.
universal education from a human rights 4
perspective, i.e. that basic education is a Being an orphan as a result of deaths of
right, something that all people should be parents from HIV/AIDS or as a
entitled to. Universal clearly indicates that consequence of war has a significant impact
all children irrespective of gender, ethnicity, on a young childs life. Social upheaval, 5
disability or poverty should have an equal relocation, economic deprivation and
opportunity to access quality education. psychological distress are variously implicated
Above some of the links of lowered access to in disruption or abandonment of schooling.
education to the processes and mechanisms Whether it is likely that the expectation of a 6
of poverty are made, and the effects of child or their new guardian paying direct or
[additional] costs of education on initial indirect costs for education in such
entry or fluctuations in access as supported circumstances will be met is explored.
by the cost sharing study outcomes are shown. T

T
s
i
h
d
t

24 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 43

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

4.1 Selection of Research Countries would have made comparison impossible;

In the original proposal to DFID it was choosing only Africa or Asia would limit the
proposed that two countries be selected generalisability of findings and miss an
from Africa and two from Asia, the selection opportunity for comparison with only small
to be based on a number of criteria savings in the effort required.
including the following;
Following consultations, including a review
1) Overall mixture of countries exhibits of the budget, and peer advice during a
markedly different patterns of cost sharing preparatory seminar in 1999, the countries
allowing wide range of comparisons chosen were Bangladesh and Nepal in Asia
and Uganda and Zambia in Africa.
2) Country has not been extensively studied Subsequently, Kenya and Sri Lanka were
before with regard to cost sharing included not as full study countries but as
supplementary studies with a purely
3) Government approval for the study likely qualitative focus.
to be easy to obtain
4.2 Selection of Research Groups
4) Reasonably good background data One of the aims of the research was to build
available from government / international capacity among local research groups for
and national agencies undertaking studies of this nature. The
selection of organisations to undertake the
5) Reasonably strong infrastructure including country studies was done through a competitive
Ministry set up and potential researchers / tendering process. The research teams selected
d research institutions were diligent and conscientious, but
unfortunately, some of the researchers proved
6) A DFID field office is established in to have rather less experience of cost studies
country, or they had active DFID and/or the use of participatory appraisal
education programme. methods than was evident from their
submissions. The limited level of support
The countries suggested for the long-list were: from international consultants built into the
study proved to be less than desirable in some
Asia: Bangladesh, Nepal, Pakistan, Sri Lanka. instances: the international researchers were
not in all cases able to provide the level of
Africa: Ethiopia, Kenya, Uganda, Zambia. additional training and support the research
teams needed. Changes in research team
The rationale for selecting at least two case personnel and problems with communications
studies in both Africa and Asia was to contributed to slow progress in some of the
illustrate the similarities and differences in study sites. In two of the six countries the
household decision making in countries with final reports were not as comprehensive as
different endowments and different approaches expected: the local teams would have benefited
to educational provision. Less than two cases from additional time and support to reach

D DFID 25
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 44

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

the same quality standards as the others. decided by the participants themselves. From B
Nonetheless, our aim of building local this ranking the households were chosen on
capacity was addressed, and skills have been a random basis11. 4
enhanced for further studies of this nature.
Additional external support would still be 4.5 Selection of Research Households I
needed, especially at the analysis stage. In each of the four selected sites o
approximately 75 semi-structured interview u
4.3 Selection of Research Sites (SSI) questionnaires were administered. g
Guidance was given to the study teams on In order to have some comparison, the i
site selection. Given the in-depth nature of selected households were divided through g
the study purposive sampling was used, while the wealth ranking process into the poorest o
trying to achieve a balance between urban off (PO) and the slightly better off (SBO). f
and rural sites. Study teams were advised to Approximately 75% of questionnaires were
select three rural sites and one urban site administered to the first group and 25% to 4
using a range of available sources on poverty the second. I
and education standards. For example, in o
Nepal the 75 districts were classified according Within these groups a further division was to t
to their Human Development Index (HDI) administer questionnaires separately to both w
and the lowest ones selected (allowing for the household head and spouse (where o
access and research team security). In Zambia available). This was done not as a validation c
selection was made using Central Statistical of the information given (though that was a h
Office classifications of poverty areas by-product) but in order to gauge i
triangulated with a 1999 Comparative Poverty differences between the head and spouse in
Study by the Poverty Advisory Group. attitudes to the education of children and 4
especially to see if there were gendered I
Following general site selection, the assistance differences. Thus in total (and given that in r
of provincial/regional/district authorities some sites there were a large number of p
was used to identify possible communities/ single headed households) about 35-45 t
villages for study. Communities were households were sampled in each site. p
approached about their participation in the The definition of whether a child12 was of
study and final sites were selected accordingly. school age was set by each research team, 4
taking into account the age at which I
4.4 Wealth Ranking children were supposed (by law or custom) s
The first activity after explaining the purposes to go to school, the fact that some children o
of the study to leaders within the study area start school earlier than the law requires, and f
was to do a wealth ranking exercise using a the prevalence of repetition, leading to over t
representative sample of key informants and age children studying in school. In Uganda i
definitions of poor, relatively poor etc. and Zambia the ages selected were 5-25, in N

11
In the case of Nepal the households were chosen randomly but provided they contained at least one school-going child.
This was a misinterpretation of the guidelines that required households with at least one child of school-going age.
1
12
Although the term child is used we appreciate that the age ranges used include young adults.

26 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 45

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

m Bangladesh and Nepal 5-1913. 4.10 Methodology

The study was deliberately designed to focus
4.6 Selection of Participatory on the household and to try to complement
Approaches Groups (PAGs) quantitative data from SSIs with qualitative
In each study site triangulation and enrichment data collected from individual interviews and
of the data collected in the SSIs were PAGs. Secondary sources of data such as
undertaken by the study teams through Living Standard Measurement Studies (LSMS)
group sessions. These groups were reports, project reports, sector studies etc.
interviewed separately and comprised: were used as additional sources of data for
groups of boys and girls both in and out context and background.
t of school; primary and secondary teachers;
fathers and mothers, and community leaders. The difficulties of researching into
household costs are well known to those
4.7 Community Triangulation with experience of the field (e.g. reluctance
In Kenya and Sri Lanka, where studies relied and/or inability of respondents to give true
on PA methods and SSIs were not administered, income figures, tendencies to forget certain
o the researchers held a one-day workshop members of the household, difficulties in
with a nearby community to test the validity quantifying expenditures). Our research
of their findings this was very instructive in methods tried to compensate for some of
confirming/refuting certain conclusions and the known weaknesses through participatory
highlighting differences between communities group discussions and the use of techniques
in the same geographic area. that did not require literacy including pie
(chapati) diagrams, Venn diagrams, matrix
4.8 Feedback to Communities ranking, timelines, and seasonal calendars.
In most of the study sites feedback of the The notes of the group sessions were written
results of the country studies to the up in the field and charts and diagrams
participants was given after production of recorded or brought back to the research
the draft report. In cases where this was not headquarters.
possible it was due to security issues or time.
This was successful in involving a wide range
4.9 Country Seminars of informants and allowing researchers to
In the four full-study countries a national deepen their understanding and refine their
seminar was held to disseminate the findings probing questions for the SSIs. It was also
of the drafts reports and seek comments successful in deepening understanding of the
d from other professionals with an interest in burdens on poor households in the context
the field. The participants of these seminars of the communities they live in. It was more
included Ministry Officials, Agency and difficult to record this wealth of information
NGO representatives, and other stakeholders. and summarise it in a form which was easily
transmissible and usable in the synthesis report.
.
13
The official enrolment age for primary school in Uganda, Bangladesh and Nepal is 6 years. In Zambia it is 7 years.

D DFID 27
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 46

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

Sample Characteristics Rural Sites Mymensingh, Patuakhali, Dinajpur 4

The sections below summarise some of the Three areas in the districts of Mymensingh,
salient features of the selected communities. Dinajpur and Patuakhali were selected to cover T
The tables presenting primary and secondary some of the major ethnic groups living in t
attendance patterns for each country counter Bangladesh. In Haluaghat upazila (formerly /
-intuitively show that the proportion of children named a thana and a police station) of m
who have never attended secondary school is Mymensingh district the Garo tribe are included h
lower than those who have never attended in the sample. Ghoraghat upazila of Dinajpur m
primary school. This could be explained by district in the plain mainland includes the
the variation between the official enrolment ethnic group known as Saotal. In the past
age in each country and the ages which the they used to live on hunting of wild animals
T
research teams included in their primary and but nowadays they have started limited
secondary samples. These samples included cultivation. They are poor relative to the rest
students who were officially too young of the mainland population. In Kalapara
(below grade one age) and too old (beyond upazila in the coastal district of Patuakhali,
upper secondary grade age) to attend school. there is a tribe belonging to a different
ethnic group known as Rakhain. They are
4.11 Bangladesh poor, and engaged in agriculture and fishing.

Urban Site Mohammadpur Thana (Ward 12) 4.11.1 Selected Sample, Characteristics S
This is an urban thana of the Dhaka District The sample of households in the four study
Corporation. It is the seventh largest thana sites is shown in Table 4.1. There were 189
in Dhaka in terms of population (316,203 in households and 300 respondents altogether, T
1991) and the households of the ward 10% of whom were from single headed
studied are almost all Muslim14 . households. In PAGs a total of 368 persons
were involved including students in and out
of school, parents, teachers and community
leaders.

Table 4.1: Characteristics of Households Sampled: Bangladesh

Poorest Slightly Better Off

HSH SHH HSH SHH
MH FH MH FH MH FH MH FH Total
Mymensing 22 3 5 4 7 1 1 43
Dhaka 24 7 4 6 2 1 44
Patuakhali 21 3 6 10 40
Dinajpur 24 3 8 2 22 3 62
Total 91 16 5 14 31 4 23 5 189
Source: Bangladesh Data. Key: SHH = Single Headed Household; HSH = Household with Spouse and Head; S
FH = Female Headed; MH = Male Headed K
14
The 1991 Bangladesh Bureau of Statistics (BBS) Census Report shows that 97.99% of the population is Muslim S
(Bangladesh Report, p4-4).

28 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 47

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

r 4.11.2 Primary Activities of 4.11.3 School Attendance

Household Heads The attendance of the children in the
r The primary activities of household heads in sampled households is shown in Tables 4.3
the households surveyed were in agriculture and 4.4. Girls attendance appears to be
/animal husbandry and production work for better than boys at both primary and
men, as is shown in Table 4.2. Female secondary levels.
d household heads were to be found working
mainly in production work or housework.

Table 4.2: Main Activities of Respondent Household Heads by Gender (%)

t Activities Male Female

PO % SBO % PO % SBO %
Agriculture /animal husbandry 13 3 14
Production workers 63 72 47 43
Small Business 8 19 13 -
g. Handloom worker 3 14
Housework 23 29
Source: Bangladesh Data. Key: PO = Poorest Off; SBO = Slightly Better Off

Table 4.3: Primary School Attendance (%)

Type of HH & School Attendance Patterns (% of the age group)

Area of Study Boys Girls Boys and Girls
Never Attended Currently Never Attended Currently Never Attended Currently
attended in the past attending attended in the past attending attended in the past attending
All HHs 14.5 8 77.5 16.7 5.3 78 15.6 6.7 77.8
PO 13.8 9.2 77.1 16.5 5.8 77.7 15.1 7.5 77.4
SBO 17.2 3.4 79.3 17.2 3.4 79.3 17.2 3.4 79.3
HSH 12.7 7.6 79.7 14.5 4.3 81.2 13.6 6 80.4
SHH 25 10 65 33.3 13.3 53.3 28.6 11.4 60
Male Heads 14.9 7.4 77.7 17.8 4.2 78 16.3 5.9 77.8
Female Heads 11.8 11.8 76.5 7.1 14.3 78.6 9.7 12.9 77.4
Mymensingh 24.3 5.4 70.3 32.5 5 62.5 28.6 5.2 66.2
Dhaka 3.1 9.4 87.5 6.5 3.2 90.3 4.8 6.3 88.9
Patuakahli 11.8 14.7 73.5 11.4 5.7 82.9 11.6 10.1 78.3
Dinajpur 17.1 2.9 80 11.5 7.7 80.8 14.8 4.9 80.3
Source: Bangladesh Data, children of primary school age (5 to 10).

Key: PO = Poorest Off; SBO = Slightly Better Off; HSH = Household with Spouse and Head;

SHH = Single Headed Household; HH = Households

D DFID 29
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 48

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

Table 4.4: Secondary School Attendance (%) R

W
Type of HH & School Attendance Patterns (% of age group)
Area of Study F
Boys Girls Boys and Girls
R
Never Attended Currently Never Attended Currently Never Attended Currently
attended in the past attending attended in the past attending attended in the past attending
All HHs 6.8 47.7 45.5 4.7 37.2 58.1 5.7 42.5 51.7 E
PO 8.6 48.6 42.9 5.4 37.8 56.8 6.9 43.1 50 r
SBO 44.4 55.6 33.3 66.7 40 60
HSH 6.5 51.6 41.9 5.1 38.5 56.4 5.7 44.3 50 I
SHH 7.7 38.5 53.8 25 75 5.9 35.3 58.8 C
Male Heads 5.7 45.7 48.6 5.4 40.5 54.1 5.6 43.1 51.4 8
Female Heads 11.1 55.6 33.3 16.7 83.3 6.7 40 53.3 f
Mymensingh 36.4 63.6 10 10 80 4.8 23.8 71.4 p
Dhaka 25 41.7 33.3 7.7 61.5 30.8 16 52 32 s
Patuakahli 33.3 66.7 40 60 36.4 63.6 t
Dinajpur 66.7 33.3 33.3 66.7 50 50
Source: Bangladesh data (children aged 11 -19). Key: PO = Poorest Off; SBO = Slightly Better Off; I
HSH = Household with Spouse and Head; SHH = Single Headed Household; HH = Households g
4.12 Nepal S
Key indicators for the four study sites are The study area was in 6 settlements in 3 of i
given in Table 4.5. the poorest of the districts 14 wards, none N
of which yet has electricity. About 15% of D
Urban site 15
Kapilbastu Municipality in the the population in these settlements are
Western Region (Terai) Indian migrants. Education is relatively T
young in this area the first primary school
Kapilbastu is predominantly Hindu (83%) was established as recently as 10 years ago.
and about 75% of the population are involved The current Net Enrolment Rate (NER) is
in farming. Its HDI and per capita income 16% with only 90 girls attending school of
are below the Terai average (1998 figures). an estimated 4,000 children.

Table 4.5: District Level Indicators

Popn. HH HDI HDI Annual Girls Girls

avg. size avg. for per capita NER NER
area incomeRs. primary lower sec.
Kapilbastu 478,000 6.6 .286 .244 6,541 31% 13% S
Chandhika 232,000 7 .235 .357 5,035 30% 14% F

Darakh 570,000 10 .299 .344 7,319 37% 12%

Dhunche 45,000 not .246 .271 7,111 62% 17%
available
Source: Nepal Report Chapter 4.

Key: HH = Households; HDI = Human Development Index; NER = Net Enrolment rate
1

a
15
The definition of urban in Nepal is as much political as social or demographic. Hence, peri-urban may be a better description. a

30 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 49

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

Rural sites Chandhika VDC, Achham, Far problem for continued attendance at school
Western Region (Hill); Darakh VDC Kailali, beyond primary grades, and educational
Far Western Region (Terai); Dhunche VDC attainments are generally low.
Rasuwa, Central Region (Mountains)
y
g In Dhunche the major caste is Tamang;
Each site has different ethnic, caste and other groups include Sherpa, Chhetri,
religious dimensions. Newar, Ghale, Gurung, Damai, Sarki and
Sunar. The majority of the population is
In Chandhika the major castes include involved in farming. The study area was in
Chhetris (45%) and Brahmins (20%). About Wards 5,6,7 and 8 of Dhunche VDC16.
80% of the population are involved in
farming. The study area was in 3 of the 4.12.1 Nepal Selected Sample
poorest of the districts wards. The caste Characteristics
system is believed to be a very strong barrier The sample of households in the four study
to access to school. sites is shown in Table 4.6 below. There
were 180 HH and 300 respondents
In Darakh the major caste is Tharus, other altogether, 33% of whom were from single
groups include Biswakarma, Damai and headed households. In the PAGs a total of
Sarki. About 90% of the population are 398 persons were involved including
involved in farming. The study area was in students in and out of school, parents,
Nimdi Village (Ward 4) of Darakh VDC. teachers and community leaders.
Distance to the secondary school is a

Table 4.6: Characteristics of Households Sampled: Nepal (Absolute numbers)

Districts/Sites Characteristics of Households Sampled

& HH Types Poorest Slightly Better Off
HSH SHH HSH SHH
MH FH MH FH MH FH MH FH Total
Achham(Chandika) 14 6 23 5 5 3 56
Kailali (Darakh) 26 1 2 8 1 3 41
Kapilbastu 24 3 5 8 2 1 43
Rasuwa (Dhunche) 25 2 3 9 1 40
Total 89 12 33 30 1 8 7 180
Source Nepal report. Key: SHH = Single Headed Household; HSH = Household with Spouse and Head;
FH = Female Headed; MH = Male Headed

16
A problem with the selection of Dhunche was discovered as fieldwork progressed: it did not appear that the wards
advised by VDC leaders were really the poorest. As it was too late to rectify the problem, during writing-up some of the
analyses (especially income and expenditures) were done without Dhunche to avoid skewing of results.

D DFID 31
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 50

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

4.12.2 Primary Activities of At secondary level (Table 4.9) the gender T

Household Heads pattern of primary attendance is repeated
The primary activities of household heads in and accentuated averaging 47% of girls in
the households surveyed were in agriculture all households who have never been to
/animal husbandry and production work for secondary school (again with the notable
men, as is shown in Table 4.7. exception of female headed households).
There appear to be no boys who have
4.12.3 School Attendance dropped out (attended in the past) at the
The attendance of the primary children in the primary level. This is surprising and it could
sampled households is shown in Table 4.8. Girls be that parents were not willing to admit
attendance is notably lower than boys that their boys had dropped out.
averaging 65% for girls and 81% for boys (with
the exception of Female Headed Households).

Table 4.7: Main Activities of Respondent Household Heads by Gender (%)

S
Activities Male Female K
H
PO % SBO % PO % SBO %
N
Agriculture /animal husbandry 44 47 52 63
f
Production workers 34 16 3 -
s
Professional/clerical 9 24 - -
a
Housework 2 3 43 37
w
Source: Table 5.3 Nepal report. Key: PO = Poorest Off; SBO = Slightly Better Off
i
f
Table 4.8: Primary School Attendance (%)
4
Type of HH & School Attendance Patterns (% of the age group)
Area of Study K
Boys Girls Boys and Girls
g
Never Attended Currently Never Attended Currently Never Attended Currently
attended in the past attending attended in the past attending attended in the past attending
All HHs 18.8 - 81.2 28.9 5.6 65.6 23.2 2.4 74.4
PO 20.7 - 79.3 30.2 4.8 65.1 24.5 1.9 73.5 T
SBO 12 - 88 25.9 7.4 66.7 19.2 3.8 76.9
SHH 14.6 - 85.4 10.7 3.6 85.7 13.2 1.3 85.5
HSH 21.7 - 78.3 37.1 6.5 56.5 29 3.1 67.9
Male Heads 22.7 - 77.3 36.2 5.8 58 28.7 2.5 68.8
Female Heads 6.9 - 93.1 4.8 4.8 90.5 6 2 92
Source: Nepal Report Table 6.7 proportion of primary school age children (5 to 10).
Key: PO = Poorest Off; SBO = Slightly Better Off; SHH = Single Headed Household;
HSH = Household with Spouse and Head

32 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 51

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

Table 4.9: Secondary School Attendance (%)

Type of HH & School Attendance Patterns (% of the age group)

Area of Study Boys Girls Boys and Girls
Never Attended Currently Never Attended Currently Never Attended Currently
attended in the past attending attended in the past attending attended in the past attending
All HHs 8.7 21.7 69.6 47.4 15.5 37.1 26.4 18.9 54.7
PO 9.7 23.7 66.7 54.8 17.7 27.4 27.7 21.3 51
SBO 4.5 13.6 81.8 34.3 11.4 54.3 22.8 12.3 64.9
SHH 2.5 20 77.5 33.3 11.1 55.6 17.1 15.8 67.1
HSH 12 22.7 65.3 55.7 18 26.2 31.6 20.6 47.8
Male Heads 10 22.2 67.8 52.1 15.5 32.4 28.6 19.3 52.2
Female Heads 4 20 76 34.6 15.4 50 19.6 17.6 62.7
Figures below for area are for primary and secondary combined %
Accham 9.7 8.6 81.7 23.2 8.7 68.1 15.4 8.6 75.9
Kailali 24.2 9.7 66.1 54.5 9.1 36.4 39.8 9.4 50.8
Kapilbastu 10.4 14.3 75.3 38.5 15.4 46.2 21.7 14.7 63.6
Source: Nepal Report table 6.8 proportion of secondary school age children (11 to 19).
Key: PO = Poorest Off; SBO = Slightly Better Off; SHH = Single Headed Household;
HSH = Household with Spouse and Head

Note that the figures on school attendance Urban site Mbale

for the households included in the research
sample are not representative of the country Mbale District
as a whole, nor their own area. Only those Mbale District is one of the main urban
with school-going children were selected for districts in the Eastern Region of Uganda
inclusion in the study, thus distorting the with a population density of 284 persons per
figures upward. square km. The age group distribution
indicates a high proportion of children in
4.13 Uganda each household (about 46%) and a high
Key indicators for the four study sites are dependency ratio of 50%.
given in Table 4.10.
y
g

Table 4.10: District Level Indicators

Population HH Literacy rate GER (P&S) Secondary

avg. Enrolment as %
size of primary
Enrolment
Mbale (Urban) 931,500 N/A 56%: (63%M; 49%F) 88% 6%
Apac 566,400 N/A 43%: (70%M; 36% F) 87% 3%
Kibaale 300,821 5 N/A 84% 3%
Kiboga 170,000 4 N/A 92% 3%
Source: Table 4.2 Uganda report, and 1998 Educational Statistical Abstract

D DFID 33
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 52

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

The main economic activities in the district There are many cases of early marriages in i
are agriculture, trade, construction, small scale the area. The HIV/AIDS epidemic influences c
and medium-scale processing in cottage marriage patterns. Marrying young women m
industries. Transport infrastructure and is a preferred option, and sexual intercourse
distribution facilities such as markets are with young virgins is understood by some 4
poor due to the bad terrain. men to confer protection against HIV and
other STDs. Early marriage and unwanted T
Rural Sites Apac, Kibaale, Kiboga pregnancies are the main reasons why many s
girls do not complete primary education or w
Apac District proceed beyond primary level. 3
Apac District is located in the Northern Region h
of Uganda. About 80% of the population are In the district women have been mobilised 2
peasant farmers. Ibuje, which is the area of in groups under the Community Mobilisation i
study, is one of the poorest Sub-Counties due Programme, which aims to strengthen the c
to dry spells. The main economic activity in capacity of women through skills training to
the area is fishing in Lake Kwania and the River enable them to benefit from credit schemes.
Nile. The fish is dried and sold in Kampala,
Masindi and also into the neighbouring Kiboga District
countries of Sudan and the Democratic The district is located in central Uganda. T
Republic of Congo. This activity provides a About 66% of the people in Kiboga live in
very quick source of money for boys, and temporary dwellings. The majority of the
therefore, lures them from schooling. people in this area are mainly peasant
farmers. Coffee and tobacco are the main
The extended family system is predominant cash crops while bananas, maize, potatoes,
and men with wealth many livestock practise beans, cassava, vegetables, groundnuts and
polygamy. With the introduction of UPE in fruits are grown as food crops with the
1997, the enrolment of pupils in Apac increased surplus for sale.
by 30% on average. The number of boys
S
enrolled is still greater than that of girls. In the past, families were polygamous but
S
with the changes in the socio-economic
Kibaale District conditions, most men have turned to T
Kibaale District is located in the mid-western monogamous marriages, except in a few
part of Uganda. Population density is low, at 68 isolated cases. The extended family system is
persons per sq. km. However, the population is still prevalent.
unevenly distributed. Most families are extended
in nature, leading to economic pressure on Before 1997, 25,000 children were attending
household heads. Polygamy in the district is a school. However, with the introduction of
widespread phenomenon and practised by men UPE in 1997, the enrolment more than
of all religious denominations. More than 50% doubled to approximately 58,000 children17.
of the total population are children below 15. However, the areas where nomadic pastoralism
S
17
District Education Officer Kiboga District

34 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 53

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

is practised have a higher proportion of 4.13.2 Primary Activities of

s children who do not attend school. As parents Household Heads
migrate their childrens schooling is disrupted. The primary activities of household heads in
the households surveyed were as shown in
4.13.1 Uganda Selected Sample Table 4.12. There are considerable regional
Characteristics variations within the four study sites which
The sample of households in the four study are masked by these average figures.
sites is shown in Table 4.11 below. There
were 195 HH and 300 respondents altogether, 4.13.3 School Attendance
32% of whom were from single headed The attendance of the children in the
households. In the PAGs a total of about sampled households is shown in Tables 4.13
200 persons were involved including students and 4.14. (overleaf) Both girls and boys
n in and out of school, parents, teachers and attendance is high averaging 81% for the
community leaders. sample as a whole the difference between
boys and girls is minimal for each category: a
possible indication of a successful new
education policy.

