43
SOUTH AFRICA: POVERTY AND INEQUALITY IN FORMAL DISCUSSION PAPEPR SERIES 19330 February1999 Poverty and Inequality in the Distribution of Public Education Spending in South Africa Florencia Castro-Leal WORLD BANK COUNTRY DEPARTMENT I AFRICA REGION Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized

Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

SOUTH AFRICA: POVERTY AND INEQUALITYIN FORMAL DISCUSSION PAPEPR SERIES

19330February 1999

Poverty and Inequality in theDistribution of Public Education

Spending in South Africa

Florencia Castro-Leal

WORLD BANKCOUNTRY DEPARTMENT IAFRICA REGION

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

Page 2: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

SOUTH AFRICA: POVERTY AND INEQUALITYIN FORMAL DISCUSSION PAPER SERIES

* Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

Florencia Castro-Leal, February 1999

* The Impact of Public Health Spending on Povertyand Inequality in South Africa

Florencia Castro-Leal, February 1999

* Review of Public Expenditures: Efficiency andPoverty Focus

Gurushri Swamy and Richard Ketley, February 1999

* Safety Nets and Income Transfers in South Africa

Harold Alderman, February 1999

* Poverty Issues for Zero Rating Value-Added Tax (VAT)in South Africa

Harold Alderman and Carlo del Ninno, February 1999

* Violence and Poverty in South Africa: Their Impacton Household Relations and Social Capital

Caroline Moser, February 1999

* Women Workers in South Africa: Participation, Pay andPrejudice in the Formal Labor Market

Carolyn Winter, February 1999

Page 3: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

FOREWORD

This paper is one of a series of informal discussion papers on poverty andinequality issues in South Africa, which were produced as contributions to the Povertyand Inequality Report (PIR). The PIR was commissioned by the Deputy President'sOffice of the Government of the Republic of South Africa (and was published in 1998 byPraxis Publishing, South Africa). As these papers were written at different times overthe years 1996-1998, the analysis in each paper covers different periods; however, forease of reference, they are now being disseminated in one series.

A complementary report, which gathers the views of the poor themselves, waswritten by a team of South Africans and also published by Praxis Publishing. "TheExperience and Perceptions of Poverty in South Africa" (1998) gives voice to the poor,who describe what poverty is to them, how they get trapped in it, and how they mightescape from it. This study was initiated and funded by the World Bank (through a DutchTrust Fund) and by the Overseas Development Administration of the U.K. Government.

The papers in this series were written under the direction of Ann Duncan (TaskManager) and under the overall guidance of Pamela Cox (Country Director) and Ruth Kagia(Sector Manager). The series was edited by Barbara Koeppel, and the final presentation wasmanaged and executed by Lori Geurts.

Country Department IThe World BankFebruary 1999

Copyright 1999The World Bank1818 "H' Street, N.W.Washington, DC 20433, U.S.A.

This is an informal study by World Bank staff publishedfor discussion purposes. It is not an official World Bankdocument.

Page 4: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

ACKNOWLEDGEMENTS

This paper was written by Florencia Castro-Leal, Economist, PRMPO. Theauthor gratefully acknowledges the support received from Ann Duncan in producing thispaper. The author also wishes to thank Stephan Klasen, Janet Leno, Gurushri Swamyand Carolyn Winter for helpful comments on an earlier draft. Yisgedullish Amde andKalpana Mehra provided excellent computational analysis of household survey data.Many thanks to Precy Lizarondo for providing quick and efficient word processingassistance.

The views expressed in this study are those of the author and should not beattributed to the World Bank, members of its Board of Executive Directors, or thecountries they represent.

Page 5: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

ABSTRACT

From 1987 to 1994, South Africa's public education resources increased from5.8% to 7.3% of GDP. The annual growth rate of education spending in real termsshowed an absolute per capita expansion. This paper explores the distribution of theseresources across socioeconomic and demographic groups, using Benefit IncidenceAnalysis, which measures how well public services have been targeted.

The analysis revealed spending disparities across income groups, regions andraces: the shares of public education resources benefiting the poor and ultra-poor aresubstantially lower than their shares of school-age population. Inequality in thedistribution widens by educational level from primary to tertiary for all populationgroups.

The analysis of household direct expenses suggests that if the costs of uniforms,transport and meals were reduced for poor primary and secondary school-age children,and scholarships were expanded for tertiary education in order to exempt the poor fromschool fees, enrollment among the poor could rise. These programs could beaccompanied by cost recovery mechanisms that do not discriminate against the poor, andwould free up public education funds to finance the additional expenses. Regionally,important gains could be achieved in allocating public education resources to the poorand the ultra-poor if the Eastem Cape, KwaZulu Natal and Northern Province weretargeted.

Page 6: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

ABBREVIATIONS AND ACRONYMS

DBSA Development Bank of Southem AfricaDET Department of Education and TrainingGDP Gross Domestic ProductH-Africans Africans living in the former HomelandsNH-Africans Africans not living in the former HomelandsSALDRU Southern Africa Labour & Development Research UnitSGT Self-Governing TerritoriesTBVC Transkei, Bophuthatswana, Venda, Ciskei

CURRENCY EQUIVALENTS

Rand/US$

1990 2.6

1991 2.7

1992 3.1

1993 3.4

1994 3.5

1995 3.6

1996 4.7

1997 4.9

1998 5.9

Page 7: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

TABLE OF CONTENTS

EXECUTIVE SUMMARY ............................................................ I

I. INTRODUCTIONR.

II. SOCIOECONOMIC AND DEMOGRAPHIC GROUPS ................... 2....................2

Ill. SCHOOL ENROLLMENT ................... .......................................... 3

IV. PUBLIC SPENDING ON EDUCATION ............................................................ 10

V. THE DISTRIBUTION OF PUBLIC EDUCATION SPENDING ............ ............. 14

VI. HOUSEHOLD SPENDING ON EDUCATION ...................................... ............. 21

VII. CROSS-COUNTRY COMPARISONS ............................................................. 23

VIII. POLICY IMPLICATIONS ............................................................. 26

REFERENCES ............................................................ 28

APPENDIX A ............................................................ 29

Page 8: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency
Page 9: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty andInequality in the Distribution of Public Education Spending in South Africa

EXECUTIVE SUMMARY

South Africa's public education spending has increased substantially in recentyears. The expansion outpaced population growth in the early 1990s and the allocationof state expenditures to public education grew by more than the rate of growth in theGross Domestic Product (GDP). However, the education system still exhibits strongdisparities across income groups, regions and races.

Indeed, the shares of public education resources benefiting the ultra-poor and thepoor are substantially lower than their shares of school-age population (see Table A). In1993, poor households received 40% of public education spending for 60% of theschool-age population, while the share for the richest was 23%, who accounted for only8% of school-age children. Ultra-poor households received 21%, while accounting for34% of the school-age population.

