Education Sector Public Expenditure Tracking

and Service Delivery Survey in Zambia

EDUCATION GLOBAL PRACTICE

Education Sector Public Expenditure Tracking and Service Delivery Survey in Zambia

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Education Sector Public Expenditure Tracking and Service Delivery Survey in Zambia

December 2015

© 2016 International Bank for Reconstruction and Development / The World Bank1818 H Street NW, Washington, DC 20433Telephone: 202-473-1000; Internet: www.worldbank.org

Some rights reserved1 2 3 4 19 18 17 16

This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries.

Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved.

Rights and Permissions

This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free  to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions:

Attribution—Please cite the work as follows: World Bank. 2016. “Education Sector Public Expenditure Tracking and Service Delivery Survey in Zambia.” World Bank, Washington, DC.

Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation.

Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank.

Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images.

All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org.

Cover photo: © World Bank. Further permission required for reuse.Cover design: Bill Pragluski, Critical Stages, LLC.

Contents iii

Contents

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixAbbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1Education Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2Education Sector Financing and Management: PETS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3Education Inputs: QSDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4Summary of Policy Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9Country and Sector Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9Major Findings from PETS-QSDS 2002 and 2006 and a Brief Comparison with PETS-QSDS 2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10Policy Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14

2 PETS-QSDS Study Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152014 PETS-QSDS Survey Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152014 PETS-QSDS Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18

3 Education Sector Finance and Management: PETS . . . . . . . . . . . . . . . . . . . . . . . . .19Education Budget Process and Expenditure Flow (General Education) . . . . . . . . . . . . . . . . . . . . . .19Flow of General Education Expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21School Finance and Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30

4 Education Sector Performance and Service Delivery: QSDS . . . . . . . . . . . . . . . . .31Education Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31Education Inputs and Service Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44Teacher Quality and Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50Student Performance and Education Inputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64

5 Major Findings and Policy Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65Education Sector Financing and Management: PETS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65Access, Internal Efficiency, and Service Indicator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66

iv Contents

Teacher Management and Student Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70

Appendix A Personalities and Motivations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71

Appendix B Literature Review on School Inputs and Learning Outcome . . . . . . . . .81

Boxes4.1 Calculation of Gross and Net Enrollment Rates Using LCMS, ZDHS, and ESB Data . . . . . . . . . .344.2 Measuring the Career and Prosocial Motivations of Head and Other Teachers

in the 2014 PETS-QSDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .594.3 Classroom Size and Contract Teachers: Kenya Extra Teacher Program . . . . . . . . . . . . . . . . . . . . .614.4 Measuring Soft Skills in the 2014 PETS-QSDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62

Figures3.1 Budget Process in 2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .203.2 Primary Responsibility for Budgeting at the School Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .213.3 Decision Making on Financial Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .213.4 Flow of Public Funds in General Education (Grades 1–12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .223.5 Government Expenditure Flow in General Education, 2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .233.6 Primary Teacher Salary as a Ratio to GDP per Capita Compared across African Countries . . . . . . 233.7 School Revenue per Pupil and Share of Public Funding in Primary

and Secondary Education, by Income Level of Students . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .243.8 School Revenue per Pupil, by Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253.9 School Expenditure for Primary and Secondary Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253.10 Percentage of Schools That Received a Grant or Textbooks, by Province. . . . . . . . . . . . . . . . . . . .263.11 Percentage of Grant Amount Explained by the Budget Allocation Rule . . . . . . . . . . . . . . . . . . . .283.12 School Grants per Pupil and Share of Schools Receiving a Grant, by Family

Income of Students . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .283.13 Percentage of Schools Reporting Shortages and Making Requests . . . . . . . . . . . . . . . . . . . . . . . . .294.1 Number of Schools Offering Each Grade Level, 2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .324.2 Number of Basic (Grades 1–9) and Secondary (Grades 8–12) Schools

and Enrollment, 2008–13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .334.3 Enrollment per School, by Grade Level and Province, 2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .334.4 Gross Enrollment Rate, by Grade Level and Gender, 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .354.5 Gross Enrollment Rate, by Grade Level and Province, 2010 and 2015 . . . . . . . . . . . . . . . . . . . . . .364.6 Percentage of Out-of-School Children, by Age and Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .374.7 Completion Rates, by Grade Level and Gender, 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .374.8 Transition Rates, by Grade Level and Gender, 2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .384.9 Transition Rates, by Grade Level and Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .384.10 Reasons That Girls Drop Out of School, by Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . .394.11 Reasons That Boys Drop Out of School, by Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . .394.12 Pregnancy among School-Age Girls, 2008–14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .404.13 Percentage of Teachers Asking Students to Leave Because of Pregnancy or

Nonpayment of School Fees, by Education Level and Rural-Urban Location . . . . . . . . . . . . . . . .404.14 Trends in Grade 5 Student Learning Assessment, 1999–2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41

Contents v

4.15 Grade 5 and 9 Learning, by Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .424.16 Grade 5 Learning, by Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .424.17 Grade 5 and 9 Learning, by Income Tercile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .434.18 Distribution of Grade 5 and 9 Learning, by Students’ Family Income . . . . . . . . . . . . . . . . . . . . . .434.19 Distance from the School to a Commercial Bank, by Education Level . . . . . . . . . . . . . . . . . . . . .454.20 Travel Time to School, by Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .454.21 Schooling Hours and Number of Shifts, by Grade Level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .464.22 Schooling Hours, by Grade Level and Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . .464.23 Multigrade Teaching in Grade 5 and 9 Schools, by Urban-Rural Location . . . . . . . . . . . . . . . . . .474.24 Number of Pupils per Classroom, by Urban-Rural Location and Education Level . . . . . . . . . . .484.25 Pupil-Teacher Ratio, by Grade Level and Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . . . .484.26 Pupil-Classroom Ratio, by Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .494.27 Pupil-Teacher Ratio, by Grade Level and Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .494.28 Pupil-Textbook Ratio (per Five Students), by Subject, Education Level,

and Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .504.29 Qualification of Grade 5 and 9 Teachers, by Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . .504.30 Qualification of Grade 5 Teachers, by Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .514.31 Teacher Assessment in Grades 5 and 9, by Subject and Rural-Urban Location . . . . . . . . . . . . . .514.32 Relationship between Teacher Qualifications and Subject Knowledge Assessment . . . . . . . . . . .524.33 Official Attendance Rate (June), by Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .524.34 Teachers’ Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .524.35 Number of Absent Days (during a Sample Week) among Grade 5 and 9 Teachers,

by Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .534.36 Reasons Given by Grade 4 and 9 Teachers for Being Absent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .544.37 Type of Teacher in Government and Grant-Aided Primary Schools, by Province . . . . . . . . . . . .544.38 Number of Teachers and Attrition Rate, by Level of Education, 2006–13 . . . . . . . . . . . . . . . . . . .554.39 Transfer Rates of Grade 5 and Grade 9 Teachers between Urban and Rural Schools . . . . . . . . . .554.40 Percentage of Grade 5 and 9 Teachers Wanting and Requesting to Transfer,

by Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .564.41 Percentage of Teachers Receiving Training in 2013, by Qualification and Education Level . . . .564.42 Percentage of Teachers Receiving Training in 2013, by Grade Level,

Urban-Rural Location, and Remoteness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .574.43 Training Location, by Grade Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .574.44 Action Taken by Head Teachers to Address Teacher Absenteeism, by Education Level . . . . . . .574.45 Lessons and Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .594.46 Minute-by-Minute Teaching Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60A.1 Motivations of Grade 5 Teachers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71A.2 Big 5 Personality Traits of Grade 5 Teachers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .72A.3 Motivations of Grade 5 Students: Inventory of School Motivation Score. . . . . . . . . . . . . . . . . . . .72A.4 Big 5 Personality Traits of Grade 5 Students . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73

TablesES.1 Short- and Long-Term Strategies for Achieving the Overarching Goals for

Zambia’s Education Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71.1 Comparison of Key Findings from PETS-QSDS in 2002, 2006, and 2014 . . . . . . . . . . . . . . . . . . .111.2 Indicators of Major Education Inputs in 2014 QSDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13

vi Contents

2.1 Data Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .162.2 Survey Modules, Content, and Respondents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .162.3 Planned and Final Sample Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173.1 School Revenue (Private and Public Sources) in Primary and Secondary Education, 2014 . . . .243.2 School Grants Received by Public or Grant-Aided Schools (Annual), by Urban-Rural

Location and Education Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .263.3 Annual Disbursement and Receipt of Primary School Grants, by Province . . . . . . . . . . . . . . . . .274.1 Zambian Education System Change, 1996–Present . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .324.2 Number of Schools Offering Each Grade Level, 2013 and 2014 . . . . . . . . . . . . . . . . . . . . . . . . . . .324.3 Gross and Net Enrollment Rates, by Grade Level and Data Source, 2010–14 . . . . . . . . . . . . . . . .344.4 Repetition and Dropouts, by Grade Level and Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .374.5 Student Absence, by Grade Level, Urban-Rural Location, and Gender . . . . . . . . . . . . . . . . . . . . .414.6 Grade 5 and 9 Learning Assessment, by Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .424.7 School Infrastructure, by Education Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .454.8 Infrastructure at the Primary Level, by Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .464.9 Teaching Hours in Grade 5 and 9 Schools, by Urban-Rural Location . . . . . . . . . . . . . . . . . . . . . .474.10 Teachers’ Use of Pedagogical Tools, by Grade Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .594.11 Intra-Cluster Correlation: Variance in Text Scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .605.1 Short- and Long-Term Strategies for Achieving the Overarching Goals for

Zambia’s Education Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70A.1 Teacher Transfer and Attendance and Physical School Environment . . . . . . . . . . . . . . . . . . . . . . .74A.2 Teacher Transfer and Attendance and Teaching Intensity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74A.3 Head Teachers’ and Other Teachers’ Motivation Questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . .74A.4 Head Teachers’ and Other Teachers’ Motivations and Teacher Transfer and Attendance . . . . . .75A.5 Relationship between School-Level Input and Type of Teacher Contract and

Student Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .76A.6 Relationship between Student-Level Inputs and Student Learning . . . . . . . . . . . . . . . . . . . . . . . .76A.7 Student Learning Outcomes in Relation to with Teacher Subject Knowledge . . . . . . . . . . . . . . . .77A.8 Student Learning Outcomes in Relation to Student Personalities and Motivations . . . . . . . . . .78A.9 Student Learning Outcomes in Relation to Teacher Personalities and Motivations . . . . . . . . . .78

Foreword vii

Foreword

The Public Expenditure Tracking Survey (PETS) and Quality Service Delivery Survey (QSDS) report is a product of a microscopic analysis of the performance of the Zambian Government in the education and skills sector. The report brings out independent viewpoints on the status of the education sector in Zambia; that is, it sheds light on the achievements that the sector has made in implementing the strategies and interventions intended to provide quality education service to the young citizens.

The PETS-QSDS report comes at an oppor-tune time when the country is in the process of assessing its performance in the achievement of the Millennium Development Goals (MDGs) and Education for All (EFA) goals. While numerous milestones toward the realization of the above-stated goals have been attained, there is still more that needs to be done not only to enhance donor but also public confidence in service provision and management of resources in the sector.

The report is cognizant of the huge invest-ment that the Zambian Government has made and continues to make in the education sector through the allocation of a significant propor-tion of its national budget to the sector. It is the

desire of the Government, therefore, that these resources are utilised prudently in order to achieve the targets and goals outlined in the education policy document, “Educating Our Future,” Vision 2030 as well as the Revised Sixth Development Plan (2011–16).

On behalf of the Zambian Government, I would like to thank the World Bank for profes-sionally undertaking the surveys whose findings and recommendations constitute this report. The Ministry of General Education commits itself to study the findings and recommendations herein proposed in order to restrategize the manner in which service is provided at all management lev-els of the Ministry. It is evident that some of the findings and recommendations will require pol-icy implementation repositioning to enhance the attainment of the Sustainable Development Goals (SDGs) on education by the year 2030.

Hon. Dr. John J. N. Phiri, MP. Minister of General Education

Acknowledgments ix

Acknowledgments

The Public Expenditure Tracking Survey-Quantitative Service Delivery Survey report was  prepared by the Education Global Practice of  the World Bank Group. The authors are Hiroshi Saeki, Seo Yeon Hong (lead author), Lauren Marie Gardner, and Shinsaku Nomura. Kundhavi Kadiresan and Sajitha Bashir pro-vided guidance and suggestions throughout the study. Tsuyoshi Fukao provided excellent sup-port on finalizing the report and coordinating between the stakeholders. As peer reviewers, Vijay Pillai, Dhushyanth Raju, and James Habyarimana (Georgetown University), and Vaikalathur Ravishankar and Nava Ashraf (Harvard University) provided excellent com-ments for the report and survey design. Celia Dos Santos provided timely and helpful admin-istrative support to the team.

This report was financially supported by the Department for International Development of the Government of United Kingdom and benefited from comments provided by Tanya Zebroff. The guidance of members of the Public Expenditure Tracking Survey Steering Committee enriched the report, offering with their insights into educa-tion in Zambia. The PETS Steering Committee is led by Owen Mgemezulu (Ministry of Education, Science, Vocational Training, and Early Education) and is composed of the Department for International Development, Japan International Cooperation Agency, United Nations Children’s Fund, the Embassy of Ireland, and representatives of civil society. This report relied on critical support from the National Assessment Survey teams of the Examination Council of Zambia and the RuralNet Associates.

Abbreviations xi

Abbreviations

AWPB annual work plan and budgetAY academic yearCBU Copperbelt UniversityCSO Central Statistical OfficeDEBS District Education Board SecretariesECZ Examinations Council for ZambiaESB Education Statistical BulletinETP Extra Teacher Program, KenyaGDP gross domestic productLCMS Living Conditions Measurement SurveyMESVTEE Ministry of Education, Science, Vocational Training and Early EducationMTEF medium-term expenditure frameworkNAS National Assessment SurveyPEO Provincial Education OfficePETS Public Expenditure Tracking SurveyPPS probability proportional to sizePTA parent-teacher associationQSDS Quantitative Service Delivery SurveyTEVET Technical Education, Vocational Entrepreneurship TrainingTEVETA Technical Education, Vocational Entrepreneurship Training AuthorityUNZA University of ZambiaZDHS Zambia Demographics and Healthy Survey

Executive Summary 1

Executive Summary

BackgroundZambia’s economy has experienced steady growth over the past years, and the Ministry of Education, Science, Vocational Training and Early Education (MESVTEE)1 continues to maintain the highest share of government expen-diture. Zambia’s economic growth was robust at around 6.5 percent (estimate) in 2014, exceeding growth of 4.6 percent for the whole of Sub-Saharan Africa and 2.6 percent for the world (World Bank 2014).2 The main drivers of the country’s strong economic growth are a bumper maize harvest, rapid expansion in the construc-tion industry, and strong growth in services. As the economy grew, expenditure on education also increased, from K 1.5 billion in 2006 to K 9.4 billion in 2015 (budget) in nominal terms, and from 3.9 to 5.0 percent of gross domestic product (GDP). Education expenditure as a percentage of total government expenditure also reached more than 20 percent, the single largest share of the proposed budget in 2015.

As government education expenditure grew, student enrollment and some education inputs improved in general education (grades 1–12). Enrollment in primary school increased from 2.9 million students in 2008 to 3.1 million in 2013;3 enrollment in secondary education grew from 617,394 students to 743,1754 during the same period (MESVTEE 2013). School infrastructure also improved. More schools now have a library, science laboratory, electricity, and latrines than in 2006, when the last Public Expenditure Tracking Survey (PETS)–Quantitative Service Delivery Survey (QSDS) was conducted. The number of pupils per classroom declined from 97 to 70 between 2008 and 2014, before the intro-duction of teaching shifts. By introducing teach-ing shifts, schools have managed to maintain the number of pupils per classroom at around 33 without reducing the number of school days and  teaching hours. The number of pupils per

teacher also declined from 59 to 40 during the same period.

The MESVTEE has accelerated the educa-tion   reforms to continue improving education outcomes. The education system in Zambia is in the midst of reforms at all levels. General educa-tion is now transforming from basic (grades 1–9) to primary (grades 1–7) education and from high (grades 10–12) to secondary (grades 8–12) edu-cation. Curriculum reforms have made good progress over the past few years, introducing and accommodating new interventions, such as information and communication technology, soft skills, and technical and vocational skills. New syllabi are being implemented at various levels of early childhood education and at grades 1, 5, 8, and 10. In 2013 the MESVTEE established the Higher Education Authority under the Higher Education Act of 2013 to coordinate all higher education. In addition, the Teaching Council of Zambia was established under the Teaching Profession Act of 2013 in order to reg-ulate teachers, their practice, and their profes-sional conduct and to accredit and regulate colleges of education. To improve the quality of teachers, three teacher training colleges are being upgraded to university status. The government intends to enhance results-based management in all ministries, and introduction of the output- based budget (OBB) is one step toward this. The MESVTEE was chosen to pilot the feasibility of such a budgeting system. Instead of being allocated to directorates, resources will now be allocated to programs and subprograms and linked to outputs. This shows the MESVTEE’s strong commitment to final deliverables rather than just inputs.

While the government’s efforts and reforms are commendable, they have yet to be translated fully into student learning outcomes. Student test scores at grades 5 and 9 have stagnated for a decade and there has been virtually no change in student scores in English, mathematics, science,

2 Executive Summary

and Zambian language, which have remained below the benchmark of 40 percent. In 2007 Zambia was ranked the second lowest on student achievement by the Southern and Eastern Africa Consortium for Monitoring Educational Quality. The government is cognizant of the stagnation in student learning outcomes, and the Examinations Council for Zambia (ECZ) has been conducting biannual assessments of student learning and teachers, with rigorous analysis of the data for the past years. It has gradually accumulated evi-dence and factors relevant to explaining the low learning outcomes of students and the low com-petency of teachers.

This PETS-QSDS report aims to supplement the current government’s efforts to gather evi-dence on the ground and to incorporate this evi-dence into policy discussions as well as the output-based budget. Through an analysis of data collected from the PETS-QSDS, this report aims to answer the following policy questions: (a) Does MESVTEE efficiently and effectively allocate its public funding to the areas most in need? and (b) What factors potentially keep stu-dent learning outcomes low? The report addresses the first question using data on education sector financing and management from the PETS and the second question using data on education inputs collected from the QSDS.

Education PerformanceEducation performance has been characterized by rapid growth in enrollment and low internal efficiency over the past decade. Total enrollment and the number of schools have expanded, and primary education can accommodate almost all primary school-age children (gross enrollment rate of 99 percent in 2010 and 102 percent in 2013, reflecting a problem in the official calcula-tions). However, secondary education has to improve its capacity to absorb more students (gross enrollment rate of 61 percent in 2010 and 59 percent in 2013). Equitable access to primary education is still in progress. Furthermore, delayed enrollment and repetition are still

common in primary grades, and student absen-teeism remains high.

As universal primary education makes prog-ress, net enrollment and completion rates for pri-mary education, in particular, need to be improved. While gross enrollment rates for both primary and secondary are relatively high, at 99 and 61 percent, respectively, net enrollment rates remain low, at 73 and 40 percent, respectively, according to the 2010 Living Conditions Measurement Survey (LCMS). This reflects some improvement in primary education, but little improvement in secondary education over the last decade. In 2015, according to preliminary findings from the 2015 LCMS, primary and sec-ondary net enrollment rates were 78 and 43 percent, respectively. According to the LCMS and the 2006 Public Expenditure Review, the net enrollment rate for primary education was 68 percent in 1998 and 75 percent in 2002. In 2010, 72 percent of children ages 16–18 completed pri-mary education (grade 7), 48 percent of individ-uals ages 18–20 completed lower-secondary education (grade 9), and 24 percent of individu-als ages 20–22 completed secondary education (grade 12). If completion rates were calculated by taking into account only the age cohorts for pri-mary and secondary education, the rates would be even lower.

Student learning outcomes have also been persistently low, despite improvement in some of the quantitative indicators of education service, especially infrastructure. Possible reasons for low learning outcomes include inefficient teacher management, poor-quality teaching (such as insufficient subject knowledge, especially at the secondary level), and a serious shortage of text-books. There has been no improvement in stu-dent learning at grade 5 since the first national learning assessment was conducted in 1999 (34.3 percent in math and 33.2 percent in English in 1999 compared with 35.3 and 31.4 percent, respectively, in 2014).

In addition, it is important to recognize that each province is at a different stage of educational development and requires solutions specific to its needs. There are stark differences across provinces

Executive Summary 3

in educational access and service indicators. For  example, provinces with almost universal education (for example, Copperbelt Province) face a shortage of physical education inputs and suffer from crowded classrooms and high pupil-teacher ratios, while Eastern Province lags far behind in educational access (gross enrollment rate of 85 percent in primary in 2010) and has the longest travel time to school (more than one hour). In provinces with almost universal educa-tion, expanding the existing school structure and using temporary classrooms and contract teach-ers would be realistic solutions. In provinces lag-ging behind in educational access, it is important to identify the cause, and the central government should support the province in solving the prob-lem. Simple statistics regarding school location and travel time reveal that long distance to schools may be a barrier to achieving universal education. Supporting large numbers of small-scale commu-nity schools may be a realistic solution, especially given the fact that public funding is very tight. However, all of these suggestions should be designed and evaluated carefully prior to full application.

Education Sector Financing and Management: PETSMajor Findings

School revenues (from both public and private sources) are characterized by inefficiency and inequity. There are substantial differences in revenue per pupil at the school level, depending on the number of rich (or poor) students. Schools with more rich students are likely to have more revenue (both public and private sources), while those with more poor students are likely to have less revenue. For instance, at the primary level, schools with more poor stu-dents have revenue of K 35 per pupil per year, while those with more rich students have reve-nue of K 46 per pupil per year. At the secondary education level, the difference is stark: K 144 per pupil per year at poor schools and K 390 per pupil per year at rich schools.5 There is also a

huge disparity in school revenue per pupil across provinces. While schools in Lusaka Province, on average, have revenue of K 75 per pupil per year, schools in Western Province have revenue of only K 7 per pupil per year.

Zambia has a formula for allocating school grants, but grants are not distributed according to the allocation formula, which causes ineffi-ciency and inequity. First, the MESVTEE, includ-ing provinces and the District Education Board Secretaries (DEBS), does not distribute public funding efficiently: 19 percent of grant-eligible schools do not receive any public funds, and 28 percent of primary schools and 70 percent of secondary schools do not receive school grants. School grants for primary education are a critical element of the government’s free education pol-icy. Schools reportedly do not receive school grants primarily because the provinces do not seem to follow the grant allocation formula. Further, there is a wide gap in the percentage of schools receiving public funds (including text-book grants) across the provinces, with 97 percent of schools in Lusaka Province having received a grant or textbooks, compared with only 69 percent of schools in Eastern Province.

The primary school grant is pro-poor, while the secondary school grant seems to support schools with richer students. At the primary level, on average, schools with relatively poor students receive grants averaging K 16 per pupil, while schools with relatively rich students receive grants averaging K 7 per pupil. At the secondary level, schools with relatively poor students receive K 15 per pupil, while schools with rela-tively rich students receive K 19 per pupil. Further, while poor students in primary schools have more chance to receive school grants, poor students in secondary schools have less chance to receive them.6

Policy Implications

Public funding needs to be targeted more strategi-cally to poor students and schools with more poor students. At the primary level, despite the pro-poor funding of school grants, the inequity in

4 Executive Summary

school finance persists. This is especially a prob-lem when poor students have lower learning out-comes than rich students. In Zambia, the scores for students from lower-income families are con-centrated around the lower mean, with little dif-ference among poor students. When schools with more poor students have fewer resources to spend, the learning gap between poor and rich students gets larger. At the secondary level, school grants should be distributed strategically toward the more economically vulnerable students. Unlike primary school grants, secondary school grants do not have a formula for compensating poorer schools. Having well-developed guidelines for grant formulas at the Provincial Education Offices (PEO) and schools could dampen the inequity in finance at secondary schools.

Financial decentralization (that is, depositing school grants from the Ministry of Finance directly into school accounts) is strongly recom-mended. In the meantime, the process of distrib-uting and receiving school grants should be formalized and strictly enforced, with strong accountability at all levels. At the beginning of the academic year, the official document describ-ing the school grant formula in each province along with the expected timing of disbursement should be publicly available. Furthermore, at the end of the academic year, the actual disburse-ment of school grants at the DEBS level and the disbursement dates should be publicly available. All of this information could be distributed pref-erably through official notice via the MESVTEE website but also through formal notice to the PEO, DEBS, and schools via the mail. Further, it is strongly recommended that the government expedite the financial decentralization of school grants by disbursing grants from the Ministry of Finance directly to school accounts. This would reduce any potential leakage and increase the accountability of the DEBS for school grants. Currently, the Ministry of Finance distributes grants directly to secondary schools, but not to primary schools, where remoteness and bank charges for withdrawals hinder direct deposit. These bank charges should be eliminated before direct deposit is implemented.

In addition, the central government needs to improve the predictability of funding from the Ministry of Finance to the MESTVTEE, to the PEO and DEBS, and then to schools. Unpredictable delivery of school grants and funding for train-ing  worsens school resource management and adversely affects the results of potentially success-ful programs. Output-based funding focuses on education outputs; however, without a steady and reliable funding flow, the expected outcomes will be difficult to achieve.

Education Inputs: QSDSMajor Findings

As a result of the government’s efforts to improve education services by increasing education expenditures and recruiting more teachers, some service indicators have improved over the past decade. There are now more schools with access to a library and electricity (13 and 36 percent, respectively, up from 5.6 and 19 percent) and fewer students in each classroom (70 pupils per classroom and 40 pupils per teacher, down from 97 and 59, respectively). However, access to drinking water has deteriorated since 2007.

The low pupil-textbook ratio is a serious issue, especially given that textbooks and a library are strong predictors of student learning outcomes. Five primary school students share less than one textbook for each subject (1.0 for mathematics, 0.9 for English, and 0.9 for science) and five sec-ondary school students share between 1.0 and 1.5 textbooks depending on the subject (1.0 for math, 1.7 for English, and 1.0 for science). Given the fact that having a textbook raises a students’ math scores by 3 percent, the low pupil-textbook ratio should be at the center of education policy discussions.

The low pupil-textbook ratio reflects deficien-cies in current textbook policy, including insuffi-cient funding for textbooks, misalignment within textbook policy, and lack of funds for textbook delivery. First, the current textbook budget can-not fully cover the procurement of all of the text-books needed. Second, the timing of curriculum

Executive Summary 5

development and procurement policy is not coor-dinated. In 2013 the delivery of textbooks was sig-nificantly delayed due to the development of new curricula and the lack of procurement capacity in decentralized units (DEBS and schools). The textbooks with the revised curriculum were pro-cured centrally and not published and delivered until the middle of the 2013 academic year. Third, there is no secured budget for textbook delivery, and the DEBS tend to deduct the cost of textbook delivery from the school grant amount.

Further, there is evidence that teacher management is inefficient: (a) teacher attrition is high, and a large percentage of teachers want to and do transfer schools, (b) teacher absentee-ism  is also high and virtually unchanged over the past decade,7 and (c) teacher subject knowl-edge is insufficient, especially at grade 9. While approximately 50 percent of teachers want to transfer, especially teachers in rural schools, 53 percent of actual transfers happen at grade 5; at grade 9, 44 percent of actual transfers are from urban to rural schools. In a given month, 16 percent of primary school teachers are absent for more than half of school days. With regard to teacher subject knowledge, 50 percent of grade 5 teachers score below 90 percent in grade 5 sub-ject materials, and 50 percent of grade 9 teachers score below 70 percent in grade 9 subject materials.

Teacher management (attendance and trans-fer) is strongly correlated with the motivation of the head and other teachers. In general, schools where the head and other teachers have higher pro-social motivation and intend to remain in the community have lower rates of transfer requests and teacher absenteeism than schools where they do not.