Table 4.11: Characteristics of Households Sampled: Uganda

Poorest Slightly Better Off

HSH SHH HSH SHH
MH FH MH FH MH FH MH FH Total
Kibaale 27 0 0 4 9 1 0 0 41
Kiboga 16 1 1 17 10 1 0 1 47
Apec 26 0 0 9 10 0 1 1 47
Mbale 17 3 2 24 11 0 2 1 60
Total 86 4 3 54 40 2 3 3 195
Source Ugandan data. Key: HSH = Households with both Head and Spouse; FH = Female Headed;
SHH = Single Headed Household; MH = Male Headed

Table 4.12: Main Activities of Respondent Household Heads by Gender (%)

Activities Male Female

PO % SBO % PO % SBO %
Agriculture /animal husbandry 70 61 49 30
Production workers 15 21 3 10
g
Estate /seasonal worker 5 - - -
Self empolyed street vendor 2 2 16 10
Housework - 2 12 -
Resting 50
m
Source: Table 5.3 Uganda Report. Key: PO = Poorest Off; SBO = Slightly Better Off

D DFID 35
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 54

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

Table 4.13: Primary School Attendance (%) 4

Type of HH & School Attendance Patterns (% of the age group)

Area of Study U
Boys Girls Boys and Girls
K
Never Attended Currently Never Attended Currently Never Attended Currently
attended in the past attending attended in the past attending attended in the past attending (
All HHs 17.8 1.4 80.8 17.4 1.4 81.2 17.6 1.4 81 a
PO 17.9 1.3 80.8 16.6 1.8 81.7 17.2 1.6 81.3 M
SBO 17.5 1.6 81 20.5 79.5 18.7 0.9 80.4 p
HSH 18.8 2.6 78.6 20.2 2.6 77.2 19.5 2.6 77.9 w
SHH 12.9 87.1 7.7 92.3 10.4 89.6 o
Male Heads 19.9 2 78.1 21.1 2.1 76.4 20.7 2 77.3 f
Female Heads 12.7 87.3 8.7 91.3 10.6 89.4 a
Kibaale 24.4 2.4 73.2 21.1 3.5 75.4 22.4 3.1 74.5 T
Kiboga 10.7 1.8 87.5 9.3 2.3 88.4 10.1 2 87.9 i
Apac 14.9 2.1 83 18.8 81.3 16.8 1.1 82.1 (
Mbale 21.4 78.6 18.5 81.5 20 80 r

Source: Ugandan data: children of primary school age (5 to 12). Key: PO = Poorest Off;
p
SBO = Slightly Better Off; HSH = Households with both Head and Spouse; SHH = Single Headed Household
p
l
C
Table 4.14: Secondary School Attendance (%)
Z

Type of HH & School Attendance Patterns (% of age group) r

Area of Study m
Boys Girls Boys and Girls
Never Attended Currently Never Attended Currently Never Attended Currently
attended in the past attending attended in the past attending attended in the past attending K
All HHs 0.6 24.6 74.3 2.1 39.6 58.3 1.3 31.5 66.9 s
PO 0.8 25.4 73.8 1 41 58 0.9 32.3 66.9 p
SBO 22 75.6 4.5 36.4 59.1 2.4 29.4 67.1 p
HSH 1.1 18.7 80.2 2.9 42.9 54.3 1.9 29.2 68.9 b
SHH 26.7 71.1 34 66 30.4 68.5
Male Heads 0.8 24.2 74.2 2.9 41.7 55.3 1.8 32.2 65.6 f
Female Heads 25.6 74.4 34.1 65.9 29.8 70.2 i
Kibaale 25 75 41.2 58.8 32.4 67.6 i
Kiboga 33.3 66.7 2.9 42.9 54.3 1.3 37.7 61 f
Apac 1.6 19.7 78.7 41.7 58.3 0.9 29.4 69.7 a
Mbale 22.7 75 4.5 34.1 61.4 2.3 28.4 68.2 h
Source: Ugandan data (children aged 13-25)

Key: PO = Poorest Off; SBO = Slightly Better Off; HSH = Households with both Head and Spouse;

SHH = Single Headed Household

R
p
G
V

36 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 55

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

4.14 Zambia Historically, in Yuka village there has been an

intermittent inflow of people from Angola.
Urban Site Kapoto (Kitwe, Copperbelt) The settlement is an outlier, one of nine
Kapoto is a large squatter settlement villages in the district. The main economic
y
g (estimated population 35,000) which started activities are subsistence cultivation of
as a farming area for the residents of Kitwe. cassava, millet, maize, sweet potatoes and
Many people came to settle in Kapoto as a pumpkins, animal husbandry, fishing, craft
place for retired miners and industrial making (baskets and mats), and vending.
workers. Later, the City Council embarked Culturally and ethnically, the area is mixed
on an unsuccessful attempt to evict residents and comprises Lozi, Mwenyi, Makome and
from the area. The majority of the people Mbunda. There are also the Kwandi,
are involved in informal sector activities. Kwangwa, Youma and Mbalango ethnic
They buy and re-sell such groceries as food groups from Angola. Yuka village consists of
items and soap. Some depend on piecework a population of some 1,200 people widely
(casual short-term menial jobs). While the scattered over semi-Kalahari sandy areas of
residents of Kapoto come from different Kalabo District.
parts and language groups of Zambia, the
predominant language spoken in Kapoto, Munyama village is one of the settlements
like in other parts of Kitwe and the created after the displacement of the people
Copperbelt, is Bemba. Like the majority of of the Valley following the construction of
Zambians, the social organisation of Kapoto Lake Kariba in 1958. Economically, the area
residents is matrilineal. Most practice is underdeveloped. Fishing is the main
monogamy and some form of Christianity. economic activity of the area. Commercial
y fishing is monopolised by outsiders
g Kitwe district has 24 lower and middle basic (Zambians from other parts of the country
schools, 23 upper basic, 9 secondary and who have established fishing camps in the
private schools. There are a total of 65,680 area) and indigenous people are engaged in
pupils and 1,643 teachers in primary and fishing for subsistence. Farming is equally of
basic schools, giving a teacher-pupil ratio of a subsistence type: there is limited cultivatable
1:40. The majority of teachers (68.2%) are land. The countryside is hilly and the
females. The year 2000 saw a 2.43 % lakeshores sandy. Sorghum is grown in alluvial
increase in the death rate of trained teachers soils. Goats and cattle are reared and are the
in the district with a loss of 13 male and 27 main source of wealth for the local people.
female teachers. In secondary schools, there
are 291 teachers; the figure is inclusive of The dominant ethnic groups are the Tonga,
head-teachers and their deputies. followed by the Goba. The population is
sparse and estimated at 3,000 people
Rural Sites - Yuka village (Kalabo, Western scattered over a wide area. Facilities found in
province); Munyama Village (Siavonga, the village include a primary school, a clinic
Gwembe Valley, Southern province) Mulanga and the headquarters for Harvest Help, an
Village ( Chinsali, Northern province) NGO that built the school and is responsible

D DFID 37
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 56

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

for running the clinic. Harvest Help is is sparse (c. 2,500) and distributed in 4
central to the lives of the people of hamlets which form the village area. The
Munyama and other lakeshore settlements. area comprises three villages of Nsofu, T
The organisation has built five other schools Chuma, and Kalanguluka. A Roman t
along the lakeshore and provides relief food Catholic Mission serves the villages T
during drought. According to the District providing a rural health centre, staff houses
Education Officer (DEO), it is estimated and a church building. Village houses are 4
that between 10 to 15 % of school-going age built out of sun-dried bricks and roofed with S
children are out of school in the District. thatch or iron sheets. i
The teacher-pupil ratio is estimated at 1:47. b
4.14.1 Zambia Selected Sample
Mulanga village is a rural settlement located Characteristics T
70 km south of Chinsali District in the The sample of households in the four study
Northern Province. The whole area is sites is shown in Table 4.15 below. There
economically depressed. Groundnuts are the were 190 HH and 291 respondents
main cash crop of the area but the local altogether, 28% of whom were from single
people find it difficult to market this headed households. In the PAGs a total of
produce. The people of the area traditionally 382 persons were involved including
grow millet and cassava for subsistence. The students in and out of school, parents,
main ethnic groups found in Mulanga are teachers and community leaders.
the Bemba and Bisa people. The population
S

Table 4.15: Characteristics of Households Sampled: Zambia T

Poorest Slightly Better Off

HSH SHH HSH SHH
MH FH MH FH MH FH MH FH Total
Kitwe 13 1 18 5 4 3 44
Chinsali 11 1 9 23 2 46
Siavonga 25 1 2 2 11 2 43
Kalabo 23 7 2 12 12 1 57
Total 72 9 5 41 51 4 8 190
Source: Zambian data. Key: HSH = Households with both Head and Spouse; SHH = Single Headed Household;
MH = Male Headed; FH = Female Headed

S
H

38 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 57

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

4.14.2 Primary Activities of as never having attended school. In many

Household Heads cases this could be because children are
The primary activities of household heads in considered by their parents to be too young
the households surveyed were as shown in to go to school even though of school age
Table 4.16. a possibly convenient excuse if the real
reasons are economic. Rates of attendance
4.14.3 School Attendance are similar for boys and girls (except in
h School attendance figures (Tables 4.17 & 4.18) Siavonga). Differences between SBO and
indicate a substantial number of children at PO households are pronounced at the
both primary and secondary levels reported secondary level.

Table 4.16: Main Activities of Respondent Household Heads by Gender (%)

Activities Male Female

PO SBO PO SBO
Agriculture /animal husbandry 74 72 50 36
Production workers 8 9 4 9
Estate /seasonal worker 2 - -
Self empolyed street vendor 7 2 17 -
Housework 1 - 20 27
Self empolyed sales worker 3 4 18
Source: Zambian data. Key: PO = Poorest Off; SBO = Slightly Better Off

Table 4.17: Primary School Attendance (%)

Type of HH & School Attendance Patterns (% of the age group)

Area of Study Boys Girls Boys and Girls
Never Attended Currently Never Attended Currently Never Attended Currently
attended in the past attending attended in the past attending attended in the past attending
All HHs 37.3 9.8 52.9 38.6 9.7 51.2 38 9.7 52.1
PO 38 10.2 51.8 43.1 11.5 45.4 40.4 10.9 48.7
SBO 35.8 9 55.2 31.2 6.5 61 33.3 7.6 58.3
HSH 38.8 8.6 52.6 40.9 6.7 51.7 39.9 7.6 52.2
SHH 32.7 13.5 53.8 32.8 17.2 50 32.7 15.5 51.8
Male Heads 35.4 8.8 55.8 42.3 7.7 49.3 38.8 8.3 52.6
Female Heads 42.1 12.3 45.6 30.8 13.8 55.4 36.1 13.1 50.8
Kitwe 25.5 15.7 58.8 32.6 15.2 50 28.9 15.5 54.6
Chinsali 25.4 8.5 66.1 23.9 13 63 24.8 10.5 64.8
Siavonga 48.6 51.4 56.9 1.7 41.4 53.7 1.1 45.3
Kalabo 52.6 12.3 35.1 36.8 10.5 52.6 44.7 11.4 43.9
Source: Zambia Data, children of primary school age (5 to 13). Key: PO = Poorest Off; SBO = Slightly Better Off;
HSH = Households with both Head and Spouse; SHH = Single Headed Household; HH = Households

D DFID 39
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 58

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

Table 4.18: Secondary School Attendance (%) S

Type of HH & School Attendance Patterns (% of age group)

Area of Study K
Boys Girls Boys and Girls
6
Never Attended Currently Never Attended Currently Never Attended Currently
attended in the past attending attended in the past attending attended in the past attending v
All HHs 4.8 44.9 50.3 8.3 44.8 47 6.7 44.8 48.5 K
PO 7.8 53.3 38.9 10.1 56 33.9 9 54.8 36.2 p
SBO 31.6 68.4 5.6 27.8 66.7 3.1 29.5 67.4 w
HSH 5.3 39.8 54.9 9.8 44.7 45.5 7.8 42.4 49.8 l
SHH 2.9 61.8 35.3 4.1 44.9 51 3.6 51.8 44.6 s
Male Heads 4.6 39.8 55.6 8.3 44.2 47.5 6.6 42.1 51.3 m
Female Heads 5.1 61.8 35.9 8.2 45.9 45.9 7 51 42
Kitwe 73.5 26.5 7.5 47.5 45 4.1 59.5 36.4 T
Chinsali 34.2 65.8 31.3 68.8 32.6 67.4 c
Siavonga 12 12 76 19 38.1 42.9 15.2 23.9 60.9 o
Kalabo 8 50 42 11.1 54.2 34.7 9.8 52.5 37.7 p
Source: Zambia data (children aged 14-25). Key: PO = Poorest Off; SBO = Slightly Better Off;
w
HSH = Households with both Head and Spouse; SHH = Single Headed Household; HH = Households
w
4.15 Kenya economic activity of Mandani village is farming p
No SSIs were administered in the Kenya but it has a scarcity of water (mainly rain water) a
study all fieldwork was done through PA for successful farming. When rains fails (it had v
sessions with different groups of villagers. failed for two years prior to our visit) families m
are devastated. Only a few families who have one s
Site One: Mandani village: Mwingi district of their kin in paid employment are in a better T
position of getting regular income. It is this
Mandani village is in Mwingi district. There situation which leads to a desire for education T
are a total of 331 people living in 58 households as a means of escaping from poverty by being s
The village is settled by people of Akamba able to access formal paid employment. p
ethnic group whose main practice is small- K
scale farming, weaving and bee keeping. The Results of wealth ranking for Mandani village a
general economic status of people is poor. show that 19 households are categorised as T
Available statistics, for example, the 1994 very poor, 17 households were categorised p
Welfare Monitoring Survey Report (Central as poor, 11 households were categorised as s
Bureau of Statistics, 1996) states that average middle while 11 other households were t
annual income per person in the rural areas is categorised as rich. The first school in the i
Ksh 8,505 (US$117). However, in Mandani area was built by a missionary in l946 at w
incomes for the poorest persons are much Ithumbi, twelve kilometres away from the
lower: Ksh. 100 (US$1.25) per month. village. During the late sixties, a primary
school was built at Mbondoni, eight
There are several reasons cited for the poor kilometres away, and by the mid-seventies a
status of these villages. For example, the main primary school was started in the village.

40 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 59

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

Site two: Kapsinendet village: Bomet district: Kapsinendet increases the preference for
boys education. A historical time-line for
Kapsinendet village has a total of approximately Kapsinendet village was produced by
600 people living in 72 households. The participants which helps to explain the local
y
g village is populated predominantly by the demand for education based on the long
Kipsigis ethnic group whose traditional history of exposure to European education
practice has been cattle keeping. However systems and available institutions and the
with increasing population and decreasing impact of cost sharing on effective demand
land size, there has been a gradual shift to (Figure 4.1 overleaf).
small-scale farming with the production of
maize as the main crop. Results of wealth ranking for Kapsinendet
village were similar to those of Mandani,
There is only one village road; other although the proportion of the very poor is
communication ways are only footpaths. Part smaller. Those villagers categorised as very
of this village is served by a community water poor were 10, those categorised as poor
pipe that was initiated when the government's were 32, those categorised as middle were
water supply system failed. The community 23 while those categorised as rich were 7.
water supply serves less than half the villages Not only did the informants indicate that
g population leaving the majority without Mandani village has more villagers in the
) adequate water supply. Just as in Mandani very poor category than Kapsinendet, but
d village, the study team observed that the their definitions of poverty in the two
s middle and rich persons live near the roads, villages also differ, perhaps reflecting
e shopping centres, schools and other services. differences in perceptions, based on the
The poor live far off at the corners of the village. somewhat different socio-economic
structures.
n The 1994 Welfare Monitoring Survey Report
g states that Kapsinendet village has a monthly The first school, initially attended by boys
per capita income of between Ksh 350 to only, was built in the neighbouring village in
Ksh 500 (US$ 5-7) compared to the national 1956. By the beginning of 1960 a few girls
annual average of Ksh 8,505 (US$117). started going to school but most dropped
The village suffers long dry spells, and this out soon after. Villagers interviewed during
periodic drought affects potential milk this study attributed this to pressure at the
supply and crop production. This contributes time from the community for the girls to get
to low income potential and in turn circumcised and eventually settled in
influences school attendance, encouraging marriage.
withdrawal from school. Cultural practices in

D DFID 41
 Figure 4.1: Historical Time Line for Kapsinendet Village

42
The coming of
Europeans and the
introduction of Cost sharing
European education extended to colleges.
and religion. Construction of a Liberalisation of
Mission was built A few boys school at Kipsuter, economy. Shortage
50km away. continued with
7km away. More of drugs and health
Churches were school. There was
people went to facilities. There are
constructed and no payment. Girls
school, but the Kapsinendet no funds to meet
used as schools. were still not
majority still did primary school school levies and
People did not allowed to go
not want to go was built. More many have been
accept education to school to school children enrolled forced to drop out
1910-1920 1930-1940 1950-1960 1980-1990
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 60

4 Methodology and Sample Characteristics

1900-1910 1920-1930 1940-1950 1960-1970 1990-Onwards

A few boys joined More churches More pupils 8.4.4 system of
Reaching the Poor - The costs of sending children to school

school but girls constructed. joined Kipsuter education

were excluded. Introduction of primary school. introduced and
Churches were modern clothes. First girls went to cost sharing
used as schools There were no school but most demands increased.
modern schools, did not remain for Children further
those interested many years as they dropped out of
travelled to a dropped out for school due to lack
boarding school marriage of funds and food
50km away

DFID
l
l
i
i

t
t
s
s
s

a
c
c
a
c

e
4

g
p

d
d

A
A
T
T
A

V
A

G
D
N

m
m
m

M
M
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 61

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

4.16 Sri Lanka The total number of families in Andimune is

No SSIs were administered in the Sri Lanka approximately 600. A cross-section of 133
study all field work was done through PA families was chosen as a sample for the study.
sessions with different groups of villagers. A wealth-ranking exercise was conducted
with a group of individuals (both male and
Andimune Community female) from the sample families. 23 families
(17%) were classified as rich; 28 (21%) as
Andimune is a fishing village on the west average, and 82 (62%) as poor.
coast of Sri Lanka located north of Colombo
in Udappuwa Division, part of Puttalam Udappuwa, including Andimune, is well
district. The entire population of Udappuwa known among Sri Lankans as a district with
Division around 2,000 families is Tamil, a unique identity, of comparatively recent
predominantly Hindu, with some Roman South Indian origin and with strong
Catholics. Andimune village includes more religious and cultural traditions. Indeed
than 600 families, many of whom were during the study it was observed that the
settled under a special housing scheme. social life of the people in Udappuwa is
The main economic activity in the village is closely associated with the village Kovil
fishing. In Udappuwa Division, there are (temple) and the fortunes and misfortunes
about 1,500 families engaged in the fishing of life are attributed to the acts of Kovil gods.
industry. The fishing season on the west
coast of Sri Lanka lasts from October to The position of women in the community is
May. During the off-season, most fishermen governed by traditional Tamil values. Girls
migrate to the East coast, particularly to are married young, usually soon after they
Trincomalee. However, in recent years the reach puberty. The attainment of puberty by
civil war has limited this seasonal migration. girls is celebrated in the traditional way.
Many people in Andimune, particularly older Many girls of school-age do not attend
men and women, work in prawn ponds as school regularly.
labourers or watchers, earning about Rs150
a day. However, it is said that when women Rohanapura Community
go to work in the ponds, their children
dont go to school. Also, some children are Rohanapura village is located in the Suriya-ara
employed as night watchers which affects Division of Uva Province. Originally settled
their attendance. by migrant families, it now has 101 families
and a total population of 475, 232 men and
Andimune is served by two schools, 243 women. Among the adults in the
Andimune Vidyalaya and Udappuwa Maha village, 40% are illiterate, 45% had studied up
Vidyalaya. Udappuwa MV goes up to to primary level, and 5% had received
Grades 12/13, offering GCE AL, and is the secondary education. All are Sinhala
larger school. Both are Tamil medium, Buddhists.
mixed schools.

D DFID 43
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 62

Reaching the Poor - The costs of sending children to school

4 Methodology and Sample Characteristics

There is no electricity in the village. Water is water. The major problems encountered by 5
scarce, its quality is poor, and the fluoride the farmers are the high cost of inputs such I
content of the water is high. The gravel as hiring tractors, purchasing agro-chemicals b
roads in the village were constructed and marketing of their produce. Many i
through shramadana, with financial farmers take out credit to fund these inputs,
assistance from NORAD and World Vision either from money-lenders, the shop owner
International. The village is about 2 km or the mudalali (village entrepreneur) at
from the main road, and the villagers use exorbitant interest rates.
public transport to reach nearby towns.
A wealth-ranking exercise was carried out
The village is served by Sevanagala Usbim with the villagers to identify the socio
school. The medium of instruction is Sinhala. economic differences in the village. In
Rohanapura, there are 11 rich families
The main economic activity in the village is (10%), 35 average families (35%), 33 poor
highland agriculture (chena cultivation). The families (33%) and 22 (22%) very poor
main crops are green gram, cowpea, families. However, this is all relative, as I
kurakkan (millet), black gram, ground nuts, except for 6 families in the rich category o
maize and vegetables such as pumpkins, who owned hand tractors (3) and small f
brinjal, bitter gourd, snake gourd, cultivated boutiques (3), all the other families are f
on the family plots. Farmers also try to grow recipients of government Samurdhi s
perennial crops such as coconuts, cashew subsidies. w
and lime with little success, due to scarcity of t
e
t
c
t
e
t
N
e
d

I
b
a

44 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 63

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

5.1 Introduction benefits (as perceived by the respondents)

In this section findings are presented from that education brings to the individual
both the SSI and PAGs. The areas covered and/or household (i.e. the social and
include: economic returns to education) are also
The value and benefits of education; considered.
Constraints to access and barriers to
education; (i) Value of Education
Quality of education; Do Parents Value Education?
Income and expenditure; The attitudes of poor parents and their
The costs of education; children towards the benefits of education
Ability and willingness to pay for are seldom well understood. Throughout
education; the fieldwork many teachers and community
Household sacrifices; and leaders (though by no means all) expressed
Equity issues in education. the view that the poorest parents (who they
believe to be generally uneducated) have
It is worth bearing in mind from the outset little or no understanding of the benefits of
of this section that the comparative report education and many children do not attend
faces the problem of any attempt to generalise school (or attend irregularly) because their
from data collected from heterogenous parents do not value education, particularly
sources. The individual country studies, upon secondary education.
which this report is based, have demonstrated
the complexity and the heterogeneity of the Even during post-fieldwork dissemination
experiences of households within and between workshops18 (e.g. Bangladesh, Zambia and
the different communities in each of the study Nepal) involving ministry officials, teachers,
countries, and between the study countries and donors this viewpoint was echoed, albeit
themselves. As such caution needs to be to a lesser extent. However, this study clearly
exercised when considering the relevance of demonstrates that, on the whole, the poorest
these findings to a wider context. parents and their children do indeed value
Nonetheless, some general patterns are education and usually have clear and rational
evident and these are also included in the reasons for not participating, or participating
discussion. infrequently. Some interesting gender
differences between the qualitative and
5.2 Value and Benefits of Education quantitative findings qualify these remarks.

5.2.1 Household Perceptions on the When parents were asked Do you think
Value of Education primary and secondary education is important
In this section the value attached to education for boys and/or girls? the overwhelming
by parents and children (both school-going response was that it is equally important for
and out of school children) is examined. The both as demonstrated in Table 5.1 overleaf.

18
In each of the four main study countries the findings of the study were disseminated through a seminar.

D DFID 45
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 64

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Table 5.1: Importance of Primary and Secondary Education (% of Respondents) T

d
All respondents Importance of Primary Education Importance of Secondary Education
(
Boy Girl Both Boy Girl Both
w
Nepal 15.1 1.8 83.1 20.4 2.2 77.3
c
Bangladesh 19.3 3 79.7 20 2.7 77.3
i
Zambia 6.4 1.8 91.8 8.6 0.7 90.7
s
Uganda 4.7 1.7 93.6 10 2.7 87.3
S
Source: SSI Q 32
c
At both primary and secondary levels a higher Purdah19. In contrast, communities in the b
percentage of respondents in the African hills and mountains are traditionally (
countries tend to favour the education of Buddhist and although physical access can f
both sexes than those in the Asian countries. be difficult, particularly in the mountains, c
However, in all countries it is shown that and enrolment of children in these areas is t
where respondents have a preference for boys low, drop out rates also tend to be low. In p
or girls only it is nearly always the boys who the Terai areas the situation is almost b
fare better, and this is particularly marked in reversed with relatively high enrolment at a
Nepal and Bangladesh. Moreover, the pattern the primary level followed by a rapid drop p
is accentuated at the Secondary level where out out of girls, and very low transition rates f
there is a greater preference for boys than girls of girls to secondary school (M. Sibbons,
than at the primary level. Results from Kenya 1998)20.
and Sri Lanka where only qualitative research
techniques were used give a similar picture. (ii) Gendered Perceptions
During PAGs discussions it was generally
There is some variation of results when the acknowledged in all sites that in the past boys
same question is analysed by area. For example, education was preferred to girls by both
in Nepal, the area of Achham (rural) attaches mothers and fathers. The economic reasons
a very high value to the importance of were overriding, with boys money earning
education of both children 98.7% and potential after an education being enhanced,
96.0% for primary and secondary respectively. which was not the case with girls. There is
This compares to Kapilbastu (urban/Terai) evidence from all the studies that these
where the figures are 61.3% for primary and attitudes are changing and that girls education
56.0% for secondary. This reflects a different is increasingly given value. However, there
attitude towards the value of education for are considerable differences between the
boys and girls in Terai communities than countries, and even within the countries,
those in the hills and mountains. In the with different locations showing a variation I
Terai areas communities are traditionally in support for and reasoning behind the d
Hindu and Muslim, which still practice educating of sons or daughters. e
s
19
The segregation and veiling of females post-puberty. The segregation restricts girls access to schools.
h
20
M. Sibbons, Approaches to Gender Awareness Raising experiences in a government education project in Nepal, in Gender

and Development, Vol 6, No.2, July 1998, Oxfam Journal

46 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 65

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

The findings suggest that although parents reasons (school fees being due, no clothes)
do value education for both boys and girls affect both sexes equally, girls are needed for
(as supported by the quantitative findings) household chores and not boys. Change in
when a household is faced with financial perceptions of the value of education for
constraints priority would be given to boys boys and girls may therefore be evident in
in Uganda, Zambia and Kenya. (This the equal numbers in primary school, but
situation is more pronounced at the underlying attitudes do still persist as shown
Secondary schooling level.) The reasons in the few girls in secondary school and their
cited in these countries were that girls will involvement in household chores.
become pregnant or get married very young
(these reasons were cited by mothers, Time spent on household chores very
fathers, schoolgirls and schoolboys). The probably also contributes to the failure of
children (both boys and girls) involved in girls to do as well in school as boys (see the
the PAGs confirmed that boys were given next section), as they are given less time and
priority in times of hardship. Out-of-school less support to carry on their learning after
boys from Kiboga and Kibaale (Uganda) school hours. The PAGs with school
also acknowledged early pregnancies as a students support this view, indicating that
problem for secondary girls and had the the need for girls to collect water and fuel
s following to say: for the household affects their school
attendance. A vicious circle is established in
Parents should provide necessities to poor these situations where girls have to be
children especially the girls who cannot work exceptional to continue their schooling.
and earn on their own to meet all that they
require. If this is done it will enable them to In Nepal, where gender differences in access
concentrate on their education instead of are very marked, responses to questions about
getting involved in affairs with men with a the relative importance of education for boys
hope of getting what they need. This has always and girls showed a greater preference for
resulted in pregnancy and early marriages. boys education, at both primary and
, Out-of-school boys of Kiboga and secondary level and by men and women.
Kibaale, Uganda However, there was some difference in the
reasons for education being valued, with men
n Researchers observation giving more emphasis to economic priorities
secondary school boys of Kiboga spent and women to social and welfare concerns
most of their monies on girls. (leading a better life, being more responsible
in the family). These perceptions of relative
In Bangladesh there is a distinct gender value and worth of girls and boys education
difference in the reasons given for the had transferred to the next generation: the
estimated 30% of boys and girls not attending same views were expressed by school boys and
school. Although nearly half are absent for school girls about the value of their education
health reasons, and other economic-related as those expressed by parents.

D DFID 47
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 66

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

In Sri Lanka there were significant variations For example, in Uganda, parents, generally T
of opinion on the value of education for girls speaking, claimed not to differentiate between
and boys. At the primary level there appear sons and daughters in the support they give
to be very few gender differences, however, to their education saying they are both equally
at the junior secondary level and senior important. However, talking to the children
secondary level education was thought to be a different reality emerged. Girls in Kiboga
less important for girls (who would simply for example said that their parents paid for
become housewives) . In Nepal and
21
their brothers education first, and only paid
Bangladesh the PA findings tended to match for girls if they had enough money left, and
with the quantitative findings, with the this was repeated by out-of-school girls in
reasons for valuing education being more or Apac (both in the Uganda study). Parents
less similar for boys and girls. Gender differences had said they valued education equally for all
were apparent in the emphasis on expected their children. The childrens perceptions do
benefits which boys and girls receive from appear to fit reality rather better than parents
S
education (this is discussed further in the stated views. These same differences emerged
following section on the benefits of education). in Kenya with students (boys and girls) T
acknowledging gender differences in decision
It is important to note that although the value making by their parents.
of education was generally found to be equal
for boys and girls in Sri Lanka, Bangladesh Boys are aware of the reasons for girls not
and Nepal, here as in the other countries attending school and of the differences that
when households are faced with financial exist between them and their female peers.
difficulties it is the girls who suffer. This For example, in reference to secondary
point came through in the discussions with education in Kiboga (Uganda) out-of-school
children, more so than with adults. Although boys said: Lack of secondary schools nearby has
out-of-school boys and girls in Nepal were discouraged many children from continuing with
themselves aware of the value of education secondary education. This has greatly affected girls
they could not attend school this was not a who fear walking long distances and disturbances
matter of choice, they explained, but the result by men on the way to and from school. Some girls
of their households poor economic situation. have got pregnant as a result of getting lifts from
rich men.
(iii) Childrens Perceptions
Childrens views have been referred to several (iv) Changing Attitudes
times above. In many of the study sites an In each country it is apparent that parental
interesting feature has been the differences attitudes have changed over time in relation
that emerged between parents and childrens to the value attached to education. Table 5.2
perceptions of gender differences in support below demonstrates the percentage of all
for education. respondents who stated that their attitudes

S
21
Note that as our study targeted disadvantaged communities, some bias is to be expected. National statistics from Sri
Lanka show that girls are actually doing as well as boys in terms of access and achievement.

