A Comparative Institutional Analysis of Government, NGO and Private Rural Primary Schooling in Pakistan

Abstract

This paper draws on the experience of private sector and NGO schools to identify lessons for the government sector, the main provider of basic education in Pakistan. In a principal-agent framework, we compare the institutional effectiveness of rural primary schooling delivery by the government with the NGO and private sectors. Our main findings are that the NGO schools were the most successful in many respects and that ‘good management’ and/or ‘good leadership’ are the key factors for sound schooling. Further, if meaningful ‘participation’ is to be achieved in government schools, the power relations between administrators, teachers and parents need to be addressed.

S'appuyant sur des expériences d'écoles du secteur privé et d'ONG, cet article identifie des leçons pour le secteur gouvernemental, le principal pourvoyeur d'éducation de base. Par le biais d'un cadre ‘agent-principal’, nous comparons l'efficacité institutionnelle du gouvernement dans l'offre des écoles primaires rurales, à celle des ONG et du secteur privé. Nos principales conclusions sont que pour de nombreux aspects, ce sont les écoles des ONG qui ont le plus de succès et que la ‘bonne gestion’ et/ou le ‘bon leadership’ sont les ingrédients clés d'une scolarisation solide. De plus, si une “participation” significative doit être atteinte dans les écoles gouvernementales, les relations de pouvoir entre les administrateurs, les enseignants et les parents doivent être abordées.

Notes

Shahrukh Rafi Khan is Visiting Professor of Economics, Mount Holyoke College, 50 College Street, South Hadley, MA 01075, [email protected] Sajid Kazmi is Head of the Advocacy Unit, Sustainable Development Policy Institute; Zainab Latif is a research associate at the Vera Institute of Justice, New York. Thanks are due to The Asia Foundation, Pakistan, for supporting this research. Thanks are also due to Eric Jensen for encouraging the research, and to Haris Gazdar and the anonymous referee of this journal for very useful comments on earlier drafts.

 1. UNDP [1999: 136] and [2002: 151].

 2. UNDP [2002].

 3. UNDP [1998: 163].

 4. While the ‘non-government’ sector has made major inroads in the provision of primary schooling in urban areas, the government sector still accounts for the bulk of total rural primary enrollments as will be documented below.

 5. Interest in this subject has grown. Refer to Kardar [2001], Andarabi et al. [2002] and Burnett [2002].

 6. We utilized sampling frames made available by the Society for the Advancement of Education, SAHE [1997] and Trust for Voluntary Organizations, TVO [1994].

 7. Formal NGO schools are likely to be found in the larger localities and government schools in such localities tend to perform better than those in smaller localities. Thus, there may be an upward bias in the performance of government schools in our sample. About three-quarters (96 of the 129) of the schools were mixed, 18 were all-girl and 15 were all-boy schools.

 8. A complete description will be made available by the authors on request.

 9. In a nutshell, the format required the field-team to report whether or not they considered the school to be a success or failure on criteria including the state of discipline and confidence of students, motivation, dedication, training and experience of teachers, whether students and/or teachers cheated in the tests, physical facilities of the school, availability of school supplies, and the quality of school administration and management. In addition, they were to look for other household, school and community factors that could contribute to success or failure.

10. For a concise description refer to Stiglitz [1998].

11. The corporate scandals of 2002 in the US show how difficult it is to align principal-agent interests.

12. Hirschman [1970].

13. girls and boys respectively. By 2001–02, rural primary boys and girls enrollments in government schools for the upper quintile were 64 and 69 per cent as a percentage of total primary enrollments. By contrast, these numbers were respectively 92 and 90 per cent for the lowest quintile. Government of Pakistan [1998, 2002].

14. Refer to the section called Comparative Analysis: Quantative for details on the socio-economic background of students by school type.

15. Refer to Khan [2001] for an account of devolution and education in Pakistan.

16. NGOs face considerable donor pressure to rapidly attain financial sustainability.

17. All schools are subject to the government curriculum unless they are preparing children for the British-administered O and A level exams, which is the case for elite urban schools. Independent analysis by Nayyar and Salim [2003] demonstrates that the current curriculum distorts history, spawns communalism, and promotes war and the military and a narrow and militant version of Islam. The last minister of education buckled under pressure from Islamic groups in an attempt to reform the curriculum and made matters worse while the new minister of education, a former head of Pakistan's Inter Services Intelligence Bureau, appears to be more forceful and has put the ministry of education on notice regarding a review of education policy, textbooks and curriculum for reform by start of school term in the summer of 2006. Thanks are due to A.H. Nayyar for this information.

