69

Universal primary enroll-ment is another one of theMDGs related to child welfare. The millenniumdevelopment goal is to ensurethat, by 2015, all children arein school, the net primaryenrollment ratio is 100%, andthat all the pupils enteringgrade 1 are retained until grade5 (typically the last year ofprimary school).

The numerous benefits ofschooling are well known andhave been widely discussed inthe literature on economicdevelopment. Schooling is oneof the most powerful instru-ments for reducing poverty,unemployment and inequality;improving health andnutrition; and promotingsustained, human develop-ment-led growth. It is also self-perpetuating acrossgenerations, with educatedparents much more likely toprovide schooling to theirchildren. Both the pecuniaryand non-pecuniary returnsfrom schooling have been well-documented in the literaturefor several countries, includingIndia.

Overall TrendsLevels and trends

India has made rapid stridesin schooling during the last 4-5decades. The gross primaryenrollment rate, which was only 43% in 1950-51, reached100% by 1990-91, and has fallen slightly since then (Figure IV.1).62

As in many othercountries, gross enrollmentrates obtained from schooladministrative data in Indiatend to overstate the actualenrollment rate.63 House-

hold survey data showmuch smaller net, gross andage-specific enrollments.The most recent source ofnationally-representativedata on enrollments is the55th round of the NationalSample Survey (NSS)undertaken in 1999-2000.These data suggest thatonly 78% of children aged6-11 years were attendingschool in 1999-2000. At52.5%, the net primaryenrollment rate wassignificantly lower.64 Thereare many reasons for the

������������� �

43

62

7981

100

104

9295

30

40

50

60

70

80

90

100

1950-51 1960-61 1970-71 1980-81 1990-91 1995-96 1998-99 1999-2000

Figure IV.1 : Gross primary enrollment rates, India,1950-51 to 1999-2000

62 Note that “primary” schooling refers to grades 1-5 in this chapter. The term “lower primary” is sometimes used in India todenote grades 1-5, while “upper primary” refers to grades 6-8. “Elementary” education refers to grades 1-8.

63 Note that age-specific or net primary enrollment rates are not available from the school-based administrative data.64 The gross primary enrollment rate for 1999-2000 was 61%.

70

Attaining the Millennium Development Goals in India

discrepancy betweenhousehold survey-basedand school administrativerecords-based enrollmentrates, including the factthat household surveystypically obtain infor-mation on schoolattendance during a shortreference period (thequestion to householdstypically is: “is your child currently attendingschool?”), while admini-strative data refer tostudents actually enrolledin the registers of theschool at the beginning ofthe school year. Inaddition, gross enrollmentrates from administrativerecords are very sensitive toincorrect estimates of thepopulation of school-agedchildren. Finally, there are

incentives for schooladministrators and districtofficials to overstate thenumber of enrolledstudents, since many typesof government educationexpenditure allocations to

schools and districts areoften based on the numberof enrolled students.65

Inter-state variationsAn average gross primary

enrollment rate of 95% for thecountry in 1999-2000 for thecountry masks wide variationsacross states. Gross primaryenrollment rates vary from alow of 65% in Uttar Pradesh toa high of 139% in Sikkim(Figure IV.2). Some of thenumbers seem counter-intuitive, such as the relativelylarge rate for Rajasthan and thelow rates for Kerala and Punjab,but these likely reflectdifferences in the extent of private-school enrollmentacross states. School-basedenrollment data often do notinclude students attendingprivate schools.

60 70 80 90 100 110 120 130 140

U.P.Goa

J & KashmirBihar

PunjabHaryana

PondicherryDelhi

KeralaH.P.

NagalandManipur

IndiaW Bengal

TamilnaduA.P.

OrissaTripura

KarnatakaRajasthan

GujaratMaharashtra

AssamM.P.

MeghalayaArunachal

Sikkim

Figure IV.2 : Gross primary enrollment rates, by state, 1999-2000

65 Even so, the discrepancy between the gross primary enrollment rate of 95%, as reported by school administrative data, and61%, as calculated from household surveys, is overly large. It is unlikely that it can be accounted for by students who areenrolled in school but not attending.

0

50

0

10

00

15

00

20

00

25

00

30

00

35

00

40

00

45

00

50

00

55

00

60

00

65

00

70

00

75

00

West BengalBihar

Andhra PradeshUttar Pradesh

Madhya PradeshOrissa

Jammu & KashmirPunjab

RajasthanKarnataka

HaryanaMaharashtra

AssamTripura

Tamil NaduKerala

ManipurGujarat

GoaMeghalayaNagaland

Himachal PradeshArunachal Pradesh

MizoramSikkim

Per student enrolled in elementary school

Per child aged 6-14 years

Figure IV.3 : Public spending on elementary education per child aged6-14 years and per student enrolled in elementary school, 1998-99

Economic Growth,Public Spending,and SchoolEnrollmentsPatterns of public expenditureon elementary education

As Figure IV.3 shows,states spend very differentamounts on elementaryeducation. In 1998-1999,Sikkim, Mizoram, ArunachalPradesh and HimachalPradesh spent upwards of

Rs. 3,000 per child ofelementary school age (i.e., 6-14 years). At the other end,Bihar, Uttar Pradesh, AndhraPradesh and West Bengalspent less than Rs. 1,000 perchild 6-14. Thus, the stateshaving low school enrollmentor attendance rates are alsothe ones that spend less perchild and per enrolled child.Note that some of the inter-state variation in the

discrepancy between theamount of public expenditureper child and that perenrolled child arises due tovariations in the proportionof children attending privateschools. In Goa, Kerala,Punjab and Haryana, forinstance, private schooling ismore common amongchildren 6-14, with the resultthat public spending per childenrolled in public school issignificantly larger thanpublic spending per childaged 6-14.

Figure IV.4 shows theincrease in public spending on education over time and across states. In most states, real growth in aggregate public spending onelementary education increasedimpressively between 1980-81and 1999-2000 – exceeding thegrowth in the elementaryschool-aged population (ages 6-14) as well in elementaryenrollments. As a result, realpublic spending per child 6-14

71

Primary Schooling

annual growth in aggregate public spending on elementary educationannual growth (%) in real public spending per child 6-14annual elementary enrollment growth (%)annual growth in number of children aged 6-14 (%)

-1

0

1

2

3

4

5

6

7

8

9

Punja

b

And

hraPra

desh

Ker

ala

Bihar

Wes

t Ben

gal

Mah

aras

htra

Kar

nata

ka

Utta

r Prade

sh

Har

yana

Guj

arat

Rajas

than

Tamil

Nad

u

Mad

hyaPra

desh

Oris

sa

Figure IV.4 : Annual % growth in elementary enrollments, populationaged 6-14, and real government expenditure on elementary

education per child aged 6-14, by state, 1980-99

Himachal Pradesh

58 60

38 33

5 7

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Students Public spending

Elementary Secondary University

Andhra Pradesh

63

50

33

30

4

20

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Students Public spending

Elementary Secondary University

Figure IV.5 : Distribution of enrolled students and of public spendingon education in two states, by level, 1998-99

increased in virtually every state– at annual rates of 2% or more.

An important issue thatarises is the allocation of publicspending on education acrosslevels. Since the unit cost ofuniversity education is muchgreater than that of elementaryand secondary schooling, theshare of universities and highereducational institutions inpublic spending on educationtypically exceeds their share ofthe total student population inmost countries. This is true ofIndia as well, but themismatch between student andspending shares varies a greatdeal across states. An exampleof a state where the mismatchis large is Andhra Pradesh. Theelementary education sectoraccounts for 63% of all

students in the state but foronly 50% of all publicspending on education (FigureIV.5). In contrast, the higherschooling sector in AndhraPradesh commands 20% of allpublic resources going intoeducation but accounts foronly 4% of the students. Atthe other extreme is HimachalPradesh, where the shares ofthe elementary educationsector in both the number ofstudents and in total publicspending on education areapproximately similar.

It may be instructive to seehow inter-state changes in public spending on elementaryeducation per child arecorrelated with changes inenrollment rates over time. Anadvantage of the administrative

data on enrollments is that theyare available at the state level inIndia going back a number ofyears. In order to analyze theassociation between economicgrowth, public spending onelementary education, and grossprimary enrollment rates acrossstates and over time, we havemerged the state-level data ongross primary enrollment ratesover the period 1980-99 withstate-level data on real gross state domestic product per capita (GSDP) and real public spending on elementaryeducation per child over thesame period.66

A plot of the data suggests apositive association betweenthe gross primary enrollmentrate and real governmentexpenditure on elementaryschooling per child aged 6-14years (Figure IV.6). However,since there is no control forother variables that may alsoinfluence school enrollments, itis difficult to make much of the association observed in the figure.

Another manner in whichto view the data is to examinethe inter-state increases inelementary enrollments overthe period 1980-81 to 1999-2000 relative to the growth ofper child public spending onelementary schooling over thesame period. This is shown in

72

Attaining the Millennium Development Goals in India

66 Data on government elementary education expenditure were obtained from detailed budget demand documents of individualstates. Although government expenditure on primary education is of interest in this paper, state government expenditure inIndia are available only for the elementary level, comprising the lower primary (grades 1-5) and upper primary (grades 6-8)levels. Note that the state government expenditure data used in this paper includes expenditure incurred by a state governmentout of its own revenues as well as central government allocations to that state.

