While it is widely believed that social motivation can promote efficiency in public service delivery, its role in provision of schooling is little understood. Using Nepal as a case study, this article provides empirical insights into how socially motivated, not-for-profit private schools could enhance excellence in schooling – both in absolute term, and in terms of student performance relative to school expenditure per student.
Many prominent economists have argued that social motivation can be an important driver of efficiency in public goods provision (Andreoni 1989, Besley and Ghatak 2005, Bennett and Iossa 2010). In developing countries, not-for-profit and other private schools co-exist with public schools, catering to various income groups from the very rich to the very poor. A relevant and unexplored question is: How good are not-for-profit private schools relative to other schools, in terms of delivering secondary education? We examine this question in the context of Nepal (Pal and Saha 2018).
Non-profit vs. for-profit private schools
All private, unaided schools in India, whether established by trusts, societies (educational, charitable, or religious), or private entrepreneurs, are registered as not-for-profit organisations or charities under the Societies Registration Act, 1860 or similar state acts, even though it is common for many of these schools to seek profit in practice. Such pursuance of profit is often masked by a two-tier legal structure by setting up a trust that runs the school, and a company that owns the assets (land, building, etc.) and provides services (management and technology), thus receiving payment in return.
In Nepal, however, thanks to the fifth amendment of the Nepalese Education Act, 2028 (that is, 1971) enacted in 1992, there is an official distinction between for-profit and not-for-profit private schools. A private school must be registered as either a (not-for-profit) ‘trust’ or a profit-seeking (and thus tax-liable) ‘company’. Neither would receive any government support including subsidy on purchase of land.
Private schools that were set up before 1992 and were operating formally or informally were migrated into one of these two registration categories. In this reassignment of status, schools that were previously using government-owned or ‘privately donated’ building or land, were forced to be trust schools. This particular aspect of the reform allows us to regard the registered trust schools as reliably socially motivated. Schools that were set up after 1992, had the choice of being company schools, if their motivation was profit-seeking. Schools that were set up before 1992 with some public support and wanted to continue to make profits, had to alter their operations and objectives, due to the force of regulation. If the trust status was a convenient disguise to make profit, we would have seen faster growth of trust schools than company schools. But that is not the case.
Analysing efficiency across different types of private schools in Nepal
We analyse Nepal’s School Leaving Certificate (SLC) exam data on nearly 7,000 students from 2002 to 2004, across 14 trust schools, 108 company schools, 216 fully government-aided schools, and 94 partially aided government (PA) schools. Not surprisingly, trust schools are in minority; their share in total schools is 3.2% and in total number of students is 8.5%. In comparison, company schools have only 11% of the student share, despite having 25% share in total schools. Trust schools are drawn from 11 districts representing the whole country, including the remote Far Western Development Region.
On average, company schools are the most expensive. They charge an annual fee which is nearly twice that of trust schools and seven times that of government schools (which are not entirely free, but inexpensive). The student-teacher ratio is smallest in company schools and largest in government schools. Their spending per student is also the highest, but only 10% more than that of trust schools, on average. A comparison of the student fees earned and school expenditure shows a small deficit for trust schools, but nearly 100% profit for company schools. It also turns out that trust schools and government schools have more trained teachers (as percentage of their total teachers) than company schools. Looking at the SLC marks data in isolation, one may be tempted to conclude that company schools are the best because students from these schools get higher marks, on average, than any other school. But such conclusions may be hasty and even erroneous, unless we take into account all other factors that contribute to student exam performance; the main aim of our study is to discern the contribution of various factors and isolate the pure school effect1.
We obtain a measure of ‘absolute efficiency’ (AE) for each school type. We also define an alternative ‘value for money’ measure of efficiency, which we call ‘relative efficiency’ (RE), by dividing the student (standardised) test score by the school expenditure per student. These estimates summarised in Table 1 below indicate that company schools are, in general, not the most efficient schools. In fact, in terms of both AE and RE scores, trust schools are the best schools (their coefficient estimate being the largest, positive, and significant), with company schools coming a distant second in most cases. Columns (1) and (4) show that relative to government schools, trust schools’ coefficient is the largest’, and it is significantly greater than that of company schools. PA schools are at the bottom of the ranking.