48 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 67

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Table 5.2: In your View Has Education Always been Equally Important for
n Boys and Girls (% of respondents)?

Gender of respondents Primary Secondary

y
Yes No Not Not Yes No Not Not
n
Sure Applicable Sure Applicable
Nepal Male 72.1 27.9 67.5 32.5
Female 65 35 57 43
Bangladesh Male 78 22 85.1 14.9
Female 77.7 22.3 86.2 13.8
Zambia Male 89 11 92.9 7.1
Female 86.1 13.9 83.1 16.9
l
Uganda Male 74 17.6 1.5 6.9 62.9 25.8 11.4
o
Female 77.8 16.8 5.4 71.3 19.8 9
s
Source: SSI Q34
d

Table 5.3: Reasons for Changes in Attitudes (all Respondents)

n
Nepal Bangladesh

Primary Primary
Demonstration effect of general increase Everybody should receive education (45.3%)
in girls schooling (33.7%) At present female education is a must (34%)
Social awareness favouring girls
education (26.5%)

Secondary Secondary
l Importance of education realised (27.4%) Everybody should receive education (45.7%)
as Demonstration effect of general increase After receiving education they can respect all
h in girls schooling (26.3%) (28.5%)
ls
Zambia Uganda
s
ls Primary Primary
Now men and women have equal Increased sensitisation towards the value of
m
opportunities (35.5%) education for boys & girls, including financial
People have lost trust in marriage (25.8%) incentives (42.6%)
Availability of empowerment opportunities to
both (13.1%)

Secondary Secondary
Girls now go for/want higher education Increased sensitisation towards the value of
(35.5%) education for boys & girls (22.8%)
2 Now girls also need education (22.6%) Availability of empowerment opportunities to
both (14%)
Reduction in cases of pregnancy (14%)
Parents now benefit economically from boys
and girls (14%)
Source: SSI Q35 and 39

D DFID 49
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 68

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

towards the education of boys and girls have In Kenya workshop participants also agreed w
always been the same. that the value of education had changed i
over time, reaching a climax around the mid w
In all countries except Uganda, a greater 1980s, from which time it has gradually d
percentage of female respondents than male fallen to a lower level. This is because, although s
state that their own attitudes have changed education is still considered to be a necessary
i.e. they have not always considered it equally and valuable tool in the Kenyan study areas T
important to send boys and girls to school. it is no longer one that can guarantee Z
In Nepal the changes in attitude amongst employment any investment in education r
females appears to be greatest (35% and 43% is therefore made only after great consideration. h
at primary and secondary respectively). Nepal In Sri Lanka parents stated that their opinions m
also demonstrates a greater change in attitudes on the value of education had changed but I
between the PO and SBO households. for some it had not changed in a positive r
way. Like Kenya there are fears about the b
The reasons for the changes in attitudes vary returns to the investment, i.e. employment is s
from country to country. However, in each no longer guaranteed in the locations studied. b
country the reasons relate in some way to a Moreover, traditional views and beliefs
positive shift in favour of female participation continue to influence decisions particularly I
in education this is encouraging. Table with regard to the education of girls. s
5.3 shows the most common responses. b
5.2.2 Benefits of Education s
The PA findings support the quantitative As noted in the earlier section the gender 8
findings with parents (and in some cases gap in education does seem to be narrowing r
teachers and community leaders) agreeing at least where parental attitudes to the value i
that their attitudes towards educating boys of education are concerned. Nonetheless, it e
and girls have changed. is apparent that there is still a stronger B
See below for examples: interest in the education of boys. r
International research would suggest that m
Parents had no interest in the schooling of this is closely related to different perceptions h
their children in the past. But now they are about the economic returns to be derived a
more aware of the need for education for it is from educating a boy as opposed to a girl c
necessary for prosperity and enjoyment of life. (and to that extent the changing of attitudes r
(Primary Teachers, Bangladesh) cited above suggests changes in these
economic returns also) although socio T
You know our parents (ancestors) long time cultural and religious factors may impact on i
ago used to marry polygamously in order to schooling decisions to varying degrees, t
bear many children who would look after them depending on place and time. b
in their old age. The function of polygamy has b
been taken over by education. (PAGs with The research investigated both the economic t
females, Zambia) and non-economic benefits of education for
2

this sample but, in particular, it explores c

50 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 69

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

whether the benefits are realised by the better marriage prospects and acquire
individual and/or the household and knowledge and skills respectively.
d whether this has any effect on household
decision making to support childrens It is only in Bangladesh that the quantitative
h schooling. results distinguish between the different
y benefits realised by boys and girls, however,
The majority of respondents in Bangladesh, the qualitative findings for all countries
Zambia and Uganda (56%, 69% and 53.3% provide more detail on this issue. In Nepal,
respectively) were of the opinion that the Bangladesh and Sri Lanka parents motivation
. household benefits from educating a child for boys education relates mainly to the
s more than the individual child him/herself. economic benefits (better jobs with higher
In Nepal however, the majority of incomes) whilst the motivation for girls
respondents (59.7%) felt that the individual relates to better family care and marriage
benefited more from education. Figure 5.1 prospects. In contrast in Zambia, Uganda
s summarises the responses relating to the and Kenya although the majority of PAGs
. benefits to the household. participants held the view that the benefits
for boys are mainly economic and that for
In all countries except Zambia financial girls they relate to marriage and/or family
support to the household is seen as a major care, there were some participants (female)
benefit of education, with Bangladesh who held the view that boys and girls can
showing the highest percentage (83% and both benefit economically (and that girls do
80% for male and female respondents not simply benefit through better marriage
g respectively)22. There is very little variation prospects and/or better family care).
e in responses between male and females in
each of the countries, although in According to PAGs participants this change
Bangladesh and Uganda the female in attitude can be attributed to an increase in
respondents have provided responses relating employment opportunities for women and
more directly to family care (better the fact that both men and women quite
s housekeeping, better education for children often have to try to get employment to be
and support to young siblings). The other able to meet the financial needs of the
category shown is made up of a variety of household. In Zambia there appears to be a
s responses. greater incidence of marriage break-down
with a consequent decline in reliance on
The individual benefits of education are shown lifetime marriage as a source of security.
in Figure 5.2. The majority of responses relate Women are having to look after themselves
to getting a better job or being economically and their families more so than in the past.
better off, except for the responses provided
by females in Bangladesh and Uganda where
c the highest percentage of responses was

22
In Zambia respondents indicated that helping the family is the major benefit to the household. However, what is not
clear from the quantitative results is whether this help is financial and/or otherwise.

D DFID 51
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 70

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings
Figure 5.1: Household Benefits F
Napal - Male respondents Napal - Female respondents N
19%
14%

39% 9% 30%

10%

15%

11%
1

14% 8% 19%
12%

Provide financial support to family Become resonsible member of family

Enhance social prestige of HH Ensure old age support to parents
Gain knowledge, skills and manners Other
B
Bangladesh - Male respondents Bangladesh - Female respondents
2% 5% 8%
4%
4% 9%
2%
3%
3
3
83% 80% 3

Can sustain family with job Will be able to provide better education to their children
Can be educated and established Provide better care for old people
Daughter can be good housekeeper Other
Can serve society better

Zambia - Male respondents Zambia- Female respondents

Z
21% 22% 23% 21%

16% 3
15% 19%
16%

18% 8%
14% 7%

When educated they can help the family They look after their parents
Family benefits through material support Help family financially
It depends on the indivuals personality Other

Uganda - Male respondents Uganda - Female respondents U

21%
27%

38% 39%

9%

6%
9%
8%

12% 10% 12%

9%
Get job to help family Support young siblings
Provide financial support to family Building permanent/better house
Prestige/respect/happiness Able to provide basic health care
Improve general welfare of house Other
Source: SSI Q42

52 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 71

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings
Figure 5.2: Individual Benefits
Napal - Male respondents Napal - Female respondents

10%
12%
6%
6% 12%
42% 39%
7% 7%
9%
14%
20% 14% 16%

Get employment/earn money Lead a better life

Become self-reliant
Able to support parents

Earn social prestige

Gain knowledge and skills
Other
Bangladesh - Male respondents Bangladesh - Female respondents
6%
3% 3% 3%
3%
3%
6% 3%

29%

6% 7%

3% 48%
3% 7%
3%
3%
3%

19% 39%

Can bring more prestige to own and spouses family Boys can develop their family

ldren
Can Sustain family with job
Boys will be established

Daughters can have better marriage prospects

Can be good housekeeper

Provide better care for old people

Creates awareness about health and hygiene
Boys will be wisemen Boys can have business and family

Boys can improve their knowledge

Zambia - Male respondents Zambia- Female respondents

4% 7%
15%
21%
19%

32%
36%
4%

55%
7%

They can educate their siblings Individual is able to lead an economically better life
Family benefits through material support Get well paid jobs
Help support parents
Uganda - Male respondents Uganda - Female respondents
5% 5%
5%
9%
9%

31% 9% 39%

9%
9%

14% 27% 9%

23%

Provide financial and material support to HH Self-sustenance

Prestige/respect/happiness Build a better/permanet house
Help family with food and clothing Get a job
Acquire knowledge and skills Reduces burden on parents
Easy interaction/communication Source: SSI Q42

D DFID 53
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 72

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

The school children involved in the PAGs 5.2.3 Value and Benefits of
also had their views on the benefits of Education vs Participation
education as demonstrated below. Thus far it has been seen that the parents
and children (even the poorest) do value
Benefits of Education:
education this is supported by both the
High School Girls (x11, class 8 to 10)
qualitative and quantitative information
Patuakhali, Bangladesh
from each of the study countries. Moreover,
Education is necessary to get a job the benefits of education as perceived by
It is necessary for national development parents and children have been investigated.
It is necessary for maintaining accounts The reasons why participation in education
Education creates awareness of social values is so low, despite such positive views of
Education helps to serve the country education in each of the countries studied,
They can teach brothers and sisters and are discussed at length in subsequent sections
well as neighbours of this report. Some of the points for
It leads to good marriage consideration include:
Superstitions are eliminated
through education Although parents may be convinced of the
value of education and the benefits derived
Benefits of Education:
from it, they may feel that the schooling
Primary School Boys (x9, class III to V)
system fails to provide the kind of education
Patuakhali, Bangladesh they are looking for (either through the
Education will lead to honour and provision of poor quality education or a
respect of parents curriculum that is not seen to be relevant).
Helps to get a good job The issue of quality is discussed further
Improve financial condition of family in Section 5.4.
Enables to write letter
Helps to be a business man Households facing financial constraints are
making rational choices when deciding not
The school children in Bangladesh cited above to send their children to school. The study
did not discuss whether they thought the has looked at the economic and non
benefits of education accrue to the household economic motives for demanding
and/or individual, nor did they rank the education or not (the benefits/returns). 5

perceived benefits in order of priority. In addition the economic climate in each

Nonetheless, it is apparent that the economic of the study countries provides a further A
benefits are thought to be important for both disincentive. Many respondents (in both e
groups, although more so for boys. It is also the qualitative and quantitative results) c
interesting to see that the girls include better made it clear that they did not send their
marriage prospects, as did their parents. children to school because there would be
Results from the PAGs in Uganda generated no job at the end of it (i.e. no return on
similar findings, although there was no their investment). This viewpoint came
mention of better marriage prospects. through very strongly in Sri Lanka, Zambia,

54 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 73

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Uganda and Kenya. Also, when households school regularly, and leaving school
suffer from internal shocks (e.g. death of prematurely, and to identify the main reasons
income earner) a child or children are which affected any of these education
withdrawn from school as a coping parameters. In order to determine if factors
mechanism. The seasonal variation in impact differently on rural or urban
income often results in households households, on one sex more than the other,
having no cash to pay for fees and this can and on the poorest more or less than the
also impact negatively on households slightly better off, data was disagreggated by
ability to send children to school. This is these variables. The following notes summarise
discussed further in section 5.6. the findings from the four countries of the
main study. Aggregated and summary tables
Even when parents are keen to send their only are presented in this section, although
s children to school they may be unable or comments are made based on the full data
unwilling to do so for a variety of reasons set. More details and other factors which
including but not limited to: lack of present barriers to access are outlined in
financial resources, safety issues, physical appropriate sections. The following section
e accessibility or need for childs household on quality provides further information on
d labour etc. These issues are explored in perceptions of quality of schooling and the
subsequent sections. quality factors which influence demand.
n Similarly, sections related to income, costs of
Parental attitudes on the value of and education and household sacrifices look at
benefits from education are not independent resource limitations of access.
. of other parents views. That is to say, one
persons views are influenced by others in 5.3.1 The Effect of Poor Quality as a
the family and/or community. For example Barrier to Education
in Nepal some parents said that girls Section 5.4 discusses quality issues in some
e are not supposed to go to school in our detail. This section emphasises the points
t community. Non-enrolment of girls in which effectively act as a barrier or a constraint
y some locations of Nepal was seen to be in to children attending school. The first point
conformity with the traditional social norm. to make is the high level of dissatisfaction
with the schools available to children. Even
5.3 Constraints to Access and when parents state that they would not change
Barriers to Education school, this is often not for the reason that
Amongst other things this study sought to they are satisfied with the current school,
explore lifestyle patterns which impact on a but that they cannot afford to go elsewhere,
childs likelihood of entering school, attending or there is no other school accessible.

D DFID 55
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 74

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Table 5.4: Would You Change School and Why? p

w
Would you change school? Present Want school For better Distance Costs
a
(female respondents) school with secondary learning related
Yes% No% not good grades
Nepal 18.3 (13.2) 81.1 (86.2) 71.8 13.6 2 E
Bangladesh 24.3 (26.2) 75.7 (73.8) 40 8.6 20 a
Uganda 32.9 (26.8) 66.1 (72.8) 51.3 1.8 13.6 0.7 f
Zambia 32.9 (35.3) 67.1 (64.7) 42.5 7.8 21.6 24 3 t
p
Source: SSI Q27, 28
s
t
Table 5.5: Why Would You Not Change School?
f
Would you change school? No better Cannot afford Loss of time t
(female respondents) School near extra costs for other chores c
Yes% No% b
Nepal 18.3 (13.2) 81.1 (86.2) 49 46 2.8
s
Bangladesh 24.3 19.2 30
r
Uganda 32.9 (26.8) 66.1 (72.8) 46.1 13.3 7.9
Zambia 32.9 (35.3) 67.1 (64.7) 54.8 26.9 1
D
Source: SSI Q29
i
p
Table 5.6: What are the Problems with the School?
b

Are there Punishment Teacher related* Too far, poor Cannot pay M
problems /strictness infrastructure and fees/afford m
Yes % TLMs lacking dress t
Nepal 34 5 26.9 45.6 25.8 w
Bangladesh 60.1 3.1 12 44 35 h
Uganda 59.2 6.9 34.5 57.5 4.1 m
Zambia** 65.1 12.1 23.2 29.3 w
* Untrained/no teachers/insufficient numbers/irregular/poor teaching i
** 33.4% of respondents provided reasons which fall into other categories. The researchers have not provided a
breakdown of what these included. Any one or all of the summary categories could be deflated because of this d
missing 33.4% a
r
The summarised information from the country infrastructure, water and sanitation, desks and 2
studies presented here on the problems that chairs, availability of teaching and learning w
respondents considered there to be in their local materials (text books, pens, paper) and d
school is supported and reiterated in more location of the school; money costs; and p
detail in the next section (5.4:2 perceptions of punishment) some clear patterns emerge. It
quality). When grouping information (all should be noted that this grouping is rather 5
points related to teachers and teaching; and arbitrary and using other combinations might A
all information relating to the physical produce a different emphasis e.g. including i

56 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 75

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

pens, paper and text books in the costs column reasons determine why some children never
will increase the significance of that group as go to school. The primary GER figures
a constraint to education. (Table 5.29) show that in the countries of
d
the study enrolment rates vary considerably
Even though reading the results with caution between countries and by gender. Our study
and taking into account information available locations in the poorest areas of the countries
from other sections of the questionnaire and have lower enrolment rates than the national
the PAGs it can generally be seen that most figures, but even these figures hide the poor
parents are dissatisfied with most aspects of attendance patterns and the limited
schooling, and costs impact on the decision likelihood of completion of the primary cycle.
to school children. The exception to the Later in this section, some of the reasons for
first point is Nepal, where only 34% said irregular attendance and dropping out of
there were problems with the school their school are explored.
children attended. As is noted below, the
bias in the Nepal sample (including only When asked why their children had never
school-attending childrens families) may be attended school, an unexpectedly high
resulting in these different results. percentage of respondents for primary school
children in Bangladesh, Uganda and Zambia
Disaggregated data from the SSIs show that claimed that their child was too young to
in Nepal and rather strangely, there are greater attend school. The researchers asked
problems with teachers attending on a regular subsidiary questions to this answer as this
basis in urban areas than in rural schools. question was only asked about school-going
More to be expected is the fact that costs are age children. Parents responded that in the
more significant a barrier to rural parents early years small children should not be
than urban ones. Given that the Nepal sample expected to walk so far that actually it was
was biased as only school-going childrens the distance to school that was the barrier,
households were included in the sample it not the age of the child. While this may well
might be anticipated that if households be so for the youngest children, it is also the
where children are not attending were case that parents are deferring school costs
included in the sample the results might for a period of time which may well be an
display even wider disparities between urban added incentive. The latter thought is
and rural. Certainly, this finding is not prompted by the fact that costs (direct and
replicated elsewhere, for example in Uganda indirect) are a significant reason for children
d 21% of respondents in rural areas said there never attending school. In Nepal one of the
was a problem with teachers attending each main reasons for never attending school is
day compared to 5.7% of respondents in cited to be the need for them to work at
peri-urban areas. home, which includes both productive and
reproductive activities, boys and girls, but
5.3.2 Reasons for Never Attending School greater for primary girls and for secondary
t As might be anticipated from the comments boys. This is replicated for dropping out of
in the previous paragraphs, these sets of school and being absent in the last month.

D DFID 57
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 76

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

The lower NERs in Nepal23 than the other be explained by the fact that so few children T
countries in the study demonstrate support are in school in the first place in comparison
for the suggestion that education is of less with the other countrys locations).
importance for households in Nepal than
meeting basic needs with the support of all When asking the reason for them not attending
members of the household. school, health and fee payment are the main
reasons given. Obviously, if the questionnaire
Further exploration of the data demonstrates was administered at a time just after the
that the problem of distance to school is of demand for school fees this would feature N
greater significance in some places than others, more prominently than at other times. o
and is a particular problem for secondary However, the fact that the result of having f
schools. Even though many countries have to pay fees removes children from school, b
policies about ensuring access to primary however temporarily, provides an important s
schools within certain distances or walking insight into the ability of families to pay for p
times there are clearly persistent problems schooling. The annual cycles which dominate
with school locations especially for the rural household lives mean that fees can more T
younger children and in more remote villages. readily be paid at some times and not others. c
The cycles also result in higher illness rates a
5.3.3 Other Reasons for Poor Attendance during some months, and lean times of the r
and Leaving School year when children may go hungry. In other i
Respondents were asked if children had been words, it is possible that the answers to these
absent from school at any time in the last questions may vary depending on the time
month, and then asked the reasons for absence. of the year they are asked, but irrespective of T

As can be seen, in all countries other than this, the high levels of sickness and the
Nepal, which is again an exception, nearly dominance of fee demands as reasons for
half of the children had been out of school failure to remain in school continuously are
at some point in the last month. (Nepal may worthy of note24.

Table 5.7: Out of School in the Last Month?

Has (your child) been out of Nepal Bangladesh Uganda Zambia

school during the last month?
Yes % 22.2 43.3 46.7 43.7
No % 76.9 56.7 52.4 56.3
Source: SSI Q22

23
1999 NER figures in Nepal were 72 in primary, 31 lower secondary (yrs 6-8), and 21 in secondary school. Girls are
fewer in number and represent about 40% of the intake at all levels aggregated across the whole country. Wide disparities
exist between regions and between the Terai and the hills/mountains and this aggregated data hides the very significant
educational disadvantage experienced in some locations. Although Table 5.30 on children from the study households
currently attending school (in equity section) does not clearly demonstrate significant differences between the Nepal study
and the other countries, the selection of households in the Nepal study only included those with school-going children
and therefore is not representative of the population as a whole.
24
Further disaggregation by level of education and household poverty level of the information in Table 5.8 can be found
in Table 5.17. S

58 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 77

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Table 5.8: Health and School Fees as Reasons for Non-attendance in the Last Month
(as % of All Reasons for Non-attendance).

Nepal Bangladesh Uganda Zambia

g Health 29.6 52.4 55.4 9.8

School fees due 1.6 2.1 25 62.9
e Assisting household 47.2 4.3 6.7 0.8
Source: SSI Q23

Note that helping the household with domestic Taking these sets of information together
or agricultural work was the dominant reason provides a picture of highly unsatisfactory
for Nepali children not to attend school school quality, poor households finding it
but, as stated, fewer children in the Nepal difficult to sustain payments (or start payments)
survey interrupted their schooling in the for schooling, and illness impacting on
previous month. regularity of attendance. In such a situation
e it would seem logical for parents to consider
e The reasons for the permanent removal of educating their children through the formal
. children from school were also explored, and education system as an unnecessary income-
a similar set of answers was given to those in draining act. They gain little from education,
response to the question on lack of attendance children themselves are uncomfortable in the
r in the last month. often violent and uncomfortable surroundings
e of the school, and their time can be spent

f Table 5.9: What is the Main Reason for Pupils to Leave School?

Nepal Bangladesh Uganda Zambia

Lack of money to pay school expenses 14.9 61.9 48.1 55.3
Didnt want to continue 24.1 11.5 6.8 14.4
Difficulties with school work - 3.5 0.8 -
Expelled from school - 0.9 0.8 0.5
Earning money 13.8 2.7 0.8 -
Needed to work at home 13.7 5.3 0.8 -
Illness 4.6 3.5 9.8 2.9
Marriage 8 - 7.5 1
Pregnancy - - 9.8 1.9
Death in the family 1.1 0.9 2.3 1
Failed a grade and would have to repeat 4.6 2.7 1.5 11.5
Too old to go to school 1.1 - 1.5 -
Completed schooling (primary and secondary) 5.7 - 4.5 3.4
Transfer - - 0.8 -
y
Withdrawn by parent/guardian 3.4 - 0.8 -
(corporal punishment)
Source: SSI Q19

D DFID 59
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 78

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

valuably assisting the household with Subsidiary information from talking to e

reproductive or productive activities. children in Uganda and Zambia provides a f
different insight into the situation in these c
5.3.4 Impact of Social Reproductive two countries. Both boys and girls say that q
Activities on Education Access girls are the most involved and that it does w
Although it seems from the SSIs that everyday have a considerable impact on their schooling. m
social reproductive activities do not have a large It makes them late for school, which then u
impact on schooling, this is not wholly confirmed results in punishment which then puts them h
by the complementary qualitative data. off going to school. The next section on c
quality issues (5.4) demonstrates the r
Respondents to the SSIs were asked if problems faced by students resulting from h
collecting water affected the ability of the violence in the school: punishment is often d
child in their household to go to school, and severe and considerably out of proportion to
the same question about the collection of fuel. the misdemeanour, and girls are exposed to N
The vast majority of parents who were sexual violence from male teachers. Whether c
respondents to the SSIs claimed that neither parents are unaware of the impact that r
water or fuel collection had much impact on spending time and energy on fuel and water f
schooling; the results from these questions collection has on their childrens schooling r
can be seen in the following tables (5.10 and or consider it less important than do the
5.11). In the two African countries more children themselves, it does seem to be an W
than 10 % of children were affected, which is issue which would do well to be raised when m
a greater impact than in Bangladesh and considering the removal of barriers to education. a
slightly more than in Nepal. Being late for e
school, having less time for studying at home Other examples from the study were found w
or doing homework and, in the Nepal case of parents giving different responses to those l
particularly, missing school altogether were of their children; when asked if they had a d
consequences of having to contribute in this preference for educating their sons or t
way to the household daily domestic duties. daughters their responses did not reflect the i

Table 5.10: The Effect of Water Collection on Childrens Education T

Water Collection Late for Less time for home Missing Walking long
affects schooling School studies/fetches school distance
water after school
Yes % Yes % Yes % Yes % Yes %
M F M F M F M F M F
Nepal 6 6 75 56 - 11 25 33
Bangladesh 1.2 0.7 N/A N/A N/A N/A N/A N/A N/A N/A
Uganda 11 9 12.5 31 62.5 46
Zambia 8 12 22.2 14.6 1.9 6.3 14.8 29.2
M: Male respondents F: Female respondents M
Source: SSI Q97 and 98 S

60 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 79

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

experiences of sons and daughters. Bias in is the SBO households children who are
favour of boys was claimed not to be the more affected by water collection. In Zambia,
case by parents, but when they were their responses indicate that although the
questioned both boys and girls said that sons poorest children are more likely to be out of
were given preference if choices had to be school, it is the SBO child who is more likely
g. made. Barriers to girls education from to be involved in water and fuel collection.
unconscious (or unacknowledged) There is no obvious explanation for this
household bias in favour of boys clearly counter-intuitive response pattern.
combines with other household gendered
resource and task distributions and Even more than with water collection,
household preferences to create educational generally fuel collection affects poorest
disadvantage for girls. households more than SBOs, rural more than
o urban, and some locations more than others.
Note that information on the effect of water In Uganda it is the SBO households who
r collection on specific aspects of schooling have a slightly higher percentage affected
relied on the questioner using probe questions than the poorest households.
following the initial yes/no question. Not all
researchers produced this information. 5.3.5 Summary
Although there are wide variations in the level
Water collection affects poorest households of impact of particular factors on entry to
n more than SBOs, rural more than urban, and attendance in school, in general for the
. and some locations more than others. The poorest household, schooling all their children
exceptions to this pattern are Bangladesh is a difficult and dispensable proposition.
where both water and fuel collection is even Limited formal employment possibilities
e less significant than elsewhere, but where it discourage children, especially boys, from
does influence schooling it is the boy rather wanting to continue in schools which offer
than the girl who is more likely to be very little of satisfaction and little chance of
involved; and Uganda and Zambia where it academic achievement. Girls are equally

Table 5.11: The Effect of Fuel Collection on Childrens Education

Fuel Collection Late for School Less time for home Missing school
affects schooling studies/collects
after school
Yes % Yes % Yes % Yes %
M F M F M F M F
Nepal 6 8 25 25 - - 75 75
Bangladesh 0.6 none N/A N/A N/A N/A N/A N/A
Uganda 7.6 6.0 100 100
Zambia 15 17.3 33.3 17.0 1.9 5.7
M: Male respondents F: Female respondents
Source: SSI Q101 and 102

