Professional Documents
Culture Documents
Abstract
A census and survey of schools in the slums of East Delhi, India, explored the nature and extent of private education
serving low-income families, and compared inputs to public and private schooling. Around two-thirds of all schools were
private unaided, with more unrecognised private than government schools. Teaching activity was found to be considerably
higher in private unaided than government schools, although teacher absenteeism was lowest in government schools. Most
inputs showed either comparable levels of provision in government and private unaided schools, or superiority in private
unaided schools. Possible implications are explored, concerning targeted vouchers, increased regulation and self-
regulation.
r 2006 Elsevier Ltd. All rights reserved.
0738-0593/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijedudev.2006.08.002
ARTICLE IN PRESS
206 J. Tooley, P. Dixon / International Journal of Educational Development 27 (2007) 205–219
areas’ often located adjacent to a government 15% and 17%, respectively (World Bank, 2004;
school (Aggarwal, 2000, p. 20). In Uganda and Habyarimana et al., 2004).
Malawi private schools have ‘mushroomed due to Public education for the poor is also reported to
the poor quality government primary schools’ suffer from inadequate conditions. One government
(Rose, 2002, p. 6; Rose, 2003, p. 80), while a ‘large school highlighted by the World Development
increase’ of ‘fly-by-night’ private schools has been Report 2004, in north Bihar, India, describe
reported in South Africa (Rose, 2002, p. 4). In ‘horrific’ conditions (World Bank 2003, p. 24).
Kenya ‘the deteriorating quality of public education Facilities in government primary schools in Calcut-
y created demand for private alternatives’ (Baurer ta were reported ‘by no means satisfactory’ (Nam-
et al., 2002). bissan, 2003, p. 20): of 11 primary schools only two
Reasons given for this ‘mushrooming’ highlight had safe drinking water for the children, nine had a
the low quality of government schools for the poor, general toilet, and only five had a playground.
including problems of teacher absenteeism and lack Listing major problems in their schools, head-
of teacher commitment. In government primary teachers included the lack of electricity, space and
schools in West Bengal it is reported that ‘teachers furniture (p. 21). A study of 100 government
do not teach’ and ‘teaching is the last priority for primary schools in Bangladesh found that 81%
the teachers’ (Rana et al., 2002, p. 64 and 67). The had water, 39% electricity, 97% toilets, 76% a
Probe Team found that in their sample, only 53% of playground and only 0.4% a library, while the
government schools was there any teaching going average pupil–teacher ratio was 69:1. (Chaudhury et
on at all (The Probe Team, 1999). The Human al., 2004b). The Probe Team in India found that out
Development Report 2003 notes that in India and of 162 government primary schools, 59% had no
Pakistan ‘poor households cited teacher absentee- functional water supply, 89% had no toilets, and
ism in public schools as their main reason for only 23% had a library, 48% a playground. The
choosing private ones.’ (UNDP, 2003, p. 112). A average pupil teacher ratio was 68:1 (The Probe
comprehensive survey of teacher absenteeism con- Team, 1999).
ducted by the World Bank in India (Kremer et al., Finally, private schools in India provide (or
2004) looked at a nationally representative sample purport to provide) English medium instruction,
of 20 Indian states, involving 3750 schools. which is desired by parents; government schools
Although public and private schools were investi- teach in state languages, not usually teaching
gated, like was not compared with like—urban and English until about Class 5 (Nambissan, 2003; De
rural government schools were compared with rural et al., 2002). Moreover, in some countries public
private schools only. In government schools, schools have limited places, because of an increase
absenteeism rates were 25.2% in rural and 22.9% in the number of school-age population without an
in urban schools, while in the rural private schools, increase in government spending (Rose, 2002;
absenteeism was about 22.8% (Kremer et al., 2004, Nwagwu, 1997).
p. 5 and p. 9). In 257 government, mosque and However, whilst this literature indicates that one
private schools in Pakistan there was an absence of the reasons low-income parents send their
rate of 20% when researchers physically checked the children to private schools is the perceived low
attendance status of one randomly chosen teacher at quality of public education, concerns are also
the school. In this same study the ‘official’ records expressed about the quality of the private schools
showed only an absence rate of 5% (Ali and Reed, to which parents turn as alternatives, especially
1994). Studies of government teacher absence in six those that are not recognised by government. The
countries—Bangladesh, Ecuador, India, Indonesia, Oxfam Education Report, for instance, notes that
Peru and Uganda—found teacher absence rates to while ‘there is no doubting the appalling standard of
be between 11% and 27% (Chaudhury et al., 2004a; provision in public education systems’, the private
Rogers et al., 2004, p.142; Akhamadi and Surya- schools that poor parents are using instead are of
darma, 2000). A study in two rural districts of ‘inferior quality’, offering ‘a low-quality service’
Kenya found that government teachers were absent that will ‘restrict children’s future opportunities.’
