Speaker: Priyadarshani Joshi, researcher at the Global Education Monitoring Report Discussant: Claire Galante, Project Manager at Agence Française de Développement Moderator: Michaela Martin, Programme Specialist at IIEP-UNESCO Wednesday, 4 October 2017, 4 p.m. to 6 p.m

1. How Does Private Schooling Growth Affect
the Public System and Educational Equity?
IIEP Strategic Debate
October 4, 2017
Priyadarshani Joshi
Research Evidence from Nepal

2. A public-private education system
Parents
Students
Teachers, principals
Education bureaucracy
(national, district, sub-district)
Private school
associations

3. Research Questions
• How and why do parents select schools? How do they
continue to engage with schools after choosing them?
• What is the public sector’s experience of and response
to private competition?
• How do private school actors view their role in the
education system?
3

4. Research Motivation and Perspective
• Choice a global trend and academic focus (charters, vouchers, home
schooling, private schools, etc.)
• To focus on process: In developing countries, there is a lack of focus
on “analyzing the processes through which schooling decisions are
made, and how households interact with schools once they make that
choice” (Srivastava)
• To focus on local experiences: We need a better understanding of
local education markets to understand how competition is experienced
and whether there is any choice (Betts, 2009; Lubienski and Weitzel,
2010)
4

5. Investigate key aspects of the choice debate
Expected Benefits and Concerns
5
References: Chubb and Moe (1990), Schneider,Teske and Marschall (2000), Lubienski andWeitzel (2010); Hsieh and Urquiola (2006);
Levin (2002)
Better match children to schools, more engaged and satisfied
Sorting by income and ability (who chooses, who loses?)
Face competitive pressures, and respond with improvement efforts
Inappropriate to view as a market, difficult to incentivize reform
Responsive to parental demands, need to provide quality
Shifts focus from social good aspects of education; focus on
getting “most desirable”
Parental
Choice
Public
schools
Private
schools

6. 6
Research Design
& Context

7. Mustang
Chitwan
Kathmandu
Kavre
Dadeldhura
Supplementary
7
JhapaIn-depth
Research Design
field sites for primary data collection in Nepal
Sarlahi

8. 8
PRINCIPAL SURVEY (KATHMANDU, CHITWAN)
IN-DEPTH DATA SUPPLEMENTAL DATA
Kathmandu
Chitwan
National
DadeldhuraSarlahiMustang Kavre Jhapa
District-Level Private
School Records
(Kathmandu, Chitwan) National
School
Records
National
Exam
Records
Population
Census
QuantitativeSECONDARY SOURCES
PRIMARY SOURCES
Parent focus
groups, surveys
(6th grade)
data
Qualitative
Living
standards
survey
SLC
study
2004
Research Design

9. • Low-income, landlocked/difficult topography, high ethno-linguistic diversity
• Political instability, high migration
• Private schooling growth fueled by demand and supply side factors
• Public education reforms: per-child funding, decentralization
Sources: Population Census 2011, National Examination Center records, EMIS national administrative records, NLSS surveys I, II and III
the development of a two-tiered system
Private Market Shares Student pass rates
(High School Exams)
21
6
0
10
20
30
2005 06 07 08 09 2010
public
private
% Dalit
(disadvantaged population)
Country Context
3,3
7,8
12,7
31,6
48,2 52,1
0
20
40
60
1995-96 2003-04 2010-11
rural urban
47
90
0
20
40
60
80
100
1999-00 2004-05 2009-10
public
private
9

10. 10
“Protests in private schools
over raising school fees”
“5 public schools displaced
due to shortage of students”
(Nagarik National Daily | May 15, 2011) (Nagarik National Daily | May 4, 2012)
Issues with two-tiered education system

11. 11
Parent
decision-making
Why and how do parents select schools?
How do parents continue to exercise choice after
selection?

12. 12
Framing the process of parent decision-making
Schneider, Elacqua and Buckley (2006); Schneider,Teske and Marschall (2000); Hirschman (1970); Srivastava (2007).

