Improving Primary School Education in India: An Impact Assessment of DPEP I
Jyotsna Jalan
Indian Statistical Institute, SASPR jjalan@isid.ac.in
Elena Glinskaya
SASPR eglinskaya@worldbank.org
Outline of talk
! ! ! ! ! ! !
Introduction DPEP program Evaluation problem Data Results Interpretation of results Conclusions
What is common amongst these?
!NFE !OBB !TLC !MDM !SK !LJ
Non-formal education (1982) Operation Blackboard (1986)
Total Literacy Campaign (1988)
Mid-day meals (1982)
Shiksha Karmi (1987) Lok Jumbish (1988)
Have these programs had an impact?
Limited evidence!
Census 1991 " National literacy rate: 52 percent " National female literacy rate: 39.3 percent " Female literacy rate in rural areas: 30.6 percent " Some rural areas, female literacy rate: 11.6
percent
" Scheduled caste literacy rate: 37 percent " Scheduled tribe literacy rate: 32 percent
Government of India launches the District Primary School Education Program (DPEP) in 1993-94
! Provide access for all children to primary school or its
equivalent non-formal education
! Reduce overall dropout rates ! Increase average learning achievement levels ! Reduce gaps in enrollments, dropouts, and learning
among gender and social groups
! Establish capacity at the district, state and national
level to plan, manage and monitor program
Funding of the DPEP
! 85 percent borne by the central government and remainder by
the individual states
! Central governments contribution funded by World Bank, EU
(Madhya Pradesh), DFID (HP, Orissa, WB and AP) and UNICEF (Bihar) and Netherlands (Gujarat)
! By 2001, US $1.5+ billion committed to the program and 50
million children covered under the program
! States had to maintain 1991-92 expenditures on elementary
education at the state level but had discretion to re-allocate
expenditures across districts
! At the district level, maximum allowed allocation was US $8
million (over 7 years) irrespective of school-age population
Interventions?
Focus on interventions to improve quality of education with construction expenses restricted to 24 percent and management costs to 6 percent
#
Community mobilization and participation
$ Establish
Village
Education
Committees,
Mother-Teacher
Associations to increase awareness about the importance of primary school education
" Textbook recreation
$ Develop and introduce textbooks based on minimum levels of
learning (MLLs)
$ Stressed on importance of availability of textbooks in major tribal
languages
Professional development of teachers
$ Most teachers in DPEP schools possess a one year pre-service
teacher certification degree.
DPEP introduced regular in-
service training ranging from 3-20 days.
" Improved classroom practices
$ Teachers
were
encouraged
to
make
classrooms
less
intimidating and more sensitive to needs of female and minority children
#
Provide early childhood education
$ The expectation was that enrollments and retention of girls
would increase by providing alternate sibling care during school hours Reform in educational management and planning
$ Strengthen educational planning and management capacities in
the project states
Did the interventions work? ! Objective of this paper is to assess the impact of
the DPEP in terms of broad education indicators
! Focus on DPEP Phase-1 districts implemented in
1993-94 in 42 districts across 7 states (Assam, Haryana, Karnataka, Kerala, Madhya Pradesh, Maharashtra, TN)
! 310.5 million dollars were committed over 7 years
covering a little over 5 million students
Evaluation problem
! Nave estimator focuses primarily on changes in educational
outcomes that have happened within DPEP districts -- probably seriously overestimates the impact
! Simple estimator compares project with non-project districts -still incorrect, because project districts were not chosen
randomly
! Net impact of the program -- improvements in primary school
education in a DPEP district that can be ascribed to the program in addition to improvements that would have happened in natural course of events
Problems faced in evaluating DPEP
Data
! No data collected under the program on nonproject districts
! Need to use secondary data. This means that
we are able to assess the program in terms of general education indicators like enrollment rates and educational attainments rather than evaluate specific components of the project
Program Design and Evaluation Methodology
! Program placement is not random so comparison
of project with non-project districts not feasible.
! Use Propensity Score Methods to adjust for
observed differences across treatment and control groups
! Difference in Difference matched estimator:
(PT Pc)post-program - (PT Pc)pre-program
! Institutional differences in educational systems
across states so have to find controls within states
Which districts were chosen as treatments?