The 2014 PETS-QSDS investigated factors correlated with student learning outcomes in Zambia at the school, teacher, and student levels. At the school level, a library, a lower pupil-teacher ratio, and a longer school day are correlated with higher student learning outcomes. At the teacher level, the use of contract teachers significantly increases the learning outcomes of students, and a teacher’s subject knowledge has a strong positive

correlation with student learning outcomes. In addition, teachers who want to be respected by students and the community and those who are characterized by neuroticism are positively cor-related with student learning outcomes. At the student level, family income is a major factor in determining student learning outcomes: low- income students have lower learning outcomes than high-income students. Several important factors such as textbook ownership and school attendance are correlated with students’ family income as well as learning outcomes. Students who want to be recognized by teachers and their peers have higher scores than those who care less about recognition. Further, students who can agree with others (agreeableness) have higher scores than those who have less agreeableness.

Policy Implications

Concretizing and clarifying the textbook policy to ensure on-time delivery of textbooks should be the first priority of MESVTEE. The misalign-ment between textbook distribution and curric-ulum design and between procurement and delivery can cause serious delays or even no text-book delivery in a given academic year. Such delays critically harm student learning, as shown by the strong correlation between textbook avail-ability and student learning outcomes. The pro-cess of textbook distribution is unclear, especially with regard to the centralization or decentraliza-tion of procurement. Before implementing full decentralization of textbook procurement, the government needs to diagnose the availability of textbooks in local markets and the transparency of textbook prices as well as the capacity of schools. This information is a prerequisite for the successful decentralization of textbook procure-ment. Furthermore, the lack of funds for deliver-ing textbooks to remote schools also delays or obstructs the delivery of textbooks. It is strongly recommended that the government (a) increase public funding and collaborate with partners to ensure that public funding is sufficient to procure all of the necessary textbooks, (b) create a har-monious system of curriculum development and

6 Executive Summary

textbook procurement and delivery, (c) secure the funds for textbook delivery both at the cen-tral and at the DEBS levels, and (d) inform and build the capacity of all stakeholders involved in the process, including the central government, the PEO, the DEBS, schools, publishers, and others.

To improve teacher management, the govern-ment is advised to conduct a thorough evalua-tion of the recruitment and deployment process as well as introduction of a teacher performance feedback program. The importance of hiring teachers with pro-social motivations who view teaching as a calling should be highlighted dur-ing recruitment, performance feedback, and management training. This report finds a posi-tive correlation between the personality and motivation of the head and other teachers and the rate of teacher transfer requests and atten-dance as well as student learning outcomes. Further investigation of whether and how factors such as motivation, personality traits, and home-town influence student learning outcomes is needed to help the MESVTEE to design innova-tive approaches to the recruitment and deploy-ment of government teachers. At the same time, monitoring of teacher performance and a proper feedback mechanism are needed.

In order to implement monitoring and feed-back, it is necessary to develop teacher perfor-mance metrics that capture multiple aspects of teacher quality (not only student performance but also teacher satisfaction, motivation, peda-gogical style, and subject knowledge) that are proven to be effective in the Zambian context. The ECZ has turned its attention to teacher performance; the data collected through PETS-QSDS could be used as a baseline, and the MESVTEE could conduct a robust analysis, such as an impact evaluation, of motivation and per-sonality traits. Further, it is essential to develop the capacity of central, provincial, and district education officers and resource centers to inter-link teacher performance with the feedback mechanism.

The use of an efficiently managed contract teacher program could potentially address the

issues of deployment and absenteeism as well as the shortage of rural teachers. International stud-ies show that hiring young teachers for the short term can be effective in three ways: reducing the pupil-teacher ratio, lowering absenteeism, and improving student learning. However, if the gov-ernment intends to introduce such a program, it should be contextualized for Zambian schools.

Key to the success of a contract teacher pro-gram is transparency of the recruitment process at the local level. This transparency could be achieved by empowering school management and parent-teacher associations and training them to recruit teachers who have the proper education level as well as sufficient knowledge of the subject matter they will be teaching. At the same time, a contract teacher program could be a gateway to recruiting high-quality civil service teachers who would stay in the local area after receiving positive performance evaluations for several years. Furthermore, examining the applicant’s motiva-tion and work incentives during recruitment would help schools to select better-suited contract teachers, especially in remote rural areas.

Teacher training could reinforce the efficient use of contract teachers and community schools. Currently, the government supports in-service training of government teachers, and, partly due to this, more-educated teachers (those with at least a certificate) receive more in-service training than less-educated teachers (with only a general certificate of education or lower degree), who are likely to be contract or community school teachers. As mentioned earlier, limited school access and some aspects of teacher management could be addressed by setting up more small-scale com-munity schools and hiring young contract teach-ers, especially in remote areas. These teachers likely will need proper training to be effective.

Understanding the personality and motiva-tion of both teachers and students is needed to inform the design of better classrooms and peda-gogical styles and to improve student learning. This report’s rudimentary findings can be a start for concretizing and formalizing the develop-ment of soft skills. At the same time, motivation is a good indicator of how long head and

Executive Summary 7

classroom teachers will stay in the school and of student achievement. If measured and tracked properly, these indicators would be helpful for improving teacher management policy in the long term and deepening understanding of how Zambian students respond to teachers’ motiva-tion and personality; this information would be helpful for informing the teacher training and recruitment process in the long run.

Besides handling education inputs, school financing, and school management, the MESVTEE should continue to build its capacity to collect and use statistics. For instance, it is well known that the current Education Statistical Bulletin has some limitations that need to be addressed (for example, the net enrollment rate goes beyond 100 percent). It is recommended that the MESVTEE along with the cooperating partners take full advantage of the survey data available in the country (such as the LCMS) and periodically verify the consistency of

data across data sets. The ECZ is working on building its statistical capacity, and ECZ officials have benefited from several international training sessions on statistics. Now they have the capability to conduct large-scale surveys—from designing to sampling to analyzing data.

Summary of Policy StrategiesThis report recommends the following overarch-ing goals for Zambia’s education policy:

• Have each province devise solutions specific to its own needs

• Improve the predictability of funding from the Ministry of Finance to MESVTEE and to the PEO, the DEBS, and schools.

Table ES.1 divides these goals into short (one to two years) and medium (three to five years) strategies.

TABLE ES .1 Short- and Long-Term Strategies for Achieving the Overarching Goals for Zambia’s Education Policy

Type of policy Short-term strategies (1–2 years) Medium-term strategies (3–5 years)

Textbook policy • Review textbook procurement capacity at both DEBS and the MESVTEE and establish clear guidelines regarding textbook policy, including book selection procedures and distribution mechanisms and budgets

• Ensure that sufficient budget is released to schools for distributing all textbooks with the new curriculum

• Inform and build the capacity of all stakeholders involved in textbook procurement

School grant distribution • Enforce the school grant formula used by the DEBS to distribute school grant funds to primary and basic schools and monitor how well it is followed, including the public dissemination of transparent and clear guidelines regarding grant distribution

• Decentralize financial disbursement of school grants from MESVTEE or the Ministry of Finance directly to school bank accounts

Teacher recruitment, deployment, and attendance

• Conduct more research on teachers (including untrained, “contract” teachers and volunteers) to review how to improve attrition, effectiveness, and attendance

• Conduct a pilot on how the motivation of the head teacher and classroom teachers can improve the retention of quality teachers

• Prioritize training on leadership and management of head teachers to improve teacher attendance and delivery

• Have the Ministry of General Education Permanent Secretary send a circular to the PEO, the DEBS, and head teachers advising them to enforce teacher attendance and refer to human resources and public service guidelines

• Using results of the teacher study, revise, implement, and enforce teacher deployment regulations and explore other types of teachers who could teach more effectively

• Design and implement a teacher recruitment and deployment policy based on the findings of the pilot study on teacher motivation

• Revise the monitoring and evaluation strategy to include explicit measurements of teacher attendance

table continues next page

8 Executive Summary

Notes 1. The MESVTEE was split into two ministries—the

Ministry of General Education and the Ministry of Higher Education—in September 2015. This report, however, continues to consider the MESVTEE as one ministry.

2. However, as of October 2015, Zambia was experi-encing economic difficulties, with a high inflation rate (14 percent) and a copper export industry hit by a power shortage (due to low hydropower), low demand, and weakening currency.

3. About 85 percent of children 5–14 years old. 4. About 49 percent of children 15–19 years old. 5. Poor schools are defined as the bottom 33 percent

of schools in average family income of students, and rich schools are defined as the top 33 percent of schools in average family income of students.

6. In principle, under the free primary education policy, all primary students are supposed to receive primary school grants.

7. The 2014 PETS-QSDS uses the same definition of  teacher absenteeism as the World Bank

services delivery indicators definition of “absence from school” to distinguish absentee-ism from attendance (attended days) based on administrative records: “To measure absence, in each school, 10  teachers were randomly selected from the list of all teachers during the first visit to the school. The whereabouts of these 10 teachers was then verified in a second unan-nounced visit. Absence from school is defined as the share (of a maximum of 10 teachers) who could not be found on the school premises dur-ing the unannounced visit.”

ReferencesMESVTEE (Ministry of Education, Science,

Vocational Training and Early Education). 2013. Educational Statistical Bulletin. Lusaka, Zambia.

World Bank. 2006. “Zambia Education Sector Public Expenditure Review.” World Bank, Washington, DC.

———. 2014. Global Economic Prospects. Washington, DC: World Bank.

TABLE ES .1 continuedType of policy Short-term strategies (1–2 years) Medium-term strategies (3–5 years)

Teacher performance and training

• Review teacher evaluation systems (such as the Annual Performance Appraisal System) to see what is fit for purpose

• Develop a new way of assessing teachers (including head teachers) by developing teacher performance metrics

• Review and revise teacher recruitment procedures (at central and decentralized levels)

• Devise an effective teacher evaluation system for recruitment, performance evaluation, and feedback on all teachers and administrators at the school and district levels

• Continuously monitor teacher performance and provide feedback with a form of teacher training

Note: DEBS = District Education Board Secretaries; MESVTEE = Ministry of Education, Science, Vocational Training, and Early Education; PEO = Provincial Education Offices.

Introduction 9

Chapter 1

Introduction

Country and Sector BackgroundZambia’s economy continues to experience sta-ble growth and is expected to continue growing through 2018. Zambia’s economy is estimated to grow at around 6.5 percent in 2014, despite the modest global growth of 2.6 percent (World Bank 2014) Zambia’s growth exceeds Sub-Saharan Africa’s 2014 growth rate of 4.6 percent. The main drivers of strong growth are a bumper maize harvest, rapid expansion in the construc-tion industry, and strong growth in services. The outlook for 2015 remains positive, despite certain risks due to both domestic and external factors. Economic growth is expected to reach 6.7 percent in 2015, but this positive economic prospect is subject to both the political envi-ronment after the death of President Michael Sata in October 2014 and the global financial markets.

Youth in Zambia play an important role in maintaining the current level of economic growth; however, their human capital is not being used efficiently. Zambia is one of the youngest coun-tries in the world. In 2012 the country’s popula-tion was estimated at 14.4  million. Of the total population, 45.3 percent were below 15 years of age. Youth between ages 15 and 24 account for more than 40 percent of the total population, rep-resenting a youth bulge that has the potential to yield a demographic dividend if complemented by economic and social opportunities for youth. While youth constitute 30 percent of the total labor force, their unemployment rate is more than 15 percent, which is almost twice the average unemployment rate of 7.9 percent. This means that, of the unemployed population of 459,132, almost 60 percent are youth ages 15–24 (Zambia Central Statistical Office 2013). One of the rea-sons for the high level of youth unemployment is that youth tend to lack the skills, competencies,

and attributes required by the labor market (Koyi, Masumbu, and Halwampa 2012).

The government puts strong emphasis on skills development and allocated the single larg-est share of the proposed budget to the education sector in 2015. Job and employment creation, especially for youth, is at the center of govern-ment policies. The policies of the Fifth National Development Plan 2006–10 and the Sixth National Development Plan 2011–15 address the need for high-quality education as the founda-tion for high-quality human capital. This com-mitment is reflected in the steady increase in public education expenditure over the past decade. Education expenditure as a percentage of GDP rose substantially from 3.9 percent of gross domestic product (GDP) in 2006 to 5.0 percent in 2015 (budgeted). Education expenditure as a percentage of total government expenditure also reached more than 20 percent, the single largest share of the proposed budget in 2015.

Besides a budget increase, the education sector has experienced significant changes at all levels of the education system over the past decade. To name a few, the basic and high school education system was introduced in 1996, but after 15 years, the sys-tem returned to the primary and secondary educa-tion system. The government now promotes free  primary and secondary education policies. There was a major policy reform in technical education and vocational entrepreneurship training (TEVET) in 1996, and the Technical Education and Vocational Entrepreneurship Training Authority (TEVETA) was established in 1998 to ensure that TEVET institutions adhered to the newly estab-lished quality assurance system. As  part of the reform, the TEVET system was decentralized and institutions became semiautonomous, with almost full academic, administrative, and financial inde-pendence. Hence TEVET trainers are no longer

10 Introduction

public servants. At the higher education level, the Higher Education Authority was recently estab-lished. It aims to improve the quality of higher edu-cation through structured quality assurance, governance and regulation, and registration mechanisms.

The education reforms have had a positive influence on the education sector. While many education reforms are currently under way, they have achieved some positive outcomes. For instance, after the initiation of free primary and secondary education, enrollment in general edu-cation (grades 1–7 in primary education and grades 8–9 in secondary education) increased from 3.5 million in 2008 to 3.8 million in 2013. In particular, enrollment in secondary education continues to expand, mainly due to the growing number of graduates from primary education. While the TEVET sector is underfunded com-pared to other education subsectors, the number of public TEVET institutions is increasing and, in general, showing good performance in their financial statements. Relatively larger public TEVET institutions yield a surplus almost every year, and even smaller institutions cover most of their operational costs. The higher education sys-tem, too, has experienced rapid growth. There are three public universities in Zambia: University of Zambia (UNZA), Copperbelt University (CBU), and Mulungushi University, which was established in 2008. UNZA and CBU account for the majority of student enrollment in higher education, with total enrollment of almost 6,000 students in 1994 and approximately 30,000 in 2013 (World Bank 2015).

The strong financial commitment of the gov-ernment to the education sector and the continued expansion of the education system have not yet fully translated into strong student learning out-comes. The Sixth National Development Plan addresses the importance of focusing on quality improvement in, among others, the supply of teachers (recruitment, deployment, and retention), teaching and learning materials, infrastructure development, school governance, teacher manage-ment, quality assurance, teachers’ continuing pro-fessional development, and pedagogical support.

However, student learning outcomes remain low in all assessed subjects (mathematics, English, life skills, and Zambian language for grade 5 and mathematics, English, and science for grade 9). Zambia’s scores were all below the benchmark scores of 40 percent in 2014. Indeed, Zambia is close to the bottom in the rankings of the Southern and Eastern Africa Consortium for Monitoring Educational Quality.

Major Findings from PETS-QSDS 2002 and 2006 and a Brief Comparison with PETS-QSDS 2014Zambia conducted a Public Expenditure Tracking Survey (PETS)–Quantitative Service Delivery Survey (QSDS) in Lusaka, Copperbelt, Northern, and Eastern provinces in 2002 and in all prov-inces in 2006, focusing on basic education. Since the education system, including type of schools (for example, basic and primary education) and budget allocation mechanism, has changed sub-stantially over the past decade, some results from PETS-QSDS in 2002 and 2006 cannot be directly compared with results from PETS-QSDS in 2014. For instance, basic (grades 1–9) and high (grades 10–12) schools are now in transition to become primary (grades 1–7) and secondary (grades 8–12) schools, respectively. The budget allocation mechanism has changed from a formula based on a fixed amount to a formula based on different characteristics of school such as location and a size of school. Despite the changes in the educa-tion system, the 2014 PETS-QSDS sought to build on the major findings from 2002 and 2006. Major findings of the previous PETS-QSDS in 2002 and 2006 are summarized in table 1.1.

2002

Undertaken in 2002 for the first time, the educa-tion sector PETS-QSDS in Zambia provided a baseline assessment of the country’s funding structure and service delivery. The survey found major differences in efficiency and equity based

Introduction 11

on the type of funding. Rule-based funding at the central level seemed to be efficient and dis-bursed progressively, with more resources reach-ing poorer schools (more than 90 percent) than richer schools. However, larger discretionary funds disbursed at the province and district lev-els had high leakages, with less than 25 percent of schools receiving any discretionary funding. Richer schools also had a higher probability than poorer schools of receiving discretionary funds. With all funding sources considered, public school funding appeared regressive, with almost 30 percent higher allocations to richer schools.

For households, school-related expenditures other than fees were seven times their school fees and a major source of inequality.

Based on the success of rule-based funding in reaching poorer schools, the report recom-mended that the government increase the per-centage allocation of this type of funding. Public funds should also be targeted at the school level to improve funding to poorer districts. Due to crowding out of private expenditures by public funding, the report concluded that focusing solely on increasing public funding may not have the most desired impact and recommended

TABLE 1 .1 Comparison of Key Findings from PETS-QSDS in 2002, 2006, and 2014Indicator 2002 2006 2014

Rule-based allocation to schools Fixed amount Formula-based Formula-based

US$600 or US$650 fixed amount per school

“Basic school grants” + “free basic education requisites”

Primary school grants

High school grants Secondary school grants

More than 90% of grants reach schools

Textbook grants (sector pool) Textbook grants

Basic school: about 98% of government-owned schools report receiving grant funding from government or a donor sector pool in 2006

Primary school: 81% of grant-eligible primary schools (government and grant-aided schools) receive a school grant, including textbook grants, from MESVTEE during the 2013 academic year

High school: 90% of disbursed funds reach schools

Secondary school: 30% of secondary school grant-eligible schools receive the granta

52% of textbook grants disbursed at headquarters reach the DEBS

23% of textbook grants are disbursed from the Ministry of Finance to MESVTEE, and 82% of textbook grants are disbursed from MESVTEE

Infrastructure (% of schools, unless otherwise noted)

Library — 5.6 13

Science lab — — 3

Potable water — 92 74

Electricity — 19 34

Number of students per latrine — 77.1 84

Number of girls per latrine — — 76

Number of pupils per classroom — 97.1 70

Number of teachers per house — 3.1 1

Spot absence rate 17 22 18

Number of pupils per textbook — Basic school: math, 5; English, 4; science, 6; Zambian language, 9

Primary school: math, 5; English, 5; science, 6

Pupil-teacher ratio 42 (urban); 61 (rural) 59 40

Sources: 2002, 2006, and 2014 PETS-QSDS.Note: — = not available.a. Sample schools are both basic and secondary or high schools offering a grade 9 education. This captures previous basic schools not registered as secondary schools. Therefore, many schools eligible for secondary school grants probably do not receive them.

12 Introduction

concentrating instead on improving inputs, such as teacher quality, that households cannot provide on their own.

2007

A second round of PETS-QSDS in 2006 traced changes made based on the previous recommenda-tions and identified remaining or new gaps. Notable improvements occurred in public funding after 2002. Allocations per pupil rose, donor funding became streamlined, public funds incorporated more rule-based funding, and discretionary fund-ing became slightly more progressive. However, many of the same issues remained. Teacher salaries continued to be regressive due to higher pupil-teacher ratios in remote and poorer areas. Schools still received a low proportion of basic education funding despite decentralization of the flow of funds from the province to the district level. Disbursement delays actually increased, keeping more discretionary funds from schools. Institutional development lagged behind the increases in public funds, with shortages of learning materials, limited school infrastructure, low supervision, and use of double-shift school days. Both teacher and student absenteeism remained high: the average absence rates were 20 percent for teachers and 25 percent for students.

Policy recommendations from the 2007 report focused on increasing accountability throughout the flow of funds. Problems with an unclear fund-ing formula and poor communication could be addressed by simplifying and publicizing the “rule” for school grants, such as using a uniform capitation grant. On disbursements, the report suggested harmonizing the timing and require-ments for funding. In 2008 the uncoordinated and unpredictable pool of sector funding and central government resources impeded schools from having integrated school development and operational plans. Leakages could be minimized by improving the District Education Board Secretaries (DEBS) and district-level manage-ment capacity, while further deepening decen-tralization to the communities. The report also recommended basing DEBS grants on the total

amount of school grants in the district rather than on activities being implemented in the sector plan, which was not followed in practice. Overall, the 2006 PETS-QSDS findings reinforced those of the 2002 survey: there was a need to focus fur-ther on implementing reforms and improving inputs and resources at the school level.

Brief Comparison of PETS-QSDS in 2002, 2006, and 2014

While the direct comparison with previous PETS-QSDS in 2002 and 2006 is difficult, table  1.1  summarizes some changes in PETS-QSDS since 2002.

PETSThe rule of budget allocation to schools changed from a fixed-amount allocation to a formula-based allocation between 2002 and 2006. Further, the formula-based allocation was modified slightly in terms of school types—from basic school to primary school and from high school to second-ary school—between 2006 and 2014. The share of primary (or basic) schools receiving any grant and learning materials from the government declined from 98 percent in 2006 to 90 percent in 2002 and finally to 81 percent in 2014.1 The schedule of fund disbursement from the Ministry of Finance to MESVTEE was still unpredictable, impeding the planning of school operations. Rule-based allocation (that is, school grants in 2014) remained pro-poor in primary education. The regressive nature of teacher salaries mea-sured by the pupil-teacher ratio in 2002 and 2006 was not found in 2014.

QSDSSchool infrastructure has improved since 2006, except for accessibility to potable water. For exam-ple, primary schools with a library increased from 5.6 to 13 percent of all schools, those with electric-ity increased from 19 to 36 percent, and the number of pupils per classroom declined from 97 to 70. Teacher absenteeism has shown virtually no improvement since 2002, with absenteeism of 17  percent in 2002 and 18 percent in 2014. The

Introduction 13

pupil-teacher ratio improved, reaching 40 in 2014, down from 42 in urban and 61 in rural schools in 2002 and from 59 in 2006. However, access to potable water deteriorated from 92 percent in 2006 to 74 percent in 2014.

Other indicators of major education inputs found in 2014 QSDS are shown in table 1.2.

Policy QuestionsThis report evaluates three aspects of general education in Zambia: (a) educational perfor-mance, (b) public expenditure in relation to school-level financing and equity, and (c) various school inputs, including physical inputs, teachers, and soft skills of both students and teachers. It also attempts to show the link between educa-tional performance and educational inputs and to offer constructive recommendations for the government of Zambia to consider. The analysis of this report is confined to general education (primary education consisting of grades 1–7 and

secondary education consisting of grades 8–12). Early childhood education, TEVET, and higher education are beyond the scope of this report.

First, the report discusses the financial aspects of general education. The PETS-QSDS analyzes the overall flow of general education funds, whether public funds are reaching schools for the intended purpose, and whether public funds, especially school grants, are equitable. Despite the large amount of spending on primary and sec-ondary education, school grants are insufficient for the operation of schools, based on the revenue and expenditure analysis at the school level. Furthermore, there is evidence that public school grants are not being used for their intended pur-pose and that there is a delay in receiving school grants. However, more primary school grants are distributed to poorer schools, indicating the pro-poor nature of primary school grants. Meanwhile, secondary school grants are not equitable, and more resources are directed to richer schools.

Second, the report compiles the various indicators of education performance over the

TABLE 1 .2 Indicators of Major Education Inputs in 2014 QSDS Indicator Description

School absence rate for pupils (% of days absent during the previous week of the survey date)

Grade 5, 15%. Grade 9, 8%

School attendance rate for teachers (% of teachers present during unannounced visit and number of days present in June according to official attendance book)

82% of primary school teachers were present in school during unannounced school visit. Primary: 17 days present out of 21 school days (18.5 urban, 17.2 rural). Secondary: 18 days present out of 21 school days (17.7 urban and 18.2 rural)

Classroom absence rate (% of teachers absent in observation exercise)

During the survey visit (classroom observation), 6.5% of classes were not held. Reasons were (a) teachers were absent from school (20%); (b) teachers were in school but did not show up for class (31%); and (c) class was canceled without notice (48%)

Time spent teaching per day Grade 5 teachers (primary): 5.5 hours. Grade 9 teachers (secondary): 4.2 hours

Knowledge among teachers (scores on the same examination questions that were put to their students)

Grade 5 teachers: mathematics, English, and life skills, more than 90%. Grade 9 teachers: math, English, and science, approximately 70%

Infrastructure availability Primary school: potable water, 75%; electricity, 36%; boys per latrine, 83; girls per latrine, 75; pupils per classroom, 70. Secondary: potable water, 93%; electricity, 56%; boys per latrine, 73; girls per latrine, 67; pupils per classroom, 58

Teaching equipment availability Primary school: library, 13%; science lab, 3%. Secondary school: library, 28%; science lab, 22%

Share of pupils with textbooks 20% for primary and secondary. Primary school: mathematics, 20%; English and science, 18%. Secondary school: math and science, 20%; English, 34%

Pupil-teacher ratio 40 pupils per teacher. Grade 2 schools: urban, 45; rural, 51. Grade 5 schools: urban, 43; rural, rural 37. Grade 7 schools: urban, 39; rural, 31

Student learning (learning assessment score) Grade 5 scores: math, life skills, and Zambian language, 35%; English, 32%. Grade 9 scores: math, 29%; English and science, 36%

14 Introduction

past  decade. Student learning outcomes have stagnated since the first national learning assess-ment, conducted in 1999. Student scores on mathematics and English at grade 5 were 34.3 and 33.2, respectively, in 1999. The scores were almost the same in 2014: 35 in mathematics and 32 in English. Student learning assessment at grade 9 only started in 2013. Student scores on mathe-matics, English, and science at grade 9 were 29, 36, and 36, respectively, in 2014, barely changed from 29.0, 35.8, and 36.3, respectively in 2013.

Third, the report looks closely into education inputs that might be related to student outcomes. The government has put a lot of effort into strengthening the education sector over the past decade, including giving it the highest share of the budget (World Bank 2015). These efforts have started bearing fruit in the form of improvement in some education inputs. For instance, the aver-age pupil-teacher ratio at grades 1–9 is now about 40, compared to 59 in 2006. Student classroom ratio also improved from 97 to 70 (or 33 with two teaching shifts) from 2006 to 2014. Physical inputs have also improved since 2006. The percentage of primary schools with libraries increased from 5.6 to 13 percent, those with electricity increased from 19 to 36 percent, and the number of pupils per classroom declined from 97 to 70.

However, other education inputs and factors seem to be keeping student learning outcomes low. This low improvement in student learning outcomes, despite good progress in education inputs, gives rise to two questions: (a) Does the MESVTEE allocate its public funding efficiently and effectively to the areas most in need? and (b) What factors potentially keep student learn-ing outcomes low? While PETS-QSDS cannot identify direct or causal links between student learning outcomes and the quality of education inputs or financial resources, this report, based on analysis of the data from PETS-QSDS in 2014, aims to identify several areas that, if improved,

might help to improve student learning out-comes. The report discusses locations and physi-cal inputs, schooling and teaching hours, number of classrooms and teachers, textbooks, quality of teachers, and motivation and personality traits of teachers and students. In addition, the report addresses efficiency and equity by looking at the details of expenditure and revenue at the MESVTEE, provincial, DEBS, and schools levels.

The report is organized as follows. Chapter 2 discusses the survey methodology of the PETS-QSDS. Chapter 3 discusses the overall finance and management of general education. Chapter 4 describes the country’s recent education perfor-mance, presents major findings of education inputs from the QSDS, and analyzes the relation-ship between education performance and inputs. Chapter 5 summarizes the major findings and discusses policy implications.

Note 1. Direct comparison across the years is not recom-

mended due to changes in the allocation rule for school grants and representativeness of the data sets: in 2002, the data were only representative in four provinces, while in 2007 and 2014, the data were representative at the province level.

