Do larger school grants improve educational attainment? Evidence from urban Mexico

ABSTRACT

We study the effects of larger cash grants on the educational attainment of low-income middle and high school students in Mexico. Starting in 2009, school grants from the Oportunidades conditional cash transfer programme increased by 27 percent for females and 30 percent for males in 263 of 551 urban localities. Using a difference-in-difference analysis of longitudinal programme registries linked to national standardised tests, we find that students with larger grants experienced lower dropout rates in middle school and were more likely to graduate from high school on time. Specifically, the likelihood of graduation increased by 38.7 percent for females and 41.3 percent for males, suggesting an elastic response to the larger grants.

KEYWORDS:

• Cash transfers
• scholarship
• grants
• school dropout
• high school graduation
• Mexico
• Progresa
• Oportunidades
• Prospera

Acknowledgments

We thank the Prospera Program for providing program administrative data used for this analysis. In particular, we thank Armando Jerónimo Cano, Angélica Castañeda, Martha Cuevas, Rogelio Grados, and José Solis for their support and valuable discussions that have informed this research. We thank Daniel Hernández, Estefanía Moleres, and Proceso Silva at the Ministry of Education (SEP) for providing access to ENLACE data. For comments we thank Orazio Attanasio, Marta Dormal, Paul Gertler, Amanda Glassman, Santiago Levy, Ferdinando Regalia, Norbert Schady, Solis Winters, and seminar participants at Prospera and the Inter-American Development Bank. Laura Dávila provided assistance with the initial assembly of the data sets and guidance over the program operation. All opinions in this paper are those of the authors and do not necessarily represent the views of the Government of Mexico, or of the Inter-American Development Bank, its Executive Directors or the governments they represent.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed here.

Notes

1. Garcia and Saavedra (2017) present the most recent meta-analysis for CCTs impacts on educational outcomes. Forty-seven CCTs impact evaluations (for 31 countries) met their quality criteria. Fifty-three percent of them were implemented in Latin America, 32 percent in Asia and 15 percent in Africa.

2. There is a body of literature analysing the ‘sheepskin effect’, whether an educational degree yields higher returns than the same amount of studying without the possession of a certificate. Hungerford and Solon (1987) and Jaeger and Page (1996) are seminal papers on this topic. Publications on sheepskin effects of a highschool diploma, for example, show heterogeneous results, yielding an average additional return on a schooling degree of 19.8%, with Brazil, where most studies on this topic have been carried out, showing greatest additional returns. In the meta-analysis of completion effect of highschool diplomas developed by Mora Rodriguez and Muro, indeed, identify a 8% sheepskin effect, suggesting thus that not only the years of education received are important for an increased wage, but also a diploma (completion certificate) (Mora Rodríguez and Muro 2015).

3. Filmer and Schady (2008) describe that the amount transferred by the Cambodian programme was very small compared to those transferred by other CCTs: 2 percent of the consumption of the median recipient household in Cambodia, while the comparable value was 22 percent for recipients of the CCT programme in Mexico. In Malawi, Baird, McIntosh, and Ozler (2011) transferred varied between $1USD and $5USD per month.

4. The Oportunidades CCT programme began as Progresa in 1997, was renamed Oportunidades in 2002, and became Prospera in 2014. Given that the policy change studied in this paper took place under Oportunidades, we use this name to refer to the programme. The programme closed down in 2019 and the school grants component was replaced with the ‘Benito Juarez’ Universal Grants for Wellbeing programme in public high schools.

5. See Dávila-Lárraga (2016) for a detailed description of the grant structure. While the relative increase in grants was slightly higher for males, the transfer levels remained higher for females. In addition to school grants, the graduation grant ‘Jovenes con Prospera’ was also increased by 29.5 percent in treatment localities.

6. Assumes exchange rate of $16 Mexican Pesos per USD as an approximate average in July 2015.

7. Educación Secundaria and Educación Media Superior, are the Spanish names for middle and high school, respectively, in Mexico.

8. The authors hypothesised that amongst the less poor, the transfer value was too low relative to their opportunity cost to comply with conditions. They found that administrative changes in the program’s operational processes that increased participation costs for the beneficiaries could translate into large numbers of families being removed from the program’s roster.

9. See Oportunidades (2009) and Dávila-Lárraga (2016) for a description of the proposed reforms under the Urban Model and their adoption by the programme. The alternative targeting scheme was tested against the existing model but was never adopted by the programme or scaled up. As such, the targeting criteria and eligibility rules remained the same in all urban communities. The other three initiatives (adjustment of health component, payments through financial institutions and service desks) were adopted and quickly scaled up to all urban communities. By 2011, for example, all urban Oportunidades beneficiaries were receiving payments through savings accounts (Masino and Niño-Zarazúa 2018).

10. Dávila-Lárraga (2016) describes the program’s geographical targeting process. The selection of Urban Model localities in the 2009 enrolment cycle was partially documented in an internal programme memo obtained by the authors, and we conducted informational interviews with former programme staff to clarify details of the selection process.

11. The 2005 inter-census survey updated population statistics in between the 2000 and 2010 national census. See https://www.inegi.org.mx/programas/ccpv/2005/ (accessed 5/28/2019).

12. See https://www.coneval.org.mx/rw/resource/coneval/med_pobreza/3967.pdf (accessed 5/28/2019) for a description of the poverty rate estimates.

