Vouchers for Private Schooling in ColombiaPDF version

Joshua Angrist
Eric Bettinger
Erik Bloom
Elizabeth King
Michael Kremer
Juan Saavedra
1,600 applicants to school voucher program
1998 - 2004
Policy Issue: 
Student Learning
Policy Issue: 
Student Participation
Policy Issue: 

In many countries, per capita income can be an important determinant of the quality of public education a child receives. This can have significant consequences in developing countries, where income levels tend to be low and as a result, educational resources, scarce. The view that private schools function better than public schools in low income countries has prompted calls for governments to experiment with programs to give the poor access to private schooling, often through vouchers. Little is known about the effects of providing private school opportunities for poor families in developing countries.

Context of the Evaluation: 

Many people in Colombia have low levels of education. Public education is widespread, and 78 percent of the poorest primary-school aged children are enrolled in school. But only 55 percent of those students subsequently enrolled in secondary school as of 1993.

The government established Programa de Ampliación de Cobertura de la Educación Secundaria (PACES), one of the largest voucher programs to date, providing over 125,000 students from poor urban neighborhoods with vouchers that cover more than half of the cost for private secondary schools in Colombia. PACES vouchers were worth only about US$190 at the time of the survey, initially set to correspond to the average tuition of low-to-middle cost private schools in Colombia’s largest cities. But because they did not keep up with inflation, the amount voucher recipients paid for private education averaged $340. This compares with only $58 dollars for public school parents. Participating private schools tended to serve lower-income pupils, and to have lower tuition than non-participating private schools. Schools with a vocational curriculum were also over-represented in the program. Pupil-teacher ratios and facilities were similar in public and participating private schools, and many of the teachers in the PACES private schools are retired public school teachers.

Details of the Intervention: 

Because the demand for PACES vouchers exceeds the supply, eligibility is determined by a random lottery, generating a natural experiment to examine the impact of school choice on educational and other outcomes. Survey data was collected in 1998 from 1,600 previous applicants from three cohorts primarily from Bogota, three years after they had started high school. The sample was stratified so that half those sampled were lottery winners and half were lottery losers, meaning that half of the students surveyed had been given the opportunity to attend a private high school. Surveys were conducted by telephone, as the winners and losers groups had similar levels of telephone access, as well as being statistically similar in age and sex across cohorts. Field workers asked about further schooling and drop-out rates, amount of time spent on education, later monetary investments made toward education, and public or private choice for secondary school.

Results and Policy Lessons: 

Impact on Education: Preliminary results showed no significant differences in school enrollment between lottery winners and losers three years after application. But lottery winners were 15 percentage points more likely to have attended private school, had completed an average 0.1 more years of schooling, and were about 10 percentage points more likely to have finished 8th grade, primarily because they were less likely to repeat grades. The results also show that lottery winners worked about 1.2 fewer hours per week than lottery losers.

Impact on Economic Returns: Analysis of the economic returns to the additional schooling attained by winners after three years of participating in the program suggests that the benefits from increased income greatly exceeded the tuition costs to families and the government. Annual earnings of parents in our sample were about $2,400 per worker, who had on average only 5.9 years of education. Since the average PACES applicant had already completed 7.5 years and was still in school at the time of the survey, it can be assumed that their expected earnings are $3000, even without the benefit of private schooling. If higher test scores have a grade-equivalent payoff, lottery winners’ wages can be expected to rise by $36 to $300 per year, for estimated earnings of up to $3,300. Discounted over applicants working lives, these benefits easily outweigh the combined educational resource cost to the government and household per lottery winner, which are estimated at no more than $195.

Long-Term Effects: Later, in 2005, researchers examined administrative records on registration and test scores from a government college entrance examination. Lottery winners were more likely to take the university entrance exam, a good predictor of high school graduation, since 90 percent of all high school graduates take the exam. Within the sample of voucher applicants, about 28 percent of students took the ICFES exam, and vouchers raised the exam registration rates about 7 percentage points. While test scores suggest vouchers had more positive effects than simply reducing grade repetition, no strong conclusions were able to be drawn about the long-term effects of private schooling on test scores.  

Related Papers Citations: 

Angrist, Joshua, Eric Bettinger, Erik Bloom, Elizabeth King, Michael Kremer and Juan Saavedra.  2002. “Vouchers for Private Schooling in Colombia: Evidence from a Randomized Natural Experiment,” American Economic Review 92(5): 1535–58. 

Angrist, Joshua, Eric Bettinger, and Michael Kremer. 2006. "Long-Term Educational Consequences of Secondary School Vouchers: Evidence from Administrative Records in Colombia." The American Economic Review 96(3): 847-62.


Bettinger, Eric, Michael Kremer, and Juan E. Saavedra. 2010. "Are Educational Vouchers Only Redistributive?" The Economic Journal 120 (August): F204-28.