ReferencesKoyi, G., G. Masumbu, and A. Halwampa. 2012.

Understanding Youth Labor Demand Constraints in Zambia. Lusaka, Zambia.

World Bank. 2014. Global Economic Prospects. Washington, DC: World Bank.

———. 2015. “Education Public Expenditure Review in Zambia.” Washington, DC: World Bank.

Zambia Central Statistical Office. 2013. 2012 Preliminary Labour Force Survey Report. Lusaka, Zambia.

———. 2013. Zambia Labour Force Survey Report 2012. Lusaka, Zambia.

PETS-QSDS Study Methodology 15

Chapter 2

PETS-QSDS Study Methodology

The Public Expenditure Tracking Survey (PETS) and the Quantitative Service Delivery Survey (QSDS) report is based on various sources of data and government documents. The primary data for this report are from the 2014 PETS-QSDS—the data collected in the second term of  academic year (AY) 2014. These data cap-ture  information about the schools, students, and teachers in AY 2014 and the financial information of AY 2013 in order to paint a complete picture of financing for the entire school year. Data from the National Assessment Survey (NAS) of 2014, the Living Conditions Measurement Survey (LCMS) for 2010 and 2015, and the Zambia Demographics and Health Survey (ZDHS) for 2013–14 are used to complement the 2014 PETS-QSDS with regard performance indicators. This report also refers to preliminary findings of the 2015 LCMS, which were published by the Central Statistical Office in November 2015. If the information is not available in the preliminary findings for 2015, this report uses those from the 2010 LCMS for analysis (table 2.1).

Other major data sources for the report include government policy documents and statis-tics, including the Sixth National Development Plan, the National Implementation Framework, the Joint Annual Review, the Education Statistical Bulletin 2013 (ESB 2013), and the Yellow Book and Blue Book from the Ministry of Finance for the government budget and public expenditure, as well as interviews and meetings with govern-ment officials and cooperating partners. Data from the 2013 ESB, 2010 and 2015 LCMS, and 2013–14 ZDHS were used to cross-check perfor-mance indicators such as enrollment and com-pletion rates. Due to some inconsistency in the gross and net enrollment rates in ESB for 2013,1 this study reports enrollment rates using data from 2010 and 2015 LCMS and 2013–14 ZDHS.

2014 PETS-QSDS Survey ModulesThe 2014 PETS-QSDS is designed to capture the management of Provincial Education Offices (PEO), District Education Board Secretaries (DEBS), and schools, the performance of schools, teachers, and students, physical inputs for schools such as infrastructure and classroom size, and soft inputs for schools such as textbooks, teacher quality, student and teacher motivations, and classroom activities. Each survey module was developed based on previous PETS-QSDS (in this case, the 2002 and 2007 surveys), the World Bank service delivery indicator survey, and vari-ous documents from a literature review on soft skills and motivations (table 2.2).

2014 PETS-QSDS SamplePETS-QSDS is a representative sample of schools with grade 5 enrollment (primary schools) and schools with grade 9 enrollment (secondary schools).2 The data are representative at the prov-ince level for primary education. Originally, 500 schools were selected: 300 primary and 200 sec-ondary schools. The sampling method was multi-stage cluster sampling stratified by location (urban and rural) in each province. This sampling method is employed by the NAS in Zambia, and the PETS-QSDS uses the same sample as the 2014 NAS in order to be comparable to its grade 5 student assessment survey results. The sample frame for schools (cluster) is the list of schools with grade 5 enrollments for primary school and the list of schools with grade 9 enrollments for secondary schools in the Annual School Census 2013. PETS-QSDS defines “secondary” as a school with grade 9 enrollment—this means that some of these schools were previously basic schools offering

16 PETS-QSDS Study Methodology

grades 1–9. This sample includes government (public), grant-aided, community, and private schools. For the first stage, 500 schools were sam-pled using probability proportional to size. For the second stage, based on the number of streams in each grade, 10–20 students were randomly selected in each school (cluster) for primary schools and 10 students were randomly selected in each school (cluster) for secondary schools.

The field survey coincided with the national examinations being held for grades 7, 9, and 11. Certain schools were turned into examination centers and therefore could not be surveyed. This resulted in a final sample of 272 primary schools and 174 secondary schools,3 and the

TABLE 2 .1 Data SourceData source and year Description

PETS-QSDS 2014 See 2014 PETS-QSDS survey modules 2014

NAS 2014 National assessment of grade 5 students and teachers

LCMS 2010 and 2015; ZDHS 2013–14

Gross enrollment rate, net enrollment rate, and out-of-school children

ESB 2013 Enrollment and school numbers, repetition and dropout rates, teacher statistics

Yellow Book and Blue Book

Government financial statement C and budget

Interviews and meetings Government officials and cooperating partners

Note: PETS = Public Expenditure Tracking Survey; QSDS = Quantitative Service Delivery Survey; NAS = National Assessment Survey; LCMS = Living Conditions Measurement Survey; ZDHS = Zambia Demographics and Health Survey; ESB = Education Statistical Bulletin.

TABLE 2 .2 Survey Modules, Content, and Respondents Module Primary respondent or source Description

Teacher list Supervisor based on registry (teacher register) A listing of all teachers (grades 2, 5, 7, 9, and 11) and their information

Student and teacher selection module

Supervisor A listing of sample students and teachers and mapping of students and teachers

Teacher attendance I (first visit)

Head teacher based on registry (attendance book)

Basic information on teachers and teacher attendance for all teachers listed

General school

Part A Head teacher based on registry (student and teacher registers and attendance books)

General school information, location, facilities, and enrollment or repetition

Part B Head teacher or financial administrator based on registry (accounting books)

School financing, fund flow, expenditure, and decision making

Head teacher

Part A Head teacher Head teacher information

Part B Head teacher Head teacher personality and motivation

Classroom teacher

Part A Sample teacher (up to 3 teachers) Detailed teacher information and characteristics

Part B Sample teacher (up to 3 teachers) Teacher personality and motivation

Student

Part A Sample students (up to 20 students) Detailed information and characteristics of selected students

Part B Sample students (up to 20 students) Student personality and motivation

Household Parents of sample students Household demographics, education, and economics status

Classroom observation Observer Observation of grade 5 classroom of sample teachers

Teacher attendance II (second visit)

Observer Second unannounced school visit to check teacher attendance of 10 random sample teachers

PEO Supervisor PEO office information and PETS

DEBS Supervisor DEBS office information and PETS

Grade 9 assessment

Student assessment Sample students selected for interview module 7 Grade 9 student assessment

Teacher assessment Sample teachers selected for interview module 6 Grade 9 teacher assessment

Note: Survey mode was a combination of paper and tabulate. PEO = Provincial Education Offices; DEBS = District Education Board Secretaries; PETS = Public Expenditure Tracking Survey.

PETS-QSDS Study Methodology 17

province-level analysis at the secondary level was not possible due to the small sample size. The original sample weight was calculated using the equal probability of selection method (self-weighting) since the probability of select-ing schools given stratum (province and loca-tion) was equal for all schools. However, due to the constraints caused by the national examina-tions schedule and national holidays during the second term of AY 2014, some sampled schools were partially surveyed, which caused problems related to missing observations across modules. This missing observation problem was resolved by reweighting the sample schools using the inverse probability weight (IPW = w*1/p) pro-posed by Wooldridge (2007). The probability of being surveyed for each module was estimated using logistic regressions with the assumption of “missing at random” conditional on certain characteristics of schools (urban and rural, size of school, and province).

Another challenge that the 2014 PETS-QSDS faced was the difficulty of collecting reliable financial information from the PEO and the DEBS due to a lack of systematic financial record keeping that resulted in a large amount of miss-ing information and inconsistency between school-level financial records and PEO and DEBS financial records. For this reason, the report uses only two sources to track financial flows: (a) the Ministry of Education, Science, Vocational Training, and Early Education (MESVTEE) financial statement C, which records financial transactions from headquarters

to the PEO and DEBS, and (b) school-level finan-cial records from the 2014 PETS-QSDS.

The final sample includes 272 schools with 3,715 students and 295 teachers at the primary level and 174 schools with 1,833 students and 347 teachers at the secondary level. At the pri-mary level, 78 percent are government schools, 4 percent are private or church schools, 3 percent are grant-aided schools, and 15 percent are community schools. At the secondary level, 75  percent are government schools, 18 percent are private or church schools, 4 percent are grant-aided schools, and 1 percent are community schools at the secondary level (table 2.3). By loca-tion, 85 percent of primary schools were rural, compared with 67 percent of secondary schools. The average number of grade 1–7 students per school was 412, and the average number of grade 8–12 students per school was 201.

Notes 1. Estimates of gross and net enrollment rates require

reliable estimates of the school-age population; the EBS noted that the population estimates could be flawed, and enrollment size was overestimated.

2. Since the Zambian education system has been changing and schools are offering different grades than before (see chapter 4), this report defines pri-mary school as a school with grade 5 enrollment and secondary school as a school with grade 9 enrollment.

3. This sample size is calculated based on the num-ber of schools that have complete information in general school parts A and B.

TABLE 2 .3 Planned and Final Sample Size

ModulePrimary (grade 5) Secondary (grade 9)

Planned Final Planned Final

Student and teacher selection 300 292 200 186

General school 300 272 200 174

Head teacher 300 258 200 161

Teacher 450–500 295 600 347

Student 4,500–5,000 3,715 2,000 1,833

Household 4,500–5,000 2,790 2,000 1,400

Teacher attendance (school random visit) 200 185 0 0

Classroom observation 300–400 268 0 0

18 PETS-QSDS Study Methodology

ReferencesWooldridge, Jeffrey M. 2007. “Inverse probability

weighted estimation for general missing data problems”, Journal of Econometrics.

Zambia Central Statistical Office. 2013. 2012 Preliminary Labour Force Survey Report. Lusaka, Zambia.

———. 2013. Zambia Labour Force Survey Report 2012. Lusaka, Zambia.

MESVTEE (Ministry of Education, Science, Vocational Training and Early Education). 2013.  Educational Statistical Bulletin. Lusaka, Zambia.

Education Sector Finance and Management: PETS 19

Chapter 3

Education Sector Finance and Management: PETS

This chapter discusses financial aspects of general education (primary and secondary level), includ-ing budgeting process, expenditure flow from the Ministry of Education, Science, Vocational Training, and Early Education (MESVTEE) to schools, and school-level financing.

Education Budget Process and Expenditure Flow (General Education)Zambia’s budgeting process follows both top-down and bottom-up approaches. The MESVTEE decides the priorities for national education pol-icy every five years and allocates the budget for the program and activities through indicative planning figures. In the meantime, all public edu-cation institutions in the country are involved in the budgeting process, planning and budgeting their own operation and programs (figure 3.1).

The budgeting process of the MESVTEE is guided partly by the Call Circular as well as by the budgeting committees’ recommendations:

• The Call Circular provides guidelines for the preparation of estimates of revenue and expen-diture and the three-year medium-term expen-diture framework (MTEF). The guidelines are intended to ensure that the ministry prepares the budget in line with Cabinet decisions as out-lined in the Green Paper on the three-year MTEF and the budget. The MESVTEE prepares a budget including the MTEF budget, which outlines priority programs and activities for the coming year, and the annual work plan and budget (AWPB), which is used as a benchmark for implementing programs and activities.

• The process of developing the AWPB uses the top-down and bottom-up planning approach

with three-year rolling plans. The ministry prepares the guidelines and sets indicative planning ceilings for each institution.

• The District Education Board Secretaries (DEBS) disseminate the budget preparation guidelines and priorities to all basic and com-munity schools; colleges and high schools dis-seminate the same to all of their departments to commence planning and budgeting.

• Using the prescribed formats and tools, all insti-tutions develop their AWPBs. The budget com-mittee is expected to review the AWPB from all institutions and departments prior to consoli-dation of the institutional and district AWPB.

• The Provincial Education Offices (PEO) review the AWPBs submitted by all institutions and consolidate the provincial budget after col-leges, the DEBs, and secondary schools have finished developing their budgets.

• The PEO and department heads defend their budget submissions, and the MESVTEE Budget Committee, chaired by the minister, gives final approval.

• The MESVTEE submits the MTEF and AWPB to the Ministry of Finance and National Planning. At this stage, the MESVTEE finalizes the Budget Framework Paper and the Activity Based Budgeting for submission to the Budget Office of the Ministry of Finance and National Planning.

The government embarked on a pilot out-put-based budget (OBB) system for the education sector in 2015. To enhance the potential efficiency of public financing, the government identified the MESVTEE as one of the pilot ministries to be migrated to the OBB system in 2015. Under this system, the educational outputs or targets are identified, and expenditures are aligned to each of

20 Education Sector Finance and Management: PETS

the targets. Although it was difficult to identify the education level to which some of the activities belong, especially activities that are included under the heading “expenditure” of the MESVTEE headquarters, this transition to OBB enables planners to allocate resources to different subsec-tors of education and to see the links between the budget and the outcomes. This PETS-QSDS, however, focuses on the 2013 budget and expen-diture since the OBB was initiated only in 2015.

The MESVTEE sets the budgeting process and financial guidelines for most schools, and about half of the district offices interact with pro-vincial offices to discuss budgeting. A little less than a half of DEBS officers said that they had met with the PEO to discuss district budgets. Among schools, 89  percent of primary schools

and 94  percent of secondary schools reported having a budgeting process and 92  percent reported following financial guidelines set by the MESVTEE headquarters (2014 PETS-QSDS).

Budget preparation is separated from budget approval in schools: 49  percent of schools reported that the operating budget was prepared primarily by the financial committee, while 59  percent of schools reported that the final budget was approved by the head teacher. The school management committee and parent-teacher association (PTA) are also involved in the school budgeting process: 16  percent of schools involve the PTA in budgeting and 25 percent of schools have a school management committee, which is responsible for preparing the budget (figure 3.2).

FIGuRE 3 .1 Budget Process in 2014

Notes: MESVTEE = Ministry of Education, Science, Vocational Training, and Early Education; MTEF = medium-term expenditure framework; AWPB = annual work plan and budget; PTA = parent-teacher association; DEBS = District Education Board Secretaries; PEO = Provincial Education Offices; ABB: Activity Based Budgeting.

Ministry of Finance

MESVTEE

1) Call Circular(budget guideline)

2) Budget preparation including setting indicative planning figures (ceiling

set for institutions based on MTEF)

DEBS3) Disseminate budget planning guidelines

Schools4) Prepare and submit AWPB

Departments4) Prepare and submit AWPB

PEO

5) Review AWPBs submitted by all institutionsand consolidate provincial ABB

MTEF• Set national priorities• Determine the use of funds• Justify programs and activities• Prepare the budget, including setting etc.

AWPB• Operational budget and cash flow• Top-down and bottom-up approach• Budget committee consists of all

institutions, province or district develop-ment coordinating committees, and PTAs

7) MTEF and AWPB submitted for approval

6) PEOs and departmentheads submit to the committee

Colleges3) Disseminate budget planning guidelines

Education Sector Finance and Management: PETS 21

The PTA has an active role in school manage-ment and finance. More than 90  percent of schools had held a PTA meeting at least once and  40  percent of schools had held more than three PTA meetings. More than half (59 percent) of PTA meetings were about school infra-structure. PTA members have significant deci-sion making powers on areas such as decision on PTA/community fees, financial support to a child, and providing housing for teachers or per-manent classroom (figure 3.3).

Flow of General Education ExpenditureExpenditure on general education consists of three items: staff and teacher salaries (personal emolu-ments), school grants for education materials and for free primary and secondary education, and infrastructure development, mainly for school construction. These three items capture govern-ment policy priorities regarding teachers, free pri-mary and secondary education, and the supply of

FIGuRE 3 .2 Primary Responsibility for Budgeting at the School Level

Source: 2014 PETS-QSDS.

70

60

50

40

30

% o

f sc

ho

ols

20

10

0School management

committee

25

12

48

8

Financialcommittee

14

59

Head teacher

16

3

Head teacherand teacher

16 16

Parent-teacher association

Preparation of budget Approval of budget

FIGuRE 3 .3 Decision Making on Financial Management

Source: 2014 PETS-QSDS.Note: PETS-QSDS asked the head teacher who is the main decision-making party for the areas mentioned. PTA = parent-teacher association.

1

57

20

51

22

29

36

23 23 25

PTA/communityfees

Exempt child from PTA fee

% o

f d

ecis

ion

mak

ing

po

wer

Financial supportto a child

Housing for teachers Providing permanentclassroom

60

50

40

30

20

10

0

CommunityMESVTEE DEO Head teacher Teachers PTA executive Parents

22 Education Sector Finance and Management: PETS

schools. Personal emoluments and secondary school grants flow from the Ministry of Finance to individual accounts and secondary school accounts, and infrastructure funds for construc-tion flow from the MESTVEE to individual schools. In primary, the school grants and textbook funds are disbursed through decentralized units—the DEBS. In secondary, textbooks are supposed to be delivered directly to schools from publishers, and funds flow to the publishers (figure 3.4).

In 2013, 89 percent1 of expenditure in general education went to salaries of school teachers and staff; the second major expenditure item, con-struction and upgrading of secondary schools, accounted for 7.5 percent of the budget. In 2013, out of total spending on primary and secondary education (ZMW 4,034 million, 77  percent of total education expenditure), 89 percent went to salaries of school staff and teachers, 8.5  percent went to infrastructure development, especially to construction and upgrading of secondary schools (7.5  percent), and only 2.5  percent of spending was for school grants (figure 3.5). Only 0.5 percent of general education expenditure was spent on education materials for secondary schools, 0.5  percent was spent on grants for DEBS, and 1.5 percent was spent on school grants for primary schools. Figure 3.5 does not include the budget for

textbooks. In 2013, the actual expenditure for textbooks was about ZMW 8 million (0.2 percent of the general education expenditure), which was significantly lower than the amount budgeted (about ZMK 42 million) because of the serious delay in textbook procurement and delivery caused by new curriculum development.

Teacher salaries in Zambia are relatively high compared to other countries in Sub-Saharan Africa. Civil service reform in 2013 raised the sal-aries of all civil servants by 45  percent (World Bank 2015). Today, 46,403 primary school teach-ers receive on average ZMW 61,404 per year, which is approximately US$9,520, and 24,091 secondary school teachers receive ZMW 65,088 per year. In comparison to GDP per capita, the average teacher salaries for primary and second-ary teachers are 6.7 times or 7.1 times higher. An international comparison of primary teacher sal-ary from 35 Sub-Saharan African countries (from different years) shows that the ratio of 6.7 times higher belongs to a relatively higher end of the distribution (figure 3.6).2 In addition, because of an increase in salary in 2013, the ratio to GDP per capita increased from 3.67 in 2011 to 6.7 in 2014.

School Finance and ManagementTotal Revenue and Expenditure at the School Level

More than 50  percent of primary schools and approximately 93  percent of secondary schools charge school fees (tuition, PTA) to finance their operations. Despite the free education policies in  primary and secondary education, schools continue to charge students tuition, fees, or both. Sources of revenue differ for primary and sec-ondary schools. While 55  percent of primary schools charge school fees, only 34 percent of stu-dents actually pay them (table 3.1). For second-ary schools, school fees are a key source of revenue, with 63 percent of students paying them, and the share of public funding is only 10 percent. Primary schools charge ZMW 37 annually, while

FIGuRE 3 .4 Flow of Public Funds in General Education (Grades 1–12)

Note: Textbook procurement is decentralized; however, due to the new curriculum, the textbooks were centrally procured in 2013 and delivered to schools in 2014. DEBS = District Education Board Secretaries; PEO = Provincial Education Offices.a. Infrastructure for secondary school building.

Teachers and nonteaching staff

Secondary schoolsBasic or primary schools

DEBS PEO

Grants, textbooks, and infrastructure

Grants, textbooks,and infrastructure

Infrastructurea

and textbooksSecondary

school grants

Personal emoluments

Ministry of Finance

MESVTEE (headquarters)

Education Sector Finance and Management: PETS 23

FIGuRE 3 .5 Government Expenditure Flow in General Education, 2013

Source: MESVTEE Financial Statement C, 2013.Note: This includes expenditure related to general education (primary and secondary education) only. For all expenditures, see Public Expenditure Review 2015.

Infrastructure(8.5%)

General education expenditure(ZMW 4.034 billion)

Personal emoluments(89%)

Basic school(73%)

High school(16%)

Secondaryschool(7.5%)

Provincial level:secondary school

(0.5%)

District level:DEBS grants

(0.5%)

District level:Primary school

(1.5%)

Primaryschool(1.5%)

Grants(2.5%)

FIGuRE 3 .6 Primary Teacher Salary as a Ratio to GDP per Capita Compared across African Countries

Source: UNESCO Institute for Statistics 2011. The information for Zambia is from the 2014 budget.Note: Teachers’ salaries in Zambia consist of a basic salary, housing allowance, and transport allowance for all teachers. Other allowances, such as a remote allowance for teachers deployed in remote areas, a responsibility allowance for teachers with a diploma teaching in secondary schools, and a double-class allowance for primary school teachers teaching two shifts, are not included.

Pri

mar

y te

ach

er s

alar

y as

ara

tio

to

GD

P p

er c

apit

a

0

2

4

6

8

0.9

Con

go, R

ep. (

2007

)A

ngol

a (2

003)

Gui

nea

(200

5)

Sey

chel

les

(200

3)S

udan

(200

3)

São

Tom

é an

d P

rínci

pe (2

006)

Cap

e Ve

rde

(200

9)R

wan

da (2

008)

Mad

agas

car (

2006

)Li

beria

(200

8)

Cam

eroo

n (2

007)

Cen

tral A

frica

n R

epub

lic (2

007)

Erit

ria (2

003)

Bur

undi

(200

7)

Zam

bia

Nig

er (2

008)

Mal

awi (

2008

)

Zim

babw

e (2

003)

Togo

(200

7)

Cha

d (2

003)

Keny

a (2

004)

Bur

kina

Fas

o (2

006)

Leso

tho

(200

4)

Nig

eria

(200

3)

Côt

e d’

lvoi

re (2

007)

Uga

nda

(200

7)

Sen

egal

(200

4)

Gha

na (2

007)

Gui

nea-

Bis

sau

(200

6)

Sie

rra

Leon

e (2

004)

Mal

i (20

08)

Moz

ambi

que

(200

3)

Con

go, D

em. R

ep. (

2005

)

Gam

bia,

The

(200

3)

Ben

in (2

006)

1.5 1.7 1.72.2 2.3 2.5 2.6

2.9 3.03.2 3.3

3.73.63.9 4.0 4.2 4.2 4.4

4.7 4.7 4.7 4.9 4.9 5.05.3 5.3 5.4

6.1 6.16.3

6.6 6.7

7.6 7.7

24 Education Sector Finance and Management: PETS

secondary schools charge ZMW 384 annually. A large percentage of primary schools raise revenue through the PTA fee (58 percent of school fees), while secondary schools raise revenue through tuition (46 percent of school fees).

Schools with poor students have lower reve-nues per pupil (combining both private and pub-lic sources) than schools with rich students, especially at the secondary level. Poor primary schools have a revenue amount of ZMW 35 per pupil, and their share of public funding is

67  percent, while rich primary schools have a revenue amount of ZMW 46, and their share of public funding is 46 percent (figure 3.7). At the secondary level, schools with richer students have revenue of ZMW 390 per pupil, while schools with poorer students have revenue of only ZMW 144 per student.

Revenue per pupil differs largely by province. The gap between the lowest and the highest per  pupil revenue is ZMW 73 at the primary level. Western Province has per pupil revenue of ZMW 7, while Eastern Province has per pupil revenue of ZMW 80 (figure 3.8). There are several  potential interpretations of the vast difference in per pupil revenue. First, provinces like Eastern Province may be charging higher fees, given the fact that Eastern receives the least amount of public school grants (to be dis-cussed in a following section). Second, the enroll-ment factor in the public funding formula (that is, primary school grants) may not be appropri-ately taken into account. Third, there may be a  smaller number of students in provinces like Eastern (but not Lusaka). The fourth possi-bility could be a combination of these reasons.

TABLE 3 .1 School Revenue (Private and Public Sources) in Primary and Secondary Education, 2014

Indicator Primary Secondary

% of schools charging fees 55 96

% of students paying fees 34 63

Total school fee amount (ZMW) 37 384

% of parent-teacher association fees (out of total school fees)

58 24

% of tuition fees (out of total school fees) 14 46

Amount of revenue per pupil (kwacha) 35 250

% of public fund in revenue 64 10

Source: 2014 PETS-QSDS.Note: The sample includes all primary and secondary schools.

FIGuRE 3 .7 School Revenue per Pupil and Share of Public Funding in Primary and Secondary Education, by Income Level of Students

Source: 2014 PETS-QSDS.Note: A poor primary school is defined as the bottom 33 percent of schools in average income of students’ families, and a rich primary school is defined as the top 33 percent of schools in average income of students’ families.

Rev

enu

e p

er p

up

il (Z

MW

)

% o

f p

ub

lic f

un

din

g in

rev

enu

e

35

21%

67% 390

46%

4611%

58%

144

30 3%

149

Poor Middle Rich

450

400

350

300

250

200

150

100

50

0

80

70

60

50

40

30

30

20

10

0

PrimaryPrimary Secondary Secondary

Education Sector Finance and Management: PETS 25

Further study of the huge difference in per pupil revenue is probably needed. If the large differ-ence undermines equity, some mitigation mea-sures should be prepared.

The expenditure pattern is similar in primary and secondary schools. A large percentage of expenditure goes to co-curriculum activities and construction (figure 3.9). The bulk of expenditure

on textbooks (development and procurement) happens at the MESVTEE headquarters and the DEBS. Textbook expenditure at the school level is likely to be sourced through a private fund such as PTA fees.

Public grants to schools include school grants, learning materials, and textbooks in 2013, yet 19  percent of grant-eligible primary

FIGuRE 3 .8 School Revenue per Pupil, by Province

Source: 2014 PETS-QSDS.

0

10

20

30

40

50

60

70

80

90

Sch

oo

l rev

enu

e p

er p

up

il (Z

MW

)

Wes

tern

Luap

ula

North

weste

rn

Coppe

rbelt

North

ern

South

ern

Muc

hinga

Centra

l

Lusa

ka

Easte

rn

714

1722

2732

38

55

7580

FIGuRE 3 .9 School Expenditure for Primary and Secondary Education

Source: 2014 PETS-QSDS.Note: Includes both private and public funds. Textbook expenditure does not include textbook grants spent at the DEBS and MESVTEE levels.

30

25

20

15

10

5

0

% o

f sc

ho

ol e

xpen

dit

ure

Co-curriculumactivities

Construction Learningmaterials

Nongovern-mental

teacher salary

Maintenance Otherexpendituresfor teachers

Textbooks

24 2422

2021

12

14

68

11

6

01

19

Primary Secondary

26 Education Sector Finance and Management: PETS

schools (government or grant-aided schools) reported not having received any grant from the MESVTEE during academic year (AY) 2013. Further, a large provincial gap exists with regard to grants. Lusaka Province has the highest share of schools receiving grants or other materials from the MESTVTEE (97  percent), while Eastern Province has the lowest share of schools receiving any form of grant (69  percent) (figure 3.10).

28 percent of primary schools and 70 percent of secondary schools reported having not received a school grant. . There is a big differ-ence in grant receipt between urban and rural schools, with a different pattern for primary and secondary schools. Rural primary schools received a significantly larger school grant than urban primary schools (ZMW 16 and ZMW 7, respectively), and a larger  percentage of rural primary schools received a school grant than urban schools (73 and 65 percent, respectively) (table 3.2). The situation is the opposite for sec-ondary schools. Urban secondary schools received ZMW 18 per pupil, while rural sec-ondary schools received ZMW 7 per pupil.