13. Financial constraints and children’s interest key to school selection
Child role in decision-making
“I had no idea about this school. She said that I will
go to this school, and asked me for the admission
fees. She came back and said that the school needs
either the mother or father to come for admission.
The next day I went to the school, and got her
admitted there.”
13
62,7
13,3
72,0
21,3
66,7 69,7
54,2
4,2
0
20
40
60
80
Close to
home
Cheaper
than others
Academic
quality
English
medium
Private
Public
The top 3 reasons they selected the school
(percent that chose each attribute)
10,7
20,4
27,3
22,6
14,7
26,4
0
10
20
30
High Ed Med Ed Low Ed No Ed Private Public
Proportion of parents who chose child decision
as one of the top 3 reasons for selecting the school
How: through informal networks, local
knowledge, and child interest
N = 147
Why: proximity, ability to pay,
“quality”, English medium

14. “Average and above”
Public
Poor parents do not benefit from existing choice
“Below average”
Public
“Smaller”
Private
“Well-known”
Private
Focus groups - 2-3 public schools;
1-4 private schools in 6 locations
parents most engaged
and empowered
Less engagement in
more high demand
schools
Dissatisfaction on
English education,
and disciplinary
aspects
Schools not gauging
parental satisfaction;
parents not communicating
dissatisfaction

15. 15
Public sector
perspectives
How do public schools experience competition?
What are public schools doing in response to private competition?
What are the factors that mediate public school efforts?

16. 16
Framing the process of
competitive effects on public schools
OutcomesExperiences of
Competition
Responses to
Competition
Motivations/
Constraints
(Betts, 2009; Gauri,1998; Hess, Maranto and Milliman, 2001; Ni and Arsen, 2010; Lubienski andWeitzel, 2010; Ladd and Fiske,
2003; Rouse, Hannaway, Goldhaber and Figlio, 2007; Thapa, 2013; Zief, Maynard, Bradley, Keefe and Kralik, 2005)

17. Extent of private sector growth, local
enrollment pressures, and how
officials perceive this competition.
Competition Measures (Objective and Subjective)
The local experience of competition
• Geographic proximity: How many private secondary schools are
within a 1 km. walk from your school? (2011-12)
• Market share: Percentage private enrollment in locality
• Principal perception:
Who are you competing with? (Name 3 schools) (2011-12)
[Dummy variable: 1 = name at least 1 private school]

18. 18
144
463
pre-1950 1960 1970 1980 1990 2000 2010
67
92
pre-1950 1960 1970 1980 1990 2000 2010
Kathmandu District Chitwan District
year
Numberofsecondarylevelschools
Sources: Private school district records, Principal Survey
public
private
Competitive experience – private school growth?
More private schools

19. 19Sources: Private school district records, Principal Survey
ChitwanKathmanduNepal
30
35
40
45
50
55
60
2006-07 07-08 08-09 09-10 2010-11
public
private
30
35
40
45
50
55
60
2006-07 07-08 08-09 09-10 2010-11
public
private
30
35
40
45
50
55
60
2006-07 07-08 08-09 09-10 2010-11
public
private
The data is restricted to schools that have data for 2006-07 to 2010-11. The sample includes 22,146 public schools and 1770 private
schools in Nepal; 266 public and 554 private schools in Kathmandu; and 339 public and 84 private schools in Chitwan.
Competitive experience – enrollment pressure?
Lower enrollment

20. Response: some policy adoptions for competing
English medium Ties and belts
ABCs
Source: Principal Survey, 2011-12
Adopted
English
medium of
instruction:
73%
Added ties
and belts
to school
uniforms:
67%
“private-mirroring” strategies Instructional time strategies
EXAM Coaching
TEST
PREP
Remedial classes
REMEDIAL
SLC Coaching 96%
SLC coaching for
weaker students: 45%
Remedial
classes for
weaker
students:
35%