nearly 30% of the time and present at school but (Watkins, 2000, p. 230). Nambissan (2003) notes
not physically in the classroom 12.4% of the that in Calcutta, ‘the mushrooming of privately
time (Glewwe et al., 2004). Studies from Papua managed unregulated pre-primary and primary
New Guinea and Zambia revealed absence rates of schoolsy can have only deleterious consequences
ARTICLE IN PRESS
J. Tooley, P. Dixon / International Journal of Educational Development 27 (2007) 205–219 207
for the spread of education in general and among In India, school management type is of three
the poor in particular’ (p. 52), for the quality of the kinds: government, private aided and private
private schools is ‘often suspect’ (p. 15, footnote 25). unaided. Government schools are funded and
Save the Children, although noting that poor managed by some level of government, state or
parents in Nepal and Pakistan identify ‘irregularity, local. Private aided schools are privately managed,
negligence and indiscipline of the teachers, large but usually have 100% of teacher salaries, plus
class sizes and a lower standard of English learning’ other expenses, funded by government. Private
as ‘the reasons why they decided against public unaided schools are entirely privately managed
schools’ (Save the Children UK, South and Central and privately funded. Private unaided schools are
Asia, 2002, p. 8), is concerned that the private of two types, recognised and unrecognised. The
schools they opt for offer ‘an extremely low former have purportedly met the regulatory require-
standard of education’ (p. 13). Finally, Rose ments of the state. Unrecognised schools are in
(2002) asks why poor parents in sub-Saharan Africa effect operating in the informal sector of the
are paying ‘for poor quality education, when they economy. They have either not applied for recogni-
could be getting fee-free schooling in the state tion, or have not succeeded in gaining recognition
sector’? (p. 16); ‘the quality of education received is from the government. The research was interested in
debatable’ in the private sector (p. 7). However, the investigating all these school types.
evidence given about poor quality private provision, After consultation with government officials and
and the relative quality of public and private non-government organisations working in the city,
provision for the poor in these sources is limited— the census was conducted in North Shahdara, East
either suggesting that ‘little hard evidence’ is Delhi, reported to be one of the poorest areas of the
available (Watkins, 2000, p. 230), or basing the city.1 North Shahdara covers an area of 40 km2, but
conclusion on the observation that private schools only the ‘‘notified slums’’ (according to the Census
employ low paid teachers, in low-quality accom- of India, 2001) were researched, estimated to cover
modation, without comparing what is on offer in about half this area. Permission was granted by
the government alternative. government officials for the research to be con-
Our two-and a half year study (April 2003–Dec ducted. A team of 20 researchers recruited from a
2005) aimed to contribute to the understanding of local non-government organisation were trained in
private school provision for the poor, and its methods of gaining access to schools, the use of an
relative quality vis-à-vis government provision, interview schedule for school managers and head-
through research in selected low-income areas in teachers, and an observation schedule, which was
India (Delhi, Hyderabad and Mahbubnagar), Gha- trialled to ensure reliability of observations made.
na, Nigeria, Kenya and China. This paper reports Given that we were particularly interested in finding
on some of the findings from Delhi, India only. It is ‘unrecognised’ private unaided schools, which are,
particularly concerned with exploring some of the by definition, not on any official list, the researchers
assumptions made about the nature of private were asked to physically visit every street and
schooling for low-income families, and in compar- alleyway in the area, during the morning of a
ing the inputs in these (including facilities, teacher school day (except where it was indicated that a
commitment, etc.) with those in government schools school was operating in shifts, in which case they
in the same areas. (Further research conducted returned to this school in the afternoon too),
detailed comparisons of the achievement of students looking for all schools, primary and secondary.
in these schools, and pupil, parent and teacher (Nursery-only schools were excluded from the
satisfaction, reported elsewhere). study, as were non-formal education provision,
such as learning centres and after-school clubs).
2. Method Government lists were used to check that all
government, private aided and recognised private
The research reported here, conducted during unaided schools were found.