The amount of per pupil free primary school grant that the MESVTEE disbursed to primary schools is ZMW 22, while the per pupil school

grant that grant-eligible primary schools reported receiving is ZMW 15. The amount that the MESVTEE disbursed to each province differs across provinces: Eastern Province received the largest per pupil free primary school grant (ZMW 43) and Western and Copperbelt provinces received the lowest per pupil free primary school grant (ZMW 15) (table 3.3). The  difference

FIGuRE 3 .10 Percentage of Schools That Received a Grant or Textbooks, by Province

Source: 2014 PETS-QSDS.Note: Sample includes grant-eligible schools only. The survey asked head teachers and school accountants whether their schools received any grant or textbooks from the Ministry of Education, Science, Vocational Training, and Early Education.

100

80

60

40

20

0

% o

f sc

ho

ols

Wes

tern

Luap

ula

North

weste

rn

Coppe

rbelt

North

ern

South

ern

Muc

hinga

Centra

l

Lusa

ka

All pro

vince

s

Easte

rn

69 70 7072 75

94 95 95 97

8186

TABLE 3 .2 School Grants Received by Public or Grant-Aided Schools (Annual), by urban-Rural Location and Education Level

Primary Secondary

Location of school

Per pupil school grant (ZMW)

% of schools that received

a grant

Per pupil school grant (ZMW)

% schools that received

a grant

Urban 7 65 18 48

Rural 16 73 7 26

Total 15 72 10 30

Source: 2014 PETS-QSDS.Note: Sample schools are both basic and secondary or high schools offering a grade 9 education. This captures basic schools not registered as secondary schools. Therefore, many schools eligible for secondary school grants probably do not receive them. The % of schools that received a grant is based on the grant amount they received from the Ministry of Education, Science, Vocational and Entrepreneurship Training. If the school reported a positive amount, the school is considered to have received a school grant. The survey asked if the grant was for primary school or secondary school, and the grant for secondary school is separated from the grant for primary school and vice versa. The bottom 5% and top 5% of schools in grant amount distribution are considered as outliers and not included in the calculation. The sample includes grant-eligible schools only.

Education Sector Finance and Management: PETS 27

between the amount disbursed by the MESVTEE and the amount reportedly received by schools is substantial for Eastern and Luapula provinces: ZMW 24 and ZMW 20, respectively. This gap does not necessarily indicate leakage in school grants. A follow-up survey with government offi-cials at the PEO found that the DEBS also use school grants to pay for delivering textbooks3 to remote schools and other operational costs (espe-cially, transportation) to support schools. Since these transportation and delivery expenses are not budgeted, the DEBS deduct them from the school grant. This needs government follow-up with proper investigation.

The government’s budget allocation rule for primary school grants does not align with the

actual disbursement of school grants. According to the government’s grant allocation rule, three factors determine the amount of free primary grants to schools: (a) school location (remote-ness), (b) gender ratio, and (c) size of school (enrollment). According to the 2014 PETS-QSDS, about 18 percent of the actual amount of grants received by primary schools is explained by these factors, to varying degrees in each prov-ince. Actual grant amounts disbursed in Lusaka Province are explained by the grant allocation rule the most, and those disbursed in Western and Eastern provinces are explained the least (figure 3.11).

Primary school grants are pro-poor, while secondary school grants are pro-rich. In prin-ciple, all primary schools are supposed to receive grants under the free primary education policy. On average, at the primary level, schools with relatively poor students receive ZMW 16 per pupil, while schools with relatively rich students receive ZMW 7 per pupil (figure 3.12). Moreover, schools with poor students (75 percent) tend to receive free primary school grants more than schools with relatively rich students (60  percent). At the secondary level, schools with relatively poor students receive ZMW 15 per pupil, but schools with richer students receive ZMW19 per pupil. Further, secondary schools with relatively rich students have higher chances of receiving school grants than those with relatively poor students (37 and 46 percent, respectively).

A majority of schools (67 percent) reported delays in receiving the school grant; however, the delay in teacher salaries is negligible for both primary and secondary schools. The tim-ing of receipt of the school grant is unpredict-able due to the delay in disbursement from the Ministry of Finance to the MESVTEE. The reported delay captures the delay in disburse-ment from the Ministry of Finance down to the school level as well as the failure to receive the grant amount. The delay is positively correlated with distance from the school to the DEBS—63  percent of schools within a 5-kilometer radius of the DEBS and 69  percent of schools

TABLE 3 .3 Annual Disbursement and Receipt of Primary School Grants, by ProvinceAmount (ZMW)

Province Amount disbursed by MESVTEEa

Amount received by schoolb

Eastern 43 19

Luapula 29 9

Central 22 9

Western 15 5

Muchinga 26 19

Northwestern 20 14

Southern 19 12

Lusaka 19 14

Northern 27 25

Copperbeltc 15 24

National 22 15

Source: 2014 PETS-QSDS.Note: The bottom 5% and top 5% of schools in grant amount distribution are considered as outliers and not included in the calculation. The sample includes only grant-eligible schools. Until 2013, the free primary school grant, school requisites (notebooks or other learning materials for students), and funds for orphans and vulnerable children were considered separate expenditure items; however, since 2014, free primary school grants have included school requisites and funds for orphans and vulnerable children. MESVTEE = Ministry of Education, Science, Vocational Training, and Early Education; AY = academic year; DEBS = District Education Board Secretaries.a. Based on the total amount of free primary school grant disbursed in AY 2013 from financial statement C divided by enrollment size (grade 1–7) from ESB 2013.b. Based on the school-reported amount of primary school grants received from MESVTEE divided by the primary school enrollment number for AY 2013, using 2014 PETS-QSDS.c. In Copperbelt Province, the average amount of school grant per pupil that sampled schools reported is greater than the per pupil amount that is disbursed to the DEBS for the free primary school grant. This may happen because schools report the total amount of grant received from the MESVTEE, not just the free primary school grant.

28 Education Sector Finance and Management: PETS

located farther than 25 kilometers from the DEBS reported a delay in receiving the grant. In the meantime, the reported outstanding salary was only 0.2 month worth of salary, on average, which is cumulated for the entire period served,

for primary school teachers and 0.3 month worth of salary for secondary school teachers. The grant amount that schools should receive is available only at the DEBS and not at the school; for this reason, remote schools may not know

FIGuRE 3 .11 Percentage of Grant Amount Explained by the Budget Allocation Rule

Source: 2014 PETS-QSDS.Note: Amount of school grants received from DEBS (District Education Board Secretaries) are regressed on gender parity, enrollment, and school location (rural and distance to DEBS) in each province. R-squared of regression is shown in the figure, which indicates how much of the school grant amount can be explained by the above mentioned factors.

100

80

60

40

20

0

% o

f g

ran

t am

ou

nt

exp

lain

ed b

ybu

dg

et a

lloca

tio

n r

ule

Wes

tern

Luap

ula

North

weste

rn

Coppe

rbelt

North

ern

South

ern

Muc

hinga

Centra

l

Lusa

ka

Easte

rn

7 10

2731 31

42

51 5359

31

FIGuRE 3 .12 School Grants per Pupil and Share of Schools Receiving a Grant, by Family Income of Students

Source: 2014 PETS-QSDS.Note: A poor primary school is defined as the bottom 33.3% of schools in average income of students’ families, and a rich primary school is defined as the top 33.3% of schools in average income of students’ families. The bottom 5% and top 5% of schools in grant amount distribution are considered as outliers and not included in the calculation. The sample only includes schools eligible for a grant.

Sch

oo

l gra

nt

per

pu

pil

(ZM

W)

% o

f sc

ho

ols

th

at r

ecei

ved

a g

ran

t16

15

37%

75%69%

60%

19

46%

7

16

15%

3

Poor Middle Rich

20

18

16

14

12

10

8

6

4

2

0

80

70

60

50

40

30

30

20

10

0

PrimaryPrimary Secondary Secondary

Education Sector Finance and Management: PETS 29

the exact grant amount or time of delivery. Currently, the government is working on the direct deposit of school grants to the school accounts from the Ministry of Finance, which may reduce the unintended use of school grants at the DEBS level as well as shorten the delay in disbursement. In terms of teacher salaries, the Ministry of Finance deposits the amount directly into government teachers’ accounts.

In addition to delays or failed delivery of school grants, many schools experience short-ages of textbooks, desks and chairs, staff, and building maintenance. Of these issues, textbook shortages are the most serious. Although many schools make requests to the government through the DEBS to resolve the shortage issues, most of the problems remain unresolved (figure 3.13).

In addition, ambiguity in the responsibilities for textbook delivery and in the delivery funds is created by a mismatch between the government’s textbook procurement policy and implementa-tion. The current textbook procurement policy (Decentralized Textbook and Other Educational

Materials Procurement Manual 2012) states that the delivery of textbooks is the responsibility of publishers. Primary schools submit textbook orders to the DEBS, which convey the orders to publishers on behalf of the schools, and then publishers deliver the textbooks to the schools. Secondary schools purchase textbooks directly from publishers, and publishers are responsible for delivering textbooks to the schools. However, because the textbooks are procured centrally, contrary to the government’s official textbook procurement policy, responsibility for textbook delivery and budgetary support to local units (such as the DEBS and the PEOs) for delivering textbooks are not spelled out. For this reason, textbook delivery funds are lacking, and the DEBS deduct money from school grants to deliver textbooks to schools.

Notes 1. Analysis of this report only includes expenditure

at the general education (grades 1–12) level and excludes expenditure in technical education and

Textbookshortage

0

40

50

60

80

10091

76

61

50

3829

7471

50

137 10

76

62 61

% o

f sc

ho

ols

Desk and chairshortage

Staff shortage Late payment Buildingmaintenance

Shortage problem Request made Remaining problem or issue

FIGuRE 3 .13 Percentage of Schools Reporting Shortages and Making Requests

Source: 2014 PETS-QSDS.

30 Education Sector Finance and Management: PETS

vocational entrepreneurship training (TEVET) and higher education. The 2015 Education Public Expenditure Review in Zambia reports financial aspects for the entire education sector (2015, The Word Bank).

2. GDP per capita for 2014 is estimated as ZMW 9,135 by using linear projection.

3. Textbook delivery is the responsibility of pub-lishers, according to the decentralized textbook procurement manual (Ministry of Education, Science, Vocational Training and Early Education, 2012); however, textbooks are still centrally procured, and the responsibility for textbook delivery and the provision of funds for delivery are not clearly stated in the policy docu-ment and budget.

ReferencesDecentralized Textbook and Other Education

Materials Procurement Manual. 2012. Ministry of Education, Science, Vocational Training and Early Education. Zambia.

UNESCO (United Nations Educational, Scientific and Cultural Organization). 2010. IIEP Newsletter. Paris.

World Bank. 2015. Zambia Economic Brief. Lusaka. Zambia.

———. 2015. “Education Public Expenditure Review in Zambia.” Washington, DC: World Bank.

MESVTEE (Ministry of Education, Science, Vocational Training and Early Education). 2013. Educational Statistical Bulletin. Lusaka, Zambia.

Education Sector Performance and Service Delivery: QSDS 31

Chapter 4

Education Sector Performance and Service Delivery: QSDS

This chapter discusses primary and secondary education performance in Zambia, including enrollment, completion, and internal efficiency across gender in the provinces, student learn-ing outcomes, education inputs, and service indicators such as location, infrastructure, schooling hours, pupils per classroom, text-book availability, and quality of teachers and teaching practice.

Education PerformanceAccess

The Zambian education system experienced two  major restructurings in 1996 and 2011 (table  4.1). Education reform in 1996 trans-formed the education system from primary (grades 1–7) and secondary (grades 8–12) schools into basic (grades 1–9) and high (grades 10–12) schools in order to achieve the goal of universal education up to grade 9. However, given low conversion rates from primary and secondary to basic and high school and difficul-ties related to teacher management and teacher deployment for grades 8 and 9 in basic schools, the government reformed the system again, returning to a system of primary and secondary schools. However, the current policy (Education Our Future) and the legal system (the Education Act of 2011) are not aligned, as the legal system does not recognize primary and secondary schools and keeps the previous system of basic and high schools.

The restructuring of Zambia’s education sys-tem is still in transition, and more than half of all  schools remain in the previous structure. In 2013 the Zambian education system formally adopted  primary (grades 1–7) and secondary

(grades 8–12) levels; however, full implementa-tion of this reform will take time. Indeed, only 49 percent of schools have the current structure of primary and secondary levels (figure 4.1 and table 4.2). A large supply of community schools offer only lower-primary grades, and the supply of different grades in different schools compli-cates the situation. In 2013, at the primary level, 45 percent of schools offered grades 1–7, 8 percent of schools offered only grades 1–4, and 35 percent of schools offered both primary and lower- secondary grades (grades 1–9). At the secondary level, 4  percent of schools offered grades 8–12, and 1 percent of schools offered grades 10–12.

General education (grades 1–12) in the past decade experienced an expansion in the supply of schools and an increase in enrollment, but this expansion reflects mainly the growth of the school-age population rather than an increase in the enrollment rate. Total enrollment and the number of schools increased 8 percent between 2008 and 2013; at the same time, there were around 600 more basic schools in 2013 than in 2008, an increase from 8,195 to 8,801. Basic schools enrolled 236,000 more students in grades 1–9 in 2013 than in 2008 (figure 4.2).

In the early 2000s, it was claimed that the increase in net enrollment rates, from 68 to 75 percent between 1998 and 2003, was partially due to an increase in access through building and upgrading basic schools as well as implementa-tion of the free primary education policy (free tuition) for grades 1–7. However, in 2010, the net enrollment rate for primary education decreased to 73 percent (2010 LCMS), and, in 2015, accord-ing to preliminary findings, there was little improvement in the net enrollment rate for pri-mary education (78 percent) (2015 LCMS). Similar to basic education, there has been

32 Education Sector Performance and Service Delivery: QSDS

steady  expansion in secondary or high school education. There were 84 more secondary schools (grades 8–12) in 2013 than in 2008, with both government and private secondary schools expanding (figure 4.2). With increasing school supply, there were 119,000 more secondary stu-dents in 2013 than in 2008. For both primary and secondary schools, the government is the major service provider (62 and 76 percent, respec-tively); the second largest service provider is the community for primary and the private sector for secondary.

Student enrollment per school decreases with higher grades, with substantial increases in grades 10 and 12 (figure 4.3). This may indicate a severe shortage of schools for upper-secondary grades. On average, total enrollment in 2013 in grades 1–12 was 3,818,336, and the total number of schools was 9,484, which results in 402 students per school. Enrollment per school is U-shaped, meaning that it decreases as grades go up to the lower-secondary level (grades 8–9) and increases dramatically after that. Enrollment per school is 231 students for grades 1–4, 146 for grades 5–7, 120 for grades 8–9, and 570 for grades 10–12. There are also significant differences in enrollment per school across provinces. Copperbelt, for instance, has a large number of students per school in upper-secondary grades (789 students per school).

Primary schools in Zambia can accommodate the enrollment of most primary school-age chil-dren (ages 7–13), but the capacity of secondary education has room for improvement. While the primary gross enrollment rate is almost 100 percent (99 percent in 2010 and 105 percent in 2015, according to preliminary findings and sub-ject to error, as discussed in box 4.1), the primary net enrollment rate was only 73 percent in 2010 (78 percent in 2015, preliminary findings) (table 4.3). This implies either low internal effi-ciency due to high repetition and dropouts or late entrance of students enrolling in primary school. For secondary grades, the net enrollment rate was 40 percent for grades 8–12 in 2010 (43  percent in 2015, preliminary findings). However, the net enrollment rate for grades 8–9 and for grades 10–12 is a lot lower than the

TABLE 4 .1 Zambian Education System Change, 1996–Present

Pre-1996 1996–2011 2012–present

Primary (grade 1–7) Basic (grades 1–9)

Early childhood education Primary (grades 1–7)

Secondary (grade 8–12)

Secondary (grades 8–12)

High (grades 10–12)

TEVETa TEVET

Higher education Higher education Higher education

Note: TEVET = technical education and vocational entrepreneurship training.a. The TEVET sector had major policy reform in 1998.

FIGuRE 4 .1 Number of Schools Offering Each Grade Level, 2013

4,272

745

617337

89

3,330

8

86

Grades 1–4

Grades 1–7

Grades 1–9

Grades 1–12

Grades 8–9

Grades 8–12

Grades 10–12

Unknown

Source: Education Statistical Bulletin 2013.

TABLE 4 .2 Number of Schools Offering Each Grade Level, 2013 and 2014

Grade level 2013 2014 % change, 2013–14

1–4 745 551 −26

1–7 4,272 3,864 −10

1–9 3,330 2,821 −15

1–12 89 66 −26

8–9 8 16 100

8–12 337 403 20

10–12 86 25 −71

Source: Education Statistical Bulletin 2013 and 2014.

Education Sector Performance and Service Delivery: QSDS 33

FIGuRE 4 .2 Number of Basic (Grades 1–9) and Secondary (Grades 8–12) Schools and Enrollment, 2008–13

Central government, government aided

Community, unknown

Private, church

Private, church, community

2008

2009

2010

2011

2012

2013

2008

2009

2010

2011

2012

2013

3,500,000 3,290,000

4,790

2,994 2,576 2,851

482 498485

2,637 2,6422,896

4,709 4,903 5,016 5,219 5,420

3,352,000

3,165,0003,411,000

3,592,000

3,526,000

624,000

673,000569,000

624,000

744,000

743,000

3,000,000

2,000,000

2,500,000

1,500,000

500,000

0

1,000,000

Nu

mb

er o

f st

ud

ents

en

rolle

d

3,500,000

3,000,000

2,000,000

2,500,000

1,500,000

500,000

0

1,000,000

Nu

mb

er o

f st

ud

ents

en

rolle

d

Nu

mb

er o

f sc

ho

ols

12,000

8,000

10,000

6,000

2,000

0

4,000

12,000

8,000

10,000

6,000

2,000

0

4,000

Nu

mb

er o

f sc

ho

ols

a. Basic b. Secondary

Source: Education Statistical Bulletin 2013.

FIGuRE 4 .3 Enrollment per School, by Grade Level and Province, 2013

Centra

l

Coppe

rbelt

Easte

rn

Luap

ula

Lusa

ka

Muc

hinga

North

weste

rn

North

ern

South

ern

Wes

tern

All pro

vince

s0

100

200

300

400

500

600

700

800

900

620

789

481

590 564

486438

499

411

638

231

146120

570

Nu

mb

er o

f st

ud

ents

en

rolle

d p

er s

cho

ol

Grades 1–4 Grades 5–7 Grades 8–9 Grades 10–12

Source: Education Statistical Bulletin 2013.Note: Out of 9,484 schools, 617 schools (7 percent) are categorized as unknown grades, and these schools are excluded from calculations.

combined rate for grades 8–12, indicating that a large number of students may delay their school-ing within these grade levels. According to the  Zambia Demographic and Health Survey (ZDHS), during the 2013–14 survey year, for pri-mary, the gross attendance rate was 102 percent

and the net attendance rate was 80 percent, com-pared with 59 and 40 percent, respectively, for secondary.1 These rates are little changed from the gross and net enrollment rates of the 2010 LCMS. Box 4.1 describes issues related to the calculation of enrollment rates.

34 Education Sector Performance and Service Delivery: QSDS

TABLE 4 .3 Gross and Net Enrollment Rates, by Grade Level and Data Source, 2010–14Enrollment rate (%)

Grades 1–7 (primary) Grades 8–12 (secondary)

Source Gross enrollment rate Net enrollment rate Gross enrollment rate Net enrollment rate

LCMS 2010 99 73 61 40

ZDHS 2013–14 102 80 59 40

ESB 2013 127 107a —b —b

LCMS 2015 (preliminary) 105 78 64 43

Source: 2010 LCMS; 2015 LCMS preliminary findings; 2013–14 ZDHS; and Education Statistical Bulletin (ESB) 2013.a. The net enrollment rate cannot exceed 100%, and the problems with ESB’s calculation of the net enrollment rate is discussed in box 4.1.b. — = not available. The ESB does not report gross or net enrollment rates for secondary; instead, it reports them for grades 10–12 (high school level). The gross enrollment rate for grades 10–12 is 32.6%, while the net enrollment rate for grades 10–12 is 28%.

BOx 4 .1 Calculation of Gross and Net Enrollment Rates using LCMS, ZDHS, and ESB Data

This study refers to three data sources for reporting of gross and net enrollment rates: the Living Conditions Measurement Survey (LCMS), the Zambia Demographic and Health Survey (ZDHS), and the Educational Statistical Bulletin (ESB).

Education Statistical Bulletin 2013. The Ministry of Education, Science, Vocational Training, and Early Education (MESVTEE) publishes an annual Education Statistical Bulletin, which primarily provides information about the quantity (for example, size of enrollment and schools) and quality (for example, efficiency and school inputs) of education through the Annual School Census exercise. The information represents the official statistics for MESVTEE.

Living Conditions Monitoring Survey 2010 and 2015 (preliminary findings). Every five years, the Central Statistical Office (CSO) conducts a household survey representative at the national, provincial, and urban and rural levels. This survey provides information for various government and donor policies and programs such as poverty rates and other socioeconomic indicators (for example, health and education indicators). It covers economic activities at the household level and health and education status at the individual level. In November 2015, the CSO released the preliminary findings of 2015 LCMS.

Zambia Demographics and Health Survey 2013–14. The 2013–14 ZDHS is a national survey designed to provide up-to-date information on background characteristics of the respondents’ demographics, including education and health. The 2013–14 ZDHS was implemented by the CSO in partnership with the Ministry of Health.

Calculation of enrollment rates. The calculation of both gross and net enrollment rates requires estimating the population of the relevant age group (for example, children ages 7–13 for primary and children ages 14–18 for secondary). The ESB uses the population statistics projected by the CSO based on the census in 2000. However, the estimates do not reflect demographic changes in trends such as birth and death rates or migration over the past decade. This causes the problem of overestimation because the denominator is based on projected population, while the numerator is based on actual enrollment figures. This results in a net enrollment rate higher than 100 percent. The EBS states, “The MESVTEE is aware that enrollment for some age groups may exceed the current population estimates and has been working with the CSO to adjust the population projections.” For consistency and reliability, this report uses the LCMS and ZDHS to calculate major education indicators.

Using data from either the LCMS or ZDHS, which are for a representative sample of households in the country, gross and net enrollment rates seem to be consistent, since both denominator and numerator are based on the same survey data. This report uses data from the 2010 LCMS, 2013–14 ZDHS, and 2015 LCMS (preliminary findings).

Note: This report refers to preliminary findings of LCSM 2015 that were published by the CSO in November 2015. If the information is not available in the preliminary findings for 2015, this report uses LCMS 2010 for analysis.

Education Sector Performance and Service Delivery: QSDS 35

On average, children enroll in grade 1 with a one-year delay, and half of them repeat at least one school year during their primary education cycle. A delay of one year in entering primary school is the average; students delay complet-ing grade 7 by 1.5 years, which indicates that almost half of the students repeat one school year in the primary cycle (figure 4.4). The aver-age age of grade 1 students in 2010 was 8 years old, which is one year above the standard age of primary school entry. Grade 8 students were, on average, 15.6 years old (1.6 years above the normal secondary school starting age). The gross enrollment rate for the lower primary grades is around 100 percent, with some fluctu-ation for middle-higher primary grades. This fluctuation may be the result of population fluctuations in those age groups. The gross enrollment rate is slightly higher for boys than for girls throughout.

Equal access to primary education remains a challenge, given the substantial differences between provinces. In particular, Eastern and Luapula provinces are lagging behind other provinces. Eastern Province has the lowest access. In 2010 Eastern Province had the lowest gross enrollment rate in all grades (85 percent for

primary, 53 percent for lower secondary, and 22  percent for upper secondary); although access  in Eastern Province improved in 2015, according to preliminary findings (CSO, 2015), it remains relatively lower than in other provinces (figure 4.5). According to 2015 LCMS, Luapula Province has the lowest access of all provinces in primary school (gross enrollment rate of 94 percent). Access to upper-secondary education remains significantly low (gross enrollment rate of 50 percent in 2015), with substantial differ-ences among provinces, ranging from 32 to 72 percent in 2015.

The percentage of out-of-school children ages 7–18 is 25 percent for boys and 26 percent for girls. There is little gender gap until age 15, but after age 15 girls tend to be out of school more than boys (figure 4.6). For both girls and boys in 2010, close to half of children were not in school at age 7. This percentage declines from ages 7 to 9 and then remains under 20 percent for both boys and girls from ages 10 to 14. At age 15, a gender gap begins to build, with the percentage of out-of-school children increas-ing much more for girls than for boys. The per-centage of children not enrolled in school increases between ages 15 and 18, respectively,

FIGuRE 4 .4 Gross Enrollment Rate, by Grade Level and Gender, 2010

Ag

e (y

ears

)

Grade level

Gro

ss e

nro

llmen

t ra

te (

%)

0

20

40

60

80

100

120

140

7

9

11

13

15

17

19

19.6

15.6

14.5

8

Average age Average school ageBoy Girl

G1 G2 G3 G4 G5 G10G6 G7 G8 G9 G11 G12

Source: 2010 LCMS.Note: The information is not yet available in 2015 LCMS preliminary findings.

36 Education Sector Performance and Service Delivery: QSDS

from 20 to 36 percent for boys and from 19 to 52 percent for girls.

Girls’ completion rate is higher than that of boys in primary (grade 7); however, boys’ com-pletion rate is higher than that of girls in higher secondary (grade 12). In 2010, 72 percent of chil-dren ages 16–18 completed primary school-ing  (grade 7), 48 percent of those ages 18–20 completed lower-secondary education (grade 9), and 24 percent of those ages 20–22 completed secondary education (grade 12) (figure 4.7).

In terms of gender difference, among youth ages 16–18, 76 percent of girls and 68 percent of boys completed grade 7; among youth ages 20–22, 23 percent of girls and 26 percent of boys completed grade 12. According to the 2013–14 ZDHS, in 2013, for grade 7 (primary), the completion rate  was 79 percent (66 percent for boys and 73 percent for girls), for grade 9, it was 46 percent (47 percent for boys and 45 percent for girls), and for grade 12 (secondary), it was 23 percent (26 percent for boys and 20 percent for girls).

FIGuRE 4 .5 Gross Enrollment Rate, by Grade Level and Province, 2010 and 2015

Source: 2010 LCMS and 2015 LCMS report.Note: Authors’ calculation using 2010 LCMS. Preliminary findings from 2015 LCMS report (CSO).

0

20

40

60

80

100

120

101

59

103

72

108

79

94

64

112

86

54

109 10899

106 109101

7265

32

58

74

49

37

53

3235

Easte

rn

North

ern

Centra

l

Luap

ula

South

ern

Wes

tern

Lusa

ka

Coppe

rbelt

North

wes

tern

Grades 1–7 Grades 8–9 Grades 10–12

Gro

ss e

nro

llmen

t ra

te (

%)

b. 2015 (preliminary findings)

0

120

85

53

104

82

34

102

82

48

9485

28

102

86

43

105

87

44

9991

67

10192

34

100 98

65

22

100

80

60

40

20

Gro

ss e

nro

llmen

t ra

te (

%)

a. 2010

Easte

rn

North

ern

Centra

l

Luap

ula

South

ern

North

wes

tern

Lusa

ka

Wes

tern

Coppe

rbelt

Education Sector Performance and Service Delivery: QSDS 37

FIGuRE 4 .6 Percentage of Out-of-School Children, by Age and Gender

60

5047

19

40

30

20

10

0

Age (years)

% o

f ch

ildre

n w

ho

are

no

t in

sch

oo

l

7 8 9 10 11 12 13 14 15 16 17 18

Boys Girls

36

52

45

Source: 2010 LCMS.