Table 1. Absolute (AE) and relative (RE) efficiency estimates of heterogeneous private schools
2SLS – AE: Standardised SLC test scores |
2SLS-RE: standardised scores/ school expenses per student |
|||||
Variables | (1) All | (2) Rural | (3) Urban | (4) All | (5) Rural | (6) Urban |
IV PA School | -0.038*** | -0.028** | -0.035* | -0.010 | -0.020 | 0.030 |
(0.012) | (0.011) | (0.021) | (0.019) | (0.020) | (0.039) | |
IV Company School | 0.014* | 0.025*** | 0.006 | 0.013 | 0.008 | 0.025** |
(0.007) | (0.006) | (0.010) | (0.012) | (0.013) | (0.012) | |
IV Trust School | 0.060*** | 0.051*** | 0.031 | 0.086*** | 0.096*** | -0.005 |
(0.017) | (0.011) | (0.034) | (0.021) | (0.019) | (0.048) | |
Other controls | Yes | Yes | Yes | Yes | Yes | Yes |
Intercept | 0.013** | 0.012** | 0.030** | 0.007 | 0.020* | 0.020 |
(0.006) | (0.005) | (0.012) | (0.008) | (0.010) | (0.015) | |
District & SLC year dummies | Yes | Yes | Yes | Yes | Yes | Yes |
IV exclusion test: F(p-value) | 0.00(1.00) | 1.43 (0.24) | 0.55 (0.65) | 0.00 (1.00) | 0.19 (0.91) | 1.18 (0.31) |
Observations | 5,828 | 3,974 | 1,854 | 5,828 | 3,974 | 1,854 |
R-squared | 0.400 | 0.383 | 0.420 | 0.134 | 0.136 | 0.176 |
Notes: (i) Robust standard errors in parentheses; (ii) *** p<0.01, ** p<0.05, * p<0.1; (iii) IVs are the number of private schools, proportion of non-Hindu population, and walking time to the nearest private, aided school in the municipality, respectively for company, trust, and PA schools.
=Due to data limitations, we cannot precisely explain why trust schools are the most efficient. We may, however, conjecture that their non-profit objectives are a driver for their student-centric spending and exam-targeted efforts. PA schools are the worst performers possibly because of their inability to hire trained teachers.
Secondly, for absolute efficiency, there is a rural-urban dichotomy. In rural areas, the efficiency ranking is sharp and all four types of schools are differentiated, with trust schools at the top and the PA schools at the bottom (column 2). In urban areas, surprisingly, there is no difference between government schools, trust schools, and company schools (column 3); all three are equally efficient.
Thirdly, in terms of relative efficiency, trust schools’ performance stands out even stronger in rural areas (columns 5), as it is the only type of school that performs better than government schools. It is only in urban areas that trust schools fails to shine (column 6); company schools are the only type of school there that exhibit better value for money than government schools.
Concluding remarks
While studies (for example, Chudgar and Quin 2012, Thapa 2015, etc.) have looked at private-versus-public school efficiency in India and Nepal, the diversity of private schools with a focus on not-for-profit schools has not been explored. Our analysis shows: (i) that not-for-profit private schools can be an effective mode of delivering quality education; (ii) in Nepal, such schools appear to be better than both government schools and for-profit schools; and (iii) such schools demonstrate value for money, justifying transferring resources to them in proportion to their students.
We can also draw lessons for India. In Nepal, private schools as a whole (combining trust and company schools) perform very well in comparison to government schools. In contrast, studies in India (Chudgar and Quin 2012, Singh 2015) do not generally find strong differences between private and government schools. One reason could be that the broad private classification hides crucial heterogeneity within the private sector. It is worth considering a Nepal-type reform, where some private schools can be legally permitted to be run as profit-making ventures, so that the genuine non-profit schools gain credibility and are able to raise more funds. This is simply because policymakers and donors can be more confident if the contracted party can be trusted, even if contract monitoring is difficult. Such reform will also help in achieving transparency in private schooling. Truly non-profit schools will be able to provide good quality, value-for-money education for low-income families.
Notes:
- Any empirical estimation of school efficiency should include a model of school choice; otherwise the estimates of the school’s contribution to student performance would be biased. Given that school choice is rarely random and parents choose schools based on a number of factors such as parental characteristics, child’s ability, and so on – it is a priori difficult to predict direction of estimation bias. Hence, instead of ordinary least square (OLS), we use two-stage least square (2SLS) instrumental variable (IV) method (see note to Table 1; also see Pal and Saha 2018) that helps minimise school selection bias that could otherwise contaminate school efficiency estimates.
Further Reading
- Andreoni, James (1989). “Giving with impure altruism: applications to charity and Ricardian equivalence”, Journal of Political Economy, 97: 1447–1458.
- Bennett, John and Elisabetta Lossa (2010), “Contracting out public service provision to not-for-profit firms”, Oxford Economic Papers, 62, 784-802.
- Besley, Timothy and Maitreesh Ghatak (2005), “Competition and incentives with motivated agents”, American Economic Review, 95: 616-636, DOI: 10.1257/0002828054201413
- Chudgar, Amita and Elizabeth Quin (2012), “Relationship Between Private Schooling and Achievement: Results from Rural and Urban India”, Economics of Education Review, Vol. 31: 376-90, doi :10.1016/j.econedurev.2011.12.003. Available here.
- Pal, Sarmistha and Bibhas Saha (2018), “Enhancing Excellence: Socially Motivated Private Schools of Nepal”, Journal of Development Studies (forthcoming).
- Singh, Abhijeet (2015), “Private school effects in urban and rural India: Panel estimates at primary and secondary school ages”, Journal of Development Economics, Vol. 113, March 2015, pp. 16-32.
- Thapa, Amrit (2015), “Public and Private school performance in Nepal: An Analysis using SLC examination”, Education Economics, 23(1): 47-62. Available here.
Comments will be held for moderation. Your contact information will not be made public.