October 2004, consisted of two main parts: a census When a school was located, the researcher called
of schools and survey of inputs, aimed to discover unannounced and asked for a brief interview with
the extent of private schools in selected low-income the headteacher or school manager, taking about
areas and to compare their inputs with government
1
schools serving the same populations. http://www.pratham.org/documents/northshahdara.doc.
ARTICLE IN PRESS
208 J. Tooley, P. Dixon / International Journal of Educational Development 27 (2007) 205–219
10 minutes. After this, the researchers asked to 3.1. Proportion of schools by management type
make a school visit, where they checked the facilities
available in the school against a short check-list of The survey team found a total of 265 schools in
facilities, and visited Class 4, to observe the activity the slum areas of North Shahdara. Of these, 26.8%
of the teacher. This visit was made when a normal (71 schools) were government, 7.2% (19 schools)
lesson was timetabled—the researchers were in- private aided, and the rest—66% of the total (175
structed to wait until such a lesson was scheduled if schools)—private unaided schools. That is, a large
there were other activities (assembly, break, sports, majority of schools is private unaided. Of these, the
etc.) taking place. largest number is recognised, (102 schools or 38.5%
The total number of schools located (including of the total), while 73 private unaided schools were
schools operating in shifts) was 265; in all schools unrecognised (27.5% of the total). Hence, there are
the researcher was able to ask at least some more unrecognised private unaided schools than there
of the questions on the short interview schedule. are government schools (Table 2). This figure must
Similarly, the researchers were able to answer be taken as indicating a lower bound on the
at least one of the questions on the observation numbers of private unrecognised schools, as we
schedule in all 265 schools. (The tables below cannot be sure we found all unrecognised private
indicate the exact numbers of observations unaided schools, as there were obviously no official
made for each question). Data discussed in Section lists with which to compare our findings.
3 below came from the interview, while data
discussed in Section 4 came from the observation
schedule. 3.2. Proportion of pupils by management type
In addition, a stratified random sample of 146
schools was selected, after stratifying the schools The researchers asked school managers or head-
into approximate size bands and three management teachers for the number of children enrolled in the
categories: private unaided (unrecognised), private primary and secondary sections of the school,
unaided (recognised), and government (ignoring for checking this against the register, but not physically
these purposes the small number of private aided checking this against pupils in the school. In the 265
schools). (Table 1). This sample was primarily used schools, it was reported that 137,493 children
to elicit further data on academic performance of attended. Breaking down this information by
children, background variables, and satisfaction management type, we find that 60.4% of all children
levels (reported elsewhere). However, the discussion were at government schools, 3.7% at private aided
of philanthropy and teacher salaries used data from schools, 27.2% of children at recognised private
this stratified sample. unaided schools, and 8.8% of children at unrecog-
nised private unaided schools (Table 3). Three
caveats must be made here. First, there is the
reported propensity of government and private
3. Results: census of schools aided schools to exaggerate enrolment, as there
are clear financial and job security incentives to
The main aim of the survey was to gauge the claim larger enrolment than is actually the case
extent of private provision, and to explore some (Kingdon, 1996; Kingdon and Drèze, 1998). Sec-
facets of the private unaided schools to increase ond, school managers and headteachers from both
understanding of this sector. public and private schools informed us of possible
Table 1
Schools in stratified random sample, by management type
‘‘double counting’’ of pupils. Many children, we unrecognised 57.5% boys and private unaided
were informed, are enrolled in both government and recognised 56.9% boys (Table 4).
private schools, in order to benefit from mid-day
meals in government schools—children we were told
go to private schools in the morning, and then go to 3.4. Teachers and pupil– teacher ratio
government school for the mid-day meal. This had
the additional benefit that children are able to take Researchers obtained data on the number of
examinations as a government, rather than private, teachers from 259 of the schools, by asking for this
school pupil—which was particularly valuable if information of the school manager or headteacher.
children were enrolled at unrecognised private The total number of teachers was 3511, with 51.3%
schools. Third, again we note that all unrecognised working in private unaided schools, 5.2% in private
private unaided schools may not have been located aided schools, and 43.5% in government schools.