Internal Efficiency

With regard to repetition of grades, there is no gender difference. For both boys and girls, drop-out and repetition become an issue between grades 7 and 9. Repetition rates are at 5 percent

throughout the lower-primary grades (1–5), but start rising to reach 8 percent toward the end of primary education (table 4.4). The dropout rate is 1 percent until grade 4 and increases grad-ually to reach 4 percent by grade 9. This increase

FIGuRE 4 .7 Completion Rates, by Grade Level and Gender, 2010

Girls Boys

100

80

60

40

20

0

76

Co

mp

leti

on

rat

e (%

)

68

Grade 7(Ages 16–18)

49 47

Grade 9(Ages 18–20)

23 26

Grade 12(Ages 20–22)

Source: 2010 LCMS. The information is not yet available in 2015 preliminary LCMS report.Note: The completion rate is the number of children who completed the grade divided by the total number of children in the appropriate age group.

TABLE 4 .4 Repetition and Dropouts, by Grade Level and GenderRate (%)

Repetition Dropout

Grade Boy Girl Total Boy Girl Total

Grade 1 5 5 5 1 1 1

Grade 2 5 5 5 1 1 1

Grade 3 5 5 5 1 2 1

Grade 4 6 5 5 1 1 1

Grade 5 5 5 5 1 2 2

Grade 6 6 6 6 2 2 2

Grade 7 8 7 8 2 4 3

Grade 8 6 6 6 2 4 3

Grade 9 15 15 15 2 6 4

Grade 10 1 1 1 0 1 1

Grade 11 1 1 1 1 2 1

Source: Education Statistical Bulletin 2013.

38 Education Sector Performance and Service Delivery: QSDS

is due mainly to the high dropout rate of girls in these grades.

Only 62 percent of primary school graduates continue their education in secondary, and only 43 percent of grade 9 students continue their education in lower secondary (grades 8–9). As partially captured in the dropout rate, transition rates between grades 7 and 8 (primary to second-ary) and between grades 9 and 10 (lower second-ary to upper secondary) are quite low, at 62 and 43 percent, respectively. Transition rates are slightly lower for girls than for boys, but not sig-nificantly lower: 63 percent of boys transi-tion  from grade 7 to grade 8, compared with 61 percent of girls, and 44 percent of boys transi-tion from grade 9 to grade 10, compared with 43 percent of girls (figure 4.8). However, provin-cial differences in the transition rate are quite stark; for example, in Northern Province, the transition rate from primary to secondary is as low as 35 percent, while in Copperbelt, it is 78 percent (figure 4.9). The transition rate from grade 9 to grade 10 is as low as 29 percent in Southern Province and as high as 58 percent in Northwestern Province.

Reasons for dropping out differ between boys and girls and also between rural and urban areas. Girls drop out of school mainly

due to early marriage—81 percent in urban areas and 85 percent in rural areas (figure 4.10). The second most common reasons for girls to drop out are migration in urban areas, with 9 percent of dropouts, and poverty or economic issues in rural areas, with 4 percent of drop-outs. Boys have a wider variety of reasons for dropping out of school (figure 4.11). In rural areas, 31 percent of boys drop out because they are needed for farm work. Inability to pay school fees causes a 21 percent dropout rate in urban areas and a 13 percent dropout rate in rural areas.

FIGuRE 4 .8 Transition Rates, by Grade Level and Gender, 2013

100

80

6063 61

44 4340

20

0Boys Girls Boys Girls

Grades 7–8 Grades 9–10

Tra

nsi

tio

n r

ate

(%)

Source: Education Statistical Bulletin 2013.

FIGuRE 4 .9 Transition Rates, by Grade Level and Province

100

80

60

40

North

en

Mun

ching

a

Centra

l

Luap

ula

Easte

rn

South

ern

Lusa

ka

Wes

tern

North

weste

rn

Coppe

rbelt

35

47 4750 52

34

60

32

61

39

63

29

68

55

59

38

73

58

78

52

20Tra

nsi

tio

n r

ate

(%)

0

Grades 7–8 Grades 9–10

Source: Education Statistical Bulletin 2013.

Education Sector Performance and Service Delivery: QSDS 39

FIGuRE 4 .10 Reasons That Girls Drop Out of School, by urban-Rural Location

100

80

60

40

20Rea

son

fo

r d

rop

pin

g o

ut

(%)

0Nonpaymentof school fees

Earlymarriage

Migration(transfer)

Lack ofparentalsupport

Lack ofmotivation

Work(farm)

Work(nonfarm)

Poverty oreconomic

issue

51

8185

9

1 1 2 0 0 1 1 2 410

Urban Rural

Source: 2014 PETS-QSDS.Note: Main reasons for dropping out were obtained from head teachers.

FIGuRE 4 .11 Reasons That Boys Drop Out of School, by urban-Rural Location

Urban Rural

100

80

60

40

20Rea

son

fo

r d

rop

pin

g o

ut

(%)

0Nonpaymentof school fees

Earlymarriage

Migration(transfer)

Lack ofparentalsupport

Lack ofmotivation

Work(farm)

Work(other)

Behavioralproblem

21

13

59 10

3 47 5 7 6

31

643

Source: 2014 PETS-QSDS.Note: Main reasons for dropping out were obtained from head teachers.

In total, 14,928 girls in primary and secondary schools became pregnant in 2013, and, among pregnant girls, 84 percent became pregnant before grade 10. The pregnancy rate of girls has not decreased since 2008 (figure 4.12). Among these

pregnant girls, only 39 percent (5,829 girls) were readmitted in 2013. In 2014, the number of preg-nant girls increased to 16,378, and 7,391 girls were readmitted. Still, a significant percentage of teach-ers ask students to leave school due to pregnancy,

40 Education Sector Performance and Service Delivery: QSDS

and rural teachers do this more often than urban teachers: according to the 2014 Public Expenditure Tracking Survey (PETS)–Quantitative Service Delivery Survey (QSDS), 26 percent of rural sec-ondary school teachers asked students to leave school, compared with 13 percent of urban sec-ondary school teachers (figure 4.12).

A large share of students are absent for at least one school day a week. On the 2014 PETS-QSDS, 35 percent of grade 5 students and 22 percent of grade 9 students reported having been absent at least one school day during the previous week (table 4.5). On average, grade 5 students missed 15 percent of school days, and grade 9 students

FIGuRE 4 .12 Pregnancy among School-Age Girls, 2008–14

17,000

Nu

mb

er o

f p

reg

nan

t g

irls

% o

f p

reg

nan

t g

irls

wh

o w

ere

read

mit

ted89 88 88 89

86 84

81

13,936

15,497 15,586 15,707

14,849 14,928

16,378

16,000

15,000

14,000

13,000

12,000

11,000

10,0002008 2009 2010 2011 2012 2013 2014

0

20

40

60

80

100

% of girls who were readmittedNumber of pregnancies

Source: Education Statistical Bulletin 2013, 2014.

FIGuRE 4 .13 Percentage of Teachers Asking Students to Leave Because of Pregnancy or Nonpayment of School Fees, by Education Level and Rural-urban Location

Urban Rural

Primary

% o

f te

ach

ers

Secondary

Urban Rural

30

25

20

15

10

5

0

6

10

5

20

1113

8

26

Ask to leave due to nonpayment of fees Ask to leave due to pregnancy

Source: 2014 PETS-QSDS.

Education Sector Performance and Service Delivery: QSDS 41

missed 8 percent. By a slight margin, in both grades, urban students tend to miss more school days than rural students. The reasons for grade 5 students being absent include illness (61 percent), family reasons such as funerals or weddings (9  percent), and others such as work, family holidays, and dirty clothes.

Student Learning Outcomes

There has been little improvement in student learning since the first national learning assess-ment was conducted in 1999 (figure 4.14). Scores

for English and mathematics for grade 5 remain as low as 32 and 35 percent, respectively (2014 NAS). Assessment scores for life skills and local language are also low, at 32 percent, respectively. Grade 9 scores fare worse, with scores of 29 percent for mathematics and 36 percent for English and sci-ence (table 4.6). While there is no gender gap at the grade 5 level, girls score lower than boys in mathe-matics and science at the grade 9 level.

At the primary level, learning gaps between urban and rural schools are much more pro-nounced than those between genders. Students in rural schools have lower learning levels than stu-dents in urban schools for both grades 5 and 9 (figure 4.15). For example, in English, grade 5 stu-dents in rural schools score 29 percent, on aver-age, while grade 5 students in urban schools score 37 percent, on average. For grade 9 students, stu-dents in urban schools score 40 percent in English, while students in rural schools score 32 percent. In addition, there is some difference across the provinces; Lusaka Province has the highest scores (38 percent for math and 37 percent for English), while Central Province has the lowest (32 percent for math and 29 percent for English) (figure 4.16).

Income disparity in learning at both grades 5 and 9 is an issue. In general, average scores for

TABLE 4 .5 Student Absence, by Grade Level, urban-Rural Location, and Gender% of students

% of days absent, last week

% of students absent at least once last week

Location and gender

Grade 5 Grade 9 Grade 5 Grade 9

Location

Urban 16 10 36 25

Rural 14 7 34 20

Gender

Boys 15 8 36 22

Girls 14 9 34 22

Total 15 8 35 22

Source: 2014 PETS-QSDS.

FIGuRE 4 .14 Trends in Grade 5 Student Learning Assessment, 1999–2014

English Mathematics

19990

20

40Sco

re (

%)

34.338.5 38.5 39.3 38.3 35.8

32.634.135.334.533.933.2

60

80

100

2003 2006 2008 2012 2014

Source: 1999–2014 NAS.

42 Education Sector Performance and Service Delivery: QSDS

students from the top 33 percent of household income are higher than the scores for students from lower-income households (figure 4.17). In English, grade 5 students from the bottom 33 percent of family income score 30 percent, while students from the top 33 percent of family income score 42 percent. Of more concern is that the distribution of scores for students from high-er-income families is widely spread, whereas the distribution of scores for students from lower- income families is concentrated around the lower mean (figure 4.18). This may indicate that family income could be the major factor determining

TABLE 4 .6 Grade 5 and 9 Learning Assessment, by GenderScore (%)

Grade Math English Life skills

Science Zambian language

Grade 5

Boys 36 32 34 n.a. 36

Girls 35 32 35 n.a. 35

Total 35 32 35 n.a. 35

Grade 9

Boys 31 36 n.a. 37 n.a.

Girls 27 36 n.a. 35 n.a.

Total 29 36 n.a. 36 n.a.

Sources: 2014 PETS-QSDS; 2014 NAS.Note: n.a. = not applicable.

FIGuRE 4 .15 Grade 5 and 9 Learning, by urban-Rural Location

Math English EnglishLife skills Math Science

Grade 5

0

20

40 39 3729

3442

3136 35

30 28

40 37 3532

60

80

100

Grade 9

Zambianlanguage

Urban Rural

Sco

re (

%)

Sources: 2014 PETS-QSDS; 2014 NAS.

FIGuRE 4 .16 Grade 5 Learning, by Province

Centra

l

North

ern

South

ern

Wes

tern

Luap

ula

Coppe

rbelt

Easte

rn

North

weste

rn

Muc

hinga

Lusa

ka

32

0

20

40

60

80

100

2933 34 35 36 36 36 36 36 38 37

32 32 293331 313227

Math English

Sco

re (

%)

Source: 2014 NAS.

Education Sector Performance and Service Delivery: QSDS 43

FIGuRE 4 .17 Grade 5 and 9 Learning, by Income Tercile

100

80

60

40

20

0Math English Science

Grade 5

Sco

re (

%)

Grade 9

Math English Life skills Zambianlaguage

34 3642

Poor Middle Rich

30 31

42

32 35

48

32 35

48

35 3541

3135

45

27 283336 36 3836 36 38

Sources: 2014 PETS-QSDS; 2014 NAS.Note: Poor is defined as the bottom 33.3 percent (tercile) in family income (asset index) and rich is defined as top 33.3 percent in family income.

FIGuRE 4 .18 Distribution of Grade 5 and 9 Learning, by Students’ Family Income

A. Grade 5

Den

sity

0

0 20 40

Score (%)

60 80 100

.01

.02

.03

.04

a. Mathematics b. English reading

Den

sity

Score (%)

0 20 40 60 80 100

.01

0

.02

.03

.04

Den

sity

Score (%)

c. Life skills

0 20 40 60 80 100

0

.01

.02

.03

d. Zambian language

Den

sity

Score (%)

0 20 40 60 80 100

0

.01

.02

.03

Poor Richfigure continues next page

44 Education Sector Performance and Service Delivery: QSDS

low student learning outcomes for low-income students. Certain factors are correlated with stu-dent learning and family income. For example, textbook ownership and student attendance are positively correlated with student learning, and students from richer families are more likely to have a textbook and higher attendance than stu-dents from poorer families.

Education Inputs and Service IndicatorsLocation and Physical Inputs

More primary schools are located in remote areas compared to secondary schools. There is a big difference in the remoteness of primary schools

across provinces, which is correlated with the population density of the province. Zambia is one of the less densely populated countries (ranked 201 out of 241 countries). This means that physical access to schools is an issue in remote areas. Measuring the percentage of schools located in relatively remote areas (more than 25 kilometers from an urban center proxied by the presence of a commercial bank) found that 49 percent of secondary schools are within a 25-kilometer radius of a commercial bank, compared with 68 percent of primary schools (figure 4.19). In terms of travel time, grade 5  students reported taking 34 minutes, on average, to get to school from home, and grade 9  students reported taking 37 minutes. In gen-eral, rural students spend approximately 10 more

B. Grade 9

0

0 20 40 60

Score (%)

a. Mathematics

80 100

.01

.02

Den

sity

.03

.04

b. English

Score (%)

0 20 40 60 80

Den

sity

0

.01

.02

.03

.04

.05

c. Science

Score (%)

Den

sity

0

0 20 40 60 80

.01

.02

.03

.04

Poor Rich

Figure 4 .18 (continued)

Sources: 2014 NAS; 2014 PETS-QSDS.Note: Kernel density plot of percentage scores by subject and by rich and poor.

Education Sector Performance and Service Delivery: QSDS 45

minutes to arrive at school from home (29 min-utes for grade 5 students in urban and 37 minutes for those in rural areas; 31 minutes for grade 9 students in urban and 41 minutes for those in rural areas). In particular, students in Eastern Province travel more than an hour (69 minutes) from home to school (figure 4.20).

School infrastructure improved between 2008 and 2014. For example, the share of primary schools with a library increased from 5.6 to 13  percent, those with electricity rose from 19 to  36 percent, and the number of pupils per

classroom declined from 97 to 70 (table 4.7). Secondary schools have better infrastructure than primary schools, especially with regard to basic amenities such as accessibility of potable water, electricity, and latrines. However, in terms  of educational facilities, students still have  limited access to libraries and science labs  even at the secondary level (28 and 22 percent of schools, respectively). Regional differences in school infrastructure vary, and none of the sampled schools in Central and Muchinga provinces had a library, and none of the sampled schools in Luapula, Lusaka, Northwestern, and Southern provinces had a science lab (table 4.8).

FIGuRE 4 .19 Distance from the School to a Commercial Bank, by Education Level

–5 kilometers

5–25 kilometers

25– kilometers

15

17

68

40

0

80

60

20

100

49

29

32

Primary Secondary

% o

f s

cho

ols

Source: 2014 PETS-QSDS.

FIGuRE 4 .20 Travel Time to School, by Province

80

70

60

50

40

30

20

Trav

el t

ime

(min

ute

s)

10

0

North

ern

Luap

ula

Wes

tern

Coppe

rbelt

Muc

hinga

North

Wes

tern

Lusa

ka

Centra

l

South

ern

Easte

rn

All pro

vince

s

1721 22

25 27 29 29 30

53

34

69

Source: 2014 PETS-QSDS.

TABLE 4 .7 School Infrastructure, by Education Level% of schools, unless otherwise noted

Type of infrastructure Primary Secondary

Library 13 28

Science lab 3 22

Potable water 75 93

Electricity 36 56

Number of boys per latrine 83 73

Number of girls per latrine 75 67

Number of pupils per classroom 70 58

Number of teachers per house 1 1

Source: 2014 PETS-QSDS.

46 Education Sector Performance and Service Delivery: QSDS

School Days and Teaching Hours

Schools have, on average, 2.2 shifts for both pri-mary and secondary schools; however, the shift system in schools does not seem to have affected the number of schooling hours for students. When comparing schools with less than two shifts and those with two or more shifts, the number of schooling hours stays relatively the same for grades 1–9, with less than 0.3 hour dif-ference (figure 4.21).2 Schooling hours per stu-dent are much more affected by the school’s location—urban or rural. As seen in figure 4.22, a gap is apparent for students in rural schools. Students across grades 1–9 average 4.7 schooling hours in rural areas compared to 5.5 hours in urban areas. This gap is greatest for grades 1–4, with students in rural schools attending school for one hour less than students in urban schools, on average. For grades 10–12, rural schools have, on average, 0.4 more schooling hour per shift than urban schools.

Teachers spend a little more than 8 hours per day in school. Primary school teachers spend about 5.5 hours teaching, and secondary school teachers spend slightly more than 4 hours teach-ing; rural teachers spend slightly more time in school and teaching—per shift, teaching hours are 4.5 for primary, which is similar to students’

TABLE 4 .8 Infrastructure at the Primary Level, by Province% of schools, unless otherwise noted

Province Central Copperbelt Eastern Luapula Lusaka Muchinga Northwestern Northern Southern Western

Library 0 8 38 2 38 0 8 6 23 9

Science lab 1 1 12 0 0 1 0 10 0 0

Potable water 87 72 100 66 100 49 65 41 94 58

Electricity 32 51 30 49 68 20 22 17 37 32

Number of boys per latrine 117 141 55 89 104 91 84 65 68 63

Number of girls per latrine 105 125 60 77 97 80 73 54 64 57

Number of pupils per classroom 84 76 56 80 96 59 78 61 64 64

Number of teachers per house 1 6 1 2 3 1 1 1 1 1

Source: 2014 PETS-QSDS.Note: Since 2014 PETS-QSDS is a sample survey, the figures in the table are in statistical terms. Hence, 0% does not necessarily mean that no school in the province has the facility.

FIGuRE 4 .21 Schooling Hours and Number of Shifts, by Grade Level

7

6

5

4

3

2

1

0

Grade

1 2 33 4 5 6 7 8 9 10 11 12

3.8

6.6

5.9

Nu

mb

er o

f h

ou

rsin

th

e sc

ho

ol d

ay

Less then 2 shifts 2 or more shifts

4.1

Source: 2014 PETS-QSDS.

FIGuRE 4 .22 Schooling Hours, by Grade Level and urban-Rural Location

Grade

Urban Rural

3.9

4.9

6.6

1 2 3 4 765 8 9

5.7

Nu

mb

er o

f h

ou

rs in

th

e sc

ho

ol d

ay

1

3

4

5

7

6

2

0

Source: 2014 PETS-QSDS.

Education Sector Performance and Service Delivery: QSDS 47

schooling hours (table 4.9). At the primary level, 44 percent of teachers teach more than one shift per day (1.5 shifts, on average), and more teach-ers in rural areas teach more than one shift than teachers in urban areas (50 and 20 percent, respectively). At the secondary level, 57 percent of teachers teach more than one shift, an aver-age of 1.7 shift, and there is little difference between the percentage of rural and urban teachers teaching more than one shift per day (55 and 57 percent, respectively). Secondary school teachers (grade 9) spend more time in school than primary school teachers (grade 5) in that they spend less time teaching and more time on preparation and administration than primary school teachers.

Multigrade teaching does not seem to affect teaching hours, but it may place an additional burden on teachers due to the greater amount of preparation needed. In addition to teaching shifts, multigrade teaching is a relatively com-mon practice, especially in rural areas. Multigrade teaching and teaching hours are not particularly related, as shown by the comparison between urban and rural areas (figure 4.23). However, multigrade teaching may affect teach-ing preparation time for each grade and subject knowledge for specific grades. In rural areas, 37 percent of grade 5 teachers teach more than one grade, compared with only 10 percent of grade 5 teachers in urban areas. These grade 5 teachers in multigrade classes in rural schools teach, on average, almost one grade more than grade 5 teachers in multigrade classes in urban schools.

However, the urban-rural gap is smaller for grade 9 teachers. At the grade 9 level, teachers teach by subject rather than by grade level, so despite the rural-urban differences in the per-centage of multigrade teachers, they both teach, on average, 2.4 grade levels. Even if multigrade teachers spend almost the same amount of time teaching or conducting other activities in school and students receive the same teaching time regardless, multigrade teachers may pay less attention to students’ homework or outcomes because they have less preparation time for each grade and subject.

Number of Classrooms and Teachers

School shifting and temporary classrooms ease classroom crowding. Primary schools have a very high number of pupils per permanent class-room (92 percent for urban and 85 percent rural schools); however, after taking into account tem-porary classrooms and the school shift system, the rate declines to 33 students per classroom (figure 4.24). The rural-urban gap in the number of pupils per classroom is evident, especially at the secondary level, even after teaching shifts are considered. In secondary, rural schools have a higher number of pupils per classroom (60) than urban schools (53), while in primary, urban schools have more pupils per classroom (79) than rural schools (68).

TABLE 4 .9 Teaching Hours in Grade 5 and 9 Schools, by urban-Rural Location

Type of activity Grade 5 Grade 9

urban Rural urban Rural

Hours at school per day 7.4 8.3 8.3 8.7

Teaching hours per day 5.3 5.6 3.9 4.6

Teaching hours per shift 4.8 4.1 2.7 3.1

Nonteaching hours per day (preparation and administration) 2.9 3.7 3.8 4.4

Break hours per day 0.6 1.1 1.1 1.3

Source: 2014 PETS-QSDS.

FIGuRE 4 .23 Multigrade Teaching in Grade 5 and 9 Schools, by urban-Rural Location

Urban Rural Urban Rural

10

0

37%3.0

2.0

1.0

0.0

2.42.4

1.9

1.2 20

30

40

Grade 5 Teacher Grade 9 Teacher

Percentage multigrade teaching

Mean number of grades teaching

10%

28%

Nu

mb

er o

f g

rad

es t

hat

teac

her

s te

ach

% o

f te

ach

ers

teac

hin

gm

ore

th

an o

ne

gra

de

17%

Source: 2014 PETS-QSDS.

48 Education Sector Performance and Service Delivery: QSDS

The pupil-teacher ratio is about 40 overall, but the ratio is higher for lower grades (figure 4.25). This may be partially because the supply of school inputs such as classrooms and teach-ers has not kept up with the expansion of enroll-ment in lower grades; however, enrollment has not increased rapidly in upper-primary and

secondary grades. Also, because secondary teachers are subject teachers, the pupil-teacher ratio is lower. However, as secondary enroll-ment is expected to keep increasing as the sup-ply of schools expands and the cost of schooling declines (free secondary policy and the con-struction of secondary schools are priorities according to the 2010 National Implementation Framework), Zambian secondary students may  face higher pupil-teacher ratios in the near  future if secondary teachers are not recruited as fast as the expansion of secondary school buildings.

Provincial gaps in both the pupil-classroom and the pupil-teacher ratio are substantial. Higher pupil-classroom and pupil-teacher ratios are correlated with higher gross enrollment rates. This may indicate that the expansion of school demand is outstripping the supply of education inputs (classrooms and teachers). For instance, Lusaka Province has the second highest number of students per classroom, the highest number of  students per teacher, and the highest gross enrollment rate among 10 districts (figure 4.26). In contrast, Eastern Province has the lowest gross enrollment rate and low pupil-teacher and

FIGuRE 4 .24 Number of Pupils per Classroom, by urban-Rural Location and Education Level

100

80

60

40

20

0Urban Rural

Primary

Nu

mb

er o

f p

up

ils p

er c

lass

roo

m

Urban Rural

Secondary

Students perpermanent classroom

Students per allclassroom types

Students perclassroom and shifts

92

79

34

85

68

33

71

60

30

5653

24

Source: 2014 PETS-QSDS.

FIGuRE 4 .25 Pupil-Teacher Ratio, by Grade Level and urban-Rural Location

60

50

40

30

20

10

Nu

mb

er o

f p

up

ils p

er t

each

er

0

Urban Rural

4551

43

37 39

31

812

Grade 2 Grade 5 Grade 7 Grade 9

Source: 2014 PETS-QSDS.Note: According to the Education Statistical Bulletin, the pupil-teacher ratio is 41 for all grades.

Education Sector Performance and Service Delivery: QSDS 49

pupil-classroom ratios (figure 4.27). Therefore, low pupil-teacher and pupil-classroom ratios are not necessarily a good indication of educational performance, especially educational access.

Textbooks

Limited availability of textbooks is a serious issue, especially in primary schools. Five pri-mary school students share less than 1 textbook for each subject (1.0 for mathematics, 0.9 for

English, and 0.9 for science) and five secondary school students share between 1 and 1.5 text-books, depending on the subject (1.0 for math, 1.7 for English, and 1.0 for science). At the pri-mary level, there seems to be no urban-rural difference in the availability of textbooks; how-ever, at the secondary level, rural schools face a greater  shortage of textbooks than urban schools (figure 4.28). The textbook shortage was reconfirmed during the classroom observation conducted at the grade 5 level: 84 percent of

FIGuRE 4 .26 Pupil-Classroom Ratio, by Province

0

20

40

60

80

47

25

53

64

28 28

66 66 64

76 72

32 33 36 38 41

76

89

30 30

100

Nu

mb

er o

f p

up

ils p

er c

lass

roo

m

Easter

n

Muchin

ga

Northe

rn

Southe

rn

Coppe

rbelt

Wes

tern

Luap

ula

Northw

ester

n

Lusa

ka

Centra

l

Students per classroom and shiftsStudents per all classroom types

Source: 2014 PETS-QSDS.

FIGuRE 4 .27 Pupil-Teacher Ratio, by Grade Level and Province

39

0

20

40

60

Pu

pil-

teac

her

rat

io (

nu

mb

ero

f p

up

ils p

er t

each

er)

80

100

39 423736

50 50

354141

4451 56

60 6267

6258

44 443636

35

19

34

24302626

19

Muc

hinga

Easte

rn

Coppe

rbelt

Centra

l

Wes

tern

North

weste

rn

Luap

ula

South

ern

North

ern

Lusa

ka

Grade 2 Grade 5 Grade 7

Source: 2014 PETS-QSDS.

50 Education Sector Performance and Service Delivery: QSDS

teachers use textbooks, while only 8 percent of students use textbooks.

As shown in the textbook-pupil ratio, a large percentage of schools reported a shortage of text-books in academic year (AY) 2013: 82 percent of primary schools and 63 percent of secondary schools. Among primary schools, 70 percent requested the District Education Board Secretaries (DEBS) to supply the textbooks; however, only 26 percent of schools said that the request was resolved at least partially. Among secondary schools, 51 percent of secondary schools requested the Provincial Education Offices (PEO) to supply the textbooks, but only 22 percent of schools said that the request had been met at least partially (2014 PETS-QSDS). The high shortage of textbooks and the low level of resolution partly reflect the delay in textbook procurement at the MESVTEE headquarters caused by changes in curriculum and revision of textbooks in 2013.3

Teacher Quality and ManagementAbout 22 percent of grade 5 school teachers have a general certificate of education or lower degree.

The majority of secondary school teachers have  a  diploma, a higher degree, or at least a certificate. On average, rural teachers have lower qualifications than urban teachers (figure 4.29). In Eastern and Muchinga provinces, approximately 40 percent of teachers have a general certificate of education or lower education level (figure 4.30).