!Districts with female literacy rates below
the national average
!Districts where Total Literacy Campaigns
(community programs) implemented sponsored had been adult literacy successfully
Table 1: DPEP-I Districts (National female literacy rate: 39.4 percent)
ASSAM (Districts in 1991: 23) 1. Dhubri (23) 4. Darrang (30) 5. Karbi Anglong (30) 11. Marigaon (37) KERALA (Districts in 1991: 14) 2. Kasaragod (75) 3. Wayanad (78) 5. Malappuram (84) HARYANA (Districts in 1991: 16) 2. Kaithal (24) 3. Jind (24) 4. Hisar (24) 6. Sirsa (27) MAHARASHTRA (Districts in 1991: 30) 2. Parbhani(23) 3. Nanded (24) 6. Aurangabad (28) 8. Latur (35) 10. Osmanabad (36) KARNATAKA (Districts in 1991: 20) 2. Raichur (16) 6. Kolar (29) 7. Belgaum (31) 8. Mandya (32) TAMIL NADU (Districts in 1991: 22) 1. Dharmapuri (32) 2. South Arcot (34) 5. TSambuvarayar(36)
MADHYA PRADESH
(Districts in 1991: 47)
3. Rajgarh (9) 4. Guna (10) 5. Sidhi (11) 7. Surguja (13) 8. Shahdol (13) 11. Ratlam (14) 12. Chattarpur (14) 13. Panna (15) 15. Sehore (15) 17. Tikamgarh (15) 18. Dhar (16) 25. Mandsaur (20) 26. Raisen (20) 27. Bilaspur (21) 29. Satna (22) 30. Rajnandgaon (22) 32. Rewa (23) 33. Raigarh (23) 40. Betul (27)
Data:
! !
Data and Variables
Census 1991 National Sample Survey data for 1993-94 (pre-program) and 1999-00 (post-program)
Outcome variables:
!
School enrollment rates for 5-11 and 12-15 year old (all children, male and female children, children from scheduled caste and tribes) Dropout rate for 5-11 year old and 12-15 year old
! !
Progression from primary to higher levels for 5-9 year old in 199394 (11-15 year old in 1999-2000) Proportion of children with no education for 12-15 year old Proportion of children with completed primary education for 12-15 year old Separate MP -- 2 state level interventions -Alternative Schooling and Education Guarantee Scheme concurrently with DPEP Phase-I
! !
Table 2: Literacy rates across different subpopulations
Indicator
Female literacy rates Male literacy rates Scheduled caste (females) literacy rates Scheduled tribe (females) literacy rates
Source: Census 1991
DPEP I 26.48 (15.7) 55.66 (12.4) 21.00 (14.90) 15.89 (13.10)
Non-DPEP I 39.23 (18.1) 66.69 (11.6) 34.49 (19.50) 27.68 (19.80)
Descriptive Statistics
50th Round (1993-94) DPEP No of households Percent of children :5-11 years :12-15 years Percent of girls :5-11 years :12-15 years Percent of scheduled tribes Percent of scheduled castes 5,986 15.51 8.36 8.32 3.47 16.7 18.0 Non-DPEP 12,576 14.60 8.15 7.94 3.42 9.0 16.8 55th Round (1999-00) DPEP 5,438 15.75 8.86 8.45 3.96 18.8 19.6 Non-DPEP 13,027 14.90 8.41 7.77 3.81 10.6 19.4
Enrollment Rate among 5-11 year old
100% 95% 90% 85% 80% 75% 70% 65% 60% 55% 50% 1993 1999 All Children non-DPEP DPEP-I 1993 1999 Male 1993 1999 Female 1993 1999
Minority
non-DPEP (with below national average female literacy level)
Dropout Rate among 12-14 year old
50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 1993 1999 All Children non-DPEP DPEP-I 1993 1999 Male 1993 1999 Female 1993 1999
Minority
non-DPEP (with below national average female literacy level)
Proportion of Children among 12-15 years old with No Education
50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 1993 1999 All Children 1993 1999 Male 1993 1999 Female 1993 1999
Minority
non-DPEP
DPEP-I
non-DPEP (with below national average female literacy level)
Proportion of Children 12-15 years old with at least Completed Primary Education