21. Example:Adopting English medium ABCs
• Parental demand for English; what is attractive about private schools
• Public schools hope to stem the outflow of students by providing English medium
** sig. at 5% level; * sig. at 10% level
Logistic Regression Results (Estimated odds ratios)
Dependent variable: 1 = Adopted English medium
Perception Measure (Mentioned at least 1 private
school among competing schools)
2.21**
Proximity Measure
Medium competition 0.63
High competition 0.49*
Interpretation:
greater than 1: higher likelihood of adopting English compared to omitted
category (no perceived competition; low competition)
Does policy adoption vary by extent of competition?
Source: Principal Survey, 2011-12
Kathmandu Chitwan
Year transitioned to English medium of instruction,
public secondary schools
0
10
20
30
2000
2002
2004
2006
2008
2010
0
10
20
30
2000
2002
2004
2006
2008
2010
Rapid recent adoption of
English medium
21
Principal perception measure may capture
leadership motivation, instrumental for
policy adoption

22. necessary conditions for productive responses
revenue
time
Resources
Community support
Principal’s Leadership
Child centric
learning
&
Committed
teams
Bureaucratic
support
in adopting policies
National
District
Resource
CenterTeachers

23. constraints to responding
revenue
time
No supplemental
revenueLack community support
Start from disadvantage
ABCs
English medium
Lack
Bureaucratic
Support
political party interference

24. Pervasive role of political instability
24
“Politically, all the teachers are divided.And so are the school management
committees. And so is society.When you talk about children, they ask “whose child”
they are and “which party” they belong to.”
Deputy Director, Department of Education

25. 25Source: Qualitative datasets
A separation of roles for public and private schooling
All political party leaders’ children have gone to boarding (private)
school. So, whoever is part of the intellectual and educated circle, they
have all focused their (education related) attention on boarding
school.The government/community school has become a place to
play politics and get some additional respect
”
“
School
management
committee
member
Questions posed to public school principals on challenges % agree
• There are more politically appointed teachers who are not
concerned with teaching in public schools
65.6
• Most of the parents of public school students do not
understand the importance of education
59.4
Principals of public schools in locations with more private schools have higher
likelihood of agreeing to these challenges.

26. 26
Private sector
perspectives
What are the characteristics of the private schools in the sample?
What do private schools provide, in contrast to public schools?
What do principals think the government role is, and should be?

27. 0
100200300
qfee_gr1us68
pre-1980 1980s 1990s 2000s
Kathmandu schools – fee-based diversity
Private schooling trends
Private expansion seems to be more “market-like” in Kathmandu
Public schools = $8.6 avg. (0 to $182); n = 145
Estimatedmonthlyandannualfees
(InUS$),grade1,2011-12
Decade of establishment1st grade fee distribution
0
.002.004.006.008
.01
0 100 200 300
qfee_gr1us68
N = 607
27

28. 0
50
100150200
qfee_gr1us68
pre-1980 1980s 1990s 2000s
Chitwan schools – narrower fee bands
Private schooling trends
Private school fees lower, more controlled in Chitwan - smaller size
Public schools = $3.6 avg. (0 to $92); n = 67
Estimatedmonthlyandannualfees
(InUS$),grade1,2011-12
Decade of establishment
0
.01.02.03
0 50 100 150 200
qfee_gr1us68
1st grade fee distributionN = 113
28

29. Private school characteristics
Why was the school started?
• Employment,“clean” investment, social motivations
How?
• Group got together to operate a school (2 to 53 shareholders) (esp. in Chitwan)
• Takeover administrative control from other team esp. in Kathmandu – 150
requests for ownership changes in two year period (2010 and 2011)
Who attends?
• All strata; Most do not have the option to sort students – too much competition
in city centers, too few children outside city centers
Why more favored?
• “it’s not that private schools are very good; it’s just that public schools are bad”
• Sole provider of quality education (care, teacher effort, high stakes exam
performance) – combination of factors are better
29

30. Need productive government role in private schooling
Private school demands
• View private schools positively, recognize contributions to
national development
• Support private schools (resources, monitoring, bigger private
role)
• Increase communication and reduce hostility
• Develop a separate private education act to stabilize the sector
and view private schools as education institutions and not
companies
30

31. 31
“Best” schools
how can one identify great schools, and what
makes them special?