FIGuRE 4 .28 Pupil-Textbook Ratio (per Five Students), by Subject, Education Level, and urban-Rural Location

3.0

2.0

1.0

0.0

Primary

Ave

rag

e n

um

ber

of

text

bo

oks

that

5 s

tud

ents

sh

are

Math English Science Math English Science

Secondary

Urban Rural

0.91.0 1.0 0.9

0.80.9

1.6

0.7

2.5

1.3 1.2

0.8

Source: 2014 PETS-QSDS.

FIGuRE 4 .29 Qualification of Grade 5 and 9 Teachers, by urban-Rural Location

100

% o

f te

ach

ers

80

60

40

20

0

19

57

2410

11

79

35

51

14

Urban RuralGrade 5 Grade 9

Urban Rural

GCE or lower Certificate Diploma or higher

30

68

1

Source: 2014 PETS-QSDS.Note: GCE = general certificate of education. Certificate: primary certificate, special education certificate

Education Sector Performance and Service Delivery: QSDS 51

Subject Knowledge

The subject knowledge of secondary school teachers (grade 9) is not sufficient. Grade 9 teach-ers scored approximately 70 percent across all subjects and require more subject knowledge to teach effectively (figure 4.31).When teachers were tested using the same examination ques-tions that were put to their students, grade 5

teachers scored more than 90 percent in mathe-matics, English, and life skills. Across all subjects, rural teachers performed better than urban teachers.

Teacher qualification has little correlation with subject knowledge. At the primary level (grade 5), there is no difference in subject knowl-edge between teachers with a certificate and

FIGuRE 4 .30 Qualification of Grade 5 Teachers, by Province

Easte

rn

Muc

hinga

North

weste

rn

Centra

l

Coppe

rbelt

South

ern

Luap

ula

Lusa

ka

North

ern

Wes

tern

0

% o

f te

ach

ers

GCE or lower Certificate Diploma or higher

1

58

90

73

27

58

28

1416

64

21

16

44

40

18

56

26

28

54

18

36

42

22

39

37

24

41

47

13

20

40

60

80

100

Source: 2014 PETS-QSDS.Note: GCE = general certificate of education.

FIGuRE 4 .31 Teacher Assessment in Grades 5 and 9, by Subject and Rural-urban Location

Sco

re (

% o

f co

rrec

t an

swer

s)

67.367.569.767.073.2

62.2

87.081.9

95.793.191.7 95.2

RuralUrban

Grade 5 Grade 9

ScienceEnglishMathZambianlanguage

Life skillsEnglishMath

100

80

60

40

20

0

91.889.7

Source: 2014 PETS-QSDS.Note: Due to the small sample size in the grade 9 teacher assessment, the mean difference between rural and urban is not statistically significant, and the inference we can draw from this is limited.

52 Education Sector Performance and Service Delivery: QSDS

those with a diploma or higher. Only in mathe-matics do teachers with a general certificate of education or lower degree score lower for subject knowledge, although the difference is not statisti-cally significant (figure 4.32).4

Attendance

Official records (available at primary and sec-ondary schools) indicate 83 percent attendance rates for both primary and secondary schools (in  February and June of 2013). However, 16  percent of primary school teachers and 12  percent of secondary school teachers were absent for more than 50 percent of school days in June and were a major factor in bringing down the teacher attendance rate. There was no urban-rural gap in attendance, but there was a provincial gap: districts in Northern and Muchinga provinces have the lowest attendance rates (67  and 72 percent), while districts in Lusaka, Northwestern, and Luapula provinces have the highest attendance rates (between 87 and 92 percent) (figure 4.33).5

During random unannounced visits to pri-mary schools, the survey team found that teacher

attendance on a given day was 82 percent (on par with the official rate of 83 percent). During these visits, the team found that 54 percent of teachers were engaged in teaching activities (in class-rooms or outdoors), 8 percent were in classrooms but not teaching, and 20 percent were in school  but not involved in teaching activities (figure 4.34). About 70 percent of teachers’ time is spent teaching, and about 70 percent of

FIGuRE 4 .33 Official Attendance Rate (June), by Province

100

0

20

40

60

80

North

ern

Muc

hinga

Wes

tern

Centra

l

Coppe

rbelt

South

ern

Easte

rn

Lusa

ka

North

weste

rn

Luap

ula

6772

82 84 85 85 86 87 8792

Off

icia

l att

end

ance

rat

e (%

)Source: 2014 PET-QSDS.

FIGuRE 4 .34 Teachers’ Activities

60

0

10

20

30

40

5052

2

8

20 18% o

f te

ach

ers

Teach

ing in

class

room

Not te

achin

g

but in

clas

sroo

m

Not te

achin

g an

d

not in

clas

sroo

m Not in

atte

ndan

ce

Teach

ing

outd

oors

Source: 2014 PETS-QSDS.

FIGuRE 4 .32 Relationship between Teacher Qualifications and Subject Knowledge Assessment

869292 91

93 94

0

20

40

60

80

% o

f te

ach

ers

100

Math English

GCE or lower Certificate Diploma or higher

Source: 2014 PETS-QSDS.Note: GCE = general certificate of education.

Education Sector Performance and Service Delivery: QSDS 53

teachers who are present in school are engaged in teaching activities.

Head teachers (or senior teachers) are aware of teacher absences because most teachers receive permission to take leave from them. During a sample week, 19 percent of teachers reported taking leave, often without permission. Among grade 9 teachers, 23 percent reported taking time off during the week, compared with 18 percent of grade 5 teachers, with an average of 2.3 and 20 days, respectively (figure 4.35). For grade 9, urban teachers seem to be absent more than rural teachers.

The main reasons for being absent from school include illness, illness of others, and funerals. For grade 5 teachers, 38 percent were absent due to illness and 23 percent were absent due to the illness of others. For grade 9 teachers, 23 percent were absent due to funerals and 18 percent were absent due to illness (figure 4.36). Other reasons for being absent were, for grade 5 teachers, leave, maternity leave, studies, and per-sonal schooling and, for grade 9 teachers, meet-ings with officials, official leave, monthly woman’s leave, and accidents.

Management

Primary schools use a large number of volunteer or contract nongovernment teachers. Of all pri-mary school teachers (government and nongov-ernment schools), 13 percent of teachers are volunteer (unpaid) or contract (paid) teachers and 8 percent are part-time teachers (2014 PETS-QSDS). Among government schools and grant-aided primary schools, 7 percent of teachers are volunteer or contract nongovernment teachers, with some differences across provinces. Luapula Province has 12 percent contract teachers, and Central and Muchinga provinces have large per-centages of volunteer teachers (8 percent) in gov-ernment and grant-aided schools (figure 4.37). At the secondary level, 19 percent of all second-ary school teachers are contract teachers, and, among government schools, only 2 percent are contract teachers.

The teacher attrition rate is hovering around 11 percent, despite two major increases in teacher salaries, and the reason for this attrition is “unknown.” The government has been invest-ing in teachers, with major salary increases in

FIGuRE 4 .35 Number of Absent Days (during a Sample Week) among Grade 5 and 9 Teachers, by urban-Rural Location

3.0

2.0

1.0

0.0

30

20

10

0

Nu

mb

er o

f d

ays

abse

nt

(wit

h o

r w

ith

ou

t p

erm

issi

on

)

% o

f te

ach

ers

wh

o r

epo

rted

taki

ng

tim

e o

ff

Urban RuralUrban Rural

1.7 19

0.0

18

0.1

2.124

0.1

2.6

22

0.0

1.9

Grade 5 teacher Grade 9 teacher

Number days offwith permission

Number days offwithout permission

Percentage timeoff last week

Source: 2014 PETS-QSDS.

54 Education Sector Performance and Service Delivery: QSDS

FIGuRE 4 .36 Reasons Given by Grade 4 and 9 Teachers for Being Absent

Own illness

Funeral

School emergency

Other

Illness of others

Transportation

Grade 5 teacher

38.4%

22.7%

3.3%

10.8%

0.0%

24.9%17.8%

12.3%

11.2%

22.7%

2.1%

33.9%

Grade 9 teacher

Source: 2014 PETS-QSDS.Note: Teachers were asked to give reasons for absence. Other reasons includes official leave, maternity leave, and studies. During the Education Joint Annual Review field mission, teachers in the school visited gave the collection of salary as a major reason for being absent; however, teachers in PETS-QSDS did not give the collection of salary as a significant reason.

FIGuRE 4 .37 Type of Teacher in Government and Grant-Aided Primary Schools, by Province

100

% o

f te

ach

ers

Centra

l

Coppe

rbelt

Easte

rn

Luap

ula

Lusa

ka

Muc

hinga

North

weste

rn

North

ern

South

ern

Wes

tern

80

60

40

20

0

8

92

6

94

9

91

12

88

100

8

92

100

3

97

31

96

3

97

Government Paid contract Volunteer Full-time

Source: 2014 PETS-QSDS.Note: Sample includes government and grant-aided primary schools.

2007 and 2013. A recruitment policy aims to have 5,000 new government teachers graduat-ing from teachers college annually. The total number of teachers increased 40 percent, from 66,145 in 2006 to 93,164 in 2013, and 70,937

out of 93,164 teachers were employed in government schools in 2013 ( figure 4.38). Only 10 percent of attrition is from retirement, and  about 8 percent of attrition is due to either  death or illness. It is necessary for the

Education Sector Performance and Service Delivery: QSDS 55

government to identify the real cause of high teacher attrition.

Teacher transfers between schools are com-mon, and teachers are transferred from urban to rural schools more often than the reverse. Also, 13 percent of all teachers reportedly transferred to other schools in 2013, according to the MESVTEE: 15 percent of grade 5 and 12 percent of grade 9 teachers transferred to different schools in 2013 (2014 PETS-QSDS). The per-centage of teachers transferring from urban to rural schools is high: 53 percent of transfers happen at grade 5, and 44 percent of transfers at  grade 9 are from urban to rural schools ( figure 4.39). The main reasons for teacher trans-fers between schools are marriage (20 percent) and better location (16 percent); only 26 percent are official transfers by the ministry (ESB 2013).

In addition to the high transfer rate, a large percentage of teachers said that they wanted to transfer, but few made an official request: 31 percent of grade 5 teachers and 42 percent of grade 9 teachers wanted to transfer, but only 8 and 17 percent, respectively, requested a transfer. Among grade 5 teachers, 43 percent of teachers in rural areas wanted to transfer, compared with

30 percent in urban areas (figure 4.40). Among grade 9 teachers, 63 percent of teachers in rural areas wanted to transfer compared to 54 percent in urban areas.

In 2013, 27 percent of teachers had received any training the previous year; however,

FIGuRE 4 .38 Number of Teachers and Attrition Rate, by Level of Education, 2006–13

2008 2009 2010 2011 2012 20130.00

0.50

0.10

0.15

20072006

12%12%

15%14%

13%

BasicSecondary

0

20,000

40,000

60,000

80,000

Nu

mb

er o

f te

ach

ers

Teac

her

att

riti

on

rat

e

13,62314,717

16,497 16,822

18,639 19,615

52,522 56,895 61,811 60,865 63,052 65,014 72,967

7%

10%9%

73,549

15,412 12,947

Source: Education Statistics Bulletin 2013.

FIGuRE 4 .39 Transfer Rates of Grade 5 and Grade 9 Teachers between urban and Rural Schools

% o

f te

ach

ers

Urban0

10

20

30

40

50

60

14 13

53

20

10

22

44

24

Grade 5 teacher Grade 9 teacher

Previous ruralPrevious urban

UrbanRural Rural

Source: 2014 PETS-QSDS.Note: Previous urban (rural) means teachers who were posted in urban (rural) schools before moving to current post.

56 Education Sector Performance and Service Delivery: QSDS

low-qualified teachers received less training than high-qualified teachers. Better-educated teachers (with at least a certificate) received more in- service training than teachers with only a general certificate of education or a lower  degree (figure 4.41). This could be partly  because government teachers, who tend  to be highly qualified, are eligible for government-provided training, whereas non-government teachers such as contract teachers or community school teachers are not. In order to upgrade the skills of less-educated teachers, the government should open government train-ing to contract and community teachers, who compose about a quarter of the supply of teach-ers in primary and secondary education (26 and 27 percent, respectively).

At the secondary level, rural school teachers received significantly less training than urban school teachers: 36 percent of grade 9 teachers in urban government schools received some train-ing, compared with only 23 percent of grade 9 teachers in rural schools in 2013. At the primary level, teachers in schools located 5–25 kilometers from the DEBS office received significantly less training (14 percent) than teachers in schools

located adjacent to the office (25 percent) or farther away (26 percent) (figure 4.42).

At the primary level, training is provided largely at the local level (in school or in the dis-trict resource center) and through distance learning. At the secondary level, training is pro-vided largely through distance learning. At the primary level, rural schools use more training through the district resource center, while urban schools use more in-school as well as dis-tance learning (figure 4.43). At the secondary level, distance learning is highly used in both urban and rural schools (34 and 28 percent, respectively).

Teacher absenteeism has been a problem for several decades. At the primary level, 16 percent of head teachers reported not having taken any action against absent teachers, compared with 9  percent at the secondary level (figure 4.44). Even though a significant portion of head teach-ers do not take any action, a large percentage do report teacher absenteeism to the DEBS or the PEO (41 percent at primary and 30 percent at secondary) or warn teachers (25 percent at pri-mary and 24 percent at secondary). However,

FIGuRE 4 .40 Percentage of Grade 5 and 9 Teachers Wanting and Requesting to Transfer, by urban-Rural Location

10

0

20

30

43

30

4

11

54

63

11

21

40

50

60

70

% o

f te

ach

ers

Grade 5 teacher Grade 9 teacher

Want to transfer

Want to transfer

Requestedtransfer

Requestedtransfer

Urban Rural

Source: 2014 PETS-QSDS.

FIGuRE 4 .41 Percentage of Teachers Receiving Training in 2013, by Qualification and Education Level

Primary0

20

40

1622 24

10

34

25

60

80

100

GCE or lower Certificate Diploma or higher

Secondary

% o

f te

ach

ers

Source: 2014 PETS-QSDS.Note: % of teachers receiving any training in 2013; the percentage by education qualification excludes teachers who had already received at least primary certificate from the training.

Education Sector Performance and Service Delivery: QSDS 57

the actions taken (report or warning) do not seem to have reduced the problem of teacher absenteeism.

The Teaching Council of Zambia was estab-lished by the Teaching Profession Act of 2013

in order to regulate teachers, their practice, and professional conduct and to accredit and regulate colleges of education. According to the act, the Teaching Council was created to develop, maintain, and improve appropriate standards of qualification for the teaching pro-fession and to promote continuing professional

FIGuRE 4 .42 Percentage of Teachers Receiving Training in 2013, by Grade Level, urban-Rural Location, and Remoteness

% o

f te

ach

ers

rece

ivin

g t

rain

ing

100

22

36

25 23 2529

14

26 2634

80

60

40

20

0Urban Rural Less than 5

kilometers5 kilometers to 25

kilometersMore than 25

kilometers

By location By distance to DEBS

Grade 5 Grade 9

Source: 2014 PETS-QSDS.Note: DEBS = District Education Bureau Secretaries.

FIGuRE 4 .43 Training Location, by Grade Level

40

30

20

10

0

20 21

4

13

18

4

15

18

6 6

29

5

Grade 5

In school

Districtresource center

Provincialtraining center

Headquatertraining center

Distancelearning

PrivateInstitution

Grade 9

% o

f tra

inin

g

Source: 2014 PETS-QSDS.

FIGuRE 4 .44 Action Taken by Head Teachers to Address Teacher Absenteeism, by Education Level

50

40

30

20

10

0

None

War

n te

ache

rs

Refer

mat

ter t

o

DEBS/PEO

Suspe

nd/tr

ansfe

r

teac

hers Oth

er

41

30

4 5

15

32

2524

16

9

Primary Secondary

% o

f h

ead

tea

cher

s

Source: 2014 PETS-QSDS.

58 Education Sector Performance and Service Delivery: QSDS

development among teachers. Establishing the  authority is a  first step toward improv-ing  student learning. It will be important for  the council to use the objective informa-tion, data, and relevant studies acquired through the MESVTEE and the Examinations Council for Zambia to promote good-quality teaching.

Factors Related to Teacher Transfer and AttendanceSchool facilities and physical environment are not significant factors in teachers’ desire to transfer or actual requests for a transfer, but the location of the school (rural versus urban) is a major factor. Furthermore, unlike the con-ventional belief, teachers in a hardship posi-tion do not have a stronger desire to transfer than other teachers. Teachers in rural schools have a significantly higher probability of want-ing to transfer (15 percent higher than their urban counterparts) and requesting a transfer (6 percent higher), but the desire to transfer does not increase with the remoteness of schools. Furthermore, the physical environ-ment of school, such as the availability of elec-tricity and a library, the number of staff houses and staff rooms, and the number of latrines, does not have a significant correlation with teacher transfer after taking the location of the school into account (table A.1 in appendix A). Also, school finance (school revenue) is not correlated with teacher transfers.

Teaching intensity, such as the number of teaching hours, number of students per class-room, school shifts, and the multigrade teach-ing, is associated with teachers’ desire to transfer or request a transfer. Grade 5 teachers teaching multigrades are more likely to request a transfer than teachers teaching a single grade (7 percent) and, as daily teaching hours increase by one hour, teachers’ likelihood of wanting to transfer increases 6 percentage points. Teachers in schools with multiple shifts have a 4.5 percent higher likelihood of requesting a transfer than teachers in schools without shifts (table A.2 in appendix A).

Head Teacher and Teacher MotivationHead teachers’ intention to advance in their career and pro-social motivations are strong predictors of  teachers’ desire to transfer and their  actual requests for transfer (see box 4.2 and table A.4 in appendix A). If the head teacher has a strong inten-tion to progress through the ranks in 5–10 years, teachers in the school will be more likely to want to transfer (11 percent more likely) and to request a transfer (9 percent more likely). In contrast, if the head teacher is more interested in being in the same community in 5–10 years, teachers will also be less likely to request a transfer (10 percent less likely). At the same time, if the head teacher per-ceives that community interests and his or her self-interest overlap (inclusion of the “other” in the self), teachers in the school will be less likely to want to transfer to a different school.

Similarly, teachers’ intention to advance in their career and their pro-social motivation are strongly correlated with their desire to transfer and actual request to transfer. If teachers intend to advance in their career, they will be 28 percent more likely to want to transfer to a different school and 3 percent more likely to request a transfer. In the meantime, teachers who intend to remain in the community in the future will be 27  percent less likely to want to transfer, and those who perceive community and self-interest as overlapping will be less likely to want to trans-fer and to request a transfer than those who do not perceive them as overlapping.

Furthermore, teachers’ career and pro-social impact motivations are correlated with teacher attendance at school. Teachers with a higher calling—meaning they do not separate their work from their personal life and try to fulfill their achievement through work—higher desire for career advancement, and higher pro-social motivation, have lower probability of being absent from school. Also, teachers with higher pro-social motivation have significantly fewer absent days (table A.4 in appendix A).

Pedagogy and Classroom ObservationThe majority of teachers in the sample use various pedagogical tools. Most teachers (90–94 percent)

Education Sector Performance and Service Delivery: QSDS 59

use a syllabus, while fewer (62–63 percent) follow the curriculum (table 4.10). Only 85–89 percent of teachers use textbooks, 66–70 percent use learning objectives, and 54–68 percent use a learning plan.

Classroom observation is conducted as part of 2014 PETS-QSDS. Each class lasts an average of 38 minutes (between 30 and 45 minutes), and surveyors observe teacher and student activities every minute. During observations 90 percent of teachers introduced the lesson to start the class, and 44 percent summarized the lesson at the end of class (figure 4.45). Assignment of homework was not a common practice—only

15 percent of teachers assigned homework and  only 11 percent of teachers collected and corrected the homework—and 83 percent of teachers used the local language as a medium of  instruction. Approximately 41 percent of

BOx 4 .2 Measuring the Career and Prosocial Motivations of Head and Other Teachers in the 2014 PETS-QSDS

The 2014 PETS-QSDS captures the motivations of head and other teachers in order to identify how their motivations at work, serving community, and intention in career advancement correlate with how teach-ers perform in schools and the community.

Intention in career advancement. The 2014 PETS-QSDS measures the intention of the head and other teachers to advance in their career by asking whether they aim to be ranked more highly in 5–10 years or to remain in the community in 5–10 years.

Motivation at work. In general, individuals with high “career orientation” tend to invest in their work, to advance within the occupational structure, and to pursue monetary gain, while individuals with a high “calling” do not separate their work from their personal life and achieve personal fulfillment through their work (Bellah et al. 1998).

Pro-social motivation. Motivations in serving the community are captured in two ways: desire for having a positive social impact and desire for including the “other” in the self. Desire for having a positive social impact measures the degree to which an individual desires and benefits psychologically from the positive impact of his or her work on others. The inclusion of the other in the self measures the closeness of the relationship between oneself and the community (Aron et al. 1992).

More detailed information and questionnaires are shown in figures A.1–A.4 and table A.3 in appendix A.

TABLE 4 .10 Teachers’ use of Pedagogical Tools, by Grade Level% of teachers

Tool Grade 5 Grade 9

Syllabus 90 94

Curriculum 63 62

Textbooks 85 89

Learning objectives 66 70

Learning plan 68 54

Source: 2014 PETS-QSDS.

FIGuRE 4 .45 Lessons and Homework

90

44

1511

83

0

20

40

60

80

100

Less

on In

trodu

ction

at th

e be

ginnin

g

Summ

arize

d les

son

at th

e en

d

Teac

her a

ssign

ed

hom

ewor

k

Teac

her c

ollec

ts or

answ

er h

omew

ork

Use lo

cal la

ngua

ge

for m

edium

of

instru

ction

% o

f te

ach

ers

Source: 2014 PETS-QSDS.

60 Education Sector Performance and Service Delivery: QSDS

teaching time was spent interacting with all children as a group, and an average of 30 percent of the time was spent waiting for pupils to com-plete the task assigned (figure 4.46).

Student Performance and Education InputsStudent assessment scores seem to be largely determined by students’ characteristics, such as their socioeconomic status, motivation, and per-sonality; however, some portion of the scores can be explained by school-level and teacher charac-teristics. This is captured in the analysis of vari-ance (intra-cluster correlation); about 14–17 percent of the variance in test scores (math and English, respectively) is between schools, about 9–10 percent of the variance is between teachers, and only 1–2 percent of the variance is between provinces (table 4.11). The largest variance in test scores is between students (70–76 percent).

At the school level, the presence of a library, the amount of school grant received per child, a lower pupil-teacher ratio, and longer school hours are significantly and positively correlated with student learning outcomes. With regard to school facilities, the presence of a library is the

only one with a significant and positive associa-tion with student learning outcomes. Other infrastructure, such as having a laboratory, availability of electricity, and number of latrines per students, are not strongly related to student learning. At the same time, the increase in the amount of school grant received per student is associated with student learning, but the magnitude of this positive correlation is weak; ZMW 20 higher per student grant is associated with only 0.012 and 0.022 percent higher scores in math and English, respectively (table A.5 in appendix A).

Among all of these factors, the type of teacher position (unpaid volunteer and paid contract teacher)6 is the most important factor that is pos-itively correlated with student learning. The stu-dents of contract teachers have significantly higher learning outcomes (8 percent higher for

FIGuRE 4 .46 Minute-by-Minute Teaching Activities

50

41

3028

14 1210

40

30

20

10

% o

f te

ach

ers

0Interacts withall children as

a group

Waiting forpupils to

complete task

Writing onblackboard

Listening topupils read

Interactswith childrenone-to-one

Lectures topupils or pupils

only listen

Source: 2014 PETS-QSDS classroom observation module.Note: The percentages of time that individual teachers spent on each activity are averaged out across the teacher sample.

TABLE 4 .11 Intra-Cluster Correlation: Variance in Text ScoresCorrelation (%)

Cluster Math English

Between provinces 1 2

Between schools 14 17

Between teachers 9 10

Source: 2014 NAS grade 5 student assessment.

Education Sector Performance and Service Delivery: QSDS 61

math and 5 percent higher for English) than stu-dents of noncontract teachers (government teachers). Even after taking into account the sub-ject knowledge of contract teachers7 (table A.5 in appendix A), the type of position remains impor-tant. Given its significance for student learning, it is worthwhile to investigate the significantly higher result for students of contract teachers. Box 4.3 describes a contract teacher program that was implemented in Kenya and found to be effective (Duflo, Dupas, and Kremer 2012).

At the student level, having a textbook, the frequency of questions that a teacher asks a stu-dent, and a student’s school attendance are important factors in higher learning outcomes. Having a textbook increases student math scores by 3 percent, while having a textbook is highly correlated with wealth of the student’s family.

However, the relationship between textbooks and student learning persists even after control-ling for wealth of the student’s family. Missing one day of school in a week is associated with a 1 to 1.5 percent lower score in math and English, respectively, and the number of days missed is higher for poorer students than for richer stu-dents. Homework and frequency of homework assigned are not significant predictors of student learning outcomes, but the frequency of ques-tions that a teacher asks a student is a strong indi-cator of higher learning outcomes (table A.6 in appendix A).

In addition, teachers’ subject knowledge, personalities, and motivations and students’ personalities and motivation are important factors at the teacher and student levels. International literature shows the importance of

BOx 4 .3 Classroom Size and Contract Teachers: Kenya Extra Teacher Program

Program goals were to reduce class size by providing funds to a school committee to hire extra teachers in the local area.

Program implementation details. The program consisted of the following:

1. Hire local contract teachers to supplement a quarter of civil service teachers.2. Give the school and the parent-teacher association (PTA) authority to hire and fire the contract

teachers.3. Provide some training to help schools and PTAs to supervise the recruitment of contract teachers and

to evaluate teacher performance, including checking voluntary attendance and renewing contracts at the end of the academic year.

Program impact and implication. The impact evaluation was conducted by a team of researchers at the Massachusetts Institute of Technology. The following summarizes their findings:

1. Contract teachers had lower absent rates than civil service teachers.2. Students of contract teachers had higher scores than students of civil service teachers at the end of the

year.3. The overall positive impacts were achieved only when the PTA was empowered through training and

supervision. Civil service teachers responded to the ETP by reducing their own effort (as evident in higher absenteeism and lower exam preparation time) and by encouraging the school committee to hire their own relatives. However, when the school committee and PTA were trained and empowered, civil service teachers did not reduce their efforts and did not interfere with the recruitment of contract teachers.

4. Eventually, the high-performing contract teachers were hired as permanent teachers by the civil service. This means that some civil service teachers could be hired initially as contract teachers by the school committee with PTA supervision.

62 Education Sector Performance and Service Delivery: QSDS

head teachers (their management style and motivation), along with other school inputs such as classroom size, infrastructure, and school environment, as well as teacher and stu-dent motivation and teacher subject knowledge (see appendix B for a review of the literature). In Zambia, teacher subject knowledge and student learning outcomes have a positive relationship in grade 5, especially for English and the local language. For example, a 1 standard deviation increase in teacher subject knowledge (9 percent for English and 14 percent for local language) is associated with an increase in student scores by 2–2.5 percent in English and 4–5 percent in Zambian language, while, in math, the relation-ship between teachers’ subject knowledge and students’ learning outcomes is positive but very marginal (table A.7 in appendix A). The size of the result should be interpreted carefully; since teachers’ subject knowledge is measured using the same examination sheet, the score is highly skewed to the right side of the distribution (around 90 percentage point in score), which may downplay the magnitude of the relationship between teachers’ subject knowledge and stu-dents’ learning outcomes.

Another significant determinant of higher student learning outcomes is having a female teacher. On average, students of female teach-ers have 4.8 percentage point higher scores in

math, 3.8 percentage point higher scores in English, and 2.7 percentage point higher scores in Zambian language than students of male teachers.