by the researchers. For all these reasons, it is For the 259 schools reporting both their number of
suggested that the data here may underestimate the children and teaching staff, we calculate an average
true proportion in private unaided, especially pupil–teacher ratio of 32.8 to 1. This was highest in
unrecognised, schools. the government schools (54.7:1), followed by private
recognised unaided schools (26.1:1), then private
aided (23.4:1) and lowest in the unrecognised
3.3. Gender of pupils private unaided schools (22.7:1) (Table 5).
Table 3
Pupil enrolment by management type
Table 5
Pupil teacher ratio in each school management type
Management type Number of schools reporting Mean pupil-teacher ratio Std. deviation
Table 6
Age of schools by management type
Government 1 6 10 44 61
1.6% 9.8% 16.4% 72.1% 100.0%
Private aided 5 6 4 2 2 19
26.3% 31.6% 21.1% 10.5% 10.5% 100.0%
Private unrecognised 33 23 9 2 1 68
48.5% 33.8% 13.2% 2.9% 1.5% 100.0%
Private recognised 14 30 35 9 13 101
13.9% 29.7% 34.7% 8.9% 12.9% 100.0%
Total 52 60 54 23 60 249
20.9% 24.1% 21.7% 9.2% 24.1% 100.0%
3.5. Scope of schools ment was 2000 and 1995, respectively. While the
unrecognised schools are certainly newer than their
Over one quarter of government schools are recognised counterparts (which themselves are new-
primary-only schools, with just over an additional er than the government schools), they are certainly
half being nursery and primary providers. Around not all recently established. Table 6 gives the overall
10% provide primary and secondary education and figures for all schools, tabulated in intervals of 5
just over 10% provide all sections. The private aided years.
schools are mainly nursery and primary schools
(89.5%). For private unrecognised schools only 3.7. Medium of instruction
1.4% provide primary and secondary, the remaining
being primary and nursery providers. Some private The literature above noted that parents may
recognised provide all sections (8.8%). prefer private unaided schools because they are
English medium: our results confirmed a significant
3.6. Establishment of schools difference between private and government schools
in their medium of instruction. Of the total 265
Epithets such as ‘mushrooming’ and ‘fly-by- schools, 27.2% reported that they were English
nights’ used to describe private schools implies that medium only, 42.3% Hindi medium only, and the
such schools are in general newly established. Our remaining 30.6% English and Hindi medium.
data suggest that this is not entirely true. The mean Disaggregated into management types, we find
year of establishment for private unaided unrecog- 47.1% of recognised private unaided and 20.5% of
nised schools was reported as 1998; for private unrecognised private unaided schools reported they
unaided recognised schools the average year of were English medium, compared to only 2.8% of
establishment was 1993. The mode year of establish- government schools and 36.8% private aided
ARTICLE IN PRESS
J. Tooley, P. Dixon / International Journal of Educational Development 27 (2007) 205–219 211
Table 7
Number and % of free and concessionary seats in private unaided schools
Total seats Free seats % of free seats Concessionary seats % of concessionary seats
schools. The majority of government schools were they also offer free or concessionary seats to
Hindi medium (80.3%). Many of the private children. We explored this issue with the smaller
unrecognised schools are Hindi medium (45.2%) number of private unaided schools in the stratified
or provide both Hindi and English medium streams random sample. The researchers asked the school
(34.2%). manager how many students were admitted to the
school with free or concessionary seats, and
3.8. Fees in private unaided schools triangulated the results with questions on the
parents, questionnaire. Of the 111 private unaided
The private unaided schools were found to charge schools participating in this part of the research, 94
predominantly monthly fees. The researchers asked school managers gave information about the
school managers for details of these fees, checking number of free and concessionary places. Of schools
these where possible against written fee charges. giving information, 58% of the unrecognised and
There is a statistically significant difference in the 50% of the recognised private unaided schools offer
fees charged in unrecognised and recognised free places to some students in their schools.
schools, with the former consistently lower than Regarding concessionary places, 46% of the un-
the latter, at each level. For example, for pre- recognised and 48% of the recognised private
primary grade, mean fees in recognised private unaided schools offer these. (In both cases, the
unaided schools are Rs. 190.25 (£2.442) per month, difference between school types was not statistically
compared to Rs. 92.55 (£1.19) per month in the significant).
unrecognised schools. At primary grade, the same The total number of free seats given was stated as
figures are Rs. 227.60 (£2.92) compared to Rs. 1045 (591 in unrecognised and 454 in recognised
124.45 (£1.60). The median figures may be a more private unaided schools), while the total number of
useful figure for comparison, given that a small concessionary places was 1184 (409 in unrecognised
number of schools were found to be charging higher and 775 in recognised private unaided schools).