4.0

4.1

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

5.0

5.1

5.0 5.5 6.0 6.5 7.0 7.5 8.0

Ln real gov't exp on elementary schooling per child 6-15

Ln

gro

sslo

we

rp

rim

ary

en

roll

me

nt

rate

Figure IV.6 : Gross lower primary enrollment rates and realpublic spending on elementary schooling per child 6-14,

across states, 1980-99

Figure IV.7. The figuresuggests that over this 20-yearperiod, some poor, low-enrollment states, such asMadhya Pradesh andRajasthan, were able to enrolllarge numbers of children in elementary school withrelatively small increases in public spending onelementary education. Butother poor states, such asUttar Pradesh, Bihar andOrissa, were able to increaseelementary enrollments byonly half as much for everyadditional (public) Rupeedevoted to elementaryschooling. The differencesobserved in Figure IV.7 could,of course, reflect differences inthe quality of schooling acrossstates (as is most likely the casewith Kerala, which had thelowest number of additionalstudents in elementary schoolin relation to increasedexpenditures), but they couldalso reflect differences in the

marginal cost of expandingenrollments. In MadhyaPradesh, the launch of theEducation Guarantee Scheme(see Box IV.1), whichexpanded the number ofschools in the rural areas ofthe state at relatively low cost,may have been responsible for the large increase in enrollments relative to

changes in governmentelementary education expen-diture per child.

Figure IV.8 shows theadditional numbers ofstudents enrolled at the lowerprimary level between 1980and 1999 in several statesrelative to the real increase ingross state domestic product percapita over the same period.The poor states of Bihar,Madhya Pradesh, UttarPradesh, Orissa and Rajasthanrank among the states thatwere able to obtain the largest increases in primaryenrollments for given increasesin (real) gross state domesticproduct per capita. Of course,these data simply reflect thefact that these poor states have experienced very littleeconomic growth, butnevertheless have been able toincrease primary enrollmentssignificantly during the lasttwo decades.

73

Primary Schooling

8,5

27

3,3

11

2,7

94

2,6

44

2,6

00

2,1

47

2,0

98

2,0

08

1,7

51

1,5

29

1,3

89

1,1

43

875

720

654

571

498

477

466

385

378

358

302

117

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

Ker

ala

Guj

arat

Bih

ar

Utta

r Prade

sh

Oris

sa

Mah

aras

htra

WBen

gal

Har

yana

And

hra

Prade

sh

Kar

nata

ka

Raj

asth

an

Mad

hya

Prade

sh

Mad

hya

Prade

sh

Raj

asth

an

Kar

nata

ka

And

hra

Prade

sh

Har

yana

WBen

gal

Mah

aras

htra

Oris

sa

Utta

r Prade

sh

Bih

ar

Guj

arat

Ker

ala

Additional government expenditure (constant

1993-94 Rs.) on elementary schooling per

additional student enrolled in elementary

school, 1980-99

Additional students enrolled in elementary

school per million additional Rs. (constant

1993-94 prices) in government expenditure on

elementary schooling, 1980-99

Figure IV.7 : Inter-state increases in elementary school(grades 1-8) enrollments and in real governmentexpenditure on elementary education, 1980-99

2,6

96

1,9

48

1,5

93

888

734

486

426

227

215

105

59

50

0

500

1000

1500

2000

2500

3000

Bih

ar

Mad

hya

Prade

sh

Utta

r Prade

sh

Oris

sa

Raj

asth

an

WBen

gal

And

hra

Prade

sh

Har

yana

Mah

aras

htra

Guj

arat

Kar

nata

ka

Ker

ala

Figure IV.8 : Additional number of primary school (grades 1-5)students enrolled per (1993-94) Rupee increase in gross state

domestic product per capita, 1980-99

In Annex IV.1, an attemptis made to estimate therelationship between grossprimary enrollment rates andgovernment expenditure onelementary schooling per child6-14 more formally and withcontrols for other variables.While the results of the analysisare not unambiguously clear,there is evidence of a significantpositive association betweenthe gross primary enrollmentrate and per-child governmentexpenditure on elementaryeducation across states, evenafter controlling for per capitaincome and adult femaleliteracy. In addition, dependingupon different specifications,the positive associationbetween the gross primaryenrollment rate andgovernment elementary schoolexpenditure is observed to bestronger for the very poor statesthan for the non-poor states.

The finding that the positiveassociation between enrollmentsand government expenditure oneducation is stronger in the poorstates finds some support fromanother study on India.Lanjouw and Ravallion (1999)find that the poor in Indiatypically benefit more thanproportionately at the marginwhen there is an overallexpansion in primary schoolenrollments. This happensbecause poor students are almostalways the last to be enrolled,and the better-off are typicallyalready in school. As a result, government educationalexpenditures that expandschooling access are generallywell-targeted to the poor.

The econometric resultspresented in Annex IV.1 alsoindicate a strong positiveassociation between the grossprimary enrollment rate and

per capita GSDP, with theassociation becoming weakerat higher per capita incomelevels. Surprisingly, however,when there is control for percapita GSDP and per-childpublic spending on elementaryeducation, adult female literacyis not significantly associatedwith the gross primaryenrollment rate.

Household Survey-Based EnrollmentEstimates

As noted earlier, grossenrollment rates obtained from school administrativerecords differ significantly from household survey-basedestimates of enrollment (moreappropriately, attendance).67

Data from the latest NSS round(55th, undertaken in 1999-2000) suggest a gross primaryenrollment rate of 61% and a netrate of only 52.5% in 1999-2000. Figure IV.9 shows thepercent of children attendingschool and primary school bysingle-year age groups. Thisproportion rises from 67% at age6 to 84% at age 11, and thenstarts declining. The percentageof children who report attendingprimary school (grades 1-5) is45% at age 6, increases to 66% atage 9, and then declines to 35%at age 11. The figure indicatesthat, even at the peak attendanceages of 9-11 years, nearly 15% ofthe population does not attendschool. This is indicative of alarge class of children who neverattend school.

74

Attaining the Millennium Development Goals in India

0

10

20

30

40

50

60

70

80

90

6 7 8 9 10 11 12 13 14 15 16 17 18Age (years)

Attending any level

Attending primary

Figure IV.9 : School and primary school attendance rates,by age, 1999-2000

67 In what follows, the terms ‘attendance’ and ‘enrollment’ are used interchangeably.

TrendsDespite the grim situation,

there was some progress sincethe early 1990s. The 50th roundof the NSS (undertaken in1993-94) indicates that therewas an increase of about 7-10%in the age-specific enrollmentrate for ages 6-11 between1993-94 and 1999-2000(Figure IV.10). Additionally, thenet primary attendance rate for6-11 year olds also increased,albeit by not very much, overthe same period. The decline inthe net primary attendance rateobserved for 13-15 year olds is apositive development insofar asit reflects a reduction in the rateof over-age enrollment inprimary schools.

Inter-state differencesThere are large differences

across states in the primaryattendance rate (Figure IV.11).Attendance rates for the agegroup 6-11 exceed 90% in 9 states – Kerala, Tamil Nadu,

Maharashtra, Goa, HimachalPradesh, and the states of theNortheast. At the other end, theprimary attendance rates areonly 75% or lower in Bihar,Orissa, Rajasthan, UttarPradesh, and Madhya Pradesh.With only 53% of childrenaged 6-11 attending school,Bihar ranks as the poorest-performing state in the country.

Figure IV.11 also shows thevery large disparity between theage-specific and the net primaryattendance rate in all the states.Even in the states having highoverall primary attendancerates, net primary attendancerates are significantly lower. Forinstance, Himachal Pradeshand Kerala, which have morethan 95% of children aged 6-11attending school, have a netprimary attendance rate of only68% and 61%, respectively. InBihar, a mere 28% of childrenaged 6-11 attend primaryschool! The pervasively low netprimary attendance rate(relative to the age-specific rate)is largely the result of late entryin primary school. For instance,the NSS 55th round dataindicate that, in the country as awhole, 25% of 7-year old, 22%of 8-year old, and 15% of 9-year old school attendees wereactually attending pre-primary(instead of primary) school in1999-2000.

75

Primary Schooling

7

10

79 9

7 7

3

-2

11

7

20

4

98

9

13

7

11

-6

-14

12

-30

-20

-10

-35

-25

-15

-5

0

5

10

15

-31

6 7 8 9 10 11 13 14 15 16 17 18

Age (years)

% change in age-specific attendance rate

% change in net primary attendance rate

Figure IV.10 : Percent change in age-specific school attendance and net primary attendance rates, 1993-94 to 1999-2000

20 30 40 50 60 70 80 90 100

BiharArunachalPradesh

MadhyaPradeshUttarPradesh

RajasthanOrissa

WestBengalAssam

KarnatakaJammu&

AndhraPradeshGujarat

HaryanaManipur

MeghalayaPunjab

GoaTripura

MizoramMaharashtra

NagalandTamilNadu

SikkimKerala

HimachalPradesh

Netprimaryattendancerate Differencebetweennetprimary&age-specific(6-1 1)attendancerate

Figure IV.11 : Age-specific and net primary schoolattendance rates for 6-11 year olds, by state, 1999-2000

Figure IV.12 shows thechanges across states in the age-specific enrollment rate for 6-11year olds between 1993-94 and1999-2000. In most of the poorstates, such as Rajasthan, MadhyaPradesh, Uttar Pradesh andOrissa, there was a large increasein the proportion of children aged6-11 attending school. However,Bihar was an exception, with asmall decline in the proportionattending school between 1993-94 and 1999-2000.