Regarding motivation and soft skills, gaining social power and recognition from teachers and other students and personality (agreeableness) are predictors of higher student outcomes. Along with student motivation, teacher motivation is positively correlated with student learning and includes factors such as the respect that teachers receive from students and the community, when the teacher applied for the job, and the teacher’s personality (level of neuroticism) (tables A.8 and A.9 in appendix A). Box 4.4 discusses how to measure soft skills.

From the public policy point of view, using teacher and student motivations to improve stu-dent learning could be considered once several pilots show robust results. This could be imple-mented through teacher recruitment and deployment and a process of student-teacher performance review and feedback; the design of this process should be tested and validated thor-oughly. In addition, development and monitor-ing of indicators that effectively measure teaching pedagogy could be considered. These indicators would help to identify the best teaching practices and enable teachers to improve the quality of their teaching through informed training.

BOx 4 .4 Measuring Soft Skills in the 2014 PETS-QSDS

The 2014 PETS-QSDS incorporated various measurements of soft skills—nonacademic skills and traits—for students, teachers, and head teachers. The two key aspects of soft skills covered were personality traits and motivations. Simple descriptive graphs of personalities and motivations are shown in figures A.1 to A.4 in appendix A.

Personality. One section of the self-complete questionnaire categorized the respondent’s personality through an adapted version of Rammstedt and John’s (2007) 10-item Big Five Inventory, which assesses five broad personality traits through a widely accepted and cross-culturally robust model of personality (John and Srivastava 1999 Schmitt et al. 2007). John and Srivastava (1999) define the Big Five as follows:

• Extraversion: an energetic approach to the world (outgoing, sociable, assertive, active)

box continues next page

Education Sector Performance and Service Delivery: QSDS 63

Notes 1. The terms gross attendance rate and net atten-

dance rate are the same concept as gross and net enrollment rates when the data source is the national household survey. The ZDHS uses gross and net attendance rates to indicate the education participation rate. The net attendance rate indi-cates participation in primary schooling for the population ages 7–13 and in secondary schooling for the population ages 14–18. The gross atten-dance rate measures participation at each level of schooling among those of any age from 5 to 24 years.

2. A gap in schooling hours after grade 9 may be due to small sample size. Hence, the difference is not statistically significant.

3. The figures include all primary and secondary schools sampled (government schools and non-government schools).

4. For secondary education, the sample size in teacher assessment is too small to conduct statis-tically meaningful analysis.

5. In the United States, an approximately 90 percent teacher attendance rate is appropriate after taking into account official sick leave and assigned per-sonal leave days.

• Agreeableness: a pro-social and communal orientation toward others (trusting, altruistic, affectionate, modest)

• Conscientiousness: a socially prescribed impulse control that drives task- and goal-directed behavior (organized, thorough, follows norms and rules, careful, efficient)

• Neuroticism: a negative emotionality (feels anxious, nervous, sad, or tense easily and often)• Openness: an openness to experience (imaginative, open-minded, original, artistic).

Personality traits of both students and teachers were found to be predictors of student learning outcomes. Student assessment scores improved with greater levels of agreeableness in students and level of neuroti-cism in teachers.

Motivation. Psychometric measures adapted from a recent study on incentives for Zambian health workers by Ashraf, Bandiera, and Lee (2015) assessed the motivations of teachers and head teachers. To evaluate the motivations of the head and other teachers for choosing to apply for a teaching position, respondents ranked the following seven potential motivations in order of importance: “good future career,” “allows me to serve the community,” “earns respect and high status in the community,” “pays well,” “offers stable income,” “interesting job,” and “allows me to acquire useful skills.”

Teachers who were motivated by gaining respect and high status in the community when applying for their position were found to have better student learning outcomes.

For students, an adapted version of the Inventory of School Motivation questionnaire (Ali and McInerney 2005) created their academic motivational profile based on eight potential motivations and some sample statements used to measure them:

• Task: “I like to see that I am improving in my schoolwork.”• Effort: “The harder the problem, the harder I try.”• Competition: “I am only happy when I am one of the best in the class.”• Social power: “At school I like being in charge of a group.”• Affiliation: “I do my best work at school when I am working with others.”• Social concern: “I like to help other students do well at school.”• Praise: “Praise from my teachers for my good schoolwork is important to me.”• Token: “If I got rewards at school, I would work harder.”

Students with higher assessment scores were found to be more strongly motivated by social power and recognition or praise from their teachers.

Box 4 .4 (continued)

64 Education Sector Performance and Service Delivery: QSDS

6. For regression analysis, both unpaid volunteer and paid contract teachers are categorized as con-tract teachers. Student performance of contract teachers is compared with that of government teachers (regular teachers and senior teachers, such as head of department and head teachers).

7. Government teachers have slightly higher scores in subject knowledge than contract teachers. This may imply that the type of position is the major factor driving student learning outcomes. The regression includes teacher- and major school-level inputs that may be positively correlated with both student learning outcomes and hiring of con-tract teachers, such as teacher subject knowledge, teaching experience, number of libraries, amount of school grant, pupil-teacher ratio, schooling hours, type of school (government, community, private and grant-aided schools), and school loca-tion and remoteness.

ReferencesAli, Jinnat, and Dennis M. McInerney. 2005. “An

Analysis of the Predictive Validity of the Inventory of School Motivation.” SELF Research Centre, University of Western Sydney, Australia.

Aron, Arthur, Elaine N. Aron, and Danny Smollan. 1992. “Inclusion of Other in the Self Scale and the structure of interpersonal closeness,” Journal of personality and social psychology.

MESVTEE (Ministry of Education, Science, Vocational Training and Early Education). 2013. Educational Statistical Bulletin. Lusaka, Zambia.

MESVTEE (Ministry of Education, Science, Vocational Training and Early Education). 2014. Educational Statistical Bulletin. Lusaka, Zambia.

Zambia Central Statistical Office. 2015. Preliminary Living Conditions Measurement Survey Report. Lusaka, Zambia.

B. Rammstedt, O. P. John. 2007. “Measuring personal-ity in one minute or less: A 10-item short version of the Big Five Inventory in English and German” Journal of Research in Personality.

John, O. P., & Srivastava, S. 1999. “The Big Five trait taxonomy: History, measurement, and theoretical perspectives”. Handbook of personality theory and research.

Schmitt, David P. et al. 2007. “The Geographic Distribution of big Five Personality Traits: Patterns and Profiles of Human Self-Description Across 56 Nations.” Journal of Cross-Cultural Psychology.

Ashraf, Nava, Oriana Bandiera, and Scott Lee. 2015. “Do-Gooders and Go-Getters: Career Incentives, Selection, and Performance in Public Service Delivery.” Working Paper, Harvard Business School, Cambridge, MA.

Bellah, Robert N., Richard Madsen, William M. Sullivan, Ann Swidler, and Steven M. Tipton. 1988. Habits of the heart: individualism and commitment in American life. University of California Press, Berkeley, CA.

Duflo, Esther, Pascaline Dupas, and Michale Kremer. 2012. “School Governance, Teacher Incentives, and Pupil-Teacher Ratios: Experimental Evidence from Kenyan Primary Schools.” MIT Working Paper 12-07, Massachusetts Institute of Technology, Boston, MA.

Major Findings and Policy Recommendations 65

Chapter 5

Major Findings and Policy Recommendations

Education Sector Financing and Management: PETSFindings

There are inefficiencies and inequities in school finance (both public and private sources), as well as substantial differences in revenue per pupil at the school level, depending on the number of rich (or poor) students. Schools with more rich students tend to have more revenue (both public and private sources), while those with more poor students are likely to have less revenue. For instance, at the primary level, while schools with more poor students have revenue of ZMW 35 per pupil, those with rich students have revenue of ZMW 46 per pupil. At the secondary level, the difference is stark: ZMW 144 per pupil at poor schools and ZMW 390 at rich schools. There is also a huge disparity in school revenue per pupil across provinces. While schools in Lusaka, on average, have ZMW 75 per pupil, schools in Western Province have only ZMW 7 per pupil.

While the primary school grant is pro-poor, the secondary school grant seems to support schools with richer students more than those with poorer students. At the primary level, on average, the schools with relatively poor students receive ZMW 16 per pupil, while the schools with relatively rich students receive, on average, ZMW 7 per pupil. At the secondary level, the relationship between the school grant amount and the income of students at the school is the opposite of the situation at the primary level. Secondary schools with relatively poor students receive ZMW 15 per pupil, while those with richer students receive ZMW 19 per pupil. Further, while poor students in primary school have more chance to receive a primary school

grant, poor students in secondary school have less chance to receive a secondary school grant.1

The distribution of school grants does not consistently follow the grants allocation formula, which causes inefficiency and inequity. The Ministry of Education, Science, Vocational Training, and Early Education (MESVTEE), including provinces and the District Education Board Secretaries (DEBS), does not efficiently distribute public funding, and 19 percent of schools do not receive any public funding. In addition, 28 percent of primary schools and 30  percent of secondary schools do not receive their school grant. School grants for primary and secondary education are critical elements of the government’s free education policy. This lack of grant receipt is mainly due to the fact that provinces do not seem to follow the grant alloca-tion formula. Further, there is a wide gap in the percentage of schools receiving public funds (including all public funds such as textbook grants). While 97 percent of schools receive pub-lic funds, only 69 percent of schools in Eastern Province receive them.

Recommendations

At the secondary level, school grants should be strategically distributed toward the more vulner-able population. Unlike primary school grants, secondary school grants do not have a formula that would eventually compensate poorer schools. Development of a formula with well-developed guidelines at the Provincial Education Offices (PEO) and school levels could dampen the inequity in finance at secondary schools.

Although the primary school grant is shown to be pro-poor, public funding needs to be targeted even more strategically to poor students and

66 Major Findings and Policy Recommendations

schools with more poor students. Despite the pro-poor funding of primary school grants, the inequity in school finance persists. This is espe-cially a problem when poor students have lower learning outcomes than rich students. In Zambia, poor students’ learning outcomes are concen-trated on low scores, without much difference between students. When schools with more poor students have fewer resources to spend, the learn-ing gap between poor and rich students gets larger.

While financial decentralization (direct deposit of school grants from the Ministry of Finance into school accounts) is strongly recommended, until financial decentralization is enforced, the school grant distribution and receipt process should be formalized, strictly enforced, and followed by investigation of misuse of funds. At the beginning of the academic year, the official document describing the school grant formula in each prov-ince and the expected timing of school grant dis-bursement should be made publicly available. Furthermore, at the end of the academic year, the actual disbursement of school grants at the DEBS level and actual grant disbursement dates should be made publicly available. All of this information could be distributed preferably through official notice via the MESVTEE website as well as through formal notice to the PEO, DEBS, and schools via the mail. Further, it is strongly recom-mended that the government expedite the current planning of financial decentralization. This direct disbursement would reduce any leakage or unin-tended use of school grants at the DEBS. Currently, secondary school grants are distributed directly from the Ministry of Finance to secondary schools, while the direct deposit of primary school grants faces difficulty due to the remoteness of primary schools and bank charges for grant withdrawal. These bank charges should be addressed before direct deposit is implemented.

In addition, the central government needs to improve the unpredictability of funding from the Ministry of Finance to MESTVTEE and then to PEO, DEBS, and schools. Unpredictability worsens school resource management and adversely affects the results of potentially suc-cessful programs. Funding for training and

school grants is known to be unpredictable and intermittent. Output-based funding focuses on education outputs; however, without steady and reliable funding flows, the desired outputs will not be achieved.

Further investigation is needed on textbook distribution. This issue was raised in the previous PETS-QSDS in 2006 and 2009. While textbook distribution is also in the midst of several impor-tant changes (for example, decentralization of procurement at the DEBS level, but with prices are set at the MESVTEE headquarters), the MESVTEE and the cooperating partners could collaborate more in this area. On the MESVTEE side, investigation could reveal why there is a huge gap between “approved provision” at the Ministry of Finance and “allocated budget” at the MESVTEE. The cooperating partners could support monitoring, on a sample basis, about procurement of textbooks at the DEBS level.

Access, Internal Efficiency, and Service IndicatorFindings

Zambia’s education performance was character-ized by rapid growth in enrollment with low inter-nal efficiency over the past decade. The total enrollment and number of schools expanded, and primary education could accommodate almost all primary school-age children (gross enrollment rate of 99 percent in 2010, 102 percent in 2013, and an estimated 105 percent in 2015, preliminary findings). However, secondary education has to improve its capacity to absorb more students (gross enrollment rate of 61 percent in 2010, 59  percent in 2013, and 64 percent in 2015, preliminary findings). Equitable access to primary education is still in progress. Furthermore, school-joining delay and repetition are common in pri-mary grades, and student absenteeism is high.

As a result of the government’s efforts to improve education service through the expansion of education expenditure and hiring of teachers, some service indicators have improved over the past decade. There are now more schools with

Major Findings and Policy Recommendations 67

access to a library and electricity (13 and 36 percent, respectively, an improvement from 5.6  and 19 percent) and fewer students in each classroom (70 pupils per classroom and 40 pupils per teacher, down from 97 and 59). However, access to drink-ing water has deteriorated since 2007.

There are stark differences across provinces in educational access and service indicators, and these differences demonstrate the different issues that each province faces. For example, the prov-inces with almost universal education (for exam-ple, Copperbelt Province) face a shortage of physical education inputs, reflected in crowded classrooms and a high pupil-teacher ratio, while more rural provinces (for example, Eastern Province) are lagging far behind in educational access (gross enrollment rate of 85 percent in pri-mary) and have the longest travel time to school (more than one hour).

Recommendation

It is necessary to recognize that each province is at a different education development stage and needs to come up with solutions specific to its own needs. The provincial differences in access and education service indicators demonstrate the different issues that each province faces. For example, the provinces with almost universal education face a shortage of physical education inputs, such as crowded classrooms and high pupil-teacher ratio; in these cases, expanding the existing school structure and teaching in shifts, using temporary classrooms, and hiring contract teachers would be realistic solutions. Provinces lagging behind in educational access should identify the cause, and the central gov-ernment should support the province in solving the problem. For example, Eastern Province faces the lowest level of primary education access. Simple statistics regarding school loca-tion and students’ travel time reveal that the long distance to schools may be a barrier to achieving universal education. Supporting large numbers of small-scale community schools may be a realistic solution, especially given that pub-lic funding is very tight. However, all of these

suggestions should be carefully designed and evaluated prior to full application.

Teacher Management and Student Learning Findings

Student learning outcomes have been persistently low, despite the improvement of some quantitative service indicators. Possible reasons include ineffi-cient teacher management, low quality of teaching (such as insufficient subject knowledge), and a serious shortage of textbooks. There has been no improvement in student learning at grade 5 since the first national learning assessment was con-ducted in 1999 (34.3 percent in math and 33.2 percent in English in 1999 versus 35.3 percent in math and 31.4 percent in English in 2014).

The significantly low pupil-textbook ratio is a serious issue, especially given that textbooks and a library are strong predictors of student learning outcomes. Five primary school students share less than 1 textbook for each subject (1.0 for mathematics, 0.9 for English, and 0.9 for science), and five secondary school students share between 1 and 1.5 textbooks, depending on the subject (1.0 for math, 1.7 for English, and 1.0 for science). Having a textbook increases a student’s math score by 3 percent.

Further, there is evidence that teacher manage-ment is inefficient: (a) teacher attrition is high, and a large percentage of teachers want to and do transfer schools, (b) teacher absenteeism has remained almost the same for the past decade, and (c) teacher subject knowledge is insufficient, espe-cially at grade 9. In a given month, 16 percent of primary school teachers are absent for more than 50 percent of the school days. Approximately 50 percent of teachers want to transfer, especially teachers in rural schools; 53 percent of transfers happen at grade 5, and 44 percent of transfers at grade 9 are from urban schools to rural schools. In terms of teacher subject knowledge, 50 percent of grade 5 teachers score below 90 percent in grade 5 subject materials, and the same percentage of

68 Major Findings and Policy Recommendations

grade 9 teachers score below 70 percent in grade 9 subject materials.

Teacher management (attendance and trans-fer) is strongly correlated with motivations of the head and other teachers. In general, schools whose head and other teachers have higher pro-social motivation and intention to remain in the community have lower likelihood of transfer requests and teacher absenteeism.

The 2014 PETS-QSDS investigated factors cor-related with student learning outcomes in Zambia at the school, teacher, and student levels. At the school level, a library, a lower pupil-teacher ratio, and a longer school day are correlated with higher student learning outcomes. At the teacher level, the use of contract teachers significantly raises the learning outcomes of students, and a teacher’s subject knowledge has a strong positive correla-tion with student learning outcomes. In addition, teachers who are motivated to receive respect from students and the community are positively cor-related with student learning outcomes. Also, teachers with a characteristic of “neuroticism” also are positively correlated with student learning out-comes. At the student level, family income is a major factor in determining low learning out-comes for low-income students. Several important factors such as textbook ownership and school attendance are correlated with students’ family income as well as learning outcome. Students who want to be recognized by teachers and their peers score higher than those who care less about recog-nition. Further, students who can agree with oth-ers (agreeableness) score higher than those who show less agreeableness. In addition, teachers who are motivated to receive respect from students and the community are positively correlated with student learning outcomes.

Recommendations

Concretizing and clarifying the textbook policy to ensure on-time delivery should be the first prior-ity of government. The misalignment of the text-book distribution process from curriculum design to procurement and delivery can cause serious delays or even no textbook delivery in a given

academic year, and such delays critically harm student learning, as shown in the strong correla-tion between textbook availability and learning outcomes. However, there is confusion in the text-book distribution process, especially with regard to centralized or decentralized textbook procure-ment. Decentralization of textbook procurement highly depends on the availability of textbooks in local markets and the transparency of textbook prices as well as the financial capacity of schools. Furthermore, the lack of funds for delivering text-books to remote schools also causes the delay or no delivery of textbook. It is strongly recom-mended that the government (a) secure sufficient funds to procure all textbooks necessary through higher public funding as well as through collabo-ration with cooperating partners, (b) create a har-monious system of curriculum development, textbook procurement, and textbook delivery, (c) secure the funds for textbook delivery both at the central and at the DEBS levels, and (d) inform and build the capacity of all stakeholders involved in the process, including the central government, PEO, DEBS, schools, publishers, and others.

Teacher management policy needs a thorough evaluation of the recruitment and deployment process as well as the introduction of a teacher performance feedback program. The importance of having a pro-social orientation and viewing teaching as a calling should be highlighted as good qualities for head and other teachers during recruitment, performance feedback, and manage-ment training. This report finds positive correla-tions between the head and other teachers’ personality and motivation and teacher transfer requests and attendance as well as student learn-ing outcomes. Further investigation of whether and how these factors influence student learning outcomes is needed to help the MESVTEE to design innovative approaches to recruitment and deployment of government teachers that use information regarding their motivation, personal-ity, and hometown. At the same time, monitoring of teacher performance and a proper feedback mechanism could be introduced.

In order to implement these recommendations, it is necessary to develop teacher performance

Major Findings and Policy Recommendations 69

metrics that capture multiple aspects of teacher quality (not only student performance but also teacher satisfaction, motivation, pedagogical style, and subject knowledge) that are proven to be effective in the Zambia context. This informa-tion could be used to improve teacher quality through a properly designed teacher perfor-mance feedback and incentive system. The Examinations Council for Zambia (ECZ) has turned its attention to teacher performance; the data collected through PETS-QSDS could be used as a baseline study, and the MESVTEE could conduct a robust analysis, such as an impact evaluation, on motivation and personal-ity. Further, it is essential to develop the capacity of central, provincial, and district education offi-cers and resource centers to interlink teacher performance and feedback.

An efficiently managed contract teacher pro-gram could address issues of deployment and absenteeism as well as the shortage of rural teachers. International studies show the effec-tiveness of hiring young teachers for the short term in three aspects: reducing the pupil-teacher ratio, lowering absenteeism, and improving stu-dent learning. However, if the government intends to introduce such a program, it should be contextualized for Zambian schools.

Key to the success of a contract teacher pro-gram is transparency of the recruitment process at the local level through the empowerment of school management and the parent-teacher asso-ciation (PTA). Training of recruitment practice to employ qualified teachers, according to not only their education level but also their knowledge of the subject, should be emphasized in the pro-gram. At the same time, this contract teacher pro-gram could be a gateway to recruiting high-quality civil service teachers who would stay in the local area after receiving successful performance evalu-ations for some years. Furthermore, during recruitment, their motivation and incentives at work could be examined to select better-suited contract teachers, especially in remote rural areas. Empowering the PTA could make the contract teacher management program more transparent and therefore more effective.

Teacher training can reinforce the efficient use of contract teachers and community schools. Currently, the government supports in-service training for government teachers and, partly due to this, the more-educated teachers (with at least a certificate) receive more in-service training than the less-educated teachers (with only a general certificate of edu-cation or a lower degree). Limited school access and some aspects of teacher management can be addressed by setting up more small-scale community schools and hiring young contract teachers, especially in remote areas. However, these teachers may need proper training to be effective.

Having an understanding of personality and motivation for both teachers and students can help in designing better classroom and peda-gogical styles that improve student learning. This report’s rudimentary findings can be a start for concretizing and formalizing the devel-opment of  soft skills. At the same time, the motivation of head and other teachers is a good indicator of how long teachers will stay in the school and the learning of students. If properly measured and tracked, these indicators will inform teacher management policy in the long term and give the Zambian education sector a deeper understanding of how students respond to teachers’ motivation and personality, which can improve teacher training and recruitment in the long run.

Besides education inputs, school financing, and school management, the MESVTEE should continue to build capacity in statistics at the min-istry. For instance, over the course of this study, inconsistencies were found in the current Education Statistical Bulletin (the net enrollment rate cannot go beyond 100 percent). It is highly recommended that the MESVTEE along with cooperating partners take full advantage of the various survey data available to the country, such as the Living Conditions Measurement Survey, and periodically verify the data consistency across data sets. The ECZ is a good example of building statistical capacity. Officials at the ECZ under-went several international trainings in statistics

70 Major Findings and Policy Recommendations

TABLE 5 .1 Short- and Long-Term Strategies for Achieving the Overarching Goals for Zambia’s Education Policy

Type of policy Short-term strategies (1–2 years) Medium-term strategies (3–5 years)

Textbook policy • Review textbook procurement capacity at both the DEBS and the MESVTEE and establish clear guidelines regarding textbook policy, including book selection procedures and distribution mechanisms and budgets

• Ensure that sufficient budget is released to schools for distributing all textbooks with the new curriculum

• Inform and build the capacity of all stakeholders involved in textbook procurement

School grant distribution • Enforce the school grant formula used by the DEBS to distribute school grant funds to primary and basic schools and monitor how well it is followed, including the public dissemination of transparent and clear guidelines regarding grant distribution

• Decentralize financial disbursement of school grants from MESVTEE or the Ministry of Finance directly to school bank accounts

Teacher recruitment, deployment, and attendance

• Conduct more research on teachers (including untrained, “contract” teachers and volunteers) to review how to improve attrition, effectiveness, and attendance

• Conduct a pilot on how the motivation of the head teacher and classroom teachers can improve the retention of quality teachers

• Prioritize training on leadership and management of head teachers to improve teacher attendance and delivery

• Have the Ministry of General Education Permanent Secretary send a circular to the PEO, the DEBS, and head teachers advising them to enforce teacher attendance and refer to human resources and public service guidelines

• Using results of the teacher study, revise, implement, and enforce teacher deployment regulations and explore other types of teachers who could teach more effectively

• Design and implement a teacher recruitment and deployment policy based on the findings of the pilot study on teacher motivation

• Revise the monitoring and evaluation strategy to include explicit measurements of teacher attendance

Teacher performance and training

• Review teacher performance evaluation systems (such as the Annual Performance Appraisal System) to see what is fit for purpose

• Develop a new way of assessing teachers (including head teachers) by developing teacher performance metrics

• Review and revise teacher recruitment procedures (at central and decentralized levels)

• Devise an effective teacher evaluation system for recruitment, performance evaluation, and feedback on all teachers and administrators at the school and district levels

• Continuously monitor teacher performance and provide feedback with a form of teacher training

Note: DEBS = District Education Board Secretaries; MESVTEE = Ministry of Education, Science, Vocational Training, and Early Education; PEO = Provincial Education Offices.

and now can conduct a large-scale survey—from designing to sampling to analyzing data.

Table 5.1 suggests short-term and medium- term strategies based on the policy recommendations.

Note 1. In principle, all primary students should receive a

primary school grant due to the free primary edu-cation policy.

Personalities and Motivations 71

Appendix A

Personalities and Motivations

FIGuRE A .1 Motivations of Grade 5 Teachers

f. Allows me to acquireuseful skills

0

0.5

1.0

2.0

1.5

2 4 6 8

0

0.5

1.0

2.0

1.5

2.5

f. Interesting job

2 4 6 8

e. Offers stable income

0

0.2

0.4

0.6

0.8

0.10

2 4 6 8

d. Pays well

0

0.5

1.0

1.5

2 4 6 8

0

0.5

1.0

2.0

1.5

2.5

2 4 6 8

c. Allows me to servethe community

0

0.5

1.0

2.0

1.5

2.5

0 2 4 6 8

b. Good future career

0

0.2

0.4

0.8

0.6

0.10

2 4 6 8

a. Respect and high status inthe community

Note: Ratings are on a scale of 1 (the factor is important) to 7 (the factor is not important).

72 Personalities and Motivations

FIGuRE A .3 Motivations of Grade 5 Students: Inventory of School Motivation Score

0

0.5

1.5

1.0

2.0

1 2 3 4 8

a. Task

0

0.5

1.5

1.0

2.0

1 2 3 4 8

b. Effort

0

0.5

1.5

1.0

2.0

1 2 3 4 5

c. Competition

0

0.5

1.5

1.0

1 2 3 4 5

d. Social power

figure continues next page

FIGuRE A .2 Big 5 Personality Traits of Grade 5 Teachers

–1–2 0 1 2

e. Openness

0

1.0

0.5

1.5

0

2

1

3

–1–2 0 1 2

b. Agreeableness c. Conscientiousness

0

2

1

3

–1–2 0 1 2–1–2 0 1 2

a. Extraversion

0

1.0

0.5

1.5

d. Neuroticism

0

0.4

0.2

0.6

0.8

1.0

–1–2 0 1 2

Note: Ratings are on a scale of –2 (the personality trait is not at all likely) to 2 (the personality trait is very much likely).

Personalities and Motivations 73

0

0.5

1.5

1.0

2.0

2.5

1 2 3 4 5

e. Affiliation f. Social concern

0

0.5

1.5

1.0

2.0

1 2 3 4 5

g. Praise

0

1

2

3

1 2 3 4 5

h. Token

0

0.5

1.0

1.5

1 2 3 4 5

Figure A .3 continued

Note: Ratings on panels a and b are on a scale of 1 (the factor is not at all likely) to 8 (the factor is very much likely); ratings on panels c through h are on a scale of 1 (the factor is not at all likely) to 5 (the factor is very much likely).

FIGuRE A .4 Big 5 Personality Traits of Grade 5 Students

–1–2 0 1 20

1.0

0.5

1.5

2.0

2.5

b. Agreeableness

–1–2 0 1 20

1.0

0.5

1.5

2.0

2.5

c. Conscientiousness

–1–2 0 1 2

a. Extraversion

0

1.0

0.5

1.5

2.0

–1–2 0 1 20

1.0

0.5

1.5

2.0

2.5e. Openness

–1–2 0 1 20

1.0

0.5

1.5

2.0

2.5

d. Neuroticism

Note: Ratings are on a scale of –2 (the personality trait is not at all likely) to 2 (the personality trait is very much likely).