Table AIncidence of Public Education Spending by Household Group, 1993Percentage Share

Educational level Education spending benefiting:(School-age population) Ultra-poor Poor Richest

(poorest (poorest (richest 20%)20%) 400%)

% share of education spending 27 48 19benefiting Primary school children(% of population in age group 6 to 12 (36) (62) (8)years)

% share of education spending 18 36 25benefiting Secondary school children(% of population in age group 13 to 17 (35) (61) (9)years)

% share of education spending 11 24 32benefiting Tertiary school children(% of population in age group 18 to 22 (29) (54) (9)years)

% share of all education spending 21 40 23benefiting all ages(% of population in age group 6 to 22 (34) (60) (8)years) I_I

Source: DBSA (1993) and SALDRU (1993)

i

Page 10: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

The unequal distribution of public spending widens from the primary to tertiarylevels, across different income groups and races: Per capita spending is lowest forAfricans at every educational level (see Figure A). However, not all African students arepoor or ultra-poor. Thus, public education resources must be targeted to the EasternCape and Northern Provinces, which have by far the highest poverty rates and hold morethan half the former homelands population.

Figure AAnnual Education Spending Per Capita, by Level of Schooling and Race, 1993

1000900T

g 800 | |Primary700

Ft 600 | *SecondaryIt,500 3 0 Tertiary22400 | |AllEducation

s.300 +AEuao200

0

.

Source: DBSA (1993) and SALDRU (1993)

Based on findings from a household survey, school expenses are the mostimportant reason that keep parents from sending children to school. Most families haveto spend a significant amount for school fees, uniforms, transport, meals and books andstationary: Poor families pay more than 40% of per capita non-food householdexpenditures per child in primary school compared to only 6% for non-poor families.Education costs increase to half of per capita non-food household expenditures per childin secondary school, in contrast to 10% for the non-poor. By the time students reach thetertiary level, the costs are prohibitive: Each poor student needs more than two-and-a-half times the per capita amount spent on non-food items compared to less than one-thirdfor non-poor students.

Thus, if the costs of uniforms, transport and meals for poor primary andsecondary school-age children were reduced and scholarships were provided for tertiaryeducation (which would exempt the poor from fees), enrollments among the poor wouldlikely rise. These programs could be accompanied by cost recovery mechanisms that donot discriminate against the poor--and would thereby free up public education funds tofinance the additional expense.

ii

Page 11: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

I. INTRODUCTION

From 19871 to 1994, South Africa's spending on public education increasedsignificantly, from 5.8% to 7.3% of GDP. Indeed, the annual growth rate in real terms

was more than 5%, against 2% for the population aged 6-18.2 In other words, real percapita public funding of education expanded in absolute terms. This paper explores thedistribution of the education resources across socioeconomic and demographic groups.

Two main factors influence inequities in the distribution of public educationresources: (a) the allocation of public resources to individuals in different socioeconomicgroups, regions, races and by gender as a result of their consumption of public services;and (b) the allocation of government spending within the education sector. Thedistribution can be examined by allocating per unit public subsidies according toindividual utilization rates of public services. The methodology, Benefit Incidence

Analysis,3 measures how well public services are targeted, for example, to the poor,geographic regions and girls.

Detailed information on spending was available from a comprehensive review ofeducation expenditures published by the Development Information Group of theDevelopment Bank of South Africa and the World Bank (DBSA, 1993). Data onindividual patterns of access and use of public services were obtained from the SouthAfrica Living Standards Survey, coordinated by the University of Cape Town's SouthernAfrica Labour and Development Research Unit (SALDRU, 1993) and funded by thegovernments of Denmark, the Netherlands and Norway, with technical assistance fromthe World Bank. DBSA and SALDRU data were used to calculate per capita educationsubsidies for different socioeconomic groups, races, regions and by gender.

Section II explains the disaggregation by demographic groups used throughoutthis paper. Section III looks at the education indicators. Section IV explores publiceducation spending and the allocation of resources within the sector. Section V analyzesthe distribution patterns of public education spending across socio-economic groups andgender. Section VI presents information on household education expenditures, andSection VII draws some cross-country comparisons. Section VIII presents the policyimplications based on this study.

1 Buckland, Peter and Jolm Fielden (1994). Public Expenditure on Education in South Africa, 1987/88 to1991/92, p. 62 Ibid., p. 8

3 See Van de Walle and Nead (1994), Meennan (1979), and Selowsky (1979).

1

Page 12: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty andlnequality in the Distribution ofPublic Education Spending in South Africa

II. SOCIOECONOMIC AND DEMOGRAPHIC GROUPS

When targeting public education resources, it is critical to address theconcentration of the poor in certain geographical areas, since a poverty profile found thatnearly 70% live in the former homelands.4 This disparity in income justified studyingthe population in the former homelands separately from the rest of the country. In theseareas, Africans constitute almost 100% of the population (see Appendix Table A. 1),account for 66% of the African population (see Appendix Table A.2), 63% of householdsin the poorest quintile, and 55% in the second poorest quintile (see Figure 1).5

Figure 1Racial Composition within Quintiles

100DAfricaW/ er-

80 Homeland

*JWhite.4 60

40Ulda

20 03 Coloured

0M Africaz/Non- ex-

Homeland

Source: SALDRU (1993)

Regional income disparities suggest any analysis of education spending mustfocus on three new provinces: the Eastern Cape, Northern Province and KwaZulu Natal,where more than 75% of residents from the former homelands (H-Africans) live (seeTable 1).

4 "Key Indicators of Poverty in South Africa", RDP (1995), p. 12

5 The poverty assessment found that about 40% of South African households are poor. Households areranked by per adult equivalent expenditures, the poorest 20% constitutes the bottom quintile, and so on forthe next quintile, up to the top quintile which contains the richest 20% of all households. Since householdsin bottom quintiles are larger than households in higher quintiles, the total population found in poverty islarger than the number of poor households.

2

Page 13: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

Table 1Percentage of Former Homelands Populationby New Province

New province Share %N.Cape 0.00Free State 2.34North West 11.07Mpumalanga 10.13W.Cape 0.00N.Province 22.88E.Cape 23.42Gauteng 0.00KwaZulu Natal 30.15Total 100.00

Source: SALDRU (1993)

With DBSA data, per student subsidies were calculated separately for all racesand for H-Africans and Africans outside the former homelands (NH-Africans). TheSALDRU household survey contained information on household residence by newprovince and its racial composition.

III. SCHOOL ENROLLMENT

Many older, poor children attend primary school

Net and gross enrollment rates for primary school were similar across incomequintiles and races (see Tables 2 and 3).6 Overall, the net enrollment rate was about 90%and the gross rate was 106%. Thus, nine out of every 10 children from 6-12 wereenrolled in primary school, and the spread between the net and gross enrollment ratesindicated that many older children were in primary schools.7 The spread is four timeslarger for the poorest than for the richest income quintile: The poorest quintile had aprimary net enrollment rate of 85% compared with a 112% gross enrollment rate whilethe rates for the richest quintile were 90% and 97%, respectively. The main problem forolder children attending primary school is that they often must help support the

6 The net enrollment rate is defined as the total number of school age children currently enrolled in eacheducational level as a percentage of the total school age population at each level. The gross enrollment rateis defined as the total number of enrollments in each educational level, of any age, as a percentage of thetotal school age population at each level.7 The 87% net enrollment rate for the 6-12 age group is lower than the 92% current attendance ratereported in Table 4. This is because there are sample observations of children who are 12 years of age andreport having completed primary school.