Intra-state differencesRegional differences in

the net primary schoolattendance rate are shown inMap IV.1.68 Six regions inthe country – largely inBihar, Jharkhand, Orissa andManipur – had a net primaryattendance rate of less than40% in 1999-2000. Sixteenregions (about one-fifth of all

the regions) had a netprimary attendance rate ofless than 50%. On the otherhand, no region enjoyed a netrate of more than 76%. Evenmore discouraging is the factthat 29 regions (or more thana third of the regions in thecountry) either did not see animprovement in their netprimary attendance ratebetween 1993-94 and 1999-2000, or saw it decline (MapIV.2). Nine regions saw theirnet primary attendance rateincrease by 40% or moreduring the same period.

76

Attaining the Millennium Development Goals in India

1993-94

40

50

60

70

80

90

100

r l t r b i d

A

h h h n a m a

Ja

ir h a a a a m a u

H

h la

Bih

a

runa

chal

Prade

s

Mad

hya

Prade

s

Utta

r Prade

s

Raj

asth

aO

riss

Wes

t Ben

ga

Ass

a

Kar

nata

k

mm

u&

Kas

hm

And

hra

Prade

s

Guj

ara

Har

yan

Man

ipu

Meg

hala

y

Punja G

o

Tripur

Miz

ora

Mah

aras

htr

Del

h

Nag

alan

Tamil

Nad

imac

hal Pra

des

Ker

a

1999-2000

Figure IV.12 : Percentage of children aged 6-11 years attendingschool, by state, 1993-94 and 1999-2000

�������������� �����������

���������

��������������

���������������������������������

Map IV.1 : Regional estimates of net primaryenrollment rates, 1999-2000

68 As with other indicators, such as infant mortality and underweight rates, it is not possible to derive reliable estimates ofenrollment rates at the district level. Hence, only regional estimates are provided here.

Concentration ofOut-of-SchoolChildren Aged 6-11

According to the NSS 55thround data, there were nearly 30million out-of-school childrenaged 6-11 in India in 1999-2000. If the goal of policy is toreduce the absolute number ofout-of-school children, it isimportant to analyze thedistribution of these childrenacross states and sub-nationalunits. The NSS data indicatethat nearly half of all out-of-school children in the countryare found in the two states ofUttar Pradesh and Bihar (FigureIV.13). Madhya Pradeshaccounts for another 11% of all

out-of-school children aged 6-11. School non-attendance inIndia is thus more heavily

concentrated geographicallythan infant deaths or the numberof underweight children.

Disaggregating further, onefinds that school nonattendance isextremely concentrated inrelatively few villages in thecountry. A mere 10% of villages inthe country account for nearlyone-half of all out-of-schoolchildren aged 6-11, while 20% ofvillages account for three-quartersof all out-of-school children(Figure IV.14). Most likely, thesevillages are scheduled tribehabitations that do not have aprimary school of their own andare not within easy walking reachof a primary school. They are alsolikely to be poor villages where theopportunity cost of child labor (interms of agricultural work) is high.

Primary CompletionRate

Getting out-of-schoolchildren into school is only one of the schooling-relatedmillennium development goals.

77

Primary Schooling

% change in net primary enrollment rates, 1993-94 to 1999-2000

% Change

>= 4020 to 3910 to 19

0 to 9 < 0Missing

Map IV.2 : Regional estimates of the % change in the net primaryenrollment rates, 1993-94 to 1999-2000

72

58

8289

94 9798 99

24

6 6 4 4 4 3 2 1 1 1 1 0 0

11 8

23

0

10

20

30

40

50

60

70

80

90

100

Bih

ar

Utta

r Prade

sh

Mad

hya

Prade

sh

Wes

t Ben

gal

Raj

asth

an

And

hra

Prade

sh

Oris

sa

Kar

nata

ka

Mah

aras

htra

Guj

arat

Ass

am

Tamil

Nad

u

Har

yana

Punja

b

J &K

ashm

ir

Del

hi

Ker

ala

Cu

mu

lati

ve

co

ntr

ibu

tio

n(%

)

Cumulative share in nationwide number of out-of-school 6-11 year olds

Share in nationwide number of out-of-school 6-11 year olds

Figure IV.13 : Contribution of 17 states to the nationalnumber of 6-11 year olds out of school, 1999-2000

Another goal is retention ofstudents – viz., to ensure that theentire cohort of children whobegins grade 1 remains in schooluntil grade 5. School retention isan indicator – albeit imperfect –of the quality of schooling. It ispossible that in the rush toexpand access to schooling,policy makers might compromisethe quality of schooling. Thecompromise in quality wouldlikely show up in lower rates ofstudent retention and primaryschool completion.

Calculating the true primaryretention rate or completion raterequires longitudinal data onchildren, but in the absence ofsuch data, one can use householdsurvey data on children’s ever-schooled, currently-in-school,

and current grade status. TheNSS 55th round data do notreport the specific grade whichchildren attended at the timeof the survey; however, thesurvey reports the current andcompleted schooling level ofeach child.

We use the aboveinformation to calculate theprimary completion rate forchildren aged 12 years.Obviously, 12-year olds whonever attended school areexcluded from the calculationof the primary completion rate.A child is considered to havecompleted primary school ifhe/she reported being inmiddle school at the time ofthe survey and if he/she wasnot reported as never havingattended school. Nationally, in1999-2000, the primarycompletion rate, as calculatedabove, was 61.4% in 1999-2000 – only slightly up from58.7% in 1993-94.69,70

There are large variationsin the primary completionrate across states (FigureIV.15). At 92.1%, Keralahas the highest primary

78

Attaining the Millennium Development Goals in India

69 Increasing the potential pool of children from age 12 to those aged 12-14 years does not make a large difference to theestimated primary completion rate. In 1999-2000, the estimated primary completion rate goes up from 61.4% to 65% whenthe age group 12-14 years is considered.

70 Another, more widely-used method of calculating the primary completion rate is to compare the size of the first grade cohort in a givenyear with that of the sixth grade cohort five years later. Applying this method to administrative school data, reported in the variousissues of Selected Educational Statistics, for 1994-95 and 1999-2000, one also obtains a primary completion rate of about 60%.

0 10 20 30 40 50 60 70 80 90 100

Cumulative % of villages or districts (ranked by number of out-of-school 6-11 year olds)

0

10

20

30

40

50

60

70

80

90

100

Cu

mu

lati

ve

%o

fal

lo

ut-

of-

sch

oo

l6

-11

yea

ro

lds

in

the

cou

ntr

y Districts Villages

Figure IV.14 : Cumulative distribution of all out-of-school 6-11 yearolds in India across villages and districts, 1999-2000

30 40 50 60 70 80 90

KeralaGoa

MaharashtraKarnataka

Tamil NaduMizoram

Himachal PradeshAssam

GujaratAndhra Pradesh

PondicherryNagaland

Jammu & KashmirOrissaPunjab

ManipurWest Bengal

HaryanaBihar

Uttar PradeshMadhya Pradesh

RajasthanArunachal P.

MeghalayaTripuraSikkim

Figure IV.15 : Primary completion rate (%) by state, 1999-2000

completion rate, followed by Goa, Maharashtra,Karnataka and Tamil Nadu.The Northeastern states andBihar, Madhya Pradesh,Uttar Pradesh and Rajasthanrank at the bottom, withprimary completion rates ofabout 50% or lower. Theintra-state differences in the level of the primarycompletion rate are shown inMap IV.3, while percentchanges in the rate between1993-94 and 1999-2000 areshown in Map IV.4.

Socio-economicDifferences inPrimary SchoolAttendance andCompletion

As would be expected,there are large differences in the primary attendanceand completion rates acrossconsumption quintiles. Inthe top consumptionquintile, nearly 94% of all children aged 6-11 attend school, while thecorresponding proportion inthe bottom quintile is only64% (Figure IV.16). The fact that even the topconsumption quintile hasgenerally low net and grossprimary attendance rates andlow primary completion ratesis rather surprising. Itsuggests that late entry intoprimary school and overageenrollment are endemic andnot necessarily specific tolow-income groups.

79

Primary Schooling

Primary completion rates, 1999-2000

Completion Rates

>= 8070 to 79 60 to 69 50 to 59 45 to 49 0 to 44

Missing

Map IV.3 : Regional estimates of the primary completion rateamong 12-year olds, 1999-2000

% change in primary completion

rates, 1993-94 to 1999-2000

% Change

>= 3020 to 2910 to 19 0 to 9

-19 to -1<= -20Missing

Map IV.4 : Regional estimates of the % change in the primarycompletion rate among 12-year olds, 1993-94 to 1999-2000

There are also largedifferences in primary schoolattendance and completionrates across social groups, withscheduled tribes having the lowest attendance andcompletion rates of all groups,followed by scheduled castesand other backward castes(Figure IV.17).

There appears to be athreshold effect in therelationship between schoolattendance rates and adultfemale schooling (FigureIV.18).71 The data suggest thatthe major difference in schoolattendance rates occursbetween households in whichthe highest-educated adultfemale has no schooling at alland those in which the highest-educated female has someschooling. The number of years

of schooling of the adultfemale, conditional on herhaving some schooling, doesnot appear to make much of adifference to the age-specific,gross or net primary attendancerates of children in herhousehold.