fees than others. At the primary level, these were Rs. That is, in these schools, 10% of all places were
100 (£1.28) for unrecognised and Rs. 192.50 (£2.47) provided either free of charge or at a concessionary
for recognised schools. rate 5% free and 5% concessionary. Unrecognised
The minimum wage for Delhi is set at Rs. 90.00 schools were slightly more generous in this regard
(£1.15) per day (2001 figures, Labour Bureau, than recognised schools—offering 8% of seats free,
Government of India, 2005) which translates to compared to 3% in the recognised schools (Table 7).
about Rs. 2160 (£27.69) per month (assuming 24
days work per month). That is, the median fees for 3.10. Teacher salaries
unrecognised schools are about 5% of the monthly
wage for a breadwinner on the minimum wage, We also explored the issue of teacher salaries with
while recognised school fees are about 9%. the Class 5 teacher from each of the stratified
random sample of schools. The average monthly
3.9. Private school philanthropy salary of a full-time teacher in a government school
was reported to be Rs. 10,072, compared to Rs.
However, not all students pay these fees. A 1360 in unrecognised and Rs. 3627 in recognised
notable feature of the private unaided schools is private unaided schools. The average salaries in
that, although they require fee incomes to survive, government schools are more than seven times
higher than in the unrecognised, and more than two
2
£1 ¼ Rs.78/- (11th October 2005 rate). and a half times those in the recognised private
ARTICLE IN PRESS
212 J. Tooley, P. Dixon / International Journal of Educational Development 27 (2007) 205–219
Table 8
Teacher salaries per pupil
Management type Mean monthly salary Mean class size Mean monthly salary Ratio of unit costs
of full-time teacher at per pupil (private unrecognised
Grade 4 (Rs.) base)
Table 10
(20.9%)
(47.1%)
(82.3%)
(54.3%)
(55.1%)
Unavailable
Own playground Total
Available Unavailable
9
8
51
50
118
Government 49 21 70
70.0% 30.0% 100.0%
(79.1%)
(52.9%)
(17.7%)
(45.7%)
(44.9%)
Private aided 1 18 19
Available
5.3% 94.7% 100.0%
Private unrecognised 3 70 73
34
9
11
42
96
4.1% 95.9% 100.0%
47 (67.1%)
58 (80.6%)
73 (72.3%)
178 (73.3%)
Unavailable
Private recognised 16 86 102
15.7% 84.3% 100.0%
Total 69 195 264
Tape recordersd
26.1% 73.9% 100.0%
23 (32.9%)
14 (19.4%)
28 (27.7%)
65 (26.7%)
Available
Note: w2 ¼ 98:168, df ¼ 3, Significant, po0.001.
(62.3%)
(88.9%)
(63.3%)
(72.2%)
(68.4%)
Unavailable
Library for use by childrenc
‘pucca’ buildings, while 14% of the government
schools were not. Regarding the provision of
162
43
16
38
65
playgrounds, it was found that 70% of government
schools had a playground compared with only 5.3%
of private aided schools, 4.1% of private unrecog-
(37.7%)
(11.1%)
(36.7%)
(27.8%)
(31.6%)
Available
22
25
75
2
8 (11.4%)
3 (4.1%)
6 (5.9%)
17 (6.9%)
70 (95.9%)
96 (94.1%)
228 (93.1%)
7 (9.9%)
2 (2.8%)
6 (5.9%)
15 (6.1%)
Blackboard availabilitya
70 (97.2%)
96 (94.1%)
230 (93.9%)
b 2
d 2
a 2
c 2
e 2
Unavailable
11(11.0%)
18 (26.5%)
34 (13.3%)
to 81% in government and private unaided
1 (5.3%)
4 (5.9%)
schools. (Only one private aided school had tape
recorders available for teaching).
(73.5%)
(94.7%)
(94.1%)
(89.0%)
(86.7%)
For one input, government schools had superior
Available
inputs to private unaided schools (Table 11):
221
50
18
64
89
Separate toilets for boys and girls (excluding
(92.9%)
(78.9%)
(76.4%)
(54.0%)
(72.4%)
Unavailable
18% of unrecognised private unaided have
separate toilets, compared with 79% of govern-
189
65
15
55
54
ment schools.