D DFID 61
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 80

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

discouraged but for a range of different of many well-meaning publicity campaigns). E

reasons, more related to domestic Indeed, this study confirms that, although T
reproductive factors than productive, and poverty is one of the main influences on b
determined in some part by the dominant demand, other important factors also come
socio-cultural norms of their communities. into play, including the quality of schooling. a
Even where children do overcome the physical t
and economic barriers to entry into school, 5.4.2 Perceptions of Quality 5
social and schooling system factors impact Perceptions of quality are, of necessity, i
on their ability or desire to continue. The subjective. This study brings no surprises S
specific factors which dominate in any one when it reveals that teachers, parents and a
location will vary from any other location, pupils all have different views on what f
within the same country and between constitutes quality in schooling. These h
countries. Time also impacts on the differences in perceptions reflect the study s
dominance of particular factors, a feature participants own experiences. For example, S
most notable in subsistence-agriculture many of the children (but not all) would f
households who rely on seasonally available view a good school as one where teachers q
resources. However, in all places and at all did not beat or humiliate them in classrooms t
times the poorest and the most vulnerable this reflects the fact they themselves face s
have the least robust demand for schooling this problem. Similarly, cultural and historical m
and this is particularly influenced by factors related to other social stratifiers p
opportunity costs: both direct and indirect influence perceptions of acceptability, and t
monetary and non-monetary costs. what are desirable qualitative features of i
schooling. a
5.4 Quality of Education r
Our study attempted to probe the issue of
T
5.4.1 Introduction schooling quality and its effect on household
In a previous Section (5.2) on the value and decision making by asking the following
benefits of education it was suggested that questions:
parents and children in poor households do
value education, and they understand the Do good schools exist in the study areas,
economic and non-economic benefits derived and if so why are they perceived to be good?
from it. However, the attendance patterns of
the study sample demonstrate that a significant If good schools do not exist, how do the
number of children (particularly at the participants perceive quality ?
secondary level) do not attend school regularly, S
have never attended or have dropped out. If the quality of schooling were better 2

Therefore, simply valuing education or being would parents be encouraged to send t

a
convinced of the benefits of education will more or all of their children to school? 2

not necessarily lead to demand for schooling p

a
though it may influence it (hence, the failure f
c
2

62 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 81

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Existence of Good Quality Schools difficulty of providing services to more

Table 5.12 shows those respondents who remote areas with scattered populations, to
believe there to be a good school in their area25 the unwillingness of teachers to live in
when the figures generated for all respondents isolated regions, and to demands for child
are looked at it appears that over half consider labour. Public spending priorities that
there to be a good school in the area (from concentrate resources on urban areas add to
53.3% in Nepal to 62.1% in Zambia). However, the problems of rural people (p.15327).
if the analysis by type of household (PO and
SBO) and location (rural/urban) is looked More interesting differences are revealed when
at some interesting differences emerge. Apart each study location is looked at separately by
from Nepal, a greater percentage of the poorest country as shown in Figure 5.3. In Nepal the
households believe there to be good quality greatest variation was found between areas
schools in the area than the SBO households. with only 6.7% of Kapilbastu respondents
Similarly, a greater percentage of the respondents stating that there is a good quality school
from urban areas believe there to be good compared with 74.7% in Kailali. According
quality schools in their area than of those in to the Nepal Country report Kapilbastu
s the rural areas (apart from Nepal where the lags behind the national average in educational
situation is reversed26 ). This is what one achievement (p.28). Out of about 4,000
l might expect given the disparity in the children of school age, only about 600 children
provision of education in rural and urban areas are in school. The NER is less than 16 %.
throughout the developing world. As noted Girls education is extremely neglected(p.29)
in the Oxfam Education Report (2000) there
are a variety of reasons for the disparities The Nepal country report argues that there
ranging from the administrative cost and may be better parental awareness of education

Table 5.12: Existence of Good Quality School

d

Is/are there good Nepal % Bangladesh % Uganda % Zambia %

quality schools in
the area Yes No Dont Yes No Yes No Yes No
know
, All respondents 53.3 46 0.7 59.3 40.7 61 39 62.1 37.9
? Poorest 51.4 47.7 0.9 60.3 39.7 63.9 36.1 68.8 31.2
Slightly better off 59 41 56.6 43.4 53.1 46.9 50.5 49.5
Rural 68.9 31.1 56 44 52.4 47.6 57.1 42.9
Urban 6.7 90.7 2.7 69.3 30.7 86.7 13.3 79 21
Source: SSI Q159
25
It is important to note that although respondents may believe there to be good schools in their area it does not follow
that their children attend these schools, for a variety of reasons. The main reason is cost, but distance and eligibility are
also factors.
26
The urban area selected in the Nepal study (Kapilbastu) is not typical of an urban area, where resources and education
provision would generally be expected to be better. Kapilbastu district lags behind the national average in educational
achievement, and in the education resources available. It also demonstrates a considerable gender gap in provision (18%
female teachers in primary level and 6% at secondary), enrolment (33% female students) and outcomes (13.6% literacy
compared to an overall 28% in the district).
27
See Oxfam report p153-154 for examples of educational inequalities in developing countries

D DFID 63
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 82

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

issues in hill communities than in the Terai In Uganda just over a third of the respondents T
areas (Kapilbastu is a Terai area) and that many from Kibaale thought that there were good
of the respondents from Kapilbastu belong quality schools in the area. It is worth noting
to disadvantaged caste groups (p.47). As that Kibaale is the second poorest district in
discussed in the section on Equity disadvantaged Uganda, with an income index of 0.0796,
groups in Nepal are economically vulnerable, compared to the national average of 0.2098.
they are socially excluded, and some are socially At the same time, the district has only 9
prohibited from entering schools. Kapilbastus government aided secondary schools and
response to the issue of availability of good 212 primary schools thus providing limited
schools is therefore a reflection of their deep access to education. (Ugandan Country
discontent with the schools in the area. Report p.24). With such poor provision of
secondary schools it is no wonder that over

Figure 5.3: Existence of Good Quality Schools

Nepal Bangladesh
74.7

100%
90.7

80 %
69.3

S
82.7

68
62.7

6
69.3

80 %
60 %

46.7
53.3
w
60 %
40 %

30.7
37.3

T
30.7

40 %
32
25.3

17.3

20 % K
20 %
6.7

o
T
YES NO YES NO
a
Achham Kapilbastu Mymensing Patuakhali
Kailali Rasuwa Dhaka Dinajpur
W
T
Uganda Zambia
q
78.1
79

74.3

100 % 80 %
86.7

s
48

67.2

80 % w
60 %
68

s
60 %
c
52

32.8

40 %
13.3
37.3

t
62.7

25.7

40 %
21.9

A
21
32

20 %
20 %

i
YES NO YES NO t
Kibaale Apac Kitwe Siavonga
Kiboga Mbale Chinsali Kalabo
T
Source: SSI Q159 w

64 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 83

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

s Table 5.13: Characteristics of a Good Quality School (%)

Characteristics Nepal Bangladesh Uganda Zambia

g
n Experienced, motivated & trained teachers 12.7 23.3 31.4 21.3
Teachers attend classes every day 14.1 9.6 2.2 7.5

. School located near to home 16.9 15.1 2.2 7.5

School building in good condition 2.8 2.7 10.3 14.9
Latrines & water available 1.4 5.5 2.4
Provision of desks and chairs 1.4 11 0.5 4
Provision of textbooks 5.5 1.1 4.8
No corporal punishment 1.4 1.6 7.5
Good discipline and morals taught 4.3
Good performance by children 29.6 19.2 29.3 9.5
Secondary grades offered 11.3
Small class size and reputed school 5.6 5.5 3.6 4.8
Other 5.6 5.3 8 15.8
Source: SSI Q160

60% respondents in Kibaale thought there teachers in the school, teachers were concerned
were no quality schools in the area. with lack of community commitment to
helping the school: a demonstration of the
The qualitative results from Sri Lanka and persons or the groups role in respect of the
Kenya also demonstrate that there is a lack school influencing their perceptions of quality.
of good quality schools in the areas studied. In Uganda the researchers noted that the
The differences in the perceptions of quality children included in their study had the same
are discussed next. concerns as the other groups, but they added
textbook availability to their quality indicators.
What is Quality Schooling?
To probe the issue of quality schooling, There would be no value therefore in trying
questions were asked whether the current to come up with one definition of quality to
schooling situation was satisfactory and also fit all it would not be possible. However,
what the ideal school would be. The two the various perceptions of quality can be
sets of answers interlink, in that what is demonstrated, starting by looking at the SSI
currently highly unsatisfactory (such as having results which concentrate on the views of
to rely on unqualified volunteer teachers in parents.
Andimune school in Sri Lanka described as
stooges of the school principal) colours what The above table shows that in all countries
is desirable (hence appointed and qualified except Nepal, a key concern is teachers not
teachers was the parental desire in Andimune). only their experience, motivation and training,
but also their regular attendance at school.
To continue using the Sri Lanka example, Teachers attending school each day is identified
whereas the parents were concerned about by respondents as a quality indicator, either

D DFID 65
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 84

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

overtly stating that this is not the current teachers attend class every day (66.7%) and a
situation, or implying it by the high priority teachers are experienced (33.3%). In m
afforded this in their description of desirable contrast the other study sites in Nepal give o
qualities of a good school. Good performance good performance by children as the most
in school is the most frequently cited response important factor of a good quality school. In T
in Nepal this also receives a high percentage Uganda, almost half of the responses received m
of responses in Bangladesh and Uganda (46.2%) in Mbale were good performance g
though this is also intimately related to the by children. Mbale is an urban area where w
availability of teachers. access to good jobs may require qualifications t
(more so than in rural areas) this may o
Of interest are the differences between the explain the respondents desire for good e
communities for example in Nepal, Kapilbastu performance. In addition, it is often more i
gives only two responses in relation to what attractive for teachers to work in urban areas a
constitutes a good quality school. They are than rural areas, therefore the problems w
w
Corporal Punishment
i
Sometimes the older children (girls aged 11 Now it (corporal punishment) has become very
b
and 12) from the household go to school late bad. Our children refuse to go to school. You may
m
because they have to prepare food in the think they are going to school but they have
h
morning to take to school. When they go late, gone somewhere else. At school they are learning
n
teachers punish them by making them walk on to be beaten.
p
their knees for about 30 metres. One day these Mother, Chinsali, Zambia
o
children were given the same punishment and
returned home with bleeding knees. They stayed Boys in Kapsinendet (Kenya) pointed out the
T
out of school for one week. The mother reported issue of good teacher/pupil relationship, again
q
the matter to the headmaster and the concerned pointing to the issue of discipline which
g
teachers had to foot the bill for medical treatment. contributes to some boys dropping out of school.
(
Therefore, when the children are late nowadays Researchers observation, Kenya
i
they dont go to school, as they are frightened.
m
Mother, Kibaale, Uganda Many examples were given of inappropriate
e
behaviour in one instance, a teacher cut up
h
The problem of corporal punishment affects girls a pair of tennis shoes worn by a student because
c
more than boys, not only because they come to he was not wearing the required black shoes.
p
school late more often because of household duties Volunteer teachers are said to threaten
p
but also because teachers seem more interested students when they delay payment of monthly
5
in boys education than girls. fees. Teachers were also said to conduct love
a
Secondary School boys, Kibaale, Uganda affairs with female students, and some parents
r
expressed their fear of sending girls to school
r
Once late, they (teachers) beat. after they attain puberty.
i
School girl, Kapoto, Zambia Researchers observation, Andimune,
r
Sri Lanka
p
Source: Qualitative Fieldwork

66 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 85

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

d associated with teachers in the rural areas themselves experience first hand corporal
may not be experienced in the urban areas, punishment which they may not report to
or at least they may not be as pronounced. their parents. Examples related to corporal
punishment from children, parents and
There were also differences in the views of researchers, are provided on the previous page
d men and women. For example, women as a
general rule were more concerned with In addition to corporal punishment sexual
welfare aspects of schooling (the attitude of harassment was also raised as a problem in
s the teacher towards students, the provision Zambia, Uganda, Kenya and Sri Lanka.
of food, a secure and pleasant school Community leaders in Kibaale (Uganda)
environment) than men who expressed greater were aware of the problem and said how
interest in management and organisation issues, can the male teachers teach when they spend
s and teacher qualifications. Both were concerned their time harassing our girls? The issue was
with having good discipline. Boys and girls also raised by parents and teachers in Kibaale
while also agreeing with some of the same although their perceptions on the situation
issues, seemed to put greater emphasis on were quite different (see below).
being able to learn (teaching and learning Sexual Harassment
y
materials and a teacher in the classroom), on The teachers are not sympathetic, if they were,
having extra- or co-curricular activities, and a they wouldnt be impregnating our girls, and
g
non-violent environment (no corporal female teachers would not be seducing these
punishment). All agreed that the condition girls for the male teachers. And they would be
of the school buildings was important. teaching the children daily. Refusal of the girls
to fall in love with teachers results into
The results from the PAGs work support the harassment thus forcing them to drop out
quantitative findings. These do, though, offer of school.
greater insights as more groups were involved Mothers and fathers, Kibaale, Uganda

(parents, children and teachers) and the parents

in Zambia, Uganda, and Sri Lanka seemed Some parents educate their girls not because
more willing to talk about certain issues, for they have a positive attitude towards girls
example, corporal punishment and sexual education but as a way of extorting money
harassment. The groups of children did not from men who impregnate their daughters in
e
contradict the views of their parents, but they addition to compelling those responsible for the
put more emphasis on corporal punishment, pregnancies to marry the girls.
particularly in Uganda and Zambia. In Table Secondary Teachers, Kibaale, Uganda
5.13 parents include no corporal punishment Source: Qualitative Fieldwork
as a factor of good quality schools but this
received a relatively small percentage of In Zambia secondary school girls complained
responses (from 1.4% in Bangladesh to 7.5% that teachers treated them very poorly and
in Zambia). Again these differences in emphasis gave more attention to boys. However at
reflect the individual experiences of the times the male teachers flirted with them.
participants that is to say that children The Zambian researchers observed that

D DFID 67
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 86

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

teachers did not seem to be aware of the locations which reflect disparities in the A
consequences of their flirting. schools themselves: in Kapsinendet the f
supply and availability of books and school a
In summary, teachers attitudes towards their furniture emerged as greater concerns than o
pupils are also a significant factor influencing in Mandani. Moreover, some unique problems e
parents demand for their childrens education, facing girls who do not have equal access to t
and childrens willingness to continue in the school were revealed in Kapsinendet q
school. Bullying, corporal punishment, and where socio-cultural reasons see more girls e
sexual abuse were all cited as features of than boys dropping out of school.
current schooling practice which are D
unacceptable. Boys and girls are susceptible In Sri Lanka too the quality and quantity of
to bullying and corporal punishment and teachers was raised as an important issue. As t
indicate this as a reason to avoid school; girls one villager (Rohanapura) succinctly put it c
are generally more commonly exposed to teachers to our schools come inside d
sexual abuse by teachers. Schools as centres envelopes teachers are officially appointed s
of violence is a particularly disturbing finding by the Education Department and the o
of this study, and a finding which indicates school principal is notified by post. Some of o
the poor learning environment of many schools. these teachers do not turn up for work, a
The impact on the childs willingness to again illustrating the problem of teacher m
attend school, boys directly and girls because provision in remote/rural areas. a
of the unsympathetic school environment, P
will influence the parents willingness to pay 5.4.3 Quality of Education as a d
for schooling. Determinant of Demand T
The forgoing discussion indicates that the a
Apart from the issue of corporal punishment quality of education is an important determinant e
and sexual harassment, the PA findings from of demand. Quality is variously and individually F
the groups of parents concurred, on the whole, defined, but common concerns with non s
with those of the SSI findings and as in the economic aspects of demand such as teachers c
SSI findings, there were differences between attendance, conditions for study and a d
the different study locations in each country. supportive atmosphere shine through all the
In Sri Lanka and Kenya where no SSIs were studies. It is notable that relevance of education I
administered, similar quality indicators were did not feature strongly in the responses on c
discussed to those in Nepal, Bangladesh, quality, but such concerns were a feature of f
Zambia and Uganda. the reasons why parents would not continue t
the schooling of their children to higher g
In Kenya, for example, in-school issues levels or under increased cost. b
about the quality and quantity of teaching i
being provided were of concern as was the It is worth remarking that, in so-called f
environment within which the children were disadvantaged communities (e.g. in Nepal) it h
required to learn. However, there were is the nature (and language) of the education s
differences found between the two study which influences demand for education. p

68 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 87

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

As well as the social exclusion of these children desirable qualitative features of schooling,
from state schools, there are factors which and hence what or if parents are willing to
affect parents (and childrens) perceptions bear costs.
of qualitative aspects of the government
s education system and whether these are what This is not to deny the influence of poverty
they want (and will demand). The social and (as a broad definition, not simply limited to
qualitative factors add to (or over-ride) the income) on education. Even using the
economic determinants. narrow definition of income and
consumption poverty, disparities in access to
Drze and Sen make a similar case in India: schooling and educational achievement are
The point is that the willingness of parents evident. The World Development Report,
to bear these costs, such as they are, and to 2001, p.27: In some poor countries most
coax their children into going to school may children from the poorest households have no
depend crucially on the quality of the schooling schooling at all. A study of Demographic and
d services they obtain in return. The quality Health Survey data found 12 countries in which
of schooling affects not only the motivation more than half the 15 to 19 year-olds in the
of parents, but also that of children. (Drze poorest 40 % of households had zero years of
and Sen, 1998). In their study they show a schooling: Bangladesh, India, Morocco and
marked contrast between educational Pakistan, and eight countries in sub-Saharan
achievement in Kerala compared to Uttar Africa. In contrast, the median number of years
Pradesh: this is not solely the result of completed by 15 to 19 year-olds in the richest 20
differences in income as this is not significant. % of households was 10 in India and 8 in
The particularly obvious feature of Keralas Morocco. In other countries the gap in
achievements, in population, health and in educational achievement was much smaller: one
t education, is the gender equality of outcomes. year in Kenya, two in Ghana and Tanzania,
y Female autonomy increases their demand for and three in Indonesia and Uganda
social services for themselves and their Primary enrolment rates show similar gaps. The
children; their entitlement to express those enrolment rate for 6 to 14 year-olds is 52
demands publicly ensures continuing access. percentage points lower for the poorest households
than for the richest in Senegal [and] 36
n In other words, it is not poverty that is percentage points lower in Zambia. Four of
creating the gender differences in demand the six countries included in our cost sharing
for education elsewhere than Kerala, even study are cited in this quote from the World
though poverty may influence how many Bank. However, the report continues to
girls get to and stay at school compared to demonstrate within country differences
boys. The differences are not based on between urban and rural areas and between
income but on social, cultural and historical males and females. Thus this report
factors. In a similar vein, social, cultural and reinforces the point that income alone is not
t historical factors related to other social the sole determinant of effective demand for
n stratifiers (ethnicity, class, caste) influence education. The range of educational
perceptions of acceptability, and what are difference does not correlate exactly with

D DFID 69
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 88

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

levels of income and income distribution. 5.5 Income and Expenditure28 Z

Poverty influences but does not cause lack of In common with many other household (
demand for education. surveys of income and expenditure, researchers (
in all the sites found it difficult to extract h
The reason to emphasise this point is that the information that was accurate and reliable.
country studies on cost sharing demonstrate Suspicion of what the data will be used for I
that parents express willingness to pay for or whether it will find its way to the tax s
education, and make considerable sacrifices authorities leads many to be economical in e
in order to do so. This willingness is, however, their responses. The findings cited below a
qualified: certain changes need to take place must, therefore, be treated with some w
especially in the context of rising charges. caution. The major sources of income for b
The quote from Drze and Sen above on the Uganda, Bangladesh, Zambia and Nepal are g
influence of quality on demand is supported in given in Table 5.14 below. a
the studies. As noted above, in all the studies a
and in every location a large number of There is a marked contrast between the h
parents and pupils noted the quality of countries with other activities being most s
schooling to be of an unacceptable standard. important for Uganda and wages and salaries h
It is this which influences whether parents accounting for more than three quarters of h
wish to continue to pay (directly and indirectly) sources of income in Bangladesh and Nepal,
for schooling, and whether children want to but only 26% and 31% respectively for Uganda T
stay on at school. In Zambia, despite and Zambia. It is unfortunate that other (
conducting the study in the four poorest activities was not broken down further for
areas of the country, the researchers noted: Uganda as it accounts for 57% of the sources T
ability and willingness to pay for of income. However, it may be inferred from
education were causally inter-related. When the responses given by household heads on
parents see the value in education, when they see their main production activities (see Table
good quality education like good physical 4.11) that work in the informal sector (as a
infrastructure, good quality teaching/learning street vendor, selling labour, beer brewing or
they somehow find the money to pay. hawker) may account for some of this. In

Table 5.14: Sources of Income (%)

Income source (all households) Nepal Bangladesh Uganda Zambia

Wages 78 77 26 31
Agricultural/Fishing 10 9 12 39
Other activities 7 12 57 24
Gifts 0 1 3 0
Income in kind 5 1 2 6
Total 100 100 100 100
Source: SSI Q116 - 123 S
28
In both the tables for household income and expenditure the percentages shown have been calculated by using the
aggregate household figures. 2

70 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 89

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Zambia the other activities include trading fact that our respondents were from the poorest
(66%), piecework (17%), informal business HHs and usually in the poorest areas. The
s (6.4%), beer brewing (6.4%), traditional percentage of expenditure allocated to
healing (2.2%) and washing clothes (2%). education ranges from 6.5% to 15.2%, though
these figures do not include all of the indirect
It is notable that despite the fact that all the costs associated with schooling and will vary
sites surveyed are very poor developing with the number of school age children in
economies, fishing and agriculture appear to the family who are attending school29. Only
account for only around 10 % of HH incomes, in Nepal is expenditure on clothing above
with the exception of Zambia. This may be 10%, presumably reflecting the fact that
because the question concerning income climatic conditions are so different there.
generation was interpreted as money income
and did not take account of livestock, fish or 5.5.1 Expenditures on Education
agricultural products consumed by the as a Proportion of Household
household. (Section 4 indicates that in all Expenditures
sites except Bangladesh agriculture/animal The proportion of household expenditures
s husbandry was the majority occupation of devoted to education costs is one indicator
household heads). of the economic burden falling on households
and the sacrifices made to continue schooling
a The major expenditure reported by households for children. The figures given in Table 5.15
(Table 5.15 below) is on food, reflecting the come from the SSIs. In most cases these

s Table 5.15: Household Expenditure Pattern (%)

m
Expenditure Item Uganda Bangladesh Zambia Nepal
Food 37.1 65.3 62.9 41.2
Fuel/electricity 5.8 4.1 1.8 3.6
Housing 1.7 5.7 1.4 0.4
Health 9.1 4.4 1.9 6.2
Clothing 6.1 5 4.3 13.2
Education 15.2 6.5 10.6 8.2
Festival/entertainment 9.8 3.7 4.3 6.1
Alcohol/tabacco 6.1 3.2 2.7 6.3
Fertilisers etc. 0.5 0.1 5.8
Water 2.8 0.4
Loan repayment 0.9 1.1 4.7
Hired labour/land/pond/oxen 1 0.2 2.2
Other 4.4 1 6.7 3.6
Total 100 100 100 100
Source: SSI Q107 - 108

29
It was not possible to quantify all indirect costs. This is discussed further in the next section on the costs of schooling.

D DFID 71
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 90

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

figures were supported both by more detailed (but still probably underestimated) picture g
questioning on household expenditures on of the real burden of direct and indirect f
education and by PAGs where education costs on poor families, especially when it is i
regularly figured as a significant cost in remembered that these proportions are of S
household expenditures. very small amounts of money. Unfortunately, b
the figures do not distinguish between H
For example, in Uganda PAGs ranked education primary and secondary education so little d
spending on average the second or third major can be said about the costs of free education. s
item of household expenditure, in Bangladesh e
PAGs ranked education second or third on 5.5.2 Income and Expenditure by Country t
average although it was estimated at 14% of w
household expenditure. The reasons for such (i) Nepal (excluding Rasuwa) 31
r
a discrepancy may be a tendency to think The annual average per capita income was
only of direct costs when asked about Rs 3,462, with the average for the PO I
education costs within the context of households being Rs3,289 and for SBO t
household expenses or lack of appreciation HHs Rs 3,986. There was considerable p
of the true costs of education when all variation between regions with annual per a
components are taken into account. capita incomes in Achham, Kailali, Kapilbastu 5
Conversely, rough guesswork may be leading being respectively 3,656, 3,119 and 3,571. c
to overestimation in the case of Bangladesh. There was a strong negative correlation f
between per capita income and the size of i
Of more interest are the figures for the the HH. Where the HH size was 3 the i
proportion of discretionary expenditures annual per capita income was 7,239, with a s
used for education as shown in Figure 5.4 7-9 person household the annual per capita s
below30. Such figures give a more accurate income was 3,336 and where HH size was P

Figure 5.4: Proportion of Household Expenditure Spent on Education s

a
t
100 % t
83

80 %
68

67
67

(
60 %
T
40 % c
33

33
32

t
17

20 %
e
f
Nepal Bangladesh Uganda Zambia

Proportion of household discretionary expenditure spent on education T

Discretionary household expenditure

30
Discretionary expenditure has been calculated by subtracting household expenditure on food, health, fuel, water and

shelter from total household expenditure.

*
31
See footnote 16 for explanation

72 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 91

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

greater than 12 the annual per capita income The income index for the country, which is
fell to 2291. For all levels of income HH based on average income in the country
incomes of the PO were below those of the compared to a world average income of
SBO. This indicates that our sample had US$5,900, was 0.2098. The income index
, been correctly selected to reflect the poor for the sites selected are given below in
HHs and shows why poor HHs found it so Table 5.16
difficult to pay the costs of their childrens
. schooling. It is no surprise to find that HH There was a significant disparity in the sources
expenditure was closely related to income; of income as already shown in Table 5.14,
y the annual per capita HH expenditure average with significant differences across regions
was Rs 2,953 and for the PO and SBO HHs and between PO and SBO households. For
respectively it was Rs2,853 and 3,334. example, the annual average household
income in Kiboga was Ush 298,088 and in
If the pattern of spending on education is Mbale it was Ush 959,135. The annual
turned to, despite the Governments declared average household income of SBOs was Ush
policy of providing free education (for primary 200,000 greater than that of the PO HHs.
and lower secondary) HHs spent between Rs
u 569 and 751 (mean 660) on a primary school When expenditure patterns are examined, again
child and Rs 2,019 to 2,185 (mean 2102), there is a pronounced difference between the
for a secondary child per year. If that spending regions. In Kibaale 17.8 % of expenditure is
is compared to the income of the poorest HHs on food and 20% on entertainment whereas in
it amounts to 20% of income per primary Mbale the percentages are respectively 54% and
school child and 64% of income per secondary 3.4%. If HH spending on education is turned
school child. It is no surprise that both the to regional and gender differences are found.
PO and SBO HHs send very few children to These are shown in Table 5.18 in section 5.6.
secondary school and that for large HHs not
all children are sent to school and, if they are, (iii) Zambia
they attend irregularly and do not complete 73% of the people of Zambia are poor, with
the primary cycle. poverty being described in 1996 as people
whose incomes were below US$12 per month.
(ii) Uganda Although the response rate for incomes was
The areas selected for the study were low there was a very large range of incomes
characterised by extreme levels of poverty: from respondents, with annual salaries
the majority of the population live below the ranging from K5,000 to K840,000 (US$233).
established poverty line of US$7.5 per month The major source of income was other
for the poor and US$3.75 for the poorest. activities as previously mentioned. There

Table 5.16: Uganda: Income Index

Central reg. Northern reg. Eastern reg. Mbale mun. *Western reg.
0.3360 0.1260 0.1640 0.1889 0.1980
* Although a slightly better off region, it has in Kibaale the second poorest district in the country with an idex of 0.0796.