3

Page 14: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty andlInequtality in the Distribution ofPuiblic Education Spending in South Africa

household and can only attend sporadically; this reduces their motivation to completethat level of education and proceed to the next.

Enrollment disparities are wi(le in secondlary schools and widest at tertiary institutions

Although H-Africans and NH-Africans have the lowest net enrollment rates insecondary schools, disparities across races are widest at the tertiary level: Enrollmentrates for Africans and Coloreds are less than 10% compared with Indians and Whites,whose rates are four times as high.

Net enrollment rates in secondary schools and tertiary education are 60% and11%, respectively, while gross enrollment rates are 97% and 14%.8 Disparities inenrollment rates across income quintiles and races are wide at the secondary level butwidest at the tertiary level (see Tables 2 and 3). The poorest quintile has about half thenet enrollment rate of the richest quintile at the secondary level, and only 10% at thetertiary level.

The spread between net and gross enrollment rates in secondary education for thepoorest quintile is twice as large as for the richest quintile, indicating that poor childrenare often older than the norm:9 Because they complete primary school at an older age,they are also older (than the average) in secondary schools, which decreases theirprospects of reaching the tertiary level. This, in turn, affects their earning capacity.

Girls and boys are equally likely to attend primary school; girls are more likely toattend secondary school

Gender differences are not substantial in primary school (see Appendix TablesA.3 and A.4). However, males at this level have slightly higher gross enrollment ratesthan females, reflecting a higher proportion of over-aged males than over-aged females.Disparities between girls and boys are wider at the secondary level, where more girls areenrolled: 63% overall net enrollment (for girls) compared with 56% (for boys). Thispattern occurs in every income quintile. Tertiary education is the only level at whichgirls have lower enrollment rates than boys, but this occurs only among the rich.

8 Enrollment rates at the secondary level are calculated for 13-17 year olds and for 18-22 year olds at thetertiary level.9 The largest spread betwveen the net and gross enrollment rates occurs at the third quintile. This may berelated to a lower drop-out rate a_nd a higher probability of staying in school even if repetition occurs.

4

Page 15: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribuion of Public Education Spending in South Africa

Table 2Net Enrollment Rates, by Quintile, Race and Level of schooling, 1993

RacesHh quintile African/ African/ Colored Indian White All

Non-Ex- Ex-Homeland Homeland

Primary educationPoorest 84 85 88 ** ** 85II 86 87 94 ** ** 87III 88 88 87 ** ** 88IV 89 88 91 93 90 89Richest 89 ** ** 91 90 90Total 86 86 90 92 90 87

Secondary educationPoorest 37 49 49 ** ** 46la 60 56 59 ** ** 57InI 63 67 74 ** ** 67IV 73 73 79 89 80 78Richest 77 ** ** 93 84 83Total 54 56 70 89 82 60

[Tertiary educationPoorest 5 4 4 ** ** 4II 5 5 2 ** ** 5III 7 10 3 ** ** 8IV 24 23 19 ** ** 20Richest ** ** ** 41 38 38Total 10 7 9 36 32 11

** Sample <= 30 observationsSource: SALDRU (1993)

5

Page 16: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty andInequality in the Distribution of Public Education Spending in South Afiica

Table 3Gross Enrollment Rates, by Quintile, Race and Level of Education

_____ ~~~~~~~~~~~(%)Races

Hh quintile African/ African/ Colored Indian White AllNon-Ex- Ex-Homeland Homeland

Primary educationPoorest 117 110 101 ** ** 112II 107 107 113 ** ** 108III 103 105 97 ** ** 103IV 97 96 98 98 98 97Richest 108 ** ** 91 98 97Total 108 108 101 96 98 106

Secondary educationPoorest 69 85 63 ** ** 81II 104 100 68 ** ** 98III 111 112 96 ** ** 110IV 109 107 103 99 101 108Richest 105 ** ** 103 98 101Total 96 98 88 99 98 97

Tertiary educationPoorest 5 5 4 ** ** 5II 6 8 4 ** ** 7III 9 13 5 ** ** 11IV 27 30 21 ** ** 24Richest ** ** ** 52 48 48Total 11 10 12 44 40 14

** Sample <= 30 observationsSource: SALDRU (1993)

High school expenses are the key factor explaining non-enrollment

The main reason why poor children are not in primary and secondary school(more than one in five)is that expenses are too high (see Table 4).10,11 In addition, 20%of secondary school-age children who were not in school said they could not cope withschool work, while 16% said they had become pregnant. 12

10 For a large proportion of primary school-age children, almost 60%, the reason for being out-of-schoolneeds to be further explored. These children are all grouped into the "other" category as their response forbeing out-of-school.II Household education expenditures are disaggregated in Section _.

12 Explored by gender in Table 7.

6

Page 17: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

Table 4Reasons for Being Out-of-School by Age Group(as % of non-enrolled in age group)

________ Age groups ..

Reason forbeing out-of- 6 to 12 13 to 17 18to22 Allschool (%) year olds year olds year

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _o ld s_ _ _ _ _

School expenses too high 21 23 15 17Need to work - 10 19 15Illness/disabled 6 9 4 5Pregnancies - 16 15 -

No school locally 4 4 1 2Could not cope with school 8 20 12 12workSchool boycott - 2 1 1Social unrest/violence 3 1 1Completed education - 3 24 16Required for work at 1 4 3 3home/farmOther 59 7 5 15Total non-enrolled 100 100 100 100Non-enrolled as percentage ofpopulation group 8 10 53 20Source: SALDRU Survey (1993)

Those from 18-22 who were not in school said that high costs were an importantreason, but even more (20%) said they could not attend because they needed to work. Inaddition, the non-enrolled population as a share of the age group is considerably higherfor this age group: Over half are not enrolled in any type of school. These responseshighlight the need to promote early enrollment and lower repetition rates, as thelikelihood of being out-of-school increases drastically for the population over 18.

Primary school-age Africans find expenses too high andfew local schools available

Among primary school-age Africans not attending school, the two main reasonswere that school expenses were too high (for over 20%) and there was no local schoolavailable (nearly 40% of H-Africans). These explanations account for more than 60% ofH-African and 30% of NH-African school-age children currently not enrolled. Genderdifferentials were not substantial.