(21.1%)
(23.6%)
(46.0%)
(27.6%)
(7.1%)
Unavailable Available
private unaided schools were superior in provision
to government schools (Table 12):
17
46
72
5
4
(80.3%) 14 (19.7%)
Desks: In 87% of recognised and 90% of
(97.3%) 2 (2.7%)
(93.1%) 7 (6.9%)
(90.7%) 23 (9.3%)
unrecognised private unaided and private aided
Children’s toiletsd
schools, desks were available in the classroom,
compared to only 67% of government class-
rooms. That is, one third of the government
Unavailable Available
classrooms did not have desks available.
71
(89.1%) 11 (10.9%) 95
Chairs or benches for children: In 87% of
recognised and 94% of unrecognised schools,
(94.4%) 1 (5.6%)
(93.1%) 5 (6.9%)
Fans in the classroomc
(87.1%) 13 (12.9%)
(83.7%) 43 (16.3%)
Chairs in the classroomb
(94.4%) 4 (5.6%)
(90.3%) 7 (9.7%)
b 2
d 2
a 2
c 2
e 2
f 2
Gov
PA
22 (36.7%) 38 (63.3%)
25 (27.8%) 65 (72.2%)
47 (31.3%) 103 (68.7%)
Unavailable
compared to 95% of the private aided, 94% of
the private unrecognised and 89% of the private
recognised schools.
Unavailable Available
Librarye
unrecognised and recognised private unaided
schools only, for consideration in the final section.
65 (90.3%) 7 (9.7%)
It is widely acknowledged that a low-cost private
education sector is now serving low-income families
in developing countries. In part, it is suggested that
Available
this sector has emerged because of the low quality of
available government schools for the poor. How-
ever, there are also doubts raised about the
relatively quality of private provision for the poor,
0
explored the nature and extent of private schools
174 (99.4%)
73 (100%)
serving low-income families, and compared their
101 (99%)
inputs to those in government schools.
It would have been valuable to have been able to
physically count the numbers of children present
within the schools, to gain more realistic estimates
Blackboards in the classroomb
2 (2.8%)
6 (5.9%)
8 (4.6%)
more, more than one unannounced visit would have
been useful to gauge the extent to which teachers
were teaching over a prolonged period, and a
comprehensive survey of facilities within the school
7 (9.7%) 70 (97.2%)
8 (7.9%) 96 (94.1%)
15 (8.7%) 166 (95.4%)
b 2
d 2
a 2
c 2
e 2
Unavailable
more unrecognised private unaided than government
2 (2.7%)
7 (6.9%)
9 (5.1%)
schools (28% compared to 27%). Given that we
cannot be sure we found all the private unrecog-
nised schools, this gives a lower bound for private
sector involvement. The private unrecognised
71 (97.3%)
95 (93.1%)
166 (94.9%)
schools, however, were reported to be much smaller
Available
Toiletse
11 (11.0%)
Electricity in the classroomc
4 (5.9%)
15 (8.9%)
17 (9.8%)
Chairs in the classroomb
16 (9.2%)
b 2
d 2
a 2
c 2
e 2
Table 15
Inputs to schools (3), private unaided schools only—significant differences
Own playgrounda Separate toilets for boys Tape recordersc School computers used by
and girls where necessaryb childrend
inputs was there any significant difference between families unable to pay school fees are able to attend
recognised and unrecognised private unaided private schools, governments could finance their
schools—tape recorders, computers, separate toilets education through vouchers.’ (UNDP, 2003,
for boys and girls and playgrounds, two of which p. 115). The report gives examples of successful
are not subject to government regulation—suggest- schemes in Colombia and Pakistan, the latter aimed
ing that, in general, it is not government regulation at disadvantaged girls, that may be transferable to
and inspections that leads to higher quality provi- the Indian context. To a not insignificant extent, we
sion (Tables 14 and 15). found that the private schools themselves were
On other inputs, private unaided schools ap- already offering their own version of targeted
peared at least as good, or better than their vouchers—on average 10% of all places were
government counterparts. Reported pupil–teacher provided free or at a concessionary rate—although
ratios were much more favourable to private further research is required to ascertain to whom
unaided than government schools—highest in gov- these were offered, and whether concern for girls’
ernment schools, (at 55:1), more than twice as high enrolment featured. (Other research has indicated
as in both recognised and unrecognised private that gender was one factor in the distribution of
unaided schools (23:1). On four inputs, (the avail- scholarships in other parts of India, Tooley and
ability of blackboards, drinking water, library and Dixon, 2005a). Ways of extending or enhancing
tape recorders), government and private unaided such informal schemes might be considered a
schools offered a comparable range of facilities. For valuable way forward.