However, the situation isdifferent with respect to primarycompletion. In this case, thenumber of (non-zero) years ofschooling of the adult female in the household makes an important difference. Inhouseholds where the adultfemale has 1-5 years of schooling, the primarycompletion rate is 66%, but itincreases to more than 80% withmore than 8 years of schooling.

Infrastructure andSchooling

The NSS data also suggestthat access to infrastructure is associated with higherprimary school attendanceand completion rates.Districts having greaterelectricity coverage not onlyhave higher age-specific,gross and net primaryattendance rates but also

80

Attaining the Millennium Development Goals in India

Fourth

64

74

81

88

94

51

60

6567 68

44

51

5657

59

51

55

59

66

74

40

50

60

70

80

90

100

Poorest Second Third Richest

Per capita consumption expenditure quintile

Age-specific attendance rate (%)

Gross primary attendance rate (%)

Net primary attendance rate (%)

Primary completion rate for 12-year olds (%)

Figure IV.16 : School attendance rates (ages 6-11) and primarycompletion rate for 12-year olds, by per capita

cons. exp. quintile, 1999-2000

67

73

77

85

55

5860

65

4850

52

56

51

55

61

67

40

50

60

70

80

90

Scheduled tribe Scheduled caste Other backward caste Forward caste

Social group

Age-specific attendance rate (%)

Gross primary attendance rate (%)

Net primary attendance rate (%)

Primary completion rate for 12-year olds (%)

Figure IV.17 : School attendance rates (ages 6-11) and primarycompletion rate for 12-year olds, by social group, 1999-2000

71 The NSS data do not permit identification of the mother of each child in the sample (unless the sample is restricted tobiological children of the household head). We have therefore used the schooling years of the highest-educated adult femalein a household as a proxy for maternal schooling.

higher primary completionrates (Figure IV.19). Theevidence is somewhat moremixed with respect to accessto roads. Primary schoolattendance does not varymuch with better road access.However, better road accessdoes seem to improveprimary completion rates(Figure IV.20).

Teachers andPrimary CompletionRates

Since the pupil-teacher ratiois often considered an indicatorof school quality, one mightexpect this ratio to be associatedwith primary completion rates.However, the NSS 55th rounddata do now show anysignificant association between

the two variables. While theprimary completion rate for 12-year olds falls as the number ofpupils per teacher goes frombelow 40 to between 41 and 50,it is greatest in districts wherethe pupil-teacher ratio is above50. The lack of association mayreflect other factors at work,such as teacher absenteeism orthe socioeconomic backgroundof students.

One major programlaunched by the Government ofIndia to address quality issueswas Operation Blackboard.Under this program, theGovernment provided a secondteacher to all single-teacherprimary schools and a teaching-learning equipment packet toall primary schools. Between1987, when the programbegan, to 1994, when all theoriginally-targeted schools had been served, OperationBlackboard accounted for overhalf of central governmentspending on elementaryeducation and extended tovirtually every district of thecountry. On average, OperationBlackboard added two teachersfor every 1,000 primary-school-aged children and increased thestock of primary schoolteachers by 10%.

A rigorous econometricevaluation of Operation Black-board, using multiple rounds ofthe National Sample Surveys, a census of school resources (the All-India Educational Survey),and Operation Blackboardadministrative data, found that,

81

Primary Schooling

63

7780

87

50

66 6563

41

5658

56

50

54

61

69

30

40

50

60

70

80

90

100

0-25 26-50 51-75 >75

% of population in district having electricity access

Age-specific attendance rate (%)

Gross primary attendance rate (%)

Net primary attendance rate (%)

Primary completion rate for 12-year olds (%)

Figure IV.19 : School attendance rates (ages 6-11) and primarycompletion rate for 12-year olds, by access to electricity, 1999-2000

69

9194 96 95 97

56

7167 67

62 62

48

61 61 6258 58

52

66

79

8481

83

40

50

60

70

80

90

100

110

120

0 1-5 6-8 9-10 11-12 >12

Schooling years of highest-educated adult female in household

Age-specific attendance rate (%)

Gross primary attendance rate (%)

Net primary attendance rate (%)

Primary completion rate for 12-year olds (%)

Figure IV.18 : School attendance rates (ages 6-11) and primarycompletion rate for 12-year olds, by female education, 1999-2000

despite substantial misallocationof Operation Blackboard teachersby state and local governments,the program did reduce theprevalence of single-teacherschools and increase the numberof teachers per school (Chin2002). The study also concludedthat the teacher component of Operation Blackboardsignificantly raised primary schoolcompletion and literacy rates forgirls, but there was no such effectfor boys. Girls’ primary schoolcompletion increased by 3-4percentage points, and girls’literacy increased by 2-3percentage points. Thus,Operation Blackboard not onlyinduced illiterate girls whootherwise would never haveattended school to attend, but alsoinduced literate girls whootherwise would have droppedout to stay in school. Thus, thestudy suggests that improvementsin school quality – in particular,the availability of additionalteachers in schools – may be a

viable way of retaining girls inprimary school.

Teacher payTo what extent does teacher

pay influence teacherperformance and studentachievement? Using data from asample survey of 902 studentsand 172 teachers across 20government-funded and 10private schools in India,Kingdon and Teal (2002)explore the evidence for thepayment of performance-related pay and the extent to which such pay structures impact on studentachievement. They find that,even after controlling forstudent ability, parentalbackground, and indicators ofteacher and school quality,students in private schools getsignificantly better academicresults (as measured by tests ofnumeracy and literacy) whenteacher pay is related to studentachievement. However, this

association is not observed forgovernment schools, reflectingthe fact that since governmentteaching jobs in India aretypically permanent contractswith virtually no power ofdismissal, higher wages havelittle influence on teachermotivation in the public sector.On the other hand, theflexibility that the privateeducation sector has to setwages and dismiss lax teachersallows school managers to useteacher wages as incentives toenhance teacher performance.

Teacher absenteeismSince retention of students

in primary schools dependsupon the type and caliber ofinstruction they receive,primary completion rates arelikely to be more sensitive toteacher absenteeism thanenrollment rates. In a study of schools in Uttar Pradesh, Dreze and Gazdar(1997) argue that “… thespecific problem of teacherabsenteeism and shirking … isby far the most importantissue of education policy inUttar Pradesh today.” Drezeand Gazdar came to thisconclusion after finding thattwo-thirds of the teachers wereabsent from their positionsduring announced visits to 16schools in the state. A study byPROBE (1999), based on visitto over 200 primary schoolsacross the country, found thatthe head teacher (or principal)was absent in nearly a third ofthe schools on the day of thesurvey team’s visit. A recent

82

Attaining the Millennium Development Goals in India

pucca road

75 7576

85

60 6160

62

51 52 51

55

58 57

61

69

45

50

55

60

65

70

75

80

85

0-40 41-65 66-90 >90

% of villages in district having access to a

Age-specific attendance rate (%)

Gross primary attendance rate (%)

Net primary attendance rate (%)

Primary completion rate for 12-year olds (%)

Figure IV.20 : School attendance rates (ages 6-11) and primarycompletion rate for 12-year olds, by access to pucca roads, 1999-2000

national survey focusing onservice delivery in primaryhealth centers and primaryschools found that, onaverage, one-quarter of India’sprimary school teachers werenot present at the schoolswhere they are supposed toteach on any given day72 –rates that are high incomparison to those observedin other countries (WorldBank 2003). The problem ofabsenteeism appears to bemuch worse in the poorerstates, with 39% of primaryschool teachers being absenton any given day in Bihar(Chaudhury et al. 2003).

One reason for the highrates of teacher absenteeismmay be extensive involvementof teachers in politics. Kingdonand Muzammil (2003) cite thecase of Uttar Pradesh, wherethere has been significantpenetration by teachers intostate politics, in large part dueto a constitutional provision fortheir reserved representation inthe upper house of the state

legislature.73 This has led tosignificant lobbying and unionactivity by the teachers and hasearned them substantial salarygains and other monetarybenefits from the stategovernment. The salary gainshave crowded out publicspending on other items of expenditure, such asinstructional materials andsupplies and scholarships.74 Inaddition, the increasedparticipation of teachers in

regional and local politics hasdiverted them from theirteaching responsibilities, andmay in part be the reason forthe high rate of absenteeism.Of course, the problem ofteacher absenteeism goes farbeyond the politicization ofteachers; it reflects a general lack of accountability of teachers and schools –indeed, the entire educationalsystem – to students and the community.

83

Primary Schooling

12

34

56

78

910

1112

13 13

1516

1718

1920

2122

Tamil

Nad

u

Guj

arat

Kar

nataka

Mah

aras

htra

And

hra Pr

ades

h

Mad

hya Pr

ades

h

Utta

r Prade

sh

Ker

ala

Rajas

than

Bihar

Wes

t Ben

gal

Sikki

m

Ass

am

Oris

sa

Har

yana

Him

acha

l Pra

desh

Mizor

am

Nag

alan

d

Punjab

Meg

halaya

Tripur

a

Aru

nach

alPra

desh

Figure IV.21: Ranking of states in the performance of governmentschool education services, 2002

72 The absence rate is the percentage of staff that is supposed to be present but is not on the day of an unannounced visit. Itincludes both authorized and unauthorized absences. In fact, most teacher absences in schools may be authorized absences(Howes and Murgai 2004). For example, the PROBE study observed that of the 200 days of potential teacher attendanceduring a year, teachers had valid reasons to be absent from school for 50 days (or 25%): 20 days of holidays and permittedleave, 21 days of non-teaching duties (including deputation and in-service training), and 9 days for collection of salaries(PROBE 1999). Thus, the high absence rate likely reflects generous leave terms and significant non-teaching duties assignedto teachers. Whether teacher absences are authorized or unauthorized, they reflect the general lack of accountability of thepublic school system to the students they are supposed to serve. There is evidence that the amount of academic instructionaltime in a school influences student outcomes, especially in government schools. The study by Kingdon and Teal (2002)mentioned above found that a one percent increase in the minutes of academic instruction per week in a government schoolincreased student test scores in that school by almost 0.5 percent. This would seem to suggest that teacher absences, whetherauthorized or unauthorized, are likely to have a large deleterious impact on student achievement.