7

Page 18: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequiality in the Distribution ofPublic Education Spending in South Africa

Table 5Reasons for Non-Enrollment for 6-12 Year-Olds by Race(as % of non-enrolled)

RacesReason for being out-of- African/ African/ Colored Indian White Allschool (%) Non-Ex- Ex-

Homeland HomelandSchool expenses too 17 26 4 21highNeed to work - - - - -

Illness/disabled 6 8 3 6Became pregnant - 1 - - -

No school locally 10 38 6 1 4Could not cope with 6 9 - - - 8school workSchool boycott - 1 - - 0Social unrest/violence 1 1 - - - 0Completed education - - - - - -

Required for work at 1 1 - - - 1home/farmOther 58 53 94 - 96 59Total non-enrolled 100 100 100 - 100 100Non-enrolled as percentof population group 8 9 6 0 5 8

Source: SALDRU (1993)

More than one in 10 secondary school-age Africans and Coloreds are not enrolled

A much larger proportion of secondary school-age Africans and Coloreds wereout of school compared with other races--more than 10%, compared with only 4% ofIndians and 2% of Whites. Reasons vary more across races for this age group than withthe primary school-age group. Africans found school expenses too high and 20% of H-Africans said they could not cope with school work, which could be related to the factthat 10% reported illness or disability. Almost 20% of NH-Africans needed to work,while nearly 33% of Coloreds reported they either needed to work or could not cope withschool work (see Table 6)

8

Page 19: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

Table 6Non-Enrollment among 13-17 Year-Olds by Race(as % of non-enrolled)

RacesReason for being out-of- African/ African/ Colored Indian White Allschool (%) Non-Ex- Ex-

Homeland Homeland

School expenses too high 20 28 3 - - 23Need to work 18 4 27 33 - 10Illness/disabled 4 11 9 - - 9Becarne pregnant 16 18 9 - 16No school locally 8 2 3 - 24 4Could not cope with school 16 21 27 17 - 20lworkSchool boycott 3 1 - - 2Social unrest/violence 1 3 3 17 - 3Completed education 2 2 - 33 76 3Required for work at 2 5 9 - - 4home/farmOther 11 5 9 - - 7Total non-enrol led 100 100 100 100 100 100Non-enrolled as percentageof population group 13 11 10 4 2 10

Source: SALDRU (1993)

One in three secondary school-age girls is not enrolled because ofpregnancy

Moreover, one third of girls aged 13-17 were not enrolled in secondary schooldue to pregnancy: A joint analysis by gender and race revealed that Africans have higherteenage pregnancy rates than other racial groups. Among out-of-school males in thesame age group, a third could not cope with school work. These findings are similar toother countries (see Table 7).

9

Page 20: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

Table 7Reasons for Non-Enrollment for Children Aged 6-12 and 13-17, by Gender (as % ofnon-enrolled)

6 to 12 year olds 13 to 17 year oldsReason for being out-of- Female Male All Female Male Allschool (%)School expenses too high 20 21 21 23 24 23Need to work - - - 8 12 10Illness/disabled 5 8 6 9 8 9Became pregnant 1 - 0 30 - 16Noschool locally 4 5 4 4 4 4Could not cope with school 7 8 8 11 31 20workSchool boycott - 1 0 1 2 2Social unrest/violence 1 0 2 4 3Completed education - 2 5 3Required for work at 1 1 1 4 4 4home/farmOther 63 57 59 8 6 7Total non-enrolled 100 100 100 100 100 100Non-enrolled as percentage ofpopulation group 7 9 8 1 1 10 10Source: SALDRU (1993)

IV. PUBLIC SPENDING ON EDUCATION

Increases in spending for public education outpaced population growth in the early1990s

Public spending for education grew steadily in the early 1990s. The increase inreal terms was more than twice as large as the population growth rate, 5.2% comparedwith 2.2%, respectively.'3 In addition, the allocation of state expenditures to publiceducation increased by more than public education spending as a share of GDP (seeTable 8).

13 Buckland, Peter and John Fielden (1994). Op. cit., p. 8

10

Page 21: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

Table 8Public Financing of Education

Public Spending Public SpendingFiscal Year On education as On education as

% of total state % GDPexpenditures

1987/88 22.9 5.81988/89 21.9 5.51989/90 22.0 5.71990/91 22.3 6.31991/92 23.9 6.51992/93 23.8 7.31993/94 24.4 7.3Source: Buckland and Fielden (1994)

Spending increased for primary schools and lecreasedfor technical schools

Primary and secondary schools received more than 75% of public educationresources, while technical colleges, teacher's education and technikons together obtainedless than 7%. Universities obtained a considerable proportion compared with othertertiary-level programs--about 10%. Further, from 1987-1992, the share of publicresources for primary education increased while the shares for technical education anduniversities decreased:14 The primary education share rose from 38% to about 43%,secondary education spending remained constant at about 30%, university spendingdecreased from 13% to 10%, and technical education spending dropped from 2% to 1.5%(see Figure 2).

1 4 Ibid

11

Page 22: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

Figure 2Share of Public Education Spending by Educational Level, 1991/1992

Pre-Primary Pnrmary 43%AdmW=intrtion6%7 3Adult!Voc 0.3 % _

University I 0%

Technikon 2.4%Teachers 2.9%Technical 1.5X

Private/Special 3%

Secondary 31 %

Source: Buckland and Fielden (1994)

Disparities are substantial in public spending per student

Disparities in public spending per student by department (under apartheid,separate departments were created to administer the schools for the different races) weresubstantial, particularly at the primary and secondary levels. These indicated qualitydifferentials in education by region since the population in the former homelandsreceived the lowest spending per student and three of the new provinces, the EasternCape, Northern Province and KwaZulu Natal, contain over 75% of the former homelandspopulation (see Table 9).

Per student spending for the House of Assembly (the institution that administerededucation for White pupils under apartheid) was R3,099 in primary and R4,675 insecondary. For the same educational levels, per student spending for the House ofDelegates (historically for Indians under apartheid) was R2,565 and R3,353 and for theHouse of Representatives (historically for Coloreds under apartheid) was R2,308 andR2,735, respectively. Public resources per student allocated to the African Departmentsof Education were considerably lower than for any of the Houses. Per student spendingfor the Department of Education and Training (DET-historically for NH-Africans) wasR1,012 at the primary level and R1,014 at the secondary level. The SGT and TBVCDepartments of Education (historically for H-Africans under apartheid) had the lowestper student spending for primary and secondary, R660 and R790, respectively.