D DFID 73
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 92

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

were differences in the income from these Within our sample the per capita income of s
activities with Kitwe residents receiving the PO HHs and the SBO HHs was US$36 e
more, K450,000 per household per year and US$40 respectively, considerably lower a
(from this activity) than other case study than the national per capita income of US$294. o
communities. The expenditure patterns were In percentage terms the PO and SBO HHs p
similar to those in our other countries with were receiving an income of 12.2% and c
food being the dominant item of spending. 13.6% respectively of the national per capital 5
income. The total average annual expenditure f
Table 5.18 in Section 5.6 provides details of was estimated to be US$53 for the PO HHs h
spending on education for PO and SBO and US$51 for the SBO HHs. This seems g
households in the different regions. counter-intuitive, but no explanation was t
Although this spending can be supported by offered by the research team. b
the richest households, it cannot be e
supported by lower income HHs. Even for (v) Kenya I
the richest households amongst the SBO Using monthly expenditure as a proxy the j
group, those with incomes of K840,000 average household income per month was w
would find it difficult to support more than estimated at Ksh700 in Mandani and Ksh T
one child in secondary school, given the 500 in Kapsinendet. However, these amounts s
other expenses that the families have. represent the times of year when households w
are generally better off in times of great t
(iv) Bangladesh hardship incomes were said to be as low as
The annual per capita income in Bangladesh Ksh100 (US$1.25) in Mandani and Ksh350 (
is the lowest of all the countries studied, (US$5) in Kapsinendet. T
being US$294 per annum in 1999 . Half 32
p
the population is living below the The expenditure patterns for both sites in L
internationally accepted definition of poverty. Kenya are shown in Figure 5.5. below It can a
s
a
Figure 5.5: Expenditure/Income Patterns
w
Kapsinendet Mandani
i
t
a
250 Ksh 350 Ksh p
s
110 Ksh 150 Ksh
o
125 Ksh d
180 Ksh
20 Ksh e
15 Ksh
u
Food Clothing t
Health Education
s
32
Bangladesh Bureau of Statistics (BBS), 1999 c

74 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 93

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

seen that there is little available for education 5.5.3 Summary

6 expenditure despite the fact that when asked The data on income and expenditure is very
about the costs of schooling it became variable in quality and it is difficult to draw
. obvious that there were multiple mandatory any but the most qualified implications from
payments (discussed further in the school it. The data on HH income is probably the
costs section). It is also shown that nearly least reliable because of the unwillingness of
50% of household income is used to purchase interviewees to reveal their true incomes. A
e food, and of the remainder of income available further problem concerns the question of
s health and clothing expenditures are of how much income is paid in kind,
greater priority than education. To compound particularly in rural economies where the
this bad situation the seasonal calendars for exchange system is not as sophisticated,
both villages show that income and monetarised, as in the urban settings. What
expenditure patterns are not correlated. is very clear is that in most sites there was a
Income is available at the time of harvest and difference between POs and SBOs in their
just after (peaking between July and September), incomes and ability to pay the cost of their
whereas the school calendar starts in January. childrens education. For all groups in our
The net result of the income constraint and sample the costs of education were a
s seasonal labour demands present parents significant burden on the HHs and paying for
s with the need to make tough choices over education would involve significant sacrifices,
the schooling of their children. as is shown elsewhere in the report.

(vi) Sri Lanka Where the decision of HHs is to only school

The information gathered through the selected children (discussed elsewhere), it is
participatory qualitative methods used in Sri because the sacrifices that HHs would
Lanka did not include specific details on income otherwise have to make to their nutrition,
n and expenditure. However, the sample housing, clothing etc. are deemed too great.
selected involved the poorest communities, Whether the research has more than
and within those communities, through scratched the surface of knowledge of HH
wealth ranking, the poorest households were income remains uncertain. However, if the
included. The seasonal nature of their income, picture of income bears any relationship to
the reliance on non-government organisations reality, education expenditure will bear
and religious institutions to cope during heavily on the poor. It will be no surprise
particularly difficult periods indicate that a that many choose to take their children out
similar pattern of income and expenditure to of school, even while acknowledging that
other countries exists. Little or no education may be the only escape from
discretionary income, a high proportion of poverty for their children. The opportunity
expenditure on food and vulnerability to cost is perceived by many parents, rightly or
unexpected expenditure requirements are wrongly, as too great.
the likely consequences, which would have
similar results of households making tough
choices over schooling as seen elsewhere.

D DFID 75
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 94

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

5.6 The Costs of Education say that other factors do not figure in the T
equation. Indeed as has already been
5.6.1 Introduction discussed there are a number of barriers to
Education shall be free, at least in the education, for example poor quality, distance
elementary and fundamental stages. etc. Nonetheless, there is no doubt that cost
is one of the major barriers to education.
Ensuring that by 2015 all children,
particularly girls, children in difficult Specifically, this section covers the following:
circumstances and those belonging to ethnic
minorities, have access to and complete Cost as a barrier to education
free and compulsory primary education of Household direct and indirect costs of
good quality. schooling
The unexpected costs of schooling
Despite the noble sentiments expressed above33, Monetary v's non-monetary payments.
the harsh reality is that many children in
developing countries are still unable to access 5.6.2 Cost as a Barrier to Education
primary education (free or otherwise). Many Before looking at the scale of costs borne by
of those who are fortunate enough to enroll households it is worth discussing the extent
at the primary level are forced to drop out to which households view the costs of
before completion. The situation is even more education as a barrier. Interesting results
chronic at the secondary level. Even in those were found in the countries where both the
countries (e.g. Uganda) where free tuition SSIs and PAGs were conducted. Table 5.17
has been introduced (at the primary level) summarises the SSI results from the main
households still encounter problems quite study areas.
often school costs simply move from direct
to indirect categories with no overall Primary Level
reduction in costs to the household. In all countries lack of money is given as a
response for children never attending
In this section the study looks at the variety primary school although the figures vary
of primary and secondary school costs, direct from country to country. In Uganda the
and indirect, incurred by households. It figures are lowest, perhaps reflecting the
shows that the extent of cost sharing by the 34
positive effect of the UPE campaign. In
household varies from country to country. Uganda and Zambia it can be seen that the
However, what is clear in all countries is that poverty levels of the household also have an
the demand for education is sensitive to effect on enrolment, as would be expected,
price with the poorest households being with the poorest showing that lack of money
more price sensitive than the slightly better is a greater barrier for them than for the
off households in the sample. That is not to slightly better off households.

33
Quotes from the United Nations: Universal Declaration of Human Rights, Geneva, 1948, Article 26 and The World
Education Forum 2000 respectively.
*
34
As defined in Section 3.
S

76 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 95

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Table 5.17: Main Barriers to Education (%)

Reasons for children in the Nepal Bangladesh Uganda Zambia

household never attending PO SBO PO SBO PO SBO PO SBO
e
Primary
t
Lack of money 38.7 53.3 12.7 17.6 8.9 3.3 36.4 13.7
Too young 22.7 13.3 56.4 35.3 64.6 90 53.3 68.5
Needed to work at home 13.3 20
:

Secondary
Lack of money 21.1 40.7 27.3 * 25 0 35.7 44.4
Health 50 40
Death in family 0 20
Needed to work at home 31.1 22.2

Reasons for children dropping

out of school PO SBO PO SBO PO SBO PO SBO
y
Primary
Lack of money 57.1 100 60.9 50 30 0 86.4 21.4
Health 30 100
Death in family 30 0
Needed to work at home 28.6 13 0

Secondary
Lack of money 22.4 0 64 54.5 52.1 42.9 64.2 37.9
Needed to work at home 5.2 22.2
Earning money for the household 19 5.6

Reasons for being out of school

over the last month PO SBO PO SBO PO SBO PO SBO

Primary
School fees due 24.7 8.5 61.8 40.7
Health 44.4 52.7 64 56.7 64.4
Needed to work at home 19.1 55.6

Secondary
School fees due 47.1 50 31.7 31.3 77.6 66.7
Health 23.1 23.5 48.8 56.3 9 10.1
Needed to work at home 47.1 30.8
Agricultural work 11.8 3.8
* No figures available for SBO at the secondary level.
Source: SSI Q18, 19 and 23

D DFID 77
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 96

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

The reverse is true in Nepal and Bangladesh stationery, and for girls in particular, clothing, b
in Nepal one of the sites (Rasuwa) was soap and basic underwear still hinders many a
incorrectly selected with the income levels children from attending school. High costs also b
being higher than the other poorest households encourage early marriage and prevent those
in the sample. This site has been included in who would want to return to school from T
the above analysis and skews the results doing so. (Tumushabe et al, p. xx)36 w
somewhat35. (There seems to be no apparent l
explanation for the case of Bangladesh.) As It is also interesting to note that health is a m
can be seen another great barrier at the major factor in dropping out in Uganda for (
primary level is too young to go to school the poorest and slightly better off, and (
when probing was done during the SSI work having to work at home in Nepal and s
it became apparent that this issue was linked Bangladesh are barriers for the poorest. f
to distance and safety concerns by the parents. o
Many said that it is too much to expect small Irregular attendance also features as a S
children to walk the distances involved in problem in each of the countries, and it can r
going to school. Interestingly, parents made be seen that in Zambia school fees due is a o
this point for many of their children who major problem. This was confirmed during
were deemed old enough to go to school PAGs with parents and children. Children w
according to the official school age set by indicated that although some schools allow u
the Government. This was confirmed during them to continue to attend when they are a
the PAGs. unable to pay for school fees they often feel m
embarrassed that they cannot pay and t
When looking at the issue of dropouts the therefore decide not to go to school. In c
results show a dramatic increase in lack of other cases children are not allowed to I
money being a barrier. This demonstrates the attend school by the head teacher or teacher v
retention problems at the primary level, an until payment is made. In Nepal, although l
issue which governments must take on board school fees being due does not seem to feature n
particularly when looking at any policy of free the study shows that needed to work at home d
education. For example, with the introduction is the cause of irregular attendance this is a c
of UPE in Uganda in 1997 enrolments clear indicator of the opportunity costs a
almost doubled. However, studies show (e.g. related to having children in school.
Tumushabe et al, 1999) that even with free S
tuition, which may encourage enrolment, It is worth noting that in the PAGs with A
other costs hinder attendance, particularly children lack of money and having to work i
for girls. at home featured highly, with health being l
less of a problem than the parents indicated. f
Despite the abolition of school fees, direct costs In Sri Lanka and Kenya the findings were a
of schooling in the form of exercise books and similar to those above, the only difference D
n
In this section on school costs Rasuwa has been excluded from the analysis unless otherwise stated.
35

Tumushabe J., Barasa C., Muhanguzi F. and Otim-Nape J. (1999): Gender and Primary Schooling in Uganda,
36

Research Report 42, IDS, Sussex 3

78 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 97

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

, being that parents, teachers and children also for dropping out of school. Although
alike cited lack of money as the number one respondents were not willing to discuss the
o barrier to education in all instances. impact of AIDS on the households ability to
send children to school these results do give
Though not shown in the previous table an indication of the scale of the problem. It
when the data is disaggregated by each study is odd, however, that the Zambian results
location Zambia and Uganda show that the make no mention of this problem.
main barrier to education in the urban areas
(Kitwe and Mbale respectively) is cost. 96% Also of interest is the overall increase in
(Kitwe) and 86% (Mbale) of respondents responses related to children working (whether
stated that the main reason for the children at home helping with chores, earning cash
from their household being out of school or assisting with agricultural work) in Nepal.
over the previous month was school fees due. Again it is worth pointing out that the
Similarly in Mbale and Kitwe of those children involved in the PAGs in each
respondents who indicated that one or more country indicated that this was a problem for
of their children had dropped out of school them while the parents involved in the SSIs
100% and 61% respectively indicated that it in Bangladesh, Zambia and Uganda did not.
was due to financial difficulties. Although
urban areas may not suffer from physical In Kenya and Sri Lanka cost again was the
access problems (e.g. distance to school) as major barrier to education. In addition,
much as rural areas the costs they face tend circumcision is problematic in Kenya, with
to be higher than the rural areas resulting in many girls dropping out of school. In
children dropping out of attending irregularly. Andimune (a fishing village) in Sri Lanka
In Bangladesh and Nepal although some boys often drop out of school to learn the
r variation exists between each of the study fishing trade this is not simply because
locations the differences are not as marked, their labour is needed rather it is a pragmatic
e nor do they seem to demonstrate any great response to the lack of alternative economic
difference in difficulty in paying for school opportunities, as well as being an alternative
a costs between the rural and urban locations, education in the traditions and culture of the
as demonstrated in Uganda and Zambia. fishing village. In both Sri Lanka and Kenya
children also drop out to help at home with
Secondary Level siblings, housework and work in the fields.
At the secondary level the pattern of results
is slightly different to those at the primary 5.6.3 Household Direct and Indirect
level. lack of money is seen as being a greater Costs of Schooling
factor for never enrolling in Uganda which As already discussed in Section 3 the costs of
again points to the positive effect of UPE. education to the households include direct
Death in the family also appears in Uganda, and indirect costs37, and may take the form
not only as a reason for never enrolling but of monetary or non-monetary contributions.

37
As defined in Section 3.

D DFID 79
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 98

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

These costs (whether monetary or non Table 5.18 demonstrates that in almost all T
monetary) are opportunity costs they cases SBO households spend more per child
reflect all that has been given up by the than PO households do. The only
choices that have been made. This study was exceptions are for Bangladesh (secondary
interested in the expenditure choices made school costs for boys) and Nepal (primary
by the household on education. school costs for boys). This may be due to
the household selection criteria where local
(i) Household Costs per Child definitions of PO and SBO were used.
The types of direct cost looked at in the
study were many and varied between the At the primary level some gender disparities
countries. They included, but were not exist particularly amongst the poorest
limited to, school fees, boarding fees, households. However, in Uganda and Zambia
admission fees, Parent Teacher Association in almost all cases spending on girls at the
fees (PTA), examination fees, school primary level seems to be greater than
development fund, textbook rental, pens and spending on boys confirming that the costs
pencils, transport to school, uniforms etc. In of sending girls to school in these countries
addition the study looked at the indirect can be greater than sending boys. At the
costs (e.g. contributions to teachers funeral secondary level the extent of the differences
expenses), including the cost of having a in costs between boys and girls vary from
child in school rather than having him/her country to country in some instances the
helping at home or earning income. The cost is greater for boys, in others it is greater
latter of these indirect costs was not for girls. The area of Kitwe demonstrates
quantified. However, it was apparent that that spending on girls education is almost
the opportunity cost of having a child in three times that of boys, however, this figure
school was high, often prohibiting the child should be treated with caution. On further
from enrolling in school or attending analysis it appears that the average household
regularly. Both the direct and indirect costs spending in Kitwe is inflated by an outlier
include non-monetary payments such as one female secondary child attends a private
assisting with building schools, provision of day school the cost of which is significantly
building materials, payments in kind etc. greater than other types of school in the
Where possible an attempt was made to sample (e.g. government day schools). If this
quantify these non-monetary payments, case is taken out of the analysis the average
although the respondents found this difficult. figure for household spending on girls
The table right which includes direct and secondary education becomes Kwacha
(
indirect costs may therefore underestimate 157,700, twice that of the spending on S
the total per student costs of education. boys. The gender disparity in Bangladesh is
much more pronounced at the secondary
level than in any of the other countries38.

38
The Promote project in Bangladesh is designed to encourage more girls into and through secondary education.
Scholarships are provided to girls for this purpose. It is possible that the figures presented here are influenced by the
inclusion of households in the project. As the research team did not provide this information (following repeated
requests) this has to remain at the level of supposition.

80 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 99

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Table 5.18: Per Student Household Spending on Education

Household type and area of study Boys Girls Boys Girls

Nepal (Rupees)
All households 751 569 2,019 2,185
PO households 804 488 1,811 1,802
SBO households 564 694 2,624 2,824
Achham 922 718 2,383 2,163
Kailali 666 478 1,326 0
Kapilbastu 647 374 1,537 2,222
Bangladesh (Taka)
All households 723 728 4,026 2,505
a
PO households 518 503 4,449 2,486
SBO households 1,236 1,378 3,134 2,564
Mymensingh 331 315 5,580 1,969
Dhaka 1,704 1,654 5,123 3,547
Patuakhali 887 652 4,107 4,107
Dinajpur 188 282 2,082 1,735
Uganda (Shillings)
All households 21,230 25,985 273,629 245,779
PO households 20,901 24,201 251,947 202,571
r
SBO households 22,265 33,119 307,917 275,945
Kibaale 14,143 11,238 0 0
Kiboga 15,710 17,455 390,333 183,450
e
Apac 15,495 23,835 300,304 319,419
Mbale 37,214 44,989 270,990 179,429
d
Zambia (Kwacha)
All households 24,427 27,190 125,073 115,809
PO households 21,908 23,891 66,525 88,100
SBO households 28,071 31,094 175,257 129,664
Kitwe 46,344 71,526 82,750 232,750
s
Chinsali 13,620 17,796 127,525 83,300
Siavonga 22,581 9,228 169,500 94,150
Kalabo 19,590 17,648 122,270 89,843
(Note: where a zero value is given it means that no children from that area attend secondary school)
Source: SSI Q 48-79

D DFID 81
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 100

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

A number of other studies looking at the primary and secondary levels for the A
household costs of education have been done households in the sample. Uganda has the e
in Zambia. For example, the World Bank most dramatic difference with spending on
(1996)39 stated that the primary costs of secondary education being 12.9 and 9.5 l
education for households ranged between times that of spending on primary education s
US$ 30 (K 35,000) to US$ 60 (K 60,000) for boys and girls respectively. This may be T
per child. Kelly (1998) shows that school explained in part by the policy of free tuition a
costs vary from US$5 per year in very remote (UPE) in Uganda at the primary level.
areas to over $50 in urban areas at the c
primary level. Based on household income In determining school costs in Sri Lanka the a
and expenditure data from the Living researchers found it difficult to differentiate t
Conditions Survey in Zambia (1998), a between the costs that the study participants i
more recent study by Seshamani (2001), felt ought to be incurred, and actual costs e
undertaken for Oxfam and JCTR, estimates of education. It is likely therefore that the f
household primary school spending per child results indicate higher than actual expenditure. w
at K 73,668. All of these figures vary from Furthermore, participants expressed large R
our results where the lowest and highest variations in the costs per child. Where families
amounts paid per primary child are US$ 3.7 have more than one school-going child, school I
(K 9,228) and US$ 29 (K 71,526) from the equipment is usually shared by all the children d
individual sites . That is not to say that the
40
or handed down. The costs given therefore f
figures from this or the other studies are present an average. The total annual cost of g
inaccurate it simply reflects the different primary schooling per child as estimated by m
types of households being studied. In the parents was between Rs4,000 to Rs 5,000 in a
case of this study, as already mentioned, a both of the study sites; and around Rs 2,700 a
purposive sample the poorest households as estimated by children. The items included L
was selected. in these costs were clothing41, food, school s
and tuition fees, transport, exercise books c
Table 5.19 below shows the differences and school excursions. m
between per student school costs at the t

Table 5.19: Spending on Secondary Education as a Factor of Spending on Primary Education I

s
All houeholds Nepal Bangladesh Uganda Zambia r
Boys 2.7 5.6 12.9 5.1 t
Girls 3.8 3.4 9.5 4.3 m
Source: SSI Q 48-79 e
c
39
World Bank (1996), Zambia Education Sector Expenditure Review, September 1996 M
40
Using the exchange rate at the time of the field work US$1=K2,500 e
41
In Sri Lanka parents are provided with cloth from the government to enable them to make uniforms for their children,
it is therefore strange that clothing seems to be a major part of the costs of schooling. Two points should be borne in 5
mind firstly clothing includes footwear and underwear (which the government does not provide) and secondly there are
costs related to having the uniforms tailored. l

82 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 101

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

At the senior secondary level schooling was and Ksh 27,900 for primary and secondary
estimated by parents to be Rs 2,417 and Rs respectively. When compared with the
16,182 by students in Andimune. These household income (Ksh 8,400 per year in
large differences make these cost estimates Mandani and Ksh6,000 per year in
n somewhat suspicious and difficult to compare. Kapsinendet) these education costs can be
The differences in opinion seem to lie mainly put in context. Household incomes are not
n around the cost on food (parents quoted Rs adequate to meet basic needs such as food,
1,000 and students Rs 4,000) and extra health and clothing and do not stretch to
curricular activities (parents quoted Rs 170 meet the costs of education.
and students Rs 7,100). The students seem
to have presented a wish list of the costs Family size is also an issue in Kapsinendet
involved few were actually involved in the the average number of children in a household
extra curricular activities because of household is 10 to 14 and in Mandani it is 8 to 10. The
financial difficulties. In Rohanapura the costs disjuncture between income available and
. were estimated at Rs4,385 by parents, and mandatory school payment is clearly a major
Rs5,098 by students. reason for so many children of both of these
s villages either not enrolling at all, attending
l It is difficult to draw conclusions from the intermittently or dropping out. (Over half
n data gathered in Sri Lanka. However, the the households in Kapsinendet, for example,
findings do show that the costs of schooling say that they have children who have never
goes significantly beyond school fees, with enrolled or who have dropped out.)
much of the expenditure being on clothing
n and food. These costs add up to significant Direct costs are only part of the problem in
0 amounts, relative to the Government of Sri Sri Lanka and Kenya. Indirect or opportunity
d Lankas (GOSL) official poverty line. It is no costs also create responses which require
surprise that in times of hardship school students to miss school days, weeks or months.
costs represent an area where savings can be In Kenya this is reported to affect girls more
made by withdrawing children from school than boys, for example in Mandani girls are
temporarily or permanently. needed to help their mother on market days,
or to find water during periods of water
n In Kenya, the costs of education in both shortage or drought. In Kapsinendet, the
study sites were found to be far beyond the indirect costs include loss of labour in cattle
reach of the majority. Levies vary from herding, farming and taking care of young
textbooks, desks, pens, physical facilities, siblings and household chores. These activities
maintenance of equipment, costs for zonal again are for girls, but a few boys said that it
educational, security and other miscellaneous is boys who are asked to stay at home and
charges that occur from time to time. In look after cattle. In Sri Lanka, the opportunity
Mandani the annual costs of primary cost of having children in school was also
education per child were estimated at Ksh raised as an important issue. In Andimune a
5,010 and Ksh 38,100 at the secondary boy can earn Rs 500 per day from casual
e
level. In Kapsinendet they were Ksh 7,290 labour (dragging and cleaning nets and

D DFID 83
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 102

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

working in prawn ponds). If he goes to together a number of items of school costs. i

school this represents a substantial loss in Where applicable42 payments to school is e
income. In Rohanapura parents are required made up of tuition fees, boarding fees, t
to meet the teachers at the primary level and admission fees, school development funds, p
to attend shramadanas regularly leading to a PTA fees, exam fees and sports clubs. t
number of lost days earnings. There are less Education materials include textbooks, l
income earning opportunities for boys in exercise books and pens and pencils. Other p
Rohanapura, although some do sell cannabis support costs include school bags, transport,
to earn an income. Girls in both areas are lunch/tiffin, uniforms, footwear, unexpected A
frequently required to stay at home to look /indirect and other costs. Private tuition l
after their younger siblings and to attend to refers to tuition outside normal school hours. h
other household work, particularly during B
the farming season. At both the primary and secondary levels it t
can be seen that very little is spent on private i
(ii) Composition of Education Costs tuition. Given that the households included t
The composition of education costs varies at in the study are the poorest this finding is e
the primary and secondary levels. Moreover, what might be expected. In all countries the t
the composition of education costs varies main reason for attending private tuition was T
between the different study sites and from to make up for poor learning at school. The a
country to country. The figure below shows private tutors, in most cases, appeared to be i
a simplified picture of how the costs are the schoolteachers themselves who needed (
made up within each country by pooling the additional funds to boost their meagre r
b
Figure 5.6: Household Composition of Primary and Secondary Education Costs
g

100 %
e
p
n
80 %
t
m
60 %

40 %

20 %

T
Primary Secondary Primary Secondary Primary Secondary Primary Secondary
Nepal Bangladesh Uganda Zambia t
o
Payment to school Private tuition
Education materials Other support costs t
s
42
Not all of these items will be relevant to all countries but for simplification we have listed all of the items named in the SSI. r

84 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 103

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

incomes. In Zambia and Uganda, anecdotal clothes and footwear specifically for school at
evidence from discussions with parents suggests a considerable expense to the household.
that in some instances teachers were deliberately Information from the PAGs work with girls
providing poor tuition in school hours so from Kitwe in Zambia shows that the cost of
that children would have to attend private uniforms are prohibitive with the girls
lessons if they wanted to have any chance of indicating that they are the second highest
passing exams. cost item, with school fees being the most
, expensive item.
d At the primary level payment to school is
lowest in Uganda and Nepal both of which 5.6.4 Seasonality of Household Incomes
. have a policy of free tuition. Although and Education Costs
Bangladesh too has a free tuition policy at The proportion of household expenditures
the primary level, payment to school which devoted to education costs has already been
e includes school fees is nonetheless higher discussed in Section 5.5.1 on Household
than that of Nepal and Uganda. A possible Income and Expenditure.
explanation is that fees continue to be charged
though they may be given a different name. Here the study looks at the relationship
s The other payments for Uganda and Zambia between household incomes and the costs of
are considerable. On further investigation it education. It is evident from each of the
is apparent that at the primary level 51% countries studied that the issue of seasonality
(Uganda) and 76% (Zambia) of these costs of incomes and timing of school costs severely
relate to uniforms (or clothes) and footwear hampers the ability of children to go to
bought specifically for school. This is surprising, school. A few examples are presented overleaf.
given that uniforms are not compulsory in
either country. Discussions with children and The income and expenditure pattern in
parents revealed that although uniforms are Kapsinendet village, Kenya (Figure 5.7)
not compulsory some schools still insist upon demonstrates the highest incomes when
them. In Zambia one of the researchers families are able to sell their produce,
made the following comment: particularly maize, which is grown when
possible in large quantities so as to earn
Although government policy does not make income. The calendar shows that income
uniform a compulsory requirement, schools and expenditure patterns are not correlated.
have been reluctant to follow the rule because Income is available at the time of harvest
they see the uniform as a symbol of identity and just after (peaking between July and
September), whereas the school calendar
To add to this problem of schools ignoring starts in January.
the policy on uniforms is the fact that many
of the children (particularly girls) feel that In Bangladesh, during Ashar-Saraban (Jun-
they do not have adequate clothing for Aug) and Kartik-Agrahyan (Nov-Dec), it is
school, whether it is a uniform or not. As a difficult for children to go to school. In
result parents find themselves having to buy Ashar-Saraban, there are monsoon rains with

D DFID 85
 Figure 5.7: Seasonal Calender Kenya - Kapsinendet village

86
Rainfall

Labour demand

Source: Kenyan Country Report

5 Discussion of Findings

Income
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 104

Expenditure
Reaching the Poor - The costs of sending children to school

Disease

Agricultural activities
e, on ,
s & lds t ati ize
m aiz et
l ps ird fie ves r : ma llet
g: mil g cro g b m ar epa i
i n , n n ro h st pr ing m
nt ns edi epi s f rly rve nd nt ns,
Pla bea We Ke dog Ea Ha La Pla bea

DFID
y ril y e y st er er er er
uar ar y rch Ma Jun Jul gu mb tob mb mb
Jan bru Ma Ap Au te c ve ce
Fe Sep O No De

4
i
f
f

t
t
t

s
s

a
a
a
a
a
a
e
e

F
p

A
K

w
m
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 105

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

flooding limiting the opportunities of the study areas, it is evident that seasonality
employment and as a result the income of of income, coupled with the timing of
the parents. During these times any income schooling costs, creates great difficulties for
ber

that is earned is spent on food. During Ashin, households, particularly in the rural areas.
cem

