How Free is Free Education? Private Household Spending on Education in Sri Lanka: 1985-2010

Photo courtesy Rose Sri Lanka

Sri Lanka’s Free Education Scheme aims to provide each child an equal opportunity at educational success, but there is debate over to what extent it succeeds in this endeavor. Notably, some critics are concerned that private spending on education (for private and international schools, universities, and tuition) may be diluting the Free Education Scheme’s ability to provide equal opportunities to all. My doctoral dissertation is composed of three studies exploring these issues. In my first study, summarized here, I described patterns of educational stratification in Sri Lanka from 1985 to 2010. This article summarizes the second of the three studies, describing private household educational expenditure over the same period.

Does private education spending cause educational stratification?

In recent years many developing and developed countries have seen a strong trend to privatize education systems that had previously been completely run by the state. In Sri Lanka, advertisements for international schools, private universities, and tuition services fill newspapers and billboards, while heated discussion of these controversial developments appear in newspapers’ opinion pages and on television talk shows. Historically, free education has been one of the core services that the Sri Lankan state provides its citizens, and the fact that this opportunity is provided to all is a matter of national pride. For this reason the recent expansion of what some call a de facto privatization of education in Sri Lanka is contentious.

In particular, scholars and social commentators suggest that private education services may be a root cause of educational stratification, and that this mechanism may be increasing in strength over time.[i] My previous article demonstrates that, contrary to the claims of the free education scheme, not all children have equal chance of achieving educational success. Rather, the children of socioeconomically advantaged families have better educational outcomes than those of disadvantaged families. This study will gather evidence about whether it is plausible that private household spending on education causes educational stratification. First, I examine the scope of private education spending to assess if it is conceivably widespread enough to cause the educational stratification described in the prior article. Second, I analyze which groups engage in private education expenditures and which groups spend more or less in order to gauge whether the differences in spending between advantaged and disadvantaged groups are substantial enough to drive stratification. I also describe how patterns of private spending have changed from 1985 to 2010, relative to government spending.

Defining “privatization” with regard to the data

I use nationally-representative household expenditure data from 1985 to 2010[ii] which reports educational spending in a variety of categories, which I will group as: substitutes, supplements, and school choice categories. The kinds of private education spending that cause the most controversy in the Sri Lankan media are “substitutes:” international schools, private schools, and private universities[iii] that families use instead of government schools. Spending on “supplements” for government-provided schooling, such as exercise books, school facilities fees, and private tuition fees, is less controversial, but still periodically garners complaints by free education purists. Of these, spending on tuition may be the most educationally substantive because this spending pays for instruction, while the other categories secure the pre-requisites for learning.

Two categories of spending, transportation and boarding, may indicate a family has substituted the local government school that the children is zoned to attend for another school. Government schools are widely geographically distributed in Sri Lanka, so it is rare that the distance to the nearest government school requires transportation or boarding. Spending on transportation or boarding may be a signal that the family has gained entrance for their child into a school they preferred to the local school, either through excellent exam performance or through irregular means such as a falsified address or a bribe. Families likely believe the farther-away school will give their child better instruction than the local one; if they are correct, spending on these categories may impact educational outcomes and stratification. I will describe spending on transportation and boarding as “school choice categories” for the sake of brevity, although the data only make it possible, not certain, that families have indeed exercised school choice.

Descriptive statistics on educational spending

Table 1, below shows what percent of households with school-aged children report any spending in a number of categories, for each wave of the survey data.

Table 1:Prevalence of kinds of educational expenditures
Spending category 1986 1996 2007 2010
Notebooks/stationary 56% 51% 67% 65%
Exercise books 5% 6% 3% 4%
School facilities fees 18% 9% 20% 16%
Tuition fees 17% 25% 52% 55%
Pre-school fees 15%
Other educational expenses 3% 5% 24% 26%
Private school fees 5% 3% 3% 4%
Higher education 1% 1%
Professional/technical fees 2% 1%
School Choice Categories
Boarding fees 1% 1% 1% 1%
Education-related transport 15% 27% 25%
Total educational expenditure 65% 66% 81% 83%
Source: Analysis of HIES and LFSES data. Blanks indicate that data were not yet collected for that category. Calculated among all households with children aged 6-21 in all categories except pre-school, which was calculated among households with children aged 0-5.

Table 1 shows that, for any year, between 1% and 67% of households engage in any given kind of spending. Participation is low—below 25%—in all categories except notebooks and tuition. Notably, participation is below 5% in substitute categories (private school fees, higher education, professional/technical) and in boarding. Over time participation rates are stagnant for most categories, but rise for notebooks, tuition, and “other expenses.”

Table 2, below, shows the average expenditure amount per school-aged child in households with any spending, and all amounts are adjusted for inflation and shown in 2010 Sri Lankan Rupees, adjusted by a spatial price index that accounts for district level variation in pricing.