32. 32
Which are the schools identified as the “best” schools?
Survey Question: which are the best schools in the district?
(name up to 3 schools)
Of the 212 public schools, 56 were identified as being among the best schools by
other public school principals.Very few public schools were cited frequently
school
1
23
B
A

33. 33
What differentiates the best public schools?
Most frequently cited schools have better examination performance, and
perhaps as a result are in greater demand (higher enrolment, higher fees,
entrance exams at the secondary level)
We consider School B as a “guardian of all schools in this
whole locality”. And we learn from them.
“
”
Within-school
functioning
It is not enough for just us to improve – only one school
improving is not enough. Need 10 to 20 to do well to
bring children back to public school.
We need to build confidence for this to happen.
“
”
Vision for public
schooling
Targeting student performance
Motivating teachers by leading through example
Bigger student role in monitoring, providing suggestions
Involving parents in monitoring, helping motivate students
Strong reputation
Private school principal:

34. 34
What works?
I don’t say things like I am a pure teacher and have no interest
in politics. I just don’t advertise and promote it.
There are all kinds of [politically affiliated] teachers here – the
reason the team is still well formed is because none of the
teachers can say that the headmaster has particular favorites
or is biased towards someone or the other because of their
ideological orientation.
I am aware of not letting that bias to be felt. ”
“
Keep politics
outside the
classroom
[Principal,
School B]
My three children passed from this school itself. I cannot send
them to the private schools. It’s not that I don’t have the
(financial) capacity, … but I have the confidence that all the
children will pass from this community school with 1st or 2nd
division results and make their own progress.
Teach in public
schools
[Principal,
School A]
“
”

35. 35Source: Qualitative datasets
Challenge of social prestige, even in the best public schools
We surveyed 100 local guardians – why don’t they come
even though our results are good? We got three points of
feedback.
First our school has no English medium – that was the
main issue.
Second, they said that the students of the school are
children who was dishes in other people’s homes, and we
feared that they would spoil by being in bad company.
And third, most of them said that – school B is a
government school.The fear was that their social status
would fall.
Why not
attractive to
parents
despite track
record and
reputation?
”
“

36. 36
Conclusions &
Implications

37. Circling back to the choice debate
Implications from the Nepal case
37
References: Chubb and Moe (1990), Schneider,Teske and Marschall (2000), Lubienski andWeitzel (2010); Hsieh and Urquiola (2006);
Levin (2002)
• Better match children to schools, more engaged and satisfied? Yes, primarily
middle class
• Sorting? Yes, unregulated choice, based on financial capacity
• Face competitive pressures, and respond? recent focus on “private mirroring”
policies due to extreme enrollment pressures
• Difficult to incentivize reform, inappropriate to view as a market? Yes, politics and
loss of community support propel the public-private divide
• Responsive to parental demands, need to provide quality – Yes, focus on test
score outcomes, English medium
• Focus on getting “most desirable”; shifts focus from social good aspects of
education? – not all private schools super selective, but sorting and profit
motive are concerns, private schools need govt attention
Parental
Choice
Public
schools
Private
schools

38. • Build localized, process-based data systems learning from
other education data systems
o Existing data: integrate and increase the accessibility of existing school-
level data
o New data: instructional and noninstructional policies, school financing,
parental SES, different data system for private schools
• Streamline roles of local level personnel
• Improve teacher distribution and financing equity
38
Facilitating research and policymaking
data and policies

39. • Recruit effective principals: importance of good principals who can navigate the
political system
o Can effectively motivate teachers and minimize political activities within school
o Partner the best functioning public schools with public schools that require special assistance
• Fight the perception battle
o Need to switch mindset from “schooling for the poor” to accessible to the poor
o Positive publicity – teachers and principals sending their children to the public schools have
led to transformations in some examples
• Collaboration focus
o Politicians could be viewed more productively as strategic partners and not just “interference”
– some districts have smoother operations [private sector] 39
What can, and has worked?
“The plant is dry… It’s not dead yet. But, it has nothing – the leaves have fallen.You
cannot give it too much water or fertilizer.You have to protect it from sun and water
damages, and slowly improve its situation so that new leaves emerge again.
Private school principal (on the public education system)

40. 40
Global implications for the public system and equity
Can private growth make a positive difference?
• Yes, private expansion increases opportunities for the middle to lower middle class
– to choose schools; generate employment and contribute to society
What are the effects on equity?
• Loosely regulated systems will lead to significant negative consequences on equity
Can the public sector benefit from private growth?
• Public systems can suffer from stratification and then stigma; hard to reverse with
specific reforms
• To benefit public schools, public-private systems requires government commitment
and collaborations, and well-timed interventions – hard to transform after long-
term decay.