12

Page 23: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

Table 9Annual Public Recurrent Spending on Education, 1993by Level, Department, Race and Student

Department and race Primary Secondary Tertiary*

DET 1,679.3 972.1 801.8SGT & TBVC 2,954.3 1,972.8 369.2House of Assembly 1,773.4 1,819.7 1,305.8House of Representatives 1,304.3 796.5 244.8House of Delegates 370.6 402.5 176.2Total 1 8,081.9 5,963.5 2,897.7ENROLLMENTS BY RACE (1 )BSA)Africans/non-ex-homeland 1,696,522 646,542 144,716Africans/ex-homeland 4,175,150 1,559,577 67,107Whites (HoA) 537,712 369,889 234,114Coloreds (HoR) 630,200 230,442 35,411Indians (HoD) 149,285 95,894 30,847Total 7,188,869 2,902,344 512,195ENROLLMENTS BY RACE ALDRU)Africans/non-ex-homeland 1,659,454 958,539 117,638Africans/ex-homeland 4,473,855 2,497,889 195,166Whites 572,171 389,236 143,882Coloreds 565,019 291,243 36,186Indians 144,476 120,037 21,244Total 7,414,975 4,256,944 514,116

SPENDING PER STUDENT (I _ _SA)

Africans/non-ex-homeland 990 1,503 5,540Africans/ex-homeland 708 1,265 5,502Whites 3,298 4,920 5,578Coloreds 2,070 3,456 6,912Indians 2,482 4,197 5,712Average 1,124 2,055 5,657SPENDING PER STUDENT (' ,ALDRU)Africans/non-ex-homeland 1,012 1,014 6,816Africans/ex-homeland 660 790 1,892Whlites 3,099 4,675 9,075Coloreds 2,308 2,735 6,764Indians 2,565 3,353 8,293Average 1,090 1,401 5,636

* Public spending and enrollments at the tertiary level are calculated bySource: DBSA (1993), SALDRU (1993)

13

Page 24: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

V. THE DISTRIBUTION OF PUBLIC EDUCATION SPENDING

The poor received a smaller share of public spending at every educational levelthan their share of the population, but funding is still larger than the part contributed byhouseholds. This pattern is seen as weakly pro-poor because the distribution of publicspending at every educational level lies between absolute equality and the Lorenzdistribution of total household expenditures: The distribution of resources by educationallevel is shown in Figure 10 across income quintiles together with the Lorenz distributionof total household expenditures.15 The diagonal line (or 450 line) is also known as theline of absolute equality since it goes through those points where the cumulative share ofthe population equals the cumulative share of public education spending.

This study used the Benefit Incidence Analysis methodology to analyze thedistribution of public spending across different socioeconomic and demographic groups.Its main benefit is that it can measure how well public services are targeted to certaingroups in the population, such as the poor, regions, girls and women. The targeting offunds across socioeconomic and demographic groups was analyzed with information onthe pattern of government recurrent spending per student obtained from DBSA (1993)and the pattern of enrollment rates obtained from SALDRU (1993): The spending andenrollment rates were determined by existing disparities. The incidence analysis usedper-student spending, calculated on SALDRU enrollments for primary and secondaryeducation, and per-student spending calculated on DBSA enrollments for tertiaryeducation (see Figure 3).16

15 hI this section, quintiles are created by ranking individuals on the basis of per adult equivalentexpenditures, from poorest to richest and then aggregating them into five groups with 20% of individuals ineach quintile. This way of creating population quintiles, by ranking individuals instead of households, iscommonly used in most countries where incidence analysis of public spending is available and thereforecan be used for cross-country comparisons.16 Per student spending in Table 9 by educational levels and race uses public recurrent educationexpenditures and school enrollments from two different sources: DBSA (1993) and SALDRU (1993).Recurrent expenditures and school enrollments reported by DBSA (1993) are for the 1991/92 school year.School enrollments obtained from SALDRU (1993) are an update for the 1993/94 school year. Householdsurveys like SALDRU are considered a reliable method for estimating primary and secondary enrolhnentssince students usually live with their parents. However, household surveys could underestimate terdary-level enrollments because it is common for students to live in single-headed households or boardingfacilities.

14

Page 25: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

Figure 3Annual Spending per Student by Educational Level and Race, 1993

8000

6000

4000

2000

0

Source: DBSA (1993) and SALDRU (1993)

The incidence of public spending was the result of (a) public policy decisionsabout the allocation of public expenditures to and within each sector; and (b) privatedecisions, i.e., household behavior. The allocation of funds to different socioeconomicand demographic groups was determined by a combination of supply and demandfactors: Supply factors included government allocations to and within the sector for eachtype of public service, and demand factors included household behavior regardingutilization rates of public services by socioeconomic groups. Thus, the incidenceanalysis integrates two sources of information: unitary subsidies by type of service (i.e.,the per student annual subsidy for the education sector), and information on individualutilization rates of public services (i.e., total enrollments by educational level)disaggregated by socioeconomic group, region and gender. In this way, funding foreducation was distributed across socioeconomic groups subject to their enrollment ratesand the annual public subsidy per student by educational level.

Public education spending is not pro-poor

If public education funds were equally distributed across the population, everyhousehold quintile would receive a percentage share equal to its share of the population.However, as mentioned above, based on the study findings, public education spendinghas not been pro-poor because the share of resources for the poor and ultra-poor has beensubstantially smaller than their share of the population:17 While the poor constitute 53%

17 Quintiles are created by ranking households, on the basis of per adult equivalent expenditures, frompoorest to richest and then aggregating them into five groups with 20% of households in each quintile. The

15

Page 26: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribufion ofPublic Education Spending in South Africa

of the population, they received only 40% of education resources; the ultra-poor, whorepresent 29% of the population (subsumed within the 53%), received just 20%.Conversely, the richest household quintile received almost twice its share of thepopulation, 23.4% of public resources for 12.5% of the population (see Figure 4).

This pattern across household quintiles is based on supply and demand effects: (a)public education spending per student is low for poor and ultra-poor students;18 and, (b)the poor and ultra-poor have very low enrollments at the secondary and tertiary levels. 19

Public education spending was less inequitable than the Lorenz distribution oftotal household expenditures (see Figure 10). The share of all public education fundingfor the poorest quintile was 14% compared with only a three percentage share of totalhousehold expenditures. By contrast, the richest quintile received a 35% share of allpublic education spending but concentrated 64% of total household expenditures.

Funding for primary education for the poorest quintile was 19%, compared with28% for the richest; at the secondary level, it was 11% and 39%, respectively. At thetertiary level,. the distribution was almost as inequitable as that of total householdexpenditures: The poorest quintile received only a 6% share of public tertiary-leveleducation resources compared with 47% for the richest quintile.

poverty profile defined as "poor" the poorest 40% of households and as "ultra-poor" the poorest 20% ("KeyIndicators of Poverty in South Africa", RDP, 1995). The poor constitute 53% of the population living inthe poorest 40% of households and the ultra-poor consitute 29% of the population living in the poorest 20%of households. A larger concentration of the population among poorer household quintiles occurs becausepoorer households tend to have larger families than richer ones (see Appendix Figure A. 1).18 This is because the distribution of the population by race within household quintiles shows that a largemajority of the poor and ultra-poor are Africans, mainly inhabitants of the former homelands (see AppendixFigure A.2). H-Africans get the smallest per student public education spending at every level of educationand NH-Africans get the second smallest amount at the primary and secondary educational levels (see Table9).19 See Tables 6 and 7.