six other inputs (desks, chairs, fans, toilets, compu- Second, concerning the variable standards in
ters and electricity), private unaided schools, in- private unaided schools, others have argued that
cluding unrecognised ones, offered these facilities government regulation and inspection needs to be
significantly more often than government schools. strengthened in recognised private unaided schools,
Only on two inputs—playgrounds and separate and unrecognised schools brought under this
toilets for boys and girls (single-sex schools ex- regulatory remit (by ensuring that unrecognised
cluded) did government schools have superiority schools become recognised), (see for example,
over private unaided schools. Nambissan, 2003; Rose, 2002). However, other
Three ways forward are suggested by these research has suggested that, in Hyderabad, India,
findings. First, concerning the gender imbalance in government recognition may be achieved through
private unaided schools, one possible solution here, unofficial payments or bribes to inspectors,
if attendance at private unaided schools is consid- rather than achieving regulatory standards (Tooley
ered to be a valuable good that is currently and Dixon, 2005b). If this is also the case in Delhi—
dominated by boys, might be to consider the and other sources suggest that the corruption
introduction of targeted vouchers, aimed at girls. found in Hyderabad is likely to be repeated
UNDP (2003) explores this as a possible way across India (corruption is ‘endemic’ in India, Drèze
forward: ‘To ensure that children from poor and Sen, 2002, p. 53, which is ranked amongst
ARTICLE IN PRESS
218 J. Tooley, P. Dixon / International Journal of Educational Development 27 (2007) 205–219
the top ten most corrupt nations in the world Habyarimana, J., Das, J., Dercon, S., Krishnan, P., 2004. Sense
Mitra, 1998, Qp. 38), then this may not be an and Absence: Absenteeism and Learning in Zambian Schools.
immediate way of raising standards in private World Bank, Washington, DC.
Kingdon, G., 1996. Student achievement and teacher pay.
unaided schools. Third, an alternative approach Discussion Paper, No. 74, STICERD, London School of
(raised in Tooley and Dixon, 2005b), would be to Economics, August.
consider how private unaided schools could be Kingdon, G., Drèze, J., 1998. Biases in educational statistics. The
brought within systems of self-regulation and Hindu, 6 March.
accreditation to assist in school improvement, and Kingdon, G., Muzammil, M., 2003. The Political Economy of
Education in India: Teacher Politics in Uttar Pradesh. Oxford
so enhance the learning experiences of students University Press, New Delhi, Oxford, New York.
within them. Kremer, M., Mularidharan, K., Chaudhury, N., Hammer, J.,
Rogers, H., 2004. Teacher Absence in India (Available
from: /http://econ.worldbank.org/files/36660_Absenteeism.
India.Schools.June.2004.pdfS accessed October 2004).
References Mitra, C., 1998. The Corrupt Society: The Crimilization of India
from Independence to the 1990s. Viking Penguin India, New
Aggarwal, Y., 2000. Public and Private Partnership in Primary Delhi, India.
Education in India: A Study of Unrecognised Schools in Nambissan, G.B., 2003. Educational deprivation and primary
Haryana. National Institute of Educational Planning and school provision: a study of providers in the city of Calcutta.
Administration, New Delhi. IDS Working Paper 187, Institute of Development Studies.
Akhamadi, S.U., Suryadarma, D., 2000. When teachers are Nwagwu, C.C., 1997. The environment of crises in the Nigerian
absent: Where do they go and what is the impact on students? Education System. Comparative Education 33 (1), 87–95.
SMERU Field Report. SMERU Research Institute, Rana, K., Rafique, A., Sengupta, A., 2002. The Delivery of
Indonesia. Primary Education: A Study in West Bengal. TLM Books and
Alderman, H., Orazem, P.F., Paterno, E.M., 2001. School Pratichi (India) Trust, Delhi.
quality, school cost and the public/private school choices of Rogers, H.F., Lopez-Calix, J., Chaudhury, N., Hammer, J.,
low-income households in Pakistan. Journal of Human Cordoba, N., Kremer, M., Muralidharan, K., 2004. Teacher
Resources 36 (2), 304–326. absence and incentives in primary education: results from a
Alderman, H., Kim, J., Orazem, P.F., 2003. Design, evaluation, national teacher tracking survey in Ecuador. Ecuador:
and sustainability of private schools for the poor: the Pakistan Creating Fiscal Space for Poverty Reduction: A Fiscal
urban and rural fellowship school experiment. Economics of Management and Public Expenditure Review vol. 2,
Education Review 22, 265–274. 136–162 (Chapter 6).
