Enrolment and social exclusion
Enrolment rates at the primary level in Latin America do not vary much by income quintile (Figure 3.10).32 Most countries secure good compliance with mandatory primary education, through public policies to guarantee universal access and the success of conditional cash-transfer programmes. It is probably also the case that in most countries child labour for this age group is not cost-effective and relevant laws better enforced.
Figure 3.10. Enrolment rate by income quintiles
Unfortunately, by the time these children reach secondary education, enrolment rates start to exhibit a strong correlation with economic status.33 The situation deteriorates again at the tertiary level to the point that tertiary education in Latin America is still mainly associated with the affluent. Post-primary educational enrolment in Latin America is still highly related to a family’s economic background.
Box 3.1. Private expenditure on education and educational mobility in the Andean countries
Parents paying for private education is common in most Latin American countries. Private schools are perceived to provide higher quality and people in Latin America, as elsewhere, see education as an important way to move up the social ladder – 56% of them in the 2006 Latinobarómetro survey said it was the most important factor determining success in life. Middle- and high-income families back this expressed view up by devoting significant financial resources to sending their children to private establishments.
This box looks at four Latin American countries, chosen because of the availability of suitable data from their national household surveys: Bolivia (2005), Colombia (2008), Ecuador (2006) and Peru (2006). The questions it seeks to answer are: do the middle sectors make a special “financial effort” (measured as the portion of household income devoted to education related expenses), and what reward do they get for their investments, in terms of improvement in educational achievement?
Sending children to school involves costs – even if they are attending public schools. The household surveys identify these and allow them to be compared across different socio-economic groups; items included are the cost of uniforms, school supplies, books, transport, food and other linked expenses. To these can be added school registration and tuition costs, where appropriate. On the basis of these data, low-income families make the largest effort relative to income in all countries except Peru, where the proportion of income allocated to education rises with income (Figure 3.11).
Figure 3.11. Percentage of household income devoted to education
In absolute terms, each middle-sector household spends USD 57 a year in Ecuador, USD 100 in Colombia, USD 120 in Bolivia, and USD 420 in Peru (on a purchasing-power parity basis). In each country expenditure by middle-sector households is more than twice that of disadvantaged households but only around a third that of affluent. Overall, the middle sectors seem to make an intermediate investment effort in relative and absolute terms in the four countries.
What are the payoffs to these investments? Econometric analysis of the schooling gap of 15-year olds in these countries shows that household expenditures significantly decrease the schooling gap in Bolivia and Peru, while for Colombia and Ecuador the effect is not significant. However, these national results hide important differences across income groups. While in Colombia and Ecuador expenditure returns for the middle sectors are significantly higher than for the disadvantaged and affluent, in Bolivia and Peru expenditure returns for the middle sectors are not significantly different from those of the disadvantaged.
Private schools and social exclusion
Looking at the proportion of students in each income quintile that attend private schools reveals interesting differences in the pattern of enrolment (Figure 3.12). At the tertiary level, between about 35% and 50% of each income group attend private establishments. This contrasts with the division evident at both primary and secondary levels, with the affluent going to private schools and the disadvantaged and middle sectors concentrated in the public system.
Figure 3.12. Percentage of students enrolled in private establishments by income quintiles
This shape is consistent with the relatively poor performance of the region’s schools in the PISA measures of social inclusiveness (Figure 3.13).34 The six countries from Latin America are clustered at the bottom of the distribution, less inclusive than either the OECD average or most of their developing peers.
This low inclusiveness reduces inter-generational social mobility in two ways. Where private education is better – as it usually is – then the access problem for middle sector and disadvantaged children is compounded by the lower yield in the labour market for each year of their education. Then they lose again when lack of mixing across class groups compromises their social networks.
Figure 3.13. Social inclusion in secondary schools by country
There is evidence for this in data from Peru which show that returns to private education are significantly higher than to public in terms of wage-earning power, and have been increasing over the last two decades.35 The difference is greatest at the primary and secondary level, precisely where the class groups are most split. In assessing the causes of this it is difficult to disentangle the value of access to "high-value" social networks from differences in the quality of education. However, there is some suggestive evidence that both problems play their role in the region (see Box 3.2).