16

Page 27: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

Figure 4Spending by Household Quintile and Race

i~~~~~~~~~ o1 Wh18iteSource: DBSA (193 sandg I 1ia993)

atth riar chollve: lhoght23.4evels 80a thoeirsareidth

spopolaFion 540)10 0 Coloured

201

0 M~~~~~~~~~~~~~~~~~AfricaW Non-Primary-L lex- Homsenand

I

Source: DBSA (1993) and SALDRU (1993)

The poor and the ultra-poor received the highest allocation of education fundingat the primary school level: Although they received less than their share in thepopulation, they obtained almost half the funds spent for this level. The spending sharewithin quintiles by race indicated that African students constituted the large majority ofprimary enrollments amnong the poor and ultra-poor, but still received less than theirshare of the population (see Figure 5).

Figure 5Primary-Level Education Funding by Household Quintile and Race

30 U9 j0 - OAfricanlex-

~~20 E189 Spendinzg600omanUPopulation M In dian

10 - 00 Coloured

II ~~~~~~~~200 *~~~~~~~~~~~~~~~~~~~African/ Non-

0 4 0 ~~~~~~~~~~~~~~~~~~ex- Homelantd

Source: DB SA (1 993) and SALDRU (1 993)

With regard to secondary schools, less than 40% of spending was allocated to thepoor, although, as mentioned above, they represented 53% of the population. The ultra-poor received less than 66% of the spending as compared with their share of the

17

Page 28: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

population. By contrast, the richest quintile received 25% for their 12.5% of the

population. The spending share within quintiles by race indicates that Africans were thelarge majority of secondary enrollments among the poor and ultra-poor, but received less

than their share of the population (see Figure 6).

Figure 6Secondary-Level Education Funding by Household Quintile and Race

10 .10El Africani ex-

0 o .s: ex- fHomeland30 25.3 Nit

21.4 * 60

a Spending Th40 I- rndians Population

2 10 IiColoured

0 African!Non-- j ex- Homeland

Source: DBSA (1993) and SALDRU (1993)

With respect to tertiary education, although the poor accounted for more than halfthe population, they received less than 25% of public spending. The ultra-poor receivedabout 33%, when compared with their share of the population, while the richest quintilereceived 32.2%, for their 12.5% share. This, as mentioned above, is mainly due tosignificantly low tertiary-level enrollments among the poor and ultra-poor.20 Thespending share within quintiles by race indicated that African students made up most ofthe tertiary enrollments amnong the poor and ultra-poor, but they received less than theirshare in the population (see Figure 7).

20 Ibid.

18

Page 29: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty andInequality in the Distribution of Public Education Spending in South Africa

Figure 7Tertiary-Level Education Funding by Household Quintile and Race

32.2 i100*3Afiicanlex-

30 27.8 Homeland

a20 *Spending .~60

3. *Populafion * Indian~~~~ 0.~ ~ ~ ~ ~~~~4

~10 '0 20 l[ Coloured

O . _ % O % AfricanlNon-

tz ex- Homeland

Source: DBSA (1993) and SALDRU (1993)

Race-specific subsidies reveal severe inequities in the system

Severe inequities persist in the distribution of public education resources acrossincome groups, regions and races. The major reason is the substantial difference in publicspending per student by race:21 Per capita public spending is lowest for Africans at everyeducational level, due to both supply and demand effects. African departments ofeducation have the smallest allocation of public spending per student and Africans havethe lowest enrollments at every educational level (see Figure 9).

From all public education spending aggregated for all levels (primary, secondaryand tertiary) in per capita terms, H-Africans received R3 11, NH-Africans R328, ColoredsR765, Indians R899 and Whites R95 1(see Appendix Table A.5). Differences in spendingper student explained the disparities at the primary and secondary levels, whileenrollments were the main source at the tertiary level.

H-Africans in primary schools received RI 53 a year, the smallest spending in percapita terms. NH-Africans received R167, while Coloreds received R424, Indians R372,and Whites R3 84. Coloreds received the greatest chunk per capita because ofdemographics (they have large families), and enrollment patterns (they have highenrollments at the primary level). The incidence analysis by income groups suggests that

21 See Table 9 and Figure 3.

19

Page 30: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty andInequality in the Distribution of Public Education Spending in South Africa

race-specific subsidies have a significant impact on the distribution of funding acrosspopulation groups.

Although the ultra-poor constituted 33% of all enrollments, they received only20% of public education spending. By contrast, the richest income quintile accounted forless than 10% of all enrollments but received almost 25% of all public education funds.This was determined through race-specific subsidies, compiled by DBSA.Unfortunately, progress in the distribution of education resources will be increasinglydifficult to assess where race-disaggregated information is no longer being collected (seeFigure 8).

If average, instead of race-specific, subsidies had been used in the incidenceanalysis of public education spending in 1993, funds would have appeared to be equallydistributed with every income quintile receiving almost its share of enrollments (seeFigure 8).

Figure 8Public Education Enrollments and Spending by Household Quintile, 1993

35 34

30 _ average subsidies26 2 race-specific subsidies

24 225 I 23

21 ~~~~~~~~~~20*S20 18 19 19 1 7

1515 ~~~~~~~~~~~~13

15 ~~~~~~~~~~~~~~~~~~12

10

0Poorest IT III IV Richest

Source: DBSA (1993) and SALDRU (1993)

Public education spending per capita is lowestfor Africans

The poverty focus of public investments in human capital can be improved ifeducation incentives, along with a larger share of resources, are targeted to increasesecondary and tertiary enrollments of H-Africans. However, because not all H-Africanstudents are poor or ultra-poor, it is essential to identify the geographic areas within theformer homelands that are the most impoverished to improve the targeting. In thiscontext, the Eastern Cape, KwaZulu Natal and Northern Province could be targeted,

20

Page 31: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distihbution ofPublic Education Spending in South Africa

since the study found these three provinces have by far the highest poverty rates, andconcentrate more than half of H-Africans.

Figure 9Annual Education Spending Per Capita, by Level of Schooling and Race

1000900800

to 50 M Secondary400 a~~~~~~~~~~~~~~~~~~~~ Tertiary

t 300- *AI Education

2,00l00

Source: DBSA (1993) and SALDRU (1993)

Targeting the Eastern Cape, KwaZulu Natal and Northern Province can improve thepovertyfocus ofpublic education spending

Targeting public education spending to the poor and ultra-poor will requireactions on both the supply of and demand for education. The poverty focus of publicinvestments in human capital can be substantially improved if (a) incentives are targetedto increase secondary and tertiary enrollments of H-Africans and (b) a larger share ofpublic resources is allocated to them. Important gains could be made by targeting theEastern Cape, KwaZulu Natal and Northern Province.

VI. HOUSEHOLD SPENDING ON EDUCATION

lThepoor cannot afford the direct costs of education

The direct costs of sending children to school are a key factor that discouragesattendance. Most families spend a significant proportion of their resources to sendchildren to school: Direct expenses include school fees, uniforms, transport, meals, booksand paper and other items such as boarding fees, contributions to school buildings, extraamounts for teachers and extramural activities.