13. Note that while the Urban Model officially started in the third bimester of 2009, most families with school-age children enrolled in the fifth bimester of 2009, at the start of the school year. Our analysis sample thus includes families enrolled in the later period. New beneficiaries in treatment localities were not informed of the alternative grant structure at the time of recruitment and enrolment.

14. Both cities were excluded from our analysis.

15. It seems plausible that these results are explained by differences in the opportunity cost of attending high school for men and women in both cities. Mexico’s National Occupation and Employment Survey shows that, for the period in which this pilot was implemented, both the probability of finding employment and the salary received by young women in Ecatepec were substantially lower than those received by men in Ecatepec and by women and men in Puebla. In line with this possible explanation, our study analyzes gender-differentiated impacts.

16. These faults include main beneficiary failing to collect her transfer or make any bank account transaction in two consecutive bimonthly cycles, or an existing dispute over the transfers that household members should be receiving, among others.

17. The periodic re-evaluation of socioeconomic conditions through the so called ‘recertification’ process updates the household level information to confirm whether: (a) household per capita income is below a lower threshold known as the minimum welfare (MWL); or (b) is above the MWL – although below an upper threshold known as the permanent socioeconomic verification line -, but has at least one member of the household under 12 years of age, or one who can remain a scholarship holder of the programme, or a woman under 49 years of age (Dávila-Lárraga 2016).

18. We only had access to ENLACE middle-school micro data from 2009 to 2011. ENLACE was discontinued after the 2014 school year and substituted with the PLANEA exam. As with ENLACE, the 2015 PLANEA was administered at the end of the school year (June 2016) to all students in 12th grade. As of the 2016–2017 school year, PLANEA was administered on a sample basis. See http://planea.sep.gob.mx/ms/ (accessed 5/28/2019).

19. Attrition based on CURP is not correlated with treatment status. However, some baseline characteristics of the sub-sample of students without CURP are different from the analysis sample, suggesting a potential alteration to the sample composition. Table A2.4 in the Online Appendix compares demographic and wealth indicators by CURP availability status.

20. We exclude students in four cities (Reynosa, Puebla, Juarez and Ecatepec) that piloted an alternative targeting model in 2009. Puebla and Ecatepec were also the sites of the original Urban Model experiment described in Attanasio and Espinosa (2010) and Espinosa (2014).

21. As robustness checks to sample composition, we (a) restricted the analysis to localities present in the sample both years, (b) reweight the sample based on the probability of selection into treatment using inverse probability weighting, (c) limit the analysis to localities with the nearest coverage gaps and (d) use the common support sample of a 1 to 1 match on population. Results are presented in Online Appendix A6. In all cases, results are similar to our main estimates.

22. Given that treatment is assigned at the locality level, we followed the recommendation of Abadie et al. (2017) of clustering at this level.

23. School level data are from the National Institute for Education Evaluation: https://www.inee.edu.mx/bases-de-datos-inee-2019/.

24. Official dropout rates in the control group are of a similar magnitude to dropout rates reported by the education sector in middle school nationwide for 2013 (4.5 percent for females and 6.5 percent for males) (INEE 2014).

25. There were no new enrolments between the first and fifth bimesters of 2009 or in the 6th bimester of 2008. There are no payments for school grants in the 4th bimester of each year (summer break).

Additional information

Notes on contributors

M. Caridad Araujo

M. Caridad Araujo is the Chief of the Gender and Diversity Division of the Inter-American Development Bank (IDB). Prior to this role, she was a Principal Economist in the IDB’s Social Protection and Health Division, where she worked on early child development and anti-poverty programs. She worked at the World Bank and taught in Georgetown University. She holds a Doctorate in Agricultural Economics and Natural Resources from the University of California at Berkeley.

M. Adelaida Martinez

Maria Adelaida Martinez is a Ph.D. student at the Harris School of Public Policy at The University of Chicago. She received her B.A. and M.A. in Economics from Universidad de los Andes-Colombia, where she also worked as a researcher. Before starting her Ph.D., Ma. Adelaida worked as a Research Fellow at the Inter-American Development Bank (IADB), focusing on the social sector. Whilst at the IADB, she collaborated on projects related to education and early child development in Latin America. Her research focuses on development economics, education and migration.

Sebastian Martinez

Sebastian Martinez is the Director of Evaluation at the International Initiative for Impact Evaluation (3ie) and a Principal Economist (on leave) in the Office of Strategic Planning and Development Effectiveness at the Inter-American Development Bank (IDB). His work focuses on strengthening the evidence base and development effectiveness of the social and infrastructure sectors, including health, social protection, labor markets, water and sanitation and housing and urban development.  Sebastian holds a Ph.D. in Economics from the University of California at Berkeley, with a specialization in development and applied microeconomics. 

Michelle Perez

Michelle Perez is part of the Global Scale for Early Development (GSED) team at the World Health Organization (WHO). Prior to this role, she was a consultant at the Social Protection and Health Division, as well as the Office of Strategic Planning and Development Effectiveness at the Inter-American Development Bank (IDB). Her work focuses on project design, coordination, and evaluation of anti-poverty programs, particularly in terms of social protection, health, and early childhood development. Michelle holds a M.A. in Development Economics from American University.

Mario Sanchez

Mario Sanchez is a Principal Specialist in the Social Protection and Health Division of the Inter-American Development Bank (IDB). Prior to joining the IDB, he was a research fellow at the Center for Population Economics at the University of Chicago.  He holds master’s and Ph.D. degrees from the University of Chicago.