18. Thanks to this journal's anonymous referee for pointing this out.

19. Thanks to the editor of the journal for pointing this out.

20. Data on students, teachers and basic facilities in government and non-government schools are available in Government of Pakistan [1998: 66–7] and Government of Pakistan [2003: 47–62] .

21. Our findings regarding the difference between government and non-government schools endorse the findings of World Bank [1998] that partly covered some of the same ground as this study. However, the differences in Table 1 may be due to household characteristics rather than the schools per se. As Appendix Table A1 indicates, NGO and private schools cater to the better-off rural households relative to government schools. This issue is the subject of a detailed quantitative exercise in a companion paper.

22. Eleven per cent of the fathers and 27 per cent of the mothers of government school students were illiterate compared to five per cent and five per cent for NGO schools and 16 per cent and 18 per cent for private schools respectively. These numbers are way below the national average because they represent the responses of parents of schoolchildren (randomly selected Grade 5 students). Also see Appendix Table A1.

23. As per usage throughout the paper, ‘non-government’ is used by the government to refer to both private and NGO schooling. In this regard, NGO could probably be improved upon as an acronym to suggest what it is supposed to be rather than what it is not. Public Interest Organization (PIO) is gaining some currency in Pakistan.

24. Government of Pakistan [1998: 33].

25. The mean teacher salary and experience is reported in Appendix Table A4.

26. The mean salary of government school teachers of Rs 3,567 was about equal to the mean monthly salary of unskilled workers (using a straight average of the daily wage for the national and four provincial capitals and multiplying by 30) and less than half the monthly salary of skilled workers like masons and carpenters [Government of Pakistan, 1999: 143]. However, government teachers are entitled to benefits like a provident fund, health facilities and pension that are not accessible to daily-wage workers.

27. The lack of monitoring of government schools is often mentioned as a cause of poor performance. However, there was a mean of 4.23 (standard deviation, 2.78) inspections of government schools compared to 3.03 (s. d., 1.99) and 2.54 (s. d., 1.24) for private and NGO schools. Also see Appendix Table A3.

28. We generated a physical quality index based on the availability of the following: boundary wall, desks, chairs, taats (mats), indoor teaching, electricity, fans, drinking water, washrooms (both availability and quality) and library. Based on this, the quality score ranged from 0 to a maximum of 12. The mean score on this index for government, private and NGO schools was 5.2 (s. d., 2.71), 9.0 (s. d., 1.49) and 10.1(s. d., 1.16) respectively. Also see Appendix Table A2.

29. See Appendix Table A4 for teacher characteristics in a comparative context.

30. Only four out of the 43 private schools had a PTA (parent teacher association) or SMC (school management committee) while this was the case for 29 government schools (mandatory) and 22 NGO schools (optional). Parents were represented in no government schools' SMC/PTAs and in only nine percent of NGO SMCs/PTAs. Refer to Appendix Table A3 and for more details to Khan and Zafar [1999].

31. See Appendix Table A3 for the mean monthly fees.

32. While it may be difficult to ju.stify subsidizing a commercial activity, the government could ensure that tax authorities do not harass private schools as seems to be happening according to press reports.

33. It was difficult to find a completely consistent pattern for parental involvement in schooling across the three types of schools. Sometimes, very educated parents were complaisant about very poor private sector schooling as though they had done the best by their children and need not worry further. While the poor and illiterate were generally unaware and disinterested, they sometimes complained vociferously about the poor service delivery of government schools.

34. For an account of NGO schooling in general refer to Baqir [1998] and for a specific example of community based schooling to A.H. Khan [1998].

35. A case in point is the successful pilot project of District Government, Rahim Yar Khan, which turned over 44 non-functional government schools for control, use and management to the NRSP (National Rural Support Program). Refer to Uddin [2004] for examples of this and other successful public-NGO partnerships.

Additional information

Notes on contributors

Zainab Latif

Shahrukh Rafi Khan is Visiting Professor of Economics, Mount Holyoke College, 50 College Street, South Hadley, MA 01075, [email protected] Sajid Kazmi is Head of the Advocacy Unit, Sustainable Development Policy Institute; Zainab Latif is a research associate at the Vera Institute of Justice, New York. Thanks are due to The Asia Foundation, Pakistan, for supporting this research. Thanks are also due to Eric Jensen for encouraging the research, and to Haris Gazdar and the anonymous referee of this journal for very useful comments on earlier drafts.