This selectiveness in private schooling might work to society’s advantage if the private and public schools play to their respective pupils’ strengths. But plotting the inclusiveness of a country’s education system against its average PISA science test score shows this is not the case (Figure 3.14). Inclusiveness is generally associated with better overall educational outcomes, and more-detailed analysis shows that this relationship is statistically significant. Nor does Latin America buck this trend – all six countries are in the "bad" quadrant of below average performance even given their low levels of inclusiveness.36
Figure 3.14. Correlation between PISA science test scores and index of inclusion
The close association between differences in the socio-economic background of secondary school students at private and public institutions and the differences in their average science test scores perhaps show why parents persist with private education when they can afford it (Figure 3.15).37 The differences in both socio-economic background and test scores of students in Latin America are huge – even compared with other developing countries. For example, in Brazil, students in the private system on average perform better than those in the public system by a little more than 100 points. This implies that a student in the private system in Brazil has additional cognitive skills approximately comparable to almost three extra years of education.38
Figure 3.15. Private and public education: differences in performance and socio-economic status
The problem, as we have noted, is that this outperformance is not the result of private schools in Latin America being particularly good. If they did as well as the average outside the region would imply, their test score differences would be significantly higher: in Brazil the advantage would be 136 instead of 106
(a difference equivalent to almost an additional year of schooling); in Uruguay 124 instead of 80; in Mexico 125 instead of 53; in Colombia 80 instead of 38. Only in Argentina and Chile do they perform close to the average.
In summary, the current education framework in the region promotes selection for those who can afford it. But by itself selection tends to depress overall educational outcomes, and the region’s private schools compound this by failing to make the most of their privileged intake. Nevertheless, selection succeeds in boosting the relative position of those in the upper layer. A system that under-delivers and comes at the price of perpetuating inequalities will therefore continue to be something that parents aspire to – at least until policy provides them with an attractive alternative.
Box 3.2. The effect of parental background on returns to education: the case of Chile
Most household surveys in Latin America contain little information on the parental background of those people who are active in the labour market. This makes it difficult to evaluate inter-generational mobility issues and their relationship with wage earnings. However, in Chile the 2006 National Socio-economic Characterisation Survey (CASEN, Encuesta de Caracterización Socioeconómica Nacional) elicits information on the highest level of education attained by the father and mother of all surveyed individuals. This can be used to perform an econometric estimation of the return to education with the aim of exploring the effects of socio-economic background on labour-market earnings. Variables include years of education, as well as age and the square of age as a proxy for experience-related human capital and also to allow for decreasing marginal returns over time.39
The wage equations are estimated for three different levels of parental education: high (tertiary education completed), medium (secondary education completed) and low (primary completed or less). Overall, the results show significant differences across parental backgrounds (Figure 3.16). One additional year of education yields more than twice as much for a person from a high or medium background as for a similar person whose parents have a low level of education. These differences are not only statistically significant, but also significant from an economic point of view. For example a man (woman) with 12 years of education from a high-education family would earn around 1.3 times (1.5 times) as much as their analogue from a low education family. Even for those in the middle the implied differences are large: 73% for men and 85% for women.
Figure 3.16. Private returns to education by parental educational background in Chile
Of course, it is difficult to separate the effects of differences in the quality of education from other factors that may be at work, such as network effects, early childhood factors that influence the ability to learn (including pre-school education, as well as exposure to reasoning practices and language skills at home), or even plain discrimination (since parental educational background and social class is often associated with race, for example). Nevertheless, a paper by Núñez and Gutiérrez (2004) found that returns in Chile for upper-class professionals were around 50% higher than for professionals from less-favoured socio-economic backgrounds, even after controlling for ability. Even though the returns to tertiary education are significant for individuals that do not belong to the upper class – by itself some support for the idea of meritocracy – this 50% gap is larger.