Table 2:Among households with any spending, average annual household expenditure per school-aged child

(in 2010 LKR, adjusted by spatial price index)

Spending category 1986 1996 2007 2010
Notebooks/stationary 915 1,383 2,154 1,932
Exercise books 1,728 1,967 2,724 2,097
School facilities fees 446 680 643 772
Tuition fees 2,819 4,937 5,511 5,765
Pre-school fees 2,778
Other educational expenses 2,436 4,541 4,973 5,244
Private school fees 3,419 5,749 10,351 14,193
Higher education 31,931 28,127
Professional/technical fees 22,732 38,221
School Choice Categories        
Boarding fees 23,663 21,322 23,959 22,472
Education-related transport 3,663 5,388 4,830
Total educational expenses 2,439 5,013 9,379 9,736
Education expenditure as a percent of total expenditure 3% 5% 7% 8%
MOE per pupil expenditure 12,223 18,223 29,142 24,410
Source: Analysis of LFSES and HIES data, except MOE per pupil expenditure (drawn from Central Bank Annual Report statistics), added for reference purposes. Blanks indicate that data were not yet collected for that category. Calculated among all households with children aged 6-21 in all categories except pre-school, which was calculated among households with children aged 0-5.

For comparison purposes I also show the Ministry of Education’s per pupil expenditures, adjusted for inflation, and bolded at the bottom of Table 2. Expenditure in some categories is high (private school fees, boarding fees, higher education, professional/technical education), but few households spend at all in these categories. Tuition is the only category that is widespread where families spend substantial amounts that rise over time. Total educational expenditures combined across categories are also substantial and rise over time. In 1985-86 they account for 24% of MOE’s per pupil expenditures, and by 2009-10 this figure has risen to 40%. Some portion of MOE spending goes to overhead costs, where by contrast a higher proportion of private household spending may go directly to instruction. If we had data on spending that went directly to instruction for MOE and private households, household spending could account for an even larger proportion than 40%.

Characteristics of households that spend on education

I conducted regression analyses to examine the extent to which participation and spending amounts vary with socioeconomic advantage and ethnicity. This analysis can shed light on whether disproportionate spending by certain groups may impact educational stratification.

Not surprisingly, advantaged households are more likely to spend money on educational costs and, among those who spend, advantaged groups spend more. For example, households where the head has post-graduate education are 3.25 times more likely to spend money on tuition than households where the head has primary education. Further, among households where there is at least some spending on tuition, the expected spending is 123% higher in households where the head has post-graduate education than in households where the head has primary education. Because it is widespread (unlike private schools and universities) and has a strong association with advantage, tuition is plausibly an important mechanism causing educational stratification. This conclusion assumes that tuition has a positive impact on students’ educational outcomes; I will assess this question in a following article.

For most spending categories, participation rates grew faster in disadvantaged households than in advantaged households. If education expenditure causes stratification, it could be that over time stratification is lessening because the educational advantages conferred by these expenditures spread beyond advantaged households to a broader and broader group. On the other hand, spending amounts grew faster in most categories for advantaged households than for the disadvantaged. If spending more on education expenditures—for example, hiring a tuition master with better qualifications who charges higher fees—leads to better education outcomes, this trend could lead to greater stratification over time.


Overall, I conclude that private education expenditures in Sri Lanka are substantive in scope and a phenomenon worthy of further study. I sought to assess whether household educational expenditures might cause educational stratification along lines of socioeconomic advantage. I find substitutes and school-choice categories show particularly strong associations with advantage, but because they are quite rarely used it is unlikely that they have a large impact on educational stratification. Tuition, on the other hand, because it is broadly used and has a moderate association with advantage, is potentially one mechanism causing Sri Lanka’s educational stratification. My final study explores whether tuition causes students to earn higher marks on exams. Time trends show that a broader and broader swath of households is engaging in private education spending over time, but at the same time advantaged families are increasingly spending more than disadvantaged families. The potential implications for stratification are contradictory, and should be investigated through further research.

Finally, this study has shown that while household education expenditure in Sri Lanka is substantial, private education is nonetheless limited in scope. Tuition may be a mechanism of stratification, but it is a supplement that only impacts students for a few hours each week. Full substitutes for government schools and school choice options are not widely used. The bulk of children are primarily educated in government schools through regular channels. As compared to J.R. Jayawardena’s 1945 charge that “our educational structure has created… two different nations,”[iv] in terms of expenditures today’s educational structure is relatively unified and has only low levels of privatization.


[i] In Evolving Inequality of School Attainment in Sri Lanka, authors Harsha Aturupane and Anil Deolalikar, suggest that, “The intergenerational transmission of schooling is an important reason why schooling inequalities persist, or even widen, over time” (page 4), but beyond private spending and intergenerational transmission they list no further potential causes of inequality of schooling outcomes. See also Samarakoon, A.K., “Democracy in an Oligarchy,” in Groundviews, 3 October, 2012.

[ii] Household Income and Expenditure Survey (1995-96, 2006-07, and 2009-10 waves) and the Labour Force and Socio-Economic Survey (1985-86), conducted by the Sri Lanka Department of Census and Statistics. For all three waves of the HIES surveys the districts of the Northern and Eastern provinces were either not included or were sampled on a shorter timeline because these areas were most directly impacted by the ethnic conflict, so this study’s findings cannot be considered representative of all of Sri Lanka, but rather only the portions sampled.

[iii] It is not clear how to classify spending on higher education, or professional/technical education because, while in the majority of cases households are purchasing a private substitute for government services, the government also offers limited services at these levels and my data may include spending on nominal fees for these.

[iv] Quoted in Jayasuriya, J. E. (1969). Education in Ceylon before and after independence: 1939-1968 (1st ed.). Colombo: Associated Educational Publishers. Page 67