74 Personalities and Motivations

TABLE A .1 Teacher Transfer and Attendance and Physical School EnvironmentPhysical school environment

Teacher: Want to transfer

Teacher: Requested to transfer

Teacher: Absent last week

Teacher: Number of days absent

School with electricity 0.039 −0.000 −0.069 0.076

(0.057) (0.032) (0.043) (1.230)

Number of staff houses 0.008 0.006 −0.005 0.024

(0.006) (0.003) (0.005) (0.137)

Number of staff rooms 0.039 0.017 0.044 0.040

(0.038) (0.022) (0.030) (0.861)

School with library −0.071 −0.005 0.050 −0.822

(0.068) (0.039) (0.054) (1.532)

Number of latrines per pupil 0.055* 0.024 0.000 −0.184

(0.029) (0.017) (0.024) (0.515)

In hardship position 0.056 0.034 0.018 1.614

(0.061) (0.036) (0.048) (1.360)

School revenue −0.000 0.000 −0.000 −0.000

(0.000) (0.000) (0.000) (0.000)

Note: The table shows the results from multiple regressions with covariates of each school characteristic and school location and remoteness because of the high correlation within school infrastructure characteristics. For example, the likelihood of teachers wanting to transfer was regressed on indicator variables showing the availability of electricity, school location (rural), and remoteness (distance to the nearest health facility). Standard errors are in parentheses.Significance level = * 10% ** 5% *** 1%.

TABLE A .2 Teacher Transfer and Attendance and Teaching Intensity

Indicator of intensity Teacher: Want to transfer

Teacher: Requested to transfer

Teacher: Absent last week

Teacher: Number of days absent

Multigrade teaching 0.050 0.071** 0.042 −1.970

(0.057) (0.032) (0.043) (1.226)

Daily teaching hours 0.061*** 0.002 0.001 −0.010

(0.022) (0.013) (0.017) (0.484)

Number of school shifts 0.017 0.045* −0.009 −0.900

(0.042) (0.023) (0.032) (0.913)

Number of students teaching −0.001 0.001* −0.002** −0.018

(0.001) (0.001) (0.001) (0.025)

Note: The table shows the results from multiple regressions with covariates of each variable showing teaching intensity and school location and remoteness. Standard errors are in parentheses.Significance level = * 10% ** 5% *** 1%.

TABLE A .3 Head Teachers’ and Other Teachers’ Motivation QuestionnaireSubject Description

Career advancement

Future: Rank Aims to be a higher-rank educational professional in 5–10 years.When you envision yourself in 5–10 years’ time, what do you see yourself doing?Indicator variable: 1 if answered either district education board secretary, provincial education officer, Ministry of Education, Science, Vocational Training, and Early Education (MESVTEE) director (of any department), MESVTEE permanent secretary, or not in the education field

Future: Community Aims to remain in the same community.When you envision yourself in 5–10 years’ time, where do you most see yourself?Indicator variable: 1 if answered remaining in the same community where I am now.

table continues next page

Personalities and Motivations 75

TABLE A .3 continuedSubject Description

Motivation at work

Career orientation Composite variable of three questions:1. I expect to be in a higher-level job in five years.2. I view my job as a stepping stone to other jobs.3. I expect to be doing the same work as a teacher in five years (reverse).

Calling Composite variable of three questions:1. Head teachers make the world a better place.2. I enjoy talking about teaching to others.3. I would choose to apply for this position again if I had the opportunity.

Serving the community

Inclusion of the “other” in the self Applicants are asked to choose between sets of pictures, each showing two circles (labeled “self” and “community”) with varying degrees of overlap, from nonoverlapping to almost completely overlapping. This variable equals 1 if the respondent chooses the almost completely overlapping picture, 0 otherwise.

Desire for positive pro-social impact Composite variable of three questions:1. I do my best when I’m working on a task that contributes to the well-being of others.2. I want to have positive impact on others through my work.3. One of my objectives at work is to make a positive difference in other people’s lives.

TABLE A .4 Head Teachers’ and Other Teachers’ Motivations and Teacher Transfer and Attendance

Indicator Teacher: Want to transfer

Teacher: Requested to transfer

Teacher: Absent last week

Teacher: Number of days absent

Head teacher career advance: higher rank in the future

0.119* 0.099** −0.065 −2.062

(0.070) (0.038) (0.054) (1.330)

Head teacher career advance: community in the future

−0.095 −0.107*** 0.071 1.301

(0.058) (0.032) (0.045) (1.104)

Head teacher motivation: career orientation

0.030 0.022 0.002 −0.757

(0.027) (0.015) (0.021) (0.526)

Head teacher motivation: calling

0.022 −0.009 −0.077** 0.430

(0.040) (0.022) (0.031) (0.765)

Head teacher: pro-social impact

0.046 0.010 −0.068* −2.901***

(0.050) (0.028) (0.040) (0.976)

Head teacher: inclusion of others in self

−0.141** 0.056 −0.015 −0.614

(0.062) (0.036) (0.050) (1.213)

Teacher career advance: higher rank in the future

0.282*** 0.060 −0.135** −1.375

(0.072) (0.043) (0.055) (1.391)

Teacher career advance: community in the future

−0.271*** −0.052 −0.008 −0.281

(0.063) (0.037) (0.048) (1.193)

Teacher motivation: career orientation

0.047 0.034** 0.013 −0.326

(0.029) (0.017) (0.022) (0.556)

Teacher motivation: calling

0.022 −0.009 −0.077** 0.430

(0.040) (0.022) (0.031) (0.765)

Teacher: pro-social impact

0.046 0.010 −0.068* −2.901***

(0.050) (0.028) (0.040) (0.976)

Teacher: inclusion of others in self

−0.112** −0.082** −0.012 −1.401

(0.056) (0.033) (0.042) (1.035)

Note: The table shows the results from multiple regressions with covariates of each motivation variable and with school location and remoteness controlled. Standard errors are in parentheses.Significance level = * 10% ** 5% *** 1%.

76 Personalities and Motivations

TABLE A .5 Relationship between School-Level Input and Type of Teacher Contract and Student Learning

Indicator Math English

School-level infrastructure

Schools with library 1.864** 3.669***

(0.934) (1.115)

Schools with laboratory 1.094 −0.725

(1.280) (1.482)

Schools with electricity 0.724 −1.523

(0.857) (1.002)

Number of latrines per students −0.011 −0.087

(0.297) (0.367)

School-level financial input

Per student school revenue −0.000 0.000

(0.000) (0.000)

Per student school grant 0.006*** 0.011***

(0.002) (0.003)

School-level soft input

Pupil-teacher ratio −0.040*** −0.033**

(0.012) (0.014)

Schooling hours per shift 0.714** 0.819**

(0.308) (0.384)

Number of school shifts −1.981** −1.924**

(0.802) (0.951)

Teacher contract +

Contract teacher (reference: government teacher)+ 8.097*** 5.456***

(1.316) (1.511)

Contract teacher (after controlling subject knowledge)+ 8.213*** 11.826***

(1.438) (2.183)

Note: The table shows the results from multiple regressions with covariates of each input variable presented in the table and with school location and remoteness controlled. In addition, regressions in teacher contract (+) control for major school-level inputs such as number of libraries, schooling hours per shift, pupil-teacher ratio, school type (government, grant-aided, private, and others), and grant amount received. Contract teacher includes paid contract teacher and unpaid volunteer. Standard errors are in parentheses.Significance = * 10% ** 5% *** 1%.

TABLE A .6 Relationship between Student-Level Inputs and Student Learning Indicator Math English

Have a textbook (student level) 3.193** 1.829

(1.267) (1.290)

Have a textbook (student level), after controlling for wealth of students 2.962** 1.438

(1.273) (1.295)

Days missed school last week −1.130*** −1.522***

(0.305) (0.384)

Given homework −0.810 −0.393

(0.590) (0.732)

Frequency of questions asked by teacher 1.604** 1.583*

(0.717) (0.909)

Note: The table shows the results from multiple regressions with covariates of each input variable presented in the table and with school location and remoteness controlled. Standard errors are in parentheses.Significance = * 10% ** 5% *** 1%.

Personalities and Motivations

77

TABLE A .7 Student Learning Outcomes in Relation to with Teacher Subject KnowledgeMath English Zambian language

Indicator Score (mean) Score (25th) Score (75th) Score (mean) Score (25th) Score (75th) Score (mean) Score (25th) Score (75th)

Teacher subject knowledge 0.046 0.077** 0.063 0.228*** 0.254*** 0.271*** 0.297*** 0.128*** 0.400***

(0.052) (0.035) (0.084) (0.067) (0.040) (0.076) (0.060) (0.035) (0.068)

Female student 0.104 0.171 0.403 1.359 −0.282 0.833 −0.035 0.852 −0.000

(0.827) (0.712) (1.339) (1.321) (1.047) (1.798) (0.837) (0.811) (1.431)

Female teacher 4.797*** 0.684 5.922** 3.801** 0.788 3.869 2.691** 1.128 0.667

(1.641) (1.094) (2.402) (1.678) (1.067) (2.495) (1.301) (1.329) (2.013)

Teaching experience (year) −0.100 −0.085 −0.062 −0.083 −0.196** −0.179** −0.006 −0.025 0.000

(0.098) (0.069) (0.109) (0.113) (0.092) (0.089) (0.090) (0.087) (0.151)

Rural −2.424 −3.162** −2.067 −6.933*** −3.302*** −12.381*** −0.713 1.053 −4.667*

(1.647) (1.229) (2.261) (1.825) (1.149) (3.787) (1.503) (1.409) (2.815)

Government school −1.623 −1.709 −2.005 −0.554 −0.110 0.298 −0.444 −2.030* 1.333

(1.746) (1.125) (2.410) (1.989) (1.301) (2.629) (1.446) (1.189) (2.195)

Constant 33.191*** 24.103*** 38.770*** 13.845** 1.033 22.024*** 9.326* 13.459*** 12.000**

(4.412) (3.391) (7.292) (6.176) (3.650) (7.722) (5.203) (3.165) (6.114)

Sample size 1543 1543 1543 1111 1111 1111 1522 1522 1522

R-squared 0.051 0.033 0.051 0.068 0.043 0.065 0.038 0.022 0.031

Note: Regression results at the mean, quantile regression results at 25th percentile, and quantile regression results at 75th percentile are shown. Standard errors are in parentheses.Significance = * 10% ** 5% *** 1%.

78 Personalities and Motivations

TABLE A .8 Student Learning Outcomes in Relation to Student Personalities and Motivations

Math English

Indicator Score (mean) Score (25th) Score (75th) Score (mean) Score (25th) Score (75th)

Extraversion (student) −0.718* −0.352 −0.532 −1.393** −0.045 −2.409***

(0.390) (0.459) (0.751) (0.587) (0.507) (0.890)

Agreeable (student) 1.222*** 1.132** 1.399* 1.882*** 0.459 2.660***

(0.465) (0.499) (0.715) (0.525) (0.540) (0.887)

Inventory of School Motivation (social power and recognition) 0.941** 0.772* 0.488 0.535 0.157 −0.345

(0.382) (0.400) (0.745) (0.644) (0.513) (1.076)

Teacher subject knowledge 0.048 0.030 0.078 0.199*** 0.256*** 0.140

(0.042) (0.069) (0.076) (0.076) (0.057) (0.263)

Female student 0.427 0.052 1.396 2.075 −0.621 2.969

(0.963) (1.123) (1.690) (1.628) (1.335) (2.352)

Female teacher 4.453*** 1.114 5.456** 4.889** −0.089 6.390**

(1.621) (1.627) (2.551) (1.961) (1.308) (3.194)

Teaching experience (year) −0.092 −0.041 −0.153 −0.032 −0.033 −0.129

(0.092) (0.081) (0.151) (0.134) (0.096) (0.156)

Rural −2.530 −2.315 −2.506 −6.121*** −1.708 −13.797***

(1.602) (1.600) (2.361) (2.073) (1.436) (3.697)

Government school −2.179 −1.714 −2.748 −0.692 −0.213 −0.860

(1.786) (1.578) (2.992) (2.464) (1.697) (3.719)

Constant 31.230*** 25.316*** 37.931*** 14.429* −0.136 37.224

(4.043) (6.880) (6.821) (7.784) (5.262) (25.743)

Sample size 1,006 1,006 1,006 743 743 743

Note: Regression results at the mean, quantile regression results at 25th percentile, and quantile regression results at 75th percentile are shown. Standard errors are in parentheses.Significance = * 10% ** 5% *** 1%.

TABLE A .9 Student Learning Outcomes in Relation to Teacher Personalities and Motivations

Math English

Indicator Score (mean) Score (25th) Score (75th) Score (mean) Score (25th) Score (75th)

Teacher motivation (respect from community) 0.536 0.217 0.880** 0.285 0.241 −0.280

(0.333) (0.269) (0.433) (0.551) (0.256) (1.004)

Neuroticism (teacher) 1.626* 0.378 2.265** 2.596** 2.102*** 3.183*

(0.891) (0.724) (0.955) (1.098) (0.782) (1.673)

Teacher subject knowledge 0.065 0.095** 0.084 0.196* 0.100 0.191

(0.078) (0.045) (0.066) (0.102) (0.096) (0.285)

Female student 0.179 −0.000 −0.000 2.950* 1.003 3.535

(1.167) (1.044) (1.585) (1.611) (1.407) (2.654)

Female teacher 6.457*** 2.121 8.470*** 6.555*** 2.082 8.529**

(1.745) (1.430) (2.270) (1.976) (1.463) (3.432)

table continues next page

Personalities and Motivations 79

TABLE A .9 continuedMath English

Indicator Score (mean) Score (25th) Score (75th) Score (mean) Score (25th) Score (75th)

Teaching experience (year) −0.127 −0.113 −0.003 0.109 −0.084 0.149

(0.147) (0.084) (0.289) (0.153) (0.129) (0.262)

Rural −1.763 −3.511** −1.440 −5.189** −2.595 −10.256**

(2.050) (1.508) (2.564) (2.478) (1.696) (4.171)

Government school −3.198 −3.601** −5.281* −0.235 −0.703 0.684

(2.349) (1.612) (2.768) (2.038) (1.554) (3.178)

Constant 30.030*** 22.544*** 34.940*** 12.745 15.059* 26.018

(6.097) (4.244) (6.588) (9.554) (8.561) (27.495)

Sample size 961 961 961 727 727 727

Note: Regression results at the mean, quantile regression results at 25th percentile, and quantile regression results at 75th percentile are shown. Standard errors are in parentheses.Significance = * 10% ** 5% level *** 1%.

Literature Review on School Inputs and Learning Outcome 81

Appendix B

Literature Review on School Inputs and Learning Outcome

Improvements in basic school inputs, such as funding, infrastructure, and resources, increase school enrollments effectively, but the quantity of education has not always translated into quality education (Glewwe and Kremer 2006). School-level inputs such as teacher and head teacher quality, and school management, however, are important contributors to student achievement. Both previous Public Expenditure Tracking Survey (PETS)–Quantitative Service Delivery Survey (QSDS) reports for Zambia have recom-mended greater policy focus in these areas, which, in some cases, are more significant than school materials or infrastructure (Behrman et al. 1997).

Attributes of high-quality teachers are diffi-cult to measure (Jepsen 2004; Luschei and Chudgar 2011). Yet many studies have shown that observable or résumé characteristics—such as qualifications, experience, and pay—are weak predictors of student achievement (Aslam and Kingdon 2011; Behrman et al. 1997; Hanushek and Rivkin 2006; Hein and Allen 2013; Kingdon 2006; Rawal, Aslam, and Jamil 2013; Winters, Dixon, and Greene 2012). Instead, less observed characteristics, such as pedagogical style, subject knowledge, attitude, and behavior, contribute more to a teacher’s value added (Aslam and Kingdon 2011; Hein and Allen 2013; Lavy 2011; Lopez-Acevedo 2002; Marshall and Sorto 2012; Opdenakker and Van Damme 2006). Efforts to target teacher quality through incentives such as pay for performance have been successful at increasing test scores, but not as much at improv-ing teacher absence (Chaudhury et al. 2006; Glewwe, Ilias, and Kremer 2010; Muralidharan and Sundararaman 2011). Rogers and Vegas (2009) suggest that incentives for improved teacher quality should focus on accountability and intrinsic rewards as well as performance.

Effectiveness at the head teacher level has more indirect but significant effects on student learning. Through their experience, leadership style, and relationship with the community, head teachers can create a positive working environ-ment for teachers with a learner-centered focus (Anderson 2008; Branch et al. 2013; Clark, Martorell, and Rockoff 2009; Lassibille 2012; Mattar 2012; Oplatka and Gurion 2004). Management within a school that allows for proper support, supervision, control, and auton-omy relates to the quality of teachers and head teachers (Glewwe et al. 2011; Vegas 2002). In developing countries especially, decentralization at this level would be beneficial to counteract challenges of low autonomy for head teachers and autocratic leadership styles (Kitavi and Van  Der Westhuizen 1997; Oplatka and Gurion 2004). Decentralization has been shown to improve accountability, management, and academic outcomes, yet is not always pro-poor and is most effective when initiated by the community (Bruns, Filmer, and Patrinos 2011; CERID 2009; Faquet and Sánchez 2008; Galiani, Gertler, and Schargrodsky 2008).

Unobservable characteristics, such as per-sonality and motivation, have been shown to have significant effects on student success and teacher and head teacher quality. Self, or intrin-sic, motivation for students has a positive effect on their performance (Gamboa, Rodriguez, and Garcia 2013). The level of students’ aca-demic motivation and goals can predict their academic achievement and improve quality of learning (Ali and McInerney 2005; Amrai et al. 2011; Covington 2000; McInerney 2008). Students’ personality traits can affect their learning styles and outcomes. Understanding these effects can be important pedagogical tools

82 Literature Review on School Inputs and Learning Outcome

for teachers in tailoring lessons to the needs at the individual student level (Chowdhury 2006). Teachers are also key moderators of student motivation and can use student incentives, sup-port mechanisms, and goals to target motiva-tion (McInerney 2008).

Success of teachers can be attributed to their own strength of motivation, affecting their effort and the level of classroom support they give to stu-dents (Adrabi et al. 2009; Opdenakker and Van Damme 2006). Long-term teacher motivation is determined by their goal orientation and experi-ence of a calling, or meaningful passion, to teach (Dobrow and Tosti-Kharas 2011; Malmberg 2006). Public service motivation and certain personality traits—such as conscientiousness—can predict the attendance, integrity, and leadership effectiveness of teachers and head teachers (Callen et al. 2013; Rushton, Morgan, and Richard 2007).

Targeting these unobservable characteristics for teachers and head teachers through strategic recruitment can increase the number of qualified applicants and improve retention and perfor-mance (Ashraf, Bandiera, and Lee 2013; Callen et al. 2013). For current teachers and head teach-ers, improving working conditions where moti-vation can be the lowest, such as in rural areas, and enhancing school systems with management and accountability reforms can increase motiva-tion (Bennell and Akyeampong 2007; Oplatka and Gurion 2004).

ReferencesAdrabi et al. 2009. Andrabi, Tahir, Jishnu Das, Asim

Ijaz Khwaja, Tara Vishwanath, and Tristan Zajonc. 2009. Learning and Educational Achievements in Punjab Schools (LEAPS) Report. Washington, DC: World Bank.

Ali, Jinnat, and Dennis M. McInerney. 2005. “An Analysis of the Predictive Validity of the Inventory of School Motivation.” SELF Research Centre, University of Western Sydney, Australia.

Amrai, Kourosh, Shahrzad Elahi Motlagh, Hamzeh Azizi Zalani, and Hadi Parhon. 2011. “The Relationship between Academic Motivation and Academic Achievement Students.” Procedia 15 (February 7): 399–402.

Anderson, Joan B. 2008. “Principals’ Role and Public Primary Schools’ Effectiveness in Four Latin American Cities.” Elementary School Journal 109 (1): 36–60.

Ashraf, Nava, Oriana Bandiera, and Scott Lee. 2013. “Do-Gooders and Go-Getters: Career Incentives, Selection, and Performance in Public Service Delivery.” Working Paper, Harvard Business School, Cambridge, MA.

Aslam, Monazza, and Geeta G. Kingdon. 2011. “What Can Teachers Do to Raise Pupil Achievement?” Economics of Education Review 30 (3): 559–74.

Behrman, Jere R., Shahrukh Khan, David Ross, and  Richard Sabot. 1997. “School Quality and Cognitive Achievement Production: A Case Study for Rural Pakistan.” Economics of Education Review 16 (2): 127–42.

Bennell, Paul, and Kwame Akyeampong. 2007. “Teacher Motivation in Sub-Saharan Africa and  South Asia.” Educational Paper 71, DFID, London.

Branch, Gregory F., Eric A.  Hanushek, and Steven G. Rivkin. 2013. “School Leaders Matter.” Education Next 13 (1): 62–69.

Bruns, Barabara, Deon Filmer, and Harry A. Patrinos. 2011. Making Schools Work: New Evidence on Accountability Reforms. Washington, DC: World Bank.

Callen, Michael, Saad Gulzar, Ali Hasanain, Yasir Khan, and Arman Rezaee. 2013. “Personalities and  Public Sector Performance: Experimental Evidence from Pakistan.” NBER Working Paper 21180, National Bureau of Economic Research, Cambridge, MA.

CERID (Research Centre for Educational Innovation and Development). 2009. “Community-Managed School: An Innovative Approach to School Management.” Formative Research Project 33, CERID, Kathmandu, Nepal.

Chaudhury, Nazmul, Jeffrey Hammer, Michael Kremer, Karthik Muralidharan, and F. Halsey Rogers. 2006. “Missing in Action: Teacher and Health Worker Absence in Developing Countries.” Journal of Economic Perspectives 20 (1): 91–116.

Chowdhury, Mohammed. 2006. “Students’ Personality Traits and Academic Performance.” College Quarterly 9 (3).

Clark, Damon, Paco Martorell, and Jonah E. Rockoff. 2009. “School Principals and School Performance.” Working Paper 38, National Center for Analysis of Longitudinal Data in Education Research, Washington, DC.

Covington, Martin V. 2000. “Goal Theory, Motivation, and School Achievement: An Integrative Review.”

Literature Review on School Inputs and Learning Outcome 83

Annual Review of Psychology 51 (February): 171–200.

Dobrow, Shoshana R., and Jennifer Tosti-Kharas. 2011. “Calling: The Development of a Scale Measure.” Personnel Psychology 64 (4): 1001–49.

Faquet, Jean-Paul, and Fabio Sánchez. 2008. “Decentralization’s Effects on Educational Outcomes in Bolivia and Colombia.” World Development 36 (7): 1294–316.

Galiani, Sebastian, Paul Gertler, and Ernesto Schargrodsky. 2008. “School Decentralization: Helping the Good Get Better, But Leaving the Poor Behind.” Journal of Public Economics 92 (10-11): 2106–20.

Gamboa, Luis, Mauricio Rodriguez, and Andres Garcia. 2013. “Differences in Motivations and Academic Achievement.” Lecturas de Economia 78: 9–44.

Glewwe, Paul W., Eric A. Hanushek, Sarah D. Humpage, and Renato Ravina. 2011. “School Resources and Educational Outcomes in Developing Countries: A Review of Literature from 1990 to 2010.” Working Paper Series 17554, National Bureau of Economic Research, Cambridge, MA.

Glewwe, Paul, Nauman Ilias, and Michael Kremer. 2010. “Teacher Incentives.” American Economic Journal: Applied Economics 2 (3): 205–27.

Glewwe, Paul, and Michael Kremer. 2006. “Schools, Teachers, and Education Outcomes in Developing Countries.” In Handbook of the Economics of Education, vol. 2, edited by Eric Hanushek and Finis Welch, 945–1017. Amsterdam: North-Holland.

Hanushek, Eric A., and Steven G. Rivkin. 2006. “Teacher Quality.” In Handbook of the Economics of Education, vol. 2, edited by Eric Hanushek and Finis Welch, 1051–78. Amsterdam: North-Holland.

Hein, Christopher F., and Rebecca Allen. 2013. “Teacher Quality in Sub-Saharan Africa: Pupil Fixed Effects Estimates for Twelve Countries.” Working Paper 12-08, Department of Quantitative Social Science, Institute of Education, London.

Jepsen, Christopher. 2004. “Teacher Characteristics and Student Achievement: Evidence from Teacher Surveys.” Journal of Urban Economics 57 (2005): 302–19.

Kingdon, Geeta G. 2006. “Teaching Characteristics and Student Performance in India: A Pupil Fixed-Effects Approach.” Working Paper 059, Global Poverty Research Group, Oxford.

Kitavi, Mwaya Wa, and Philip C. Van Der Westhuizen. 1997. “Problems Facing Beginning Principals in Developing Countries: A Study of Beginning Principals in Kenya.” International Journal of Educational Development 17 (3): 251–63.

Lassibille, Gerard. 2012. “Teachers’ Engagement at Work in a Developing Country.” Journal of African Economies 22 (1): 52–72.

Lavy, Victor. 2011. “What Makes an Effective Teacher? Quasi-Experimental Evidence.” Working Paper 16885, National Bureau of Economic Research, Cambridge, MA.

Lopez-Acevedo, Gladys. 2002. “Teachers’ Incentives and Professional Development in Schools in Mexico.” World Bank, Washington, DC.

Luschei, Thomas F., and Amita Chudgar. 2011. “Teachers, Student Achievement, and National Income: A Cross-National Examination of Relationships and Interactions.” Prospects 41 (4): 507–33.

Malmberg, Lars-Erik. 2006. “Goal-Orientation and Teacher Motivation among Teacher Applicants and Student Teachers.” Teaching and Teacher Education 22 (1): 58–76.

Marshall, Jeffrey H., and M. Alejandro Sorto. 2012. “The Effects of Teacher Mathematics Knowledge and Pedagogy on Student Achievement in Rural Guatemala.” International Review of Education 58 (2): 173–97.

Mattar, Dorine M. 2012. “Factors Affecting the Performance of Public Schools in Lebanon.” International Journal of Educational Development 32 (2012): 252–63.

McInerney, Dennis M. 2008. “Personal Investment, Culture, and Learning: Insights into School Achievement across Anglo, Aboriginal, Asian, and Lebanese Students in Australia.” International Journal of Psychology 43 (5): 870–79.

Muralidharan, Karthik, and Venkatesh Sundararaman. 2011. “Teacher Performance Pay: Experimental Evidence from India.” Journal of Political Economy 119 (1): 39–77.

Opdenakker, Marie-Christine, and Jan Van Damme. 2006. “Teacher Characteristics and Teaching Styles as Effectiveness-Enhancing Factors of Classroom Practice.” Teaching and Teacher Education 22 (2006): 1–21.

Oplatka, Izhar, and Ben Gurion. 2004. “The Principalship in Developing Countries: Context, Characteristics, and Reality.” Comparative Education 40 (3): 427–48.

Rawal, Shenila, Monazza Aslam, and Baela Jamil. 2013. “Teacher Characteristics, Actions, and Perceptions: What Matters for Student Achievement in Pakistan?” Working Paper 2013-10, Centre for the Study of African Economies, Oxford.

Rogers, F. Halsey, and Emiliana Vegas. 2009. “No More Cutting Class? Reducing Teacher Absence and Providing Incentives for Performance.” Policy Research Working Paper 4847, World Bank, Washington, DC.

84 Literature Review on School Inputs and Learning Outcome

Rushton, Stephen, Jackson Morgan, and Michael Richard. 2007. “Teacher’s Myers-Briggs Personality Profiles: Identifying Effective Teacher Personality Traits.” Teaching and Teacher Education 23 (4): 432–41.

Vegas, Emilana. 2002. “School Choice, Student Performance, and Teacher and School Characteristics:

The Chilean Case.” Policy Research Working Paper 2777, World Bank, Washington, DC.

Winters, Marcus A., Bruce L. Dixon, and Jay P. Greene. 2012. “Observed Characteristics and Teacher Quality: Impacts of Sample Selection on a Value-Added Model.” Economics of Education Review 31 (2012): 19–32.

SKU K8641

Education Sector Public Expenditure Tracking and Service Delivery Survey in Zambia