Kartik and Agrahyan (Sep-Dec), the food Faced with expenditure choices households
De

stock of households is exhausted and are responding rationally by spending on

ber

expenditure on food is the households major food and health before spending on education.
vem

priority. Thus, in these months, many children

No

are unable to attend school. 5.6.5 The Unexpected Costs of Schooling

er

Table 5.18 (per student costs of schooling

tob

In Zambia peak expenditures for both food primary and secondary) included all costs
Oc

and education are experienced during the related to schooling including unexpected
months of January to March when incomes items, however, it is worth mentioning the
ber

are at their lowest in rural areas. To add to range of items included in this category. In
ptem

the burden of households these months are Zambia the items included funeral expenses
Se

also the peak malaria season and thus funds for teachers, independence day celebrations,
st

are also needed for medicines. When faced aid to teachers, contributions for a school
gu
Au

with expenditure choices between food, health guard, stationery fund, and school building
and education it is not surprising that education contributions. Approximately 18% of the
y

suffers. Ironically, from June to August when children attending school made such a
Jul

incomes are at their highest (after selling contribution with an average cost per student
farm produce) school costs are minimal. of K 3,300. In Uganda there was a greater
range of items including welcoming new
e
Jun

Although these are only a few examples from teacher, farewell to teachers, teachers funeral

Figure 5.8: Seasonal Attendance, Bangladesh43

y
Ma

Bangladesh: Seasonal difficulties in sending children to school

rl
Ap

Jun-Jul

Oct-Nov
Level of difficulty

Jul-Aug
Nov-Dec

Sep-Oct
rch

Dec-Jan
Ma

Aug-Sep
Jan-Feb Mar-Apr

May-Jun
ar y

Feb-Mar
bru

Apr-May
Fe

Apr-May May-Jun Jun-Jul Jul-Aug Aug-Sep Sep-Oct Oct-Nov Nov-Dec Dec-Jan Jan-Feb Feb-Mar Mar-Apr
y
uar
Jan

43
Source: PA work with female participants, Mymensingh research site

D DFID 87
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 106

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

expenses, entertainment of visitors, end of in Zambia and Uganda the respondents c

year celebration, school building, transportation indicated that in 63% and 87% of the cases d
of building materials, and telephone connection. respectively the non-monetary contributions r
However, a smaller percentage (9%) of were not voluntary. The extent of non g
students attending school made such a monetary payments also varies with a low of o
payment. The average payment made per 12.5% in Nepal and a high of 50% in Uganda. t
student was approximately Ush 2,900. In o
Bangladesh, 14% of children in school made SSI respondents in Nepal, Uganda and p
unexpected payments (averaging Tk 90). Zambia (75%, 57% and 68% respectively) m
The cost items included contributions for showed a preference to pay non-monetary C
picnics, damaged or lost textbooks, and contributions over monetary contributions. t
farewell lunches. Only 1.5% of school-going Respondents in Bangladesh on the other i
children in Nepal made unexpected payments hand preferred to pay neither wishful o
averaging Rs 93 per child. The items thinking or a misunderstanding of the i
included contributions for picnic, painting question posed to them concerning types of
school building, school furniture, and payments. There was little difference in the
equipment for science lab. responses from the poorest and slightly T

better off households with the exception of

5.6.6 Monetary vs Non-monetary Zambia where the poorest (and respondents
Payments from rural areas) showed a greater interest in
As noted earlier the respondents found it paying non-monetary contributions than the
difficult to quantify the non-monetary slightly better off households. The main
payments that they made towards the schooling reason that parents thought it was good to
of their children. Nonetheless interesting data make non-monetary contributions in all study
was gathered on the types of non-monetary sites was to improve the quality of the
payments and the attitudes of parents towards education their children receive. However,
these contributions vis a vis monetary payments. those who were not in favour of non-monetary
contributions stated that the household was
Table 5.20 shows the extent and types of already burdened, with little more time to
non-monetary payments made for those commit to such things as school building.
children in school. The data collected does
not show any gender disparities in the activities 5.6.7 Summary
undertaken by boys and girls in Nepal and Even in those countries where education is
Bangladesh. However, the picture is quite tuition free households have to pay a
different in Uganda and Zambia, with girls considerable amount to school their children.
tending to get more involved in cleaning the In Uganda for example the study showed
school and playground. Respondents were that a considerable number of children have
also asked whether the contributions were to pay non-monetary costs and an array of
voluntary or non-voluntary almost all unexpected costs. If tuition fees at the
respondents indicated that they were primary level were not abolished, would
N
voluntary in Nepal and Bangladesh, however, parents have to make the range of other
m

88 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 107

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

contributions that have emerged? This study 5.7 The Ability and Willingness to
does not propose that tuition fees should be Pay for Education
reintroduced. But it is important for In an introductory section the economic
governments to recognise that the removal theories which underpin the study approach
f of tuition fees may result in schools having to determining ability and willingness to pay
. to insist on other charges (monetary or for education were outlined (section 3.3).
otherwise), with the overall result that
parents end up paying the same or even The evidence tended to support both the
more than when tuition fees were paid. predictions derived from general economic
Costs do act as a barrier to education and theory about the factors that would influence
the disjuncture between seasonality of the ability and willingness to pay for education
incomes and the timing of schooling costs (Belfield, 1999). Perhaps more important,
often makes a bad situation worse. Flexibility the findings resonate closely with those from
in payments is therefore suggested. earlier studies around the world (World Bank,
1990,1995). In Nepal the evidence is that

Table 5.20: Types of Non-monetary Payments made by Children and/or Parents

Country % of children and/ Type of contribution

or parents making
n non-monetary
contributions
e
Nepal 12.5 Provision of food or cooking for school meetings (P)
Clean playground and school building (C)
Upkeep of school building (P)
y
Bangladesh 16.5 Clean playground and school building (C)
Upkeep of school building (P)
Makes bricks (P)
y
Collects water (C)
Collects firewood (C)
Uganda 50 Clean playground and school building (C)
Upkeep of school building (P)
Building work for new facilities (P)
Provides grain (P)
Makes bricks (P)
Provides building material (P)
.
Collects water (C)
Collects firewood (C)
Provide cleaning material (P)
Maintains garden (C)
Zambia 20 As above for Uganda
Note: In the final colunn (P) represents contributions by parents and (C) represents contributions that children
make themselves.

D DFID 89
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 108

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

(with reference to respondents) about 16 Very similar results were found in Zambia, b
and 19% of respondents were not willing to with 77% of parents prepared to pay more a
pay more for primary and secondary for girls education of the present quality and a
education respectively (Nepal report p65), 79% for boys. However, they were not p
with poverty being the main factor affecting prepared to pay a lot more (85%). There t
their unwillingness. In this case willingness was little difference between SBO and poorer h
does appear to be a function of ability. parents in their willingness to pay more for t
education. Not surprisingly the main reasons i
However, this finding also meant that over given for not paying more were to do with w
80 % were prepared to pay more for the lack of resources (98.6%). A rather lower e
existing quality of education. Although there percentage was prepared to pay more for w
was some difference between regions when secondary education (44%); with there being s
asked about their willingness to pay more for little disparity between regions. n
improved quality education 78% were prepared t
to pay more. This is very similar to the earlier In Uganda over 75% of HHs were prepared
figure and may indicate that around 20% of to pay more for the current quality of primary C
our sample were unable to pay any more for education. However, only 50% were prepared r
education irrespective of any quality changes to pay more for secondary education. One a
to it. The evidence that 24% of households reason for this difference between primary e
would be forced to withdraw a child from and secondary education may be that many r
school supports this if costs increased by 50%. HHs did not have children attending
secondary school. There was a significant i
The evidence from Bangladesh is that both difference between Kiboga region and the d
male and female parents are prepared to other regions in HH willingness to pay for p
work harder for boys than girls education. primary education, with willingness to pay p
Figures are given in Table 5.21 below. for boys education being just over 50% of e
that for other regions. (No explanation is p
Parents were also prepared to spend more offered for this difference.) The difference was e
money on their childrens education, with less pronounced for girls education, but it d
45% prepared to spend more on the current was still around 75% of the willingness to pay r
quality of primary education and 64% on more from that found in the next lowest region. r
current quality of secondary education. The c
slightly better off were prepared to pay a The picture for willingness to pay more for i
little more than the poorest families. secondary education is similar across all regions, m
p
d
Table 5.21: Parental Willingness to Work Harder
c

Male respondents % Female respondents%

Girls education 83 88
4
Boys education 86 91
d
Source: Bangladesh report, Table 6.4.1 n

90 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 109

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

but the sample size is so small as to make Our evidence also suggests that the position
any generalisations extremely speculative. In and slope of the demand curve may differ
d all the regions over 80% of HHs were between regions and across different income
prepared to take on more work to pay for groups. Policy makers would need to bear
their childrens education. Over 80% of this in mind when determining their pricing
r households would be more willing to send levels, otherwise pricing policies successful in
their children to school if schooling quality certain regions and for certain groups may
s improved, but again for Kiboga the figure result elsewhere as a deterrent to school
was significantly lower, at 61%. The only attendance. For all groups there appears to
explanation offered for Kiboga being different be a contrast between the willingness to pay
was that NGOs had already improved a lot more and a little more for schooling44,
g schooling quality and that the reasons for with most HHs prepared to pay a little more
not sending them to school were unrelated but not a lot more. This suggests that a
to quality. kinked demand curve for schooling exists,
with there being a price level at which the
y Certain inferences can be drawn from these demand curve becomes considerably more
d results. There seems to be a readiness inelastic. This is illustrated in Figure 5.9.
amongst HHs to pay more for their childrens
education than they do currently and this The demand curve has the same shape as in
readiness increases if coupled with an Figure 3.3 until the price falls to P* when its
improvement in the quality of education that slope becomes steeper (more inelastic). This
is being provided. Elsewhere, in the report a means that for any fall in price below P* there
discussion on what is meant by quality is will be a smaller increase in the quantity (of
presented, but the principal concern of schooling) bought than with the demand
parents is related to their perceptions of the
Figure 5.9
effectiveness of the teaching input. As to
Demand for Education
parental willingness to pay more for
s education, earlier the possibility that the
demand function for education may be P1
y related to the ways in which the revenues
P*
. raised from cost sharing were used was
P2
considered: if the revenues raised are used to Private demand
for education
improve the quality of schooling then HHs
, may be prepared to pay not only existing
prices, but even higher ones. In terms of our q1 q* q2 Quantity
diagram this is the equivalent of the demand
curve shifting to the right. curve above P*. If such a point exists then it
would be useful for policy makers to know

44
It was left to the research teams in each country to define on location the meaning of a lot and a little. The
definition is more to do with the perceptions of the respondents in relation to their own circumstances than to a
notionally objective definition

D DFID 91
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 110

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

approximately where it is because changes in opportunity cost the expenditures made h

price beyond this point will significantly on education are done so at the expense of o
affect participation rates, almost certainly other spending. In that sense they may be (
having a greater effect for the poor. Quite said to represent a sacrifice the household
clearly as the response to price changes, (the makes forgoing one thing to realise another. S
price elasticity of demand changes), so also o
will the revenues that can be raised from A distinction needs to be drawn however, u
school fees. If revenues raised are used to between sacrifices and choices. Expenditures c
enhance schooling quality the demand curve on education made by those in higher level o
shows how the quality of schooling may be income groups tend to represent a o
affected by changes to the price of schooling. consumption/investment choice45 (e.g. private t
schooling or a new house) and almost certainly s
The case studies in all countries show how the choice is from within discretionary o
responsive HHs are to price changes and to expenditure. In lower income groups decisions d
perceptions of the quality of education. In about spending on education may also be o
Nepal, for example, only two percent of from discretionary expenditures, but the w
respondents said that they would withdraw poorer the household the more likely the p
their children from primary school if present decision making will enter a grey area between m
schooling costs increased by 10 %. However, discretionary and non-discretionary spending. c
this figure rises to 24 % if costs rose by 50 %. Such choices can properly be termed sacrifice. p
The elasticity of demand was also calculated A
and shows that it increases as the price is This study has amassed considerable d
increased and is greater for secondary evidence of sacrifices being made by the p
education than for primary education; poorest income groups to school their h
Demand for education is inelastic for both children. Such sacrifices include non w
primary and secondary education (Nepal discretionary expenditures such as food and t
report p70). Price elasticity is greater for clothing as well as sacrifices of capital
PO households than SBO households and assets.The components of non-discretionary A
there are also regional differences. income have no fixed definition and the r
basket of goods it represents will differ from T
5.8 Household Sacrifices country to country and even household to c
household. But, fundamentally, non m
5.8.1 Introduction discretionary income describes that w
Not only are the poor prepared to pay the proportion of household income needed to p
costs of education both direct and indirect ensure survival of the household. Those
they also make significant sacrifices to give essentials will at a minimum include: food O
educational opportunities to their children. and water, shelter, and health care. Education p
Of course, schooling of any child for the is not part of non-discretionary income; d
parents of any income bracket represents an i
t
45
Education has characteristics of both consumption and investment goods.

92 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 111

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

health may fall into both definitions depending by the gender, birth order of the child or
on the perceived importance of the condition whether they are an orphan). It indicates
(e.g. fever vs. pulmonary tuberculosis). that in some cases they do this at the
expense of the long term and short term
r. Such a spare definition is useful in an analysis health of family members, of diet and of
of household spending decisions, but their limited opportunities for recreation. It
unfortunately, somewhat restricted in also suggests that in some cases the poorest
capturing well-being, or a rounder definition households place such a great valuation on
of what is required for a life with dignity. In education as an escape route from the
other words, it would not encompass what intergenerational cycle of poverty that they
e the poor themselves would define as essential are even prepared to invest household assets
y spending. Within this study there is evidence in trying to break the vicious circle.
of households who make sacrifices in their
s diet (e.g. eating meat less frequently) in The forgoing comments have focused on
order to save money for school expenses, but what sacrifices households make to keep
who nonetheless would not contemplate not children in school. Of course, one of the
providing a dowry for their daughter in major coping strategies for households in
n marriage or spending on funerals. A very economic difficulties or suffering from
g. clear message from the recent literature on economic shocks is the withdrawal of children
. poverty (e.g. WB Voices of the Poor. Can from schooling either temporarily or
Anyone Hear Me?) is that the poor have their permanently. This decision is often a
dignity often expressed through gendered one (and in some countries related
participation in social activities - and that to whether they are orphans) this is further
household spending on these activities discussed below.
would not be classified as discretionary by
the poor themselves. At the other end of the spectrum of sacrifice,
our research shows that there are families
A further distinction must be made with who are unwilling to sacrifice anything for
respect to the capital assets of the poor. the sake of their childrens education when
These may be limited some chickens or a others around them are doing so. Reasons
cow, tin roofs, the labour of household range from a lack of valuation of education
members - but the sacrifice is all the greater believing it to be a wasted investment
when such assets are sold / used in order to because of poor quality, lack of relevance etc.
provide education opportunities. to unwillingness to take the painful steps
required in squeezing the family budget to
Our study presents clear evidence that the addiction to alcohol or other drugs which
n poorest households make sacrifices in both consume available funds before they can be
discretionary and non-discretionary spending devoted to education.
in order to school children (noting however,
that such decision making may be affected

D DFID 93
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 112

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

5.8.2 Background One of the communities investigated in e

Before questioning respondents on household Kenya and one in Sri Lanka have special o
sacrifices related to education, a general customs and social practices which involve T
question was asked about how much considerable expenditure. In Kenya circumcision v
households spent on social functions such as rites involve parties and marriages which affect o
weddings, funerals, festivals etc., and what household resources. Families in the community
sacrifices they had to make to meet the are large (averaging 6-10 members) so the W
expenses of these. The answers offer some effect on households, though irregular, is t
context in which responses about sacrifices constant. In one community in Sri Lanka the t
made for education can be judged. spending on religious festivals and donations
t
to the temple consumes time and household
w
In almost all study areas analysis of household resources it was reported by researchers
d
expenditure revealed significant spending on that families spend lavishly on religious
a
social functions when compared to education festivals at the cost of meeting childrens
b
spending (see Table 5.22 below). It is school requirements (Sri Lanka report).
r
possible that these figures are underestimates
b
as spending on alcohol and tobacco However, to present spending on education
m
commonly associated with such occasions and on social functions such as weddings,
A
is reported separately (e.g. if these items funerals, religious festivals etc. as a choice
were included in Nepal, the figure would be over which households have a significant w

1.5 times). In Africa the effect of ARCs is degree of control, is a tempting but false t
noticeable expenses on funerals in addition comparison. If anything, the poor families o
to the opportunity costs of attending them. interviewed in these studies regard such d
t
Table 5.22: Spending on Social Functions as a Factor of HH Spending on Education
a

Study area Bangladesh Nepal Uganda Zambia

5
Spending on social functions 1.5 times 0.8 times 0.53 times 1.07 times
as a factor of household I
spending on education
t
Source: SSI Q 107-115 b
Table 5.23: Sacrifices/Responses Made to Meet Costs of Social Expenditures (%) i
e
Nepal Bangladesh Uganda Zambia
t
Extra labour 31.1 17.8 11.7 33.5 o
Borrowing 32 19.2 2.8 5.6 r
Use of savings 16.5 54.8 q
Sale of land 5 w
Curtail expenditures 5.5 7.4 2.2 r
Selling livestock 16.4 11.2 b
Reducing consumption of food 23.1 10* c
Reducing clothes 7.5 6
No sacrifices needed 11 12.8
Source: SSI Q 115

94 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 113

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

expenditure on social functions as even more was the main beneficiary. Interestingly, there
of a necessity than expenditure on education. was a marked difference in response between
This does not necessarily indicate a low urban and rural areas in Nepal and Uganda
n valuation of education, but rather a different in both countries urban areas sampled
t ordering of priorities. suggested benefits of education accrued
y mainly to the household (despite Nepal as a
When households were asked what sacrifices country favouring the individual). This may
they made to meet these social expenditures be a particular characteristic of urban poverty
e there were some clear differences between in certain areas - households may see
s
the Asian and African study sites, but even education as helping to create household
d
within these sites there were markedly capital where in some rural areas education
different approaches between communities may be seen more as a route out of poverty
and differences between the poorest and for individuals and therefore of potentially
better off households. In Asia the principal less benefit to the household (see Section
responses were that households either 5.2 on Value and Benefits for further
borrowed or undertook additional work to discussion of this topic).
meet these expenses (sacrificing leisure). In
Africa, while additional labour featured strongly, 5.8.4 The Findings
what was noticeable was the emphasis given In both the SSI and the PAGs work
to reduction of food consumption and sale questions were asked about the sacrifices
of livestock as response strategies indicating households currently made to school
different patterns of asset ownership and access children. Three questions were asked:
to credit/borrowing. Table 5.23 illustrates the
averaged responses in each country. 1. What sacrifices households currently make?
2. What household responses would be in
5.8.3 Household/Individual Benefits the face of a hypothetical 50% or 100%
It is useful to consider the responses made to drop in the costs of educating their children?
these questions in the light of the perceived 3. What household responses would be to a
benefits of education to either households or hypothetical increase in household
individuals, since it would be reasonable to education costs in three steps (which were
expect that households would be more willing country specific)?
to make sacrifices if they perceived the benefit
of such sacrifice to accrue to the household The first and second set of questions allow a
rather than to the individual. When this certain degree of validation, since, by inference,
question was asked in the four countries household responses to the hypothetical drop
where SSIs were administered the majority in education costs can also indicate what is
responses were in favour of the household already being forgone. The third question
being the main beneficiary of schooling of tested the willingness and ability to pay of
children, with one exception Nepal where poor households exploring the level of
60% of respondents thought that the individual financial constraints they already felt. The

D DFID 95
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 114

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

following sections set the responses to these In Sri Lanka respondents reported that A
questions within the framework of children had to miss school excursions and a
discretionary and non-discretionary buy rubber slippers for school instead of w
household expenditures. shoes (Andimune, Sri Lanka report, p60). s
One villager in Rohanapura commented: t
5.8.5 Current Sacrifices Made h
for Education We postpone our house construction work t
In commenting on payments for schooling We postpone visiting and seeing our relatives b
children most poor households indicated (and sometimes we use the money we have for i
that they were making difficult choices within buying seeds and agro-chemicals for our farm).. m
discretionary income for this purpose. It is (Sri Lanka report, p63) t
difficult to show in a direct and unambiguous
way that spending on education is the reason Reduced spending on seeds and agro T
for such choices but, at least in the minds chemicals can lead to lower yields than f
of respondents there seems to be a very high would be possible and therefore to lower
correlation between education costs and household income or difficulties in J
spending choices the household needs to maintaining subsistence. t
make (though this could be courtesy bias h
where respondents favour education in their Sacrifices from Non-discretionary Expenditures t
responses knowing it is a research study Examples, gathered from both SSIs and f
about education). This is further confirmed PAGs, of sacrifices currently being made
by the response to the hypothetical questions from non-discretionary expenditures include: S
mentioned above. In Uganda, Bangladesh m
and Nepal a clear majority of respondents Not buying, or not having enough clothes
said they would need to make further and this affecting regular attendance at school T
sacrifices if education costs rose (see below). (Bangladesh and Uganda). Not using i
cooking fat a source of vitamins and protein i
Sacrifices from Discretionary Expenditures (Kenya). Reducing consumption of sugar l
The sacrifices made from discretionary and salt (Uganda). General reduction of food
expenditures include the opportunity costs (all countries) and skipping of meals (Kenya). 5
of additional work in order to meet Deferring treatment of illnesses (Sri Lanka).
education costs (sacrifice of leisure time), T
investment choices between schooling and In Sri Lanka, in Rohanapura, one respondent a
other investments such as improved said that: s
agricultural yields through use of fertilisers t
(sacrifice of alternative investments) and We sacrifice having a square meal and eat i
social investments such as the time or thalapa..We postpone buying clothesWe a
expenses needed to socialise with other postpone taking medicine for our illnesses (if it b
community members, for example attending is not urgent).(Sri Lanka report, p63) q
the funerals of those dying as a result of p
ARIs (sacrifice of social capital). t

96 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 115

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

As indicated in this comment treatment of what trade-offs were already being made
apparently non-urgent illnesses is one area within the household budget in order to
where temporary savings can be made. The send children to school.
study shows that this is a pattern across all
the countries in the sample. These sacrifices Overall, the patterns of response were similar
have short term benefits in terms of ability to the first question, a focus on satisfying
to pay school fees or purchase school items, basic needs, to the second question a shift
but the longer term costs in terms of towards investment strategies (e.g. income
increasing severity of health problems may in generation or putting more children in
many cases lead to more drastic effects on school). The implication that many of these
the household. households are already making sacrifices of
both discretionary and non-discretionary
Two case studies in Kenya contained the income as a result of the direct and indirect
following statements: costs of education, is unavoidable.
Table 5.24 overleaf summarises the data
Janet Nzioka, Mandani. She informed us according to broad classifications of
that she had not had breakfast and neither responses into investment/consumption
had she dinner the previous nightthe last categories. The pattern of responses in all
two (children) were in the local primary school cases clearly changes in favour of investments
for which she has to struggle to raise fees. as costs hypothetically drop.

: Sarah Syombu, Mandani. The family skips Table 5.25 extends the analysis a little further
meals frequently as a way of coping. showing the percentage of responses
indicating that satisfaction of basic needs would
l There is plenty of international evidence to be their first priority to such a drop in costs.
indicate that dietary deficiencies especially
n in the young leads to poor concentration, However, there were interesting differences.
lack of stamina and ultimately low achievement. With respect to the first question, in both
Uganda and Zambia there was less emphasis
. 5.8.6 Responses to a Drop in the Costs on food requirements than in Bangladesh
of Education and Nepal. For example, the purchase of
Two questions were asked in our survey livestock in Uganda accounted for 27% of
t about household responses to a change in responses, while in Zambia spending on
school costs. The first question asked what schooling costs was prioritised in 24% of cases.
the household response would be to a drop
in school costs by 50%, the second question One of the responses in all countries was to
asked what the household response would send another child to school. Table 5.26
t be to a drop in school costs by 100%. The indicates that the issue seemed least
questions were not asked separately for important in Uganda, perhaps because of
primary and secondary education. One of the recent policy changes there. There was
the purposes of these questions was to infer no clear pattern in other countries, but it

D DFID 97
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 116

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Table 5.24: Responses Catagorised as Investment/Consumption d

h
50% drop in costs 100% drop in costs
o
PO SBO PO SBO
t
Nepal
m
Investment 53% 53% 57% 66%
b
Consumption 47% 47% 43% 34%
d
Bangladesh
s
Investment 41% 36% 57% 63%
w
Consumption 59% 64% 43% 37%
b
Zambia
Investment 25% 28% 67% 65%
5
Consumption 75% 72% 33% 35%
Uganda
T
Investment 58% 68% 77% 79%
h
Consumption 42% 32% 23% 21%
t
Source: SSI Q 156 and 157
a
s
Table 5.25: Percentage of Respondents Identifying Basic Needs l
(Food, Health, Clothing, Shelter) a
o
50% drop in costs 100% drop in costs
t
PO SBO PO SBO
p
Nepal 38% 29% 32% 20%
r
Of which food 15% 11% 8% 4%
T
Bangladesh 34% 39% 12% 10%
a
Of which food 34% 39% 12% 10%
Zambia 47% 47% 11% 20%
U
Of which food 26% 25% 1% 1%
s
Uganda 26% 15% 24% 9%
s
Of which food 11% 11% 7% 4%
s
Source: SSI Q 156 and 157
i
b
Table 5.26: Another Child Can be Sent to School s
c
50% drop in costs 100% drop in costs
Nepal 12% 12%
I
Bangladesh 12% 8%
t
Zambia 4% 11%
Uganda 5% 4%
4

Source: SSI Q 156 and 157

p

98 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 117

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

did not feature strongly as a priority in these was to first of all increase labour hours, then
hypothetical circumstances (compared to to take loans and only finally to withdraw
other responses), again lending weight to children from school (c. 30% of responses at
the thesis that poor families are already the highest level of hypothetical costs). In
making trade-offs within the household Bangladesh withdrawing children from
budget, both within discretionary and non school was hardly mentioned at all focus
discretionary expenditures, in order to was exclusively on controlling households
school children and would, faced with a cash expenditures or taking on more work. But,
windfall, prioritise better quality of life in Uganda 38% and 46% of respondents for
before further investment in education. primary and secondary education respectively
indicated the temporary withdrawal of children
5.8.7 Responses to a Rise in the Costs would be their main response to the highest
of Education level of hypothetical costs. Borrowing was
The study also asked a question about the mentioned rarely in Uganda and Zambia
household response to a hypothetical rise in perhaps indicating very low levels of cash
the costs of education. Three rises were posited incomes and availability of credit in these sites46.
and the responses to each one recorded
separately. The hypothetical rises were set Qualitative data collected in field sessions
locally by the research teams taking into indicated that these figures do not capture
account the real direct and indirect burdens the complexity of some of the decision making
on households. The aim of the question was especially in relation to the gender and
to try to test the willingness and ability to birth order of the children to be withdrawn.
pay of respondents and to gauge their In Nepal it was commented that girls were
responses to changes in the price of schooling. sometimes withdrawn from school in order
The question was asked separately for primary to school another child (usually a boy) a
and secondary education. decision which may be compounded by the
stage of schooling reached and the higher
Unsurprisingly, the range of responses was costs met in higher grades of primary and the
similar to that expressed earlier in relation to transition to secondary school. In Kenya the
sacrifices the household already made for case studies of individual families reveal a wide
social functions (see earlier section). These range of reasons within the same family at
included taking on additional work, cutting different times - related to opportunity costs,
back on household expenses, borrowing, health, low achievement and birth order for
selling assets and withdrawing one or more withdrawal or drop out.
children from school.
Some of the other specific strategies mentioned
In Nepal the pattern of responses with respect in the SSIs and group sessions were:
to both primary and secondary education

46
Supporting evidence comes from Kenya where it was noted that in both communities borrowing took place through
bursary groups, but that access by the poorest to these groups is difficult because they cannot keep up with the regular
payments needed.