41. 41
Low and lower middle income countries (WB)
Landlocked countries Fragile political climate (EFA GMR;WB conflict report)
At least 10% private school enrollment (WDI – UNESCO UIS)
Burkina Faso
Central African Republic
Chad
Congo Republic
Democratic
Republic of
Congo
Côte d'Ivoire
Ethiopia
Guinea
Guinea-Bissau
Liberia
Mali
Niger
Pakistan
Paraguay
Philippines
Togo
Uganda
Bolivia
Nepal
Global relevance – comparative possibilities

42. Should the global community invest in
for-profit chains that aim to provide private schools for the poor?
Locally grown solutions are important,
Even more so in unequal societies
42

43. 43
• Dissertation committee: Professors Emily Hannum, Richard
Ingersoll, Jere Behrman,Amita Chudgar
• Field research: Participating schools, parents and officials; field
research teams (quantitative survey; qualitative fieldwork)
• Fieldwork and Secondary Data Approval: Department of Education
(Nepal)
• Funding: World Bank (Nepal), Dean’s Fellowship Penn GSE;
Amherst College Fellowship for Further Graduate Study
• My father, Janardan Joshi
• Graphic design: AshutoshTamrakar
• Family and friends
Acknowledgements

44. works.bepress.com/pjoshi
• Joshi, P. (2014a). Continuing to exercise choice after school selection in Nepal.
International Journal for Educational Development (IJED), 37, 57-67.
• Joshi, P. (2014b). Parent decision-making when selecting schools: the case of Nepal.
UNESCO Prospects, Quarterly Review of Comparative Education, 44(3), 411-
428.
• Joshi, P. (2016). Experiencing and responding to private competition: the importance
of subjectivity and intermediate outcomes. Comparative Education Review, 60(3),
571-600.
• Joshi, P. (2017). Identifying and investigating the “best” schools: A network-based
analysis on Nepal’s public education system. Compare:A Journal of Comparative
and International Education, 1-18.
• Joshi, P. (2018a). Competitive effects (quality and sorting) of private competition on
public school outcomes. Under review at IJED.
• Joshi, P. (2018b). The political, bureaucratic and societal challenges faced by public
schools in low-income countries. Under review at Journal of Development
Studies
• Joshi, P. (2018c). The role of the private schooling system in Nepal: private sector
perspectives. In progress.
44

45. 45
Thank you!
Comments: pjoshi@alumni.princeton.edu; p.joshi@unesco.org

46. 46
Private market share, 2010-11
Mathscores,2010-11
0
20
40
60
80
100
-20 0 20 40 60 80 100
Private market share, 2010-11
Relativemathscores,2010-11
0,4
0,6
0,8
1
1,2
0 20 40 60 80 100
r = -0.67r = 0.21
School-average math test scores
Locality-average public school outcomes/ locality-
average public and private school outcomes
The relative outcomes are lower in high competition
regions than in low competition regions
No statistically sig. association between public
school test scores and private market share
Public school exam outcomes are slightly better in high private market
share regions, but are relatively much worse than private schools

47. 47
Implications for the public system and educational equity
Historical developments
Growing distrust in government provision, private school growth
fewer children and more focus on investing in children
Parents choose by quality and
financial capacity,
determines social standing
Over the long term, unregulated education systems lead to educational inequity, and
public systems suffering from stratification and stigmatization
• Private schools are unlikely to automatically induce public school improvements
• Recent quasi-private policies suggests competitive & accountability pressures can incentivize changes
cumulative
effects
sorting of students
• Despite policy efforts, some public schools may not be able to improve outcomesIn high competition regions:
• Despite policy efforts, some public schools may not be able to improve outcomes
• Even if outcomes improve in some schools, other parents may not be interested
(public-private gap too large)
Over time, from sorting to stigma
of public schooling