73 Government school teachers are barred from contesting for elections to the lower house of the state legislature. However,teachers in government-aided, private schools are not similarly debarred from elections.

74 It is estimated that a Rupee spent on providing a full packet of instructional materials in schools improves test scores by 14times as much as a Rupee spent on teacher salaries and 12 times as much as a Rupee spent on school facility improvement.

In contrast to Chaudhury etal. (2003), a national surveyconducted by the Public AffairsCentre, which ranked theperformance of governmentschool education services on anumber of indicators, such asaccess to a government primaryschool within a kilometer fromthe place of residence, propor-tion of households sendingtheir children to governmentprimary schools, and usersatisfaction with the behaviorof teachers and with the qualityof school facilities, found thatthe poor states (with theexception of Orissa) did notquite rank at the bottom of allthe states in the country. TheNortheastern states, Punjab,and Haryana ranked lowerthan Bihar, Madhya Pradesh,and Uttar Pradesh in terms ofthe quality of their publicschool education services(Figure II.21) (PAC 2002).75

Reducing teacher absentee-ism and making schoolsaccountable to students and thecommunity is no simple task, however. As the WorldDevelopment Report 2004 pointsout, it requires broad-ranginginstitutional reform, incorporat-ing, among other things,empowerment of citizens andcommunities who can hold the state accountable forperformance, devolution ofadministrative and financial

powers to communities, greater autonomy to schools,involvement of parents in schoolmanagement, and ensuring themotivation of front-line workers.A good example of anintervention that seeks tomotivate front-line providers toimprove the delivery of services is the Learning GuaranteeProgram, launched in the poor and backward districts of Karnataka by the state government and the Azim Premji Foundation. Theprogram aims to improve accessand learning outcomes byintroducing competition amonggovernment schools forperformance-related awards (SeeBox IV.2 for a description of theprogram).

Multivariate Analysisof PrimaryAttendance andCompletion

To examine the likelihood ofthe various states in Indiaattaining the child schoolingMD goals, we have estimated amultivariate model of schoolattendance and primaryattendance for 6-11 year oldsand primary completion for 12-year olds, using the NSS 55thround unit record data (at thechild level).76 The multivariatemodel has the advantage ofcontrolling for several variablesthat are simultaneouslyassociated with child schooling.

The estimation results arereported in Annex Tables 4 and5, while only the broad findingsof the empirical analysis arediscussed here.

The multivariate modelconfirms many of the bivariaterelationships shown in thegraphs above. While theprobability of attending schoolor primary school increases withage, girls at each age have asignificantly lower probabilityof school and primary schoolattendance. Further, the genderdisparity in attendance ratesappears to increase with age.Since primary completion isconsidered only for 12-yearolds, age comparisons inprimary completion are notrelevant. However, the resultsdo suggest the importance ofgender, with 12-year old girlsbeing nearly 4% less likely tocomplete primary school than12-year old boys.

The schooling of thehighest-educated adult male orfemale in a household is significantly associated with school attendance andcompletion – more so than theschooling of the householdhead or the head’s spouse.Adult female schoolingtypically has a smallerassociation than adult maleschooling with school andprimary school attendance, but

84

Attaining the Millennium Development Goals in India

75 This is in contrast to the finding of the same (PAC) survey that the poor states ranked in the bottom half of all the states inthe country in terms of the quality of their public health services.

76 Since the dependent variables in all three models are dichotomous (i.e., assuming a value of zero or one), the models have beenestimated by the maximum-likelihood probit method.

the opposite is true withrespect to primary completion.

Even after controlling for logof monthly consumptionexpenditure per capita,historically-disadvantaged socialgroups, such as scheduled castes,scheduled tribes and otherbackward castes, havesignificantly lower attendanceand completion probabilitiesthan forward castes, withscheduled tribes being generallythe worst-off group. Likewise,Muslims consistently have lowerattendance and completion ratesthan non-Muslims. Thedifferences across social andreligious groups are large; forinstance, children aged 6-11belonging to scheduled tribeshave a 16% lower probability ofattending school than childrennot belonging to scheduledtribes, and Muslim children aged12 years have a 21% lowerprobability of having completedprimary school than non-Muslim children.

A household’s livingstandards, as proxied by the log of its monthlyconsumption expenditure percapita, are strongly associatedwith school attendance and primary completion.Likewise, the log of grossdomestic product per capita

in an individual’s state ofresidence has significantpositive associations with allthree measures of schoolenrollment and completion.Even after control for grossstate domestic product percapita, the log of governmentexpenditure on elementaryschooling per child 6-14 years in a state is strongly associated withprimary school attendanceand primary completion butnot with overall schoolattendance.77

We have merged data onthe number of cognizablekidnappings of women andgirls per capita in a districtwith the NSS householdsurvey data to examine theassociation between crimeagainst women and theschool attendance andcompletion rates of girls. Theempirical results suggest thatcrime against women isassociated with a significantdecline in female schoolattendance, female primaryschool attendance, andprimary completion rates.The magnitude of theestimated coefficients are,however, not overly large.

Infrastructure haspredictable associations with

school attendance andcompletion, with access toelectricity having the strongestassociations. Access to roads isnot significantly associatedwith school and primaryschool attendance among 6-11year olds, but it is significantlyassociated with the probabilityof children completingprimary school.

An interesting question isthe extent to which thequantity and quality ofschool infrastructure in acommunity are associatedwith school attendance and completion rates. The availability of primaryschools per 1,000 childrenaged 6-11 in a district – anindicator of schoolingquantity – is stronglyassociated with school andprimary school attendanceamong 6-11 year olds, but ithas no significant associationwith primary completionrates. On the other hand,lowering the pupil teacherratio at the primary levelin a district – an indicatorof increased schoolingquality– is associated with higher rates of school attendanceand primary completion.This suggests that school attendance iscurrently constrained by the

85

Primary Schooling

77 As an alternative, state fixed-effects models were estimated, wherein a set of 23 state dummy variables replaced the two state-level variables (viz., log of gross state domestic product per capita and log of state government elementary expenditure per child6-14). Although the state dummy variables were significant as a whole at the 5% level, the explanatory power of theregressions, as measured by a pseudo R-squared measure, did not increase much with the substitution of the state fixed effectsfor the state-level variables. The pseudo R-squared measures increased from 0.20 to 0.21, 0.06 to 0.07, and 0.09 to 0.13 inthe case of school enrollment, primary enrollment and primary completion, respectively. This suggests that the state fixedeffects models are not superior to the models reported here in terms of goodness-of-fit.

availability of primaryschools,78 while both schoolattendance and primarycompletion would likelybenefit from school qualityimprovements in the form ofa reduction of the pupilteacher ratio.79,80

Simulations to 2015Based on the multivariate

probit models discussedabove, we have undertakensimulations of the schoolattendance, primary schoolattendance and the primarycompletion rate for the poorstates under differentintervention scenarios. Theseare shown in Table IV.1 below.If the poor states were simplybrought up to the nationalaverage in terms of the nineinterventions considered –viz., better road and electricity access, increasedadult male and femaleschooling, household income(consumption) growth, growthin public spending onelementary education perchild, reduced crimes againstwomen, expanded number ofprimary schools per 1,000children, and reduction in thepupil teacher ratio at the

primary level – the cumulativeincreases in the schoolattendance, primary schoolattendance, and primarycompletion rates would be7.3, 7.1, and 6.9 percentagepoints (or 10%, 14.3%, and12.4%), respectively. If the magnitude of theinterventions were to be scaledup, so as to bring the poorstates to the average level ofthe non-poor states, thecumulative increase in schoolenrollment and completionwould be significantly larger.The school attendance,

primary school attendance,and primary completion ratesin the poor states wouldincrease by 20.7, 19.8, and20.6 percentage points (or28.4%, 39.8%, and 38.3%),respectively. The largestincreases in school attendanceand primary completion areobtained with improved livingstandards, expanded access toelectricity, and increasedgovernment expenditure onelementary schooling.

We have also undertakensimulations of the increase in

86

Attaining the Millennium Development Goals in India

0

5

10

15

20

25

30

35

40

45

50

19

99

20

00

20

01

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

Reducing the primary pupil teacher ratio

Expanding the number of primary schools per 1,000 children aged 6-11Reducing crime against women

Growth in hh cons exp per capIncreasing adult female schooling

Increasing adult male schoolingExpanding electricity access

Intervention

Increase in enrollment rate needed to

attain universal primary enrollment

Figure IV.22 : Increase in projected % of children aged 6-11attending school in the poor states, 1999-2015,

under different intervention scenarios(graph shows cumulative effect of each additional intervention)

78 Naturally, given the linear prediction, this result would hold only up to some limit.79 Lanjouw and Ravallion (1999) find that government educational expenditures that expand access are better targeted to poor

people than resources that exclusively raise quality, since poor students are almost always the last to be enrolled and the better-off are typically already in school. A study for Kenya that looked explicitly at quantity (number of schools) and quality(teacher-pupil ratio) improvements also found that an expansion in the number of school facilities increased the enrollmentof children in the poorest expenditure quintiles but had no impact on the enrollment of children in the top quintiles. On theother hand, an improvement in the teacher-pupil ratio increased the enrollment rate of children in the top quintiles, butactually reduced the enrollment of children in the poor quintiles (Deolalikar 1998).