Of the total spent on public education, households contributed 20% of the cost pereach child in both primary and secondary schools, and close to 40% at the tertiary level:Families spent R267 per primary student annually compared with RI, 149 allocated from

21

Page 32: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

public funds, R484 per secondary student compared with R2,055 from the state, andR3,3 18 compared with R,657 at the tertiary level.

Public spending per student at every level was larger than the direct costs incurredby poor families sending their children to school; for students at the primary level, it wastwo and a half times the amount spent by poor families on per capita non-food items (seeTable 10). The state contribution to education was almost three times per secondarystudent and about eight times per tertiary student the amount spent by poor householdson non-food items per child. Thus, although the government contribution to educationwas larger than the families' share, on average, households contributed a significantproportion of the total education cost for each child in school.

Reduced education costs could increase enrollments among the poor

Uniforms, transport and meals were the most important expenses for poorfamilies sending children to primary or secondary school--nearly 75% of householdeducation expenses; by comparison, non-poor households spent only 15%.

At the tertiary level, school fees were the largest household expense for both thepoor and the non-poor, on average 56% and 69%, respectively. And, based on thehousehold survey, school fees discouraged tertiary-level enrollment among the poor:Each poor student enrolled at this level needed more than two and a half times the percapita amount spent on non-food items compared with less than one-third for the non-poor. The analysis of enrollment rates by income quintile revealed that the poorest threequintiles had significantly lower tertiary-level enrollment compared with the richest twoquintiles (see Table 10).

Poor families had a total direct cost of more than 40% of per capita non-foodhousehold expenses per child sent to primary school compared to only 6% for non-poorfamilies. Direct costs increased to half of per capita non-food household expenses forchildren in secondary school, compared to 10% for the non-poor (see Table 10).

One way to reduce the direct costs of tertiary education would be to increase thenumber of scholarships that would exempt students from paying school fees. Also, costrecovery mechanisms that do not discriminate against the poor could be introduced tofree up public education funds to finance the additional expense.

22

Page 33: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South Africa

Table 10Household Expenses on Education by Educational Level, 1993 Total and Per Student

Household expenses on educationEducational Share of School Uniforms, Books & Misc. Total Total as a Publiclevel and enrollments fees transport stationary * share of spendingenrollment & meals per capita per stadentgroup non-food as a share

Hh of perexpenses capita non-

food Hhexpenses

PRIMARY _

All primary 100% 35% 47% 6% 12% 100% 12% 52%enrollmentsPoor primary 60% 17% 73% 5% 5% 100% 41% 246%enrollmentsNon-poor 40% 40% 39% 6% 14% 100% 6% 23%primaryenrollments

SECONDARY _ __ IAl secondary 100% 31% 47% 7% 14% 100% 15% 62%enrollmentsPoor secondary 52% 14% 71% 5% 10% 100% 50% 297%enrollments INon-poor 48% 36% 40% 8% 16% 100%|10% 38%secondaryenrollments

TERTIARY l l l_____Allteriary 100% 67% 12% 9% 111% 100% 29% 49%enrollmentsPoor tertiary 25% 56% 32% 12% 0% 100% 258% 798%enrollments I INon-poor 75% 69% 10% 9% 12% 100% 29% 44%tertiaryenrollments

Vi. CROSS-COUNTRY COMPARISONS

The poor in South Africa receive a smaller share of public education spending atevery educational level than their share of the population but the shares of spending arelarger than their total household expenditure share. This pattern is characterized as

23

Page 34: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Education Spending in South AJrica

weakly pro-poor because the distributions of public education spending at everyeducational level lie in between absolute equality and the Lorenz distribution of totalhousehold expenditures. The distributions of education resources by educational level areshown in Figure 10 across income quintiles together with the Lorenz distribution of total

household expenditures.2 2 The diagonal line (or 450 line) is also known as the line ofabsolute equality since it goes through those points where the cumulative share of thepopulation equals the cumulative share of public education spending.

Figure 10Distribution of Public Education Spending and Lorenz Distribution, 1993

100v

80 Primery ,.

60- AullAuction -

i 20 Household

l! , S , ,> ~~~~~eapenitresO i ~~~~~~Tertiary

0 20 40 60 80 100

Cumulative share ofpopulition

Source: DBSA (1993) and SALDRU (1993)

Public education spending in South Africa is less inequitable than the Lorenzdistribution or distribution of total household expenditures. The share of all publiceducation resources going to the poorest quintile is 14 percent compared with only athree percentage share of total household expenditures. By contrast, the richest quintilereceives a 35 percentage share of all public education spending but concentrates 64percent of total household expenditures.

Public primary education spending going to the poorest quintile is 19 percentcompared with 28 percent for the richest quintile. At the secondary level, the poorest

22 In this section, quintiles are created by rankdng individuals on the basis of per adult equivalentexpenditures, from poorest to richest and then aggregating them into five groups with 20 percent ofindividuals in each quintile. This way of creating population quintiles, by ranldng individuals instead ofhouseholds, is commonly used in most countries where incidence analysis of public spending is availableand therefore can be used for cross-country comparisons.

24

Page 35: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribufion ofPublic Education Spending in South Africa

quintile receives an 11 percent and the richest quintile a 39 percent of public educationresources. The distribution of public spending in tertiary education is almost asinequitable as the distribution of total household expenditures. The poorest quintilereceives only a six percentage share of public tertiary-level education resourcescompared with a 47 percent share going to the richest quintile.

South Africa has one of the worst distributions ofpublic education spending

In South Africa, public education spending is weakly pro-poor when compared tocountries like Chile, Colombia and Uruguay, which are strongly pro-poor. The cross-country comparison is particularly significant with Chile because the Lorenz distribution,or distribution of household expenditures, is as inequitable as in South Africa. However,Chile has launched educational reforms to progressively distribute public funds (seeTable 1 1).

Except for Madagascar, South Africa has about the most inequitable distributionof public education spending among other African countries. Also, significant disparitiesexist across races for public education spending which are wide at the secondary leveland widest at the tertiary level.

Table 11Incidence of Public Education Spending on the Poorest and Richest Quintiles,Selected Countries, Percentage Share

Education spending benefiting:Country Year the poorest 20% of the the richest 20% of th

l__________ population (%) population (%)AFRICACote d'Ivoire 1993 14 35Ghana 1992 16 21Kenya 1993/94 17 21Madagascar 1993 8 41Malawi 1994/5 16 25South Africa 1993 14 35Tanzania 1993 13 23LATIN AMERICAChile 1986 25 17lColombia 1992 23 14Mexico 1992 14 27Uruguay 1989 33 15

Sources: Dayton (1995b); Bernier, Chao, and Demery (1994); Castro-Leal (1995); Demery andVerghis (1994); World Bank (1995).Note: If public education spending were equally distributed across population quintiles, then thepoorest and richest income quintiles (every number in the table above) would receive a 20percentage share of spending.