D DFID 99
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 118

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

selling milk to raise additional funds (Kenya) In Kenya it was reported that villagers in d
both study sites made compromises with l
not allowing children to continue to household assets: p
secondary education (all countries) a
many houses in Mandani village are p
asking older children to drop out so that made of cheap earth material and corrugated o
younger children could be schooled sheets. In Kaspinendet they are mud houses h
(Nepal, Uganda, Zambia) that are grass thatched. While parents would c
wish to live in more permanent structures e.g. c
not allowing children to attend tuition made of stone, they are not able to do so. b
classes (Sri Lanka, Zambia) (Kenyan report, p4-19) t
o
cutting out some leisure expenses such as In an Oxfam sponsored study of cost sharing i
alcohol or luxury goods like soap and undertaken in Kilimanjaro, Tanzania, it was (
sugar (Uganda, Nepal). commented that:

(
5.8.8 Sacrifices of Household Assets Women in Uru North stated that they are I
The value placed by some parents on the often faced with the choice of buying a bag of
a
education of their children as a solution to the sand to repair the house or pay the school fees
e
continuous cycle of poverty is also revealed by (Maarifa N. Ufinyuo, 2001, p35).
b
the attitude to the sale of the limited assets
d
possessed by households. In the Kenya study Schools often ask for voluntary labour from
c
there were examples of households selling land community members to erect buildings, clean
D
and cows to continue keeping children in school the school grounds etc. For example, in Nepal
t
(one respondent said she periodically sells small discussions with parents and children revealed

parts of her 3 acres of land in order to finance that parents and older children gave an
t
her childrens education). In Zambia and average of 5-7 days voluntary labour each
r
Uganda sale of livestock was a common per year to schools this too is an opportunity
c
response to household financial difficulties. cost of labour which could be used
e
In Nepal respondents said they would sell productively for the household.
g
jewellery and ornaments.
f
5.8.9 Other Issues
m
An interesting feature of the examples in
s
Kenya was the attitudes of the older, more (i) Timing of Payment Demands
s
established community, Mandani, where the The timing and cycles of household income
m
valuation of education had changed as a and expenditure also affect the choices about
b
result of disillusionment with the benefits to spending on education. There are both yearly
be derived. But in the other community, cycles and patterns which span more than one
Kaspinendet, these negative messages seem T
year. Yearly cycles may be affected by weather,
to be less apparent. A
harvests availability of labour etc. More
e
infrequent patterns may result from births,
f

100 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 119

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

deaths and marriages within the household Not only will this impact on teachers and
leading to increased expenditures of a class sizes directly, but also on children.
predictable but unavoidable nature. There is
also ample evidence that the timing of dropout rates among pupils are also likely to
payments for school costs can have an impact increase because of illness or because students,
d on participation in education and the sacrifices especially girls, are needed at home to look
households have to make in order to school after the sick (Zambian report)
children. In some cases it is not the level of
costs that causes difficulty to households, 1998 figures reported for Zambia suggested
but the timing by the time school fees come there were at that time already 470,000
to be paid available cash has been used for AIDS orphans. Although our study tried to
other purposes. Further discussion of this investigate and probe for the effects of ARCs
g issue can be found in the section on equity on attendance and the additional costs borne
(5.9) and costs of education (5.6). by households there was general reluctance
to talk openly and to attribute illnesses/deaths
(ii) Severe Shocks to AIDS. The reports of fairly high levels of
In addition to all the above there are sudden spending on health within household
and unpredictable shocks to the household expenditures may be one symptom of the

economy caused by ill health or death of problem (in Zambia health expenditures
breadwinners, disease affecting livestock, were second only to food at 9.7%, in Uganda
drought or floods and macro-economic they ranked fifth after food, clothing, fuel
conditions affecting demand for produce. and alcohol).
n
Depending on the nature of these shocks
l
there may be different household responses Other natural disasters such as floods
d
households may be in a position where (regular in Bangladesh, Nepal), droughts
they have to save or to earn and their and famine (Kenya, Zambia, Uganda) have
responses to the schooling of individual equally severe effects though the incidence
y
children will be different as a result. For may fall on limited numbers.
example, an economic slowdown which
gradually constrains cash income (through Mukesh Miya (a boy of seventeen) from
falls in prices or loss of labour opportunities) Chandhika VDC in Achham was fortunate
may result in girls being withdrawn from to be enrolled in a school. He had all the good
school occasionally or permanently. A sudden dreams of life, which were shattered to pieces
shock such as the death of a key income earner three years ago when his father, the only earner
may result in the permanent withdrawal of a in the family died, leaving behind his mother,
t
boy to take up an income earning role. four sisters and a brother. Mukesh had to
y
prepare himself as the bread winner of the
e
The effect of HIV/AIDS, especially in the family at the age of fourteen.
,
African sites is particularly alarming. For (Nepal report, p81)
example, in Zambia 26% of the teaching
force is thought to have ARCs (Kelly, 1998).

D DFID 101
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 120

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

(iii) Unwillingness to Make Sacrifices Not only does the annual agricultural cycle d
There were also occasional examples of create seasonal fluctuations in wealth, labour c
households who were unwilling to make demand, hunger and illness, years of drought p
sacrifices, despite the knowledge that other exacerbate the situation. Failure of crops o
households around them were doing so. In means no food and no income, and death c
some cases a feeling of powerlessness had led of livestock. c
to a fatalistic view of the households ability a
to squeeze more from the already constrained The lack of material goods and inability to H
budget. In other cases alcoholism or poor store food or increase livestock in good years t
budgeting had reduced the money available. as insurance for the lean ones increases the p
In others still a low valuation of education vulnerability of these households. They are e
led parents to decide that the sacrifices more at risk of falling into severe destitution.
required to school their children were not Those who own their own land, have larger T
worth the low economic return. areas to cultivate and have larger numbers of a
cattle were described as the better off. a
5.9 Equity Issues Social capital and human capital are also t
This study has in large part been concerned significant dimensions of vulnerability, and t
with exploring of the impact of schooling the latter, particularly the level of education e
costs on demand for education. This section of individuals within the household, a
brings together some of the equity themes determines how robust the household might p
highlighted earlier and examines in more be in withstanding shocks. e
detail the way in which the poorest and most t
vulnerable are often disadvantaged in their In Bangladesh, the haor areas of the country f
attempts to participate equally in schooling. experience annual inundation from monsoon o
An earlier section (3.4) defines equity, rains. Although this can be planned for (it O
poverty and gender. is regular) it still leaves the large populations T
who live in these flood-prone areas without
5.9.1 Poverty ready access to facilities, and particularly
T
Who are the poor, and how is poverty vulnerable to hunger, disease and lack of
manifested in the study countries? Vulnerability shelter. Schools close, or are washed away,
and lack of protection from external shocks teachers fail to get to the schools that remain
is one common factor. Not only does this viable, and education officers are not given
increase the likelihood of the perpetuation of the necessary travel facilities to supervise the
poverty, vulnerability also increases the chances schools. Cyclones destroy the livelihoods
of the non-poor becoming poor. In other and homes of those in the coastal cyclone
words, poverty is not a static state, but a prone areas; tidal surges have a similar
dynamic situation which fluctuates (sometimes devastating effect.
dramatically, at other times regularly but less
severely) year on year and within a single Despite attempts to improve the situation,
S
year. For example, both Kenyan study for example through the construction of
locations were drought and/or famine prone. schools on stilts to keep the classrooms dry

102 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 121

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

during the wet season, and cyclone shelter failure of the service providers to give
r construction, these do little to help the regularly available schooling (absence of
t poorest manage risk and pull themselves out teachers is high), and the poor perceived
of perpetual poverty. There have been quality of the schools all add to the
considerable increases in the numbers of disincentives to continue to the end of year
children in Bangladesh who enrol in school, five and the scholarship examinations.
and this applies to girls as well as boys.
However, schooling achievement is low, with 5.9.2 Local Perceptions of Poverty
s the percentage of children completing the The communities included in the research
primary cycle being far lower than of those studies were purposively selected to represent
enrolling in class one. poorer or the poorest population sub-groups.
. They are small samples and in locations which
The gap between girls and boys may not be representative of the norm.
f achievement is particularly noticeable, Their perceptions of poverty are subjective
although not well documented. There are and will therefore reflect the local situation,
two examinations in year five. For one of not be a definition of poverty that necessarily
these, the end of cycle examination which is travels well to other places. For example,
entered by the majority of children, results references to specific food items only eaten
are not nationally collated. Results are by the poor will not be universally
t published for the other, scholarship applicable, although the fact that in many
examination. The few children who proceed places the type of food eaten will identify the
to enter the scholarship examination in year socio-economic status of people can be a
y five are a low percentage of the whole, and general statement, as will not being able to
n of these only 45 % are girls (Table 5.27). eat regularly. Fishing communities will refer
Only 22 % of those girls enter the exam pass. to lack of boats, and livestock ownership will
The inability to regularly attend school, the be relevant to pasturalist communities.

Table 5.27: Bangladesh Primary Scholarship Examination, 1999

Division Total enter Girls enter % Girls Total pass Girls pass % Girls % Passed
n enter entered girls
passed
Rajshahi 1077818 46390 43% 32253 12508 27% 39%
Barisal 30414 14181 47% 5894 2348 17% 40%
Dhaka 123475 56557 46% 28376 11956 21% 42%
Khulna 51075 22079 43% 17772 6913 31% 39%
Sylhet 25479 12721 50% 4575 2041 16% 45%
Chittagong 99218 45902 46% 17495 7248 16% 41%
Total 437479 197830 45% 106365 43014 22% 40%
Source: Ministry of Education, Department of Primary Education, Dhaka, Bangladesh

D DFID 103
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 122

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

The urban poor will be more concerned off. The significant features of difference d
with sanitation and the rural poor with access between these two groups are the level of n
to or ownership of good growing land. The social capital accessible, and the degree of T
table below compiles all of the studys poverty voicelessness perceived. Both groups suffer s
definitions, and puts their references into from lack of physical assets, but the more h
general categories rather than specific terms. destitute a household the greater their lack p
By doing this some general features can be of social status and access to social resources s
determined which are of relevance when c
considering who might be excluded from 5.9.3 Poverty, Gender and t
obligations to pay for schooling, or be in Government Legislation t
receipt of scholarships to ensure that Uganda provides a useful example against t
children might get to school. Note also that which to judge the effectiveness of cost sharing
the list is compiled from the defining and the means available to overcome cost/price W
features of local perceptions of both the constraints of enrolment. The introduction s
poorest households and the slightly better of free schooling in 1997 saw an immediate e
w
Table 5.28: Local Perceptions of Poverty T

Group/catagory Feature
H
Economic Daily wage earner/work for others (kamaiya - bonded
labour - in Nepal) [

Subsistence agriculture/with poor grade land t

No livestock c

Receive public assistance/depend on others d

Capital goods/infrastructure Wattle and daub huts r

Usually landless or very small holdings s

No toilet
(No capital resources such as boats, land pukka housing) 5

May have to pawn or sell belongings to meet urgent needs I

Poorly dressed/second-hand clothes (Have a mad p

appearance because of their tattered clothes) g

Social features and other Cannot participate in social affairs/community activities d

consequences i
Have large families
Widow-headed household c

Migrate or travel to find work f

Hungary seasons (barely meet basic needs), food a

insufficiency, eat only poor peoples food e

No medical treatment e

difficult to educate children e

Alcohol addiction (usually male) s

Illiteracy A

Source: study reports of wealth-ranked exercises

104 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 123

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

doubling of enrolment rates, with a particularly exclude children from formal schooling.
noticeable effect on the enrolment of girls. Similarly, the Nepali Government has legislated
The government removed the payment of against bonded labour, but indebtedness
school tuition fees for four children in each perpetuates a de facto bonding system.
household, with at least 2 girls being given
priority. Despite this positive move, the In Sri Lanka the plantation Tamils have for a
study revealed that households still pay a long time suffered in comparison to mainstream
considerable amount to send their children Sinhalese. Our study showed marked
to school. These costs are associated with differences between the selected Tamil and
teaching and learning materials, travel and Sinhalese community though both were
time (see section 5.6). equally poor (a situation exacerbated in the
g North as a result of the war). In Zambia and
e What is of note in the areas included in the Uganda tribal differences lead to members
study is the seasonal variation in income and of some communities being ostracised or
e expenditure, with two rainy seasons coinciding treated as second class.
with households inability to pay for schooling.
The researchers summarise the situation: The use of aggregated data for planning
Insecurity, drought and epidemics (mainly purposes in any country will overlook the
HIV/AIDS) have impacted negatively on the needs of population sub-groups. Recognition
[households in the] communities ability to school of the multi-dimensional nature of vulnerability,
their children. Aggregated figures for the and poverty, and the diversity of forms in which
country as a whole show a falling off in poverty and vulnerability can be manifested
demand for school after the initial good requires flexibility and responsiveness of
response to government legislation change systems. This is particularly difficult when
see Table 5.30. certain common and universal norms are
required and when standards of quality have
5.9.4 Ethnicity to be determined and maintained. Cost sharing
In some districts in Nepal 20 % of the is no more exempt from this need for diversity
population belong to so-called disadvantaged of responses to fit the local situation than
groups: specific ethnic groups who are any other aspect of the service provision.
disadvantaged by most measures of well-being, What emerges clearly from the studies is the
including income, material assets, human fact that one size does not fit all: the
capital and access to health and education imposition of a single, unitary, and uniform
facilities. Their levels of education are low policy across all schools will effectively exclude
and they have the widest gender gap in significant numbers of sub-groups of children.
education parameters. Not only are they These sub-groups will vary from place to
economically vulnerable, they are socially place, and from time to time. Vulnerability
excluded, and some ethnic groups are will determine the latter variations over time,
socially prohibited from entering schools. and persistent poverty associated with other
Although legislation now prevents social stratifiers such as gender and ethnicity
untouchability social rules persist and may be associated with particular places.

D DFID 105
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 124

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

5.9.5 Orphans and HIV/AIDS countries, there were estimated to be over 1 5

HIV/AIDS was not an explicit focus of the million AIDS orphans in Zambia in 2000. T
study. However, sub-Saharan African i
countries are particularly severely affected by In Sri Lanka the civil conflict has also created i
the pandemic, with multiple knock-on effects a body of orphans with very poor access to f
on the education system. The death of schooling. In the active conflict zone, the t
young adults has two very obvious effects: uncleared areas, of North Eastern Province s
the decline in the number of qualified the physical infrastructure of schools is u
teachers (in Kitwe district in Zambia, for severely damaged, with many schools closed. g
example, in 2000 a 2.43 % death rate of (See An appraisal of the education system in d
trained teachers was noted a total of 13 the Vanni area of Sri Lanka, 1998, MOES,
male and 27 female teachers died) and the NEP and SCF(UK).) The number of E
increase in the number of orphans, child- qualified teachers available is limited, and the e
headed or elderly-headed households. support mechanisms of the education system s
Although the case has been made that is severely curtailed. Orphaned children are d
poverty is not the cause of lowered effective most affected in terms of education access in d
demand for schooling it has to be said that it a country where education has been e
would be difficult to sustain that argument acclaimed an international success story, s
for these especially vulnerable households. given the level of GDP. Grandparents are d
Evidence shows that where orphans are often the only relatives that children have p
housed by relatives following the death of recourse to. Where their grandparents are e
their parents they tend to be marginalised displaced (a feature of the lives of many in d
within those households. The additional NEP) they have no insurance mechanisms w
mouths to feed and bodies to sustain and against the economic shocks of additional e
clothe are often too many for the household family members: no land, no crops, no i
economy to maintain. Schooling is often livestock, no income, and often poor health. s
the most dispensable luxury, and orphans They do not have the where-with-all to send i
will be the last to benefit. their grandchildren to school. a
s
Similarly, where the eldest child has to take Despite the desire to attend school, and an W
responsibility for the rest of the family their understanding of the benefits that this might c
continued education is most unlikely. In bring, orphans and the households that p
addition they will not have access to support them are the least likely to be able t
resources or work sufficient to support the to pay (in cash or in kind) for education. e
well-being of siblings and their education. Cost sharing schemes have therefore to d
Even where government provision or voluntary accommodate the needs of these disadvantaged
organisation provision is available their ability children. Bursary schemes, scholarships or
to access these resources is often minimised similar strategies go part way, but they do
through powerlessness, lack of voice, and not address the underlying social and structural 4

s
poor communications. To give some sense disabilities that face these children in their d
L
of the scale of the problem in the study pursuit of education. p

106 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 125

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

5.9.6 Gender and Access to Education Gender as an analytical concept permits the
Throughout the study, and again very much exploration of male and female relationships
in line with other research findings, over time and place. Exploration of this
d irrespective of socio-cultural and economic kind shows that significant differences do
factors, girls fare less well than boys in access exist both between countries and within
to, involvement in and outcomes from countries. In Nepal, for example, many of
schooling. The study made attempts to the Terai communities practice purdah, the
understand gender differences in access, separation of women from men, and the
. gender differences in perceptions, and gender effective removal of women from public
n differences in outcomes. space. This does have an impact on
schooling, particularly at higher grades of
Earlier it was noted that the gender gap in primary or secondary school. As girls reach
e education achievements is related to dominant puberty it is no longer acceptable for them
m social constructs. Poverty may exaggerate to be taught in the same space as boys.
differences, for example when income
n declines and it is deemed necessary to reduce There are ways of tackling this through the
expenditure on education, it is daughters not schooling system, but success of initiatives
sons who stay home. But the gendered such as separate streams for girls and boys in
decision making is not a consequence of the same schools (as is practised in
poverty, but of gendered social and economic Bangladesh secondary grades), in part
expectations. Each of the countries studies depends on the socially determined demand
demonstrates this phenomenon, although for female education. In the villages of the
with varying degrees of robustness. For hill or mountain areas of Nepal women are
example, in Zambia it was noted: while there not restricted in the same way from
is an increasing number of females going to movement in public space.
school now, the trend on the part of most parents
d is to prefer to send their sons where they can Problems of access to school are certainly
afford to send one child. A very similar there: physical distance, trafficking of the girl
statement was made in the Nepal study: child47, and household needs, but the gender
Whether the reason is high cost or dependence on constructs are less prohibiting in these
t child labour or inaccessible schooling facilities or villages than they are in the Terai. Disincentives
poor quality, the girls are denied schooling more for poor households of rising education costs
than the boys when there is a low demand for are more likely to affect girls from poor
education. In many families, the education of households in both the Terai and hill or
daughters is easily renounced if a difficulty arises. mountain villages than they will affect boys.
d

l 47
The abduction or kidnapping of girls on the way to school primarily for the purpose of being sold as sex slaves is a
significant problem in Asia. This acts as a considerable disincentive to parents in situations where children have to travel
distances to school, and in remote areas. For further information on this issue visit the Asia Pacific Forum on Women,
Law and Development at www.apwld.org, where they usefully list organisations dealing with this problem, and titles of
papers on the topic.

D DFID 107
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 126

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

There are also clear changes that have been of attendance and is skewed higher when per I
and are taking place over time in what a man capita funding is available to schools, they t
or a woman can do, what is socially accepted provide a rough comparative indicator. c
and encouraged. The rise in the number of
women in the formal workforce in European Note the decline in enrolment in Kenya over v
countries is testimony to change, a change this five year period a decline which a
which is being replicated in some of the continued into the twenty-first century. In the w
South East and East Asian economies. Kenyan Governments interim PRSP strategy o
Along with these changes associated with paper, for example, they note: n
economic structures and diversification are r
changes in the profile of education statistics, After the high enrolments of the two post o
with, for example, womens literacy rates independence decades, there has been a reversal i
moving towards or overtaking mens. at all levels of education characterized by non- 5
enrolment, high level of drop-outs/low P
5.9.7 Enrolment Rates completion rates particularly among girls, and l
The situation in the countries included in poor transition rates from one level of education t
the study on cost sharing shows wide to the other. This is attributed to the high cost U
variations in enrolment rates. Although it is of education worsened by the burden of cost g
accepted that enrolment rates are a poor sharing which has had a negative impact on
indicator of process and success in reaching access, equity and quality of education. S
international development and education n
targets (primary completion rates are a better Although there appears to be no difference p
measure), generally enrolment rates are the between male and female enrolment, the h
only readily accessible measure across all IPRSP does state that girls in particular are
countries, and over time. Despite their affected by the negative education situation. T

inaccuracy, particularly when based on This is clearly demonstrated through this

extrapolated population data, and despite the cost sharing study.
fact that enrolment does not indicate levels

Table 5.29: Gross Enrolment Rates for the Six Study Countries, 1990 and 1995 S

1990 1995
T
GER M F GER M F
f
Nepal 108 132 81 109 126 91
Sri Lanka 106 107 105 113 115 112
Uganda 74 83 66 74 81 68
Zambia 99 N/A N/A 88 91 86
Bangladesh 72 76 66 N/A N/A N/A
Kenya 95 93 97 85 85 85
Source: World Bank: World Development Indicators, 2001 S

108 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 127

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

r In Uganda later national data are affected by countries where the main cost sharing study
the introduction of free education for four was conducted: Table 5.31 below.
children (two of whom should be girls) in
1997. This was immediately followed by a Similarly to the 1995 GER figures the most
r very rapid increase in the number of boys distinct gender differences of attendance are
and girls in school but the GER figures shown in Nepal. However, the selection of
e were distorted by the sudden influx of households for inclusion in the study as those
y overage children. It is claimed that the with school-going children clearly raises the
number of girls enrolled in primary schools percentage of children attending compared
rose particularly rapidly with the introduction to the national or regional figures. In
of free tuition (compared to boys), and this Bangladesh, as many if not more, girls are said
l is borne out by national NER figures (Table to attend school as boys, again similar to the
5.30). As the GOU acknowledge in their GERs available from the WB development
PRSP some of the gains in NER have been report statistics. In Uganda, however,
lost since the initial surge in numbers, and respondents provide a different picture to that
n there is still some way to go in meeting the presented in the PRSP: there are considerably
Ugandan Governments UPE goal, although fewer girls attending than boys. The locally
gender parity in enrolment has remained. collected figures for attendance of the
respondent households in Zambia are also
Some comparisons can be made between the different to the international database figures,
national GER figures in Table 5.29 and the showing a considerably lower attendance
patterns of attendance of children from the overall than the GER figures would suggest.
households of the study sites in the four These differences may reflect several things:

Table 5.30: Uganda Net Enrolment Rates

1992 1997 2000

All Girls Boys All Girls Boys All Girls Boys
Uganda NER 62 60 65 84 83 85 76 76 77

Source: Government of Uganda, Poverty Reduction Strategy Paper, 2000 (from IMF poverty web site)

Table 5.31: Currently Attending School: All Respondents to SSI: % of School Age Children
from Respondent Households Attending School

Boys Girls
Nepal all household 75 54
Bangladesh all household 67 70
Zambia all household 80 69
Uganda all household 52 51
Source: SSI Data

D DFID 109
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 128

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

the selection of the poorest households for young women can return to school after b
the study child-birth is an institutional factor which is a
change over time as recent GER figures indicated as a barrier removal for girls c
were not available for all countries education. Changes in attitudes have gone as s
attendance being different to enrolment far as one father stating that: I think a girl w
distortions to either datasets as should be educated more than a boy these days
respondents provide the answers they because men like themselves too much, and now 5
think the researchers desire, or there is girl-child education even if you become T
school-based data is inaccurately returned. pregnant you can go back and continue o
learning. This was not a common finding e
For more detailed breakdown of figures on the vast majority stated no preference but w
primary and secondary enrolment see the may be indicative of the generally more v
methodology section (4). positive attitude towards girls education. t
n
(Note that the smaller studies conducted in In Uganda, despite evidence of traditional w
Sri Lanka and Kenya did not provide this preferences for boys, especially in times of S
information.) economic hardship, alternative views were
also heard. A mother said they would prefer H
A further point to note about enrolment to educate their daughters during times of e
rates is that the most commonly available financial hardship because they are more o
figures are gross enrolment. This figure also likely to give material and financial help to C
includes over-age children and children who their parents than boys who will only think s
are repeating years. This distortion of the of their wives. s
figures complicates any assessment being a
made, and GER figures need to be read in In Nepal, where gender differences in access
partnership with figures on repetition rates. are very marked, it was notable that I
traditional attitudes had transferred to the p
5.9.8 Reasons for Fewer Girls Attending: next generation: the same views were
Gendered Perceptions expressed by school boys and school girls n
Although in all study sites boys education about the value of their education as those p
was preferred to girls in the past, there appears expressed by parents. c
to be evidence of changing attitudes. f
There were few differences noted between s
In Zambia, high rates of divorce are cited as male and female perceptions of the
a social change which necessitates the economic importance of education. There was some 4

independence of women, and hence an equal s

indication that female headed households, p
imperative for their education. The increasing despite their generally higher levels of n
c
opportunities for women to engage in poverty, were more determined to make sure i
p
income earning, and the harsh economic that their children would get an education. g
U
climate necessitating both marriage partners The understanding of education as a means w
earning income are also cited. The fact that p
of overcoming poverty was often referred to i

110 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 129

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

by women. What men and women value better indicator of success than exam results,
about education and the factors they and certainly more reliable, even taking into
consider determine the quality of schooling account problems with population statistics48.
s showed some gender differences, and these
were evident in the different studies. As was noted above, students themselves
suggested that children being able to read
5.9.9 Gender Differences in Outcomes and write were indicators of a good school.
The first point to determine is the ways that Their ambitions are rather more modest than
outcomes might be measured. In the studies passing public exams, and perhaps indicated
examination results are used as outcomes, their concern that this was not being achieved
which reflects the majority of respondents in their local schools. This is overtly stated
views that school quality is to be judged in by a participant in Rohanapura, Sri Lanka,
this way. For example an increase in the who claimed that students of grade five
number of students passing public exams cannot even write a complete sentence.
was quoted from the FGD in Rohanapura, In this same study parents and teachers both
Sri Lanka, as an indicator of a good school. claimed that children had not developed
intellectually and psychologically sufficiently
However, end point results such as to sit public exams. Large classes, lack of
examinations do not provide any indication qualified teachers, poor infrastructure and
of process or value added by the school. lack of teaching and learning materials
Confidence in the quality of examination contribute to this. As is seen in many
systems is low, with problematic or non-existent countries, particularly in rural and poorer
standardised testing, difficulties with cheating communities and especially after the
and the continuing reliance on rote learning. introduction of compulsory enrolment at
school, students enrol in relatively high
In the World Bank draft education and numbers but stay for the first three years of
poverty report (2001) they suggest that the primary cycle. There is then a rapid
100% primary completion rates, that is the drop-out between year three and year five.
number of children who enter school and
progress through to the end of the primary This pattern is demonstrated in the enrolment
cycle, is the most important long-term goal and attendance graphs from Sevanagala
for any low-income country. This might be school in the Sri Lanka study (Figure 5.10).
seen to imply that completion rates are a

48
The primary completion rate is defined in the World Bank Education and Poverty chapter as the total number of
students completing (or graduating from) the final year of basic education divided by the total number of children in the
population whose age is equal to the official graduation age. They acknowledge that: this measure is not perfect, as the
numerator will include under-aged and over-aged (late entry and repeater) students. Despite this imperfection, they
consider this a better measure than the often used GER which may show increases as the system become more inefficient
i.e. as there are more repeaters, and does not provide any indication of the number who may go on to graduate from
primary schooling. They suggest that: Achieving 100% primary education completion is the most important long-term
goal for any low-income country i.e. they do not refer to examination results as the most important measure or goal.
Universal testing of student learning as a means of comparing within and between countries is problematic at this time,
when standardised testing is not well developed, when rote learning is the norm and excessive stakes to student
performance create overwhelming pressures for cheating. Although other means of testing are recommended as an
interim measure (sample-based, national learning assessments), these are not currently practised.