80 Another study by Duflo (2001) found that the construction of 61,000 primary schools by the Indonesian Government between 1973and 1978 – one of the largest school construction programs on record – led to a large increase in schooling. On average, children aged2-6 in 1974 received 0.12-0.19 more years of schooling for each school constructed per 1,000 children in their region of birth.

87

Primary Schooling

Table IV.1 : Projected increase in school attendance and primary completion rates(percentage points) with various interventions in the poor states

Bringing poor states to the level of the Bringing poor states to level of the non-national average poor states’ average

Primary PrimaryIntervention School school Primary School school Primary

attendance attendance completion attendance attendance completion

Male schooling (years) 0.5 0.3 0.1 1.5 1.0 0.9

Female schooling (years) 0.6 0.2 0.6 1.8 0.6 2.1

Monthly per capita consumption expenditure 3.4 3.6 2.9 10.6 11.1 9.0

Annual government expenditure on elementaryeducation per child 6-14 years 0.5 1.3 1.4 3.6

Number of cognizablekidnappings of womenand girls per 100,000population in district 0.0 0.0 0.0 0.1 0.1 0.3

% of villages in districtconnected by pucca road 0.3 0.7

Access to electricityin district (%) 1.6 1.1 1.2 4.9 3.7 3.4

Number of primary(grades 1-5) schoolsper 1,000 children aged 6-11in district (x 1000) 0.9 0.6 1.4 0.9

Pupil teacher ratio inprimary schools (grades 1-5)in district (x 1000) 0.4 0.7 0.4 0.4 1.0 0.6Sum of all interventions 7.3 7.1 6.7 20.7 19.8 20.6

Starting Assumed Endingvalue in change per value in

Intervention 1999-2000 year 2015

Adult male schooling (years) 4.5 0.25 8.5

Adult female schooling (years) 2.0 0.3 7.8

Government expenditure on elementaryeducation per child aged 6-14 (Rs.) 955 4% 1,789

Consumption expenditure per capita (Rs.) 409 3% 656

Population coverage of electricity (%) 42.5 1% point 58.5

Crime against women (no. of femalekidnappings and rapes per 100,000 pop.) 1.65 -0.05 0.85

No. of primary schools per 1,000 children aged 6-11 5.1 0.2 8.3

Pupil teacher ratio in primary schools 91 -1.0 75

Table IV.2 : Assumptions about various interventions to increase the school attendance andthe net primary school attendance rate in the poor states, 1999-2000 to 2015

88

Attaining the Millennium Development Goals in India

the school attendance and thenet primary school attendancerate in the poor states of thecountry under the assumptionthat eight interventions arepursued simultaneously andgradually every year up to2015. The nature andmagnitude of these assumedinterventions are shown inTable IV.2. As noted in earlierchapters, the policy scenarioassumed for the simulations ishypothetical. Figure IV.22shows changes in the projectedschool attendance rate for 6-11year olds in the poor states to2015, under the assumptionthat all of the interventions arepursued simultaneously. Theschool attendance rate isprojected to increase by 31.5percentage points by 2015 –enough to attain universalprimary enrollment. Most ofthe interventions, with theexception of a reduction in theincidence of crimes against

women and a reduction in thepupil-teacher ratio in primaryschools (both of which havestatistically significant butnumerically small associa-tions), are associated withappreciable increases in theschool attendance rate.

However, the situation isvery different with respect to thenet primary attendance rate.Figure IV.23, which shows theprojected increases in the netprimary attendance rate in thepoor states, shows that rateincreasing by only about 27percentage points by 2015 –well short of the 50 percentagepoints needed to attain the MDgoal. The relative ranking ofdifferent interventions in raisingthe net primary attendance rate is similar to that ofinterventions to raise the overall school attendance rate,with the exception of per-childgovernment expenditure onelementary schooling, whichhas a significant associationwith the net primary attendancerate but not with the overallschool attendance rate.

These results suggest that whileit may be possible to get all children

19

99

20

00

20

01

20

02

20

03

20

04

20

5

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

0

0

5

10

15

20

25

30

35

40

45

50

55

60

Reducing the primary pupil teacher ratioExpanding the number of primary schools per 1,000 children aged 6-11Reducing crime against womenGrowth in hh cons exp per capIncreasing adult female schoolingIncreasing adult male schoolingExpanding electricity accessIncreasing gov't exp on elementary schooling per child 6-14

Intervention

Increase in net primary enrollment rate needed to attain universal net primary enrollment rate

Figure IV.23 : Increase in projected % of children aged 6-11attending primary school in the poor states, 1999-2015,

under different intervention scenarios(graph shows cumulative effect of each additional intervention)

0

5

10

15

20

25

30

35

40

45

19

99

20

00

20

01

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

Reducing the primary pupil teacher ratioReducing crime against womenImproving road accessGrowth in hh cons exp per capIncreasing adult female schoolingIncreasing adult male schoolingExpanding electricity accessIncreasing gov't exp on elementary schooling per child 6-14

Intervention Increase in primary completion rate needed to attain MDG

Figure IV.24 : Increase in projected primary completion rate (%) inthe poor states, 1999-2015, under different intervention scenarios

(graph shows cumulative effect of each additional intervention)

89

Primary Schooling

aged 6-11 in the poor states intoschool with concerted action onmany fronts, raising the net primaryenrollment or attendance rate to100% by 2015 will be extremelychallenging. Of course, the relativeimportance of ensuring that allchildren aged 6-11 are in schoolversus ensuring that all childrenaged 6-11 are attending primaryschool is debatable. Certainly, forthe poor states of India, evenensuring that all children aged 6-11are in school (irrespective of level)will be a worthwhile and laudatoryachievement.

The findings of a comparableanalysis of primary completionrates in the poor states of thecountry are depicted in FigureIV.24. The interventions consid-ered for primary completion arelargely the same as those forprimary enrollment, with twoexceptions. First, as the numberof primary schools per 1,000children aged 6-11 is notsignificantly associated withprimary completion, it isdropped from the projectionsshown in Figure IV.24. Second,as the availability of village roadsis significant in explainingprimary completion (but notprimary enrollment), it isincluded in the projections. Weassume that the percentage ofvillages connected to a puccaroad in the poor states increasesby one percentage point eachyear – from 60% in 1999 to76% in 2015.

As observed in Figure IV.24,the net primary completion

rate is projected to increase byabout 29 percentage points by2015, which is well short of the46 percentage point increasethat would be needed to attainthe MD goal of 100% primarycompletion, but neverthelessrepresents significant progress.The largest improvements inthe completion rate occur withan increase in living standards,followed by an expansion offemale adult schooling andpublic spending on elementaryeducation per child 6-14.Increases in adult maleschooling and improvedelectricity coverage also areassociated with smaller, butsignificant, associations withthe primary completion rate.

The main implication of thefindings presented in thischapter is that attaining theMDGs relating to universal netprimary enrollment and 100%primary school completion willbe extremely challenging in thepoor states. Nevertheless, largegains in both of these indicatorsare possible, given economicgrowth, improved infrastruct-ure, and increased publicspending on elementaryeducation. In addition, the samepackage of interventions shouldbe successful in getting allchildren aged 6-11 yearsattending school.

Why do the simulationssuggest attainment of theeducation-related MDGs to be amore challenging task than the attainment of the

infant mortality- and childunderweight-related MDGs? Inpart, this reflects the largediscrepancy between the goalsand the current status of thevarious MD indicators in thepoor states of the country. Thenet primary enrollment rate inthe poor states was only about50% in 1999-2000. To attain universal net primaryenrollment, the poor states in thecountry would have to doublethe net primary enrollment rateby 2015. In contrast, the infantmortality-related MDG calls fora reduction of about 60% ininfant mortality in the poor states(from a level of 76 infant deathsper 1,000 live births in 1998-99to a level of 31 by 2015).Likewise, the underweight-related MDG calls for theproportion of underweightchildren in the poor states to bereduced by about 46% by 2015 –from a level of 51% in 1998-99to 27% by 2015. Thus, theabsolute nature of the education-related MDGs (viz., the targetbeing unrelated to the startingpoint) creates a large challenge forthe poor states in the countrythat have very low rates of netprimary enrollment and primarycompletion.