Ali, M., Reed, T., 1994. A school and Parental Survey of Book Rose, P., 2002. Is the non-state education sector serving the needs
Provision Issues in NWFP. International Book Development of the poor? Evidence from East and Southern Africa. Paper
Ltd. prepared for DfID Seminar in preparation for 2004, World
Baurer, A., Brust, F., Hybbert, J., 2002. Entrepreneurship: a case Development Report.
study in African enterprise growth, expanding private Rose, P., 2003. From the Washington to the Post-Washington
education in Kenya: Mary Okelo and Makini Schools. Consensus: the influence of international agendas on educa-
Columbia Business School. Chazen Web Journal of Interna- tion policy and practice in Malawi. Globalisation, Societies
tional Business, Fall. and Education 1 (1), 67–86.
Census of India, 2001. Primary Census Abstract: Haryana, Delhi Save the Children UK, South and Central Asia, 2002. Private
and Rajasthan, vol. 2. Government of India, Delhi. sector involvement in education: a perspective from Nepal
Chaudhury, N., Hammer, J., Kremer, M., Muralidharan, K., and Pakistan. Submission to ‘‘The Private Sector as Service
Halsey Rogers, F., 2004a. Provider Absence in Schools and Provider and its Role in Implementing Child Rights’’, Office
Health Clinics (Available from: /http://worldbank.orgS of the High Commissioner for Human Rights, Geneva, 20
accessed October 2004]. September.
Chaudhury, N., Hammer, J., Kremer, M., Mularidharan, K., The Probe Team, 1999. Public Report on Basic Education in
Rogers, H., 2004b. Roll Call Teacher Absence in Bangladesh. India. Oxford University Press, Oxford and New Delhi.
World Bank, Washington, DC. Tooley, J., Dixon, P., 2005a. An inspector calls: the regulation of
De, A., Majumdar, M., Samson, M., Noronha, C., 2002. Private ‘budget’ private schools in Hyderabad, Andhra Pradesh,
schools and universal elementary education. In: Govinda, R. India. International Journal of Educational Development 25,
(Ed.), India Education Report: A Profile of Basic Education. 269–285.
Oxford University Press, Oxford and New Delhi, Tooley, J., Dixon, P., 2005b. Is there a conflict between
pp. 131–150. commercial gain and concern for the poor? Evidence from
Drèze, J., Sen, A., 2002. India: Development and Participation, private schools for the poor in India and Nigeria. Economic
second ed. Oxford University Press, New Delhi, Oxford. Affairs 25 (2), 20–27.
Glewwe, P., Illias, N., Kremer, M., 2004. Teaching incentives. Tooley, J., Dixon, P., 2006. ‘De Facto’ privatisation of education
National Bureau of Economic Research Working Paper. and the poor: implications of a study from sub-Saharan
Government of India, 2005. Labour Bureau, Minimum Wages, Africa and India, Compare 36(4).
(Available from /http://labourbureau.nic.in/wagetb/htmS) Tooley, J., Dixon, P., 2007. Private education for low income
accessed 10.10.05. families: Research from a global research project. In:
ARTICLE IN PRESS
J. Tooley, P. Dixon / International Journal of Educational Development 27 (2007) 205–219 219
Srivastava, P., Walford, G. (Eds.), Private Schooling in Less World Bank, 2002b. India, Karnataka: financing education in the
Economically Developed Countries: Asian and African context of economic restructuring. Human Development
Perspective. Symposium Books, Didcot. Sector Unit, South Asia Region, Report No. 24207-IN,
UNDP, 2003. Human Development Report 2003. United June 13.
Nations Development Programme, New York. World Bank, 2003. Making Services Work for Poor People:
Watkins, K., 2000. The Oxfam Education Report. Oxfam in World Development Report 2004. World Bank/Oxford
Great Britain, Oxford. University Press, Washington, DC/Oxford.
World Bank, 2002a. India, Karnataka: secondary education and World Bank, 2004. Papua New Guinea: Public Expenditure
the new agenda for economic growth. Human Development and Service Delivery (Discussion Draft). World Bank,
Sector Unit, South Asia Region, Report No. 24208-IN, Washington, DC.
June 13.