D DFID 111
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 130

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Figure 5.10: Drop-out Patterns, Sri Lanka T

g
s
o
Enrolment and attendance of Male Students in the Sevanagala School in the Year 1997
Average of Enrolment/Attendance

40
T
i
t
30
r
p
20
Attendance
t
Enrolment
t
10 a
a
t

10

11

12
1

9
de

de

de

de

de

de

de

de

de

de

de

de
ra

ra

ra

ra

ra

ra

ra

ra

ra

ra

ra

ra
G

G
b
s
l
t
Enrolment and attendance of Female Students in the Sevanagala School in the Year 1997 s
50 o
r
Average of Enrolment/Attendance

40 r
p
o
30
r
T
20
Enrolment c
Attendance
g
10
d
t
a
10

11

12

g
1

9
de

de

de

de

de

de

de

de

de

de

de

de
ra

ra

ra

ra

ra

ra

ra

ra

ra

ra

ra

ra
G

t
i

(Note that boys re-appear in time to attempt enrolled in year 6 are swelled by those
I
the year five examinations, whereas the girls coming from primary grade only schools,
s
do not seem to do so. The student numbers enrolling for the secondary school.)
w

112 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 131

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

These graphs indicate a similar number of and rates did fall somewhat subsequently. In
girls and boys enrolling for school, and a the study areas a consistent pattern was noted
similar pattern of average attendance. In the of more boys than girls entering school, and
other studies some different patterns emerge. girls dropping out from school in greater
numbers. The relationship to poverty was
The Zambian study shows a clear distinction noted, with limited incomes making it
in the pattern of school attendance between difficult for families to educate all their
the urban location of Kitwe and the other children. Preference is given to boys. By
rural locations. Contrary to the national secondary grades, the gender gap widens,
picture which is described above, and contrary demonstrating a poor schooling outcome for
to the claims of neutrality over preference for girls. The figures appear to demonstrate that
the education of one sex over another, there female headed households are more likely to
are fewer girls attending school than boys, see a persistence of both boys and girls in
and there is a greater drop-out of girls before school through primary grades. Figures
the end of the seven year primary cycle. provided by the study team of overall
(Incidentally, 9 girls are reported to have attendance do not reflect year on year
been removed from one of the study site change, nor seasonal variations. However,
schools as they were pregnant [only 2 boys information on attendance immediately
left for health reasons]. It was also reported preceding the study demonstrates that nearly
that in Kitwe district 2.4% of the teaching half the children were not in school in the
staff died in the previous year, mostly because previous month. With such poor attendance
of AIDS related conditions. As it was also figures it is likely that results from schooling
reported that many of the pregnancies are the will be poor.
result of sexual abuse of the teachers on
pupils a very disturbing picture emerges, and Nepal illustrates the widest gender gap in
one which could certainly explain the schooling attendance patterns, in drop-out
removal of girls at puberty from school.) and progress through grades. By secondary
The fewer girls who do complete primary school (grade 6 and above) over half the
cycle are less likely to progress to secondary girls in the poorest households are not in
grades than boys, although there is a high school. The poor completion rates of
drop-out rate for both sexes. So, even primary education are particularly marked
though it is reported that changes in attitudes for girls. By the end of secondary grades
are taking place, and that more girls are being and the school leaving certificate in the
given the opportunity of schooling than in country as a whole less than half the number
the past, there is still a persistent gender gap of girls as boys enter the examination, and of
in outcomes. those that do enter a third of the number of
girls pass the exam compared to boys.
In Uganda the introduction of UPE in 1997 Outcomes from schooling by any measure
saw a rapid rise in enrolment rates, but as are very poor for the majority of girls.
was noted earlier, this success was not stable

D DFID 113
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 132

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

Table 5.32: Nepal School Leaving Certificate Results 1995 d

c
SLC Year Appeared: girls as % of total Passed: girls as % of total
K
1995 44.5 35.3
l
Source: HMG Nepal, Education Statistics 2000
b
In Bangladesh, although the figures provided persistence, and this is reflected in the
for the study area households show about examination figures which show fewer girls
75% of children continuing in school and no entering the state primary exam and girls
gender gap in those enrolled (and a similar achieving a lower score. This was attributed
number of boys and girls dropping out of by parents and teachers to the fact that girls
school), the real situation in terms of those were involved in many errands after school,
attending school demonstrates a poorer hence having less time for study.
situation with only 27% actually in school.
Unfortunately this figure has not been broken These Kenyan study communities are living
down by gender. Earlier, in the section on marginal existences with little or no social or
poverty (5.9.1) table 5.27 shows the primary economic capital and are vulnerable to
scholarship examination results. This external shocks. When exposed to such
demonstrates the persistent gender gap in shocks survival strategies include elimination
exam achievement throughout the country, of excess expenditure of household
and there is no reason to consider this will resources which include time, human
not be the case in the study areas. resources, energy (food consumption) and
money on education. Education becomes
Primary schooling in Kenya includes grades superfluous to survival in the immediate,
1 to 8, and the rapid decline in enrolment even if desirable in the longer term.
and attendance is seen between grades 5 and
8. There appears to be a greater fall in the What perhaps can be learned from this
numbers of girls persisting to the end of the example reiterates a point made earlier, which
primary cycle than in the number of boys. is that no one answer to cost sharing will fit
However, it is difficult to determine a pattern all situations and all locations. Policies must
of school attendance which reflects any be responsive to local needs, and those local
changing perception of the value of schooling needs may be very localised indeed. A further
for either boys or girls in a situation where point is also worth making, which is that no
natural disasters play such a significant role matter how good the schooling system, and
in disrupting behaviour patterns. As is noted no matter what redistributive measures
for one village (Kapsinendet) the fluctuations might be introduced through cost sharing
in enrolment year on year are a reflection of unless the vulnerability of communities such
the socio-economic pattern, with children as those included in the Kenyan study and
being removed from school in the hardest elsewhere, the vulnerability of poverty, is
years and returning in subsequent years. The tackled education will be a dispensable
second village setting demonstrated greater commodity. Frequently, but perhaps with a
gender disparities, in initial enrolment and downward trend, girls education has greater

114 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 133

Reaching the Poor - The costs of sending children to school

5 Discussion of Findings

dispensability than that of boys, but in severe

crises (such as the drought in 1998 in the
Kenyan villages) both boys and girls will
leave school, sometimes on a temporary
basis, sometimes permanently.

D DFID 115
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 134

Reaching the Poor - The costs of sending children to school

6 Policy Issues

6.1 Quality as Well as Quantity mentioned. For some, inappropriateness of

One of the consistent findings throughout the curriculum is the quality factor that
the studies and in all locations is the pushes the decision not to attend school (the
importance placed on quality of education students) or not to send their children to
as a determinant of demand. Quality is more school (parents).
important, or at least as important as price.
Assertions of desirability for particular school
Although there are some households in our features are also gendered. That is to say,
poverty-biased sample who find it difficult to mothers and fathers display some differences
send their children to school whatever the in what they consider to be important, and
quality, in the main households find it possible what they want for their sons and want for
to make considerable sacrifices in order to their daughters. Boys and girls require
school their children. They are more inclined different things of their schools.
to continue to make these sacrifices even in
periods of considerable adversity if they As has been found in Uganda, where costs of
consider the quality of education provided education were reduced, increases in the 6

makes it worthwhile to do so. They also demand for schooling attendant on that price T

indicate that they would consider paying more reduction have not been robust: the quality s

again with the proviso that the quality of of the education provided did not meet d

the school has to meet their needs. expectations and desires. The opportunity p

costs of schooling outweighed the benefits of c

However, quality is subjective, and what attendance. w

people are looking for from their schools

(whether this is the children/students It needs to be said too, that although there is I

themselves or their parents) is something consistency in the importance placed on e

more than currently exists from their local adequate quality, there is considerable m

establishment i.e. their definition of quality is heterogeneity in what is deemed to be a quality f

based on what they observe is missing from issue. Context specificity is significant: i

their own local school. quality issues vary from place to place. e
d

For example, where teachers are habitually Policy Implications s

absent from school, a better quality school is r

one described as having regularly attending The main quality issues highlighted in the s

teachers. Where teachers are thought to offer study are related to: the availability of m

a poor standard of teaching (where children teachers (attendance in school, numbers i

appear not to be or are not learning), the employed and their distribution the s

demand is for qualified teachers. Absence of Teacher: Student Ratio) and inadequacies y

violence or corporal punishment, better in teacher quality (subject knowledge, h

physical facilities, more and better teaching pedagogy and attitudes to students). This a

and learning materials, and nearer location leads us to conclude that greater attention i

(especially for secondary schools), are all to distribution, training, academic 4

116 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 135

Reaching the Poor - The costs of sending children to school

6 Policy Issues

supervision and management of the Socio-economic vulnerability is greater in

teaching staff should be a primary concern, some locations. The impact of flood,
e if children are to be encouraged to enrol drought, cyclones (e.g. in Bangladesh),
and attend school on a sustained and harvest failure and livestock loss results in
regular basis, at a minimum through to seasons of desperation: in some places
the completion of the primary cycle. irreversible poverty results, in others seasonal
ol poverty, but both have implications for
Unless quality improves to the satisfaction demand of education. Vulnerability does not
of those who are potential users of the necessarily mean poverty, but it does mean a
service any expectation of payment (cash tendency towards poverty during periods of
or in-kind) from them will not be met, difficulty, and does influence the likelihood of
progress towards achieving education continuous and regular attendance at school
enrolment targets will not be made and being interrupted.
children will be denied education.
f Poverty is not only lack of income, but lack
6.2 Diverse Contexts, Diverse Policies of food, reduction in resources or inability to
e This study provides clear evidence that access resources, homelessness or loss of
socio-economic and cultural issues impact on shelter, and an inability to engage in wealth
demand for education. These issues are creating activities. The loss or lack of social,
particularly context-specific and display human and financial capital which are
f considerable variations, within countries as indicators of poverty are not necessarily
well as between. evident in vulnerable households, but the
threat of these things becoming reality is
s In Nepal for example, the attitudes towards high, particularly when natural and physical
education of families who live in the hill and capital is destroyed or unavailable.
mountain areas are at variance to those of Borrowing from the sustainable livelihoods
y families in the Terai and this has a considerable approach49, recognising the
impact on demand. At secondary level this multidimensionality of peoples lives and the
effect is even greater, as puberty of girls dynamic nature of poverty is a pre-requisite
differently affects their ability to attend for the development of workable strategies
school. In the Terai, cultural considerations of policies of relevance to the population.
result in female students being removed from
school at puberty, and in the hills and The impact of health crises on vulnerable
mountains socio-economic and geographical households has the same potential to create
influences are more significant. The lack of irreversible or transient poverty as other
secondary schools (currently defined as from natural disasters, and to result in a decline in
year six) within close proximity to their education access. The HIV/AIDS pandemic
homes prevents or constrains daily travel, is an obvious example which is having a
and entry into schools with boarding facilities particularly devastating effect in East and
is a luxury few can or wish to afford. Southern Africa, and a lesser but growing

49
See the www.livelihoods.org for the sustainable livelihoods website

D DFID 117
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 136

Reaching the Poor - The costs of sending children to school

6 Policy Issues

impact in other places. This is still a disease school feeding programme which is targeted
which people find difficult to talk about, and to the poorest communities and the poorest
not surprisingly therefore, it did not feature children in schools in those communities, a
very strongly in responses to the SSIs. system which is complemented with a
stipend programme in other poorer
However, references to the impact of communities not included in the food
HIV/AIDS were made by research groups in programme. Other health related initiatives
Zambia and Uganda in their context in Africa have concentrated on health
descriptions. Informal discussions with education, including sexual health, many of
researchers and respondents provided other them through innovative programmes such
insights which have not necessarily been as child-to-child activities, peer learning
reflected in the final country reports. efforts and specialist NGO involvement in
Changes to household structures and their school activities.
impoverishment (loss of income earners,
grandparents taking on grandchild rearing Judging how effective these programmes
obligations, eldest child in orphaned have been in either reducing the levels of
households becoming head of household, illness, or preventing the further spread of
orphans being relocated) and the effect on HIV infection, and whether they have any
education system staff (increasing numbers effect on school attendance (as is claimed
of premature deaths, regular and costly with the Bangladesh food initiative) is
funeral attendance, dealing with students beyond the remit of this study. This
whose households are affected), clearly have information would make an interesting
significant effects on both supply and addition to the study findings, relating as it
demand sides of the education system. does to determining the most effective policy
to pursue in encouraging greater and more
The effect of socio-economic deprivation on effective school attendance.
health with a consequent decline in school
attendance and/or achievement is also evident Policy Implications 6
through the study. Much of this illness will A
be related to food or nutrition, water, sanitation As is implied in this section, the complexity w
and other basic needs. The interwoven and and dynamic nature of poverty cannot be w
complex synergistic relationships of the various addressed through education alone. While M
factors linking illness to poverty impact on a there are some education policy approaches i
childs ability to access and benefit from school. that target resources to the most needy r
children which might have positive t
Some countries have introduced intervention impacts, perhaps the more significant d
programmes in recognition of the link policy implication is the need for inter-
between poverty and education access, and sectoral coordination of efforts. The role A
also the relationship of food availability and of NGOs or CSOs could valuably be w
cognition. Bangladesh has a comprehensive explored in this context. o

118 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 137

Reaching the Poor - The costs of sending children to school

6 Policy Issues

Government responsiveness to exceptional counted within the poorest group of households,

circumstances needs to be strengthened, have a greater desire to see their children
including the need to identify communities educated, including their daughters, and will
who are particularly vulnerable and likely make considerable sacrifices to do so. It is
to be affected during years of especial not possible to state that women irrespective
adversity. Even though community of their marital situation or their level of
involvement is advocated in identifying education are always more supportive of
needs and managing resource distributions, education for their children, but their does
this should not result in Governments appear to be a tendency for this to be the case.
absolving themselves of responsibility to
provide additional supports which are It should also be noted that the education of
designed to prevent vulnerable households all children in a household is not treated
moving into poverty. identically. First order children will be given
preference if choices have to be made on
These different lived experiences of these who should benefit from schooling. In some
heterogeneous but vulnerable households locations where the education of boys is a
need to be understood and accommodated higher priority overall, first order boys and
in education financing policy. One size will then other boys are placed above all girls in a
not fit all - that is to say, a universal policy household when the situation demands that
prescription will not be successful in all not all children can attend school. In others,
places, indeed may be counter to if the first child is a girl she may be given
requirements to meet the EFA or UPE preference over lower order boys. There is no
goals. Identification of, and responses to, one pattern which will invariably be followed.
y location specific needs, must be built into
education planning skills and policy The policy implication of these household
development. distributional issues is to ensure that the
consequences of imposing payment
6.3 Intra Household Diversity requirements on households are understood,
Another finding which has policy implications, and that policies can be modified to ensure
y which is also an issue of heterogeneity, is that that negative outcomes do not occur for
within households there are differences. lower order children in poor households, and
Mothers and fathers do not necessarily share for girls in most locations. Where parents are
s identical views on the value of education, the to be included in policy change
reasons for educating their children, and dissemination, or if participatory needs
their attitude towards educating sons and/or assessments are to be conducted, it is
daughters also varies. advisable to have separate male and female
activities. It cannot be assumed that mothers
An interesting trend emerges from the study and fathers will respond in the same way, or
which suggests that women who are heads will identify the same issues to be addressed.
of households, despite their tendency to be

D DFID 119
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 138

Reaching the Poor - The costs of sending children to school

6 Policy Issues

With most poorest countries engaging in the learning, then clearly there will be a considerable T
production of poverty reduction strategy reluctance to continue making these sacrifices, l
papers, which require full stakeholder to contribute to schooling. p
participation and participatory poverty s
assessments, it is hoped that this would be Developing systems which satisfy parents or g
the case anyway. This research confirms the guardians need to know how their money is m
importance of differentiating between male being used would seem to be a valuable i
and female stakeholders within the same contribution in the pursuit of satisfactory s
socio-economic, ethnic or cultural group, cost sharing approaches. If well-designed, m
and the significance of age and birth order to these open and accountable systems could a
experiences and opinions. also be a substantial move towards providing
a greater sense of local ownership.
Policy Implications P
The study confirmed that the costs of schooling
Within the framework of a PRSP or (taking all direct, indirect and opportunity
equivalent, and in line with international costs into account) can be very high for the
development targets, national policy poorest income groups, even in a situation
should be developed which responds to where the government is not demanding
the education needs of the poorest and payments. Most countries (including those
most disadvantaged. in this study) subscribe to the international
EFA suggestion that basic and primary
Policy makers at all levels need to be aware education should be free. However, the
of the uncertainty of household responses reality is that schools are generally permitted
to cost sharing policies and to be aware and in many cases need to raise funds locally
that responses may well be shaped by the in order to cover recurrent costs.
gender of the household head and the
dynamics of household decision making. Although limits are put on the amount that
can be raised in this way, in reality it is
6.4 Local Decision Making, unlikely, and indeed difficult in remote
Accountability and Transparency locations, for government officers to ensure
One of the issues related to raising funds that regulations are being adhered to.
locally was identified to be the failure of the Government secondary schooling financing
education system to use methods which practices vary, but in general there are much
satisfied households need to know where higher costs for children to continue their
their money was going, what it was being education beyond the primary years.
used for. If quality improvements are not
evident, if head/teachers or government officers Where parents have no mechanisms or rights
appear to be benefiting while the school does to audit or manage the finances of schools, 5

f
not, if sacrifices are being made by households head teachers are in a position to exploit the n
but children are still reporting no teachers in p
situation for their own ends, or for what L
school and do not display any evidence of they deem to be the good of the school.
s
S

120 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 139

Reaching the Poor - The costs of sending children to school

6 Policy Issues

They may not abuse the system, and it is Learning from international experiences,
s, likely that the majority do not, but systems such as the establishment of
perceptions may be otherwise where no open School Management Committees should
systems of accountability exist. Encouraging be introduced or strengthened, and linked
greater community involvement in school into the whole school development
management, including financial management, approach referred to above. Their
is not an easily introduced strategy, but if responsibilities should extend to
successfully pursued may well provide the management of school financing, through
mechanisms for greater openness and which appropriate local resource raising
accountability, as well as encouraging more and distribution may be encouraged.
g ownership of schools by the local population. Appropriate training, supervision and
support will be required if any such
Policy Implications approach is to be successful50.
g
In order to respond to local needs relevant 6.5 Summary
decision-making should be devolved to None of the policy issues raised have easy
the level closest to the household: the policy solutions. The most significant issues
school. Whole school development are the need for the voices of the poor to be
planning can be encouraged with heard and responded to, for flexible systems
commensurate training and resources. to exist capable of responding to localised
These school development plans have the needs but without compromising standards
potential to enable schools in partnership of provision, and for a sense that the school
with communities (primary stakeholders) is responsible to the community and the
to prioritise their development programmes children it serves as much as it is to the higher
in response to local needs, and allow administrative levels of the system. Far from
heterogeneity of needs to be recognised the poor not being aware of the benefits of
and responded to. education, they in fact appear to make
sophisticated choices related to financial and
Financial allocations systems need to non-financial allocations to education. More
consider the introduction of mechanisms effective supervision of schools capable of
which respond to differential development supporting a continuing improvement to the
needs, with greater resources being quality of education being delivered, and
distributed to those with greater educational ensuring the safety of children when they
and socio-economic disadvantage. attend school, is a pre-requisite if parents and
children are to demand schooling on a regular
and sustained basis. Raising revenue, or

s
50
This report does not offer detailed information on policy prescriptions, but remains at the level of reporting findings
from the comparative study. These clearly have policy implications as we have, in very general terms, outlined. There is
now a considerable body of knowledge and information relating to whole school development, school development
planning and school improvement planning, to which the reader should turn for more detailed discussion (for example:
Levin and Lockheed, 1993; Hopkins, West and Ainscow, 1996; West and Ainscow, 1991; Cassells School Development
series of publications; Papers relating to the World Bank project Effective schooling in Rural Africa; Saunders (ed)
School Improvement Worldwide, the British Council, 2000)

D DFID 121
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 140

Reaching the Poor - The costs of sending children to school

6 Policy Issues

seeking non-monetary inputs for schools is A

possible, even from the poorest households P
but only if what is being provided meets 2
their and their childrens needs. Even when
they acknowledge the poor quality of the B
local school, parents indicate that they are E
making sacrifices to school their children. E
Increased commitment to the local school
clearly seems to be possible if the school (or B
the education system) reciprocates. a
L
G
a
s
S

B
o

B
R
c

B
P
E
W

B
E
W

B
S
F
P
E
U

122 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 141

Reaching the Poor - The costs of sending children to school

References

Annan, K. 2000, Millennium Report We the Colclough, Christopher, with Keith Lewin,
Peoples: The Role of the United Nations in the 1993, Educating All the Children: strategies
21st Century, United Nations, New York for primary schooling in the South, Oxford
University Press
Belfield, C. 1999, Economic Principles for
Education: Theory and Evidence, Cheltenham, Cumming, C.E., A. Brock, P.K. Kayiso, and
Edward Elgar. C. Opio-Awalu, 1995, Secondary Education
Costs and Finance Study. Cambridge UK
Bellew and King, 1993, in Leo-Rhynie, Elsa, Cambridge Education Consultants and
and the Institute of Development and Labour Kampala: Serefaco Consultants
Law, University of Cape Town, 1999,
Gender Mainstreaming in Education: Drze Jean and Amartya Sen, (eds) 1998,
a reference manual for governments and other Indian Development, Selected Regional
stakeholders, Gender Management System Perspectives, Oxford University Press
Series, Commonwealth Secretariat
Economic Commission for Africa, United
Blaug, M. 1990, Introduction to the Economics Nations Childrens Fund, The World Bank,
of Education Penguin, Middlesex June 1997, Consensus on Principles of Cost
Sharing in Education and Health in Sub-
Booth, David and Henry Lucas, 2001, Initial Saharan Africa, Report of an International
Review of PRSP Documentation, Report Forum held in Addis Ababa, Ethiopia.
commissioned by DFID, ODI, London Sponsored by: Ministry of Foreign Affairs,
The Netherlands, Swedish International
Bray, M. 1996 (a), Counting the Full Cost, Development Agency, Department for
Parental and Community Financing of International Development (United Kingdom),
Education in East Asia, Washington DC: The United States Agency for International
World Bank in collaboration with UNICEF Development

Bray, M. 1996 (b), Decentralization of Fine, B., and P. Rose, Education and the post-
Education: Community Financing, Washington Consensus in Development Policy
Washington DC: World Bank in the Twenty-first Century: beyond the
post-Washington consensus, 2001, Fine, B.,
Bray, M. 1999, The Private Costs of Public C. Lapavitsas and J. Pincus. London.
Schooling: Household and Community Routledge.
Financing of Primary Education in Cambodia.
Paris: UNESCO International Institute for Gertler P, and Glewwe P, July 1989, The
Educational Planning in collaboration with Willingness to Pay for Education in Developing
UNICEF Countries: Evidence from Rural Peru, LSMS,
Living Standards Measurement Study
Working Paper No. 54, The World Bank,
Washington, DC, USA

D DFID 123
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 142

Reaching the Poor - The costs of sending children to school

References

Government of Kenya, July 2000, Interim Mehrotra, S. and Delmonica E., 1998, P
Poverty Reduction Strategy Paper GOK (On Household Costs and Public Expenditure on P
IMF web site: www.imf.org) Primary Education in Five Low Income P
Countries: A Comparative Analysis. D
Hopkins, David, Mel West, Mel Ainscow, International Journal of Educational
1996, Improving the quality of education for Development, Vol. 18, No. 1, pp. 41-61. S
all; David Fulton Publishers. a
Mingat A and Tan J(1986) 'Expanding T
Igwe, S. O., 1988, Community Financing Education Through User Charges: What can
of Schools in Eastern Nigeria In Mark Bray be achieved in Malawi and other LDCs?' S
with Kevin Lillis, eds. Community Financing Economics of Education Review Vol. 5(3) , I
of Education: Issues and Policy Implications in 273-286
Less Developed Countries, Oxford, Pergamon S
Press. Ministry of Education and Science, Sri A
Lanka, NEP and SCF UK, 1998, g
Kelly, M., 1998, Primary Education in a An Appraisal of the Education System in the i
Heavily Indebted Poor Country: The Case of Vanni area of Sri Lanka, SCF Colombo J
Zambia in the 1990s, A Report for Oxfam
and UNICEF, Lusaka. Monk D., 1990, Educational Finance T
New York, McGraw-Hill G
Levin, Henry M, Marlaine E Lockheed, R
(eds) 1993, Effective Schools in Developing Narayan, D., Chambers, R., Shah, M, a
Countries; The Falmer Press, London. Petesch, P., 2000, Voices of the Poor: Crying A
Out for Change, World Bank, New York: M
Lewin, K and Wang Ying Jie, 1994, Oxford University Press
Implementing Basic Education in China: T
Progress and Prospects in rich, poor and national Narayan, D., Patel,R, Schafft, K., I
minority areas Paris, IIEP. Rademacher, A., Koch Shulte, S., 2000, J
Voices of the Poor: Can Anyone Hear Us?, V
Lipsey R., 1963, An Introduction to Positive World Bank, New York: Oxford University Press
Economics, Third Edition, London, T
Weidenfield and Nicolson. OECD, 1999, DAC Guidelines for Gender C
Equality and Womens Empowerment in E
Maarifa Ni Ufunguo, 2001, Cost Sharing: A Development Cooperation, OECD M
Case Study of Education in Kilimanjaro, Study O
Sponsored by Oxfam GB. Penrose, P., 1993, Affording the Unafffordable:
Planning and Financing Education Systems T
in Sub-Saharan Africa, Occasional Papers on O
Education 7. London. Overseas S
Development Administration. I

124 DFID
 23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 143

Reaching the Poor - The costs of sending children to school

References

Penrose, P., 1998, Cost Sharing in Education: UNESCO, 1990, World Declaration on
Public Finance, School and Household Education for All and Framework for Action to
Perspectives. Serial No. 27, London: Meet Basic Learning Needs, World Conference
Department for International Development on Education for All, Thailand

Seshamani, V, 2001, Cost Sharing In Zambia, United Nations, 1948, Universal Declaration
a report for Oxfam and the Jesuit Centre and of Human Rights, Geneva
Theological Reflection (JCTR)
Watkins, K., 2000, The Oxfam Education
Saunders, Lesley (ed), 2000, School Report, Oxfam, Redwood Books, Bath
Improvement Worldwide; The British Council
West, Mel and Mel Ainscow, 1991,
Sibbons, M., 1998, Approaches to Gender Managing school development: a practical
Awareness Raising experiences in a guide; David Fulton, London
government education project in Nepal,
in Gender and Development, Vol6, No.2, World Bank, 1990, World Development Report
July 1998, Oxfam Journal 1990: Poverty, New York: Oxford
University Press
Thin, Neil, Mary Underwood and Jim
Gilling, 2001, Sub-Saharan Africas Poverty World Bank, 1995, Priorities and Strategies for
Reduction Strategy Papers from Social Policy Education, Washington, World Bank
and Sustainable Livelihoods Perspectives,
A report for DFID, Oxford Policy World Bank, 1996, Zambia Education Sector
Management, Oxford. Expenditure Review, September 1996

Tilak, J., 1991, Family and Government World Bank, 2000, World Development
Investments in Education International Report, 2000/2001: Attacking Poverty,
Journal of Educational Development New York: Oxford University Press
Vol 2(2) 91-106
s World Bank, 2000, Overview of evidence and
Tilak, Jandhyala B. G., 1997, Lessons from issues concerning Effective Schools; Effective
Cost Recovery in Education. In Marketizing Schooling in Rural Africa Project.
Education and Health in Developing Countries:
Miracle or Mirage? edited by Colclough C. World Bank, 2001, World Development
Oxford: Clarendon Press, pp. 63-89 Indicators 2001, Washington D.C.
:
Tumushabe J., Barasa C., Muhanguzi F., and World Education Forum, 2000, The Dakar
Otim-Nape J., 1999, Gender and Primary Framework for Action. Education for All:
Schooling in Uganda, Research Report 42, Meeting Our Collective Commitments,
Institute of Development Studies, Sussex UNESCO, France

D DFID 125
23650 DFID 47 reaching the poor 1/5/03 9:37 am Page 144