As part of its commitment to universalizing access toelementary schooling andensuring completion ofelementary schooling by 2010,the Government of India haslaunched the National Programof Universal ElementaryEducation (UEE), known in

Hindi as Sarva Shiksha Abhiyan(SSA). (See Box IV.3 for adescription of the SSA.) The Program not only isconsistent with the MillenniumDevelopment Goals (MDGs)for education, but goes beyond itin covering 8 years of primaryschooling, under a much tightertime frame. Among thequantitative goals of the SSA are:all children to complete five yearsof primary schooling by 2007; allchildren to complete eight yearsof schooling by 2010; no genderand social disparities in primary

schooling by 2007 and inelementary schooling by 2010;and universal retention ofchildren in elementary schoolsby 2010. The SSA is a centrally-sponsored scheme (CSS)whereby the Union Governmentprovides incremental resources ina cost-sharing arrangement withthe states: 85:15 in the 9th Plan(FY 2001), 75:25 in the 10thPlan (FY 2002-06), and 50:50 inthe 11th Plan (FY 2007-10).The SSA anticipates that Rs. 600billion of additional resourceswill be required from the budgets

of the central and stategovernments over the next 10years to attain the SSA goals.

The simulations shown inFigure IV.22-IV.24 suggest thatit will be very difficult to meetthe government’s goal of havingall children in India enrolled inprimary schools and completingthe full course of primaryschooling by 2007. The goal ofgetting all children aged 6-14years to complete eight years ofschooling by 2010 would appearto be even more challenging.

90

Attaining the Millennium Development Goals in India

Madhya Pradesh is one of the poorest states in India. It also has one of the lowest levels of school enrollment inthe country. In order to remedy this situation and as part of its wide-ranging decentralization program, thegovernment of Madhya Pradesh began the Education Guarantee Scheme in 1997 to ensure access to primaryschooling to every child in the State. The EGS represented a community-centered and rights-based approach touniversalizing primary education in a quick, time-bound manner. Under the EGS, the Government gave aguarantee to provide a primary school facility to any habitation which did not have such a facility within a distanceof a kilometer. Further, the state government was obligated to provide the school within a period of 90 days afterreceiving the demand for it from the local community.

The EGS represents a three-way partnership among the state government, local governments (typically the villagecouncil or panchayat), and the community. The community raises the demand, identifies a local resident who couldbe the teacher in the new school, and provides the space for a school facility. The local government (panchayat)appoints the teacher and oversees the functioning of the school. The state government supports the school througha grant to cover the teacher’s salary, arranges for the training of the teacher, and provides teacher-learning materialand other inputs for quality.

Within a year of its introduction, the EGS created a primary school facility in every habitation of the State. Indeed,during 1997 – the first year of the scheme’s operation – there were 40 new primary schools that opened every day,suggesting the enormous unmet demand for schooling that existed in the state. By August 1998, Madhya Pradeshhad a primary school facility in every habitation. Thus, within a short period of 18 months, the state eliminatedthe historical backlog of schooling – at one-third the usual cost of establishing new school facilities. The EGS didnot merely provide physical access; more importantly, it provided social equity, with the new school facilities beinglargely demanded to bring primary schooling to scheduled tribes and girl children. The EGS has become a nationalmodel for a community-based approach to universalizing primary schooling in India.

While there have been no rigorous evaluations of the EGS, some studies suggest that schooling access for thepoorest 40% of households increased more rapidly in Madhya Pradesh between 1992-93 and 1998-99 (“beforeEGS” and “after EGS” years for which national survey data are available) than in the rest of India (McCarten andVyasulu 2003). An econometric evaluation of the central government’s District Primary Education Program(DPEP-I) also found that Madhya Pradesh was the only state (of a total of 7 states considered) which showedsignificant impacts of DPEP-I on school attendance rates among 6-10 year olds (Jalan and Glinskaya 2003).

Box IV.1: The Education Guarantee Scheme of Madhya Pradesh

91

Primary Schooling

In 2002, the Azim Premji Foundation, in collaboration with the Government of Karnataka, launched the LearningGuarantee Program, which is aimed at guaranteeing the learning of children in government schools. Allgovernment higher and lower primary schools (approximately 10,000) situated in the 43 blocks of the northeasterndistricts of Karnataka State (identified as most backward) were eligible to participate in this competition. Between2002-03 and 2004-05, when the program will end, participating schools will have three opportunities to enter thecompetition and improve their capabilities.

The program is essentially a competition, with rewards and recognition schemes for teachers and children ofschools which demonstrate an improvement in student learning. To encourage schools to both expand access andimprove quality of learning, the program rewards schools on the basis of learning outcomes in the school’s habitation(not the school’s population). If a minimum of 60% of all children in the school’s habitation demonstrate, in anindependent assessment, that they have acquired 90% of the prescribed competencies in mathematics andKannada (the regional language of Karnataka), the school would receive an award of Rs. 5,000. With 70% of thechildren in the habitation acquiring the expected learning outcomes, the reward would be Rs. 10,000. Finally, aschool that demonstrates 80% of children in the habitation achieving the expected learning outcomes qualifies forthe highest reward of Rs. 20,000. Thus, a school can win a maximum of Rs. 60,000 in the three rounds of thecompetition.

In the first year of the program, nearly 1,000 schools offered themselves for evaluation. Many more are expectedto opt into the program in the second and third years.

The Learning Guarantee Program is an attempt to encourage government schools to deliver learning outcomes andto make them accountable for the services they deliver. It is expected that the LGP will enable primary stakeholders(parents and the community) to put pressure on their children’s schools to become a certified Learning Guaranteeschool. It is also expected that ‘Learning Guarantee’ schools will motivate out-of-school children to enroll in schooland encourage the children in school to continue (and not drop out). By fostering healthy competition among allcommunities, the program is likely to both improve access and quality.

Source: Azim Premji Foundation (2002).

Box IV.2: The Learning Guarantee Program in Karnataka

92

Attaining the Millennium Development Goals in India

In accordance with the constitutional commitment to ensure free and compulsory education for all children up tothe age of 14 years, provision of universal elementary education has been a salient feature of national policy sinceindependence. This resolve has been spelt out emphatically in the National Policy since independence (NPE), 1986and the Program of Action (POA) 1992. A number of schemes and Programs were launched in pursuance of theemphasis embodied in the NPE and the POA. These included the scheme of Operation Blackboard (OB); NonFormal Education (NFE); Teacher Education (TE); Mahila Samakhya (MS); State specific Basic EducationProjects like the Andhra Pradesh Primary Education Project (APPEP); Bihar Education Project (BEP), LokJumbish (LJP) in Rajasthan; National Program of Nutritional Support to Primary Education (MDM); and theDistrict Primary Education Program (DPEP).

The Sarva Shiksha Abhiyan (SSA) is a historic stride towards achieving the long cherished goal of Universalizationof Elementary Education (UEE) through a time-bound integrated approach, in partnership with States. SSA,which promises to change the face of the elementary education sector of the country, aims to provide useful andquality elementary education to all children aged 6-14 years by 2010.

The SSA is an effort to recognize the need for improving the performance of the school system and to providecommunity-owned quality elementary education in the mission mode. It also envisages bridging of gender andsocial gaps.

Objectives of the SSA� All children in school, Education Guarantee Centre, Alternative School, ‘Back to School’ camp by 2003;

� All children complete five years of primary schooling by 2007;

� All children complete eight years of schooling by 2010;

� Focus on elementary education of satisfactory quality with emphasis on education for life;

� Bridge all gender and social category gaps at primary stage by 2007 and at elementary education level by 2010;

� Universal retention by 2010.

Structure for ImplementationThe Central and State governments will together implement the SSA in partnership with the local governmentsand the community. To signify the national priority for elementary education, a National SSA Mission is beingestablished with the Prime Minister as the Chairperson and the Union Minister of Human Resource Developmentas the Vice Chairperson. States have been requested to establish State level Implementation Society for UEE underthe Chairmanship of Chief Minister Education Minister. This has already been done in many States.

The SSA will not disturb existing structures in States and districts but would only try to bring convergence in allthese efforts. Efforts will be made to ensure that there is functional decentralization down to the school level inorder to improve community participation. Besides recognizing PRIs/Tribal Councils in Scheduled Areas,including the Gram Sabha, the States would be encouraged to enlarge the accountability framework by involvingNGOs, teacher, activists, and women’s organizations.

Coverage and PeriodThe SSA will cover the entire expanse of the country before March 2002 and the duration of the Program in everydistrict will depend upon the District Elementary Education Plan (DPEP) prepared by it as per its specific needs.However, the upper limit for the Program period has been fixed as ten years, i.e., up to 2010.

Strategies Central to SSAInstitutional reforms – As part of the SSA, institutional reforms in the States will be carried out. The state will haveto make an objective assessment of their prevalent education system including educational administration,achievement levels in schools, financial issues, decentralization and community ownership, review of state

Box IV.3: The Sarva Shiksha Abhiyan:A Program for Universal Elementary Education in India

93

Primary Schooling

Education Act, rationalization of teacher deployment and recruitment of teachers, monitoring and evaluation,education of girls, SC/ST and disadvantaged groups, policy regarding private schools, and ECCE. Many Stateshave already implemented institutional reforms to improve the delivery system for elementary education.

Sustainable Financing – The SSA is based on the premise that financing of elementary educationinterventions has to be sustainable. This calls for a long-term perspective on financial partnership between theCentral and the State governments.

Community ownership – The Program calls for community ownership of school-based interventions througheffective decentralization. This will be augmented by involvement of women’s groups, VEC members, andmembers of Panchayati Raj institutions.

Institutional capacity building – The SSA conceives a major capacity building role for national- and state-levelinstitutions like NIEPA, NCERT, NCTE, SCERT, and SIEMAT. Improvement in quality requires a sustainablesupport system of resource persons.