100% 95% 90% 85% 80% 75% 70% 65% 60% 55% 50% 45% 40% 1993 1999 All Children non-DPEP DPEP-I 1993 1999 Male 1993 1999 Female 1993 1999
Minority
non-DPEP (with below national average female literacy level)
Table 4: Net impacts on enrollments
5-11 years All children Girls Boys Minorities 1.439 (1.02) .035 (.02) -.624 (-.37) 4.561** (1.65) MADHYA PRADESH All children Girls Boys Minorities 1.799 (.75) -1.359 (-.41) 2.856 (1.00) 17.143* (4.55) 3.382 (1.07) 4.889 (1.04) 3.844 (1.04) 9.626* (2.02) 12-15 years -1.241 (-.65) .481 (.75) -1.186 (-.52) -3.527 (.99)
Table 5: Net impacts on dropouts
5-11 years All children Girls Boys Minorities 12-14 years -3.263** (1.70) -3.831 (1.28) -.554 (.24) 18.14* (3.20)
-.173 (.13) 1.358 (.71) .724 (.44) -7.471* (1.96)
MADHYA PRADESH -1.147 (.48) 2.506 (.75) -5.087** (1.79) -9.168**
All children Girls Boys Minorities
-6.136* (1.90) -10.225* (1.98) -3.613 (.94) 23.23*
Table 6: Net impacts on cohort progression
Difference between proportion of 11-15 year olds in 1999-00 enrolled in middle or high school and 5-11 year olds in 1993-94 enrolled in primary or pre-primary school
DPEP I states All children Girls Boys Minorities
Madhya Pradesh 4.162 (1.48) 4.287 (1.12) 6.080* (2.19) 19.149* (4.63)
2.651 (1.52) 1.568 (.66) 5.490* (2.44) 7.324* (2.32)
Table 7: Net impacts on educational attainment (12-15 years old)
No education All children Girls Boys Minorities At least completed primary .600 (.35) 4.147 (1.50) .497 (.21) -9.744* (2.72) 5.414** (1.72) 12.480* (2.65) 5.249 (1.314) -1.175 (.25)
-3.837* (2.40) -1.452 (.58) -5.139* (2.85) 1.235 (.37)
MADHYA PRADESH -6.450* (2.25) -4.072 (.91) -10.243* (3.28) -2.411 (.52)
All children Girls Boys Minorities
Interpretation of coefficient estimates
! Optimal amount of schooling is given by:
1 s =T ln[ ] (r g) (r g)
s=years of schooling, T=age of retirement, r=interest rate, =returns to education, g=growth rate, =ratio of private direct costs of schooling to opportunity cost of school time
! DPEP provided additional $9.1 per student per year in program districts. If
average annual cost of educating a primary school student for one year is between $20-$50, then additional $9 would lower costs by 20 to 40 percent.
! Calibrated estimates of DPEP on years of schooling: (T=60, r=.05, g=.025)
DPEP funds represent: 20 percent fall = .1 = .2 1.7 0.8 Translating into increase in enrollment rates 27.2 percentage points 12.8 percentage points
Potential problems with this interpretation
! Assumption is that DPEP funds represented a net increase in school spending in the program districts ! There is evidence that this is true at the state level, but no
information about districts within the state. It was within the rights of the States to re-allocate expenditures across districts
! DPEP emphasized improvements in school quality only with limited funds devoted to new school construction ! Improve already enrolled students welfare without translating
into increased enrollments
! District Collector effect ! District collectors may move between districts in a state and
carry the DPEP human capital that diffuses program effects
Conclusions
! Net program impacts on minority children are the most
impressive especially in the state of Madhya Pradesh
! No impacts on girls especially on their enrollments.
Limited impacts on other indicators
! Was 300 million dollars (cost of DPEP-I) wasted? NO.
DPEP introduced a new approach to primary school interventions in India that cannot be well assessed from secondary data
! Interventions must be accompanied by adequate data
collection