25

Page 36: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution ofPublic Educafion Spending in South Africa

VII. POLICY IMPLICATIONS

Reduce inequities by geographically targeting spending to the poor

Inequality in the distribution of resources widens by educational level fromprimary to tertiary across different income groups and races; and, public educationspending per capita is lowest for Africans at every educational level. The poverty focusof public investments in human capital can be substantially improved if educationincentives are targeted to increase secondary and tertiary enrollments of ex-Homeland-African students and if a larger share of public education resources is allocated to them.

However, because not all African students are poor or ultra poor, policies totarget resources to the poor need more precise means of identifying them. Since thepoverty profile found that nearly 70% of the poor live in the former homelands, gains inallocating education resources to the poor and ultra-poor could be achieved by targetingthe Eastern Cape, KwaZulu Natal and Northern Province, which have the highest povertyrates by far (which is linked to poverty in the former homelands). In addition, theseProvinces contain more than half the former homelands population.

Reduce education expenses on uniforms, transport and meals for poor primary andsecondary students

Targeting public education spending towards the poor and ultra-poor will requireactions on both the demand and supply side. The direct costs of sending children toschool are a critical factor determining the demand for education. Most families spend asignificant amount for school fees, uniforms, transport, meals, books and paper to sendtheir children to school; for poor families, this amounts to more than 40% of per capitanon-food expenditures, compared with only 6% for non-poor families. Moreover,education costs rise to half of such expenditures per child in secondary school, comparedwith 10% for the non-poor.

As mentioned earlier, poor families spend close to 75% of their householdeducation budget on uniforms, transport and meals for children in primary and secondaryschools in contrast to only a 15% spent by non-poor households. This suggests that if thecosts of these items for poor primary and secondary school-age children were lowered,enrollments could increase.

Create a higher education scholarship program and develop a credit market

Public resources dedicated to tertiary education largely favor non-poor students.About one-third of these funds benefit the richest income quintile, which contains only9% of the 18-22 year olds. Thus, public tertiary education is prohibitive for the poor:

26

Page 37: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

Each poor student at this level needs more than two-and-a-half times the per capitaamount spent on non-food items to pay for school expenses compared with less than one-third for non-poor students. This suggests that exempting poor students from payingschool fees could increase their enrollment at this level.

To reduce the direct costs of tertiary education for the poor, the number ofscholarships could be increased. Also, cost recovery mechanisms that do notdiscriminate against the poor could be simultaneously introduced to free up publiceducation funds to finance the additional expense: One option to free public educationfunds and simultaneously increase cost recovery is to develop tertiary education creditmarkets. Such programs have been designed in other countries and it appears they dobest when student loans are channeled through commercial credit institutions to ensureloan recovery, increase credibility and maintain cost-effectiveness.

27

Page 38: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

REFERENCES

Buckland, Peter and John Fielden. 1994. Public Expenditure on Education in SouthAfrica, 1987/88 to 1991/2. Center for Education Policy Development,Johannesburg and The World Bank, Washington, DC.

Castro-Leal, Florencia, Julia Dayton, Lionel Demery and Kalpana Mehra. 1998. PublicSocial Spending in Africa: Do the Poor Benefit? The World Bank, Washington,D.C. (mimeo).

Development Bank of Southern Africa (DBSA). 1993. Public Expenditure on Educationin South Africa, 1987/8 to 1991/2. Volume 1: Financial Information and Volume2: Contextual Information. Development Information Group of DBSA and theWorld Bank. South Africa.

Meerman, Jacob. 1979. Public Expenditure in Malaysia: Who Benefits and Why? OUPfor the World Bank, Oxford.

Ministry in the Office of the President: Reconstruction and Development Programme(RDP). Key Indicators of Poverty in South Africa. South Africa, 1995.

Southern Africa Labour and Development Research Unit (SALDRU), University of CapeTown. 1993. South Africa Living Standards and Development Survey. SouthAfrica.

Selowsky, Marcelo. 1979. Who Benefits from Government Expenditure? A Case Study ofColombia. A World Bank Research Publication, Washington, D.C.

Van de Walle, Dominique and Kimberly Nead. 1994. Public Spending and the Poor:Theory and Evidence. PRDPE. World Bank, Washington, D.C.

28

Page 39: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

APPENDIX A

Table A.1Racial composition of former homelands

Race l_ _

African 99.95Colored 0.01Indian 0.02White 0.02Total 100.00

Source: SALDRU (1993)

Table A.2Africans by area of residence

Race __ _

H-Africans 65.68NH-Africans 34.32Total 100.00

Source: SALDRU (1993)

29

Page 40: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

Table A.3Net enrollment rates, by quintile, genderand level of schooling, 1993

Gender| Hh Female Male All

r________ Primary educationPoorest 87 84 85II 87 87 87III 88 88 88IV 87 91 89Richest 91 89 90Total 87 87 87

Secondary educationPoorest 51 40 46II 61 52 57III 68 67 67IV 81 76 78Richest 87 79 83Total 63 56 60

Tertiary educationPoorest 5 3 4II 6 4 5III 10 6 8IV 19 21 20Richest 32 44 38Total 11 10 11

Sources: SALDRU (1993)

30

Page 41: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

Table A.4Gross enrollment rates, by quintile, genderand level of schooling, 1993

GenderHh Female Male All

Primary educationPoorest 109 114 112II 105 111 108III 101 105 103IV 95 100 97Richest 96 99 97Total 104 108 106

Secondary educationPoorest 81 81 81II 104 92 98EJ 117 102 110IV 115 100 108Richest 104 97 101Total 100 93 97

Tertiary educationPoorest 6 4 5II 8 6 7III 12 8 1 1IV 23 25 24Richest 40 57 48Total 14 13 14

Source: SALDRU (1993)

31

Page 42: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequality in the Distribution of Public Education Spending in South Africa

Figure A.1Population distribution by household quintiles

30.0 29.0

1!~~~~2. Houlsehold share

20.0

It|0.0~~~~~~~~~~~~~~~~~~~~~~~i-

Source: SALDRU (1993)

Figure A.2Distribution of the population by race withinhousehold quintiles

1 V

Homeland

60 S (1993

40 MIndian

20 O Colmodee

0EAfticaWNorn-gx-ii ~~~~~~ ~Homeland

Source: SALDRU (1993)

32

Page 43: Poverty and Inequality in thedocuments.worldbank.org/curated/en/517971468781194575/... · 2016-08-06 · Florencia Castro-Leal, February 1999 * Review of Public Expenditures: Efficiency

Poverty and Inequaity in the Distribution of Public Education Spending in South Africa

Table A.5Annual education spending, by level of schooling and race, 1993(Rand per capita)

Education spending by raceEducational H-African NH- Colored Indian White All

Primary 153 167 424 372 384 213Secondary 102 97 259 404 394 157Tertiary 56 65 81 122 174 76

All education 311 328 765 899 951 446

Source: DBSA (1993), SALDRU (1993)

33