Improving mainstream educational administration – The Program will have a community-based monitoringsystem. The Educational Management Information System (EMSI) will correlate school level data withcommunity-based information from micro planning and surveys. Besides this, every school will have a noticeboard showing all the grants received by the school and other details.

Habitation as a unit of planning – The SSA works on a community-based approach to planning withhabitation as a unit of planning. Habitation plans will be the basis for formulating district plans.

Accountability to community – SSA envisages cooperation between teachers, parents and PRIs, as well asaccountability and transparency.

Education of girls – Education of girls, especially those belonging to the scheduled castes and scheduled tribes,will be one of the principal concerns of the SSA.

Focus on special groups – There will be a focus on the education participation of children from scheduledcastes and tribes, religious and linguistic minorities, disadvantaged groups, and disabled children.

Pre-Project phase – SSA will commence throughout the country with a well-planned pre-project phase thatprovides for a large number of interventions for capacity development to improve the delivery and monitoringsystem.

Thrust on quality – SSA lays a special thrust on making education at the elementary level useful and relevantfor children by improving the curriculum, child-centered activities, and effective teaching methods.

Role of teachers – SSA recognizes the critical role of teachers and advocates a focus on their developmentneeds. Setting-up of BRC/CRC, recruitment of qualified teachers, opportunities for teacher development throughparticipation in curriculum-related material development, focus on classroom process, and exposure visits forteachers are all designed to develop the human resources of teachers.

District Elementary Education Plans – As per the SSA framework, each district will prepare a DistrictElementary Education Plan, reflecting all the investments being made in the education sector, with a holistic andconvergent approach.

Components of SSAThe components of the SSA include appointment of teachers, teacher training, qualitative improvement ofelementary education, provision of teaching learning materials, establishment of Block and Cluster ResourceCenters for academic support, construction of classrooms and school buildings, establishment of educationguarantee centers, integrated education of the disabled, and distance education.

Source: Government of India (2002)

Box IV.3 (cont’d): The SSA

94

Attaining the Millennium Development Goals in India

We use twenty years ofstate-level data on grossprimary enrollment rates(GPERs), real gross statedomestic product per capita(GSDP), and real publicspending on elementaryeducation per child 6-14 yearsto analyze the associationbetween the gross primaryenrollment rate on the onehand and public spending onelementary education andeconomic growth on theother.81,82

The results of this analysisare reported in Annex TableIV.1 below. OLS estimatesindicate a very strong positiveassociation between the grossprimary enrollment rate andgovernment expenditure onelementary education per childeven after control for per capita GSDP, with a onepercent increase in governmentelementary education spendingper child being associated witha 0.17 percent increase in thegross primary enrollment rate.Surprisingly, when control for unobserved heterogeneityacross states – in the form of

time-invariant state fixedeffects – is introduced in themodel, the magnitude of thepositive association increases(from 0.17 to 0.23). Theinclusion of state fixed effects, however, renders theassociation between the grossprimary enrollment rate andper capita GSDP insignificant,indicating that state hetero-geneity is positively associatedwith GSDP. An interactionbetween the log of governmenthealth expenditure and the logof per capita GSDP is negativeand significant in the fixed-effects model, indicating that the positive associationbetween the gross primaryenrollment rate and publicspending on education isweaker at higher levels of percapita state income. It isinteresting that these resultsmirror those obtained forinfant mortality (and reportedin Annex II.1), also using cross-state, time-series data over thesame period.

Unfortunately, time-seriesdata on other covariates ofgross enrollment, such as

female adult literacy, aresimply not available at thestate level for the period1980-99. To rule out thepossibility that the significantassociation between the grossprimary enrollment rate and government educationalexpenditure per child couldbe due to the omission offemale adult literacy from themodel, we have calculatedstate-level averages of theadult female literacy rate fromfour rounds of the NationalSample Survey – for 1983-84,1987-88, 1993-94 and 1999-2000. The inclusion of femaleliteracy means that the grossprimary enrollment rateregression can only beestimated for these four years,which reduces the sample sizefrom 266 to merely 56.However, with one exception,the results continue toindicate a strong positiveassociation between the grossprimary enrollment rate andpublic spending on educationand a negative and significantcoefficient on the interactionbetween public spending andper capita GSDP.

Annex IV.1: Gross Primary Enrollment Rates, Government Elementary Education Expenditure,and Per Capita Income across States, 1980-99

81 Data on government elementary education expenditures were obtained from detailed budget demand documents of individualstates. The expenditure variable includes expenditure incurred by a state out of its own revenues as well as central governmenteducational allocations to that state.

82 A state-level regression of the gross primary enrollment rate on government elementary education spending is based on theassumption that public spending is not endogenously distributed across states (Rosenzweig and Wolpin 1986). Such anassumption may not always hold, especially if government educational expenditures are allocated across states based on literacyconditions. However, the inclusion of state fixed effects in the regression effectively controls for endogenous programplacement. We allow for a central planner to distribute public spending on elementary education across states according to anunobserved education/literacy attribute; in other words, states that have innately low literacy rates may spend more onelementary education than those having innately high literacy. Since the fixed-effects model purges the unobserved (time-invariant) literacy attribute from the model, it effectively addresses the endogenous program placement problem.

95

Primary Schooling

83 A study for Kenya that looked explicitly at quantity (number of schools) and quality (teacher-pupil ratio) improvements also found that an

expansion in the number of school facilities increased the enrollment of children in the poorest expenditure quintiles but had virtually no

impact on the enrollment of children in the top quintiles. On the other hand, the same study found that an improvement in the teacher-pupil

ratio increased the enrollment rate of children in the top quintiles, but actually reduced the enrollment of children in the poor quintiles

(Deolalikar 1998).

The finding that there is apositive association betweenthe gross primary enrollmentrate and real governmentelementary education expen-diture per child should notcome as a surprise. Additionalexpenditures increase schoolcapacity (in the form of bothclassrooms and teachers), andto the extent that enrollmentsare capacity-constrained,additional expenditures willallow more children to beenrolled in a larger number ofschools and classrooms.

What may be moresurprising is that the observedpositive association betweenpublic spending on elementaryeducation and the grossprimary enrollment rate isstronger in the poor statesthan in the non-poor states.This may simply reflect thefact that schools and schoolspaces are more limited in thepoor states than in the non-

poor states, so that anexpansion in school capacityresults in a greater increase inenrollments in the formerthan in the latter. As notedearlier in this chapter, there issome support for such afinding from another study onIndia. Lanjouw and Ravallion(1999) find that the poor inIndia typically benefit morethan proportionately at themargin when there is anoverall expansion in primaryschool enrollments. Thishappens because poor studentsare almost always the last tobe enrolled, and the better-offare typically already in school. As a result, Lanjouwand Ravallion (1999) arguethat government educationalexpenditures that expandschooling access are generallywell-targeted to the poor.83

Surprisingly, the econo-metric results presented inAnnex IV.1 do not indicate a

strong association betweenthe gross primary enrollmentrate and adult female literacyrates in a state, once there iscontrol for both per capitaGSDP and per-childeducation spending.

It is important, however, tointerpret these results in thecontext of the earlierdiscussion in this chapter. The results do not considerthe counter-factual – viz., theimprovements in educationaloutcomes that would occur ifthe quality of governmenteducation expenditures weresimultaneously improved viagreater accountability inservice delivery and bettergovernance. All that theseresults indicate is the urgentneed for investments inadditional schools, classrooms,and teachers, to be targeted tothe low-enrollment states viaan effective and accountability-based delivery system.

96

Attaining the Millennium Development Goals in India

Independent OLS Fixed-effects Fixed-effects with adult female literacy*variable Coeff. t-ratio Coeff. t-ratio Coeff. t-ratio Coeff. t-ratio Coeff. t-ratio

Ln real gov’telemen. educ. per child 6-14 (G) 0.173 4.90 0.226 5.67 2.211 5.23 2.490 2.02 1.344 0.87

Ln real grossstate domestic product percapita (Y) 0.229 6.74 -0.034 -0.47 1.605 4.53 1.951 1.90 1.971 1.93

Ln G x Ln Y -0.225 -4.71 -0.268 -1.95 -0.139 -0.81

Ln adult femaleliteracy (F) 0.175 0.86 2.160 1.33

Ln F x Ln Y -0.238 -1.23

Time trend (t) -0.001 -0.46 0.003 0.96 -0.004 -1.04 -0.007 -0.60 -0.005 -0.44

Intercept 1.504 5.61 3.452 5.33 -10.946 -3.51 -13.896 -1.52 -13.671 -1.51

Number of observations 252 252 266 56 56

F-test for model 35.91 37.72 38 2.98 2.83

R-squared 0.422 0.808 0.8247 0.7971 0.8057

F-test for stateeffects 36.00 40.216 4.663 4.465

Annex Table IV.1 : Fixed-effects, log-linear regression of the gross primaryenrollment rate, pooled data for 1980-1999 across Indian states

Notes: All regressions (with the exception of those marked *) use pooled data across 14 states and 20 years.Dependent variable is log of the gross primary enrollment rate. Data are obtained from various issues ofthe Selected Educational Statistics, state government budget demand documents, and CSO statistics onGSDP of states. All values are in constant 1993-94 Rupees (using state-specific GSDP deflators). Standarderrors are corrected for heteroscedasticity using the Huber-white method. Figures in bold are statisticallysignificant at the 10% or lower level.

*Data on female literacy are available only for 1983-84, 1987-88, 1993-94 and 1999-2000, so theseregressions are restricted to these four years.